Clinical practice parameters for hemodynamic …...Pediatric Special Article Clinical practice...

of 23 /23
Pediatric Special Article Clinical practice parameters for hemodynamic support of pediatric and neonatal septic shock: 2007 update from the American College of Critical Care Medicine* Joe Brierley, MD; Joseph A. Carcillo, MD; Karen Choong, MD; Tim Cornell, MD; Allan DeCaen, MD; Andreas Deymann, MD; Allan Doctor, MD; Alan Davis, MD; John Duff, MD; Marc-Andre Dugas, MD; Alan Duncan, MD; Barry Evans, MD; Jonathan Feldman, MD; Kathryn Felmet, MD; Gene Fisher, MD; Lorry Frankel, MD; Howard Jeffries, MD; Bruce Greenwald, MD; Juan Gutierrez, MD; Mark Hall, MD; Yong Y. Han, MD; James Hanson, MD; Jan Hazelzet, MD; Lynn Hernan, MD; Jane Kiff, MD; Niranjan Kissoon, MD; Alexander Kon, MD; Jose Irazusta, MD; John Lin, MD; Angie Lorts, MD; Michelle Mariscalco, MD; Renuka Mehta, MD; Simon Nadel, MD; Trung Nguyen, MD; Carol Nicholson, MD; Mark Peters, MD; Regina Okhuysen-Cawley, MD; Tom Poulton, MD; Monica Relves, MD; Agustin Rodriguez, MD; Ranna Rozenfeld, MD; Eduardo Schnitzler, MD; Tom Shanley, MD; Sara Skache, MD; Peter Skippen, MD; Adalberto Torres, MD; Bettina von Dessauer, MD; Jacki Weingarten, MD; Timothy Yeh, MD; Arno Zaritsky, MD; Bonnie Stojadinovic, MD; Jerry Zimmerman, MD; Aaron Zuckerberg, MD *See also p. 785. The American College of Critical Care Medicine (ACCM), which honors individuals for their achieve- ments and contributions to multidisciplinary critical care medicine, is the consultative body of the Society of Critical Care Medicine (SCCM) that possesses rec- ognized expertise in the practice of critical care. The College has developed administrative guidelines and clinical practice parameters for the critical care prac- titioner. New guidelines and practice parameters are continually developed, and current ones are system- atically reviewed and revised. Dr. Brierley received meeting travel expenses from USCOM Ltd. Dr. Nadel has consulted, received hono- raria, and study funding from Eli Lilly. Dr. Shanley has received a research grant from the National Institutes of Health. The remaining authors have not disclosed any potential conflicts of interest. For information regarding this article, E-mail: [email protected] Copyright © 2009 by the Society of Critical Care Medicine and Lippincott Williams & Wilkins DOI: 10.1097/CCM.0b013e31819323c6 Background: The Institute of Medicine calls for the use of clinical guidelines and practice parameters to promote “best practices” and to improve patient outcomes. Objective: 2007 update of the 2002 American College of Critical Care Medicine Clinical Guidelines for Hemodynamic Support of Neonates and Children with Septic Shock. Participants: Society of Critical Care Medicine members with special interest in neonatal and pediatric septic shock were identified from general solicitation at the Society of Critical Care Medicine Educational and Scientific Symposia (2001–2006). Methods: The Pubmed/MEDLINE literature database (1966 – 2006) was searched using the keywords and phrases: sepsis, septicemia, septic shock, endotoxemia, persistent pulmonary hy- pertension, nitric oxide, extracorporeal membrane oxygenation (ECMO), and American College of Critical Care Medicine guide- lines. Best practice centers that reported best outcomes were identified and their practices examined as models of care. Using a modified Delphi method, 30 experts graded new literature. Over 30 additional experts then reviewed the updated recommenda- tions. The document was subsequently modified until there was greater than 90% expert consensus. Results: The 2002 guidelines were widely disseminated, trans- lated into Spanish and Portuguese, and incorporated into Society of Critical Care Medicine and AHA sanctioned recommendations. Cen- ters that implemented the 2002 guidelines reported best practice outcomes (hospital mortality 1%–3% in previously healthy, and 7%– 10% in chronically ill children). Early use of 2002 guidelines was associated with improved outcome in the community hospital emer- gency department (number needed to treat 3.3) and tertiary pediatric intensive care setting (number needed to treat 3.6); every hour that went by without guideline adherence was associated with a 1.4-fold increased mortality risk. The updated 2007 guidelines continue to recognize an increased likelihood that children with septic shock, compared with adults, require 1) proportionally larger quantities of fluid, 2) inotrope and vasodilator therapies, 3) hydro- cortisone for absolute adrenal insufficiency, and 4) ECMO for refrac- tory shock. The major new recommendation in the 2007 update is earlier use of inotrope support through peripheral access until central access is attained. Conclusion: The 2007 update continues to emphasize early use of age-specific therapies to attain time-sensitive goals, specifically recommending 1) first hour fluid resuscitation and inotrope therapy directed to goals of threshold heart rates, normal blood pressure, and capillary refill <2 secs, and 2) subsequent intensive care unit he- modynamic support directed to goals of central venous oxygen saturation >70% and cardiac index 3.3– 6.0 L/min/m 2 . (Crit Care Med 2009; 37:666 – 688) KEY WORDS: guidelines; sepsis; severe sepsis 666 Crit Care Med 2009 Vol. 37, No. 2

Embed Size (px)

Transcript of Clinical practice parameters for hemodynamic …...Pediatric Special Article Clinical practice...

