American Th racic SocieMEDIcAL sEcnoN oF THE AMERICAN LUNG Rss6clRnoN

standards for the Diagnosis and care of patients withChronic Obstructive pfrlmonary/ DiseaseTXIS OTTICINL STATEMENT OF THE AVTRICNru TNONNCIC SOcIew wAS APPRoVED BY THE AT5 Bonno or DIRecroRs, MARCH 1995

CONTENTS

Definitions, Epidemiology, Pathophysiology, Diagnosis, andStaging

Definit ionsEpidemiology of COPDRisk Factors for COPDThe Natural History of COPDPathology and Structural EffectsClinical Features of COPDLaboratory Findings and Diagnostic TestsPrognosisStaging: Prospects and Proposal

Comprehensive Outpatient Management of COpDSmoking CessationPharmacologic TherapyLong-term Oxygen TherapyPulmonary Rehabil itation

Inpatient Management of COPDEmergency EvaluationHospitalization CriteriaPharmacotherapy of COPD ExacerbationsMobilization of SecretionsAssisted VentilationInpatient Oxygen Therapy

Surgery and the COPD PatientPreoperative EvaluarionRisk of Surgical Procedures in COPD PatientsSummary of ApproachThe Future of Surgery for COPD Patients

Additional ConsiderationsSleep and the COPD PatientThe Role of Nutrit ionCOPD and Air Tiavel

Ethical Issues in COPD

This Statement was supported by an educat ional grant f rom Boehr inger In-gelheim Pharmaceut icals.

Commit teei Barto lome R. Cel l i , M.D. (Chairman); Cordon L Snider. M.D.;john Heffner, M.D., Br ian Tiep, M.D.; t rwin Ziment" M.D.; garry Make, M.D.;Sidney Braman, M.D.; Cerald Olsen, M.D.; yancy phi l l ips, M.D.

Am, Respir Cr i t Care Med Vol 152. pp S77_5t2O, 1995

Definitions, Epidemiologn Pathophysiology,Diagnosis, and StagingDEFIN lT IONS

Chronic obstructive pulmonary disease (COPD) is defined asa disease state characterized by the presence of airflow obstruc-tion due to chronic bronchitis or emphysema; the airf low ob-struction is generally progressive, may be accompanied by air-way hyperreactivity, and may be partially reversible.

In the past, asthma - viewed as a condition in which increasedresponsiveness of the tracheobronchial tree was the most promi-nent feature-was generally subsumed under COPD. Now. in-flammation with participation of complex cellular and chemi-cal mediators is considered the salient characteristic of asthma.It therefore seems prudent and practical to separate these condi-tions. That has been done in this statement, bearing in mind thatthe obstruction in many parients with COPD may include a sig-nificant reversible componenr and that some patients with asthmamay go on to develop irreversible airf low obstruction indistin-guishable from COPD.

' Chronic bronchilis is defined as rhe presence of chronic produc-tive cough for 3 mo in each of two successive years in a patientin whom other causes of chronic cough have been excluded (l).

Emphysema is defined as abnormal permanent enlargementof the airspaces distal to the terminal bronchioles, accompaniedby destruction of their walls and without obvious fibrosis. De-struction is defined as lack of uniformity in the pattern of respi-ratory airspace enlargement; the orderly appearance of the aci-nus and its components is disturbed and may be lost (2).

Note that chronic bronchitis is defined in clinical terms andemphysema in terms of anatomic pathology (see Figure l, a non-proportional Venn diagram showing the relationships amongchronic bronchitis, emphysema, asthma, and airflow obstruction).

EP IDEMIOLOGY OF COPD

Knowledge of the prevalence of COPD is incomplete. It is esti-mated that approximately l4 mill ion persons in the United Statessuffer from COPD - abour 12.5 mill ion from chronic bronchitisand about 1.65 mill ion from emphysema. The estimated num-ber of those with COPD has increased 41.590 since 1982. Esti-mates of diagnosed emphysema or chronic airf low obstructionin populat ion-based studies in the Unired Srares range f rom 4 to6Vo of adult white males and from I to 3go of adult white fe-males (3).

Age-adjusted prevalence rates for men rose only slightly overthe per iod f rom 1979-1985, wi th a prevalence of l l0 per 1,000in 1985. Among women, however, prevalence rates increased byover 3090 during that period, with a prevalence of l19 per 1,000in 1985 (4) .

Mortality

ln 1991, there were 85,544 deaths due to COpD and all ied condi-tions, a death rate of 18.6 per 100,000 persons; this categoryranked as the fourth leading cause of death. The mortality raterose nearly 32.990 between 1979 and 1991. In 1985. COpD wasthe underlying cause f or 3.60/o of all deaths and was a contribu-tory factor in an additional 4.3V0 oi deaths. Men and womenhave similar COPD mortality rates before the age of 55, but therate for men rises thereafter; at age70, the rate for men is morethan double that for women' and at 85 and older, the COpDdeath rate for males is more than 3.5 times that for females (4).

The age-adjusted death rare for COPD rose Tlgo between 1966and 1986; over those same two decades, the death rate from all

causes declinedby 220/0, and the rates for heart and cerebral vas-cular disease declined by 45 and 5890, respecrively (3). Observedincreases in mortality and morbidity appear to be related to pasttrends in cigarette smoking. Part of the rise in morbidity andmortality may also be due to increasing numbers of people l iy-ing longer; the increases in morbidity and mortality are particu-larly striking among older people who continue to smoke. Sincesmoking frequency has fallen over the past 30 yr, there shouldbe a decrease in COPD mortality in coming decades (5).

AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE VOL I52 1995

CHRONICBRONCHITIS

T _

EMPHYSEMA

COPD

iK_

ro IA|RFLOW;OBSTRUCTIONI

I

l

ASTHMA

Figure l. Schema of chronic obstructive pulmonary disease. This non-proport ional Venn diagram shows subsets of patients with chronicbronchit is, emphysema, and asthma. The subsets comprising COpDare shaded. Subset areas are not proport ional to actual relat ive sub-set sizes. Asthma is by definit ion associated with reversible airf lowobstruction, although in variant asthma special maneuvers mav benecessary to make the obstruction evident. patients with asthma whoseairf low obstruction is completely reversible (subset 9) are not con-sidered to have COPD. Because in many cases i t is virtual ly impossi-ble to dif ferentiate patients with asthma whose airf low o6structiondoes not remit completely from persons with chronic bronchit is andemphysema who have part ial ly reversible airf low obstruction witharrway hyperreactivity (55), patients with unremifting asthma are clas-s i f ied as hav ing COPD (subsets 6 ,7 , and g) . Chron ic b ronch i t i s andemphysema with airf low obstruction usually occur together (26) (sub-set 5), and some patients may have asthma associated with thesetwo disorders (subset 8). Individuals with asthma exposed to chronicirr i tat ion, as from cigarette smoke, may develop chronic productivecough, a feature of chronic bronchit is (subset 6). Such patients areoften referred to in the united states as having osthmotic bronchit isor the osthmotic form of CO?D. persons with chronic bronchit is (56)andlor emphysema (57) without airf low obstruction (subsets l . 2,and I ' l ) are not classif ied as having COpD. patients with airway ob_struction due to diseases with known etiology or specif ic pathoiogy,such as cystic f ibrosis or obl i terat ive bronchiol i t is (subset 10), are r iotinc luded in th is de f in i t ion .

RISK FACTORS FOR COPD

The primary cause of COPD is without question exposure rotobacco smoke.

American Thoracic Society

Cigarette Smoking

The major risk factor is cigarette smoking. Smokers have higherdeath rates for chronic bronchitis and emphysema; they also havea higher prevalence of lung-function abnormalit ies, respiratorysymptoms, and all forms of chronic obstructive airway disease.Cigarette smokers also have a greater annual rate of decline inFEV,. Differences between cigarette smokers and nonsmokersincrease in direct proportion ro quantity of smoking. (Pipe andcigar smokers have higher morbidit l, and mortality rares forCOPD than nonsmokers, although their rates are lower than thosefor cigarette smokers.) Age of starring, total pack-years, and cur-rent smoking status are predictive of COPD mortaliry. For un-known reasons, presumably related ro consrirutional differences,only about 1590 of cigarette smokers develop clinically signifl-cant coPD (3, 6, 7) .

Overall, tobacco smoking accounrs for an estimated 80 to 9090of the risk of developing COPD (6). The smoking-attriburablefraction of COPD mortaliry in the United States during the 1980swas 0.850 for men and 0.694 for women (8). The only other riskfactor of comparable importance for rhe individual is homozy-gous alpha,-antitrypsin (AAT) deficiency, bur thar heritable con-dition accounts for less than l9o of COPD cases (see further com-ment below).

Passive Smoking

Also known as environmental tobacco smoke (ETS) or,,second-hand smoke," passive smoking is the exposure of nonsmokersto cigarette smoke. Children whose parents smoke have a higherprevalence of respiratory symptoms and respiratory disease andappear to have small but measurable deficiencies in rests of pul-monary function when compared with children of nonsmokers.These deficiencies may presage airway hyperreacrivity and lessthan maximal attainment of lung funcrion in adulr l i fe, alrhoughthe significance of these findings in relarion ro rhe furure devel-opment of COPD is unclear. Despite lhese uncerrainties, chil-dren should be protecred from exposure ro tobacco smoke (9).

Ambient Ai r Pol lu t ion

High levels of urban air pollurion are demonsrrably harmful topersons with hearr or lung disease, but the role of environmen-tal air pollution in the eriolog)' of COPD in rhis counrrf is un-clear; its role appears to be small when compared rvith that ofcigarette smoking. Reported respirarorl. s! 'mproms, bur nor lungfunction deficiencies, have been associared with indoor nirro-gen dioxide levels and damp housing. The use of various solidfuels for cooking and heating withour adequare venrilarion, how-ever, may result in high levels of indoor air pollurion and ac-count for the developmenr of COPD (9).

Hyperrespons ive Ai rways

Asthma, atopy, and nonspecific airwal' hy,perresponsiveness ma),possibly play a role in COPD. In 1960, Orie and colleagues (10)from the Netherlands proposed rhan an "asrhmaric consti lution"(a predisposition to aropic disease, airual' hvperresponsiveness,and eosinophil ia) underlal ' the developmenr of chronic airf lou,obstruction; smoking, the)'suggested, u.as onll one extrinsic fac-tor that, superimposed on this consrilurional suscepribil i ty', couldlead to chronic airf low obsrrucrion.

Many studies have since focused on "rhe Durch hypothesis."It has become evident thar skin resr reacrivir).ro allergens, elevaredserum IgE, and eosinophil ia, u hile ali markers of aropl', are ap-parently not interchan_eeable and have differenr relarions ro clin-ical manifestations such as aslhma and ha1 fever. ln contrastto asthma, nei ther the d iagnoses of chronic bronchi r is or em-physema nor the presence of venrilarorl impairment in the ab-sence of asthma is related to ase-se.\ srandardized serum leE

levels. Smokers iend, in fact, to be less atopic than nonsmokersbut to have h igher serum levels of IgE ( l i ) .

. The possibil i ty thar nonspecific airway hyperreactivity mightbe a risk facror for COPD was first raised ai i part oi tiri Ouicnhypothesis. Such hyperreacdvity is inversely related to FEV,, andevidence is steadily accumulating that it is predictive of an ac_celerated rare of decline of lung function in smokers. But its pos_sible role as a risk factor that may predispose to the deveiop_ment of COPD in smokers or others is unclear. Nonspecifichyperreactivity might also stem from the airway inflammationtypically seen with the development of smoking-related chronicairflow obsrruction (12-14).In the Lung Health Siudy, nonspecificairway hyperreactiviry was noted in a significantly highir per-centage of women (85.190) than men (58.990). Moreover, neirlytwice as many women as men responded to < 5 m/ml ofmethacholine (46.6 and 23.9V0, respectively). In both sexes, de-gree of airf low obstruction was highly correlated with severitvof airf low obstruction but not with age (15).

Sex, Race, and Socioeconomic StatusEv.en conrrolling for smoking, there is a higher prevalence of re-spiratory symptoms in men (7). Mortality rates for COpD arehigher in whites than in nonwhites, but the difference is decreas-ing in males (3). Morbidity and mortality rates are inversely relatedto socioeconomic status and are higher in blue-collar than white-collar workers (3, 6, 7, l6). Apart from homozygous alpha,-anti-trypsin deficiency, COPD may aggregate in fimilies-(17).

Occupational Factors

Occupational factors give rise to increased prevalence of chronic1l!o* obsrruction, increased rates of FEV, decline, and higherCOPD mortality. Interaction between cigarette smoking and-jobexposure to hazardous airborne substances results in higher iatesofCOPD. However, smoking effects greatly exceed occupationaleffecrs (18).

Alphar -antitrypsin Deficiency

,A.lpha,-antitrypsin (AAI) is the only known genetic abnormal-ity that leads to COPD; AAT deficiency accounts for less thanl9o of COPD in the United States. Also known as alpha,-proteaseinhibitor, AAT is a serum protein produced by the liver ind nor-mally found in the lungs; its main role is the inhibit ion of neu-trophil elasrase (19). It is a glycoprotein, coded for by a singlegene on chromosome 14. The serum protease inhibitor pheno_type (Pi type) is determined by the independenr expression ofthe two parenral alleles. The AAT gene is highly pliomorphic.

Some 75 alleles are known, and the states they produce havebeen classified, as: normal, associated with normal serum levelsof normally functioning ANf; deficienl, associated with serumAAT levels lower than normal; nlld associated with undetect-able serum AAT; and dysfunctional in which AAT is presentat normal levels but does not function normally (20). The vari-ants of AAT occur because of point mutations that result in asingle amino acid substitution. The normal M alleles occur inabout 9090 of persons of European descent with normal serumAAI levels; their phenotype is designared pi MM. Normal valuesof serum AAT are 150 to 350 mgldl (commercial standard) or20 to 48 pM (true laboratory standard).

More than 9590 of persons in the severely deficient categoryare homozygous for the Z allele, designated pi ZZ, and havi se-rum AAI leveis of 2.5 to j pM (mean, l6go of normat). Mostof these persons are Caucasians of northern European descent,because the Z allele is rare in Orientals and blacks. Rarely ob-served phenotypes that are associated with low levels of serumAAT include Pi SZ and persons with nonexpressing alleles, pi-null. The latrer occur in homozygous form, pi null-null, or in

AMERICAN IOURNAT OF RESPIRATORY AND CRITICAL CARE MEDICINE VOL I52 I995

heterozygous form with a deficient aliele, Pi ZZ-null. Personswith phenotype Pi SS have AAf values ranging from l5 to 33pM (mean, 52Vo of normal). The threshold protective level ofl l pM or 80 mgldl (3590 of normal) is based on the knowledgethat Pi SZ heterozygotes, with serum AAT values of 8 to 19 pM(mean, 37Vo of normal), rarely develop emphysema.

Severe AAT deficiency leads to premature emphysema, oftenwith chronic bronchitis and occasionally with bronchiectasis. Theonset of pulmonary disease is accelerated by smoking: dyspneabegins at a median age of 40 yr in smokers, and a median ageof 53 yr in nonsmokers. Panacinar emphysema usually beginsat the bases. The severity of lung disease varies markedly. Per-sons classified as nonindex cases (those discovered in popula-tion surveys) tend to have better lung function, whether the;-smoke or not, than patients classified as index cases (those dis-covered because they have lung disease). People identif ied asnonindex cases may live into their eighth or ninth decade. Air-f low obstruction occurs more frequently in men; asthma, recur-rent respiratory infections, and familial factors are also concom-itant risk factors for airf low obstruction (21).

The diagnosis of AAI deficiency is made b1' measuring theserum AAT level, followed by Pi typing for confirmation. Thecircumstances in which the tests should be ordered are l isted inThble l.

Pi MZ heterozygotes have serum AAT levels that are inter-mediate between Pi MM normals and Pi ZZ homozygotes (12to 35 pM; mean, 5790 of normal) . In populat ion studies, theydo not appear to be at increased risk for COPD; in family studiesand in surveys of some populations of COPD patients, however,there has been an increased frequency of heterozl"-eotes (22). Thereis some evidence that MS heterozygotes ma.v have an increasedfrequency of nonspecific airway hyperreactivit l ' (23).

THE NATURAL HISTORY OF COPD

The FEV, in nonsmokers without respiratorl ' disease declincsby 25 to 30 ml per year beginning at about age 35 (see Figure2). The rate of decline of FEV, is steeper for smokers rhan fornonsmokers, and the heavier the smoking the sleeper rhe rare.The decline in function occurs along a slo*'11'acceleraring curvi-l inear path. In most persons, the loss occurs uniforml)'; in some.it develops in stages, with relatively steep declines. There is a di-rect relationship between init ial FEV, level and rhe slope of FE\',decline (24). There is also a somewhat stronger associarion be-tween a low FEV,/FVC and subsequent decline in FE\', in menbut not in women (25). Age, which cannor be separared fromthe number of years of cigarette smoking, is clearlv a risk facrorfor more rapid decline of lung function; so are l iferime smokinghistory and the number of cigaretres currentl l ' smoked (26).

Indiv iduals wi th COPD have more f requenr l l 'acure chesr i l l -nesses, which generally decrease lung funcrion for abour 90 d(24) . The ro le of mucus hypersecret ion is unclear . Ear l ier srudies

TABLE 'I

SCREENINGFOR ALPHAT-ANT ITRYPSIN DEF IC IENCY: tND tCAT IONS

. Chron ic b ronch i t i s w i th a i r l low obs t ruc taon in a n€ver -5more ,

. Bronch iec tas is , espec ia l l y in the absence o l c lear r rsk lac rors to -the d isease

' Premature onset o f COPD, w i th modera te o r severe rmoa: rmen: .by o r be fore age 50

. A predominance o l bas i la r emohvsema

. Deve lopment o f unremi t t ing as thma, espeoa l lv rn a person unCe, aOe 5{(screentng is ind ica ted even in the presence o f a topr ,

' A f a m i l y h i s t o r y o f a l p h a t - a n t r t r y p s i n d e l i c i e n ( v o r o i C O p 3 c ^ r e :be lo re age 50

. Cr r rhos is w i thout apparent r i sk lac to rs

A G E _ Y E A R S

Figure 2. Relationship of FEVr, age, and smoking. Nonsmokers loseFEVr at an accelerating rate with age; the average loss is about 30ml/yr. Smokers of 30 cigarettes per day average a slightly greater rateof decline and have FEVr values slightly below average when firststudied at age 40 yr. A small proportion of susceptible smokers (about15%) lose function much more rapidly, approximately 150 ml/yr, withFEVr of 0.8 t by age 55, a level so low that they experience dyspneain the course of ordinary daily l iving. Susceptible smokers who stopsmoking at age 50 do not regain lost function or regain only a l itt le,but they subsequently lose function at the same rate as never-smokers;dyspnea with ordinary activity wil l not develop unti l the mid-20s.Reprinted with permission from Snider, G. 1., L. l. Faling, and S. l.Rennard. 1994. Chronic bronchi t is and emphysema. ln l . F. Murrayand J. A. Nadel, editors. Textbook of Respiratory Medicine. W. B. Saun-ders, Phi ladelphia. 1342.

shou,ed thar ir was not associated with mortality from COpD(27), but recenr studies have shown a weak correlation (28,29).

The small airways, those less than 2 mm in diameter, are im-portant sires of airf low obstruction. The relative contributionof peripheral airway disease and loss of elastic recoil from em-phl,sema may vary. Indices such as the closing capaciry and theslope of the alveolar plateau derived from a single-breath nitro-_een test are unable ro identify individuals susceptible to chronicairrlav obsrruction with cigarette smoke exposure (30). Testsreflecting emphysema (single-breath diffusing capacity, FRC, to-tal lung capacity) predict survival in a relatively minor way (31).

After cessation of smoking, a small amount of lung functionis regained. In the Lung Health Study placebo group, for exam-ple, the 3590 of subjecrs who had stopped smoking ar rhe firstannual visir showed an increase in mean postbronchodilator FEV,of 57 ml, rvhile placebo parricipanrs who had not stopped smok-ing showed a mean decline in postbronchodilator FEV, of 3gm l (32 ) .

Thereafrer, the rate of lung function decline slows to approx-imate that seen in never-smokers of the same age (33, 34). Datafrom several studies suggest that the susceptible subgroup ofsmokers having more rapid decline of lung function can be de-fined by their grearer loss of FEV, or FEV,/FVC ratio (25, 35).Smoking cessarion improves prognosis regardless of age (36).

aEt!F 3

1, , 2II

4 5 5 53 5 6 5 7 5

\ . -31S--"i'*. . ^-

':'o{'

O Y S P N E A - A C T I V I T I E S O F O A I L Y L T V I N G

American Thoracic Society

PATHOLOGY AND STRUCTURAL EFFECTS

ln COPD, there is enlargemenr of bronchial mucous glands, withdilation of gland ducts. Gobler cell frequency is increased. Focalsquamous metaplasia and hypertrophy of airway smooth mus_cle may be present. The mucous gland enlargement of chronicbronchitis is nonspecific; similar changes occur in other diseases,such as cystic f ibrosis (37).

The respiratory bronchioles display a predominantly mono-nuclear inflammarory process (38). Membranous bronchioles lessthan 2 mm in d iameter show vary ing degrees of p lugging wi thmucus, goblet-cell metaplasia, inflammation. increased smoothmuscle, and distorrion due to fibrosis. These changes and theloss of alveolar attachments from the destructive process of em-physema cause loss of cross-sectional area (39. 40).

Three types of emphysema can be distinguished (41). The firsttype, centriacinar emphysema, begins in the respiratory bron-chioles and spreads peripherally. Centrilobular emphysema is theform of centriacinar emphysema associated with longstandingcigarette smoking. This form predominantly involves the upperhalf of the lungs. Focal emphysema is the form of cenrriacinaremphysema that occurs in coal workers' pneumoconiosis.

The second type is panacinar emphysenra, which involves theentire alveolus uniformly and predominates in the lower half ofthe lungs; this is the type of emphysema generally seen with homo-zygous AAI deficiencl'. Focal panacinar emphysema ar the lungbases may accompany centrilobular emphysema in smokers (42).

The third type, distal acinar emphysema, also known as para-septal emp hysema, preferentially involves distal airway structures,alveolar ducts, and sacs. The process is localized adjacent tofibrous septa or to the pleura. Apical bullae may cause sponra-neous pneumothorax; occasionally, giant bullae cause severe com_pression of relatively uninvolved lune. With this form of em_physema, airf low is frequently well f,reserved.

Airspace enlargement with puimonary fibrosis is commonlyseen as an inconsequential lesion adjacent to scars, but it maysometimes be severe with extensive fibrosing disease, such as sar-coidosis or tuberculosis. The underlying fibrosis is usually evi-dent radiographically, wirh extensive l inear or nodular shadowsaccompanying increased transradiancy or bullae.

The structural basis of airf low obstrucrion in COpD (43-45)may be summarized as follows: alterarions in the glands of thecentral airways have litt le effect on spirometrv. Alterarion of rhesmall airways is a major cause of airf lou, otsrrucrion. Mono-nuclear cell inflammation in rhe respiratory bronchioles is theearliest lesion in young smokers. Inflammarion, f ibrosis, goblercell metaplasia, and smooth muscle hypertrophl,in terminal bron-chioles are important causes of airf lolr ' obsrrucrion; loss of alve-olar attachments to bronchioles due to rhe destructive changesin emphysema is also a major cause.

Airflow obstruction in COpD cannol be explained enrirel),on a struclural basis; bronchoconstriction is another mechanism.A significant increase in FEV, afrer an inhaled bera_adrenergicagonist has been observed in up ro one th i rd of COpD par ienrsdur ing s ingle test ing sessions and in up ro two rh i rds dur ing ser ia ltesting (a6). The response varies greatl l 'berween resrs and corre-lates only weakly with parient fearures such as demographics.smoking h is tory, symptoms, or measuremenrs of lung funct ion.

In summary, a i r f low obstruct ion in COpD is pr imar i ly . i r re-versible and is caused by disease of rhe small airu.a1,s. lr.hich isdue in par t ro the ef fecrs of in f lammar ion in rhose a i ru.a1.s andin par t to the loss of a lveolar sepral rerher ing of smal l a i r*avsthat accompanies the destruct ive changes o i emphl sema. Bron-choconstriction due to inflammarion accouni_( to: a l imired amounlof revers ib le a i r f low obsrrucr ion.

CLIN]CAL FEATURES OF COPDHistory

Patients with COPD have usually been smoking ar least 20 ciga_rettes per day for 20 or more years before symptoms develop.They commonly presenr in the fifth decade with productiu..oughor an acute chest i l lness. Dyspnea on effort usually does not oc_cur unti l the sixth or seventh decade.

Sputum producrion is insidious, init ially occurring only in themorning; the daily volume rarely exceeds 60 ml. Sputum is usu-ally mucoid but becomes purulent with an exacerbation. Acutechest i l lnesses characterized by increased cough, purulent. spu-tum, wheezing, dyspnea, and occasionally fever may occur in_termittently. The history of wheezing and dyspnea may lead toan erroneous diagnosis of asthma.

With disease progression, the intervals between acute exacer-bations grow shorter. Late in the course of the disease, an ex-acerbation may give rise to hypoxemia with cyanosis, the latteraccentuated by erythrocyrosis. Morning headache suggests hyper_capnia; hypercapnia, with more severe hypoxemia, is often pres_ent in end-stage disease. Weight loss occurs in some patients. Corpulmonale with right heart failure and edema may develop inpatienrs with hypoxemia and hypercapnia,

. Since bronchogenic carcinoma occurs with increased frequencyin smokers with COPD, an episode of hemoptysis raises thl poj-sibil i ty that carcinoma has developed. Most episodes of hemop-tysis, however, are due to mucosal erosion and not to carcinoma.

Physical Examination

Early on, examination of the chest may reveal only slowed expi-ration and wheezing on forced expiration. As obstruction pro_gresses, hyperinflation becomes evident, and the anteroposteriordiameter of the chest increases. The diaphragm becomes limitedin its motion. Breath sounds are decreased, expiration is pro_longed, and heart sounds often become distant. boarse crac'klesmay be heard at the lung bases. Wheezes are frequently heard,especially on forced expiration.

Patients with end-stage COpD may adopt positions rhar re-l ieve dyspnea, such as leaning forward with arms outstretchedand weight supported on the palms. The accessory respiratorymuscles of the neck and shoulder girdle are in full use. Expira-tion often takes place through pursed lips. paradoxical indraw-ing of the lower interspaces is often evident. Cyanosis may bepresenr. An enlarged, tender l iver indicates heart failure; neckvein disrenrion, especially during expiration, may be observedin the absence of heart failure, due to increased intrathoracicpressure. Asterixis may be seen with severe hypercapnia.

LABORATORY FINDINGS AND DIACNOSTIC TESTSChest Radiography

Because emphysema is defined in anatomical rerms, radiographicimages of the lungs provide the clearesr evidence of irs prisence.ln frontal and lateral chesr radiographs, distention of ihe lungsis indicated by a low, flat diaphragm, an increased retrosrernalairspace, and a long, narrow heart shadow. Rapid tapering ofthe vascular shadows, accompanied by hypertraniradiancy oithelungs, is a sign of emphysema; bullae, presenring as radiolucenlareas larger than I cm in diamerer and surrounded by arcuatehairl ine shadows, are proof of its presence. Bullae, however, re-flect only locally severe disease and are not necessarily indica-tive of widespread emphysema. With complicaring pulmonaryhypertension and right veritricular hypertrophy, the hilar vascu-lar shadows are prominent, and the heart shadow encroaches onthe retrosternal space as the righr ventricle enlarges anteriorly

AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE VOL I52 I995

(47). The eniargement may become evident only on comparisonwith previous chest radiographs.

Studies correlating lung structure and the chest radiographshow that emphysema is consistently diagnosed when the dis-ease is severe, is not diagnosed when the disease is mild, and isdiagnosed in about half the instances of moderate disease (47).

Computed Tomography

Computed tomography (CT), especially high resolution CT (col-l imation of 1-2 mm), has much greater sensitivity and specific-ity than standard chest radiography (48). It may even identifythe specific anatomic type of emphysema. Because this infor-mation rarely alters therapy, CT has no place in the routine careof patients with COPD. It is, however, the main imaging toolto predict the benefit of pulmonary resection for giant bullousdisease and for diagnosing complicating bronchiectasis.

Pulmonary Function Measurements

These measurements are necessary for diagnosis and assessmentof the severity of disease and are helpful in following its prog-ress (49). Airf low obstruction is an important indicator of im-pairment of the whole person and of the l ikelihood of blood gasabnormalit ies. The FEV' is easily measurable, has less variabil-ity than other measurements of airway dynamics, and is moreaccurately predictable from age, sex, and height. Roughly com-parable information can be obtained from the peak flow mea-surement or from the forced expiratory flow-volume curve. Noneof these tests can distinguish between chronic bronchitis and em-physema.

Lung voiume measurements show an increase in total lungcapacity, functional residual capacity, and residual volume. Thevital capacity may be decreased. The single-breath carbon mon-oxide diffusing capacity is decreased in proportion to the severityof emphysema because of the loss of alveolar capil lary bed. Thetest is not specific, nor can it detect mild emphysema. Arterialblood gases reveal mild or moderate hypoxemia without hyper-capnia in the early stages. As the disease progresses, hypoxemiabecomes more severe and hypercapnia supervenes. Hypercapniais observed with increasing frequency as the FEV, falls belowI L. Blood gas abnormalit ies worsen during acute exacerbationsand may also worsen during exercise and sleep.

As noted earlier, up to 3090 of patients have an increase inl59o or more in FEV, after inhalation of a beta-agonist aerosol.The absence of a bronchodilator response during a single testnever justif ies withholding bronchodilator therapy.

Erythrocytosis is infrequently observed in patients l iving atsea level who have Pa6, levels of ) 55 mm Hg; the frequencyof erythrocytosis increases as Pae, levels fall below 55 mm Hg.

Sputum Examination

ln stable chronic bronchitis, sputum is mucoid, and the predom-inant cell is the macrophage. With an exacerbation, sputum usu-ally becomes purulent, with an influx of neutrophils. Cram's stainusually shows a mixture of organisms. The most frequent patho-gens cultured from the sputum are Slreptococcus pneumoniaeand Haemophilus influenzae. Other oropharyngeal flora, suchas Moraxella catarrhqlis, have been shown to cause exacerba-tions. ln the outpatient setting, however, cultures or even Gram'sstains are rarely necessary before instituting antimicrobial ther-apy (50). The approach to diagnosis of COPD is summarizedin Thble 2.

PROCNOSIS

Predictors of morta l i ty in pat ients u i th COPD are advancingage, severity of airf low obstruction, as indicated b1 FEV,, severity

TABLE 2

DIAGNOSIS OF COPD

History. Smoking: age at in i t iat ion, quant i ty smoked per day, whether or not 5t i l l

smoker ( i t not , date of cessat ion). Environmental (chronological) : may disc lose important r isk lactors. Cough (chronic, product ive) : f requency and durat ion, whether or not

product ive (especia l ly on awakening), presence or absenc€ of b lood. Wheezing. Acute chest i l lnesses: l requency, product iv€ cough, wheezing, dyspnea,

feve r. Dyspnea

Physical Examinat ion. Chest

Air l low obstruct ion evidenced by-Wheezes dur ing auscul tat ion on s low or forced breathinq-Prolongat ion of forced expiratory t ime

Severe emphysema indicated by-Overdistent ion of lungs in stable state, low diaphragmat ic posi t ion-Decreased intensi ty of breath and heart sounds

Severe disease suggested by (characteristic, not diagnostic)-Pursed- l ip breathing-Use ot accessory respiratory muscles- lndrawing of lower interspaces

. Other: unusual posi t ions to re l ieve dyspnea at rest , d ig i ta l c lubbing(suggests possib i l i ty o l lung cancer or bronchiectasis) , mi ld dependentedema (may be seen in absence of r ight heart fa i lure)

Laboratory. Chest radiography: diagnostic only of severe emphysema but essential

to exclude other lung diseases. Spirometry (pre- and post-bronchodilator): essential to confirm presence

and revers ib i l i ty of a i r f low obstruct ion and to quant i fy maximumlevel of vent i latory Iunct ion

. Lung volumes: measurement of more than forced vital capacity notnecessary except in specia l instances (e.9. , presence of g iant bul lae)

o Carbon monoxide ditfusing capacity: not necessary except in specialinstances (e.9., dyspnea out ol proportion to severity of airflow limitation)

. Arterial blood gases: not needed in stage I airflow obstruction (FEV1 )50% predicted) essent ia l in stages l l and l l l a i r f low obstruct ion (FEV1 <50% predicted); in very severe airflow obstruction, maior monitoring tool

ofhypoxemia, and the presence ofhypercapnia. In persons withmoderate airway obstruction, but with an FEVr ) 1.0 L, thereis a slight excess mortality at l0 yr in comparison with an age-and gender-matched population. Recent data suggest that markedreversibility of airway obstruction is a favorable prognostic fac-tor (26).

In persons with FEV, values ( 0.75 L, the approximate mor-tality rate at I yr is 3090 and at l0 yr 950/o (51). However, lon-gitudinal studies have shown that some patients with severe air-way obstruction may survive for many years beyond the average,some for as long as l5 yr. The reason appears to be that deathin COPD generally occurs as a resulr of a medical complicarion,such as acute respiratory failure, severe pneumonia, pneumotho-rax, cardiac arrhythmia, or pulmonary embolism.

STAGING: PROSPECTS AND PROPOSAL

The approach to COPD would be grearly facilitated by a stagingsystem, which would allow standardized categorization of theheterogeneous population of patients with this common disor-der. Such a system would be useful for epidemiologic and clini-cal studies, health resource planning, prognostication, and theapplication of clinical recommendations such as those in thisdocument.

Ideally, a staging system would offer a composite picture ofdisease severity based on the interrelationship of the sensationof breathlessness, impairment in airf low, and derangement ingas exchange (52, 53). These factors and their magnitude are in-teractive but not necessarily additive. At present, there are no

American Thoracic Society

data defining their interrelationship in a quantitative manner;the best correlate with mortality and morbidity is decrease inFEV,. Consequently, COPD severity may be staged on the basisof the degree of airflow obstruction, using.the criteria of the AISstatement on interpretation of lung function (54): stage I is FEV'2 500/o predicted; stage II is FEV, 35 to 490/o predicted; and stageIII is FEV' 135s/o predicted. Patients with FEV, >- 500/o do notusually have severe hypoxemia, and.arterial blood gas measure-ments are not required. Patients in stages II and III should havearterial blood gas measurements breathing air, and the oxygenand carbon dioxide tensions should be stated.

Stage I COPD comprises the great majority of patients. Inthese patients, COPD has minimal impact on health-related qual-ity of life and results in only modest per capita health care ex-

penditure. Patients in stage I will usually be cared for on a con-tinuing basis by a generalist. The presence of severe dyspnea ina patient with stage I COPD warrants additional studies andevaluation by a respiratory specialist.

Stage II COPD includes a minority of patients. In these pa-tients, COPD has significant impact on health-related qualityof life and results in large per capita health care expenditure Pa-tients with stage lI COPD usually merit evaluation by a respira-tory specialist and may receive continuing care by such a specialist.

Stage III COPD also includes a minority of patients. In thesepatients, COPD has profound impact on health-related qualityof life and results in large per capita health care expenditure. Pa-tients with stage III COPD will usually be under the care of arespiratory specialist.

AMERICAN IOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE VOL I52 I995

comprehensive outpatient Management of GopDOnce the diagnosis of COPD is established, the patient shouldbe educated about the disease and should be encouraged to par-ticipate actively in therapy and, especially, in preventive care (e.g.,immunizations, including pneumococcal and annual influeniavaccines) and maintenance of an active lifestyle. Above all, a pa-tient who sti l l smokes must be encouraged and supported in aneffort to quit.

An overall algorithmic approach to management is illustratedin Figure 3, which forms a framework for the detailed guidelinesin this section.

SMOKING CESSATION

Reported data are not encouraging. Continuous abstinence, evenin pulmonary patients after counseling, may be as low as270/oin follow-up periods ranging from 6 mo to 7 yr (58). Factors foster-ing smoking continuation, which may vary among patients, in-clude the additive potential of nicotine, conditioned responseto smoking-associated stimuli (e.9., work and social situations),dnd psychosocial problems, such as depression, poor education,low income, and forceful promotional campaigns by the tobaccoindustry.

Just as the causes of smoking init iarion and continuation aremultifactorial, successful solutions most often involve multipleinterventions (59-62). Elements of successful smoking cessationprograms, which are summarized in Table 3, are discussed in de-tail in the following paragraphs,

Yes

Physician lntervention

The clinician should always express strong and continued interestin smoking cessation for the patient with COPD. physician coun-seling frequently makes the difference between successful andunsuccessful efforts to quit. As a first step, the physician andpatient should identify the patient's stage of readiness for smok_ing cessation.

There are five stages of change or transition from smokingto nonsmoking status (63). These stages are precontemplation,contemplation, preparation, action, and maintenance. The cli_nician's role is to help the patient move through these stages andensure that interventions are tailored to the patient's stage.

A strong social support system is associated with cessationand long-term abstinence. Support may come from professionalsas well as from family and friends. The intended quitter shouldtake positive steps to recruit support and cooperation, includinginsistence on other smokers in the home environment absentinethemselves from the premises when they insist on smoking. Thismoker should also devote thought to other ways to auoid cir-cumstances likely to prompt relapse, including coping with per-sonal and interpersonal stresses.

Setting a "quit date" may be helpful. For most, though notall, patients, quitting "cold turkey', is more likely to be success-ful than gradual withdrawal. It is helpful for the physician andother health care providers to participate in setting the target dateEncouraging telephone calls to the patient on that dati and atfollow-up intervals thereafter may be helpful.

Establish Diagnosisand

Assess Symptoms

Encourageexercise, healthylifestyle andprovideimmunizations

Treat obstruction Educate patient onoptimal medication use.