  • Pediatric Special Article

    Clinical practice parameters for hemodynamic support of pediatricand neonatal septic shock: 2007 update from the American Collegeof Critical Care Medicine*

    Joe Brierley, MD; Joseph A. Carcillo, MD; Karen Choong, MD; Tim Cornell, MD; Allan DeCaen, MD;Andreas Deymann, MD; Allan Doctor, MD; Alan Davis, MD; John Duff, MD; Marc-Andre Dugas, MD;Alan Duncan, MD; Barry Evans, MD; Jonathan Feldman, MD; Kathryn Felmet, MD; Gene Fisher, MD;Lorry Frankel, MD; Howard Jeffries, MD; Bruce Greenwald, MD; Juan Gutierrez, MD;Mark Hall, MD; Yong Y. Han, MD; James Hanson, MD; Jan Hazelzet, MD; Lynn Hernan, MD; Jane Kiff, MD;Niranjan Kissoon, MD; Alexander Kon, MD; Jose Irazusta, MD; John Lin, MD; Angie Lorts, MD;Michelle Mariscalco, MD; Renuka Mehta, MD; Simon Nadel, MD; Trung Nguyen, MD; Carol Nicholson, MD;Mark Peters, MD; Regina Okhuysen-Cawley, MD; Tom Poulton, MD; Monica Relves, MD; Agustin Rodriguez, MD;Ranna Rozenfeld, MD; Eduardo Schnitzler, MD; Tom Shanley, MD; Sara Skache, MD; Peter Skippen, MD;Adalberto Torres, MD; Bettina von Dessauer, MD; Jacki Weingarten, MD; Timothy Yeh, MD; Arno Zaritsky, MD;Bonnie Stojadinovic, MD; Jerry Zimmerman, MD; Aaron Zuckerberg, MD

    *See also p. 785.The American College of Critical Care Medicine

    (ACCM), which honors individuals for their achieve-ments and contributions to multidisciplinary criticalcare medicine, is the consultative body of the Societyof Critical Care Medicine (SCCM) that possesses rec-ognized expertise in the practice of critical care. TheCollege has developed administrative guidelines and

    clinical practice parameters for the critical care prac-titioner. New guidelines and practice parameters arecontinually developed, and current ones are system-atically reviewed and revised.

    Dr. Brierley received meeting travel expenses fromUSCOM Ltd. Dr. Nadel has consulted, received hono-raria, and study funding from Eli Lilly. Dr. Shanley hasreceived a research grant from the National Institutes

    of Health. The remaining authors have not disclosedany potential conflicts of interest.

    For information regarding this article, E-mail:[email protected]

    Copyright © 2009 by the Society of Critical CareMedicine and Lippincott Williams & Wilkins

    DOI: 10.1097/CCM.0b013e31819323c6

    Background: The Institute of Medicine calls for the use ofclinical guidelines and practice parameters to promote “bestpractices” and to improve patient outcomes.

    Objective: 2007 update of the 2002 American College of CriticalCare Medicine Clinical Guidelines for Hemodynamic Support ofNeonates and Children with Septic Shock.

    Participants: Society of Critical Care Medicine members withspecial interest in neonatal and pediatric septic shock wereidentified from general solicitation at the Society of Critical CareMedicine Educational and Scientific Symposia (2001–2006).

    Methods: The Pubmed/MEDLINE literature database (1966–2006) was searched using the keywords and phrases: sepsis,septicemia, septic shock, endotoxemia, persistent pulmonary hy-pertension, nitric oxide, extracorporeal membrane oxygenation(ECMO), and American College of Critical Care Medicine guide-lines. Best practice centers that reported best outcomes wereidentified and their practices examined as models of care. Usinga modified Delphi method, 30 experts graded new literature. Over30 additional experts then reviewed the updated recommenda-tions. The document was subsequently modified until there wasgreater than 90% expert consensus.

    Results: The 2002 guidelines were widely disseminated, trans-lated into Spanish and Portuguese, and incorporated into Society ofCritical Care Medicine and AHA sanctioned recommendations. Cen-

    ters that implemented the 2002 guidelines reported best practiceoutcomes (hospital mortality 1%–3% in previously healthy, and 7%–10% in chronically ill children). Early use of 2002 guidelines wasassociated with improved outcome in the community hospital emer-gency department (number needed to treat � 3.3) and tertiarypediatric intensive care setting (number needed to treat � 3.6); everyhour that went by without guideline adherence was associated witha 1.4-fold increased mortality risk. The updated 2007 guidelinescontinue to recognize an increased likelihood that children withseptic shock, compared with adults, require 1) proportionally largerquantities of fluid, 2) inotrope and vasodilator therapies, 3) hydro-cortisone for absolute adrenal insufficiency, and 4) ECMO for refrac-tory shock. The major new recommendation in the 2007 update isearlier use of inotrope support through peripheral access until centralaccess is attained.

    Conclusion: The 2007 update continues to emphasize early use ofage-specific therapies to attain time-sensitive goals, specificallyrecommending 1) first hour fluid resuscitation and inotrope therapydirected to goals of threshold heart rates, normal blood pressure, andcapillary refill 70% and cardiac index 3.3–6.0 L/min/m2. (Crit Care Med2009; 37:666–688)

    KEY WORDS: guidelines; sepsis; severe sepsis

    666 Crit Care Med 2009 Vol. 37, No. 2

    bhattrClinical practice parameters for hemodynamic support of pediatric and neonatal septic shock: 2007 update from the American College of Critical Care Medicine: Erratum

    In the article on page 666 in the February 2009 issue, there were errors in the spelling of some author names. The correct names are: Saraswati Kache, not Sara Skache; and Jose Irazuzta, not Jose Irazusta.

    Also, on page 675 of this article, second full paragraph, the seventh sentence should read:Low-dose arginine vasopressin (in doses ≤0.04 units/min) as an adjunctive agent has short-term hemodynamic benefits in adults with vasodilatory shock.

    This erratum is published in the April 2009 issue of the journal.

  • Neonatal and pediatric severesepsis outcomes were alreadyimproving before 2002 withthe advent of neonatal andpediatric intensive care (reduction inmortality from 97% to 9%) (1–4), andwere markedly better than in adults (9%compared with 28% mortality) (3). In2002, the American College of CriticalCare Medicine (ACCM) Clinical PracticeParameters for Hemodynamic Support ofPediatric and Neonatal Shock (5) werepublished, in part, to replicate the re-ported outcomes associated with imple-mentation of “best clinical practices”(mortality rates of 0%–5% in previouslyhealthy [6–8] and 10% in chronically illchildren with septic shock [8]). There aretwo purposes served by this 2007 updateof these 2002 clinical practice parame-ters. First, this 2007 document examinesand grades new studies performed to testthe utility and efficacy of the 2002 rec-ommendations. Second, this 2007 docu-ment examines and grades relevant newtreatment and outcome studies to deter-mine to what degree, if any, the 2002guidelines should be modified.


    More than 30 clinical investigators and cli-nicians affiliated with the Society of CriticalCare Medicine who had special interest in he-modynamic support of pediatric patients withsepsis volunteered to be members of the “up-date” task force. Subcommittees were formedto review and grade the literature using theevidence-based scoring system of the ACCM.The literature was accrued, in part, by search-ing Pubmed/MEDLINE using the followingkeywords and phrases: sepsis, septicemia, sep-tic shock, endotoxemia, persistent pulmonaryhypertension (PPHN), nitric oxide (NO), andextracorporeal membrane oxygenation(ECMO). The search was narrowed to identifystudies specifically relevant to children. Bestpractice outcomes were identified and de-scribed; clinical practice in these centers wasused as a model.

    The clinical parameters and guidelineswere drafted and subsequently revised using amodification of the Delphi method. Briefly,the initial step included review of the litera-ture and grading of the evidence by topic-based subcommittees during a 6-month pe-riod. Subcommittees were formed accordingto participant interest in each subtopic. Theupdate recommendations from each subcom-mittee were incorporated into the preexisting2002 document by the task force chairperson.All members commented on the unified up-date draft, and modifications were made in an

    iterative fashion until �10% of the task forcedisagreed with any specific or general recom-mendation. This process occurred during a1-year period. Reviewers from the ACCM thenrequested further modifications that wereconsidered for inclusion if supported by evi-dence and best practice. Grading of the liter-ature and levels of recommendations werebased on published ACCM criteria (Table 1).


    Successful Dissemination,Acceptance, Implementation,and Outcome of 2002Guidelines

    The 2002 guidelines were initially dis-tributed in the English language with of-ficial sanctioning by the Society for Crit-ical Care Medicine with publication inCritical Care Medicine. The main pediat-ric algorithm was included in the Pediat-ric Advanced Life Support (PALS) manualpublished by the American Heart Associ-ation. In addition, the pediatric and new-born treatment algorithms were pub-lished in whole or part in multipletextbooks. The guidelines were subse-quently published in Spanish and Portu-guese allowing for dissemination inmuch of the American continents. Therehave been 57 peer-reviewed publicationssince 2002 that have cited these guide-lines. Taken together these findings dem-onstrate academic acceptance and dis-semination of the 2002 guidelines (Tables2 and 3).

    Many studies have tested the observa-tions and recommendations of the 2002guidelines. These studies reported evi-dence that the guidelines were useful andeffective without any evidence of harm.For example, Wills et al (9) demonstratednear 100% survival when fluid resuscita-tion was provided to children with den-

    gue shock. Maitland et al (10) demon-strated a reduction in mortality frommalaria shock from 18% to 4% whenalbumin was used for fluid resuscitationrather than crystalloid. Han et al reportedan association between early use of prac-tice consistent with the 2002 guidelinesin the community hospital and improvedoutcomes in newborns and children(mortality rate 8% vs. 38%; numberneeded to treat [NNT] � 3.3). Every hourthat went by without restoration of nor-mal blood pressure for age and capillaryrefill �3 secs was associated with a two-fold increase in adjusted mortality oddsratio (11). Ninis et al (12) similarly re-ported an association between delay ininotrope resuscitation and a 22.6-foldincreased adjusted mortality odds ratio inmeningococcal septic shock. In a ran-domized controlled study, Oliveira et al(13) reported that use of the 2002 guide-lines with continuous central venous ox-ygen saturation (ScvO2) monitoring, andtherapy directed to maintenance of ScvO2�70%, reduced mortality from 39% to12% (NNT � 3.6). In a before and afterstudy, Lin et al (14) reported that imple-mentation of the 2002 guidelines in aU.S. tertiary center achieved best practiceoutcome with a fluid refractory shock 28-day mortality of 3% and hospital mortal-ity of 6% (3% in previously healthy chil-dren; 9% in chronically ill children). Thisoutcome matched the best practice out-comes targeted by the 2002 guidelines(6–8). Similar to the experience of St.Mary’s Hospital before 2002 (7), SophiaChildren’s Hospital in Rotterdam also re-cently reported a reduction in mortalityrate from purpura and severe sepsis from20% to 1% after implementation of 2002guideline-based therapy in the referralcenter, transport system, and tertiarycare settings (15). Both of these centers

    Table 1. American College of Critical Care Medicine guidelines for evidence-based medicine ratingsystem for strength of recommendation and quality of evidence supporting the references

    Rating system for referencesa Randomized, prospective controlled trialb Nonrandomized, concurrent or historical cohort investigationsc Peer-reviewed, state of the art articles, review articles, editorials, or substantial

    case seriesd Nonpeer reviewed published opinions, such as textbook statements or official

    organizational publicationsRating system for recommendations

    Level 1 Convincingly justifiable on scientific evidence aloneLevel 2 Reasonably justifiable by scientific evidence and strongly supported by expert critical

    care opinionLevel 3 Adequate scientific evidence is lacking but widely supported by available data and

    expert opinion

    667Crit Care Med 2009 Vol. 37, No. 2

  • also used high flux continuous renal re-placement therapy (CRRT) and fresh frozenplasma infusion directed to the goal of nor-mal international normalized ratio (INR)(prothrombin time). In a U.S. child healthoutcomes database (Kids’ Inpatient Data-base or KID) analysis, hospital mortalityfrom severe sepsis was recently estimatedto be 4.2% (2.3% in previously healthy chil-dren, and 7.8% in children with chronicillness) (16), a decrease compared with 9%in 1999 (4). Taken together, these studiesindirectly and directly support the utilityand efficacy of implementation of the time-sensitive, goal-directed recommendationsof the 2002 guidelines in the emergency/delivery room and pediatric intensive careunit/neonatal intensive care unit set-tings.