Assess for hypoxemiaand treat if indicated

Assess response to therapy:> 2 hospital or emergency room visitsSevere symptoms or decreasedfunctional capacity.

Refer to Mult idiscipl inaryRehab Program

Medication

PeriodicReassessment

Figure 3. Management of coPD (adapted from reference , l94) Used with permission from Barry Make,

M.D. and Nat iona l jew ish cenrer fo r lmmuno logy and Resp i ra to ry Med ic ine , 1995.

Amer ican Thorac ic Soc ie t ,

TAEIE 3

PROTOCOL FOR SMOKINC CESSATION' I

. Ina t ia t ion

Phys ic ian or hea l th care worker 5hou ld in i t ia te qu i t t ing , exp la in ing r i sks o fc i g a r e t t e 5 m o k i n g l o r t h e r n d i v i d u a l a n d i n c l u d r n g s t r o n g a d m o n i t i o n . E n -courage es tab l i shment o t a de l rn r te qur t da te . Of fe r re fe r ra l fo r se l f -he lpor g roup program.

2 . Ear ly to l low-up

Te lephone pa t ren t on or w i th rn 3 to 5 d a f te r qu i t da te . Rev iew progressand counse l regard ing recru i tment o f suppor t person. Ca l l aga in I to 2 wkat te r qu i t da te . Repeat as needed.

3 . Cont inu ing re in fo rcement

Fur ther fo l low 'up shou ld be ar ranged by phys ic ian or hea l th care worker .Next regu la r v is i t shou ld be less than 2 mo a t te r in i t ia t ing cessat ion prc -gram; assess progress w i th CO and exp i red a t r and lo t co t rn rne in u r rne ,b lood, o r sa l i va . l f abs t inent , rev iew and reward success and re in fo rcepr io r warn ings . May lo l low by phone month ly un t i l nex t v is i t ; con t inuefo l low-up a t inc reas ing in te rva ls fo r

'12 mo a l te r qu i t da te .

4 . Fa i lu re o r rec id iv ism

l f pa t ien t fa i l s to ach ieve abs t inence or does so bu t re lapses , phys ic ian orhea l th care worker shou ld rev iew program wi th pa t ien t , emphas iz inge lements o t success and ident i f y ing c i rcumstances o f fa i lu re ; exp lo re a l te r -na t ives . N ico t ine rep lacement may be o f le red to cont ro l w i thdrawal sy rnp-toms; in t requent ly , o ther pharmaco log ic therapy , such as c lon id ine orbusp i rone, may be d iscussed. Hypnos is may be cons idered bu t i s o f l i t t leva lu€ when as a s ing le -sess ion recourse . Acupuncture is no t recommended.

Group smoking cessation clinics are offered by many hospi-tals and in many work sites as well as by voluntary agencies (theAmerican Lung Association's Freedom from Smoking clinics,for example). Such programs may play an imporrant supporr rolefor patients who are artempting to quit smoking, as well as forrecent quitters. They are especially important when the physi-cian or other primar"v caregiver is insufficienrly skil led in smok-ing cessation techniques. They ma-v also effectivel-v integratebehavioral therap,', additional counseling, adjuncrive pharmaco-logic treatment, and relapse prevention.

Pharmacologic Intervention

Nicotine is the ingredienr in cigarerres primaril! ' responsible forthe addict ion of smoking (61) . Wirh each c igarerre smoked, be-tween I and 2 mg of nicorine is delivered ro rhe lungs. Becauseof rapid absorpt ion in to the b loodsrream and a hal f - l i fe of 2 hr ,regular dayt ime smoking can cause n icor ine accumulat ion overan entire 24-h period.

Nicot ine is metabol ized chief l1 ' b l rhe l iver . Cor in ine, a pr i -mary metabol i te of n icor ine, has a longer hal f - l i fe , can be de-tected in the ur ine, and can conf i rm smoking conr inuar ion aswell as exposure to secondhand smoke.

Withdrawal from cigarerres causes unpleasanr side effecrs inmany, perhaps most, quitters. Among those reported have beenanxiety, irritabil iry, diff icu lry concenlraling, anger, farigue, drow -siness, depression, and sleep disruprion: rhese reactions are mostl ike ly to occur dur ing rhe f i rsr week fo l lou ing cessar ion.

Nicotine replacement after smoking cessation reduces \\ ' i th-drawal symptoms for those u'ho are addicred and enhances ab-stinence in a dose-dependenr manner. Highll dependenr smokerscan be identif ied as those *,ho smoke o\ er one pack of cigarer resper day, require their f irst cigarerre * irhin 30 min of arising. andfind it diff icult to refrain from smoking in places u here i i is for-bidden. Physical dependence can also be assessed by a form suchas the Fagerst rom to lerance quest ionnaire (6{} .

Nicot ine polacr i lex is avai lable in the iorm oi cheuing gum,provid ing 2 mg of n icor ine per p iece. l rs ef fecr i reness comparedwi th p lacebo has been demonsrrared. especia l l ! in h ighl l ad-dicted, self-referred smokers. A recenrl) approred dosage of 4 mg

per piece is even more efficacious in those who are highly ad-d icred to n icot ine (62) .

Transdermal nicotine patches are also readily available forreplacement therapy (65). Short-term success rares have variedwidely between l8 and 77clo,but in general, rates have been abouttwice those reported for placebo patch users. Long-term (6 moor more) success rates are considerably lower (22-420/o) but aresti l l consistently better than those for placebo nials (2-26V0'1.Nicotine patches are well tolerated; mild erythema or other localskin reaction may be seen in up to 50q0 of patients but can beminimized by rotating the patch among different skin sites andtry ing d i l ferenr brands.

Nicotine replacement regimens combined with adjuvant pro-grams such as individual counseling or group therapy producehigher success rates than nicotine replacement alone. lt has alsobeen shown that smoking behavior early in a nicorine-patch pro-gram can often predicr the outcome (66); abstinence during thefirst 2 wk suggests probable success. Smoking during the secondweek of treatment is a consistently powerful predictor of failureat the end of a 6-mo trial. Patients who fai! during the first 2wk of therapy should be offered more intense pharmacologicor adjuvant therapy. The ideal duration of therapy for each nico-tine replacement dosage has not been established, although ithas been suggested that nicotine patch therapy beyond 6 to 8 wkmay not be necessary (66,67\.

Clonidine, an alphar-adrenergic agonist, may enhance absti-nence in the short term, but enduring effects have not beendocumented. The anxiolytic buspirone has been shown to reducewithdrawal symptoms and may enhance abstinence.

Other Techniques

Hypnosis may be an effective adjunct to a basic smoking cessa-tion program. It is of little or no value, however, as a single-session"cure." Acupuncture is not recommended, as there is l i tt le evi-dence that it contributes to smoking cessarion beyond its placeboor demand characreristics (68).

PHARMACOLOCIC THERAPY

Outpatient pharmacorherapy should be organized according tothe severity of disease and the patient's tolerance for specificdrugs. The therapy goals are to induce bronchodilation, decreasethe inflammatory reacrion, and facil irate expectoration. Ingeneral, a stepwise approach should be considered (69, 70). Guide-lines for use of currently employed drugs in relation to signs ands)'mptoms, disease severity, and other variables are outl ined inTable 4.

Bronchodilators

The pharmacotherap)' of COPD is similar ro that of asthma,but important differences are recognized. Betar-agonists produceless bronchodi la tat ion in COPD than in asthma; in some pa-tients. spirometric changes may be insignificant despite symp-tomatic benefit. COPD parients, as an older group, may exhibitless tolerance for sympathomimetic-induced tremor, nervousness,and card iac s ide ef fecrs.

Betat-agonisls. There is no evidence that early, regular useof pharmacotherapy can alter the progress of COPD. Thus, ina pat ienl wi th in termi t tent symproms i t is reasonable to in i t ia temetered-dose inhaler therapy of a beta2-selective bronchodila-tor only when needed for relief of symptoms. Albuterol, pir-buterol, metaprorerenol, terbutaline, or isoetharine*each ofu'hich is preferable to less selective drugs, such as epinephrinc,isoproterenol, and ephedrine - should be used up to a maximumof three or four times a day or as prophylaxis before exercise.A spacer should be used, if indicared, ro improve aerosol dqliv-er-v and reduce side effects.

AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE VOL . I 52 1995

TABLT 4

STEP.BY-sTEP PHARMACOLOGIC THERAPY FOR COPD

1. For mi ld , var iab le symptoms:. Se lec t ive p2-agon is t metered dose inha le r (MDl ) aeroso l , 1 -2 pu f fs every

2-6 h as needed no t to exceed 8-12 ou f fs oer 24 h2 . For mi ld to modera te cont inu ing symptoms:

. f p ra t rop ium MDI aeroso l , 2 -6 pu ' t f s every 6 -8 h ; no t to be used moret requent ly

p l u s. Se lec t ive p-agon is t MDI aeroso l , 1 -4 pu f fs as requ i red four t rmes da i l y

fo r rap id re l ie f , when needed, o r as regu la r supp lement3 . l f response to s tep 2 i s unsat is tac to ry , o r there is a mi ld to modera te in -

c rease in symptoms:. Add sus ta ined re lease theophy l l ine , 200-400 mg tw ice da i l y o r 400-800

mg a t bedt ime fo r noc turna l b ronchospasmand lor. Cons ider use o f sus ta ined re lease a lbu tero l ,4 -8 mg tw ice da i l y , o r a t

n igh t on ly

a n o / o r. Cons ider use o f mucok ine t ic agent

4 . l f con t ro l o f symptoms is subopt ima l :. Cons ider course o l o ra l s te ro ids (e .9 . , p rednrsone) , up to 40 mg/d fo r

1 0 - 1 4 d- l f improvement occurs , wean to low da i l y o r a l te rna te-day dose,e . 9 . , 7 . 5 m 9- l t no improvement occurs , s top abrupt ly- l f s te ro id appears to he lp , cons ider poss ib le use o f aeroso l MDl , par -t i cu la r ly i f pa t ien t has ev idence o f b ronch ia l hyper reac t iv i t y

5 . For severe exacerbat ion :. fnc rease p2-agon is t dosage, e .9 . , MDI w i th spacer 6 -8 pu l l s every Vz-2 h

or inha lan t so lu t ion , un i t dose every Vz-2 h o r subcutaneous admin i -s t ra t ion o f ep inephr ine or te rbu ta l ine , 0 .1 -0 .5 ml

ano l o l. Inc rease ip ra t rop ium dosage, e .9 . , MDI w j th spacer 6 -8 pu f fs every

3-4 h o r inha lan t so lu t ion o f ip ra t rop ium 0 .5 mg every 4 -8 ha n d. Prov ide theophy l l ine dosage in t ravenous ly w i th ca lcu la ted amount to

br ing serum leve l to 10-12 pg lml

a n do Prov ide methy lp redn iso lone dosage in t ravenous ly g iv ing 50- ' lO0 mg im-

media te ly , then every 5 -8 h ; taper as soon as poss ib lea n d a d d :. An an t ib io t i c , i t ind ica ted. A mucok ine t ic agent j f spu tum is very v iscous

The rapid onset of action of beta-agonist aerosols may leaddyspneic patients to favor them for regular use. At present, thereis l i tt le evidence ro suggest rhat their regular use is harmful inCOPD, although some studies suggest that ir results in a slowdecline in FEV, (71,72) (as has been reported for asthma). Thereis no evidence to suggest that regular use of beta-agonists signif-icantly curtails survival of patients with COPD. The potentialfor arrhythmias necessitares careful dosing in patients with prob-able or known cardiac disease, although serious cardiac compli-cations are rare with conventional doses.

In more advanced disease, it may be reasonable to use slow-release oral albuterol, but the value and acceptabil ity of suchformulations have not been established. Similarly, the role of thenew, long-lasting aerosol salmeterol in COPD needs to be deter-mined, but it has been shown to prevent nocturnal bronchospasm.Both ol these agents improve compliance, but further studiesare required to establish their place in management of broncho-spast ic d iseases in o lder pat ienrs.

Anticholinergic agents. Topical adminisrration of an anri-cholinergic aerosol may be more effective than a topical beta-agonist in COPD, and the side effects tend to be less trouble-some; currently, ipratropium is the only merered product. Oncethe patient suffers from daily symptoms, regular use of ipratro-pium via a metered dose inhaler is recommended. The drug hasa slower onsel and longer action than such betar-agonists as al-buterol and thus is less suitable for use on an..as needed" basis.

Appropriate dosage is two to four puffs three or four times aday, but some patients require and tolerate larger doses, e.g., sixto eight puffs three times a day.

Inhaled anticholinergic bronchodilators do not appear to in-fluence the long-term decline of FEV, (32). However, there isno substantial evidence to suggest that regular use of anricholin-ergic therapy, with or without a beta2-agonist, leads to a wors-ening of spirometry or to exacerbations or premature death inCOPD. It is thus appropriate at rhis time to use iprarropium ther-apy and add a betar-agonist as ofren as needed, up to four timesa day. The use of a combination of albuterol and ipratropiumin the same merered dose inhaler wil l help simplify therapy.

Theophylline. Duringthe decade following the introductionof sustained-release theophylline, the agent was greatly favoredfor the management of both asthma and COPD, particularly incases involving nocturnal bronchospasm. Theophylline's poten-tial for toxicity led to a decline in its popularity, but it still re-tains an important role in COPD (73). It is of particular valuefor less compliant or less capable patients who cannot use aero-sol therapy optimally because they can readily take long-actingtheophyll ine once or twice a day.

The ability of theophylline to improve respiratory muscle func-tion, stimulate the respiratory center, and enhance activities ofdaily living can be important to patients who are severely limitedby COPD. Theophylline can also improve cardiac output, reducepulmonary vascular resistance, and improve the perfusion of isch-emic myocardial muscle. Thus, there are several advantages totheophylline therapy in parients wirh COPD who also havi car-diac disease or cor puimonale (74).

Theophylline has been shown to have some anti-inflammatoryeffects (75). Regular use of theophylline has not been shown tohave either a progressively beneficial or a detrimental effect onthe course of COPD. Recent information suggests that addingtheophylline to the combination of albuterol and ipratropiumcan resulr in maximum benefit in stable COPD (76). Careful dos-ing is needed, parricularly in the presence of other disease or otherinteracting drugs in the therapeutic regimen (Table 5). precau-tions that should be taken when using beta agonists, ipratropium,and .theophyll ine are shown in Table 6.

Anti-inflammatory TherapyIn contrast to their value in asthma management, the role of anti-inflammatory drugs in the routine treatment of COpD is lessclear. Cromolyn and nedocromil have not been established asuseful agents, although they could possibly be helpful ifthe pa-tient has associated respiratory tract allergy. Corticosteroids maymerit more careful evaluation in individual patients on adequatebronchodilator therapy who fail to improve sufficiently o, *hosedisease undergoes exacerbation.

Corticosteroids. At present, rhere is no evidence to suggesr thatpatients with COPD who are being treated with regular bron-chodilator therapy require the ,,protective" effects of added ste-roid therapy as used in asthma. In outpatients, exacerbations maynecessitare a course of oral steroids, but it is important to weanpatierrts quickly, since the older COpD population is suscepti-ble to complications such as skin damage, cataracrs, diabeies,osteoporosis, and secondary infection. These risks do not ac-company standard doses of steroid aerosols, which may causethrush but pose a negligible risk for causing pulmonary infection.

Most studies suggesr thar only 20 to 300/o of patients withCOPD improve if given chronic oral steroid therapy (7a). Thedangers of steroids require that careful documentation of theeffectiveness of such therapy be recorded before a patient is puton prolonged daily or alternate-day dosing; the latter regimenmay be safer, but its effectiveness has not been evaluated inCOPD. It is possible that steroids can slow the deterioration ratein COPD, but this suggestion has not been adequately evaluated

Amer ican Thorac ic Soc ie ty

TABLE 5

FACTORS THAT MIGHT AFFECT THEOPHYLLINEDOSACE REQUIREMENTS

Larger Theophy l l ine

Dosage Requ i red 'Smal le r Theophy l l ine

Dosage Requared

Maior change (26-50%)

Lesser change (10-25%)t

Cigare t te smok ing

Phenyto in

Ri lampin

lsopro tereno l l .V

Phenobarb i ta l

CarbamazeprneAminog lu te th imrde

Low carbohydra te , h igh

pro te in d ie t

Charcoa led food

I son ia2 id

Ketoconazo le

Hepat ic insu f f i c iency

Hear t fa i lu re

Cor pu lmona le

V i ra l pneumonia

Cimet id ine

Mex i le t ine

Cipro l loxac in , o ther quano loner

A l lopur ino l

Erythromycin

In t luenza vacc ina t ion

Triacetyloleandomycin (TAO)

Propranolol

Oral contraceptives

Verapami l

N i fed ip ine

Tetracycli ne

Hydrocortisone

Aluminum hydrox ide

Magnes ium hydrox ide

Thiabendazole

When increa : ing dosage, serum leve ls must b€ used fo r gu idance.Data suppor t ing these changes are less we l l -documented.

(11,72). It is possible, too, that aerosol steroids could be usedin place of low-dose oral steroids, but again there is insufficientdocumentation to support such therapy; several multicenter studiesare evaluating the course of COPD treated with aerosol steroids.Adding an inhaled steroid to a regimen that includes regular aero-sol therapy with ipratropium and a beta-agonist results in theprobabil ity of noncompliance being increased. If patients areplaced on long-term oral corticosteroids, supplemental calciummay help to prevent osteoporosis. Precautions required with ste-roid therapy are noted in Table 6.

Drugs Affecting Mucus

The only controlled study in the United States suggesting a valuefor a mucokinetic agenl in the management of chronic bronchi-tis was a multicenter evaluation of organic iodide (77). This studl'did demonstrate symptomatic benefits, but objective evidenceof benefit was deemed insufficient, and the FDA required thatmarketing of the drug be discontinued. The values of otheragents, including water, have not been clearly demonstrated, al-though some established drugs, such as oral acetylcystdine, arefavored in Europe for their anri-oxidant effects in addition totheir mucokinetic properties. Geneticall l ' engineered DNAse hasbeen proved to be useful in cystic fibrosis. The efficac-v of DNAsein the t reatment of COPD is under invesr igar ion.

Antibiotics are not of proven value in the prevention or treat-ment of COPD exacerbations unless there is evidence of infec-tion, such as fever, leukocytosis, or a change in the chest radio-graph (78-80). ln cases of recurrenr infection, parricularly inwinter, prolonged courses of antibiotics. u'hether continuous orintermittent, may be useful.

When an acute bacrerial infecrion of spurum is believed robe present, antibiotic therapy ma1' be jusrif ied. The decision isusually made clinically because sputum culture is nor cost-effec-tive. The major bacteria to be considered are Streprccoccus pneu-moniae, Hemophilus influenzae, and Moroxello catarrhalis. An-tibiotic choice wil l depend on local experience, supponed bysputum culture and sensitiviries in rhose cases where the patientis moderately ill or needs to be admirted to rhe hospiml. ln general,fiscal concerns should be taken inro accounr. and older, less costly

agents - tetracycline, doxycycline, amoxicillin, erythromycin, tri.methoprim-sulfamethoxazole, cefaclor, etc. - are often as effec-tive as newer, more expensive ones.

Other Pharmacologic Agents

Alpha,-antitrypsln augmentation therapy is appropriate in non-smoking, younger patients with severe alpha,-antitrypsin (AAI)deficiency and associated emphysema (see the earlier discussionof this subject); such therapy is not indicated in the commonform of COPD.

Respiratory stimulanls, used in some other countries, are notcurrently recommended in the United States.

Psychoactive agents are often sought by older patients withCOPD to treat depression, anxiety, insomnia, or pain. Theseagents can be helpful with appropriate care and with particularawareness of potential depressant effects on the respiratory cen-ter (81). Benzodiazepines do not have a marked effect on respi-ration in mild or moderate COPD but can be suppressive in se-vere disease, especially during sleep. The safer hypnotics for usein insomnia include sedating antihistamines and chloral hydrate;antidepressants may also have the advantage of improving sleep.

Cardiovascular theropy may be needed in severe COPD andcor puimonale; agents may include diuretics, angiotensin-con-verting enzyme (ACE) inhibitors, or calcium channel blockers.Digoxin is occasionally useful, while beta-adrenergic blockersare generally contraindicated. All such drugs must be used cau-tiously to avoid precipitating electrolyte imbalance, dehydration,hypotension, myocardial ischemia, and arrhythmias. Becausemost patients requiring such therapy are elderly or have impaireddrug clearance, alertness to potenrial side effects is vital; if theyoccur, the drug regimen must be promptly modified.

Immunization with pneumococcal and influenza vaccines isrecommended to prevent infectious complications involving therespiratory tract. Although the available vaccines are not totallyeffective, there is evidence that COPD parients are l ikely to ben-efit from their use.

Adjustment of Therapy

The relationship between patient and physician must be dynamic

TABLE 6

PRECAUTIONS IN DRUG USE

Precaut ions when us ing be ta-agon is ts. Watch fo r non improvement o r paradox ica l de ter io ra t ion w i th

aeroso l use. Use spacers to improve compl iance and reduce sys temic s ide e f fec ts. Avo id overuse; check number o f metered dose inha le rs (MDls)

used per month aga ins t number o f pu f ts per MDI (200

to 300+ depend ing on brand) .. lns t ruc t pa t ien t on max imum number o { pu f fs per day (usua l ly 8 - '12)

and on number a l lowed dur ing an exacerbat ion (e .9 . , 12-24 over

3-4 h) be fore add i t iona l in te rvent ion is requ i red. l f a long-ac t ing agent i s used, caut ion pa t ien t tha t f requent use must

be avo ided. Home updra f t nebu l izers w i th inha lan t so lu t ions tha t p rov ide la rge

dosages are ra re ly needed

Precaut ions when us ing theophy l l ine. ln i t ia te t rea tment w i th a low dose (e .9 . , 400 mg/d) and ad jusr a f te r a

few days. Reduce dosage i l d rug c learance is l i ke ly to be impa i red because o l

i l l ness , l i ve r ma l lunc t ion , o r concomi tan t d rugs .. Do no t a l low any add i t iona l theophy l l ine prepara t ion to be taken. Drug must be taken a t the same t ime each day w i th r€spec t to mea ls. When symptoms change, acu te i l l ness deve lops , new drugs are added,

or symptoms suggest ive o f tox ic i t y deve lop , check serum leve l o f

theophy l I i neo A i m f o r a s e r u m l e v e l o f 8 - 1 2 p g / m l ; a d j u s t d o s a g e a n d f o l l o w s e r u m

leve l when ind ica ted

P r e c a u t i o n s w h e n u s i n g i p r a t r o p i u m. Pat ien ts shou ld genera l l y use a spacer and shou ld avo id spray ing in to

' Be prepared to inc rease dose i f necessary f rom 2-3 pu f fs 3 -4 t imes a day

to 6 -8 pu f ls -3 -4 t imes a day , i f to le ra ted. Caut ion pa t ien t tha t onset o f e f fec t i s re la t i ve ly s low; add i t iona l doses

shou ld no t be taken fo r acu te symptom re l ie f. Mon i to r fo r s ide e f tec ts , e .9 . , tachycard ia , d ry mouth , g laucoma,

pros ta t i sm, o r b ladder neck obs t ruc t ion

Precaut ions when us ing ora l s te ro ids. Reduce dosage to lowest e f fec t i ve da i l y dose or to a l te rna te-day dosang

as qu ick ly as symptoms a l low. Mon i to r fo r hyper tens ion , d iabetes , we igh t ga in , menta l changes,

in fec t ions , cen t ra l po la r ca tarac ts , sk in th inn ing , purpura , os teoporos is ,

and os teonecros is. D is t ingu ish psycho log ica l benef i t f rom t rue pu lmonary ben€f i t by

lo l low ing FEVl fo r 2 wk a f te r in i t ia t inq therapy. Admin is te r p rophy lac t ic ca lc ium therapy to women; t rea t os teoporos is

a pprop na te ly. S tero id -dependent pa t ien ts requ i re s te ro id coverage dur rng any c r is is

lo r many months a t te r s topp ing s te ro ids. Repeated ly eva lua te pa t ien t to de termine i f s te ro id th€rapy can be

d i scont i nuedP r e c a u t i o n s w h e n u s i n g a e r o s o l s t e r o i d s

. Seek ob iec t ive ev idence o f the va lue o f th is therapy because r ts usemay decrease compl iance w i th o ther aeroso l usage

. S tandard dos ing (2 -4 pu l fs 2 -4 t imes a day) shou ld no t be exceeded

. Use a spacer ; mon i to r fo r o ra l th rush and la ryngea l dys func t ron

. 8e aware tha t aeroso l s te ro id s ide e f fec ts may occur In sk rn , bone, e tc .

. When in t roduc ing aeroso l s te ro ids In a pa t ien l on an ora l s te rord , wean

s l o w l v o f f t h e o r a l d r u g

and participative. Since there are a number of therapeutic choicesavai lable, par ients must become educated and take par t in thei ro\\ 'n care. This dynamic interaction, known as collaborotive self-tnonagement, allows patients to modify therapy according ro dis-ease severit), and variations in symptoms. Patients ma),, for ex-ample, add regular doses of iprat ropium when the pers is tenceoi s) 'mptoms lessens the usefu lness or safety of beta-agonists .Simi lar ly , courses of ant ib iot ics or cor t icostero ids could be in i -: ia ted by sel f -managed pat ients who are knowledgeable aboutrheir disease and the need to introduce specific t-vpes of rreatmenr.

Further investigation is needed in a number of areas. Optimaljosages of betar-agonist bronchodi la tors, inc luding ne$ long-

AMERICAN 'OURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE VOL 152 1gg5

acting aerosols and oral preparations, and the value of their com-bination with ipratropium or other anticholinergic agents, needto be determined. Some studies that suggest regular use of theseagents on a daily basis may lead to more rapid deterioration ofCOPD point to the need for further study. The value of currenttheophyll ine preparations and new derivatives, either alone orin combined regimens, has to be evaluated. The most importantarea for research is to define clinical criteria for the use of oralor aerosol steroids in COPD and determine if their use can re-tard the rate of deterioration.

Additionally, the possible value of cromolyn, nedocromil, andother yet-to-be marketed anti-inflammatory drugs needs to bedetermined for various rypes of COPD and associated chroniccough. Mucus therapy and antioxidant therapy are strongly fa-vored in Europe, but definit ive studies are sti l l required for bothold and newer drugs. lt is important to determine whether ornot appropriate muco-active drug therapy wil l reduce the l iabil-i1y of patients to pulmonary infections and their consequencesbecause antibioric rherapy is of l imited benefit in the typical,less severe exacerbation. Respiratory stimulants may prove to bebeneficial, but the value of European drugs such as almitrineis uncertain. Genetically engineered agents such as alfa-dornasemay have an important role in future therapy, and this avenueshould continue to be exDlored.

LONG-TERM OXYGEN THERAPY

COPD is commonly associated with progressive hypoxemia. Hyp-oxemia can rapidly lead to damaging cellular hypoxia, and theadministration of oxygen can be l ife preserving. Thus, the cor-rection or prevention of hypoxemia assumes high priority. Inpreserving cellular oxygenation, the clinician must also addressother oxygen transport variables, including adequate hemoglo-bin, cardiac output, and tissue perfusion.

Oxygen Therapy, Survival, and Quality of LifeLong-term oxygen therapy will reverse secondary polycythemia,increase body weight, alleviate right heart failure due to cor pul-monale, strengthen cardiac function, enhance neuropsychologi-cal function, and improve exercise performance and activit iesof dai ly l iv ine (82, 83) .

Two-studies have shown that long-term oxygen therapy im-proves survival in hypoxemic patients with COPD. The Brit ishMedical Research Council (MRC) srudy, which compared hypox-emic patients receiving oxygen for l5 h/d, including the hoursof sleep, with patients receiving none, showed a significant reduc-tion in mortality (8a). The National Heart, Lung, and Blood ln-stitute's Nocturnal Oxygen Therapy Trial (NOTT), comparingcontinuous oxygen therapy (average 19 h/d) versus 12 h/d, in-cluding the hours of sleep, showed a further reduction in mor-ta l i ty us ing conr inuous oxygen (85) .

The mechanism for improved survival has yet to be completelydelineated; however, it appears l ikely that pulmonary hemo-dynamics may be involved (86). Chronically hypoxemic parientstend to have increased pulmonary artery pressures; when theyreceive oxygen, they experience a fall in those pressures. Whilesuch correction in pulmonary artery pressure is not complete,lowering it wil l reduce cardiac work and improve tissue oxygendelivery.

The overriding concern is to prevent t issue hypoxia. Simplycorrecting hypoxemia may not be sufficient to treat or preventtissue hypoxia. The clinician must be concerned about the fullscope of oxygen transport and delivery to the tissues. Accord-ingli ' , lung function ought to be maximized, including correc-tion of bronchospasm, control of bronchopulmonary infection,and resolution of congestive heart failure. The transport vari-

American Thoracic Society

ables, inc luding hemoglobin and card iac output , must a lso beaddressed.

There has been overemphasis on the fear that oxygen therapymight lead to respiratorl ' drive depression, hypercapnia, and re-spi ratory ac idosis . As a consequence, c l in ic ians may be over lyt imid about prescr ib ing oxygen. l t is now understood that , whi leCO, retent ion does occur (87, 88) , i t is of ten caused by vent i la-tion-perfusion (V/Q) mismatching rather than by respiratory cen-ter depression (89). Therapeutic preference must be given to cor'recting the hypoxemia.

When oxygen is delivered via nasal cannula, a relatively lowflow of 10090 oxygen is mixed into a much higher inspiratoryflow of room air (20.9V0 oxygen). The net effect is to increasethe Pag, by an amount dependent on gas exchange variables,respiratory pattern, and arteriovenous shunting. Unless the shuntfraction approaches 5090, low-flow oxygen is very effective incorrecting hypoxemia. On the negative side, oxygen therapy it-self may be associated with such adverse effects as absorptiveatelectasis and reduction of hypoxic vasoconstriction that mightextend V/q mismatch. In general, however, the net effect of oxy-gen therapy is to reverse hypoxemia.

Measurement of arterial blood gases (ABG) is the usual meansof assessing gas exchange status. The methodology for arterialsampling has been standardized, it is safe, and complicationsare uncommon and relatively minor. With frequent requirementfor ABC assessment, an arterial l ine may be considered. Sourcesof error in ABG measurements include improper sample site(vein), inadequate blood handling techniques, high white bloodcell count (decreases the Pa6r), and instrumentalion error.

The arterial blood gas analyzer measures Pa6r, Pas6r, andpH. Bicarbonate and Sao, are calculated using standard al-gorithms. Sag, can also be measured directl l ' , using a co-oximeter,which also measures carboxyhemoglobin and hemoglobin.

Noninvasive pulse oximetry measures the transmission of twowavelengths of l ight through the skin, typically' of the finger orearlobe. The red band represents oxygenated hemoglobin, whilethe infrared band represents unbound hemoglobin. Pulse oxim-etry generally correlates well with co-oximetrl ' measured fromarterial blood, with 1 ro 2Vo error. Inaccuracies are created by'high levels of methemoglobin, carbox)'hemoglobin, bil irubin,dark sk in p igment , inadequate t issue per fus ion, and movementartifact. In spite of these error sources. pulse oximetrf is a goodmethod for following ox!'gen saturation and can be used for titrat-ing the oxygen flou' setting. Oximetrl is not considered suffi-ciently accurate to replace ABC in an init ial assessment and can-not be used to determine ac id base sratus. A b lood gas Sa6, canbe used to corroborate the o.rimerrl Saq,.

Transculaneou.i Po, measures local o\)'gen level across thedermal sur face. l t a l lous conr inuous assessment and, in contrastto oximetry, measurement of hl peroxia. ln the neonatal patient.this measurement is accurate anci srrongll correlates *' i th Pa6r.In the adul t . however. the sk in sur face i r uneren. th icker , andsubject to changes in local blood flog. Transcutaneous Po, hasnot been widely used because there is l i r r ie agreement on i ts ac-curacy or in terpretat ion in adul rs . Addir ional l l . r t requi res aheated e lect rode that can occasional i l ;ause

'Du:ns and heat

blisters.In summary, arterial blood gas measurem.n: rs recommended

for in i t ia t ion of oxygen therapl as sel i as ro ie le: 'mine Pco, andacid-base status. ABG ma;- a lso be usec inrr ra i i l to conf i rm theaccuracy of pulse ox imetr f in ad_ius l ing ih : or lgen f lou set t ing.Noninvasive pulse ox imetr l is usefu. lo : as iessrne Sa6, and ad-just ing oxygen f low set t ings: i t is les. :e. ra l i r : . s \erc lse studiesthan at rest , and ABG at the beginnr :g a i : : r 'n . : o i an exerc isesession may provide confirmato:1 ia:a T=:::-ulaneous measure-ments of Por are inaccurate an; a: r : : .a :a: r r . : ' lo : r 'cc()mmcndcdin adu l t s .

Oxygen Therapy during Sleep

Patients who are hypoxemic while awake are l ikely to be hypox-emic during sleep as well. l f the patient does not have sleep-disordered breathing due to other causes, the administration ofnocturnal oxygen therapy wil l correct nocturnal hypoxemia. ltis often recommended that I L/min of continuous flow oxygenbe added to the daytime resting prescription to allow for addi-tional hypoventilation and gas exchange.abnormalit ies duringsleep, but continuous oximetry monitoring during sleep wil l sup-porl a more accurare prescription.

Oxygen Therapy for Exercise-induced Hypoxemia

COPD patients are encouraged to remain active. Due to the un-derlying lung disease and inefficient gas exchange, continuarionof an active l ifestyle becomes increasingly diff icult. Many pa-tients with COPD who are hypoxemic at rest worsen during ex-ertion, while others develop hypoxemia only during exertion.Home supplemental oxygen is commonly prescribed for the lat-ter group, even though studies designed to determine the long-term benefit of oxygen soiely for exercise have yet to be conducted.

Some short-term studies have shown that supplemental oxy-gen during exercise can prevent transient increases in pulmonaryartery pressure (PAP) and pulmonary vascular resistance (PVR)(90). The immediate benefits of oxygen during exercise are reduc-tion in dyspnea and improvement in exercise tolerance at sub-maximal workloads (91). There are data suggesting that oxygencan also increase peak exercise level via a reduction in minuteventilation at submaximal workloads (92).

In determining the exercise flow setting for home oxygen, ar-terial blood gases may be measured while the patient is carryingon a level of exertion equal to or slightly greater than that nor-mally experienced during daily activities. Repeated exercise testsmay be necessary to titrate oxygen administration to increase Sag,to greater than 9090 during usual exertion. There wil l be casesin which rhis goal is not achievable; hence, clinical decisions wil lbe required to determine optimal oxygen flow, delivery method,and acceptable Saor. Patients should be evaluated using the samedelivery devices that they expect to use at home, e.g., nasal can-nula or oxygen-conserving device. Exercise testing is best car-ried out using muscle groups where dyspnea is most pronounced(legs versus arms, for example).

ri/hile ABG measurement remains the gold standard for de-termining arterial oxygenarion, Sao2 may be measured via nonin-vasive pulse oximetry. Unfortunately, absolute values may notbe accurate during exercise, particularly in patients with poorperipheral perfusion. Verif ication of oximetry accuracy can beaccomplished by obtaining ABG before and afrer exercise. Theadvantages of pulse oximetry are that it is noninvasive and pro-vides continuous read-out of Sas, which allows for Sa6, track-ing. Following Sao, during exercise tesring helps to dererminethe speed and extent of the desaturation resDonse.

Opt imal Medical Regimen

One of the goals of any medical regimen is to oprimize V/Qmatching as a means of correcting hypoxemia. This is particu-larlf important during and after an acute exacerbation. If thereis evidence of congestive heart failure due to cor pulmonale, di-uretic therapy may be advisable. The adequacy of oxygen deliv-er)'to the tissues may be improved by correcting anemia or poly-c1'themia. An exercise program may improve oxygen transportto *'orking muscles. lf the patient nevertheless remains hypox-emic, long-term oxygen therapy is conrinued, probably for thel i fe of the par ienr .

One aspect of good medical management is the oxygen ther-apy itself. Recent reports suggest that oxygen may have a repara-tive effect by reducing pulmonary arter)' vasoconstricrion and

AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE VOL I52 1gg5

improving V/Q matching (86, 93). If this is rhe case, withdrawalof oxygen because of improved Pae, may be distinctly detrimental.

Hazards of Oxygen

Oxygen therapy is generally safe. Hazards the clinician mustrecognize include oxygen toxicity, COr retention, and the possi-bil i ty of accidents during the storage and handling of oxygen.

Oxygen toxicity is related tofree radicols. The major end prod-uct of normal oxygen metabolism is water. Some oxygen mole-cules, however, are converted into highly reactive species calledradicals, which include superoxide anions, perhydroxy radicals,hydroxyl radicals, and hydrogen peroxide; rhey are toxic to alve-olar and tracheobronchial cells.

Normally, anrioxidanr enzymes, including metaloproreins(superoxide desmutase), catalase, and glutathione peroxidase,protect cells by scavenging the Oz radicals. But in the face ofprolonged exposure to high concentrations of oxygen, the an-tioxidant system becomes overwhelmed, permitting oxidative de-struction of lung tissue. Acute changes may be manifested as tra-cheobronchial irritation, impaired mucociiiary clearance, anddiminished vital capacity secondary to edema and reabsorptionatelectasis. As exposure continues, leaky capil laries, alveolarhemorrhage, surfactant dysfunction, and alveolar edema will de-velop. Adult respiratory distress syndrome may follow, with pul-monary infiltrates and fibrosis or death (94).