    New Major Recommendationsin the 2007 Update

    Because of the success of the 2002guidelines, the 2007 update compilation

    and discussion of the new literature weredirected to the question of what changes,if any, should be implemented in the up-date. The members of the committeewere asked whether there are clinicalpractices which the best outcome prac-tices are using in 2007 that were notrecommended in the 2002 guidelines andshould be recommended in the 2007guidelines? The changes recommendedwere few. Most importantly, there was nochange in emphasis between the 2002guidelines and the 2007 update. The con-tinued emphasis is directed to: 1) firsthour fluid resuscitation and inotropedrug therapy directed to goals of thresh-old heart rates (HR), normal blood pres-sure, and capillary refill �2 secs, and 2)subsequent intensive care unit hemody-namic support directed to goals of ScvO2�70% and cardiac index 3.3–6.0 L/min/m2. New recommendations in the 2007update include the following: 1) The 2002guidelines recommended not giving car-diovascular agents until central vascularaccess was attained. This was becausethere was and still is concern that admin-istration of peripheral vasoactive agents(especially vasopressors) could result inperipheral vascular/tissue injury. How-ever, after implementation of the 2002guidelines, the literature showed that, de-pending on availability of skilled person-nel, it could take two or more hours toestablish central access. Because mortal-ity went up with delay in time to inotropedrug use, the 2007 update now recom-mends use of peripheral inotropes (notvasopressors) until central access is at-tained. The committee continues to rec-ommend close monitoring of the periph-eral access site to prevent peripheralvascular/tissue injury; 2) The 2002 guide-lines made no recommendations on whatsedative/analgesic agent(s) to use to facil-

    itate placement of central lines and/orintubation. Multiple editorials and cohortstudies have since reported that the useof etomidate was associated with in-creased severity of illness adjusted mor-tality in adults and children with septicshock. The 2007 update now states thatetomidate is not recommended for chil-dren with septic shock unless it is used ina randomized controlled trial format. Fornow, the majority of the committee usesatropine and ketamine for invasive pro-cedures in children with septic shock.Little experience is available with ket-amine use in newborn septic shock andthe committee makes no recommenda-tion in this population; 3) Since 2002,cardiac output (CO) can be measured notonly with a pulmonary artery catheter,but also with Doppler echocardiography,or a pulse index contour cardiac outputcatheter catheter, or a femoral arterythermodilution catheter. Titration of ther-apy to CO 3.3–6.0 L/min/m2 remains thegoal in patients with persistent catechol-amine resistant shock in 2007 guidelines.Doppler echocardiography can also beused to direct therapies to a goal of su-perior vena cava (SVC) flow �40 mL/min/kg in very low birth weight (VLBW)infants; 4) There are several new poten-tial rescue therapies reported since the2002 guidelines. In children, enoximoneand levosimendan have been highlightedin case series and case reports. Unlikevasopressin, which had been suggested bysome as a vasoplegia rescue therapy,these agents are suggested by some asrecalcitrant cardiogenic shock rescueagents. In newborns, inhaled prostacyclinand intravenous (IV) adenosine were re-portedly successful in refractory PPHN.The 2007 update recommends furtherevaluation of these new agents in appro-priate patient settings; and 5) The 2002guidelines made no recommendation onfluid removal. Although fluid resuscita-tion remains the hallmark and first stepof septic shock resuscitation, two cohortstudies showed the importance of fluidremoval in fluid overloaded septic shock/multiple organ failure patients. The 2007update recommends that fluid removalusing diuretics, peritoneal dialysis, orCRRT is indicated in patients who havebeen adequately fluid resuscitated butcannot maintain subsequent even-fluidbalance through native urine output.This can be done when such patients de-velop new onset hepatomegaly, rales, or10% body weight fluid overload.

    Table 2. American College of Critical Care Medicine hemodynamic definitions of shock

    Cold or warm shock Decreased perfusion manifested by altered decreased mental status,capillary refill �2 secs (cold shock) or flash capillary refill(warm shock), diminished (cold shock) or bounding (warmshock) peripheral pulses, mottled cool extremities (cold shock),or decreased urine output �1 mL/kg/h


    Shock persists despite �60 mL/kg fluid resuscitation (whenappropriate) and dopamine infusion to 10 �g/kg/min


    Shock persists despite use of the direct-acting catecholamines;epinephrine or norepinephrine

    Refractory shock Shock persists despite goal-directed use of inotropic agents,vasopressors, vasodilators, and maintenance of metabolic(glucose and calcium) and hormonal (thyroid, hydrocortisone,insulin) homeostasis

    Table 3. Threshold heart rates and perfusion pres-sure mean arterial pressure-central venous pres-sure or mean arterial pressure-intra-abdominalpressure for age

    Threshold RatesHeart

    Rate (bpm)

    Mean ArterialPressure-CentralVenous Pressure

    (mm Hg)

    Term newborn 120–180 55Up to 1 yr 120–180 60Up to 2 yrs 120–160 65Up to 7 yrs 100–140 65Up to 15 yrs 90–140 65

    bpm, beats per minute.Modified from The Harriet Lane Handbook.

    Thirteenth Edition and National Heart, Lung,and Blood Institute, Bethesda. MD: Report of thesecond task force on blood pressure control inchildren - 1987 (306, 307).

    668 Crit Care Med 2009 Vol. 37, No. 2

  • Literature and Best PracticeReview

    Developmental Differences in the He-modynamic Response to Sepsis in New-borns, Children, and Adults. The predom-inant cause of mortality in adult septicshock is vasomotor paralysis (17). Adultshave myocardial dysfunction manifestedas a decreased ejection fraction; however,CO is usually maintained or increased bytwo mechanisms: tachycardia and re-duced systemic vascular resistance (SVR).Adults who do not develop this process tomaintain CO have a poor prognosis (18,19). Pediatric septic shock is associatedwith severe hypovolemia, and childrenfrequently respond well to aggressive vol-ume resuscitation; however, the hemody-namic response of fluid resuscitated vaso-active-dependent children seems diversecompared with adults. Contrary to theadult experience, low CO, not low SVR, isassociated with mortality in pediatric sep-tic shock (20 –29). Attainment of thetherapeutic goal of CI 3.3–6.0 L/min/m2

    may result in improved survival (21, 29).Also contrary to adults, a reduction inoxygen delivery rather than a defect inoxygen extraction, can be the major de-terminant of oxygen consumption inchildren (22). Attainment of the thera-peutic goal of oxygen consumption (VO2)�200 mL/min/m2 may also be associatedwith improved outcome (21).

    It was not until 1998 that investiga-tors reported patient outcome when ag-gressive volume resuscitation (60 mL/kgfluid in the first hour) and goal-directedtherapies (goal, CI 3.3–6.0 L/min/m2 andnormal pulmonary capillary wedge pres-sure) (21) were applied to children withseptic shock (29). Ceneviva et al (29) re-ported 50 children with fluid-refractory(�60 mL/kg in the first hour), dopamine-resistant shock. The majority (58%)showed a low CO/high SVR state, and22% had low CO and low vascular resis-tance. Hemodynamic states frequentlyprogressed and changed during the first48 hrs. Persistent shock occurred in 33%of the patients. There was a significantdecrease in cardiac function over time,requiring addition of inotropes and vaso-dilators. Although decreasing cardiacfunction accounted for the majority ofpatients with persistent shock, someshowed a complete change from a lowoutput state to a high output/low SVRstate (30 –33). Inotropes, vasopressors,and vasodilators were directed to main-tain normal CI and SVR in the patients.

    Mortality from fluid-refractory, dopamine-resistant septic shock in this study (18%)was markedly reduced compared withmortality in the 1985 study (58%) (29), inwhich aggressive fluid resuscitation wasnot used. Since 2002, investigators haveused Doppler ultrasound to measure CO(34), and similarly reported that previ-ously healthy children with meningococ-cemia often had a low CO with a highmortality rate, whereas CO was high andmortality rate was low in septic shockrelated to catheter-associated bloodstream infections.

    Neonatal septic shock can be compli-cated by the physiologic transition fromfetal to neonatal circulation. In utero,85% of fetal circulation bypasses thelungs through the ductus arteriosus andforamen ovale. This flow pattern is main-tained by suprasystemic pulmonary vas-cular resistance in the prenatal period. Atbirth, inhalation of oxygen triggers a cas-cade of biochemical events that ulti-mately result in reduction in pulmonaryvascular resistance and artery pressureand transition from fetal to neonatal cir-culation with blood flow now being di-rected through the pulmonary circula-tion. Closure of the ductus arteriosus andforamen ovale complete this transition.Pulmonary vascular resistance and arterypressures can remain elevated and theductus arteriosus can remain open forthe first 6 wks of life, whereas the fora-men ovale may remain probe patent foryears. Sepsis-induced acidosis and hyp-oxia can increase pulmonary vascular re-sistance and thus arterial pressure andmaintain patency of the ductus arterio-sus, resulting in PPHN of the newbornand persistent fetal circulation. Neonatalseptic shock with PPHN can be associatedwith increased right ventricle work. De-spite in utero conditioning, the thickenedright ventricle may fail in the presence ofsystemic pulmonary artery pressures. De-compensated right ventricular failure canbe clinically manifested by tricuspid re-gurgitation and hepatomegaly. Newbornanimal models of group B streptococcaland endotoxin shock have also docu-mented reduced CO, and increased pul-monary, mesenteric, and SVR (35–39).Therapies directed at reversal of rightventricle failure, through reduction ofpulmonary artery pressures, are com-monly needed in neonates with fluid-refractory shock and PPHN.

    The hemodynamic response in prema-ture, VLBW infants with septic shock(�32 wks gestation, �1000 g) is least

    understood. Most hemodynamic informa-tion is derived only from echocardio-graphic evaluation and there are few sep-tic shock studies in this population.Neonatology investigators often fold sep-tic shock patients into “respiratory dis-tress syndrome” and “shock” studiesrather than conduct septic shock studiesalone. Hence, the available clinical evidenceon the hemodynamic response in prema-ture infants for the most part is in babieswith respiratory distress syndrome orshock of undescribed etiology. In the first24 hrs after birth during the “transitionalphase,” the neonatal heart must rapidly ad-just to a high vascular resistance state com-pared with the low resistance placenta. COand blood pressure may decrease when vas-cular resistance is increased (40). However,the literature indicates that premature in-fants with shock can respond to volumeand inotropic therapies with improvedstroke volume (SV), contractility, and bloodpressure (41–54).

    Several other developmental consider-ations influence shock therapy in the pre-mature infant. Relative initial deficien-cies in the thyroid and parathyroidhormone axes have been reported andcan result in the need for thyroid hor-mone and/or calcium replacement.(55,56) Hydrocortisone has been examined inthis population as well. Since 2002, ran-domized controlled trials showed thatprophylactic use of hydrocortisone on day1 of life reduced the incidence of hypo-tension in this population, (57) and a7-day course of hydrocortisone reducedthe need for inotropes in VLBW infantswith septic shock (58 – 60). Immaturemechanisms of thermogenesis require at-tention to external warming. Reducedglycogen stores and muscle mass for glu-coneogenesis require attention to main-tenance of serum glucose concentration.Standard practices in resuscitation ofpreterm infants in septic shock use amore graded approach to volume resus-citation and vasopressor therapy com-pared with resuscitation of term neonatesand children. This more cautious ap-proach is a response to anecdotal reportsthat preterm infants at risk for intraven-tricular hemorrhage (�30 wks gestation)can develop hemorrhage after rapid shiftsin blood pressure; however, some nowquestion whether long-term neurologicoutcomes are related to periventricularleukomalacia (a result of prolonged un-der perfusion) more than to intraventric-ular hemorrhage. Another complicatingfactor in VLBW infants is the persistence

    669Crit Care Med 2009 Vol. 37, No. 2

  • of the patent ductus arteriosus. This canoccur because immature muscle is lessable to constrict. The majority of infantswith this condition are treated medicallywith indomethacin, or in some circum-stances with surgical ligation. Rapid ad-ministration of fluid may further increaseleft to right shunting through the ductuswith ensuant pulmonary edema.

    One single-center randomized controltrial reported improved outcome with useof daily 6-hr pentoxyfilline infusions invery premature infants with sepsis (61,62). This compound is both a vasodilatorand an anti-inflammatory agent. A Co-chrane analysis agrees that this promis-ing therapy deserves evaluation in a mul-ticentered trial setting (63).

    What Clinical Signs andHemodynamic Variables Can beUsed to Direct Treatment ofNewborn and Pediatric Shock?