Pathophysiologic changes include decreased lung compliance,reduced inspiratory airflow, decreased diffusing capacity, andsmall airway dysfunction. While these changes are well-recognizedin the acute care setting of mechanically ventilated patients receiv-ing Fl6, in excess of 5090, little is known about the long-termeffect of low flow (24 to 28c/o) oxygen. In one uncontrolled au-topsy series, patients with chronic obstructive lung disease whohad been treated with chronic low-flow oxygen therapy showedexudative and proliferative changes consistent with oxygen tox-icity (95); there was no indication that these changes had im-paired survival. Animal studies have demonstrated that long-termexposure to 8090 Fro2 may induce tolerance, perhaps in the formof greater levels of protective antioxidant enzymes. It is widelyaccepted that the increased survival and quality-ofl ife benefitsof long-term oxygen therapy outweigh the possible risks.

CO, relention may pose a threat in patients with impaired COzventilatory drive. Depression of hypoxic drive by supplementaloxygen may induce worsening hypercarbia, respiratory acidosis,and CO, narcosis. While this complication is not common rn

low-flow oxygen therapy, it is best avoided by titration ofoxygendelivery 1o maintain pao, between 60 and 65 mm Hg. When hyfer_carbia is present, initial oxygen delivery settings should be titiitedusing serial ABG assessments rather than relying solely on ox-imetry.

Some CO: retention is tolerable; patients with an intact renalsystem are capable of reabsorbing enough bicarbonate to com_pensare for the extra CO, and maintain the acid_base balance.The abil iry of patients to tolerate CO, retention may be an adap-tive mechanism that lessens the work of breathing, since an in-creased alveolar Pco, results in increased CO, excretion with eachbreath.

The major physical hazards of oxygen therapy are fires or ex-plosions. Most fires have been caused by patients lighting ciga-rettes as oxygen is flowing into their noses (96). While the can_nula is constructed of fire-retardant plastic, both the patient'snose and the cannula wil l burn vigorously in the presence of ox_ygen in high concentrations. patients, famity, and other caregiversmust be warned not to smoke. Oxygen containers should not bestored near water heaters, furnaces, or other sources of heat orflame. Serious freeze burns can occur ifthe patient does not takeproper precautions while transfilling liquid oxygen. Finally, acompressed oxygen cyiinder can be accidentally knocked over,causing explosive disconnection of the regulator and renderingthe cylinder a dangerous missile. In general, major accidents aslsociated with oxygen therapy are rare and can be avoided by goodpatient and family training and common sense.

Oxygen Systems

Oxygen comes packaged in three types of systems: compressedgas, liquid, and oxygen concentrators. The trade-offs include sizeand weight of the device, storage capacity, cost, and transfilla_bility. The fearures of the systems are compared in Table 7.

Oxygen Delivery MethodsThe continuous flow dual-prong nasal cannula is the standardmeans of oxygen delivery for the stable hypoxemic patient. Itis.simple, reliable, and generally well-tolerated. In some patientswith advanced disease and severe hypoxemia, however, a facemask may be required to increase the fraction of inspired oxy_gen (Fter).

The nasal cannula delivers a low flow of pure oxygen entrainedin a much larger volume of atmospheric air (20.990 oxygen). Eachliter per minute of oxygen flow adds about 3 to 4go to ttre Fr6r.

TABLE 7

COMPARISON OF GAS, LIQUID, AND CONCENTRATOR OXYGEN SYSTEMS

. lnl__ Concentrator

Avar lab i l i t y

Re l ia b i l i t y

CostWeigh t + use t ime 2 L /m

Sta t iona ry

Por tab le

Por tab le + OCD(02-conserv ing dev ice)

Tra ns f i | |

Power

Anrbu la to ry use

Common

Cenera l l y good;

Bourdon gauges may

become inac(ura te

Modera te

H cy l inder :

200 tb . , 2 .5 dI cy l inder :

20 rb., 5 hFiber /a lum:

4 . 5 l b . , 1 2 hL imi tedN o n e

Cood w i lh conserver

L imi ted

Poss ib le inaccura te

se t t rn9 , t reez ing

of ConnectorH i g h

Reservo i r :' t 2 0

r b . , 8 . 9 dPor tab le :

6 t b . . 4 h

OCD sys tem:5 .5 tb . , 8 h

Exce l len t

None

Cood a lone andwi th conserver

Common

6ood, bu( needs

regu lar serv tce

Low

Sta t ionary 35- lb . un i t ;no s to rage capab i l i t y

N/A

1 1 0 v o l t s A CPortable, but not

su i tab le fo r

ambula tory use

American Thoracic Society

Determination of Continued Need

Standards for continuing ox),gen therap] differ depending onwhether it is prescribed for the first rime during an acute exacer-bation or at a time when the patient is relatively' stable and receiv-ing opt imal therapy (93) . Some pat ients n ' i th COPD who werenot hypoxemic before an exacerbarion u'i l l eventually reach apoint when they no longer need oxygen. For such palients, theneed for long-term oxygen should be reassessed in 30 to 90 d,when the pat ient is c l in ica l ly s table and receiv ing adequate med-ica l management. I f the pat ient does not meet b lood gas cr i ter iaat that t ime, oxygen therapy can be d iscont inued.

The major i ty of pat ients r r 'ho are c l in ica l ly s table wi l l con-tinue to meel prescribing criteria for LTOT. Some patients, how-ever, experience an improvement in Pa6, and, as a consequence,may no longer meet criteria for such rherapy. This can occur instable patients receiving optimal therapy who have been on oxy-gen for months or years (86, 93).

Improved oxygenation in patients with stable COPD is be-lieved to result from the "reparative" effects of oxygen, includ-ing reversal of hypoxic pulmonar;- vasoconstriction, and it shouldnot be used to just i fy d iscont inuat ion of therapy (93) . Once theneed for LIOT has been established in a stable patient on op-timal therapy, LTOT most l ikell, represents a l ifetime commit-ment. Measurements of oxygen sarurations ) 90Vo should notbe used as a rationale for discontinuing the therapy.

Reimbursement Criteria and Documentation (HCFA Criteria)

Oxygen is prescribed by the physician based on physiologic find-ings and clinical judgm'ent. ln the United States, reimbursementwil l not be approved unless patienrs meer established physioiogiccriteria. It is therefore essential for the ph1'sician to provide ap-propriate documenlation that oxygen is medically necessary andmeets physiological crireria as shown in Table 9. Most privateinsurance carriers and the Department of ! 'eterans Affairs fol-low Medicare guidel ines for re imbursement .

For a prescription of LTOI a certif icate of medical necessity(Health Care Financing Adminisrrarion. HCEA form 484) mustbe completed. The documenr evolved in an attempt lo be certainthat the physician, not a home medical equipment supplier, isresponsible for decisions concerning rherap)'. HCFA requires thephysic ian or an emplovee of rhe phls ic ian, rather than the sup-pl ier , to complete form 484; guidel ines for complet ing the formhave recent ly been summarized (109) .

Pat ient educat ion and moni tor ing of compl iance are essen-tial to assure the success of LTOT. \lanl parients harbor fearsregarding the therapl'. Some, for example. associate a need forlong-term oxygen with profound dereriorarion rather than prolon-gation of l i fe and enhancement of qualirl of l i fe; the)' ma), there-fore exper ience anger or denia i . Some ma1'bel iere oxy ' -een is anaddict ive substance and mal rherefore aro id i ts use as much aspossib le. These and other concerns need ro be explored and d is-cussed so thal the pat ient is reassured of the helpfu l nature ofthe therapy.

Future technological advances in orygen e\r ract ion f rom theatmosphere should lead to rhe developmenr of pract ica l por ta-ble concenlrators. Another probabil ir l is thar or1'gen stora_ee con-ta iners wi l l become l ighrer . smal ler . anci more easi l l t ransf i l la-ble, not only from liquid bur from gas and concentrators as *'ell.Oxygen del ivery and conserv ing devices wi i l be rendered moreef f ic ient , less cost ly , and more accessib ie ro a l i uho need them.

PULMONARY REHABI LITATION

The three major goals in the managemeni o: COPD are: ( / ) tolessen a i r f low l imi tat ion; (J) r t r prere: : : i : t . i i rca i secondar i 'med-ica l compl icat ions, such as hrporem: i enu rnfcct ions; and ( - l )

to decrease respiratory symptoms and improve quality of l i fe.Many patients may be unable to enjoy l ife to the fullest becauseof shortness of breath, physical I imitations and inactivity im-posed b1'that fear-provoking symptom, and mechanical impair-ment of their respirarory system. Pulmonary rehabil iration (l10,I I I ) comprises a constellation of therapeutic modalit ies in an on--eoing, multidimensional continuum of services designed to pro-vide a framework for quality-of-l i fe improvement.

Recent comprehensive review articles and texts provide exten-sive references delineating the components of pulmonary reha-bil itation and expected results (l l2-l l5). ln evaluating the bene-fits of pulmonary rehabil itation, it is diff icult to distinguish therelative contributions of specific therapies because they are in-tegrally related. Almost any treatment (e.g., education, exercise)given by well-trained personnel wil l inevitably provide elemenrsof psychosocial support and motivation for chronically i l l anddisabled patients. As a result of rehabil itation, improvementshave been demonstrated in objective measures of quality of life(l l6), well-being (117), and health status (l l8), including reduc-tion in respiratory symptoms, increase in exercise tolerance andfunctional activit ies such as walking, increased independence,enhanced abil ity to perform activit ies of daily l iving, improvedpsychological function with less anxiety and depression, and in-creased feelings of hope, control, and self-esteem.

Health Care Util ization

Pulmonary rehabil itation also results in substantial savings inhealth-care costs. Several studies have analyzed COPD patients'hospital and medical resource utilization before and after rehabili-tation (119, 120). l,ertzman and Cherniack (120) reported an av-e.rage decrease of 20 hospital days in the year after pulmonaryrehabilitation, resulting in an estimated savings of $2,000 per pa-tient. ln a randomized controlled study, Jensen (l2l) found thatpulmonary rehabil itation led to significanrly fewer hospitaliza-tions over a 6-mo follow-up of patients with COPD who have"high-risk" markers for psychosocial problems.

Several reports have followed patients for more than I yr af-ter rehabilitation. Hudson and coworkers (122) surveyed hospita-l izations for pulmonary disease over 4 yr after a pulmonaryrehabil itation program for patients with COPD, For the 44 pa-tients sti l l alive at the end of that period, hospital days were re-duced from 529 d in the year preceding the program to 145,270,278, and 207 d in the 4 yr after rehabil irarion. This benefir wasmost striking in the 14 patients hospitalized in the year beforethe program; hospital days in this group decreased from an aver-age of 38 d per parient in the prerehabil iration year ro l0 d inthe vear af rer rehabi l i ta t ion.

Survivalstudies of patients with COPD afrer pulmonary re-habil itation have shown variable resulrs (123-125). There havebeen no prospective randomized controlled studies that provideconclusive evidence regarding survival; reported increased sur-vival may relate to the use of historical controls.

Application, Resources, and OrganizationAn algorithm for the outpatient management of COPD is shownin Figure 3, including pulmonary rehabil iration in the therapeu-t ic schema. Af ter the d iagnosis of COPD and assessment of thepatient's symptoms, the physician should actively encourage ahealthy l ifestyle, including regular exercise, weight control, andsmoking cessation. lmmunizations againsr pneumococci and in-f luenza should be rout inely adminis tered. Assessment of the pa-t ient 's percept ion of h is or her qual i ry of l i fe should fo l low theinst i tu t ion and evaluat ion of response to rhe t radi t ional aspectsof medical rherapy.

Referral ro a comprehensive pulmonary rehabil itation pro-gram is indicared in COPD patients who have been placed on

American Thoracic Society

A rough approximarion is that 1L/min increases the Fts, to 24V0,2 L/min to 28V0, and 3 L/min rc 32V0. These small increasesare usually sufficient to increase the arterial oxygen content toacceptable clinical levels. The actual Fr6, for any'particular pa-tient is variable, depending on the anatomy and parency of thenares and moment-to-moment variation in respiratory rate andpattern, as well as the underlying pathophysiologic process. TheFt6, is inversely related to the inspiratory rate; a more rapid in-spiratory rate dilutes the oxygen flowing into the nares with moreroom air, thereby reducing the Fr6r.

Some studies indicate rhat mourh breathing impairs oxygendelivery, while others show no such reduction (97, 98). It appearsthat some nasal oxygen flows over the top of the i ip into themouth. And most mouth breathers have some nasal airf low aswell. Since only a small nasal inspiratory flow is necessary, nasaloxygen delivery is sri l l beneficial to these patients.

Oxygen-conserving devices funcrion by delivering all of theoxygen during early inhalation. These devices were developedin an effort to improve the portability of oxygen rherapy by reduc-ing the size and weight of the oxygen sysrem. Other advantagesinclude a reduction of overall costs of home oxygen therapy andthe ability to more effectively treat refractory hypoxemia. Thereare three distinct devices; their characteristics are summarizedin Thble 8.

Reservoir cannulas operate by storing oxygen in a small cham-ber during exhalation for subsequenr delivery during early-phaseinhalation; they are driven by the patient's nasal inspiratory andexpiratory pressures. They are avaiiable in two configurarions,oxymizer and pendant (99, 100). The deiivery efficacies of thetwo are roughly equivalent. Compared with continuous flow oxy-gen, reservoir cannulas are two to four times as efficacious. Theyreduce oxygen usage by lowering oxygen flow setting to 25 to5090 of that required for continuous flow oxygen to achieve equiv-alent Saq,. The advanrages of reservoir cannulas l ie in their sim-plicity, reliabil iry, and low cosr. Their disadvantage is thar theyare larger and more noticeable than demand devices.

Demand pulsing oxygen delivery devices deliver a small bo-lus of approximately 10090 oxygen at the onset of inhalation (l0l).Interposed between patient and oxygen source, they serve as sen-sor switches, permitting a bolus of oxygen to flow immediatelyafter sensing the inspiratory effort pressure signal through ei-ther a standard or modified nasal cannula. Because oxygen de-livered during early inhalation reaches the alveoli, small oxygenpulses are very effective in oxygenating the patient. The devices

vary in their design fearures, including delivery strategy, missingbreath alarms, and battery life. Pulse demand devices have alsobeen combined with a transtracheal oxygen catheter, which fur-ther improves the delivery efficacy of transrracheal therapy (102);the overall delivery efficacy of this combination is about equiva-lent to the most efficacious pulsed demand nasal delivery.

Tianstracheal oxygen (TTO) delivery involves insertion of acatheter percutaneously between the second and third trachealinterspaces. Oxygen enters the trachea via this catheter. patientshave been reported to use 37 to 58go less oxygen during TTOdelivery compared with continuous flow nasal oxygen (103-105).

Although TTO is considered in the category of oxgyen-con-serving devices, it is considerably different from the mechanicaltypes discussed earlier. TTO is invasive and requires special train-ing by the physician. The procedure has risks as well as medicalbenefits and is of limited applicability. In addition ro the medi-cal indications and contraindications nored beloq TTO is of dis-tinct benefit to patients who wil l not accept the intrusive natureof the nasal cannula and mechanical oxygen-conserving devicesin the context of social interaction.

Not all parienrs are suitable candidates for TTO. The idealcandidate for TTO has a strong desire to remain active; is willingto foilow the care protocol; is not experiencing frequent exacer-bations; has a caregiver who is willing to assist with problem solv-rng and details of care; and lives within 2 h of the institutionor has equivalent follow-up in the home community. It is impor-tant to note the TTO is instituted in phases and requires activeparticipation by the patient; not every patient is ableand wil l ingto learn the required care.

Relative contraindicarions include high-dose steroids () 30mg,zd) and conditions that predispose to delayed healing, e.g.,diabetes mellitus, connective tissue disease, or severe obesiiy.,4D-solute contraindications include subglottic stenosis or vocal cordparalysis, herniation of the pleura into the insertion site, severecoagulopathy, uncompensated respiratory acidosis, and inabil-Ity to practice self-care.

TTO reduces inspired minute venrilation (106). High flow viaa transtracheal catheter reduces total dead space volumes in anamount proportional to the increase in flow rate. pleural pres-sure-time index and tension-time index for the diaphragm de-crease, changing the respiratory pattern of the diaphragm to aless demanding one (106), which may account for the dicreasein dyspnea and increase in exercise tolerance seen in these pa-tients. TT0 delivery has also led to a decrease in hematocrit, prbb-

CHARACTERISTICS OF

TABLE 8

OXYCEN-CONSERVINC DEVICES

Method Re servor r Demand Pu lse Trans t rachea l

Mechan ism o{ conserva t ion

Sav i ngs

Cosmet ics

Comfor t

Re l lab i l i t y

Cost

Spec i f i c advantages

S t o r e d u r r n g

exha la t ion

2 : 1 t o 4 : l

Obt rus ive

Adequate

Cood; s imp le

Low

Inexpens ive , eas i l yin r t ia ted , re l rab le ,

Bu lky on lace

Ear ly insp i ra to ry

de l i very

3 : 1 t o 7 : l

Adequate

Adequate

Mechan ica l

5 ign i f i can tH ighes t sav ings ,

programmable

de l ivery poss ib lewi th a la rms

Mechan ica l l y

compl ica ted ,{a i lu re poss ib le

Store a t end

exha la t ion ; bypass

oeao space

2 : l t o 3 : ' l

Bes t

CoodMucus p lug poss ib le

S ign i f i can t

Cosmetics, lack olnasa l /ear i r r i ta t ion ,good compliance,

may reduce minutevent i la t ion

Spec ia l care and

t ra rn ing , surg ica lcompl ica t ions

Spec i f i c d isadvantages

AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE VOL 152 1gg5

ably due to increased use of oxygen because of improved com-pliance and less tendency to disconnecr the delivery system (107).

Complications of TTO are infrequent and generally mild. Theyinclude catheter displacement, bacterial cellulit is, subcutaneousemphysema, hemoptysis, severed catheter, and mucus balls. Mu-cus balls can develop on the catheter due to the drying effectof the oxygen, increased sputum production, and poor adher-ence to cleaning schedules; they may cause coughing, catheterblockage, and tracheal obstruction, with serious consequences.Although daily cleaning prevents mucus ball formation in mostpatients, tracheal obstruction has also been reported in patientswho meticulously followed cleaning routines. After 2 mo, thecatheter can be removed for cieaning, and risk of mucus ball for-mation is markedly decreased.

If a patient using TTO requires endotracheal intubation, thecatheter may be capped and left in place in the tract. Irrigationis not necessary, since mucus does not form on the catheter whenoxygen is not f lowing through it. Because of the physiologic ef-fects of catheter flow, TTO delivery may be used to facilitate theweaning of patients from mechanical venti lation.

The overall goal of oxygen-conserving devices is ro improveportabil ity, mobil ity, and comfort. These devices can reduce theoverall cost of oxygen therapy in the home - not by reducing thecost of oxygen itself, which is relatively inexpensive-but byreducing the number of deliveries per monrh to the home. Thecost of the devices themselves ought to be factored in. The puls-ing devices are expensive, but the same device may be used formore than one patient over several years. Thus, the most expen-sive system might be the most cost-effective, after the init ial in-vestment iS made.

Humidi f icat ion

There is no evidence that humidification is necessary when oxy-gen is given by nasal cannula ar f lows ( 5 L,zmin (108). Thereare no differences in subjective complainrs or in severity of symp-toms over time. These findings are explained by rhe low watervapor output of bubble humidifiers and small conrribution ofoxygen flow to the patient's inspired minute venriiation, sincemost of the patient's t idal volume consisrs of atmospheric gas.Moreover, oxygen flowing through the bubble humidifier is atroom temperature; when it is raised to body temperature, therelative humiditl ' falls.

This does not apply to patienrs receiving oxygen by tracheos-lomy or TTO therapy', in whom the upper airway has beenb_vpassed by the catherer; lor rhese parienrs, humidification ofinspired gas is essential even at low now rates (l L/min). TTOpat ients at h igh r isk for mucus bal l format ion, inc l i rd ing thoseu'ith oxygen flows ) 5 L/min, large amounrs of mucus, or weakcough, may'benefit from a servoconrrolled heated humidifier be-cause th is device more ef f ic ient l l ' humidi f ies inspi red gas.

Cl in ica l Guidel ines for Long-term Oxygen Therapy

Pat ients u,hose d isease is srable on a fu l l medical regimen, wirhPae, ( 55 mm Hg (corresponding ro an Sas, ( 8890), shouldreceive long-term oxygen therap) (LTOT). A parienr u,hose Pao,is between 55 and 59 mm Hg (Saor, 89ro) and who exhib i rs s ignsof t issue hypoxia, such as cor pulmonale. erl rhrocl'rosis, edemafrom right heart failure, or impaired menral srarus, should alsoreceive LTOT. Desaturation onl! durin_s exercise or sleep suggesrsconsideration of oxygen therap) specificall l 'under those condi-tions. These guidelines are generall l accepred and have beenadopted by Medicare as reimbursemenr crireria. Some gral,areasremain, such as patienrs rvith adequare Pao. rvho have severe dys-pnea re l ieved b1 ' low-f lou o\ )gen. or par ients u 'ho are l imi tedin their exertional capacity but improre their exercise performancerv i th supplemenral oxvgen. The general i l and of f ic ia l ly recog-nized indicar ions for LTOT (1091 are summarized in Table 9.

Prescribing Home Oxygen TherapyIt is recommended that measurement of resting Pag, after 30min of air breathing, not pulse oximetry (Saor), be the clinicalstandard for initiating LTOT. Oximetry may be used to adjustoxygen flow settings over time. If hypercapnia or acidosis aresuspected, ABG analysis must be performed.

The standard of care for patients with COPD is for them tobe as active and mobile as possible. It is therefore recommendedthat the oxygen system be stationary with a portable component,unless the patient is incapable or unwili ing to be mobile. lf thepatient is immobile and wil l not move beyond a radius of 50 ft,an oxygen concentrator with 50 ft of tubing is suitable.

LTOT is the only therapy lor advanced COPD that has beenshown to decrease mortality (84, 85). For patients who are hypox-emic at rest, the more continuousiy oxygen is.administered, thegreater the benefit. It is recommended that patients be providedoxygen continuously, 24h/d, with an ambulatory capaciry to in-clude oxygen during exerrion.

Settings should be adjusted for rest, exertion, ond sleep tomeet the individual patient's needs. Ideally, the resting flow ratecan be adjusted by monitoring oximetry to Sao2 >- 900/0. Twentyminutes or more may be required for full equilibration to occur.ABG should then be used to establish initial Pae, with corroborat-ing oximetry Sasr; the oxygen flow should be titrated to a levelthat maintains a Pa6, ) 60 mm Hg during rest.

In the Nocturnal Oxygen Therapy Trial (NOTT) (85), oxygenflow rates were arbitrarily increased I L/min above the restingrate for exercise and sleep. Exercise flow rates determined in thismanner may be excessive for some patients and inadequate forothers. If the patient is using an oxygen-conserving system, Pag,should be measured using the patient's oxygen delivery system.This is particularly rrue during exercise conditions. Exercise test-ing can be accomplished via a timed walk, treadmill or birycleergometry, or a free walk at the patient's normal pace. If dys-pnea is especially prevalent during arm exertion, the patient mayalso be tested while performing arm tasks. It is recognized thatpulse oximetry measurements become less accurate during exer-cise, but they do provide useful comparative information. Thegoal is to maintain Pa6, ) 60 mm Hg or Sa6, ) 9090.

When the clinician has adjusted the oxygen delivery settingto achieve ) 9090 Sa6, during seated rest, the delivery settingfor sleep may be turned up I L,/min greater than the resting set-ting for possible sleep desaturation. If there are signs of cor pul-monale despite adequate daytime oxygenation, the patient maybe monitored with oximetry during sleep to determine the bestsleep oxygen setting.

TABLE 9

INDICATIONS FOR LONC-TERM OXYGEN THERAPY

Abso lu te. Pa62 ( 55 mm Hg or 5a9, ( 88%

ln presence o i cor pu lmona le :. Pa62 55-59 mm Hg or 5a6, ) 89%. EKC ev idence o f "P 'pu lmona le , hematocr i t > 55%, congest ive hear t

f a i l u r e

Only in spec i f i c s i tua t rons i. Pao2 > 60 mm Hg or Sa62 ) 90%. Wi th lung d isease and o ther c l in ica l needs , such as s leep apnea w i th

noc turna l desatura t ion no t cor rec ted by CpApl f the pa t ien t meets c r i te r ia a t res t , 02 shou ld a lso be prescr ibed dur ing s leep

ano exerc tse , appropr ia te ly t i t ra ted .l { the pa t ien t i s normoxemic a t res t bu t desatura tes dur ing exerc ise or s leep

(Pae2 ( 55 mm Hg) , 02 shou id be prescr ibed fo r these ind ica t ions .A lso cons ider nasa l CPAP or B iPAp.

Delnilton ol obbreviotons: CPAP = continuous positive airway prersure; EKC = ele(l ro (drdrogrdm; B i IAP = br le re l po \ r t rv€ d fwdy pressure .

optimal medical therap,y- and who: (.1) continue to display severerespiratory symptoms, including dyspnea; (2) have had severalemergency room or hospital admissions per year; (3) exhibit lim-ited functional status, restricting activit ies of daily l iving; or (4)experience impairment in quality of l i fe.

Pulmonary rehabil itation and its therapeutic modalit ies areusually employed on an outpatient basis, and comprehensive pul-monary rehabil itation programs vary in their frequency and du-ration. A national survey of 150 programs showed wide varia-tion in program delivery (126). The optimal frequenry and lengthof such programs have not been determined.

Although inpatient pulmonary rehabil itation programs havebeen described, precise indications for them have not been eluci-dated. Experience suggests, however, that some patients are ex-tremely incapacitated and have such limited abil ity to performeven very low-level activit ies of daily l iving and functional tasksthat simply traveling from home to an outpatient program is pro-hibit ively taxing; such patients cannot derive benefits from out-patient programs.

The individual components of pulmonary rehabil itationshould be provided by health professionais who are knowledge-able and experienced in the management of patients with COPDand in the therapeutic modalities to be employed. Team membersmay include a physician, a dietit ian, a respiratory care practi-t ioner, a nurse, a physical therapist, a cardiorespiratory techni-cian, an occupational therapist, a recreational therapist, a phar-macist, a cleric, an exercise physiologist, and psychosocialprofessionais (psychologists, social workers, psychiatrists). Thespecific members of the team will vary from one program to an-other depending on the availabil ity and interest of individualsin a given locale; in smaller programs, one health professionalmay perform several functions,

A rehabil itation program may include a number of compo-nents and should be tailored to the needs of the individual pa-tient. The various therapeutic modalit ies may be employed sin-gly or in combination, based on the physician's assessment ofthe patient in the physician's office, in other outpatient settings,in the hospital, and/or in the home.

Because the goal of pulmonary rehabil itation is to improvethe patient's function in the home and community setting, thebenefits accrue not only to the patient but also to family mem-bers, particularly those living in the home (l l l). ldeally, all pa-tients who complete a pulmonary rehabil itation program shouldbe given written as well as oral guidelines for home continuationof the program. Ongoing care is generally the responsibil i ty ofthe referring physician. Periodic retesting a.nd reassessment maybe extremely beneficial to the patient and may aiso provide moti-vation for regular chest physiotherapy, practice of breathing tech-niques, and exercise training. Support groups may be useful foreducational, psychosocial, and maintenance purposes.

Because most patients with COPD are older and have a his-tory of cigarette smoking, they are at risk for heart disease. Be-fore any patient with COPD enters a rehabil itation program,cardiac and pulmonary stress testing should be routinely per-formed to exclude silent cardiac disease and assure safety dur-ing exercise training.

Education is important for every patient with COPD who hasbeen placed on a therapeutic regimen, particularly when medi-cations and,/or oxygen have been prescribed. Educational efforrsshould be focused on achieving behavioral change (127) and en-hancing patient adherence to the prescribed therapl'. Ahhoughthe physician's office is the most common setting for educationand the physician is the most frequent provider of parient edu-cation, nurses, respiratory therapists, ph1'sical therapisrs, andother heal th professionals may a lso provide educar ion in ourpa-t ient . hospi ta l , or home set t ings.

AMERICAN 'OURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE VOL 152 I995

Education is a key component of comprehensive pulmonaryrehabil itation (110, l12, 128). Patients with COpD understandtheir disease better after education (129), and simiiar effects havebeen repeatedly demonstrated in patients with asthma. Researchon the efficacy of education as part of pulmonary rehabil itationprograms has been limited.

One group ofinvestigators (117) evaluated exercise and quality-of-iife outcomes in 76 patients with moderate-to-severe COPDrandomized to one of f ive groups: behavior modificarion, cog-nitive modification, cognitive-behavior modificari on, attentioncontrol, and no-treatment control; Patients met with a healthprofessional seven times over a period of 3 mo; the first and lastsessions took place ar the health care facility, with rhe remainderin the patient's home. Behavior modification therapy includeda contract signed by the patient outlining specific times for walk-ing exercise, muscle relaxation training, and breathing retrain-ing. Cognitive therapy included training to replace negative feel-ings, thoughts, and behaviors with more positive attitudes. After3 mo, the three treatment groups increased their treadmill exer-cise tolerance more than the two control groups. Similar improve-ments were demonstrared in the Quality of Well-Being HealthStatus Index and were correlated with exercise compliance andself-efficacy.

Another study found that, after a pulmonary rehabilitationprogram incorporating principles of adult education, patientsdemonstrated an increased ability to recognize and treat symp-toms, including the use of breathing and relaxation exercises (125).An educational program provided without pulmonary rehabili-tation appeared to offer little benefit (130). The educational com-ponent of pulmonary rehabilitation prepares patients and sig-nificant others to be acrively involved in providing care, improvestheir understanding of the disease process, and teaches practicalways of coping with disabling symptoms; this reliance on the pa-tient and supportive others to assume charge of care is knownas collabo rat ive self-mana gement (129).

Recognizing diversiry of learning styles, the educational com-ponent of a rehabilitation program should be custom-tailoredfor the individual parient. Group discussions and didactic presen-tations are generally most effective when they include visual in-formation in the form of videos, siides, or overheads; adjunc-tive printed materials are also helpful, perhaps included in amanual to be referenced at home. Learning does not take placeonly in the formal classroom setting. Every interaction betweenhealth professional and patient should incorporate educationalelements, and education shouid be part of the care of every pa-tient with COPD. This approach need not await the results ofthe further research needed to confirm and define the specificbenefits of education in the comprehensive pulmonary rehabil i-tation program.

Exercise Training

There is considerable evidence documenting favorable responsesto general aerobic exercise rraining in patients with chronic lungdisease (119). Benefits are borh physiologic and psychological.I n fact, patients experience increased capacity and endurance forexercise and physical acrivit5' after rehabil itation, even thoughlung function may remain unchanged. Exercise training also pro-vides an ideal opportunity for patients to learn their capacityfor physical work and to practice and use methods for control-l ing dyspnea (e.g., brearhing retraining techniques taught at thesame t lme).

Many studies have reported benefits of exercise conditioningin patients with COPD. Casaburi (l l9) reviewed 37 publishedstudies of such training, incorporating a variery of regimens inmore than 900 patients. Nearll 'unanimously, these studies dem-onstrated improvement in exercise endurance and/or maximum

American Thoracic Society

exercise tolerance. These conclusions are supported bv severalcontro l led r r ia ls documenr ing the benef ic ia l e f fecrs of exerc isetra in ing in pat ienrs wi th chronic lung d isease (114, j20, l3 l ) . S ig_nificantly more parients in the exercise groups noted improvi_ment in sense of well-being and decreased breathlessness, cough,and sputum. Significanr increases in exercise performance hivealso been seen on both l2-min walk chal lenges and maximumcycle ergomeler tests in trained patients compared with controlsu b jecrs.

The basic elements of the exercise prescription are mode, fre_quency, intensirl ' , and durarion. ln general, the guidelines for ex_ercise prescription in parients rvirh COpD are adapted from thoseestablished for healthy individuals. The program should incor_porare activit ies that involve large-muscle groups and closely re_late to daily acrivit ies, e.g., walking and cycling. Exercise ,.rslon,should take place 3 to 5 d/wk. Intensiry and duration are inter_related. Various regimens have been recommended. based on ex_perience in healthy individuals and published srudies of patientswith COPD. Experience generally indicates that a range of exer_cise intensities can be selected based on the patient's motivation,tolerable levels of dyspnea and discomfort, and whether the ses_sions are supervised. Widely differing regimens have been shownto be beneficial; further studies to establish oprimal euidelinesfor patienrs wirh COPD would be welcome,

Exercise training is a mandatory component of pulmonaryrehabilitation. Patients with COpD should regularly perform aei-obic lower-extremity endurance exercises to enhance performanceof daily activit ies and reduce dyspnea. Resisrance training is par_ticularly important when weakness of specific muscle groupJhasbeen demonstrated.

.. Upper extemity training. Many patients with COpD reporrdisabling dyspnea associated with daily acrivit ies involving theupper extremities, e.g., l i ft ing or grooming, at work levels muchlower than for the legs. This is because upper extremity exerciseis accompanied by a higher ventilatory demand for a given levelof work than lower extremity exercise. Since exercise benefits aregenerally specific to rhe muscles and tasks involved in training,basic. training primarily involving the lower extremities -uy nJtalleviate upper extremity difficulties. Special upper exrremiry ex_ercises may be important in addressing this problem.

Arm exercise endurance is significantly less than enduranceLo.].g exercise in many patients wirh severe COPD and may belimited by dyspnea accompanied by d-vsynchronous rhoracoab-dominal breathing (l l5). Simpll 'raising rhe arms creares increasedmetabolic demand (increased oxygen consumption, minute ven_tilation, and diaphragm work). Merabolic and ventilaroiy require_ments for arm exercise decrease after comprehensive Dulmonarvrehabi l i ra t ion inc luding arm exerc ise 032i . Hou,ever . no s ieni f i_cant changes on tests of venti larory muscle performance o, i i-u_lated activit ies of daily l iving have been sirou,n afrer arm exer_cise (133) .

- Upper extremity exercise trainin_e should be considered a parrof pulmonary rehabil irarion. The reducrion in oxygen consump_tion during arm acrivir) ' thar resulrs from such riainin_e migirrbe expected to improve dyspnea and a l lou. increased acr iv i r iesof daily l iving requiring use of rhe upper exrremities; furrher in_vestigations are needed to confirm rhese benefirs.

Inspiratory muscle training. The porenrial role of r,enrilarorl,muscle weakness and possible farigue as a cause of venti laror!.l imitation and respirarorl, failure in parienrs u.irh COpD hasst imulated at tempts to t ra in the venr i l i rorv muscles. \ /ar ious in-spiratory muscle trainin_e rechniques havs fsgn shown ro improvethe function of these muscles (134). lmprovemenl in eeneral ex_erc ise per formance af ter t ra in in_e of onl r rhe renr i i i ror ) . mus-cles, however, has not been consistenrlr. dsrn6n5lrared _ perhapsdue to lack of an appropr iare r ra in ing sr imulus In man).srudies.

Selected trials establishing a resistive intensity and/or con_trolling breathing straregy during training have demonstrated bothphysiologic and symptomatic improvemint (134). Decreased dys_pnea has been demonstrated after venti latory muscle traininjinpatienrs with COpD compared with untreated control subjJcts(135). One group of investigators has shown improved sympr;ms,decreased dyspnea, and improved abil ity to perform-daily ac_tivit ies when inspiratory resistive loading training was added rothe pulmonary rehabil itation regimen (136).

At present, there is no simple way ro select patients most l ikelyto benefir from inspiratory muscle training. Moreover, there hasbeen no benefit demonstrated from such training compared withaerobic lower extremity exercise conditioning. Thus, althoughsuch training may result in improved ventilaiory muscle funi_tron andlor alleviation of symptoms for some patients, it is notyet clear that this modality is appropriate for routine use in re_habil itation. Additional research is needed to guide selection ofpatients who would be expected to benefit and to define the as_sociation of improved ventilatory muscle performance with im-provement in exercise tolerance, daily activities, and other quality_of-l ife areas.

Psychosocial SupportMany patients experience anxiety, depression, fatigue, copingdifficulties, and somatic preoccupation (123), symptoms tirathaybe related ro the dyspnea associated with COpD. Inability to wori<and decreased capacity to participate in social, recreational, andsexual activities are common and often lead to depression. Neu_ropsychological deficits related to chronic hypoxemia may oc_cur. Patients with chronic hypoxemia also have decreased abil_ity to acquire and retain new information, form new concepts,think flexibly, perform complex perceptual-motor maneuvers,and engage in perceptual discrimination.

Psychosocial intervention and support are considered impor-tant components of pulmonary rehabilitation. Since patients withgOlD are frequently depressed, a careful evaluation of the pa-tient's psychological status should be performed and the usj ofantidepressant medications considered in patients with severedepression. ln addition to the informal or iormal support thatpatients may receive from others with similar conditions, socialsupport is facil i tated by psychosocial sessions with a focus onstress management techniques, including relaxation strategies andcognitive approaches to coping with emotional changes. Groupsessions incorporating spouses and.other familv members mavbe beneficial

Psl,chosocial interventions may also be directed toward model_tng appropriare methods for the patient to enlist the help of familyand friends, while helping the patienr to maintain realistic short_term and long-term goals, especially in the context of family andsocial relationships. Issues involving sex and sexuality should beexplored and discussed (137). Sexual counseling may be recom_mended when indicated, using the pLISSIT itepwise model,which includes permission-giving (p), l imited infoimation (LI),specific suggesrions (SS), and inrensive therapy (lT).

Breathing RetrainingDiminishing of dyspnea has been a consisrenr benefit of pulmo_nary rehabil itarion, possibly due in parr to the breathing ietrain_tng trequenrly included in rehabil itation programs. The goal ishelping the patient to relieve and control br.luthlrrrn.r, and rocounteract such physiologic abnormalit ies as hyperinflationrelated to chronic airf low obstruction.. Retraining.techniques include diaphragmatic and pursed_lipsbreathing to improve the ventilato.y pati..n (i.e, slow respira_tory rate and increase tidal volume), prevent dynamic airway com_

AMERICAN 'OURNAL OF RESPIRATORY AND CRITICAT CARE MEDICINE VOL 152 , I 995

pression, improve respiratory synchrony of abdominal and tho-racic musculature, and improve gas exchange. Review of clinicalstudies evaluating the effects of such techniques indicates thatimprovement in clinical symptoms is a more consistent f indingthan any measurable impact on physiologic parameters. The mostconsistent physiologic change observed has been an increase intidal volume and a decrease in respiratory rate; improved oxy-genation may also occur.

Although many patients with COPD "discover" pursed-lipsbreathing on their own, specific instructions can be given to de-crease dyspnea, as follows: breathe in slowly and deeply throughthe nose; purse the l ips I ight ly , as i f to whist le ; then breathe outslowly through the pursed lips, taking'twice as long to exhaleas to inhale. Leaning forward, with arms resting on the patient'sthighs or on a table or other hard surface, may also help somepatients in relieving breathlessness. Diaphragmatic breathing isoften taught along with pursed-lips breathing; the patient learnsto coordinate expansion of the abdominal wall with inspiration.The value of diaphragmatic breathing has been questioned bystudies documenting worsened thoracoabdominal dyssynchronywi th th is breath ing pat tern (138) .

ElLctive Ventilatory Assistance

In patients with severe COPD, the standard medical therapy of-ten does not reverse hypercapnia or alieviate the severe dyspnea,diminished stamina, and limited quality of l i fe experienced bythese patients. They have an increased mortality rate and are atrisk for hospitalization with respiratory acidosis, porentiallynecessitating mechanical venti lation to reduce Paco:.

Although there have been numerous studies evaluating theefficary of elective mechanical ventilation in patients with COPD,the results have been variable and the conclusions controversial(139). Decreased respiratory muscle acrivity has been demon-strated with use of a cuirass, pneumowrap, and ventilation de-livered per nasal mask. Thus, it appears that mechanical venti la-tion can reduce respiratory muscle activity. In addition, short-termuse of mechanical venti lation (l to 7 d) has been shown to re-duce Pa66, in patients with chronic hypercapnia due to severeCOPD. There is no relationship, however, between the degreeof respiratory muscle rest achieved with mechanical venti lationand the reduction in Pacor.

Respiratory muscle strength has improved in some srudies af-ter elective mechanical venti lat. ion but not in others, and respi-ratory muscle endurance may also improve. But improved respi-ratory muscle function may not be the mechanism responsiblefor reduction in Pasgr; it seems more l ikely that the reductionis related to the increased ventilatory response to inhaled CO,that occurs wi th mechanical vent i la t ion (140) .

Despite the observation of physiologic benefits, it has not beendemonstrated that this therapy has any impacr on parienr symp-toms. Shapiro and colleagues (l4l) randomly assigned 184 pa-tients with COPD to receive active or sham negative pressure ven-t i la t ion at home for l2 wk. These invest igators d id nor reporrany diminution in dyspnea or improvement in exercise capacityas measured by a 6-min walk rest, even in patients u,irh moresevere hypercapnia. But only 29Vo of the parienrs had a 5090 re-duction in diaphragm electromyogram acrivir)! and patients usedthe ventilator for very l imired periods of t ime during rhe da1'.

Routine elective use of venti latory support in ambulatory pa-iients with COPD with hypercapnia cannor be recommended atth is t ime. Addi t ional invesr igar ions are necessar l , to dererminethe potent ia l ro le of th is therapy in COPD. Such srudies shoulduse less cumbersome methods of venr i la t ion, such as a nasalmask; should def ine which par ienrs, i f an1. mi_ehr benef i r f romelect ive venr i la t ion; and should evaluare porenr ia l benef i rs ond1'spnea, exerc ise capaci ty , and qual i r i o f l i fe as uel l as ph1's io-logic measures.

Nonelective Mechanical Ventilation

Patients with COPD who require emergency institution of me-chanical ventilation for acute exacerbation may require more thanshort-term ventilatory assistance. Such patients may not be eas-ily weaned from support because of the severity of the underly-ing disease and associated altered reslriratory mechanics. In suchcases, a tracheostomy may be required to assure long-term air-way access. There is no simple, reliable way to predict inabil ityto wean from the ventilator, nor is there a precise definit ion ofchronic ventilator-dependency. In most cases, failure to achievesufficient 24-h ventilation without venti latory support is a con-clusion that can be reached only after multiple weaning trialsin a crit ical care unit. Thus, chronic ventilator dependency ina patient with COPD might be defined as rhe inabiliry to achievespontaneous ventilation 24h a day after repeated unsuccessfulattempts to remove the patient from the ventilator. For the bestchance of weaning success, attempts should take place when thepatient is in optimal medical condition.

Long-term management of patients who require chronic me-chanical ventilation raises several health care issues, includinghigh costs, location of appropriate sites for continued care, avail-ability of resources and personnel for that care, and prospectsfor improving quality of life. The cost of chronic care for a ven-ti lator-assisted patient with COPD in a crit ical care unit is veryhigh, in terms of both possible use of expensive and limited re-sources for a relatively stable patient (albeit a patient dependenton high technology) and financial cost to patient and insurer.

Potential sites for continued care of ventilator-assisted indi-viduals include short-term care in specialized but nonintensivecare units in acute hospitals and in general hospital medical-surgical areas. long-term care may be available in chronic hospi-tals, skilled nursing facilities, group living facilities, and thepatient's own home However, there appear to be inadequate num-bers of long-term beds in medical facilities with appropriate re-sources to care for these individuals. The most appropriate sitefor long-term care should be carefully chosen for each patientand depends on the patient's respiratory and personal care re-quirements (e.g., oxygen, suctioning, ventilation, toileting, bath-ing) and available resources (e.g., family members, skil led nurs-ing, respiratory therapy, electrical power and outlets, insuranceand financial resources) to meet those needs (142).

Home Care

COPD is the third most frequent medical diagnosis (after con-gestive heart failure and stroke) for patients receiving home care.Home care for venti lator-assisted individuals with COPD is ef-fective in selected parienrs (143), and they may be re latively inde-pendent in the home. The life expectancy of patients with COPDin the home is about 5090 at 3 yr, and these patients may requirerehospitalization for either respiratory or other problems.

The cost-effectiveness of home care for these patients wasdocumented in a recenr study (144). Seventeen patients with se-vere COPD received care from a multi-disciplinary team in ahospital-based home care program. Use of health care resourcesbefore and for ar least 6 mo after entry into the program showedsignificant savings of $328 per patient per month due to decreasedhospitalizations and emergency room visits.

These findings contrast with results of a randomized controlledstudy of 300 patients with COPD assigned to receive care forI yr using respiratory home care nurses, standard home care, oroffice care (145). Average annual health care costs for the twohome care groups were higher than those for the office care group.

Patients for the study demonstrating reduced health care costswere selected because of their high risk, based on prior hospitali-zations (laa). This was not a criterion for patient selection inthe prospective rhndomized study that did not demonstrate re-duced care costs. Thus, further research is needed.

Amer ican Thorac ic Soc ie ty

Inpatient Management of COPDAlthough an acule exacerbarion of COpD is diff icult ro defineand its pathogenesis is poorly understood, impaired lung func-tion can lead to respiratory failure requiring intubation andmechanical venti lation. Accurate diagnosis in a patient expe-riencing rapid deterioration of respiratory function may be con-founded by underlying myocardial ischemia, congestive heart fail-ure, thromboemboli, or recurrent aspiration, which can simulatean exacerbat ion of a i rway d isease.

EMERGENCY EVALUATION

The acute exacerbation frequently prompts patient evaluationand init iation of treatment regimens in an office or emergencydepartment. The key components of the history, physical exami-nation, and laboratory evaluarion that should be obtained dur-ing a moderate-to-severe acute exacerbation to assist the formu-lation of therapy and the decision for hospiral admission are listedin Thble 10.

Although acute exacerbarions of COPD represent a majorcause of hospital admissions in rhe United States, indicarionsfor hospitalization and the purpose of the hospital stay have re-ceived litt le attenrion. Traditionally', the decision to admit de-rives from subjective interpretation of clinical features, such asthe severity of dyspnea, determination of respiratory failure, short-term response to emergency therapeutic efforts, degree of corpulmonale, and the presence of complicating features, such assevere bronchitis, pneurironia, or other co-morbid conditions.This approach to decision-making is less than ideal, since up to28s/o of patients with an acure exacerbation of COPD who aredischarged from an emergency room (ER) have recurrent symp-toms within 14 d (146). Additionally, l i Vo of parienrs dischargedafter ER managemenr oi COPD will relapse and require hospi-talization (147).

Data are also lacking regarding rhe potential overuse of hos-pitalization in parients u,ith acute exacerbarions of COpD whomight have responded to more aggressive outpatient therapy.Furthermore, there is no accepred standard for deciding u'hichsurgical or diagnostic procedures in srable parienrs with COpDrequire hospitalization and which could be performed on an our-patient basis.

In shor t , there are present l ) 'no guidel ines for c l in ic ians inselect ing pat ients for hospi ta l admiss ion on rhe basis of medicalneed. The Diagnosis-Relared Group (DRG) s)'srem uses an anal-ys is of h is tor ic exper ience of s imi lar par ienrs rarher rhan par ient-specific prognosis-modifying fealures. To derermine rhe neces-sity for admission, Medicare parienr revieu' pracrices of rhe HealrhServ ices Advisory Group, lnc. , use cr i rer ia rhar have nor beenprospectively validated.

Relatively few clinical srudies have invesr igared parienr-speci f ic,objective clinical and laborarorl fearures idenrifying parienrs rvirhCOPD who requi re hospi ra l izar ion. General consensus supporrsthe need for hospitalizarion in parienrs r,r. irh severe. acure h1'p-oxemia or acute hypercarbia; hou,ever, Iess exrreme arrerial bloodgas abnormalit ies do nor assisr in the decision Drocess 048). l\tosrstudies us ing spi rometr ic va lues (FE\ ' , . FE\ ' , as a percenrage ofbasel ine, FEV, as a percenrage of predicred. or change in FE\ ' ,after emergency treatment) f ind a poor correlarion rvilh successof ER discharge (146, 141, 149) .

Murata and coworkers (149) recenr l l r 'a l idared in male pa-t ients at the Veterans Healrh Adminisrrar ion \ led ical Cenrer amul t ivar iate model thar idenr i f ied par ienrs u i rh decompensaredCOPD who had h igh and lou r isks of re lapsrng af rer ER manage-ment . Factors ident i fy ing h igh-r isk par ienrs i : r - - lu<jed a prev iousemergency v is i t wi th in I rvk, the numbe: o: doser ot -nebul ized

TABLE IO

EMERGENCY ROOM EVALUATION OF PATIENT WITH ACUTEEXACERBATION OF COPD: KEY COMPONENTS

H rs to ry :. Base l ine resp i ra to ry s ta tus. spu tum vo lunre and charac ter is t i cs. Dura t ion and progress ion o f symptoms. Dyspnea sever i tyo Exerc ise l im i ta t ions. 5 leep and ea t ing d i f f i cu l t ies. Home care resources. Home therapeut ic reg imen. Symptoms o f co-morb id acu te o r chron ic cond i t ions

Phys ica l examinat ion , ev idence o f :. C o r p u l m o n a l e. Eronchospasmo Pneumonia. Hemodynamic ins tab i l i t y. A l te red menta t ion. Paradox ica l abdomina l re t rac t ions. Use o f accessory resp i ra to ry musc les. Acute co-morb id cond i t ions

Labora tory , usua l ly inc ludes :. A B C. Chest rad iograph (year ly , la te ra l ). E K C. Theophy l l ine leve l , i f ou tpa t ien t theophy l l ine useda n d m a y i n c l u d e :o Pu lse ox imet ry mon i to r ing. EKC moni to r ing

. . Add i t iona l s tud ies as ind ica ted (e .g . , lung scan, leg venous th rombos iss tud ies)

bronchodilators required in the ER, use of home oxygen, previ-ous relapse rate, administration of aminophyll ine, and use ofcorticosteroids and antibiotics ar the time of ER discharge (149).This study requires validation in women as well as men caredfor in other institutions. In addition, the low numbers of studypatients in this and other published investigations do not allowsubgroup analysis ro determine whether hospitalization indica-tions differ for patients with chronic bronchitis and those withemphysema. It is possible that the heterogeneity of COpD andthe variety of confounding coexisting conditions may limit thevalue of hospitalization predicrion indices in individual patients.

Prediction of the required duration of hospitalization is fur-ther complicated by the numerous conditions that can compoundrespiratory dysfunction. They include bronchospasm, pulmonaryinfection, level of airrvay secretions, inspiratory muscle fatigue,malnutrit ion, hypoxemia, systemic acidosis, pulmonary embo-lism, heart failure, and metabolic derangements. It is not sur-prising, therefore, thar no correlation exists berween the lengthsof hospiral sray rhar were medically required and expected lengthsof stal ' per the assigned DRG in parienrs with acute COpD ex-acerbal ions (150) . The actual median length of s tay requi red inpat ient ser ies lo produce maximum benef i r is 6 d, u, i th a range ofI to 57 d (150) , which is longer than a l lowed by rhe DRG f inan-c ia l incenr ives to at ra in shor ter hospi ta l izat ion.

HOSPITALIZATION CRITERIA

The purpose of hospi ta l izat ion is to manage the pat ient 's acuteoecompensation and co-morbid conditions to prevent furtherdeterioration and readmission, which occurs in iaio of patientswi th in 2 wk of hospi ra l izat ion (149) . Addir ional ly , pat ients re-quire education as to the nature of their disease, correcr use ofmedications, activity modifications, and awareness of appropriatc

AMERICAN IOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE VOL 152 I995

TABLE 'I

1

INDICATIONS FOR HOSPITALIZATION OF PATIENTS WITH COPDaction in case of future emergency. Plans for clinical reassess-ment of drug regimens, use of home oxygen, or potential benefitsfrom pulmonary rehabil itation programs should be prepared.lnnger hospital stays are required in patients with worse base-line lung function and co-morbid conditions (150). Althoughbased on limited data, emerging patient profiles may allow predic-tion of duration of hospitalization when confirmed in largerstudies.

Duration of hospitalization in COPD depends at least par-tially on the presence of a multidisciplinary team directing re-spiratory management. Because of the complex management is-sues in caring for patients who have COPD with impending orfrank respiratory failure, physician speciaiists with extensive ex-perience in and knowledge of COPD should participate in thecare of hospitalized patients who present with underlying stageII or III disease, require mechanical venti lation, develop hypox-emia unresponsive to Fto, of 0.50 or new-onset hypercarbia, re-quire steroids for more than 48 h to maintain adequate respira-tory function, undergo thoracoabdominal surgery, or requirespecialized techniques to manage copious airway secretions. Thelength of hospitalization may be shortened with early dischargeplanning and development of alternate sites of care when pa-tients are being observed during the final stages of pulmonaryimprovement. A patient care coordinator can direct modifica-tions of the care plan and organize a system of home evaluationto prevent the need for early rehospitalizarion.

Guidel ines

Expert consensus identif ies indications for hospital admissionthat consider the severity of the underlying respiratory dysfunc-tion, progression of symptoms, response to outpatient therapies,existence of co-morbid conditions, necessity of surgical inter-ventions that may affect pulmonary function, and the availabil-ity of adequate home care. That consensus is reflected in Tablell. The severity of respiratory dysfunction dictates the furtherneed for admission to an intermediate care unit (ICU) (Table l2).Depending on the resources available within an institution, ad-mission of patients with severe COPD exacerbations to inter-mediate or special respiratory care units may be appropriate ifpersonnel, skil ls, and equipment exist to identify and manageacute respiratory failure successfully.

Insufficient clinical data exist to establish the duration of hos-pitalization in individual patients to achieve maximal benefit.Consensus and limited data support the discharge criteria l istedin Thble 13. Emerging data suggest that prolonged hospitaliza-tion can be avoided and good clinical outcome assured, even inpatients with severe disease, when effective discharge planningis combined with coordination of multidisciplinary care and tem-porary alternative placement, if needed. The complexity of man-agement issues requires that certain patients u'ith moderate-to-severe acute exacerbations be evaluated by a specialist trainedin the evaluation and care of patients with COPD.

Because existing guidelines lack adequate research support,additional studies are required to determine precisely' * 'hich pa-tients wirh COPD are l ikely to benefit from admission, as *ellas the ideal duration of hospitalization.

PHARMACOTHERAPY OF COPD EXACERBATIONS

Inpatient management usually begins in the emergenc)' room.Drug therapy wil l be tailored to the follou'ing variables: (/) thedegree of reversible bronchospasm; (2) prior therap)'of the pa-tient when stable; (3) recent drug usage and evidence of poten-tial toxicity; (4) the abil ity of the patient to cooperare in rakinginhaled medications; (-i) the presence of contraindications ro an1'drug or class of drugs; (6) specific causes or complications relatedlo the exacerbat ion.

2

3

Pat ien t has acu te exacerbat ion charac ter ized by inc reased dyspnea, cough,

or sputum produc t ion , p lus one or more o{ the fo l low ing :. Inadequate response o f symptoms to ou tpa t ien t management. tnab i l i t y to wa lk be tween rooms (pa t ien t p rev ious ly mob i le ). Inab i l i t y to ea t o r s leep due to dyspnea. Conc lus ion by fami ly and/or phys ic ian tha t pa t ien t cannot manage

at home, w i th supp lementary home care resource5 no t immedia te ly

ava i la b le. H igh- r i sk co-morb id cond i t ion , pu lmonary (e .9 . , pneumonia) o r non-

pu lmonaryr Pro longed, p rogress ive symptoms before emergency v i5 i t. A l te red menta t ion. worsen ing hypoxemia. New or worsen ing hypercarb ia

Pat ien t has new or worsen ing cor pu lmona le unrespons ive to ou tpa t ien t

mana9ementP lanned invas ive surg ica l o r d iagnos t ic p rocedure requ i res ana lges ics o r

sedat ives tha t may worsen pu lmonary func t ion

Co-morb id cond i t ion , e .9 . , severe s te ro id myopathy or acu te ver tebra l com-press ion f rac tu res , has worsened pu lmonary func t ion

Other ind ica t ions lo r hosp i ta l i za t ion may app ly to pa t ien ts undergo ing pu lmonary re -hab i l i ta t ion (see Pu lmonary Rehab i l i ta t ion) .

Condition and Considerations

Exacerbations may be provoked by inhalation of environmen-tal irritants, gastroesophageal reflux or aspiration, viral andbacterial infections, intercurrent illnesses leading to debility orweakness, or cardiopulmonary events, such as heart failure, ar-rhythmias, or pulmonary emboli. Increased work of breathing,ineffective cough, mucostasis, progressive hypoxemia/hypercar-bia, confusion, and fatigue may be associated with im.paired useof drug therapy and resulting overmedication or undermedica-tion. In an exacerbation, underlying general problems, such asnausea and progressive weakness, as well as specific abnormali-ties, such as pneumonia or fluid and electrolyte imbalance, mustbe addressed.

Exacerbations in COPD may cause nonspecific complaints,such as malaise, insomnia or sleepiness, fatigue, and depression,or more specific difficulties, such as increased dyspnea, produc-tive cough with altered sputum, fever, leg edema, or obtunda-tion. Management is facilitated if the patient's history, findings,and drug regimen are known; lack of this information may justifymore vigorous intervention, including early endotracheal intu-bation and ventilatory support. The physician needs to deter-mine the recent pattern of drug use and the possible overuse ormisuse of drugs such as theophyll ine, antibiotics, or sedatives.lndividual judgment is needed to determine if the patient shouldbe admitted to the hospital or to the ICU. lnabil ity of the pa-tient to cooperate in taking aerosol medications and the findingof marked impairment in vital signs are clear indications for ICUmanagement. Init ial choices in drugs may have to be empiric,but it is usually advisable to follow a step-care plan in the firstfew hours of therapy (151, 152).

TABLE 12

INDICATTONS FOR ICU ADMISSION OF PATIENTSWITH ACUTE COPD EXACEREATION

Severe dyspnea tha t responds inadequate ly to in i t ia l emergency therapy

Con lus ion , le thargy , o r resp i ra to ry musc le {a t igue ( the las t charac ter ized by

paradox ica l d iaphragmat ic mot ion)

Pers is ten t o r worsen ing hypoxemia desp i te supp lementa l oxygen or

severe /worsen ing resp i ra to ry ac idos is (pH < 7 .30)

Ass is ted m€chan ica l ven t i la t ion is requ i red . whether by means o f endo '

t rachea l tube or non invas ive techn ique

12

3

American Thoracic Society

TAELE I 3

DISCHARCE CRITERIA FOR PATIENTS WITHACUTE EXACERBATIONS OF COPD

. lnha led be ta-agon is t therapy i ! requ i red no more f requent ly than every 4 h

. Pat ien t , i t p rev iou5 ly ambula tory , i s ab le to wa lk across room

. Pat ien t i s ab le to ea t and s leep w i thout f requent awaken ing by dyspnea

. React ive a i rway d isease, i f p resent , i s under s tab le cont ro ,

. Patient has been clinically 5table, off parenteral thenpy, fo( 12-24 hr ABC has bee i r s tab le lo r 12-24 h. Pat ien t (o r home careg iver ) fu l l y unders tands cor rec t use o f med ica t ions. Fo l low-up and home care ar rangements have been comple ted (e .9 . , v is i t ing

nurse , oxygen de l i very , mea l p rov is ions). Pat ien t , fami ly , and phys ic ian are conf ident pa t ien t can manage success fu l l y

A pa t ien t who does no t fu l l y meet c r i te r ia to r home d ischarge may be cons ideredfor d ischarge to a nonacute care fac i l i t y fo r observa t ion dur inq the f ina l reso lu t ion o fsymptoms.

Most patients with COPD exacerbations respond to anti-cholinergic aerosols and beta-agonists, and adequate dosages ofthese drugs are essential components of inpatient care. Theoph-yll ine may be given as a primary drug or in addition to otherbronchodilators; dosage must be carefully calculated and exces-sive dosing avoided. Combination bronchodilator therapy canbe beneficial in severe disease. Systemic corticosteriods may beadded to the regimen if the patient is very i l l , shows an inade-quate response to bronchodilators, or has previously respondedto steroids. A change in sputum color or consistency during anexacerbation of COPD is a standard indication for antibiotictherapy; mucokinetic drugs may also be given, although theirvalue has not been firmly established in COPD. Oxygen is gener-ally needed when the Pa6, is less than 60 mm Hg, particularlyif there is cardiac irritabil i ty. Any additional problem, such asheart failure or renal failure, must be taken into account, andtreatment must be modified appropriately.

Cl in ica l Guidel ines

Stepwise therapy is recommended, with initial emphasis on drugsthat can be expected to produce a rapid response.

l. Identify cause of exacerbation, e.g., infection or sedativedrug narcosis, and direct specific therapy accordingly.

2. A beta'-agonist aerosol is usually given as the first step inmanagement of acute, severe COPD. Patients can be treated withan inhalant solution given by nebulization, or a metered doseinhaler (MDI) with a spacer can be used. These drugs have areduced functional half-life in exacerbations of COPD and there-fore they may be given every 30 to 60 min if tolerated. The safetyand value of continuous nebulization have not been establishedin COPD (153, 154). Long-acting aerosols are currently not ap-proved for use in exacerbations of COPD. Subcutaneous dosingof favorable agents is recommended only if aerosol deiivery isnot feasible; intravenous administration is not acceptable prac-tice at present, although it is favored in some countries whenaerosol therapy is not possible in a l ife-threatening situation.Dosages of sympathomimetic bronchodilators are summarizedin Table 14.

3. Anticholinergic aerosols have not been adequately assessedas a first step in COPD, although they are often favored whenthe history indicates poor responsiveness to beta-agonists. Thechosen agent is given using an MDI with a spacer or as an in-halant solution by nebulization. Ipratropium is usually employed;although an upper dosage limit has not been established, the drugis generally well-tolerated, and higher dosages than usual canbe given to a poorly responsive patient (155). The prolonged half-life, however, means that repeat doses need not usually be givenmore often than every 4 to 8 h.

4. Combination therapy with a betar-agonist and ipratropiumcan be employed (156, 157), although their combined use has notbeen clearly demonstrated to be more effective than larger dosesof each. There is evidence to suggest that they may act syner-gistically, however, and there is no increase in adverse effects withcombined usage. After the patient has improved, lower dosagesof ipratropium and the selected beta-agonist may be given lessfrequently. Any synergistic effect may not be noted after the pa-tient stabil izes (158), and one of the two drugs may be discon-tinued, particularly if side effects are a problem (159).

5. Theophylline can be added if aerosol rherapy cannot begiven or proves inadequate. The drug can be given as intrave-nous aminophylline in a severe exacerbation; suggested dosagesare noted in Thble 15. Use serum levels as a guide to avoid haz-

TABLE 14

sYMPATHOMIMETIC BRONCHODILATOR DOSAGES

M D I

Dose MDI Puf fsStandard Doses o fInha lan t So lu t ions '

Equivalent MDIPufts f rom

Inhalant Dose'

Epi neph r i ne(Adrena l in , Pr ima-

tene, € l o / . )l sopro tereno l

( l supre l , Med iha le r - tso)

lsoe tha r i ne(Bronkometer , Bron-koso l , Arm-a-Med, e t o l . )

Metapro tereno l(A lupent , Metapre l )

Terbuta l ine(Bre tha i re , Bre th ine) i

A lbu tero l(Provent i l , Vento l in )

P i rbu tero l (Maxa i r )

B i to l te ro l (Torna la te )

Sa lmetero l (5erevent )

0 . 1 6 - 0 . 1 5 m g

0.08-0.1 3 mg

0 .3a mg

0.65 mg

0.20 mg

009 mg

0.20 mg0 .37 mg0.02 m9

2-4 evety 4-6 h

2-4 eve(y 4-6 h

' l -2 every 4 h

l -2 every 3 -4 h

2 every 4-6 h

1-2 every 4 -6 h

'l-2 every 4-6 h

2 everv 4-8 h

2 every 12 h

0 .25 -0 .5 m l ( ' 1 . 25 -11 mg)

0 .25 -0 .5 m l ( 1 .25 -2 .5 mg )

0 .25 - l m l ( 2 .5 -10 mg) tor ' l .25-5 ml (2.5-5 mg)T

0 .2 -0 .3 m l ( 10 -15 m9) to r 2 -5 m l ( 10 mg l to r 2 -5 m l ( 15 mg) t

0.25-0.5 ml (0.25-0.5 m9)

0 .5 m l ( 2 .5 mg ) to r 3 .0 m l ( 2 .5 m91 t

0.25- l ml (0.5-2 mg)

5-70

1 0-30

7 -30

| ) - / 5

2 7

' U n i t d o s e i t s t a n d a r d d o l e r e c o m m e n o e o l o . a d u l t s A m o u n l r s u s u a l l y a v a i l a b l e i n 0 - 2 5 - 2 . 5 m l , a n d p e r c e n t a g e d e l i v e r e d t o l u n g 5d e P e n d s o n b o l h n e b u l r z e . a n d D r e a l h r n o r e ( h n t q u e u i e d

I D o ! a o e d e o e n d s o n D r a n c+ M a r k e t e d f o r : u b ( u t a n e o u t u r r

AMERICAN IOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE VOL 152 I995

Long Term Oxygen Therapy for the COPD Patient

Hypoxemia

:3:-:-'-lT 3:-1L?-' -11-PaO, 55-59 mm Hg with

Cor Pulmonale, Erythrocytosis

Detected as \

Y/7R""ou"ring frot\. Acute Exacerbation

on optimal medical management

Prescribe OxygenGoal: PaO, = 60-65 mm Hg

Rest -S leep-Exer t ion

Titrate liter-flowto goal at rest

Exertion: addl Um

Titrate exercise liter-fl owgoal: SaO, > 90%Titrate sleep liter-flow

8 hr sleep SaO, > 90% Sleep: add 1 Um

Recheck ABG for PaO, > 60 mm Hg in 1 to 3 monthsif hypoxemia developed during an acute exacerbation;

periodically if hypoxemia discovered as stable outpatient

Adjustoxygen PaOr > 60 mm Hg

lf stable outpatienttitrate liter-flow

to assurePO2 > 60 mm Hg

lf recoverying fromacute exacerbation

titrate liter-flowor discontinue

ardous overdosing. In most patients, a theophyll ine level of 8to 12 1tg/ml is appropriate; occasionally patients benefit fromhigher levels, up to l8 to 20 1tg/ml. When the patient improves,oral long-acting theophyll ine can be substituted, using 8090 ofthe daily dose of aminophyll ine. When the change is to be made,the appropriate dose of theophyll ine should be given at the ap-propriate time of day (e.9., 8:00 e.r' l . or 8:00 r.u.), and the intra-venous administration should then be discontinued. Theophyl-l ine has been shown to be of value in exacerbations of COPD,but overdosage and toxicity are avoided (160, l6l).

6. Corticosteroids can be useful if there is an asthmatic com-ponent in a patient who demonstrates responsiveness to beta-agonist therapy, although there is i imited information support-ing the use of intravenous or oral steroids in the managementof COPD exacerbations (162). It is vital to avoid prolonged high-dose steroid therapy in those patients who show litt le improve-ment, because such severe complications as avascular necrosisof the ends of long bones can follow after a few weeks of ther-apy. Careful observation and spirometric evaiuation are neededto prove the continuing benefit of steroids after a course of 1to 2 wk; rapid weaning must be accomplished if possible, andthe drug must be stopped if there is no clear objective evidence

TABLE 'I

5

SUGGESTED AMTNOPHYLL INE DOSAGES

I n i t i a l d o s e : 2 . 5 - 5 m g / k g ( l o a d i n g ) ; 9 r v e o v e r c o u r s e o l J 0 n t r n

Af te r 30 min : 3 mg/kg (second load ing) , i f needed; g rve s lowlv -

Subsequent ly : ma in tenance, e g . , 0 .5 mg/kg per hour , mod i ly as needed '

of improvement. For continuing therapy in a responder, steroidsare usually given orally, since intramuscular or aerosol adminis-tration have not been proved to be of value in COPD (163).

7. Abnormal mucus usually provides a rationale for a course ofantibiotic therapy; amoxicillin, trimethoprim-sulfamethoxazole,doxycycline, or erythromycin may be chosen. It has been shownthat such agents may be of some help in resolving an exacerba-tion, but they have more value in decreasing the risk of furtherdeterioration (164). ln severe exacerbations, a broad-spectrumpenicillin preparation or a cephalosporin may be preferred toprovide coverage of resistant organisms. Sputum culture can helpdirect therapy in such cases, particularly if prior antibiotic ther-apy has recently been given or the patient has l ived in a nursinghome. Mucokinetic agents, such as acetylcysteine or iodides, havenot been shown to be effective in exacerbations of COPD. Pre-liminary evidence suggests that DNAse may be of value. Failureto improve mucociliary clearance is an important factor Ieadingto endotracheal intubation and mechanical ventilation.

8. Sedation and pain management must be provided to manyseriously ill patients with respiratory depression, including thosewith COPD. Fortunately, tolerance to the effects of narcotics andsedatives in depressing the respiratory center usually occurs morequickly than does tolerance to the desired beneficial effect (165).Thus, these drugs can be used in patients with COPD and thedosage slowly increased over several days to achieve the desiredeffect. Renal insufficiency may occur as part of the aging pro-cess and can markedly prolong the effect of drugs; its presencedictates the choice of an agent with a shorter half-life. lf an over-dose is inadvertently given, the patient may require respiratorysupport and a pharmacologic antagonist. ln contrast, markedbased on symptomS or Serum leve ls

American Thoracic Society

Algorithm for Correcting Hypoxemia in the Acutely ll l COPD Patient

Assess Patient ABGs tf possibie

ABGs and SaO,

if not done earlier

Adjust O,per oximetry

to SaO, >90%

ABGS : assumingPO2 60-65 mm Hg

ABGs PO2 60-65 mm xg

sedation may be required in agitated ventilated patients, and insome cases, pharmacologic paralysis wil l be needed (166). Halo-peridol is of value for agitated or confused patients, because ithas l itt le if any respiratory depressant effect.

Because it is important to ensure the comfort and safety ofthe patient, undertreatment can be as harmful as overtreatment.ln terminal management, appropriate use of narcosis or seda-tion must be attempted when severe distress necessitates suchtreatment. Dyspnea, a major symptom, may be helped by ben-zodiazepines or narcotics; these drugs need not be withheld rvhenindicated to reiieve severe, persisting discomfort, even thoughthere is a risk of shortening.the patient's l i fe (167).

Much research remains to be carried out in this area. Moreinformation is needed regarding appropriate minimal and max-imal dosages for established bronchodilator drugs, as well as thevalue oi drug combinations. Dosage changes and optimal deliv-ery methods need to be investigated in endotracheally intubatedpatients receiving ventilator support. The possible value of muco-kinetics, surfactants, and lung lavage to correct mucostasis hasnot been adequately addressed in ventilated patients with COPD.Although COPD is associated u'ith airway inflammation, criteriafor the use of steroids and other anti-inflammatorl ' drugs havenot been esnblished. The possible roles of other agents, suchas respiratory-center stimulanls, sedatives, hypnotics, and eu-phoriants, need to be further evaluated. The use of pharmaco-logic relief for distressing terminal sr-mptoms deserves more at-tent ion.

MOBILIZATION OF SECRETIONS

Mucus hypersecretion and impaired tracheobronchial clearanceare frequent in patients rvith COPD. Although increased airwa5'secretions may correlate weakll '* ' i th airf lo* obstruction, chronicmucus hypersecretion does correlate rvith hospital admissionsfor acute exacerbat ions of COPD (168) and ma1' contr ibute tothe risk of death in patients u'ith severe ventilatorf impairment069).

During acute COPD exacerbations, many factors, such as vi-ral and bacterial airway infections, stimulate increased mucusproduction, alter mucus viscoelastic properties, and impair air-way mucociliary clearance mechanisms. Tieatment is commonlydirected toward enhancing airway secretion clearance. The effi-cacy of these techniques in patients with chronic bronchitis andemphl'sema has often been inferred from investigations of pa-tients u'ith cystic fibrosis and bronchiectasis. Scant scientific evi-dence supports their application in hospitalized patients withacute exacerbations of COPD uncomplicated by bronchiectasis'

Directed Coughing

During acute exacerbations of COPQ nonintubated patients maycough repeatedly in an ineffective manner. Patients can becoached to use a controlled cough or a forced expiratory tech-nique (known as "huff coughing"). Controlled coughing con-sists of a slow, maximal inspiration and breath-holding for severalseconds, followed by two or three coughs. The forced expiratorytechnique consists of one or two forced exhalations ("huffs") frommid to low lung volumes, with the glottis open. "Huff cciugh-ing" avoids the dynamic airway collapse, bronchoconstriction,and patient fatigue associated with undirected forceful cough-ing. Although beneficial in patients with cystic f ibrosis and ap-plied without supporr.ing data in patients with stable COPD, thesetechniques are diff icult to use with dyspneic patients during acuteCOPD exacerbations and provide unproven benefits in this clin-ical setting.

Physical Methods

Although chest physiotheropy with postural drainage, with orwithout chest percussion andlor vibration, has been considereda standard approach to patients with emphysema and chronicbronchitis, bronchiectasis, and cystic fibrosis, its value in hospital-ized patients with acure exacerbations of COPD has nol beendocumented. Trials of postural drainage in patients acutely i l l

AMIRICAN 'OURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE VOL I52 I995

SpirometryABGs

NonThoraco-

abdominal

OpenThoraco-

abdominal

TeamEvaluation

SpecificEvaluation &

Therapy

PPO.FEVlPPO.DCO

> 40% Normal

Patient DesireAlternateTherapy?

AggressiveCOPD

Therapy

tcu,Mechanical Ventilation

as needed

with COPD have failed ro show a positive effecr on spurum vol-ume, gas exchange, or spirometric measurements (170). Chestwall percussion and vibration used with postural drainage simi_larly lack_scientif ic support. Furthermore, in patients aJutely i l lwith COPD, chest percussion and vibration can cause a rran_sient decrease in FEV, and an increase in FRC as a resulr of acutebronchoconsrriction (l7l). As shown in patients with cysric f ibro_sis or pneumonia, no clear evidence supports the superiority ofmanual versus mechanical methods.

In patients with cystic f ibrosis, the benefits of postural drain-age are limited to patients who expectorate at leajt 25 ml of spu-tum per day (172). Extrapolatingfrom these observations, ii isrecommended that postural drainage, with or without chesi pei_crls;io1^o1l,ibrarion, should be limited to hospitalizea patilntswith coPD whose sputum production also exceeds 25 mlld.rrrmer.oata trom patients without COPD support the role ofchest physiotherapy in patients whose course-ii complicated bymucus plugging.with lobar atelectasis (173).

American Thoracic Society

Intermittent posirive pressure breathing (IPPB) has been aban-doned in the care of pat ienrs wi th COPD, because no sc ient i f icevidence supports i ls use as an aid either in promoting clearanceof airwal' secretions or in delivering aerosolized medications.

Positive expiratory pressure (PEP) techniques have recentlybeen prescribed to facil i tare secretion clearance bv simularing theeffects of continuous positive airway pressure face mask systems.With these techniques, parients exhale against a fixed-orifice resis-tor, generating pressures that range from l0 to 20 cm HrO (174).Although some, but not all, reports indicate improved airwai'clearance with PEP in patients with cystic f ibrosis, its value instable patients with chronic bronchitis remains unclear (175).

No objective evidence confirms benefits from inholation ofaerosolized water or saline via mask or mouthpiece. Because onl1,'a small portion of aerosolized water droplets reaches the lowerairways, sufficient volumes of water or saline cannot be ad-ministered by jet nebulizarion alone to l iquefy airway secrerions.Furthermore, bland aerosols are potentially dangerous, they mayproduce airway irritation that can increase bronchospasm andpresent a risk of nosocomial pneumonia. One study (176) ob-served marginally enhanced sputum production when 5 min ofaerosolized saline preceded chesr physiotherapy, but the samplesize was small, and other variables may have contributed to im-proved mucocil iary clearance. However, bland aerosols providebenefit in patients with artif icial airways who have COPD to pre-vent inspissation of airway secretions.