    Shock can be defined by clinical vari-ables, hemodynamic variables, oxygenutilization variables, and/or cellular vari-ables; however, after review of the litera-ture, the committee continues to chooseto define septic shock by clinical, hemo-dynamic, and oxygen utilization variablesonly. This decision may change at thenext update. For example, studies dem-onstrate that blood troponin concentra-tions correlate well with poor cardiacfunction and response to inotropic sup-port in children with septic shock (64–66). Lactate is recommended in adultseptic shock laboratory testing bundlesfor both diagnosis and subsequent mon-itoring of therapeutic responses. How-ever, most adult literature continues todefine shock by clinical hypotension, andrecommends using lactate concentrationto identify shock in normotensive adults.For now the overall committee recom-mends early recognition of pediatric sep-tic shock using clinical examination, notbiochemical tests. Two members dissentfrom this recommendation and suggestuse of lactate as well.

    Ideally, shock should be clinically di-agnosed before hypotension occurs byclinical signs, which include hypother-mia or hyperthermia, altered mental sta-tus, and peripheral vasodilation (warmshock) or vasoconstriction with capillaryrefill �2 secs (cold shock). Threshold HRassociated with increased mortality incritically ill (not necessarily septic) in-fants are a HR �90 beats per minute(bpm) or �160 bpm, and in children are

    a HR �70 bpm or �150 bpm (67). Emer-gency department therapies should be di-rected toward restoring normal mentalstatus, threshold HRs, peripheral perfu-sion (capillary refill �3 secs), palpabledistal pulses, and normal blood pressurefor age (Table 3) (11). Orr et al reportedthat specific hemodynamic abnormalitiesin the emergency department were asso-ciated with progressive mortality (in pa-renthesis); eucardia (1%) � tachycardia/bradycardia (3%) � hypotension withcapillary refill �3 secs (5%) � normo-tension with capillary refill greater than 3secs (7%) � hypotension with capillaryrefill greater than 3 secs (33%). Reversalof these hemodynamic abnormalities us-ing ACCM/PALS recommended therapywas associated with a 40% reduction inmortality odds ratio regardless of thestage of hemodynamic abnormality at thetime of presentation (68). One member ofthe committee wishes to emphasize thatthese signs are important only if the pa-tients are considered ill.

    In both neonates and children, shockshould be further evaluated and resusci-tation treatment guided by hemodynamicvariables including perfusion pressure(mean arterial pressure [MAP] minuscentral venous pressure) and CO. As pre-viously noted, blood flow (Q) varies di-rectly with perfusion pressure (dP) andinversely with resistance (R). This ismathematically represented by Q � dP/R.For the systemic circulation this is rep-resented by CO � (MAP � central venouspressure)/SVR. This relationship is im-portant for organ perfusion. In the kid-ney, for example, renal blood flow �(mean renal arterial pressure � meanrenal venous pressure)/renal vascular re-sistance. The kidney and brain have vaso-motor autoregulation, which maintainsblood flow in low blood pressure (MAP orrenal arterial pressure) states. At somecritical point, perfusion pressure is re-duced below the ability of the organ tomaintain blood flow.

    One goal of shock treatment is tomaintain perfusion pressure above thecritical point below which blood flow can-not be effectively maintained in individ-ual organs. The kidney receives the sec-ond highest blood flow relative to itsmass of any organ in the body, and mea-surement of urine output (with the ex-ception of patients with hyperosmolarstates such as hyperglycemia which leadsto osmotic diuresis) and creatinine clear-ance can be used as an indicator of ade-quate blood flow and perfusion pressure.

    Maintenance of MAP with norepinephrinehas been shown to improve urine outputand creatinine clearance in hyperdy-namic sepsis (69). Producing a supranor-mal MAP above this point is likely not ofbenefit (70).

    Reduction in perfusion pressure belowthe critical point necessary for adequatesplanchnic organ perfusion can also oc-cur in disease states with increased intra-abdominal pressure (IAP) such as bowelwall edema, ascites, or abdominal com-partment syndrome. If this increased IAPis not compensated for by an increase incontractility that improves MAP despitethe increase in vascular resistance, thensplanchnic perfusion pressure is de-creased. Therapeutic reduction of IAP(measured by intrabladder pressure) us-ing diuretics and/or peritoneal drainagefor IAP �12 mm Hg, and surgical decom-pression for �30 mm Hg, results in res-toration of perfusion pressure and hasbeen shown to improve renal function inchildren with burn shock (71).

    Normative blood pressure values inthe VLBW newborn have been reassessed.A MAP �30 mm Hg is associated withpoor neurologic outcome and survival,and is considered the absolute minimumtolerable blood pressure in the extremelypremature infant (42). Because bloodpressure does not necessarily reflect CO,it is recommended that normal COand/or SVC flow, measured by Dopplerechocardiography, be a primary goal aswell (72–82).

    Although perfusion pressure is used asa surrogate marker of adequate flow, theprevious equation shows that organ bloodflow (Q) correlates directly with perfusionpressure and inversely with vascular re-sistance. If the ventricle is healthy, anelevation of SVR results in hypertensionwith maintenance of CO. Conversely, ifventricular function is reduced, the pres-ence of normal blood pressure with highvascular resistance means that CO is re-duced. If the elevation in vascular resis-tance is marked, the reduction in bloodflow results in shock.

    A CI between 3.3 and 6.0 L/min/m2 isassociated with best outcomes in septicshock patients (21) compared with pa-tients without septic shock for whom a CIabove 2.0 L/min/m2 is sufficient (83). At-tainment of this CO goal is often depen-dent on attaining threshold HRs. How-ever, if the HR is too high, then there isnot enough time to fill the coronary ar-teries during diastole, and contractilityand CO will decrease. Coronary perfusion