Although there have been suggestions for systemically over-hydrating patients with acute exacerbarions of COPD, this prac-t ice cannot be supporred wi th data (177) .

Nasotracheal suctioning in the nonintubated patient with re-spiratory distress is seldom indicared. It is a rraumaric and un-comfortable procedure rhar mat' impair venti latory function andexhaust patients with borderline venrilatory reserve. Its very short-term or infrequent use should be l imited ro patients who mavbe able to avoid endotracheal inrubarion and mechanical venti-lation if relieved of excessive air*al secrerions.

For patients hospiralized uirh COPD who require frequenrnasotracheal suclioning. some centers advocate a percutaneoustemporary "mini-tracheotoml'," w hich provides airu'a)' access forrepeated nontraumaric suclioning u'irh a l0 Fr suction cather.er(178). Catheter misplacemenr and airual hemorrhage are im-portant potential complicarions of rhe procedure There have beenno studies examining clinical ourcome compared with noninvasivesuctioning techniques or lare airrvav complications in parientstreated with mini-tracheorom\'.

Cl in ica l Guidel ines

Controlled coughing and "huff coughing" rechniques are of pos-s ib le benef i t to hospi ta l ized par ienrs * i rh COPD u, i rh inef fec-tive coughing. Although posrural drainage. chesr percussion, andchest v ibrat ion have nor been sho*n ro benef i t pat ients wi thCOPD, these maneuvers ma) be used - rvirh careful moniroring -after pretreatment with inhaled bronchodilarors in parienrs u,irhmore than 25 ml of expectorared spurum a day.

IPPB is no longer indicared in mosr par ienrs u i rh COPD. PEPtechniques may possib l l ' enhance a i rual c learance bur requi refurther investigation. Bland aerosol rherapl and s1'sremic hydra-t ion beyond euvolemia do nor enhance l iquefacr ion of a i r* .a1 'secretions in nonintubated patienrs.

Nasotracheal suct ioning of noninrubared par ienrs should begenerally avoided or l imited lo shorr-rerm use, Some cenrers ao-vocate mini-tracheotoml' to provide remporar) access for airu.avsuct ioning in selecred par ienrs.

A summary of possib le measures and thei r prcsent s tatus isfound in Table 16. Mosr o l rhese measures lack adequare inves-t igat ive support and are used because of demonsrrared benef i t

TABLE I6

MEASURES TO MOBILIZE AIRWAY SECRETIONSIN HOSPITALIZED PATIENTS WITH COPD

. Dr rec ted cough ing , "hu f f cough ing . , . Benef i t ex t rapo la ted { rom exper ience incys t rc l rb ros ls

. Chest phys io therapy : manua l o r mechan ica l ches t percuss ion and pos tura ld rarnage. Benef i t ex t rapo la ted f rom exper ience in cys t ic f tb ros is . Can causet ransren t fa l l in F [V ] . Assumed ro le l im i ted to pa t ien ts w i th > 25 ml spu-tum per day or lobar a te lec tas is f rom mucus p lugg ing

. In te rmi t ten t pos i t i ve p ressure brea th ing ( lppB) . Not ind ica ted ; no Drovenb e n e l i t ' n C O p D

. Pos i t i ve exp i ra to ry p ressure (pEp) . Benef i t ex t rapo la ted t rom exper ience incys t ic f ib ros is . No repor ted exper ience in acu te exacerbat ions o f COpD

. B land aeroso l therapy . No demonst ra ted benef i t in COPD un less ar t i r i c ia la i rway is in p lace . May cause bronchospasm in non in tubated pa t ien t ! .

. Sys temic hydra t ion . No demonst ra ted bene l i t beyond rep le t ion o f in t ravas-cu la r vo lume to euvo lemia .

. Nasot rachea l suc t ion ing . L imi ted benet iu to le ra ted on ly fo r shor t per iods

. Mrn i - t racheotomy. Poss ib le temporary benef i t in pa t ien ts w i th pers is ten t a i rway secre t jons caus ing resp i ra to ry de ter io ra t ion

in patienrs with cystic fibrosis or bronchiectasis. Research effortsshould be focused on merhods that are both effective and inex-pensive, including patienr education and rechniques such as pEp,which can be self-administered using low-cost equipment.

ASSISTED VENTILATION

Progressive airflow obstrucrion may impair oxygenation and/orventilation to rhe degree that assisted ventilation will be required.The goals of assisted ventilation in this clinical context are tosupport the patient over the short term, during the course ofacuterespiratory failure, and to enhance gas exchange and functionalstatus in parients with chronically impaired ventilation.. . Although no objective guidelines exist for determining theideal t ime to init iate venrilarory support, that support wil l bene-fit two classes of parients: (1) those who have experienced progres-sive worsening of respiratory acidosis and/or altered mentalsta-tus,. despite aggressive pharmacologic and nonventilatory support;and (2) those with ciinically significant hypoxemia, which hasdeveloped despite the provision of supplemental oxygen by usualtechniques. In deciding to init iate assisted venrilation, the clini-cran must selecr from an array of venti latory modes and appli_ances, including endotracheal tube, oral or nasal mask ventila-tion, and negative pressure ventilarion.

Airway Access

Translaryngeal intubation provides a secure airway for positivepressure ventilarion as well as a route for tracheobronchial suc_tioning in parienrs wirh excessive airway secretions. The decisionto- intubate through an oral or nasal route depends on the skil lsof theperson in charge of the procedure and patient-specific fea-tures that may obviate one route or the other (e.g., micrognathia,nasal polyps, and bleeding diathesis with epistaxis).

Intubated parienrs with COpD may benefit from tracheotomyat some trme during rhe course of respiratory failure (179). Be_cause no srudy confirms that duration oftranslaryngeal intuba_tion correlates u,ith the incidence of laryngotracheal complica-tlons, no arbirrary l imirs to translaryngeal intubation based onsafetf issues exisr ro guide the timing of rracheotomy (lg0). Avait-able clinical data clearly dictate neither special efforts to avoidtracheotoml, based on unacceptable complications, nor are earlyapplications of the procedure jusrif ied by definire benefits.

Parients with COpD should be selected for tracheotomy whenweaning from mechanical venti lation appears distant and antic_

AMERICAN IOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE VOL I52 .I995

ipated benefits of the procedure appear to outweigh its immedi_ate surgical and long-term risks. However, it is rarely justif iablein parients with COPD to prolong rranslaryngeal inrubation be_yond l4 d, because the l ikelihood of rapid weaning markedlydiminishes after that t ime. potential benefits of traclieotomv in_clude increased patient comfort and mobilitli promotion of tians_fer to an ICU, expanded abil ity to ear, enhanced verbal commu_nication, avoidance of direct laryngeal injury, and improvedremoval of secretjons. In short, there are no clear, vaiid guide_lines. The decision to perform a tracheoromy requires cirefulclinical judgment.

Mechanical Ventilation

Assisted ventilation should be considered for patients with acuteexacerbations of COPD when pharmacologic and other nonven_tilatory trearments fail to reverse clinically significant respira-tory failure. Indications for init iating assisted ventilation durineacute COPD exacerbations include signs of respiratory muscl!fatigue, worsening respirarory acidosis, and/or deterioraing men_tal status. Although several investigators have reported su.cesswith negative pressure venrilation, most studies advocate posi_tive pressure inflation for acure exacerbarions of COpD [tgt;.Additionall l,, negalive pressure venrilation may cause upper air_way obstruction and arterial oxygen desaturation (when upperairway muscle activation is asynchronous with negative pr.isu.ebreaths). lts role in managing parients with COpD has tLereforebeen quesr ioned (181. 182) .

The main goals of assisted posirive pressure ventilation in acuterespirarory failure complicating COPD are the resting of ven_tilatory muscles and the resrorarion of gas exchange t; a stablebaseline. Ar rhe same time, the clinician musr aim to avoid com_plications associared wirh mechanical venrilation and should ini-t iate weaning and discontinuation of mechanical venti lation assoon as possible.

Major risks associated wirh assisted posirive pressure venti_lation include venrilator-associated pneumonia, pulmonarybarotrauma, and laryngorracheal complications associared withintubarion and/or rracheoroml,. In addition ro rhese general haz_1!11here are rhree specific pitfalls in venti laring pitients withCOPD. One is overvenrilarion, resulting in acure resiirarory alka-lemia, especially in patients with chronic hypercapnia. A secondis init iation of complex pulmonarl ' and caidiovascular inrerac-tions that may result in sysremic h)'porension. The third is crea_tion of intrinsic posirive end-expiratory pressure (pEEp), or"auto-PEEP," especiallf i f expirarorl ' t ime is inadequate or ifdynamic airf low obstrucrion exisrs (lg3).

Auto-PEEP has been reported to occur in up ro 39go ofmechanically venrilated parienrs (lg4). Becaus. iOpO predis_poses to the development of auto-pEEp, and auto_pEEp canhave the same adverse consequences as applied pEEp (e.g., de_creased venous return, the promotion of barotrauma, and im_paired inspiratory effort due to expirarory pressure that must beovercome before inspiration begins) recognizing and minimiz_ing auto-PEEP is importanr. Maneuvers ro di*inisi, auro_pEEpinclude: (/) trearing airf low obstrucrion; (2) increasing expira_tion time, e.g., by decreasing respiratory rate, increasing insplru_tory flow rates to avoid a disadvantageous inspiratory:expira_tory (l:E) rario, and employing a large-caliber endotracheal iube;and (J) reducing compressibre volume in the vent i raror c i rcu i r( r 85 ) .

The three ventilatory modes most widely used for managingpatients u'ith COPD are assisl-control venrilation (ACV), iruerlmittent mandatory ventilation (lMV), and pressure supporr ven_t i la t ion (PSV). Because some, rhough noi a l l , c l in ica l reporrsindicate that PSV provides increased parienr comforr (l g6), pro_motes patient synchrony wirh the ventilator, and may accelerare

weaning (187), it may be a particularly valuabre mode of venti la-tory support for stabilized patients with COpD and acute respi_ratory failure who maintain adequate ventilatory drive. No ii-rect evidence exists, however, that patient outcome is improvedwith the use of PSV compared with volume-cycled modes of me_chanical venti lation.

Noninvasive Ventilation

Tianslaryngeal intubation presents risks of nosocomial pneumo_nia, laryngotracheal injury, and bacterial sinusitis; it aiso inter_feres with the patient's capacity for verbal communication (lgg).The advent of noninvasive assisted ventilation may offer an ai_ternative to intubation in some patients with COpD and acuterespiratory failure. The weight ofavailable evidence suggests thatpositive pressure noninvasive techniques unload respirit-ory mus_cles (182) and avoid upper airway obstruction more effettivelythan negative pressure technioues.

Noninvasive positive pr.rrui. ventilation for acute exacerba_tion of COPD has been examined in several studies, using bothfacial and nasal masks in conjunction with volume-cycled venti_lation, bilevel positive airway pressure (BipAp), and piessure sup-port modes (189, 190). Although available studies suggest a poolidsuccess rate of 7 6s/o in patients with acute respiratory failuie com-plicating COPD, the reported experience ij based on small pa_tient numbers, and failure rates up to 40go have been reporiedin individual studies (l8l). Comparison studies with intu'bationand mechanical ventilation have not demonstrated an improvedclinical outcome in terms of patient survival. The maximal du-ration of successful ventilatory support using noninvasive posi_tive pressure ventilation has been brief, usrially 7 d or less.. Primary use of noninvasive techniques for respiratory failurein patients with COpD should be ieserved for centirs withadequate expertise and patient supervision to allow safe im_plementation. Patient features that ihould discourage consider-ing noninvasive ventilation for acute COPD exaceibations in_clrr.de hemodynamic instability, copious secretions, lnaUifity iLdefend the airway, poor cooperation with the technique, orirn_paired mental status.

^ The following are necessary to ensure the success and safetyof noninvasive positive pressure techniques: (/) institutional ei_perience and extensive preparedness foi noninvasive ventilationby physicians, nurses, and respiratory care practitioners; (2)ade_quate unit staffing ro manage the techniques, which requiieclosepatient monitoring, frequent bedside skillea care, and prepared_ness for emergent intubation; and (J) the patient;s mentai alert_ness, tolerance of noninvasive appliances (i.e, without agitation),hemodynamic stability, capability ro conrrol the airway ana toclear secretions, and absence of copious airway secretions.

In the absence of these safeguardi of success, intubation withpositive pressure ventilation remains the therapy of choice foracute respirarory failure. patients managed with noninvasive as_sisted positive pressure ventilation ,eqriir. close supervision toassess response ro therapy and the abil ity to proceed promptlywith intubation should noninvasive ,rpio.t 'fail ( lgl; l9l):

Weaning from Mechanical VentilationMany patienrs with CopD who undergo mechanical venti lationfor acute bronchospasm, fluid overloid, oversedation, or inad_vertent hyperoxygenation may experience successful extubationwithout goin_g through a period of weaning. However, some pa_tients with COPD intubated for respiratorylailure require gradualweaning. The most important faciors thit determine tirJaUitityof patients ro wean from prolonged mechanical venti lation areneuromuscular reserve capacity relative to respiratory load, car_diovascular- performance, oxygenation, and psychoiogicaf fac_tors. The relative contribution of each of theie iactors-in wean_

American Thoracic Society

ing varies among patients. Although some clinicians believe thatimpaired neuromuscular capacity, in the form of respiratory mus-cle fatigue, plays a major role in determining the success or fail-ure of the weaning process, no studies have confirmed this im-pression. Nor are there reliable techniques to detect and quantifyrespiratory muscle fatigue and endurance in ventiiator-dependentpatients.

A variety of objective physiological indices (maximal inspira-tory pressure, vital capacity, respiratory frequency,/t idal volume,P0.l) are designed to evaluate patients for extubation. When mea-sured values are better than anticipated from the bedside obser-vation of a patient's status, these indices may identify which pa-tients with COPD can undergo discontinuance of venti latorsupport earlier than otherwise thought possible. Conversely, se-verely deranged indices may obviate futi le weaning efforts thathave no chance of being well-tolerated. Aside from these situa-tions, however, no clear physiologic indices assist the selectionof patients for weaning, determination of the rapidity of wean-ing, or identif ication of the ideal weaning method.

Available techniques for weaning patients with COPD frommechanical venti lation include assist-control venti lation withT:piece trials, IMV, and PSV. IMV and PSV offer theoreticaladvantages: they provide partial support when the patient is con-nected to the ventilator, and they present less opportunity forbarotrauma. There is insufficient evidence, however, to estab-lish that weaning is accelerated or outcomes improved with anyparticular weaning technique now known.

General Cl in ica l Guidel ines

Translaryngeal intubation through the nose or mouth with posi-tive pressure ventilation remains the primary approach to assistedventilation in patients with acute exacerbations of COPD. Thereis no ideal t ime to convert patients from translaryngeal intuba-tion to a tracheotomy when prolonged airway access is. required.Consensus indicates that the need for tracheotomy should be an-ticipated and the procedure performed when the advantages oftracheotomy appear to outweigh its risks. Prolongation of trans-laryngeal intubation beyond l4 d is justif iable only infrequenll-v.

The mode of mechanical venti lation should be determinedon the basis of institutional experience and perceived patient ben-efit. Although PSV has theoretical benefits and some emergingclinical investigative support, further experience is needed to de-termine the relative value of the available modes of venti latorysupport. Regardless of the mode selected, special effort shouldbe expended to minimize auto-PEEP, normalize pH, avoid alkale-mia, maintain Pa6q, values equii 'alent to those achievable bythe patient when weaning is init iated, and direct aggressive re-spiratory, general medical, psychological, and physical therapl'support to promote successful weaning.

No physio logical ind ices re l iab l f ident i f l pat ients who cantolerate weaning or establish an ideal weaning technique. Sometheoret ica l benef i ts and l imi ted c l in ica l data support the possi -b le advantages of PSV for weaning pat ients u i th COPD; con-sensus indicates, however, that T-piece trials. I\1\'. and I 'ariablecombinat ions of these three techniqus5. uhen expert l l ' appl ied,are equivalent in weaning pat ients f rom mechanical vent i la t ion.

The use of noninvasive modes of posi t ive pressure vent i la t ionshould be reserved for centers with extensive erperience with thesetechniques and should be d iscouraged in pat ients rv i th hemo-dynamic instabil ity, poor airual conrroi. ercessive secretions, al-tered mentation, or an inabil it l to cooperate. Present experiencedoes not support the ro le of negat i re pressure vent i la tory tech-niques for pat ients wi th COPD and acute respi rator l ' fa i lure.

Needed Research

Creater understanding of the r isks and benet- i ts of t racheotomy

specifically related to patienrs with COPD is required to deler-mine its rolc in venti lator-v support and accelerating weaning frommechanical venti lation. Comparative trials of PSV and other newventilatory modes relative to standard positive pressure techniquesare required to identify an ideal venti latory approach to patientsu'ith COPD and acute respiratory failure. These trials should in-clude considerations of patient comfort and cost effectiveness.Although the init ial cl inical trials of noninvasive posirive pres-sure support have creared enthusiasm for these techniques, ad-ditional comparative trials with standard approaches are requiredto establish their role in management and to idenrify patient-specific features that improve their success.

Continued effort is required to discover new techniques to ac-celerate weaning from mechanical venti lation and to identifypatients v:ho can undergo weaning and discontinuance of me-chanical venti lation with a high l ikelihood of success. Greaterunderstanding is also needed of the relationship between ven-ti latory Ioad and ventilatory muscle performance, in additionto the other factors that l imit spontaneous ventilation and anabil ity to wean. Continued efforts to develop consensus amongexperts on mechanical venti lation and to disseminate informa-tion (192) promise to improve the care of patients with COPDand respiratory failure.

INPAT]ENT OXYGEN TH ERAPY

The most important consequence of hypoxemia is t issue hyp-oxia. Hence, the first responsibil i ty of the physician is to corrector prevent life-threatening hypoxemia. Ideally, an arterial bloodgas might be obtained before starting oxygen, but this is not al-ways possible. Moreover, in preserving celluiar oxygenation, theclinician must also consider the other oxygen rransport variables,including adequate hemoglobin, cardiac output, and the distri-but ion of t issue per fus ion.

Oxygen Delivery Methods

The most common oxygen delivery method in the hospital set-ting is the standard dual-prong nasal cannula. lt is inexpensive,relatively comfortable, and accepted by most patients. The Ftgr,using the nasal cannula, is affected by the geometry of the nose,mouth breathing, venti latory rate, t idal volume, respiratory pat-tern, and oxygen flow setting. The Frg, thus varies from momentto moment but can be approximately stated as follows: FIo2 =20Vo + (4 x oxygen lirer flow).

As noted, venti latory rate and pattern affect the Ftgr. Thisis because the slower inspiratory flow of room air causes lessdilution of the constant (f ixed) fiow of oxygen. Additionally, ahigh dead space-ro-r idal -volume rat io (Vd/Vt) wi l l reduce theefficacy of nasal oxygen deiivery.

The simpleJace ntask, with a mask volume of 100 to 300 ml,wil l deliver an Fts, of 35 to 5590 at 6 to l0 L,/min. Supply flou,sgreater than 5 L,/min are recommended to u'ash out CO:. Themask is similar to the nasal cannula in its dependence on ven-ti latorl 'rate and respiratory pattern to achieve its desired Ft6r.The mask is useful in patients who are strictl l , mouth breathers,as well as some patients with extreme nasal irritation or epistaxis.On the negative side, face masks are obtrusive, uncomfortable,and confining, they muffle communication, and they obstructeat ing. Not a l l mouth breathers requi re a mask. Most mouthbreathers have some nasal airf low: hence. the standard nasal can-nula can provide adequate oxygenation in many ofthese patients.

If an accurate and constant Fro, is required, perhaps due toconcern about CO, retention, a Venturi ntask can be effective.The Venturi mask, supplied by high oxygen flows, maintains af ixed rat io of oxygen to . room ai r , thus mainta in ing a constantFt6r . Typical F lo, ser t ings inc lude 24,28,31, 35. and 4090.

If higher than 4090 Fte, is required, the patient can be ad-

AMERICAN 'OURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE VOL I52 I995

,l

23

ministered oxygen viaa non-rebreathing mask with reservoir andone-way valve. Oxygen flows into the reservoir at 8 to l0 L/min,so that the patient inhales a high concentration ofoxygen. Hence,these systems can deliver Fre, up to 9090 when the mask has atight seal on the face. On the negative side, a tightly sealed maskcan be very uncomfortable and thus poorly tolerated. The higheroxygen concentration also creates a greater risk of COz retention.

Some patients who are difficult to oxygenate using the stan-dard nasal cannula can be oxygenated using a reservoir cannulaor transtracheal catheter because of their higher delivery effi-cacy. (Oxygen-conserving devices were discussed previously indetail in Long-term Oxygen Therapy.)

In general, the choice of device will depend on its effective-ness, reliability, and ease of therapeutic application as well asthe patient's ability to tolerate it. Whenever there is a change inthe delivery device, the patient wil l require an ABG or oximetry.Continued monitoring is highly advisable in the unstable Datienr.

Setting and Adjusting Oxygen FlowThe goal of oxygen therapy is correction of hypoxemia to aPae, ) 60 mm Hg or Sae, ) 900/0. These values correspond toan arterial oxygen concentration (Cao.) of about l8 vol go. Be-cause of the shape of the oxyhemoglobin dissociation curve,increasing the Pa6, to values much greater than 60 mm Hg con-fers l itt le added benefit ( l ro 2 vol go) and may, in rareiases,increase the risk of CO: retention. It is therefore recommendedthat the init ial f low setting be adjusted to bring the pae, to justabove 60 mm Hg. Recommended initial settings are summarizedin Table 17 (adapted from Mithoefer and colleagues [193]).

It is recommended that Pco, and pH be monitored while ini-tially titrating the oxygen flow setting. In general, patients re-ceiving oxygen sufficient to raise Pas, to 60 mm Hg will nor ex-perience CO, retention or acute respiratory acidosis. If CO,retention does occur, it is often minor and not accompanied bysignificant acidosis. If the patient is a chronic COr retainer, areasonable goal is to adjust the oxygen setting to correct the hyp-oxemia to a Pas, of 60 mm Hg and maintain the COr at a stable,albeit elevated, level. Again, the pH should be monitored. If ade-quate oxygenation is unachievable without progressive respira-tory acidosis, then mechanical ventilation may be required.

In titrating the oxygen flow setting, the clinician must remem-ber that it can take 20 to 30 min to achieve a steady state aftera change in Fr6, in a patient with COPD. Therefore, ABG sam-pling at closer intervals may be misleading.

Following the initial setting, adjust oxygen flow via ABC tobring Po, ) 60 mm Hg or Sa6, ) 900/0.If CO, rerention occurs,observe the pH: if i t is nonacidemic, accept the high CO, as itis chronic. If the pH is acidemig consider a Venturi mask to closelycontrol the Frs, or mechanical ventilation. All other medical treat-ments should be optimized. This includes bronchodilators, bron-chial toilet, treatment of infection, and congestive heart failure.As the patient becomes ambulatory, a walking oximetry can beused to determine an exercise flow setting.

These recommendations are summarized in Table l8 and thealgorithm shown in Figure 3 (adapted from Carter [194]).

Patients are often started on oxygen for the first time duringan acute exacerbation. As the lingering effects of the exacerba-tion take time to resolve, the patient may have to be dischargedfrom the hospital on oxygen. After 3 to 4 wk, oxygen may nolonger be required; the determination may be made by a bloodgas measurement performed at that time.

Future research should focus on maximizing oxygen trans-por.t a.nd use at all levels. Since patients become hypoxemic dueto V/Q mismatching, oxygen delivery and pharmacological ap-proaches might maximize ventilation in well-perfused lung units.Hemoglobin and other oxygen transport vehicles may be im-proved to increase the blood oxygen content. Tighter control ofoxygen delivery might be achieved through the use of closed-loopsystems in which oxygen delivery would be tied directly to oxy-gen saturation; such methodology might become a viable alter-native to Venturi masks. Techniques to direct blood flow and oxy-gen uptake into specific target tissues may also be developed.

TAELE

RECOMMENDED IN IT IALACHIEVE AMBIENT A IR

OXYCEN SETTINGS TOPas2 ) 6O mm Hg

Pao2 Breath ingAmbient A i r F to2(mm Hg) (ok)

Nasal CannulaSett ing to AchieveApproximate Fro2

(L/nin)

50454035

2428323 5

TABLE 18

AD.IUSTMENT OF OXycEN SETTTNGS BASED ON ARTERTAL BLOOD GASES

Pas, (mm Hg) Paco2 Therapeut ic Decis ion

> 6 0> 6 0> 6 0> 6 0< 6 0< 6 0< 6 0

Normal

Mi ld r i se

H i g h

Large ilse

No nse

Mi ld r i se

Large nse

Norma I

Normal

Normal

Low

N o r m a l

Normal

LOW

No change in 02 f lowNo change in 02 f low; mon i to r ABcNo chang€ in O: f low; mon i to r ABCCons ider Ventur i /mechan ica l ven t i la t ionRaise 02 ; mon i to r ABCRaise Oz; mon i to r ABCCons ider Ventur i /mechan ica l ven t i la t ion

When Pao, i s mu(h grea ter than 60 mm Hg, f low se t t ings may be reduced to ma in ta in va lues near 60 mm Hq

Amer ican Thorac ic Soc ie tv

Surgery and the COPD PatientPatients with COPD are susceptible to all the diseases that mayrequire surgical therapyr They are also more likely to develop somediseases due to age, COPD i tse l f . and pr ior tobacco use (e.g. ,atherosclerotic vascular disease and cancer). Surgery in the pa-tient with COPD can therefore be associated with increased riskof morbidity and mortality. But recent enhancements in peri-operative monitoring, pain control, and surgical techniques mayallow for surgery in patients with such severe obstructive lungdisease that their risk was formerlv resarded as prohibit ive.

PREO PERATIVE EVALUATION

Evaluation of the patient with COPD must identify the goal ofsurgery and determine whether the risk-benefit ratio makespotential attainment of the goal worthwhile (195-198). This pro-cess wil l depend on the indications for the surgery, surgery site,experience of the surgical team, type of anesthesia, and degreeof respiratory impairment (195-200). The risk factors are identi-f ied by history, physical examination, chest X-ray, and a batteryof pulmonary and cardiovascular screening tests. With that in-formation, the type of surgical approach, anesthesia, and peri-operative therapy can be determined.

The incidence of postoperative pulmonary complications(PPC) varies in the medical l i terature, depending on the defini-t ion of the term. lnvestigators often perform routine postopera-tive chest radiographs or blood gases to detect abnormalities thatmay be prevented or treated with various forms of therapy. Thefinding of a decrease in Pao, or the appearance of a basilar in-filtrate may thus be defined as a PPC. In asymptomatic patients,however, these occurrences may not increase morbidity, hospitalstay, costs of care, or mortality. True complications capable ofproducing these untoward outcomes are purulent bronchitis,atelectasis requiring bronchoscopy or respiratory therapy, pneu-monia, puimonary embolism, myocardial infarction, puimonaryedema, and cardiorespiratory failure.

Careful attention to perioperative respiratory management ofpatients with COPD can improve both outcomes and resourceutil ization (201-204). This precaution applies specifically to pa-tients with severe disease (stage II or III); patient with mild-to-moderate COPD (stage I) have operative risks similar to the generalpopulation and need no special approach (195).

The health care provider is challenged to assess risk, projectbenefit, and provide perioperative management. Education is cru-cial to the success of this endeavor. Both patient and farpily mustnot only be counseled concerning the possibil i ty of permanentdisabil ity or death (e.9., after lung resection), but they must alsobe reassured that the need for short-term postoperative mechan-ical venti lation does not necessaril l ' constitute morbidity' but ma)'be a useful postoperative care technique.

Multispecialty Consultation

Because the expertise required for oprimal management of pa-tients with stage II or I l l COPD spans several specialt ies, a mul-tidisciplinary team of consultants should be assembled to ad-vise the patient's surgeon. A respirarorl ' phy'sician ma) serve asprimary patient advocate, coordinaring all preoperative input.The consulting team might include the patient's primary care ph1.sician, an anesthesiologist, a respirator! 'rherapist, and a respi-ratory nurse specialist. Because manl'parients *ith COPD haveother i l lnesses associated with tobacco abuse. input from a car-diologist, vascular surgeon, and otorhinolarl 'ngologist is oftenhelpful. Scheduling preoperative visirs *'ell in advance of elec-tive surgical procedures assures thar all opinions are obtainedand necessary preoperative therap!' is srarred in time lo have abeneficial effect.

Preoperative specialist consultation for patients with stage IIor III COPD impairment is especially crucial when cardiothoracicor upper abdominal surgery is contemplated (196-200). Treat-ment of perioperative pain with intercostal nerve blockade orepidural anesthesia in patients with severe respiratory impair-ment often requires the expertise of an anesthesiologist/anes-thetist or other pain specialists. The primary care physician shouldinit iate therapies to reduce respiratory risk factors, includingsmoking cessation and treatment of airf low obstruction (201-204).

The respiratory specialist should participate as part of the teamand help advise the anesthesiologist/anesthetist and surgeon re-garding the type of physiologic dysfunction present and the likelyresponse to the type of procedure planned. Patients with stageII or III COPD should be admitted to the hospital at least 24 hin advance of the procedure and should have their lung diseaseaggressively treated (201). If an upper abdominal procedure canbe performed laparoscopically, every attempt should be madeto do so, because this appears to significantly reduce PPC (205).

Allied health care professionals are instrumental in teachingthe patient with stage II and III COPD maneuvers, such as breath-ing exercises and coughing techniques, to reduce postoperativerisk factors (206). Aids to lung expansion, such as incentive spi-rometry and continuous positive airway pressure (CPAP), maybe of use in extremely high-risk patients. Perioperative teachingand deep-breathing exercises have been shown to be equally ef-fective with other physical measures, including incentive spirom-etry and intermittent positive-pressure breathing (IPPB).

Assessment

In addition to routine history and physical examination, whenthe nature and severity of a patient's lung disease is unclear, phys-iologic testing should be considered. Lung resection patientsshould have pulmonary function studies, such as spirometry anddiffusing capacity, routinely performed. While there is no defini-tive study supporting the routine use of preoperative spirometryin upper abdominal surgery, this simple test can be useful in thepreoperative evaluation of patients with symptoms of COPD,especially if the test has not previously been performed (195-199).Preoperative arterial blood gas analysis in all patients with se-vere COPD is recommended if a recent test is not available; knowl-edge of preoperative blood gases may be helpful in determiningappropriate postoperative ventilator settings. An elevated arterialPco: has traditionally been deemed d contraindication to majorsurgery; more data are needed, however, since successful lungresection has now been reported in hypercapnic patients. A pre-operative chest X-ray in patients for noncardiothoracic surgeryis sensible, because patients with COPD are at increased risk ofpulmonary neoplasm. Quantitative lung scintigraphy and exer-cise testing may be helpful in determining risk of postoperativecomplications, particularly in patients undergoing lung resection.In addition, exercise testing may reveal previously unexpectedcardiovascular dysfunct ion.

RISK OF SURCICAL PROCEDURES IN COPD PATIENTSNonthoracic Procedures

In patients with reduced airf low (stage II or III), the complica-tion rate varies with the surgical site; the farther the procedurefrom the diaphragm, the lower the risk (195, 198, 199). Commonnonthoracoabdominal procedures in patients with COPD, be-cause of their age, include ophthalmologig otorhinolaryngologigand orthopedic procedures.

Ophthalmologic procedures carry a low mortality rate, un-der l9o (207). Factors that might increase this risk in patients

AMERICAN 'OURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE VOL I52 1995

with COPD include use of sedatives, narcorics, and generalanesthesia. Many of these parienrs have a cough, which may beof concern to the ophthalmologisr because of transmitted ocu-lar pressure; however, excessive cough suppression may lead toretained secretions, atelectasis, pneumonitis, or gas exchangeproblems. Narcotics may also decrease minute ventilation. Beta-blockers such as timolol, which may be used to reduce intraocu-lar pressure, may precipitate bronchospasm and cardiorespira-tory failure even when employed topically. Anticholinergics maydecrease mucocil iary clearance when given systemically, but thisrisk should be negligible if the agenrs are administered in eye-drop form. Pulmonary embolism is possible in ophthalmologicpatients when bed rest is unduly prolonged (207).

Head-and-neck procedures involving the airway may presenta risk to patienrs with COPD. Sleep-disordered hypoxemia is com-mon in COPD, for example, and nasal packing with sedationcould worsen it. Aspiration and direct inflammation, compres-sion, or compromise of the airway after nasopharyngeal, laryn-geal, or thyroidal procedures could also occur. Secrerion man-agement in the laryngectomy patient may, early on, requireadministration of humidity and agents such as beta-agonists roenhance mucoci l iary funct ion.. Orthopedic procedures pose a risk of venous thromboem-bolism (208). Deep venous thrombosis is present in the calf veinsof 40 to 6090 of normal parients not given prophylaxis with hepa-rin. Propagation of the thrombus to the i l iofemoral veins. withsubsequent pulmonary embolization, wil l lead to death in 2 to4Vo of all cases; the figures could be higher in parienrs wirhCOPD.

Urologic, gynecologic, and colorectal procedures in patientswi th s tage I I or I I I COPD that are l ike ly to be shor t andlor un-l ike ly ro involve inrraoperat ive compl icat ions should be per-formed under local or epidural anesthesia whenever possible(198). This recommendation also applies to retroperitoneal sur-gerl ', including renal surgery. Lrngthy abdominal and pelvicprocedures involving general anesthesia may be associated withan increased risk in these patients. If the risk of surgery in pa-tients with impaired lung function is greater than the potenrialbenefil, the surgeri 'should nor be undertaken. Patients with FEV,of less than I L are ar an increased risk for complications andmal,require perioperative mechanical venti lation (195, 196, 199).Hou,ever, there are no absolute contraindications to lower ab-dominal surgery, especially if the risk/benefir ratio is low, e.g.,a vaginal h-vsterecromy for a curable malignancy.

Upper abdominal surgery poses a risk of PPC for all patienrs(196, 198, 199, 206) . Compl icat ions are parr icu lar ly l ike ly in per-sons rvith predisposing facrors, such as morbid obesity, cigarettesmoking, chronic airwav obslruction, heart disease, and old age.In these par ienrs, morbid i ty may approach 8090, and rhe mor-ta l i t l ' ra te is approximarely 3 to 590. Upper abdominal surgeryshi f ts the respi rarory pump f rom the d iaphragm to rhe acces-sor l 'muscles, due ro a non-pain-re lated ref lex (198) . Inrraopera-t ive physio logical changes resul t f rom both rhe anesthesia andthe surg ical procedure i rse l f (198, 200) . Other parhophysio logi -cal changes inc lude mucus hypersecret ion wi th a i ru 'a1, c losureand lung or chest u 'a l l rest r ic t ion.

Nonimperative upper abdominal surger),, such as cholecysrec-toml for as) ,mptomar ic chole l i rh ias is , should be avoided in pa-t ients wi th srage l l or I I I COPD. l f surgery is necessary, arremprsshould be made ro prov ide carefu l anesrhesia (198, 203) . Anorherpotent ia l \ \ ,a ] , to decrease operat ive r isk ma-v be to per form theprocedure laparoscopical ly (205) .

Cardiovascular Surgery

This categor i ' inc ludes both card iac procedures and abdominalvascular surqer\/.

Cardiac procedures. COPD is the most common cause ofpreoperative pulmonary dysfunction in patients undergoingcardiac surgery (195). Cessation of smoking should precedi sur-gery by a minimum of 8 wk (204). postoperarive managemenrof patients with COPD after cardiac surgery should include treat-ment with inhaled berar-adrenergic or anticholingeric bronchodi_lators (209, 210).

If the patient has been treated with theophyll ine preopera-tively, it can be continued postoperatively if the patient's pulmo-nary dysfunction is not satisfactori ly improved with inhaledbeta-adrenergic agonists alone, and the patient's cardiac statuspermits. Monitoring of theophyll ine levels in the perioperativeperiod seems prudent. If there is postoperative bronchospasmunrelieved by the above treatment, corticosteriods should be con-sidered. Experimental studies of the effects of corticosteroidson wound healing indicate that the administration of glucocor_ticoids after the third postoperative day should not delay theu'ound healing process.

Postoperative mechanical venti lation is routine after manycardiac surgical procedures. Ventilation should maintain arterijlcarbon dioxide tension at the preoperative level with a normalpH. Weaning may require an extended time and should not beattempted unti l the patient's postoperative pulmonary status isstable. In the patient with severe COPD, gradual weaning maypermit the patient's cardiovascular status to become sufficientlystable to tolerate assumption of the full work of breathing. Ifthe patient continues to fail weaning from mechanical veniila-tion, and the spontaneous tidal volumes are less than 5 to 6 ml/kg,there may be diaphragm dysfunction secondary to phrenic nerveinjury. This problem can be diagnosed with ultrasound and inmost cases wil l improve within 6 to l2 wk of surgery.

Major abdominal vascular surgery poses a high statistical riskof postoperative respiratory failure, because patients who haveundergone these procedures have had very high incidence ofmoderate-to-severe COPD. A recent prospective study found that24Vo of patients who had such procedures required mechanicalventilation for 24 h or more after surgery (2ll). Three factorswere associared with the need for prolonged mechanical ventila_tion: a history of heavy cigarette smoking, preoperative arterialhypoxemia, and major intraoperative blood loss. Whether revers-ing these factors would change the outcome remains to be studied.

Lung Surgery

Lung surgery may encompass both diagnostic and therapeuticprocedures.

Bronchoscopy in the parienr with COpD carries slightly in-creased risk only if pulmonary dysfunction is severe. After bron_choscopy, acute respiratory failure is very rare, but it can be en-countered if sedation is excessive or airway reactivity increased.Hypoxemia during bronchoscopy is almost routine, and thus,administration of supplemental oxygen under pulse oximetrymonitoring has become routine as well, although its benefits arenot well-documented.

. Thoracoscopy is currently used for evaluation of refractorypleural effusion, performance of pleurodesis, and minilung resec-tions lo remove superficial nodules or perform biopsies. Datain patients without COPD suggesr that this procedure is less in-vasive and better tolerated than open thoracotomy (212).

Thoracotomy is performed primarily for the resection of lungmasses diagnosed or suspected of malignancy. Approximately80 to 90Vo of patients u,ith lung cancer have concomitant COpd,and 20 to 3090 have severe pulmonary dysfunction (213). Thora-cotomy alone wil l have a transient adverse effect, lasting severalmonths, on Iung funcrion; there may be as much as a 30go de-crease in vital capaciry. Pneumonectomy may perman-ently re_duce all lung function by 40 ro 50g0, which could prove devastat-

Amerrcan Thoracic Society

ing to patients with COPD. Surgeons have therefore sought toresect the smallest lung volume possible via wedge resection(0 to l09o permanent lung function decrease), segmentectomy(5 to l09o permanent loss) , and lobectomy (10 to 20% loss) . I fthe region resected has no function, then no loss other than thetemporary decrement attributable to the thoracotomy should re-sult. This principle forms the basis of the regional physiologicalassessment of lung funct ion v ia vent i la t ion or per fus ion sc int ig-raphy (213) .

I t is in the area of lung resect ional surger l ' lhat preoperat ivepulmonary funct ion studies have a wel l -documented ro le (197) .Lung funct ion studies consist ing of sp i rometry and possib lydiffusing capacity should be routinely performed preoperatively(19 '7,213,214). However, the resul ts of these tests should not beused to categorically reject a patient with an anatomically resec-table lung cancer from a potentially l i fe-saving procedure.

Whereas the risk of postoperative respiratorf insufficiencyor death is greater in adul t male pat ients undergoing a pneu-monectomy wi th a preoperat ive FEV, 12 L in an adul t man or50Vo of predicted, a maximal voluntary ventilation (MVV) ( 5090predicted or carbon monoxide diffusion in the lung (Drco)( 6090 predicted, these tests are too nonpredictive to be exclu-sionary. They should be used, however, to indicate which patientscan have resection without further testing (FEV' > 2L or ) 60Vopredicted, MVV > 5090 predicted, and,/or Drco ) 60% pre-dicted). Those patients fail ing to meet these criteria should havequantitative lung scintigraphy using ventilation or perfusion ra-dionuclide imaging to determine which lung has the better func-tion. Another approach is to count the number of bronchopul-monary segments to be resected and establish a proportion (215).A simple formula can be applied to predict the FEV ' after pneu-monectomy (213-216): Post-op FEV' = Pre-op FEV, x 9o func-t ion of contra latera l lung.

Pneumonectomy is generally chosen for this assessment, be-cause it is the largest resection l ikely to occur as a result of theintraoperative staging of the extent of tumor involvemenl. Amodification of the above equarion can be used to assess the im-pact of lesser resection (215,216).

If the calculated FE\/, afrer resection is greater than 4090 ofnormal for the patient's age, sex, and height, resecrion can becarried out with relative safety (216). lr has been shown, how-ever, that preoperative measurement of oxygen uprake (Vor) dur-ing exercise mav also be predictive of postoperative problems.The exercise test is generally performed by symprom-limited max-imum incremental cycle ergometrv. If the patient's Vo, on exer-cise is belou, l0 to l5 ml,u kg/min, complication rares are increased;i f i t is below l0 ml /kg/min, dearh is more l ike l l , (217) . Lung sc in-t igraphy assesses regional lung funcr ion, u,h i le exerc ise test ingevaluales systemic cardiopulmonar), ox),gen transport; thus, theymay be complementarr'.

Perioperative Management

General guidel ines, as noted ear l ier in rhe d iscussion of preoper-at ive p lanning, inc lude smoking cessar ion ar leasr 8 rvk preoper-atively (204) and aggressive trearmenr of lung dl,sfuncrion, usinginhaled bronchodi la tors, theoph.v l l ine, corr icostero ids, and an-t ib iot ics as indicated. Pat ienrs u, i rh srage l l or I l l COPD maybe admit ted to the hospi ta l before surger) ' for mulr id isc ip l inaryevaluat ion, pat ient educar ion, and aggressive rherapy ' (201) .

Inlraoperotive considerations. The effecrs of analgesics andregional and general aneslheric a-eenrs have been revierved (200).In general , the pat ient u, i th COPD ma1' have increased sensi r iv-ity to the ventilatory depressant effects of these agents. Bur sincemechanical vent i la t ion rv i th oxygen is general lv prov ided dur inggeneral anesthesia, and moniror ing wi th pulse ox imerr l 'and end-t idal CO, sampl ing are common. rhe inrraopcrar ive per iod ap-pears not to pose major problems.

The postoperative period. In the immediate postoperativerecovery period, a number of potential threats are present, in-cluding respiratory muscle dysfunction, acidemia, hypoxemia,and hypoventilation. It is in this delicate period that close monitor-ing and, if necessary, mechanical venti latory support become cru-c ia l to the pat ient rv i th COPD (198, 199, 218) .

The inabil ity to overcome the previously mentioned diaphrag-matic dysfunction probably explains the failure of deep breath-ing, intermittent positive pressure breathing, or incentive spirom-etr-v to completely eliminate postoperative complications afterupper abdominal surger1,. A controlled study has shown thatpostoperative complications can be reduced from 5890 (control)to approximately 20 to 25Vo (202).lncentive spirometry was fgundto be the only modality reducing length of hospital stay. Thisfinding has not been uniformly supported, however (203). In-centive spirometry has not been tested extensively in the postoper-ative management of nonabdominal surgery.

The obvious characteristic of all maneuvers to enhance lungexpansion is that they are applied intermittently. Perhaps morecontinuous intervention is required; this premise has yet to betested. In addi t ion, none of these intervent ions can susta in anincrease in FRC, because increased expiratory activity quicklydecreases FRC at the cessation of the maneuver. Indications forpostoperative mechanical venti lation are higher risk cardiotho-racic procedures, respiratory acidosis, severe hypoxemia, retainedsecretions, atelectasis, and pneumonia (198, 199). Continuationof the preoperatively prescribed antibiotics, aerosol bronchodi-lators, corticosteroids, and theophyll ine is standard therapy.

Surgical Procedures to lmprove Quality of Life

The following procedures may be contemplated for the patientrvith COPD, with the goal of either symptomatic relief or defini-t ive correct ion.

Symptom reliel Resection of the carotid body or glomectomyis a procedure designed to blunt the symptoms of dyspnea byremoving the primary source of hypoxic drive. However, init ialclaims of success were followed by problems in patients who de-veloped ventilarory failure because of insensitivity to their hyp-oxic state (219). This procedure has now been abandoned. Pro-cedures designed to change the chest wall 's response to thehyperinflation of COPD, e.g., costochondrectomy, thoracoplasty,and pleurectomy, have also been abandoned.

Surgery to improve Iung function Resection of large bullaecompressing more normal lung tissue can be helpful in relievingsevere dysfunction and dyspnea. The significant challenge is todifferentiate noncompressive destructive emphysema from bul-lae. This differentiarion is generally possible rvith computed tomo-graphl'.

Resection of bullae occupying more than one third of ahemithorax produces the best result. Pulmonary function studiesshowing primarily a restrictive defect withoul severe obstructiontend to reflecr the suspected space-occup),ing nature of the bul-lae. Severe airf low obstruction should increase concern over apossibll ' useless resection of an area of generalized emphysema.However, several studies have suggested that even with very louFEV,, rernoval of compressive bullae is indicared (220). Trear-ment of compressive bullae and recurrent pneumothoraces wilhthe application of laser or other sclerosing a_qents via thoracoscopyhas been reported, but there have been no controlled studies toc lar i f l ' rhei r ind icat ions.

Preliminary reports of the use of lung pneumecromy in pa-tients (lung volume reduction) with nonuniform emphysema havebeen encouraging (221). In select patients, the improvemenrs ins)'mptoms, FEV,, FVC, and arterial blood gases have been sig-nificant. Studies are under way to evaluate rhe role of this proce-dure in pat ients wi th COPD. Prel iminary srudies are under wayto reevaluate surgical therapy for more generalized emphysema.

AMERICAN IOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE VOL I52 1995

bcalized expiratory collapse of the major airways due to acarti laginous abnormaliry can be disabling. Surgical resecrionof the problem area or bronchoscopic placement of a stent mayimprove ventilarion and lessen symptoms (222).

Dyspnea may be relieved by the control of a large, often malig-nant, pleural effusion in the patient with COpD. pleurodesis forthis purpose or to prevent the recurrence of spontaneous pneu-mot.horaces may be performed operarively or by needle place-ment of a sclerosing agent.

Lung transplantation. The definitive procedure to ..cure" COPDwould be double lung transplantarion. This subject has been thor-oughly discussed elsewhere (223), but recenr experience suggesrsthat single lung transplanration can be life saving (224). The pro-cedure is costly, is hampered by lack of donor availabil itv. andrequires l i fe long immunosuppression.

SUMMARY OF APPROACH

The approach to surgery in rhe patient with COpD is summa-rized below.

Surgery Definitely Indicated

Lung resection. Pulmonary function studies should be performedbefore iung resection. Simple spirometry has the grearesr uri l iryin documenting physiologic operabil ity. FEV, ) 2 L in an adultman or ) 60Vo of predicted is acceptable for pneumonecromy.Values below this suggest that further studies, such as split funi-t ion assessment by quantirative lung scintigraphy and exercisetesting, are warranted.

FEV, predicted after lung resection to be less than 40 to 50goof normal for the patient's sex, age, and height suggests highermorbidity and mortaliry. An exercise VO, of less than tO to tSml,/min per kg of body weight is associared with higher mor-bidity and mortality after lung resection.

All elective surgery. Prophylaxis against deep venous throm-bosis should be given before most procedures that wil l requirepostoperative bed rest or significantly reduce mobil ity. Heparinin low doses seems well accepted for most procedures. Externalpneumatic compression of the lower legs can be used when anti-coagulants are contraindicated.

Surgery Possibly Indicated

In patients with COPD, a preoperative multidisciplinary evalua-tion that includes the primary care physician, consultant pulmo-nologist,/ intensivist, anesthesiologist/anesthetist, and operatingsurgeon is warranted. Consensus as to preoperative physiologicstatus, therapeutic preparation, and posloperative managementis essential.

Lung function srudies such as simple spirometry, includingan arterial blood gas analysis, should be performed before up-per abdominal surgery in parients exhibit ing symproms and signsof obstructive airway disease, if not previously obtained. Thereare no values that preclude a l ife-enhancing operation. Stage IIor III impairment suggests the need for intensive posroperariverespiratory therapy, which may include mechanical venli larion.

Smoking cessation should begin at least 8 wk preoperarivel).Preoperative therapy for elective surgery with antibiorics for

sputum purulence, betar-agonist, or anticholinergic bronchodi-lator aerosols (theophyll ine optional), plus education concern-ing cough and lung expansion techniques, should begin at leasr24 to 48 h preoperatively.

Therapy begun preoperatively should continue for 3 to 5 dinto the postoperative period. Incentive spirometry is noninva_sive and relatively inexpensive and may reduce the incidence ofatelectasis and shorten the hospital stay. Incentive spirometry hasnot been documented to be effective after thoracotomy and prob-ably plays no role in the postoperarive care of peripheral proce-dures.

Postoperative pain control with regional (epidural) anestheticinjection should be performed after thoracotomy. Its value afterupper abdominal surgery is less well documented.

Laparoscopic approaches to upper abdominal surgery maybe attempted. Postoperative morbidity and hospital stay appearlessened in patients without COPD. The effect on the patientwith COPD of CO: absorbed from the peritoneal cavity duringthe procedure is unknown.

Thoracoscopic approaches to small volume (e.g., biopsy) lungresection should be considered.

COPD should not be contraindication to l ife-enhancing car-diac or abdominal vascular surgery, which may be successfullyperformed even in the presence of severe disease. Respiratory careafter coronary artery bypass grafting often includes mechanicalventilatory support. Antibiotic therapy, incentive spirometry, andpain control therapies, as discussed previously, are also warranted.

Resection of bullae occupying greater than one third of ahemithorax and compressing more normal lung tissue may pro-duce improvement in dyspnea and physiologic function.. Recurrent pneumothoraces may be treated by surgical pleuro-

desis or other forms of sclerosis.Lung transplantarion may be considered for COpD in those

patients who meet established criteria.Surgical resection of noncompressive emphysema aimed to

reduce hyperinflation may be effective; further studies are un-der way.

Unproven lndications

Eiective rracheostomy performed in the stable outpatient to re-duce dead space and enhance secretion removal has been aban-doned. Incentive spirometry performed for procedures periph-eral to the thoracoabdominal cavity has no demonstrated role.

Contraindications

Glomectomy is contraindicated for reiief of dyspnea in COpD.

THE FUTURE OF SURGERY FOR COPD PATIENTS

More data are needed on the comparison of quantitative lungscintigraphy versus exercise testing for prediction of physiologiioperabil ity for lung resection. More data are needed on how re-duced exercise oxygen delivery is associated with increased mor-bidity and mortality afrer thoracotomy. We need to improve ourunderstanding of the pathophysiology of diaphragmatic dysfunc-tion after upper abdominal surgery and of therapy to reverse it.More data are needed on the physiological impact of laparoscopicand thoracoscopic procedures in COPD and on the role of re-gional anesthesia after thoracoabdominal surgery. Studies areunder way on surgical approaches to emphysema, but more in-vestigations are needed. More data are needed on the role ofpreoperative preparation techniques, including their efficacy andcomparison of their effectiveness with that of postoperative ma_neuvers.

Amer ican Thorac ic Soc ie t ,

Additional ConsiderationsSLEEP AND THE COPD PATIENT

Patients with COPD seem ro have a higher prevalence of insom-nia, excessive daytime sleepiness, and nightmares than the generalpopulation (225).

Although theophyll ine or bera-agonists could be implicaredin the insomnia, studies with these agents have failed to demon-strate any adverse effects on sleep staging or sleep efficienry. Max-imization of drug therapl' to prevent coughing and shortness ofbreath from disrupting sleep at nighr may help in coping wirhinsomnia. If the patient is severely desaturating in rapid eye move,ment (REM) sleep, supplemental oxygen may help as well. Theuse of hypnotics is controversial. One recent srudy on the useof low doses of short-acting hypnotics has shown that such agentscan be safely toierated without significant worsening of oxygensaturation (226). These agents should be used wirh caution, how-ever, especially in patienrs with CO, rerenrion.

Daytime sleepiness may stem from disrupted sleep due to pul-monary symptoms, but coexistent obstructive sleep apnea (OSA)should also be considered. One night of sleep deprivation hasbeen shown to iead ro small but statistically significant falls inboth FVC and FEV, (227).

Oxygen Desaturation during Sleep

Oxygen desaturation during sleep, especially REM sleep, has beenlong recognized in parients with COPD. Clinical parameters thathave been associated with the presence of nocturnal desatura-tion include daytime hypoxemia, blunted chemosensitivity whileawake, severe dysfunction on pulmonary function resting, andchronic co, retention. None of these characteristics has beenuseful in predicting individual REM desaturarors; in fact, pa-tients have been observed to desaturare with a daytime Po, above60 mm Hg Q28). One mechanism leading to hypoxemia is re-duced ventilation in sleep, especially in REM sleep. lr has alsobeen postulated that the hypoxemia may be related to ventila-tion perfusion imbalance, ahhou_sh this has been diff iculr roprove.

Hemodynamics and Long-term Effects

REM-associated drops in Sa6, are associared rvirh increases inpulmonary artery pressures. It is nor clear if isolated increasesin pulmonary artery pressures during sleep can lead to sustainedpulmonary hypertension. However, recenr srudies of patients u'ithCOPD with nocturnal desaturation and da1'rime Po. levels over60 mm Hg have demonsrrared higher da1'rime resuhing (229\ andexercise-induced (230) pulmonarl 'arrer) pressures in these pa-t ients than in a s imi lar group of par ienrs u 'ho d id nor desarurareat night.

Patients with COPD have increased premarure venrricular con-tractions during sleep, which ma1' decrease in frequencl uhenthese patients are given supplemenral ox1'gen (331).

The effect of nocturnal oxygen salurarion on survival in 9?pat ients wi th COPD has recenr l l 'been reporred (132) . Borh rhemean nocturnal Sas, and the Sao, nadi r dur ing s leep Neresignificantly related to survival. Ho*ever. neirher measure im-proved the prediction of survival over measurements of r ital ca-paci ty or awake Sa6r. Measuremenr of nocturnal Sa6. dur ingsleep therefore cannot be recommended in the routine clinicalmanagement of patients u'ith COPD. Orher groups found rhaisurvival was reduced in parienrs u ith COPD u rrh Pa6- ) 6O mmHg but ( 80 mm Hg, suggesting thar rhere rs a subgroup of pa-t ients wi th COPD in whom i r ma1'be uorrh*hi ie ro determinethe extent of nocturnal oxygen desarurarion (l l- i r. Ho*er.er. this

group could not demonstrate a reductron in mortality with noc-turnal oxygen supplementarion.

The role of sleep studies is controversial. The major sequelaeof nocturnal oxygen desaturation, including pulmonary hyper-tension and an increase in mortality, can usually be predictedfrom the measurement of daytime gas exchange and pulmonaryfunction. At the same time, it is diff icult to recommend routinemeasurement of nocturnal Sag, in patients with COpD with day-time Po, over 60 mm Hg.

Relation of COPD to Obstructive Sleep ApneaSeveral studies have demonsrrated that COPD and OSA can coex-ist, but there is no evidence that this coexistence is more com-mon than would be expected from the relative frequencies of thetwo conditions (230). The significance of the association seemsto be that patients with both disorders seem more l ikely to de-velop pulmonary hypertension and right-sided heart failure thando patienrs who have either condition alone. However, full poly-somnography would be beneficial in those Datienrs with COpDwith symproms suggestive of coexistent OSA.

Research should be undertaken to determine if treatment withsupplemental oxygen- of isolated falls in nocturnal Sao, in theabsence of severe daytime hypoxemia, prevents morbidity andmortality in patients with COPD. If it does, guidelines need tobe established to help identify those patients with daytime po,over 60 mm Hg who should undergo nocturnal Sag, monitoring.

Guidel ines

Nocturnal oxygen should be used in patients who have been dem-onstrated to have significant (( 88go) desaturation during sleep.This can_generally be predicted from daytime hypoxia (Fa6, (5.5 mm Hg). Measurement of nocturnal oxygen saturation in ;a-tients wirh COPD with daytime pa6, ) 60 mm Hg is not ,.com_mended, excepr in parients with unexplained polycythemia orcor pulmonale.

. Full polysomnography should be performed in those patientswith COPD symproms suggesrive of coexistent obstructive sleepapnea.

THE ROLE OF NUTRITION

As manv as 25Va of outpatients with COpD may be malnour-ished; almost 5090 of parients with COpD admitted to the hos-pital have evidence of malnutrit ion (234,235). Sixty percent ofpatients crit ically i l l with COPD and acute respiratory failureare malnourished (236). The incidence of malnutririon also variesu ith the degree of pulmonary function and gas exchange abnor-malit ies. In parienrs wirh chronic hypoxemia as well as in nor-moxemic parients with very severe airf low obstruction (FEV, <l,5fo), malnutririon occurs in 50go - but it is also presentin25Voof patienrs with only moderare airf low obstruction (237).

l\{alnurrit ion is associated with wasting of respiratory mus_cles. resulting in respiratory muscle weakness (2:g). In patientssith m-'-opathy, hypercapnia occurs when inspiratory pressuresare less than one third of the maximal inspiratorv Dressure(Pr*J (239). Bur hypercapnia is found in a majority of patientsuith COPD when inspiratory pressures are less than hatjof nor-mal (240). Thus, hypercapnia occurs with much less respiratorymuscle s'eakness in COpD than other conditions.

Nutrit ional Assessment

Nutrit ional assessment of patienrs with COpD includes severalpossible techniques (241). l,oss of body weight provides a read-

AMERICAN IOURNAT OF RESPIRATORY AND CRITICAL CARE MEDICINE VOL I52 1gg5

i ly accessible parameter of altered nutrit ion status; weight lossin excess of l09o of ideal body weight suggests malnutrit ion.Edema, of course, l imits the uti l i ty of body weight measurementin assessment of malnutr i t ion.

Hepatic secretory proteins, such as albumin, transferrin,retinol binding protein, and prealbumin, are markers of visceraiprotein stores and have been proposed as methods of nutrit ionalassessment; unfortunately, all are influenced by numerous fac_tors in addition to the nutrit ional state. Anthropometry involvesapplication of simple measurements of skin folds ani circum_ferences to assess the body's proportions of fat, muscle tissue,and skeletal mass; while techniques of measuremenr are stan_dardized, interpretations of results remain controversial. Im_munoincompetence is a common accompaniment of malnutri_tion and may be tested by cellular immunity or delayed cutaneoushypersensitivity; depression of cellular immunity is consistentlvassociated with mainutririon, and nutrit ional repletion is as_sociated with improved immunocompetence. The uti l i tv of skintesting is l imired by multiple factors, including technical appli_cation and interpretation of such tests.

Tests of muscle function are also used as markers of nurri_tional sratus. changes in twitch characteristics of the abductorpoll icus muscle occur with stimulation of the ulnar nerve in mal_nourished patients; lack of technical expertise and equipmentlimits widespread clinical application. Respiratory muscli strengthcan be assessed by maximal mouth or transdiaphragmatic p;s-sures, while endurance is assessed by maximal voluntarv venti_lation; both are reduced in malnourished patients. Limitationsof these tecl-rniques are multiple and include the need for alert,awake patients. Other potential limitations include metabolic fac-tors, such as hypercapnia or hypo.ria, as well as medications andintrinsic muscle disease.

In summary, multiple tests are available to assess nutrit ionalstatus. No simple recommendation can be given regarding the"best" test for nutritional assessment. Use of any of thJse meihodscan be appropriare, providing the l imitations are clearlv under-stood.

Dara in hospitalized patienrs generally (not COpD Datientsspecifically) suggesr that weight loss and physical f indings ofweight loss are l ikely to indicate malnutrit ion as reliably

", i.,or.

complex tests of nutrit ional status. And when assessed Uy Uoayweight gain, nutrirional support that reverses protein_calorie mai-nutrit ion may result in small improvements in respiratory mus_cle strength in both ourparient and hospitalized patienis withcoPD (242).

Patients with COPD have increased resting energl, require_ments, approximately l5% above values predicted by Hirris_Benedict equations. This ,,relative hypermetabolism" is explainedby the increased energy needs of the venrilatory musctei (Z+3).The energy cost of respiratory muscles can be approximared fromthe sever i ty of lung hyper inf la t ion.

Nutrit ional Support

Nutr i r ional supporr in COpD has been adminis tered ora l ly wi rhsupplements when sponraneous oral intake is insufficienr is u,ellas entera l ly wi th cont inuous or nocturnal feedings. Aggressiveoral nurrit ional supplementation of parienrs wlrtr COFb aoesresult in improvement of respiratory srrength, but it is laboriousand time-intensive; ofren, it cannot be maintained by the patienrQ4a). The optimal level of caloric supporr for parients wirhCOPD has not been determined; it depends on both rhe degreeof malnutr i t ion and the sever i ty of rhe i l lness.

Establ ishmenr of the mosl appropr iare subsrrare mrx lure, i .e . ,carbohydrates versus fats, for malnourished parienrs *,irh COpDis a lso compl icared. The respi raror l ,quor ienr (Ret is dependenron the substrate used. provision of toral calories that exceed meta-

bolic demand leads ro net l ipogenesis, resulting in a raised Refrom increased carbon dioxide production. Hypercapnia result-ing from increased carbon dioxide production (Vcor) i, nor*uflvavoided-by a compensatory increaie in ventilation, but individ_uals with compromised ventilatory status, such as patients withCOPD are unable ro increase ventilation approprialely. Clinicalsequelae of overfeeding may include precipitation of

-i,ypercuf_

nic respiratory failure and delayed *runing iro. mechanical ven_t i la t ion (245) .

^ It has been suggested rhat patients with COpD might benefit!r9m a high-lipid, low-carbohydrate dier, owing to tfie reducedRQ. However, the clinical benefits of altering fat_to_carbohydrateratios in patients with COpD when caloriei supplied u.. upp.o_priate have not been demonsrrated (246). lnaddition, the incrlasein ventilatory demand attributed to carbohydrate administrationhas been noted only when total calories idministered were farin excess of estimated demands (247). There are thus few data

. to support the use of diets with altered fat_to_carbohydrate ra_tios in patients with COPD. Instead, overfeeding_calories inexcess of metabolic demands _ should be avoided. Evaluationof overfeeding can be determined by metabolic rate and carbondioxide production.

_ Protein requirements in patients with COpD are generally un_changed from routine recommendations. protein, [o*auar, tr",been shown to increase minute ventilation, oxygen consumption,and ventilatory response to hypoxia and hypeic-apnia. tn patientiwith acute.respiratory failure, high leveli bf protein *dv c"uiifurther fatigue, and protein administration may need to Le tern_porally reduced. No data exist suggesring whiih patients wiihCOPD in the ICU setting are at risl-for thls potential complica-tion. If a patient is protein-malnourished,

"n "ttr*pt should be

made to correct that status.

Electrolytes and MicroelementsElectrolyte disturbances are common in patients with COpD andhave potential for significant adverse outcomes. Hypoptrosptrats-mia, hyperkalemia, hypocalcemia, and hypomagnesemia are as_sociated with diminished diaphragmatii iuncti6n, *friieiepie-tion of these elements results in improved function (24g1. Wiiiethese complications apply to all patients receiving nutritional sup_port, patienrs with copD may be at increasedrisk, because ofdecreased respiratory muscle function secondary to the primarylung disease. Monitoring electrolyte levels and pro"iaing sup_plemental elecrrolytes, especially phosphorus, in malnouiishJapatients should be routine in pit i lnts with CbpD and respira_tory failure.

General Guidel ines

Patienrs with COPD should be instructed on good dietary habits.Their weig.ht should approximare ideal body-weight. If ihet;;;malnourished, attempts should be made to restore nutrit ionalbalance. Several smaller meals a day may help maintain caloricneeds bur avoid undue dyspnea. Forced nutrit ion or special dietsare not recommended at the present t ime.

COPD AND AIR TRAVET

Commercial airl ine travel exposes passengers to hypobaric hyp_oxia, since aircraft cabins are not routinily pressurized to sealevel (249). In patients who have.o.np.nrut.i bopo at sea rever,lowering the partial pressure of oxygen in the aircraft cabin canproduce severe hypoxemia (250). The appearance of dyspnea,wheezing, chest pain, cyanosis, and righi heart failure ."n t.udto urgenr requesrs for oxygen during fl ights (251). physical exer_tion during a flight can increase the risk of exacer'bation of symp_toms (250). Death due ro purely respiratory causes .arely oicuis

American Thoracic Society

dur ing f l ight , but the f requeno'of less morbid evenrs, inc ludingworsening symptoms that present or persist after leaving theplane, may be underreported by patients. The proportion of pa-tients suffering cardiac events during air travel who have COPDas a co-morbid condir ion is nor known. .

Cabin Pressures

The cabin pressure l imit for a passenger aircraft is 10,000 ft(3,048 m). Newer aircraft are usually pressurized to between 5,000and 7,000 ft (1,524 to 2,134 m). For the prefl ight evaluation ofmost patients, clinicians should consider 8,000 feet (2,438 m) ofalt itude above sea level as a realistic "worst case scenario." Manyfactors influence the cabin alt itude for a particular f l ight. Theseinclude the type of aircraft, the fl ighr alt itude and fl ight plan,weather conditions, and the operation of the aircraft. In onereport, half the fl ights of B-767 aircraft produced cabin condi-tions equival ent to 7 ,412 ft or more above sea level (252). Shortfl ights usually reach lower cabin alt itudes, but exceptions do oc-cur. Helicopters and small aircraft, including commuter planes,often do not have pressurized cabins. Consequently, travel bythese aircraft can also entail significant exposure to alt itude con-ditions. Whenever specific information concerning the cabin al-t itude can be ascertained bv the clinician. it should be integratedinto decision making.

Prefl ight Assessment

Patients with COPD should have a prefl ight assessment that in-cludes estimation of the expected degree of hypoxemia at alt i-tude, identif ication of co-morbid disease conditions, and provi-sion of an oxygen prescription if necessarl '. Documentation ofrecent clinical condition and laboratory tests, panicularly if thepatient is traveiing abroad, and counseling are also desirable ele-ments of prefl ight care.

Estimation of the degree of hypoxia at altitude is accomplishedvia hypoxia inhalation testing (253,254) and the use of regres-sion formulae Qa9,253). The h-vpoxia inhalarion test (HIT) as-sesses the effects of hypoxemia on rhe patienr with COPD. Thesubject is exposed to a hypoxic gas mixrure equivalent to 8,000 ftof alt itude (an oxygen fraction of l5.lVo ar sea level) for a mini-mum of l5 min (255) . Dur ing the tesr , rhe par ient is assessed forclinical and electrocardiographic changes and arrerial blood gases.Whi le the HIT has the advantage of assessing Ihe acute c l in ica lef fects of hypoxia, one stud1, of par ients u ' i rh COPD compar ingthe test wi th actual hypobar ic exposures of -<.41_3 f r (1.650 m)found var iabi l i ty of up ro l l mm Hg in Pae, r 'a lues. The HITis not performed in manl'cl inical laborarories in rhe Unircd Sraresand is not recommended for routine use.

Pat ients wi th COPD who ma1'be considered candidares forthe HIT inc lude: ( / ) those wirh coexisr ing condir ions lhar ma)be affected by hypoxemia, e.g., coronar) arrer] disease; (2) thosewho have mani fested symptoms dur ing previous a i r t ravel ; (J)those recovering from acute exacerbations: (/) those s ho develophypovent i la t ion wi th oxygen adminisrrar ion: and (5) rhose shorequire additional reassurance before underraking air rravel.

Regression equat ions of fer the opporruni rv ro compare a pa-

tient with a group of patients with similar clinical characreris-tics who have been previously studied during exposure to hyp-oxia (249,253). While regression equations may provide a morephysiologic basis for rhe effects of high altitude than the HIT(249), the regression approach does not assess the individual'ssusceptibil iry to the development of symptoms or electrocardio-graphic changes during hypoxia. The alveolo-arterial oxygen ten-sion (A-a Or) gradient and a-A O, ratio generally have no advan-tages over regression equarions (249,253).

Oxygen Supplementation

It is currently recommended that the pacl, during air travel shouldbe maintained above 50 mm HgQ5q. While 2 to 3 L of oxygenby nasal cannula wil l replace the inspired oxygen partial pres-sure lost ar 8,000 ft compared with sea level (256), lesser incre-ments of oxygen wil l maintain the Pa6, above 50 mm Hg in manypatrents.

Patients with COPD receiving conrinuous oxygen ar homewill require suppiemenration during fl ight. Such patienrs shouldrecerve grearer oxygen supplementation during the fl ight thanat sea level. Increments equivalent to I or 2L of oxygen by nasalcannula during fl ight should suffice for most patients. patientswill also require additional oxygen supplementation if the eleva-tion at the destination is significantly greater than at home.

The Federal Aviation Administration requires a physician'sstatement of oxygen need for a patient to receive continuous oxy-gen during fl ight. There is no uniform airl ine request form; theairline must be contacted by the patient to determine what is re-quired. As the airl ines do not provide oxygen for ground use interminals, patients who require continuous oxygen should be ad-vised to make plans for such locations. The American Lung As-sociation provides patienr education material for individuals whotravel with oxygen ("Airline Tiavel with Oxygen"); it includes re-quirements for each commercial carrier as well as the costs ofoxygen during flight, data helpful to both physicians and patients.

Many ambularory patients with COpD nor receiving oxygenat home can rolerate PO: values below 50 mm Hg for brief periodsof t ime without serious consequences. patients with COpD butwithout co-morbid disease who have previously traveled with-out incident and who are currently clinically stable comparedwith their previous air travel may be permitted to travel by airrv i rh l iu le r isk.

Airl ine rravel is safe for most parients wirh COpD. Hypox-emic patients should be evaluated clinically, and a decision shouldbe made regarding 02 requirements. The use of regression for-mulae helps ro assess those needs. Those patients already on ox-1'gen should be instrucred ro increase their O, by I or 2L/minuhi le in f l ighr .

Additional dara are desirable. lnvestigators shouid explore theph-vsiologic consequences of allowing pa6, values continuouslybelou' 50 mm Hg for 4 ro l2 h in parienrs with uncomplicatedCOPD, as well as rhe effects of acure hypoxia on patients withCOPD who are hypercapnic or have co-morbid disease states.It is also helpful ro derermine the frequency of adverse clinicalevents occurring in patients with COpD who engage in air travel.

AMERICAN IOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE VOL I52 1gg5

Ethical lssues in COPDBecause COPD affecrs patients at more advanced ages, frequentlyprogresses, and may require highly expensive and prolonged life_preserving medical technology, management and care frequentlypresent ethical dilemmas. Although fundamenml principles ofhealth care ethics have been commonly adopted and widely dis-seminated through professional medical societies during the lastdecade, the decision to withhold, l imit, or withdraw care in pa-tients with COPD remains complex and diff icult. Deliberationsconcerning individual patients require careful analysis involv_ing COPD survival sratistics, quality o.f l i fe, community healthcare resources, and economic aspects of care.

The chief ethical concern in COPD focuses on the institutionand withdrawal of mechanical ventilation, particularly in the hos_pital setting.

Prediction and Prognosis

The onset of respiratory failure is a major event in the l ife ofa patienr with COPD. The complex decision ro intubate and me_chanically venti late such patients requires efforts to balance theimmediate l ife-saving benefits of venti latory support with thepoor long-term prognosis of severe COpD and the discomforts.potential risks, expense, and resource uti l ization of prolongedmechanical venti lation.

Clinicians commonly attempt to determine the value of me-chanical venti lation in individual parients by subjectively evalu-ating ciinical features of the acute respiratory failure and estimat-ing the l ikelihood of survival after intubation. Unfortunatelv.physicians using subjective bedside judgment cannot u.curut.iypredict the l ikelihood of survival in these circumstances. Thereis no correlation between predictive accuracy and physicians'ex-perience or level of training (257).

Furthermore, no studies have established that objective mark-ers of the severity or specific clinical features of the resniratorvfailure can predict the probabil ity of survival after mechanicilventi lation. No correlation exists, for instance, between short-term survival and admission arterial blood gas results, admis_sion spirometric values, hematocrit, patient age, or number ofprevious admissions for exacerbations of COpD (25g, 259). Mul_tivariate analysis techniques may improve predictive accuracy forlong-term survival in intubated patienrs, but they sti l l misclas_sify up to 23Vo of patients. Scoring systems such as the Simoli_fied Acute Physiology Score are weak predictors of shorr_termoutcome in patients with COpD and respiratory failure (260).

Clinical data do indicate, however, that outcome followinean episode of respiratory failure correlares with the severity oTthe underlying COPD and patient acriviry levels when the oa_tient is well or during stable periods of the disease (259). If non_pulmonary co-mqrbid conditions, such as gastrointestinal hemor_rhage, pulmonary embolism, or coronary heart disease, are alsopresent at the onset of respiratory decompensation, that factorcontributes to poor patient outcome (25't ,259\. Housebound pa_tients with severe, end-stage lung disease and co-morbid condi_tions have a worse short-term prognosis than more active pa_tients with less severe underlying pulmonary impairment, givensimilarly severe episodes of respiratory failure.

Analysis of the typical clinical outcome of patients with COpDwho present with acute respiratory failure can assist decision mak_ing for intubation and mechanical venti lation. Berween 75 and90s/o of such patients who are mechanically venti lated surviveto hospital discharge (258,259,261). Two-year survival for pa-tients followed from the onset of an episode of acure respiraroryfa i lure has ranges berween 28 and 70Vo (260,262). Indeed,among patients presenting with acute respiratort, failure due ro

a variety of causes, parients with COpD have the hiehest sur_vival rate (263).

Overall, the long-term prognosis of patients surviving mechan_ical venti lation is similar to that of patients with the saire degreeof underlying respiratory impairment who have not requiredhe_.l3ltJul ventilation (260). Thus, ir appears that patients withCOPD whose course is complicated by acute respiratory failurerequiring life-support measures do not have agenerally grim prog-nosis. Consequently, there is no fundameniol ethicai ditimiain considering all patients with COpD for intubation and me_chanical venti lation.

. Nevertheless, patients who have poor baseline function, mar-ginal nutritional status, severely restricted activity, and inexora-ble deterioration of late-stage pulmonary dysfunction may electto forgo intubation if, in the judgment of both patient ani phy_sician, it will only temporariiy interrupt the rerminal ptraje ofthe disease.

Advance Directives

Physicians have a solemn obligation ro assist their patienis withCOPD. in formulating directives in advance of respiratory decom_pensation. In counseling patients regarding the vilue of mechan_ical venti lation and the implications of

-declining l ife-support

measures, physicians have the responsibility, as outlined in theAIS statement on Withhotding and Withdraiing Life_sustainingf!":?Iy (2-64) and by the Presidenr,s Commissi,on io, rhe Stud!of Ethical Problems in Medicine and Biomedical and BehavioralResearch (265), to ensure that:

"(/) the patient has decision-making capacity; (2) the patienthas been informed regarding his or her diagnojij, prolnosis,the risks, benefits, and consequences for the iull ranie oi avail_able medical interventions including the option of rio therapy;(3) the parient has received from the physician professionairecommendations regarding the medical choices available, in_cluding the use of life-sustaining therapy, based on knowledgeof both the medical siruation and the values and goals of tf,epatient" (263).