    670 Crit Care Med 2009 Vol. 37, No. 2

  • may be further reduced when an unfavor-able transmural coronary artery fillingpressure is caused by low diastolic bloodpressure (DBP) and/or high end-diastolicventricular pressure. In this scenario, ef-forts should be made to improve coronaryperfusion pressure and reverse the tachy-cardia by giving volume if the SV is low,or an inotrope if contractility is low. Be-cause CO � HR � SV, therapies directedto increasing SV will often reflexively re-duce HR and improve CO. This will beevident in improvement of the shock in-dex (HR/systolic blood pressure), as wellas CO. Children have limited HR reservecompared with adults because they arealready starting with high basal HRs. Forexample, if SV is reduced due to endotox-in-induced cardiac dysfunction, an adultcan compensate for the fall in SV by in-creasing HR two-fold from 70 to 140bpm, but a baby cannot increase her HRfrom 140 bpm to 280 bpm. Althoughtachycardia is an important method formaintaining CO in infants and children,the younger the patient, the more likelythis response will be inadequate and theCO will fall. In this setting, the responseto falling SV and contractility is to vaso-constrict to maintain blood pressure. In-creased vascular resistance is clinicallyidentified by absent or weak distal pulses,cool extremities, prolonged capillary re-fill, and narrow pulse pressure with rela-tively increased DBP. The effective ap-proach for these children is vasodilatortherapy with additional volume loadingas vascular capacity is expanded. Vasodi-lator therapy reduces afterload and in-creases vascular capacitance. This shiftsthe venous compliance curve so thatmore volume can exist in the right andleft ventricles at a lower pressure. In thissetting, giving volume to restore fillingpressure results in a net increase in end-diastolic volume (i.e., preload) and ahigher CO at the same or lower fillingpressures. Effective use of this approachresults in a decreased HR and improvedperfusion.

    At the other end of the spectrum, athreshold minimum HR is also neededbecause if the HR is too low then CO willbe too low (CO � HR � SV). This can beattained by using an inotrope that is alsoa chronotrope. In addition to thresholdHRs, attention must also be paid to DBP.If the DBP–central venous pressure is toolow then addition of an inotrope/vaso-pressor agent such as norepinephrinemay be required to improve diastolic cor-onary blood flow. Conversely, if wall

    stress is too high due to an increasedend-diastolic ventricular pressure and di-astolic volume secondary to fluid over-load, then a diuretic may be required toimprove SV by moving leftward on theoverfilled Starling function curve. The ef-fectiveness of these maneuvers will simi-larly be evidenced by improvement in theHR/systolic blood pressure shock index,CO, and SVR along with improved distalpulses, skin temperature, and capillaryrefill.

    Shock should also be assessed andtreated according to oxygen utilizationmeasures. Measurement of CO and O2consumption were proposed as being ofbenefit in patients with persistent shockbecause a CI between 3.3 and 6.0L/min/m2 and O2 consumption �200mL/min/m2 are associated with improvedsurvival (21). Low CO is associated withincreased mortality in pediatric septicshock (20–29). In one study, childrenwith fluid-refractory, dopamine-resistantshock were treated with goal-directedtherapy (CI �3.3 and �6 L/min/m2) andfound to have improved outcomes com-pared with historical reports (29). Be-cause low CO is associated with increasedO2 extraction, (22) ScvO2 saturation canbe used as an indirect indicator ofwhether CO is adequate to meet tissuemetabolic demand. If tissue oxygen deliv-ery is adequate, then assuming a normalarterial oxygen saturation of 100%,mixed venous saturation is �70%. As-suming a hemoglobin concentration of10 g/dL and 100% arterial O2 saturationthen a CI �3.3 L/min/m2 with a normaloxygen consumption of 150 mL/min/m2

    (O2 consumption � CI � [arterial O2content � venous O2 content]) results ina mixed venous saturation of �70% be-cause 150 mL/min/m2 � 3.3 L/min/m2 �[1.39 � 10 g/dL � PaO2 � 0.003] � 10 �[1 �0.7]. In an emergency departmentstudy in adults with septic shock, main-tenance of SVC O2 saturation �70% byuse of blood transfusion to a hemoglobinof 10 g/dL and inotropic support to in-crease CO, resulted in a 40% reductionin mortality compared with a group inwhom MAP and central venous pressurewere maintained at usual target valueswithout attention to SVC O2 saturation(84). Since 2002, Oliveira et al (13) repro-duced this finding in children with septicshock reducing mortality from 39% to 12%when directing therapy to the goal of ScvO2saturation �70% (NNT 3.6).

    In VLBW infants, SVC blood flow mea-surement was reportedly useful in assess-

    ing the effectiveness of shock therapies.The SVC flow approximates blood flowfrom the brain. A value �40 mL/kg/minis associated with improved neurologicoutcomes and survival (78–82). ScvO2saturation can be used in low birthweight infants but may be misleading inthe presence of left to right shuntingthrough the patent ductus arteriosus.

    Intravascular Access. Vascular accessfor fluid resuscitation and inotrope/vasopressor infusion is more difficult toattain in newborns and children com-pared with adults. To facilitate a rapidapproach to vascular access in criticallyill infants and children, the AmericanHeart Association and the AmericanAcademy of Pediatrics developed neonatalresuscitation program and PALS guide-lines for emergency establishment of in-travascular support (85, 86). Essentialage-specific differences include use ofumbilical artery and umbilical venous ac-cess in newborns, and rapid use of in-traosseous access in children. Ultrasoundguidance may have a role in the place-ment of central lines in children.

    Fluid Therapy. Several fluid resuscita-tion trials have been performed since2002. For example, several randomizedtrials showed that when children withmostly stage III (narrow pulse pressure/tachycardia) and some stage IV (hypoten-sion) World Health Organization classifi-cation dengue shock received fluidresuscitation in the emergency depart-ment, there was near 100% survival re-gardless of the fluid composition used (6,9, 87, 88). In a randomized controlledtrial, Maitland et al (10) demonstrated areduction in malaria septic shock mortal-ity from 18% to 4% when albumin wasused compared with crystalloid. The largerandomized adult SAFE trial that com-pared crystalloid vs. albumin fluid resus-citation reported a trend toward im-proved outcome (p � 0.1) in septic shockpatients who received albumin (89). Pref-erence for the exclusive use of colloidresuscitation was made based on a clini-cal practice position article from a groupwho reported outstanding clinical resultsin resuscitation of meningococcal septicshock (5% mortality) both using 4% al-bumin exclusively (20 mL/kg bolusesover 5–10 mins) and intubating and ven-tilating all children who required greaterthan 40 mL/kg (7). In an Indian trial offluid resuscitation of pediatric septicshock, there was no difference in out-come with gelatin compared with crystal-loid (90). In the initial clinical case series