-. As observed earlier, survival afte r respiratory failure is closelylinked to baseline function and the presence oi co-morbid con_ditions. Decisions regarding l imitarions of care are best madeduring stable periods before respiratory failure or any other l i fe_threatening condition has occurred. patients who choose to forgolife-support measures should be encouraged to outl ine theirwishes in as much detail as possible in a wriiten instrument suchas a "living will." Patients with COPD should specify their healthcare preferences for the several clinical situations likely to be en_countered by those with pulmonary impairment, including intu_bation, mechanical venrilation, cardiopulmonary resuscitation,tracheotomy, and long-term life support with diff icult weaning.Assistance and open discussion should be offered by the physi_cian during the formulation of the advance directive to pieu.ntmisinterpretation and enhance proper implementation of the pa_tient's wishes. The patient should also be encouraged to sharethese preferences with a trusted family member, friend, or otherindividual who can be designated as a surrogate decision makerfor health care matters through a durable power of attorney (263).

When advance directives have been properly established orthe wishes of an informed patient with decision-making capac_ity made known, respect for patient autonomy requirei physi_cians caring for the patienr to honor the patient's right to-foigomedical intervenrion (263). Honoring such requests is consid_ered neither particjpation in assisted suicide nor active euthanasia.

American Thoracic Societv

However, a physician faced with a crit icallf i l l patient shoulddetermine that requests to forgo care are reasonable under theclinical circumstances and derive from deep-seated values andappropriate responses to the severity of underlying disease ratherthan from endogenous depression or from temporary conditionsofpain, fear, depression, or anxiety during episodes ofacute respi-ratory faiiure. A physician who has personal ethical or religiousvaiues that do not permit compliance with a patient's well-con-ceived request to forgo l ife support should transfer the patient'scare to another physician who can honor the patient's directives.

It should be recognized that there is no ethical differencebetween withholding and withdrarving l ife-support measures inpatients with acute respiratory failure. Ethical principles under-ly ing the decis ion to wi thhold in tubat ion and mechanical vent i -lation apply equally when patients or proxies request discontinu-ance of care for patients with terminal disease or a progressivedegenerative condition who have no hope for an acceptable andmeaningful recovery. Patients electing to have ventilatory sup-port withdrawn may request and receive adequate sedation andanalgesia to extinguish all pain and suffering during the dyingprocess even if such treatment accelerates their imminent death.

Basic Guidel ines

Because patients who have COPD with acute respiratory failurehave a favorable short-term prognosis after intubation and me-chanical venti iation, no fundamental ethical dilemma exists inconsidering all such patients for crit ical care support. No clini-cal features of the respiratory failure should be used to selectpatients for intubation, because these features do not correlatewith outcome. However, the severity of the patient's underlyingCOPD and the presence of co-morbid conditions do correlateto a degree with clinical outcome and allow physicians to assistpatients with advance directives when patients are well or in sta-b le condi t ion. Al though not absolute, the pr inc ip le of auton-omy directs physicians to honor requests of patients with severedisease to forgo l ife-support measures, even when that decisionwill cause a patient's death. Physicians cannot honor such re-quests without further patient evaluation, however, if the requestsappear inappropriate in the clinical circumstances or appear tobe derived from endogenous depression or acute pain, fear, oranxiety in patients u,ithout terminal disease.

Future Research

During the last decade, there have been major advances in arriv-ing at consensus regarding ethical principles of life supporr. Theseprinciples have been disseminated through professional socie-ties and frank exchanges among physicians, patients, and others,with the aim of ensuring respect for parient autonomy. Sti l l , themajority of patients with COPD have nor formulated advancedirectives and have not discussed rhese matters with their physi-c ians.

Future investigative efforrs should focus on the behavioral in-fluences that keep patienrs and physicians from exploring thesequestions in a timely fashion. Greater insights are needed notonll ' to promote these discussions among patients, families, andphysicians but aiso to develop ways ro address such intangiblesas quality of l i fe, cosl of l i fe-years saved, and impact on mortal-i ty in connect ion wi th these decis ions.

FUTURE DIRECTIONS

Knowledge is constantly evolving. This is parricularly rrue of dis-ease entit ies, such as COPD, thar affl icr and u,i l l conrinue roplague many individuals. \4/ith passage of rime, many of the con:cepts expressed here wil l change. Periodicall l ' , rhis sratement wil lbe modi f ied to inc lude those chanses rhar wi l l srand cr i i ica l sc i -entif ic scrutiny.

This statement was prepared by an Ad Hoc Commit tee of the Scient i f ic As_sembly on Cl in ical problems. Members of the Commit tee were:Barto lo-me R. Cel l i , M.D. (Chairman), Cordon L. Snider, M.D., lohn Heffner,M.D., Br ian Tiep, M.D., t rwin Ziment, M.D., Barry Make, M.D., Sidney Bra_man , M .D . , Ge ra ld O l sen , M .D . , and yancy ph i l l i p s , M .D .Members of the Sub-Commit tee were:Definition, Epidemiology, pothophysiology, Diognosis, ond Stogingsjames Stol ler , M.D., Rebecca Bascom, M.D., and Benjamin Br j r r6ws, M.D.Sleep ond COPD, Nutrition ond COpD, Air Trovel and'COpDSusan P ing le ton , M .D . , R i cha rd M i l lman , M .D . , Edward F i she r , M .D . , andThomas D i l l a rd , M .D .PharmocotheropyMichael L i t tner, M.D., Leonard Hudson, M.D., and Richard Matthay, M.D.Surgery ond COPDCordon_T Ford, M.D., Joseph Locicero, M.D., Michael Matthay, M.D., andjoseph Z ib rak , M .D .Oxygen TheropyRichard Carter , M.D., Lesl ie Hoftman. ph.D., R.N., and Eugene Fletcher, M.D.lnpotient ManogementDean Hess, M. Ed. , R.R.T. , David p ierson, M.D., and JudyTietsort , R.N., R.R.T.Outpotient Monagement ond RehobilitotionChar les Emory, M.D., Deborah Horve, p.T. , Donald Mahler , M.D., Cathy Musel-man , C .R .T .T . , Ma rga re t N ie l d , ph .D . , R .N . , and And res R ies , M .D . , M .p .H .

References

l. American Thoracic Sociery. t962. Chronic bronchitis, asthma, and pul-monary emphysema: a statement by the Committee on DiagnosticStandards for Nontuberculous Respiratory Diseases. lrz. Rev. Rapir.Dis. 85:162-768.

2. Snider, G. L. , J . Kle inerman, W. M. Thur lbeck, and Z. K. Bengal i .1985 . The definition of emphysema: repon of a National Hean, Lungand Blood lnsrirute, Division of Lung Diseases, Workshop. z4m. Rev.Respir . Dis, 132:182-185.

3. Higgins, M. W., and T. Thom. 1990. Incidence, prevalence, and mor-' tality: intra- and inrer-county differences. InM. J. Hensley and N. A.Saunders, editors. Clinical Epidemiology of Chronic Obstrucrive pul-monary Disease. Marcel Dekker, New york. 23-43.

4. Feinlieb, M., H. M. Rosenberg, J. G. Collins, J. E. Delozier, R. pokras,and F. M. Chevar ley. 1989. Trends in COpD morbid i ty and morta l -i ty in the Uni ted Stares. Am. Rev. Respir . Dis. 140:S9-Slg.

5. Speizer, F. E. I989. The rise in chronic obstrucrive oulmonarv diseasemorrality: overview and summary. Am. Rev. Respir. Dk.liO:StOe-s 1 0 7 .

6. U.S. Surgeon Ceneral. 1984. The Health Consequences of Smoking:Chronic Obstrucrive Lung Disease. U.S. Department of Health andHuman Services, Washington, D.C. DHHS publication No. g4-50205.

7. Sherri l l , D. L., M. D. Lebowitz, and B. Burrows. 1990. Epidemiologyof chronic obsrrucrive pulmonary disease. Ciin. Chest Med. ll:Ji.S-3 8 8 .

8. Davis, R. M., and T. E. Novotny. 1989. Theepidemiology of ciga-rette smoking and i ts impact on chronic obstrucrive pulmonary dis-ease. Am. Rev. Respir. Dls. 140:582-584.

9. Buist, S. A. 1994. Smoking and other r isk facrors. In J. F. Murrayand J. A. Nadel, edirors. Texrbook of Respirarory Medicine, 2nded. W. B. Saunders, Phi ladelphia. 1259-1291 .

10 . Or ie , N . G. , H . J . S lu i te r , K . De Vr ies , C. H. Tammel ing , and J . Wi r -kop. 1961. The host facror in bronchir is., In N. C. Olie and H. J.SIuiter, editors. Bronchit is. Royal Vangorcum, Assen. 43-59.

I l . Bur rows, B . , J . W. B loom, G. A . Traver , and M. C. C l ine . I9g i . Thecourse and prognosis of dif ferenr forms of chronic airway obsrruc-tion in a sample from rhe general popularion. N. Engl. J. Med.3 I 7 : 1 3 0 9 - l 3 1 4 .

I2. Redline, S., and l . B. Tager. 1989. The relat ionship ofairway reactiv-i ty to the occurrence of chronic obstructive pulmonary disease: anepidemiologic assessment. In M. J. Hensley and N. A. Saunders,editors. Cl inical Epidemiology of Chronic Obsrrucrive pulmonaryDisease. Marcel Dekker, New york. 169-199.

13. O'Connor, C. T., D. Sparrow, and S. T. Weiss. 19g9. The role of al-Iergy and nonspecific airway hyperresponsiveness in the pathogene-sis of chronic obstructive pulmonary disease. An. Ret,. Respir. Dis.t40:225-252.

14. Burrows, B. 1991. Epidemiologic evidence for different rypes of chronicairflow obsrruction. Am. Ret,. Respir. Dk. 143:1452-1455.

15 . Tashk in , D. P . , M. D. A l rose , E . R. B leeker , J . E . Conner t , R . E .Kanner, W. W. Lee, R. Wise, and rhe Lung Health Srudy Rescarch

AMERICAN IOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE VOL . I 52 1995

Group. 1992. The Lung Heal th Srud1,: a i rway responsiveness [o in_haled merhachol ine in smokers wi th mi ld to moderate ai r f low l imi_tauon. Am. Rev. Respir . Dr i . I45:301_310.

16 . Robo r , E . , P . D . So r t i e , N . J . Johnson , C . S . C love , and D . O . T reas -ure. 1988. A morta l i ry , s tudy of one nr i l l ion persons by demographic,socia l and economi,c- facrors: 1979_gl fo l lowup. U.S. LongirudinaiMorral i ty Study, NH 88. 2896.

17. Redl ine, S. , and S. T. Weiss. 19g9. Cener ic and per inaral r isk facrorsfor the development of chronic obstrucr ive pulmonary disease. /zM. J. Hensley and N. A. Saunders, edj tors. Cl in ical epia.rdologyof Chronic Obsrruct jve pulmonary Disease. Marcel I jekker, Ne"wY o r k . 1 3 9 - 1 6 8 .

18 . Backe r , S . , R . Jo rdan , and C . O . O r l ando . 19g9 . t n v i r r o ozone ex_posure inhib i ts mi togen- induced lymphocyre prol i ferar ion and IL_2production. J. Toxicol. Environ. Health 26:469_4g3.

19. Travis, J. 1989. Alpha,-proteinase inhib i tor def ic iency. /n M. Mas_saro, ediror . Lung Cel l Bio logy. Marcel Dekker, New york. 1227*t246 .

20. Brant l1, , M., T. Nukiwa, and R. C. Crystal . l9gg. Molecular basis ofa lpha,-ant i t rypsin def ic iency. Am. J. Med. g4:13_31.

21. Snider, G. L. 1989. pulmonary disease in a lpha,-ant i r rypsin def ic iency.Ann. Intern. Med. l l l :957_959.

22. Feld, R. D. 1989. Hererozygosi ty of a lpha,_anr i r rypsin: a heal th r isk?Cri t . Rev. Cl in. Lab. Sci . 2 i :46t-4g1.

23 . Tou 'n l ey , R . C . , J . G . Sou tha rd , p . Rad fo rd , R . J . Hopp , A . K . Bew ,rra, and L. Ford. 1990. Associat ion of MS pi phenorype wi th a j r_way hyperresponsiveness. Chest 9g:594_599.

24. Flercher, C. M., R. pero, C. Tinker, and F. E. Speizer. 1976. The Nar_ural Hisror) 'of Chronjc Obsrrucr ive Lung Disease in Working Menin London. Oxford Unjvers i ty press, Niw york.

25 . Bu r rows , B . , R . J . Knudson , A . E . Cam i l l i , S . K . Ly l e , and M . D .Lebowitz. 1987. The ,horse-racing

ef fecr ,and predict ing decl ine inforced expiratory volume in one second f rom si reening spirometry.Am. Rev. Respir . Dis. 135:7gg_793.

26. Burrows, B. 1990. Airway obsrruct ive diseases: pathogenet ic mecha-nisms and narural h istor ies of the disorde rs. Med. Ci in. North Am.74:547 -560.

27. Pero, R. , F. E. Speizer, A. L. Cochrane, F. Moore, C. M. Fletcher,C . M . T i nke r , I . T . T . H igg ins , R . C . C ray , S . M , R i cha rds , J . G i l _l i land, and B. Norman-Smirh. I993. The re levance in aduhs of a i r_flou' o!51rug116n, but not of mucus hypersecrerion, to mortality fromchronic Iung disease: resuhs f rom 20 years of prospectrve observa-t ion. Am. Rev. Respir . Drs. l2g:491-500.

28. Ebi-Krysron, K. L. 1989. predicr ing l5_year chronic bronchir is mor-ralirl, in the Whirehall srudy. J. Epidemiot. Commun. Health 16g_172.

29. Lange, P. , J . Nyboe, M. Appleyard, G. Jensen, and p. Schnohr. 1990.Relation of ventilarory impairmenr and of chronic mucus hypersecre-t ion ro morta l i ty f rom obstruct ive lung disease and i rom af l causes.Thorax 45:579-585.

30 . Bu i s r , A . S . , W . M . Vo l lme r , and L . R . Johnson . 19g4 . Does thes in -gle breath N, rest idenr i fy the suscept ib le indiv idual? Cftesr g5:10S.

3 l . Anthonisen, N. R. 1989. prognosis in chronic obslrucr ive pulmonai l ,disease: resuhs from multicenter clinical trials. Am. Rev. iiespir. Dii.I 33 :595 -S99 .

32 . An thon i sen , N . R . , J . E . Conne t t , J . p . K i l ey , M . D . Ahose , W. C .Ba i l e1 , , A . S . Bu i s r , \ f , . A . Conway , p . L . En r i gh r , R . E . Kanne r ,P . O 'Ha ra , P . O . Scan lon , D . p . Tashk in , R . e . W ise , f o r r he LungHeal th Studv Group. 1994. The ef fecrs of smoking rn lerventron andrhe use of an inhaled anr ichol inergic bronchodi laror on lhe rare of

- dec l i neo fFEV , : r heLungHea l rh S tudy . J .A .M .A .272 : t 49 j - 1505 .33 . F l e r che r , C . , and R . pe ro . l 9 j 7 . The na ru ra l h i s t o r y o f ch ron i c a i r -

f lon obstrucr ion. Br. Med. " / . l :1645_164g.34 . Cam i l l i , A . E . , B . Bu r rows , R . J . Knudson , S . K . Ly l e , and M . D .

Lebowirz. I987. Longirudinal changes in forced expirarory volumein one second in adui ts: ef fects of smoking and smoking cessar ion.Am. Ret ' . Respir . Dis. 135: j94- j99.

35 . H igg ins , M . U / . , J . B . Ke l l e r , M . Backe r , W . Howar r , J . R . Land i s ,H . Ro rman , J . C . Weg , and I . H igg ins . 19g2 . An i ndex o f r i s k f o robs r ruc r i ve a i rways d i sease . Am. Re r , . Resp i r . D i s . 125 :144_15 j .

36 . Pos rma , D . S . , and H . J . S l u i r e r . 19g9 . p rognos i so f ch ron i cobs t ruc_t ive pulmonary disease: the Dutch exper ience. Am. Rer, . Respir . Dis.I 40:S I 00-S I 05.

37 . Thu r l beck . W. l \ 4 . 1990 . pa rho log l , . o f ch ron i c obs r ruc r r ve pu tmona r ) ,disease. Ciln. Chesr Med. lt:3g9-404.

38 . N iewoehne r , D . E . , J . K l e i ne rman , and D . B . R i ce . 1974 . pa rho log i cchanges in the per ipheral a i ru,ays of young c igaret te smokers. N.Engl . J . Med. 291:755-758.

39. Naga i , A . ,1 . yamawak i , T . Tak izawa, and W. M. Thur lbeck . 1991.Alveolar attachm_ents in emphysema oihur"n lungs. Am. Rev.Respir. Dis. 144:8gg_g91.40. Wrighr, J. L., p. Caclg, e. Churg, T. V. Colby, and J. Myers. 1992.Diseases of the small airways. f m. Rev. niipir. Oir. t46:24O_262.41 . Thur lbeck , W. M. 1993. Overv iewof thep" i t -o iogyof pu lmonaryem-physema in the human. Ctin. Chest Uea. q:T.l_ZSO.42. Thurlbeck, W. M. 199.1 . pattrotogv of ctrr lnic

" irno* obstruction.

1l \. S Cheriack,_editor. Chronlc OUrti".ii". pulmonary Disease.Saunders, phi ladelphia. 3_20.43. Nagai, A., w. W. West, and w. M. ihurlbeck. t9g5. The NationalInst irutes of Health inrermirtenr posir ive_;;; ;sure breathing tr ial :pathology srudjes, ll:.correlarion berween mirphotogic findings, clin_ical findings, and evidence of expiratory-air'fio* obstruct ion. Am.' Rev. Respir. Dis. 132:946_953.44. Hale, K. A., S. L. Ewing, B. A. Cosnell , and D. E. Neiwoehner. 19g4.Lung disease in long+erm cigarette smokers with and without chronicairflow obstruction. Am. Rev. Respir. Dis.'izO:ll6_t2t.45. Martol i , S., S. Mianre,.F. Calabro, U. l . / .r i i r i i , A. Fasol i , and L.Allegra. 1990. Bronchial.epithelial cells ."poJ',o,ro.yanates poten-riate acrivation and proliieratio" oiT_;;il;.;rn . J. physiot. 259:L320_L327.46. Anthonisenl N. R., E. C. Wright and the IppB Trial Croup. 19g6.Bronchodilator response in chronic obstructiue pulmonary disease.4 /. banders, C. l99l . The^radiographic diagnosis of emphysema. Radrbl..- Ctin. North Am.29: l0t9:1d30.

48. Klein, J. S., c. Camsu, W. n. wJU, J. A. Colden, and N. L. Muller.I 992. High-resolurion CT aiag.nosis of impfiy.era in sympromaricparienrs wirh normal _chesr raiiographs

",iJ'i*f ",.a l";hffr;;;;

.- capaciry. Radiology lg2:gl7_g2t.49. Bates, D. V. 1999. Respirat-ory Function in Disease, 3rd ed. W. B. Saun-ders, New york. 172_1g7.50. Reynolds, H. y. 1991. Antibioric treatment of bronchir is and chroniclung disease. ./n N. S. Cherniack, Ji;;.-a;ionic obstructiue pul-

monary Disease. W. B. Saunders, phi ladelphia. 456_461.5 L Hodgkin, J. E. t990. prognosis jr .h;; ; i ; ;r , ;uct ive pulmonary dis--- ease. Clin. Chest Med. ll:555_569.52. Sweer, L., and C. W. Zwillich. t f9O. Dyspnea in rhe parient with chronic-- . _obstructive pulmonary disease. Ci^. Cnii i - i"a. ^:4l i_M5.53. Mahler, D. A., D. u. *ei"u"ie, c. x. i i rrrr, ' l ldA. R. Feinstein.1984. The measuremenr ot avipn.", lonr.nil,-intrrour.rver agree_

3e11' a1d phvsiologic correlaris of two ne* .iinicat i"ai*.i. ii"rl85 :75 I -758.54. American Thoracic Society. l99l. Lung function test ing: select ion of' reference values and-rnterpretative strategies (statement). /4m. Rev,Respir. Dis. t44tl20t_12t8.55. Dosman, J. A., S. R. Comez, and C. Zhou. t990. Relat ionship be-

:::.:.i,':111*sp.onsiveness and the developmeniof chronic obstruc_56. Higgins, M. I 991. Risk factors ass;.i";; ;;; ;;;;nic obstructive lung__ _ disease. Ann. N.y. Acad. Sci. 624:-t_17.57. Petty, T. L., W. Si lvers, and R. E, Sranford. t9gZ. Mild emphysemais associated with red_uced .f"rti. ,e.oii

"nJ i'n.-r."seO lung size, butnot with airflow limitation. Am. Rrr. i;;;;: brs. t36:867_876.58. Pederson, 1., J. Wil l iams, and N. Lefooe. I6AO. Smot<ing cessationamong pulmonary patients as_relared to type of iesprratory disease

-^ ^ 3nd demographic variabres. con. l. iuitii iealth 7r:r9r-194.59. Edmunds, M., H. Conner,.C. Jones, n. C"i"V.b, ""d

H. Waranch.1991. Evaluation of u,Tll]i..9Tpon.nt groupirnikrng cessarion pro-gram. prev. Med. 20:404_413.60. Koltke, T. E., R. n\. Bart ista, and C. H. DeFriese. l9gg. Attr ibutesof suciessful smoking cessation inrerventions irt medical practice:_ a mera-anali ,sis of 39 control led tr ials. J.A.M.A. 259:2gg2_2gg9.61. Lee, E. W., and C. E. D,Atonzo. f S93.-Cig"r; i i . s.oring, nicotine

:99i:lioL and its pharmacologic rreatm-ent. Arch. Intern. Med.I 53 :34-48.62. Tang, J. L., M. La*,, and N. Wald. 1994. How effect ive is nicoriner,ep-I1:e1e!r rherapy in helping people ro srof smotin g? Br. Med.J . 3 0 8 : 2 1 - 2 6 .63. Prochaska, J. O.. and M. C. Coldsrein. 1991. process of smoking ces_sarion. C/ln. Chest Med. l2:i27_735.64. Fagersrrom, K. O., and N. C. Schneider. 19g9. Measuring nicotinedependence: a revieu,of the Fagerstrom Tolerance euestionnaire.J. Behav. Med. t2:t'gi_tl2.6-s. Fiore, M. C., D. E. Jorenby, T. B. Baker, and S. L. Kenford. 1992.Tobacco dependence and the nicotine patct, cl inical guidel ines for

American Thoracic 5ociety

effective use. J. A. M. A. 268:2687 -2694.66 . Kan fo rd , S . L . , M . C . F i o re , D . E . Jo renb l ' , S . S . Sm i th , D . We t t e r ,

and T. B. Baker. 1994. Predict ing smoking cessat ion: who wi l l qui twi th and wi thout the nicor ine parch. J.A.M.A. 271:589-594.

67. Fisher, E. , D. Haire-Joshu, C. Morgan, H. Rehberg, and K. Ross.1990. Smoking and smoking cessat ion. Am. Rev. Respir . Dis. 142702-730.

68. Schwartz, J. L. I991. Methods for smoking cessation. C/in. Chest Med.12:737 -7 53.

69. Ferguson, C. T. , and R. M. Cherniack. 1993. Management of chronicobstructive pulmonary disease. N. Engl. J. Med. 328 1017-1022.

70. Chapman, K. R. 1991. Therapeut ic a lgor i thm for chronic obstruct ivepulmonary disease. Am. J. Med. 9 l : l7S-23S.

71 . Dompe l i ng , E . , C . P . van Schayck , P . M . van Grunsven , C . L . A .van Herwaarden, R. Akkermans, J. Molema, H. Folger ing, and C.van Weel . 1993. Slowing the deter iorat ion of asthma and chronicobstructive pulmonary disease observed during bronchodilator ther-apy by adding inhaled corticosteroids: a 4-year prospective study.Ann. Inlern. Med. l18:'770-178.

72. Yan Schayck, C, P. , E. Dompel ing, C. L. A. van Herwaarden, H.Folger ing, A. L. M. Verbeeck, H. J. M. van der Hoogen, and C.van Weel . 1991. Bronchodi lator t reatment in moderate asthma orchronic bronchi t is : cont inuous or on demand? A randomized con-trolled study. Br. Med. J. 303:.1426-1431 .

73 . McKay , S . E . , C . A . How ie , A . H . Thomson , B . Wh i t i ng , and C . J .Addis. I 993. Value of theophylline treatment in patients handicappedby chronic obstructive lung disease. Thorax 48:.227-232.

7 4. Ziment, I. I 990. Pharmacologic therapy o f obstructive airway disease.Ctin. Chest Med. ll:461-486.

75. Sul l ivan, P. , S. Bekir , Z. Jaf far , C. Page, P. Jef fery, and J. Costel lo.1994. Ant i - inf lammatory ef fects of low-dose oral theophyl l ine inatopic asthma. Lancet 343:1006-1008.

76. Karpel , J . P. , A. Kotch, M. Zinny, J. Pesin, and W. Al leyne. 1994.A comparison of inhaled ipratropium, oral theophylline plus inhaledp-agonist , and the combinat ion of a l l three in pat ients wi th COPD.Chesr 105:1089-1094

77. Pet ty, T. L. 1990. The Nat ional Mucolyt ic Study: resul ts of a ran-domized, double-blind, placebo-conrrolled srudy of iodinated glycerolin chronic obstruct ive bronchi t is . Chesr 97:75-83.

78. Anthonisen, N. R. , J . Manfreda, C. P. W. Warren, E. S. Hershf ie ld,C. K. M. Harding, and N. A. Nelson. 1987. Ant ib iot ic therap_"" inexacerbat ions of chronic obstruct ive pulmonary disease. Ann. In-rern. Med. 106:196-204.

79 . Rodn i ck , J . E . , and J . K . Gude . 1988 . Theuseo f an t i b i o t i c s i n acu tebronchi t is and acute exacerbat ions of chronic bronchi t is . l lest . J .Med. 149:34'l -351 .

80. Schl ick, W. 1992. Select ive indicarrons for use of ant ib iot ics: when andwhat. Eur. Respir. Rer,. 2:18'1-192.

8 L Robinson, R. W., and C. Zwi l l ich. 1989. The ef fecrs of drugs on breath-ing dur ing s leep. /n M. H. Kryger, T. Rorh, and \ \ ' . C. Dement,edi tors. Pr incip les and Pract ice of Sleep Medic ine. \ \ ' . B. Saunders,Ph i l ade lph ia . 501 -5 12 .

82 . K rop , A . D . , A . J . B l ock , and E . Cohen . 1973 . Neu ropsych ia t r i c e f -fects of cont inuous oxygen therap) in chronic obsrrucr ive pulmo-nary disease. Chesr 64:311-322.

83. Pet ty, T. L. , and M. M. Fin igan. I 966. Cl in ical evaluar ion of pro)ongedambu la to r y oxygen t he rap ) i n ch ron i c a i r ua l obs r ruc r i on . . 4n t . J .Med. 45:242-252.

84 . Repo r t o f t he Med i ca l Resea rch Counc i l \ \ o r l r ng Pa r r r . l 9S l . Long .term domici l iary oxygen rherap) in chronic h lpor ic cor pulmonalecompl icat ing chronic bronchi t is and emphl 'scma . Loncet I :66 | -6E5 .

85. Nocturnal Oxygen Therapl ,Tr ia l Croup. t980, Conrrnuous or noc-turnal oxygen therapf in hypoxemrc chronic obsrrucrrre lung dis-ease. Ann. Intern. Med. 93:391-196.

86 . We i t zenb lum, E . , A . Sau regeau , \ 1 . Eh rha r : . \ . l . \ t ammosse i . andA. Pel let ier . 1985. Long-rerm ox! 'gen therap) aan re\ers€ rhc progres-s ion of pulmonary hyperrension in par ienrs u r r h chror: rc obsrruct r \ ! .pulmonary disease. ,4nt . Rer. Resptr . Drs l - i l : {9 i - {96-

87 . Lopez -Ma jano , V . , and R . E . Du r ron . l 9 - i . Rcgu ra r ron o f r esp i r a -tory dr ive dur ing oxygen brearhing in chronr. - obslru.- r r re lung drs-ease. Am. Rev. Respir . Drs. 106: l j l - laC

88. Dunn, W. F. , S. B. Nelson, and R. D. Hubnrar: l9)1, Orygen-rnducedhypercarbia in obstruct ive puimonan d;scarc . . t r ' ' Rsr. R6ptr Ds.144:526-530.

89 . Aub ie r , M . , D . Mu rc i ano . \ . I . \ t i i r ; -En r ; . . c . s . t 9 i { , E i f e ; r i o : r hcadministrat ion of ox1'gen rherap! o. \ 3a:r :a: io. anc f . iood gascr i r rpat ients wi th chronic obstruci t \e pu; : : , i r r ; : ! : r j r !a. t ' r our lnS aju lc rc-spiratory f a i lure. Ant . Rer Res. : r - D: : l l i

- : - - - : :

90. Dempsey, J. A., E. H. Vidruk, and S. M. Westenbrook. 1980. pul-monary control sysrems in exercise. Fed. Proc.39:1498-1505.

91 . L iker , E . S . , A . Karn ick , and L . Lerner . 1975. por tab leoxygen inchronic obstrucrive Iung disease with hypoxemia and cor pulmonale.Chest 68:236.

92. Carter, R., and B. Nicotra. 1986. Newer insights into the managementand rehabil i rat ion of the patienr with pulmonary disease. Semin.Respir. Med. 8:2: I l3-123.

93. O'Donohue, W. J. 1991. Effecr of oxygen therapy on increasing ar-terial oxygen tension in hypoxemia patients with stable chronic ob-struct f ve pulmonary disease while breathing ambient air . Chest 100:968-972.

94. Jackson, R. M. 1985. Pulmonary oxygen toxiciry. C/resr 88:900-905.95. Peuy, T. L., R. E. Stanford, and T. A. Neff. 1971. Conrinuous oxy-

gen therapy in chronic airway obsrruction: observations on possibleoxygen toxicity and survival. Ann. Inlern. Med.75:361-36j.

96. West, C. A., and P. Primeau. 1983. Nonmedical hazards of long-rermoxygen therapy. Respiratory Care 28:9M-912.

97. Kory, R. C., J. C. Bergmann, R. D. Sweet, et a!.1962. Comparativeevaluation of oxygen therapy techniques. J.A.M.A. l'79:123-128.

98. Gibson, R. L., P. B. Comer, R. W. Beckman, et al. 1976. Acrual rra-cheal oxygen concentration with commonly used oxygen therapy.Anesthesiology 44:7 l-7 3.

99. Soffer, M., D. P. Tashkin, B. J. Shapiro, M. Lit tner, E. Harvey, andS. Farr. 1985. Conservation of oxygen supply using a reservoir na-sal cannula in hypoxemic patients at rest and during exercise. Ciesl88:663-668.

100. Carter, R., J. Wil l iams, J. Berry, et al.1986. Evaluation of the pen-dant oxygen conserving nasal cannula during exercise. Chest 89:806-8 l 0.

l0l . Bower, J. S., C. J.-Brook, K. Zimmer, and D. Davis. 1988. Perfor-mance of a demand oxygen saver system during rest, exercise, andsleep in hypoxemic patienrs. Chest 94:77-80.

102. Tiep, B. L., K. L. Christopher, B. T. Spofford, J. Goodman, p. D.Worley, and S. L. Macey. 1990. Pulsed nasal and rranstracheal oxy-gen delivery. Chest 97:364-368.

103. Heimlich, H. J. 1982. Respiratory rehabilitarion with transrracheal oxy-gen sysrem. Ann. Otol. Rhinol. Laryngol.9l:643-647.

104. Christopher, K. L., B. T. Spofford, M. D. Perrun, D. C. McCarty,J. R. Coodman, and T. L. Petty. 1987. A program for transtrachealoxygen delivery: assessment of safety and efficacy. Ann. Intern. Med.I 0?:802-808.

105. Hoffman, L. A., S. W. Wesmil ler, F. C. Sciurba, J. T. Johnson, P. F.Ferson, and J. H. Dauber. 1992. Nasal cannula and transtrachealoxygen delivery: comparison of patienl response after six months'use of each technique. Am. Rev. Respir. Dis. 145:827-831.

106. Benditt , J., M. Pol lock, and B. Cell i . 1993. Transtracheal del ivery ofgas decreases the oxygen cost of breathing. Am. Rev. Respir. Dis.141:1207 -1210.

107. Bloom, B. S., J. M. Daniel. M. Wiseman, R. S. Knorr, R. Cebul, andW. L. Kissick. 1989. Transtracheal oxygen delivery and patienrs withchronic obstructive pulmonary disease. Respir. Med. 83:281-288.

108. Campbell , E. J., D. Baker, and P. Crites-Si lver ' . 1988. Subjectiveef-fects of humidificarion of oxygen for deliverv by nasal cannula. Cresl93:289-293.

109. Tiep, B. L. 1991. Medicare regularions for oxygen reimbursement. /nB. L. Tiep, editor. Portable Oxygen Therapy: lncluding Oxygen Con-serving Methodology. Furura Publishing Co., Mt. Kisco, NY.

I 10. American Thoracic Socier1,. 1981. Pulmonary rehabil i tat ion. Ant. Rev.Respir. Dis. t24:663-666.

I l l . Fishman, A. P. 1994. NIH Workshop Summary: Pulmonary Rehabil i -tation Research . Am. J. Respir. Crit. Care Med. 149:825-833.

I 12. Make, B. J. 1986. Pulmonary rehabilitarion: myth or reality. Clin. ChesrMed. '7:519-540.

l l3 . Hodgk in , J . E . , C . L . Connors , and C. W. Be l l , ed i to rs . 1993. Pu l -monar.v Rehabilitation: Guidelines to Success. J. B. Lippincott, Phila-delphia.

l l4 . i r4ccav in , C. R. , S . P . Cupra , E . L . L loyd , e t a l .197 '7 . Phys ica l re -habil i tat ion for the chronic bronchir ic: resuhs of a control led tr ialof exercises in the home. Thorax 32l307-311.

I l -s. Cell i , 8., J. Rassulo, and B. J. Make. 1986. Dyssynchronous brearhingis associated with arm but not leg exercise in chronic obstructive pul-monary disease. N. Engl. J. Med. 314:1485-1490.

I 16. Cuyatt, G. H., M. Towns6nd, L. B. Berman, and S. D- Pugsely. 1987.Quali ty of l i fe in patients with chronic airf lou l imitat ion. Br. J. Dis.Chest 8l:45-54.

l l7 . A tk ins , C. J . , R . M. Kap lan , R. M. T imms, S . Re insch. and K. Lo l -back. 1984. Behavioral exercise programs in the management of

AMERICAN IOURNAL OF RESPIRATORY AND CRITICAT CARE MEDICINE VOL 152 1995

chronic obstructive pulmonary disease. J. Consult. Clin. psychol.52:591 -603.

l l8 . Make, B . , K . G lenn, D. Ik le , B . Bucher , D . Tu teur , U . Mason, e ra l .1992. Pulmonary rehabil i tat ion improves the quali ty of l i fe of pa_tienrs with chronic obstrucrive pulmonary disease (COpD). Am. ie,r,.Respir. Dis. 145:A'767.

I I9. Casaburi, R. 1993. Exercise training in chronic obsrrucrive luns dis_ease. , fn R. Casaburi, and T. L. petty, editors. principles and Frac-r ice of Pulmonary Rehabil i tat ion. W. B. Saunders, phi ladelohia.

120. Lertzman, M. M., and R. M. Cherniack. l9?6. Rehabil iat ion of oa-tienrs with chronic obstructive pulmonary disease.,4rz. Rev. Resnir.D is . l l4 :1145.

l2l. Jensen, P. S. 1983. Risk, protective facrors, and support ive inrerven_tions in chronic airway obsrrucrion. Arch. Gen. psychiatry 40:1203-t207.

122. Hudson, L. D., M. L. Tyler, and T. L. perry. 1976. Hospiral izarionneeds during an outpatient rehabilitation program for severe chronicairway obsrruction. Cftest 70:606.

123. Emery, C. 1993. Psychosocial considerations among pulmonary pa-t ienrs. ̂L J. E. Hodgkin, C. L. Connors, and C. W. Bell , editors.Pulmonary Rehabil i rarion: Cuidel ines ro Success. J. B. Lippincott,Phi ladelphia.

124. Bebour , D . E . , J . E . Hodgk in , E . G. Zorn , A . R. yee, and E. A . Sam-mer. 1983. Cl inical and physiological ourcomes ofa university-hos_pital pulmonary rehabiliration program. Respiratory Care 2g:146g_1473.

125. Pet ty , T . L . , L . M. Ner r , M. M. F in igan, G. A . Br ink , and p . R. Cor -sel lo. 1981. A comprehensive care program for chronic airway ob-struction: methods and preliminary evaluation of symptomatic andfunctional improvement. Ann. Intern. Med. 70:l l09-1120.

126. Bickford, L. S., and J. E. Hodgkin. 1988. National pulmonary re-habilitation survey. Respiratory Care 33:1030.

127. Mazzuca, S. A. t982. Does patient education in chronic disease havetherapeuric value? Journal o/ Chronic Diseases 35:521-529.

128. Ries, A. L. t 990. Posit ion paper of rhe American Associat ion of Car-diovascular and Pulmonary Rehabil i tat ion: scienti f ic basis of oul-monary rehabilitation. Journa! of Cardiopulmonary Rehabilitaiion' 10 :41 8-441.

129. Make, 8., 1994. Collaborative self-management strategies for patientswith respiratory disease. Respiratory Care 39:566-579.

130. Howland, J., E. C. Nelson, P. B. Barlow, G. Mchugo, F. A. Meier,P. Prenr, et ol. 1986. Chronic obstructive airway disease: impact ofhealth educarion. Chest 90:23i-238.

l3l. Toshima, M. T., R. M. Kaplan, and A. L. Ries. 1990. Experimentalevaluation of rehabiliration in chronic obstructive pulmonary dis-ease: short-term effects on exercise endurance and health sratus.Health Psychol. 9:237 -252.

132. Couser, J. I . , F. J. Mart inez, and B. R. Cell i . 1993. pulmonary re_habilitation that includes arm exercise reduces metabolic and ven_tilarory requiremenls for simple arm elevation. Chest 103:37_41 .

I 33. Ries, A. L., B. El l is, and R. W. Hawkins. t9gg. Upper exlremrry exer_cise training in chronic obstructive pulmonary disease. Ciest 93:6gg.

134. Belman, M. J. 1993. Venti latory muscle rraining and unloading../nR. Casaburi and T. L. Petty, editors. principles and practice of pul_monary Rehabil i tat ion. W. B. Saunders, phi ladelphia.

135. Harver, A., D. A. Mahler, and J. A. Daubenspeck. 19g9. Tareetedinspiratory muscle training improves respirarory muscle furfctioi andreduces dyspnea in parients with chronic obstrucrive pulmonary dis-ease. Ann. Intern. Med. l l l : l l i -124.

136. Dekhuizen, P. N. R., H. T. M. Folgering, and C. L. A. van Herwaarden.1991. Targer f low inspirat ion muscle rraining during pulmonary re_habil i tat ion in patients wirh COpD. Chest 99:l2g_133.

137. Selecky, P. A. 1993. Sexuali ty and the parienr wirh lung disease. . InR. Casaburi and T. L. Petry, edirors. principles and pracrice of pul_monary Rehabil i tat ion. W. B. Saunders, phi ladelphia.

138. Wil leput, R., J. P. Vashaudez, D. Landers, A. Nys, T. Knoops, andR. Sergysels. 1983. Thoracoabdominal morion during chesi phys-iotherapy in parrents affected by chronic obsrructive iung diseaie.Respi ro t ion 44:-204-2t 4.

139. Fernandez, E., M. Tanchoco-Tan, and B. J. Make. 1993. Merhodsto lmprove respiratory muscle funclion: seminars in respiratory medi_crne. Pu I m ona ry Re ho bil i tat io n | 4:446-465.

140. Elf iou, M. V., D. A. Mulvey, J. Moxham, et al. lggt. Domici l iarvnocturnal nasal intermittenl posit ive pressure venti lat ion in COpD:mechanisms underlying changes in arterial blood gas tensions. .Ear.Respir. J. 4:tM4-1052.

l4l. Shapiro, S. H., P. Ernsr, K. Cray-Donald, et at.1992-Effecr of neea_

trve pressure ventilation in severe chronic obstructive pulmonary dis-ease. Lancet i40:1425-1429.

142. O'Donohue, W. J., R. M. Giovannoni, A. F. Goldberg, T. G. Keens,B. J. Make, A. L. Plummer , et al.19g6. Long+erm mechanical ven-tilation: guidelines for management in the honre and alternate sires.Clesr 90:lS-375.

143. Make, B. J., and M. E. Gilmart in. 1991. Care of venri laror-assisredindividuals in the home and alrernate sites. 1, G. G. Burron, J. E.Hodgkin, and J. J. Ward, edirors. Respiratory Care: A Guide toClinical Practice. J. B. Lippincorr, phi iadelphia.

144. Haggerry, M. D., R. Stockdale-Woolley, and N.'sreedhar. 1991. Respi_Care: an innovative home care program for the patient with chronicobstructive pulmonary disease. Chest 100:60i,_612.

145. Bergner, M., L. P. Hudson, D. A. Conrad, et ot. l9gg. The cost andefficacy of home care for parienrs with chronic lung disease. Med.Care 26:566-579.

146. Murara, C. H., M. S. Gorby, T. W. Chick, and A. K. Halperin. 1989.Use of emergency medical services by patients with decompensatedobsrrucrive lung disease. Ann. Emerg. Med. lg:501_506.

147. Emerman, C. L., D. Effron, and T. W. Lukens. 1991. Spirometriccriteria for hospital admission of patients with acute exacerbation' of COPD. Ciesr 99:595-599.

148. Emerman, C. L., A. F. Connors, T. W. Lukens, D. Effron, and M. E.May. 1989. Relationship between arterial blood gases and spirome_try ln acute exacerbations of chronic obstructive pulmonary disease.Ann. Emerg. Med. 18l.523_527.

149. Murara, G. H., M. S. Gorby, C. O. Kapsner, T. W. Chick, and A. K.Halperin. 1992. A mulrivariate model for rhe predicrion of relapseafter outpatient treatment of decompensated chionic obstructive pul-monary disease, Arch. Intern. Med. l12:73-j7.

150. Mushlin, A. I . , E. R. Black, C. A. Connolly, K. M. Buonaccorso,and S. W. Eberly. 1991. The necessary length of hospital sny foichronic pulmonary disease. J.A.M.A. 266tE0_S3.

l5l. Ferguson, G. T., and R. M. Cherniack. 1993. Managcment of chronic. _^ _.obstluctive pulmonary disease. N. Engl. J. Med. 32g:1017_1022.152. Skorodin, M. S. 1993. pharmacotherapy-for asthma and chronic ob-

strucrive pulmonary disease. Arch. Intern. Med. 153:g14_929,153. Mesti tz, H.., J. M. Copland, and C. F. McDonald. 19g9. Comparison

of outpatient nebulzed vs. metered dose inharer terbutaline inihronicairflow obstrucrion. Clesl 96:12i7_1240.

154. Salzman, G. A., M. T. Sreele, J. p. pribble, R. M. Elenbaas, and D. R.Pyszczynski. 1989. Aerosolized metaproterenol in the treahent ofasthmatics with severe airflow obstruction: comparison of two deliv_ery methods. Chest 96:1017-1020.

155. Hockley, 8., and N. M. Johnson. 1995. A comparison of three highdoses of ipratropium bromide in chronic asrhma. Br. J. Dis. Chist'19:379-384.

156' Flenley, D. c. 1989. Should bronchodilarors be combined in chronicbronchit is and emphysema,! B.M.J. 295:l160_l l6l .

I 57. Postma, D. S. 1991. Inhaled therapy in COpD: whar are the benefits?Respir. Med. 85:447-449.

158. Karpel, J. P. 1991. Bronchodilator responses to anrichol inergic and. _^ - _.beta-adrenergic agents in acute and stable COpD. Chest gg:gi | _g7 6.159. Nisar, M., J. E. Earis, M. G. pearson, and p. M. A. Calverley. 1992.

Acure bronchodilator trials in chronic obstructive pulmonary dis-ease. Am. Rev. Respir. Drs. I46:555_559.

160. McKay, S . E . , C . A . Howie , A . H. Thomson, B . Whi r ing , and C. J .Addis. I 993. Value of theophylline rreatmenr in patients handicappedby chronic obsrrucrive lung disease. Thorax +g:22j_232.

16l. Thomas, P., J. A. pugsley, and J. H. Srewarr. 1992. Theophylt ineand salbutamol improve pulmonary funcrion in patients with irre-versible chronic obstructive pulmonary disease. Ciresl l0l : 160_165.

162. Hudson, L. D., and C. M. Monti . 199b. Rarionale and use of cor-trcosteroids in chronic obstructive pulmonary disease. Med. Ctin.N. Am.'14:66t-690.

163. Yernauh, J. C. 199.2.Inhaled conicosteroids in COpD: rhe EUROSCOPstudy in perspecrive. Eur. Respir. J. 5: l 169_l l?0.

164. Anthonisen, N. R., J. Manfreda, C. p. W. Warren, E. S. Hershfield,C. K. M. Harding, and N. A. Nelson. 19g7. Antibiot ic therapy inexacerbations of chronic obstructive pulmonary disease. eni.'tn-tern. Med. t06:196-2O4.

165. Hammack, J. E., and C. L. Loprinzi. 1994. Useof oral lyadministeredopioids for cancer related pain. Moyo Ctin. proc.69:3g4_390.

166. Wheeler, A. P. 1993. Sedation, analgesia and paralysis in the inren-srve care unit. Chesl 104:566_5.11.

167. Petty, T. L. 1993. Chronic obstructive pulmonary disease: Ireatmenlof advanced srages of disease. ^tn R. Casaburi and T. L. perry, edi-

American Thoracic Societv

t o r s . P r i nc i p l es and P rac r i ce o f Pu lmona rv Rehab i l i r a t i on . W. B .Saunders. Phi iadeiphia. 392-391 .

168 . Ves tbo , J . , K . M . Knudsen , and F . \ ' . Rasmussen . 1988 . The r . a l ueof mucus hypersecret ion as a predictor of morta l i ty ' and hospi ta l iza-t ion: an I l -year regisrer-based fo l low-up srudy of a random popu-lat ion sample of 8?6 men. Respir . Med. 83:207-21l .

169 . Spe i ze r , F . E . , M . E . Fa r ' , D . W. Docke r \ ' , i nd B . G . Fe r r i s , J r . 1989 .Chronic obsrrucl ive pulmonary disease morta l i ry in s ix U.S. c i r ies.Am. Ret, . Respir . Dis. 140(Suppl . ) :S49-S55.

170 . K i r i l l o f f , L . H . , C . R . Owens , R . M . Roge rs , and M . D . Mazzocco .1985. Does chesr physical rherap): uork? C/resr 88:436-444.

l 7 l . Campbe l l , A . , J . O 'Conne l l , and F . W i l son . 1975 . The e f f ec r o f ches rphysiotherapy upon rhe FEV, in chronic bronchi t is . Med. J. Ausr.l : 3 3 - 3 5 .

172. Rossman, C. M., R. Waldes, D. Sampson, and M. T. Newhouse. 1982.Effect of chest physiotherapy on rhe removal of mucus in par ienrsw i t h cys t i c f i b ros i s . Am. Rev . Resp i r . D i s . 126 : l 3 l - 135 .

173 . S t i l l e r , K . , T . Ceake , J . Tay lo r , R . G ran t , and B . Ha l l . 1990 . Acu relobar ate lectasis: a compar ison of two chest physiotherapl ' regimens.Chest 98:1336-1340

174 , H i l l i ng , L . , E . Bakow , J . F i nk , C . Ke l l y , D . Sobush , and P . A .Southorn. 1993. AARC Cl in ical Pracr ice Guidel ine: use of posi t iveairway pressure adjuncrs to bronchial hygiene rherapy. Respirator;'Core 38:516-521.

175. Chrisrensen, E. F., T. Nedergaard, and R. Dahl. 1990. Long-rerm rrear-ment of chronic bronchitis with positive expirarory pressure and chesrphysiotherapy. Chesr 97 :645-650.

176 . Su t t on , P . P . , H . C . Gemme l l , N . I nnes , J . Dav idson , F . W . Sm i th ,J. S. Legge, and J. A. R. Fr iend. 1988. Use of nebul ized sal ine andnebulized terbutaline as an adjuncr to chesr physiotherapy. ?"lrorax43:51-60.

I 77. Shim, C., M. King, and M. H. Wi l l iams. 1987. Lack of ef fecr of hydra-tion on sputum producrion in chronic bronchitis. Chest 92:679-682.

178. Corke, C. , and P. Cranswick. 1988. A Seldinger technique for mini -tracheostomy insertion. Anesth. Intensive Care 16:206-207.

I 79. Heffner, J . E. 1993. Timing of t racheotomy in mechanical ly venr i laredpalients. Am. Rev. Respir. Dis. l4'7:768-771.

180. Col ice, G. L. , T. A. Stukel , and B. Dain. 1989. Laryngeal compl ica-tions of prolonged intubarion. Chest 96:877-884.

l 8 l . H i l l , N . S . 1993 . Non invas i ve ven r i l a r i on : does i r wo rk . f o r whom.and how? Am. Rey. Respir . Dis. 147:1050-1055.

182 . Be lman , M . J . , G . W. Soo Hoo , J . H . Kue i , and R . Shadoneh r . 1990 .' Ef f icacy of posi t ive vs. negar ive pressure venr i lar ion in unloading

the respiratory muscles. Ciesr 98:850-856.183. Pepe, P. E. , and J. J. Mar in i . 1982. Occulr posir ive end-expira lor) '

pressure in mechanical l l 'vent i lated par ienrs wi th a i r f lou obstruc-t ion: the auto-PEEP ef fecr . Am. Rey. Respir . Dis. 126 166-110.

184 . B rown , D . G . , and D . J . P i e r son . 1986 . Au ro -PEEP i s common i nmechanical ly vent i lated pat ients: a srud! of Incidence. sever i t l , , anddetection. Respirator.r, Care 3l :1069-1074.

185 . Benson , M . S . , and D . J . P i e r son . 1988 . Au ro -PEEP du r i n_e mechan i -cal vent i lat ion of adulrs. Respiratory Care 33:55:-56E.

186. Maclntyre, N. R. 1986. Respirator l ' funcr ion dur ing pressure suppor lventilation, Chesr 89:67'7 -683.

187 . Esen , F . , T . Denke l , L . Te l c i , J . Kesec iogu l . A . S . Tu runcu , K . . { kp i r ,and B. Lachmann. 1992. Compar ison of pressu:e suppon venl i la-t i on (PSV) and i n re rm i r r en r manda ro r \ r en r i l a r i on t l \ t \ ' ) du r i ngweaning in pat ients s i th acure respiraror l fa i lure. .4dr. Erp. . \ led-Biol. 3l'7:371-316.

188 . 1989 . Consensus con fe rence on a r r i f i c i a l a i r ua l s t n pa r l en l s r e ;e i r t ngmechanical venr i lar ion. Chesl 96: l -6- 160.

189 . B rocha rd , L . , D . I sabe l , J . P i que r . P . Amaro . J \ t anc .bc . A . \ t es -sadi , C. Brun-Buisson, A. Rauss. F- Lemarc. and A, Harf , 1990.Reversal of acute exacerbat ions oi chronrc obstrucl r \e iung drseaseby inspiratory assisrance ui rh a iace mas! \ Ene. J t led. 32-1:I 523 - I 530 .

190. Bot t , 1. , M. P. Carrol l , J . H, Conual . e: a. l9)-r Randomrsed.^on-tro l led t r ia l of nasal venr i lar ion in a;u l : r :nt : ;a io:r far i . . : rc due rochronic obstructive airwa! s disease iance.' ,ij i : I -< j J- I -<J-

l 9 l . Chev ro l e t , J . C . , P . Jo l l i e r , B . Aba_ rc . A Tou -<s : . anC \1 . Lou rs . t 991 .Nasal posi t ive pressure rent i lat ior : ln pai len: ! Dl :h a.-ure reni i laror)fa l lure. Chest 100:775-761.

192. Slutsky, A. S. 1993. lVlechantca, !er . t : :a: i !a: { ! : l . : i . -a. Coi icgc o: ChesrPhysic ians' Consensus Con:e:er : : : Cces. tC,c : ! , : , : - 1859.

193 . M i t hoe fe r , J . C . , J . F . Xe igh ie r . a t : \ i S } . a : t : l r r l 9 -1 . Responseof the ar ter ia l PO, to or lge:- . a:=. : : - i : :a: .a: ; : . . - : :onr. - puimonaryd i sease : i n t e rp re ta t i on o i : i : ' - : : : g t : . 3 ! : J J r J ; . t 6 l a l t en l s and l 4

normal subjects. Ann. Intern. Med. 74:32g-335.I 94. Carter, R. I 991. Oxygen and acid-base status: measuremenr, inrerpre-

tarion, and rarionale for oxygen therapy. /n B. L. Tiep, editor. porr-able Oxygen Therapy, lncluding Oxygen Conserving Methodology.Furura Pub l ish ing Co. , Mt . K isco , Ny . 129-159.

195. Kroenke, K . , V . A . Lawrence, J . F . Theroux , M. R. Tu ley , and S.Hilsenbeck. 1993. Posroperative complicarions after rhoracic andmajor abdominal surgery in patients with and without obstructivelung disease. Chest 104:1445-145t.

196. Merl i , C. J., and H. H. Weitz. 1993. Appioaching the surgical pa-t ient: role of rhe medical consulranr. Ctin. Chesr Med.l4:205-210.

l9'1 . Zibrak, J. D., and C. R. O'Donnell . 1993. Indications for preoperativepulmonary funcrion resring. Cii, Chesr Med. l4,22j-236.

198. Ford , C. T . , T . W. Rosena l , F . C le rgue, and W. A. Whi te law. 1993.Respiratory physiology in upper abdominal surgery. Clin. Chest Med.t4:237 -252.

199. Cell i , B. R. 1993. Perioperative respirarory care of the parienr under-going upper abdominal surgery. Ctin. Chest Med. 14.253.261.

200. Sykes, L. A., and E. A. Bowe. 1993. Cardiorespirarory effects ofanesthesia. Clin. Chest Med. 14:2ll-226.

201. Cracey, D. R., M. B. Diverr ie, and E. p. Didier. 1979. preonerativepulmonary preparation of patients with chronic obstrucrivepulmo-nary disease. Chest 76:123-129.

202. Cell i , B. R., K. S. Rodriguez, and C. L. Snider. 1994. A control ledtrial of intermittent positive pressure breathing, incentive spirome-try, and deep breathing exercises in prevenring pulmonary compli-cations afrer abdominal surgery. Am. Rev. Respir. Dis. 130:12-15.

203. Hall, J. C., R. Tarala, J. Harris, et al. t 991. lncenrive splromerry versusroutine chesr physiotherapy for prevention of pulmonary complica-tions after abdominal surgery. Lancet 33j:953-956.

204. Warner, M. A., K. P. Offord, M. E. Warner, et at. 19g9. Role ofpreoperative cessation of smoking and other factors in postopera-tive pulmonary complications: a blinded prospective study of coro-nary arrery bypass patients. Mayo Clin. proc. 64:609-616.

205. MacFadyen, B. V., and J. L. Ponsky, editors. 1992. Laparoscopy forthe general surgeon. Surg. Clin. N. Am. j2:997-1185.

206. Ha l l , J . C . , R , A . Tara la , J . L . Ha l l , and J . Mander . 1991. A mul -tivariare analysis of the risk of pulmonary complicarions afrer lapa-roromy. Chest 99:923-927.

207. Gozum, M., B. J. Turner, and R. Tipperman. 1992. perioperativemanagemenr of rhe ophthalmology parienr. /n G. J. Merli and H. H.Weitz, edirors. Medical Management of the Surgical patient. W. B.Saunders, Philadelphia.

208. Coll ins, R., A. Scrimgeour, S. yusuf, and R. peto. l9gg. Reductionin fatal pulmonary embolism and venous thrombosis by preopera-tive adminisrration of sirbcuraneous heparin: overview of results ofrandomized trials in general, orthopedic, and urologic surgery. N.Engl. J. Med. 3t8:il62-ll73.

209. I,lauhal', M. A., and J. p. Weiner-Kronish. 19g9. Respirarory manage-menr afrer cardiac surgery. Chest 95:424-435.

210. Gerac i , J . M. , A . K . Rosen, A . S . Ash, K . J . McNi f f , and M. A .Moskon,itz. 1993. Predicting rhe occurrence of adverse events aftercoronar) arrery bypass grafiing. Ann. Intern. Med. llg:lg_24.

3l l . Jayr, C., M. A. Mauhay, J. Goldstoni, W. M. Cold, and J. p. Wiener_Kronish. 1993. Preoperative and intraoperative factors associatedwith prolonged mechanical venrilarion following major vascular sur-gery'. C hest 103: I23 l-1236.

lll. \lelrin, \\'. S., M. J. Krasna, and J. L. Mclaughlin. lg2. Thoracoscopicmanagemenl of sponraneous pneumorhorax. Chest 102:lg7 5_1g76.

l l3, \ larshal l , M. C., and G. N. Olsen. 1993. The ohvsioloeic evaluarionof rhe lung resecrion candidate. Clin. Chesi Med. li:305-320.

l l { . Ferguson, M. K . , L . L i r t le , L . R izzo , K . J . popov ich , C. F . C lonek ,A. Leff, D. Manjoney, and A. C. Liule. 1988. Diffusing capacirypredicrs morbidity and morral i ty after pulmonary resection. J.Thorac. Cardiovasc. Surg. 96:894-900.

l l -< . Kearne l ' , D . i . , T . H . Lee, J . J . Re i l l y , M. M. DeCamp, and D. J .Sugarbaker. 1994. Assessmenr of operative r isk in patienrs under-going lung resection: importance of predicred pulmonary function.Chest t05:.753-i59.

l l6 . i l l a rkos , J . , B . P . Mu l lan , D. R. H i l lman, A . W. Musk , V . F . Anr ico ,F. T. Lovegrove, M. J. Carler, and K. E. Finucane. 1989. preoper-atlve assessment as a predictor of mortality and morbidity after lungresecrron. Am. Rev. Respir. Dis. 139:902_910.

l l7. Olsen, C. N. 1989. The evolving role of exercise tesl ing prior to lungresecrion. Chest 95 :,218-225.

218. Boysen, P. S. 1993. Perioperative managemenl of the thoracotomyparienr. Clin. Chest Med. t4:321-333.

l

AMERICAN IOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE VOL 152 I995

219' whipp' B' J ' ' and S' A' ward' 1992' Physiologic changes fol lowing sumption of rhe respirarory muscres in normal and malnourishedibilateral carotid body resection in patients with chronic obstructivi parienrs with chroni. ibrr-.tiu. puimonary disea se. Am. Re,-. Respir.pulmonary disease. Chest 10l:656-661. Drs. I40:3g5_39i-tt' -';[;]:1'":'3;2,\tlrl;ln'_",t'it""' resuhs of surgerv ror bullous em- 244. Rogers, n. r',r., v.-nonahue, and J. cosranrino. r9e2. physiorogic

22r. cooper, r. D., E. p. rrurock, A. N. rrianrafiilou, c. A. patterson, :li'"'i'*"jb:i:i:.:i:ffrffli;.d1ffi:.Ti:?:i:T#,1gf f:lM. S. Poh l ' P . A . Deroney , R. S . Sundaresan, and c . L . Roper . 15 i l -1517.l995 Bilateral pneumectomy(volumereduction)forchronicobsriuc- 245. Dark,,D. s., s. K. pingleton, and c. R. Kerby. tgg5. Hypercapniat ivepulmonarydisease' J'Thorac,cardiovasc. s&rg.109:106-119. duringweaning:aconipl icat ionofnutr i t ionalsupport. chestgg:l4l-222. Gelb, A. F., N. Zamel, A. Colchen, et a!.1992. phyi iologic srudies 143.

fr j : i t l : : i lo8Tlfx.t t ." , t inairwavobstruction. Am. Riv. Respir. 246. ralpers, s. s., D. J. Romberger, s. B. Bunce, and S. K. pingleton.

223. tss3.American rhoracic Sociery sraremenr: rung rranspra mation. Am. ;;.1,Tii:i::ff1,'.in:iil'f;ffi:1i::fTig,i*;x:I:gi:lii.Rev. Respir. Dis. 147:772-776. Cnesr 102:551_555.224' Lou', D' E', E' P Trulock, L. R. Kaiser, er al 1992. Morbidity, mor- 247. Askanazi, : . , s. i . nor.nbaum, A. L. Hyman, et at. lgg0.Respira-tal i ty ' and early results of single virsus bi lateral lung transi lanta- tory changes induced by the large glucose loads of toral parenteralt ion foremphysema. J. Thorac. cardiovasc. surg. 103:l l lg-1126. nutr i t ion. J.A.M.A.243:1444-1447.225 Klink, M', and S Quan l987. Prevalence of reported sleep disrur- 248. Aubier, r t , , l . , p. v"i . i"no, y. Lecogui c, et al. l9g5. Effect of hypo-bances in a general populat ion and their relat ionship ro obitruct ive phosphatemia on diaphragmatic contracti l i ty in patients with acureairways diseases. Chesr 9l:540-546.226. Steens, R., Z. pouriot, r. Miilar, M. Kryger, and c. ceorge. ree3. ,r. 'r[';ff:T] f]lil,?;i

"l',f,,i;,{, t':6"1,t"';:?. l2.ar.n,r.r,

""a n.Effects of zolpidem and lr iazolam on sleep and respirat ioi in mild RurtenSerg. t976. Air l ine travel for chi ldren with chronic pulmo-ro moderate chronic obstructive pulmonary disease. Sleep l6:3lg_326. nary disease. pediatrics 57:40E_410.227' Phi l l ips, B., K. cooper, and r. Burke. 198?. The effeci of sleep loss 250. Di l lard, T. A., B. w. 8..g, K. R. Rajagopal, J. w. Dooley, and w. J.on breathing in chronic obstructive pulmonary disease' chest 9l: Mehm. 198i. Hvpo".r i" during air travel in parients with chronic29-32..228. Frercher, E., r. Miller, c. Divine, r. Flercher, and r. Miiler. 1e87. tr,. o,",,?"tJll]x:"fffi1,iffi1';*T.

"i:'#;{i{;r'.iii?;ifii

Nocturnal oxyhemoglobin denaturation in coPD patients with ar- patients with chronic obstructive pulmonary disease. Arch. Intern.terial oxygen rensions above 60 mm Hg. Chesr 92:604_609. Med. lll:1793- 1795.229 Levi-Valensi ' P' ' E' weitzenblum' Z' Rida, et at.1992. Sleep-related 252. cotrrel l , .1. L isgd. nt i i ,ua. exposure during aircraft f ly ing higher.denarurarion and dayrime pulmonary haemodynamics in COpD pa- Chest 92:gl_g4.t ients' Ear' Respir ' J ' 5:301-307' 253. Schwartz, J. s.,H. Z. Bencowitz, and K. M. Moser. 19g4. Air travel230. Fletcher, E., R. Luckett, T. Mil ler, and J. Fletcher. 19g9. Exercise hypoxemia with chronic obstructive pulmonary disease. Ann. In-hemodynamics and gas exchange in parients with chronic obstruc_ rern. Med. 100:473_477.tive pulmonary disease, sleep denaturation, and a daytime Pa6, above 254. cong, H., p. p. iasrrr<in, and u. S. Simmons. lgg4. Hypoxia-altitude60 mm Hg. Am. Rev. Respir. Dis. 140:123't-1245. sjmulat ion tesr. Am. Rev. Respir. Drs. 130:9g0_986.23 l Doug las ,N ' ,andD.F len ley . 1990 'Breath ingdur ings leep inpar ienrs 255. AMA comr i r r ion-on i 'n . rg .n .y Med ica l Serv ices . 19g2. Med ica lwithobstructivelungdisease. Am.Rev.Respir.Dis. l4l :1055-1070. aspecrsofrransponationaboardcommercialaircraft.J.A.M.A,24T:

232. connaughron, J., K. carreralr, R. Elron, J. Stradring, and N. Dougias. 1007-r0l l .1 9 8 8 ' D o s l e e p s t u d i e s c o n t r i b u t e t o t h e m a n a g e m e n t o f p a t i e n r s w i r h 2 5 6 . B e r g , B . w . , T . A . D i l l a r c t , K . R . R a j a g o p a l , a n d w . J . M e h m . l 9 9 2 .severe chronic obstructive pulmonary disease? Am' Rev. Respir. Dis. oiygen rupptara-nt-"tion during air travel in patients with chronicI 38 :34 I -344.

2-13. Frercher, E., C. Donner, B. Midgren, et at. tee2.survival in CopD ,5? p."r?'tilH:'i: i"]T;;ri'{,"0,"t'ffir?f?irlro,;fjj;1it;",.s of survivalpatients with a daytime Pa6, ) 60 mm Hg with and wirhour nocrur- for acure ,.rpir. iory fai lure in lhroni i obstructivl pulmonary dis-nal oxyhemoglobin desaruratjon. Chest l0l :649-655. ease. Chest St,SIS_5ZL234 Braun' s' R" N L Keim, R' M Dixon' et al 1984. The prevalence 258. Kaelin, R. M., t . Ars;macopoulos, and J. c. chevrolet. 19g7. Fai l-and determinants of nutr i t ional changes in chronic obstrucrive pul- ure to predicr six-month survival of patients with copD requir ingmonary disease' chest 86:558-563' mechanical ventilation by analysis of simple indices. cftest 92:971-235. Hunter , A . B . M. , M. A . Carey , and H. \ [ ' . Larsh . l9g l . The nur r i - 97g.

lT : ' ; : i lT : r j : i ' i l f ] ' ; f i [ :n icobs t rucr ivepu lmonarvd isease. 25e. Dardes ,N. ,s .campo,M.c .ch iapp in i ,e ra t . te86.prognos iso f

236. Drive, A. c., r\4. r. McArevy, and r. L. Smith. re82. Nurritionaias- ,lT&iiTll';jl8:#,:L?ff#ffi:. respiratorv railure. ^Enr.

sessmenlo fpa t ien tswi thchron icobs t ruc t ivepu lmonaryd iseaseand 260. Por r ie r , i . . , c .n . r " , i i r rJJ ,andJ . -F .Mui r . lgg2 .Determinanrso f im-acure respirarory fai lure. Chest 82:56g-5l '1.23?. schols, A. M., p.'8. Soerers, A. M. Dingemans, r. Mosrert, er a/. i'jll,t'r'il"jff1:l1.ll:T,'..i:#."J,:'$f:",:ffHiJfl*

1993' Prevalence and characterist ics of nutr i t ional deplet ion in pa- prospecrive mulr icenter srudy. cresl l0l :204-210.r ienrs* ' i rhs rab lecopDel ig ib re fo rp i r lmonar l rehab i l i ra t ion . ,4 rn . 261. Man in , T . R. , s . w . L -ewis , and R. K . Arber t . r9g2. Theprognos isRev ' Resp i r ' D is ' 141: l l5 l -1156 ' r raar "ar^ .^ r^ -^^ r

o fpa t ien tswi thchron icobs t ruc t ivepu lmonaryd iseaseaf te rhosp i -2 3 8 . K e l s e n , S . C . , M . F e r e n c e , a n d S . K a p o o r . l g 8 5 . E f f e c t s o f p r o l o n g e d _ - r a l i z a r i o n f o r a c u l e r e s p i r a t o r y f a i l u r e . C h e s t g 2 : 3 1 0 _ 3 1 4 .

HHi t l ; l i ! ;a ls t ruc lu reandfunc t ionof thed iaphragm.J .App l . 262. Hudson, I - .o . rsag.su t l i va l dara inpat ien tswi thacureandchron ic

239. Braun, N. M. T., N. S. nroru, and D. F. Rochesrer. r983. Respira_ lxt:$;: lTj:ruir ing mechanicar venri lat ion. Am. Rev. Respir. Dis.

r o r ! ' m u s c l e a n d p u l m o n a r l ' f u n c t i o n i n p r o x i m a l m y o p a t h i e s . r h o - 2 6 3 . D a v i s , H . , s . a . L e f r a k , D . M i l l e r , e r a t . l g g 0 . p r o l o n g e d m e c h a n i -rax38:616-623 N M T Rrqr rn roe( n . r . . - ; - - ^ , "^ r *^ . ,

.a l l1 'ass is ted vent i la r ion : an ana lys is o f ou tcome and charges .240. Rochester , D . R. , and N. M. T . Braun. 19g5. Derermjnants o f mar i - J .A .M.A. 243:43_45.

;f l 'T: i : i#i iT,1:T j ; , :1'_"#c obstructive pulmonarv disease 264. American Thoracic Society. leel. withholding and wirhdrawing l i fe-24r. wirson, D. o.. f - M. Rogers, and R. M. Horrman. re85. Nurririon ,65 p;:i::J,lit ili,ili;1"1?;,iii,fifil;r"Ji;Jf J,'j;lli; ," ;..,and chron ic lung d isease ' Am' Ret ' ' Resp i r ' D is ' 132:1347-1365. c ineandBiomedica landBehav io ra l Research . 19g2. Mak ingHea l th241. Wi fson.D.O. ,R.M.Rogers ,M.H.Sander , e ra l . , tgg6. Nur r i r iona l CareDec is ions : i t .E i i , " " f andLega l lmp l ica t ionsof ln fo rmedCon_

Hii l ;" ; , i ' , l r f f i i : : ;Tltd patients with emphvsema' Am Rer" sent in rhe parienr-pracri t ioner Relarionship, Vol. l . U.S. covern-

243. Donairue, M., R. M. Rog.rr, D. o. wi lson, et ar. r9g9.ox;-gen con- ment Print ing off ice' washington, D.c.

American Thoracic Societv

GLOSSARY

AAT abbreviarion for alpha,-anrirr).psin.ABG abbreviarion for arrerial blood gases.acidemia increased hydrogen ion concenrrarion (decreased pH)of the b lood.acinus in the lung, a pr imar l , respi rator l ' lobule, consis t ing ofa bronchiole and its branches; also used s-vnonymously wirh a/-veolus.ACV abbreviation for assist-conrrol venrilarion.alkalemia decreased hydrogen ion concenrrarion (increased pH)of the b lood.alpha,-antitrypsin a serum glycoprorein coded for by a singlegene on chromosome 14, produced by the l iver and normalll,found in the lungs; it is the major inhibitor of proreolyric en-zymes, and its deficiency is associated with the development ofemphysema.anticholinergic a parasympathetic nerve blocker.apical bullae vesicles found.ar rhe upper border of the em-physematous lung.apnea absence or cessation of respiration; "sleep apnea" denotestransient attacks occurring during sleep.asterixis a motor disturbance, known also as "flapping tremor"or "l iver f lap," characterized by involuntary jerking movementson assuming a particular f ixed posture.atelectasis pulmonary collapse, nonexpansion, or incomplete ex-panslon.barotrauma pressure injury, generally involving the eustachiantube, the eardrum, or the paranasal sinuses.beta'-agonists a subclass of agents (beta-blockers or beta-adren-ergic blocking agents) that inhibit responses ro sympatheric ad-renergic nerve activity and to epinephrine and other adrenergicamines.bronchiectasis chronic dilation of the bronchi.cenlriacinar emphysema emphysema thar begins in the respira-tory bronchioles and spreads peripherally.centri lobular emphysema the form of centriacinar emphl'semaassociated with longstanding cigarene smoking, predominanrlyinvolv ing the upper hal f o f the lungs.CO2 narcosis CNS depression, typicall l ' marked bl srupor, as-sociated with carbon dioxide retenrion due to depression of h1'p-oxic drive.cor pulmonale heart disease stemming from pulmonarv hl per-tension due to lung disease; it is characrerized b1' riehr venrricu-lar hypertrophy and, acutely or ulrimatell ' , r ighr hearr failure.cotinine a metabolic nicorine derivarive which, derecred in bodlfluids such as urine, provides evidence of exposure ro robaccosmoke, whether primary or secondarl '.distal acinar emphysema emphysema preferenriall l inrolr ing dis-tal airway structures, alveolar ducts and sacs: also kno*n asparaseptal emphysema.dyspnea diff icult or labored brearhing.emphysema a chronic condition characrerized bv abnormalll en-larged air spaces distal to the rerminal bronchioles. '* rrh dilarionof the a lveol i and destruct ion of the a lveolar *a l ls .FEV, abbreviation for forced expirator] r 'olume rn one second:the maximum that can be expi red, srarr in-e f rom marrmum in-spi rat ion.f ibros is any abnormal formar ion of f ibrous r issue: in lhe pul -

monary context, a chronic, progressive condition involving thealveolar walls, associated with chronic inflammation, and caus-ing steadily increasing respiratory dysfunction.focal emphysema the form of centriacinar emphysema that oc-curs in coal workers' pneumoconiosis.FRC abbreviation for functional residual capacity, the gas vol-ume remaining in the lungs at the end of a normal expiration.FVC abbreviation for forced vital capacitl', FEV measured with-out regard to time and with the subject exhaling as rapidly aspossib le.goblet cell a type of epithelial cell, functioning as a mucousgland, found in various mucous membranes, including those ofthe respiratory passages.hemoptysis rhe expectoration of blood.hilar of or pertaining to the hilus (or hilum) of the lung, a de-pressed area on the mediastinal surface marking the entry pointof the bronchus, nerves, and blood vessels.hypercapnia an abnormal increase in blood carbon dioxide.hypercarbia hypercapnia.hyperoxia a condition of excessive oxygen in tissues and or-gans, as from exposure to greater than atmospheric pressures.hypoventilation reduced alveolar venti lation, especially relativeto carbon d iox ide producr ion.hypoxemia abnormally low oxygenation of the blood.IMV abbreviation for intermittent mandatory ventilation.mucostasis a halt in the secretion of mucus.oximetry measurement of the oxygen saturation of arterial blood.panacinar emphysema emphysema involving the entire alveo-lus uniformly and predominating in the lower half of the lungs,the type generally seen with homozygous AAT deficiency; focalpanacinar emphysema at the lung bases may accompany cen-tri lobular emphysema in smokers.PAP abbreviation for positive airway pressure or for pulmonaryarterJ" pressure.paraseptal emphysema distal acinar emphysema (see above).PEEP abbreviarion for positive end-expiratory pressure, a me-chanical venrilarion technique in which pressure is maintainedin order to increase the volume remaining in the lungs at the endof expiration.pleurodesis the surgical creation of adhesions between thevisceral and parietal layers of the pleura, it may be used in treat-ment of recurrent pneumothorax.pneumothorax air or gas in the pleural cavity.poll 'c1'themia an abnormal increase in the number of red bloodcel ls.prolease an1' proteolytic enzyme.PSl' abbreviation for pressure support venti lation.PVR abbreviarion for pulmonary vascular resistance.respirato4 acidosis a state of abnormally high carbon diox-ide retention, also called hypercapnic acidosis or carbon diox-ide ac idosis , due ro inadequate pulmonary vent i la t ion.sarcoidosis a lung disease of unknown etiology marked by thepresence of granulomarous lesions on the walls of the alveoliand.zor bronchioles, with loss of lung volume and elasticity.spiromelrl the measurement of lung capacity.surfactant a locall l 'produced fluid essential to expansion andcontracrion of the alveoli; ir consisrs of phospholipids, chieflylec i th in and sphingomyel in .tidal volume rhe quantity of air inspired and expired in onerespiratorl, cycle during regular breathing.