    671Crit Care Med 2009 Vol. 37, No. 2

  • that popularized the use of aggressivevolume resuscitation for reversal of pedi-atric septic shock, a combination of crys-talloid and colloid therapies was used(91). Several new investigations exam-ined both the feasibility of the 2002guideline recommendation of rapid fluidresuscitation as well as the need for fluidremoval in patients with subsequent oli-guria after fluid resuscitation. The 2002guideline recommended rapid 20 mL/kgfluid boluses over 5 mins followed byassessment for improved perfusion orfluid overload as evidenced by new onsetrales, increased work of breathing, andhypoxemia from pulmonary edema, hep-atomegaly, or a diminishing MAP�centralvenous pressure. Emergency medicineinvestigators reported that 20 mL/kg ofcrystalloid or colloid can be pushed over5 mins, or administered via a pressurebag over 5 mins through a peripheraland/or central IV line (92). Ranjit et al(93) reported improved outcome fromdengue and bacterial septic shock whenthey implemented a protocol of aggres-sive fluid resuscitation followed by fluidremoval using diuretics and/or peritonealdialysis if oliguria ensued. In this regard,Foland et al (94) similarly reported thatpatients with multiple organ failure whoreceived CRRT when they were �10%fluid overloaded had better outcomesthan those who were �10% fluid over-loaded. Similarly, two best outcome prac-tices reported routine use of CRRT toprevent fluid overload while correctingprolonged INR with plasma infusion inpatients with purpura and septic shock(7, 15).

    The use of blood as a volume expanderwas examined in two small pediatric ob-servational studies, but no recommenda-tions were given by the investigators (95,96). In the previously mentioned study byOliveira et al (13) reporting improvedoutcome with use of the 2002 ACCMguidelines and continuous ScvO2 satura-tion monitoring, the treatment group re-ceived more blood transfusions directedto improvement of ScvO2 saturation to�70% (40% vs. 7%). This finding agreeswith the results of Rivers (84) who trans-fused patients with a SVC oxygen satura-tion �70% to assure a hemoglobin of 10g/dL as part of goal-directed therapybased on central venous oxygen satura-tion. Although the members of the taskforce use conservative goals for bloodtransfusion in routine critical illness, theobservations that for patients with septicshock, transfusion to a goal hemoglobin

    �10 g/dL to achieve ScvO2 �70% is as-sociated with increased survival suggeststhat this higher hemoglobin goal is war-ranted in this population.

    Fluid infusion is best initiated withboluses of 20 mL/kg, titrated to assuringan adequate blood pressure and clinicalmonitors of CO including HR, quality ofperipheral pulses, capillary refill, level ofconsciousness, peripheral skin tempera-ture, and urine output. Initial volumeresuscitation commonly requires 40�60mL/kg but can be as much as 200 mL/kg(28, 91, 97–104). Patients who do notrespond rapidly to initial fluid boluses, orthose with insufficient physiologic re-serve, should be considered for invasivehemodynamic monitoring. Monitoringfilling pressures can be helpful to opti-mize preload and thus CO. Observation oflittle change in the central venous pres-sure in response to a fluid bolus suggeststhat the venous capacitance system is notoverfilled and that more fluid is indi-cated. Observation that an increasingcentral venous pressure is met with re-duced MAP�central venous pressuresuggests that too much fluid has beengiven. Large volumes of fluid for acutestabilization in children have not beenshown to increase the incidence of theacute respiratory distress syndrome (91,103) or cerebral edema (91, 104). In-creased fluid requirements may be evi-dent for several days secondary to loss offluid from the intravascular compart-ment when there is profound capillaryleak (28). Routine fluid choices includecrystalloids (normal saline or lactatedRingers) and colloids (dextran, gelatin, or5% albumin) (6, 105–114). Fresh frozenplasma may be infused to correct abnor-mal prothrombin time and partial throm-boplastin time values, but should not bepushed because it may produce acute hy-potensive effects likely caused by vasoac-tive kinins and high citrate concentra-tion. Because oxygen delivery depends onhemoglobin concentration, hemoglobinshould be maintained at a minimum of10 g/dL (13, 84). Diuretics/peritoneal di-alysis/CRRT are indicated for patientswho develop signs and symptoms of fluidoverload.

    Mechanical Ventilation. There areseveral reasons to initiate intubation andventilation in relation to the hemody-namic support of patients with septicshock. In practice, the first indication isusually the need to establish invasive he-modynamic monitoring. In uncoopera-tive, coagulopathic infants, this is most

    safely accomplished in the sedated, im-mobilized patient. This step should beconsidered in any patient who is not rap-idly stabilized with fluid resuscitation andperipherally administered inotropes.

    Ventilation also provides mechanicalsupport for the circulation. Up to 40% ofCO may be required to support the workof breathing, and this can be unloaded byventilation, diverting flow to vital organs.Increased intrathoracic pressure also re-duces left ventricular afterload that maybe beneficial in patients with low CI andhigh SVR. Ventilation may also providebenefits in patients with elevated pulmo-nary vascular resistance. Mild hyperven-tilation may also be used to compensatefor metabolic acidosis by altering the re-spiratory component of acid-base bal-ance. Caution must be exercised as exces-sive ventilation may impair CO,particularly in the presence of hypovole-mia. Additional volume loading is oftennecessary at this point.

    Sedation and ventilation also facilitatetemperature control and reduce oxygenconsumption. Importantly but less com-monly, ventilation is required because ofclinical and laboratory evidence of respi-ratory failure, impaired mental state, ormoribund condition.

    Sedation for Invasive Procedures orIntubation. Airway and breathing can ini-tially be managed according to PALSguidelines using head positioning, and ahigh flow oxygen delivery system. A re-port published since 2002 supports earlymanagement of dengue shock using highflow nasal cannula O2/continuous posi-tive airway pressure (115). When intuba-tion or invasive procedures are required,patients are at risk of worsening hypoten-sion from the direct myocardial depres-sant and vasodilator effects of inductionagents as well as indirect effects due toblunting of endogenous catecholaminerelease. Propofol, thiopental, benzodiaz-epines, and inhalational agents all carrythese risks. Yamamoto (116) and others(7, 15) suggest using ketamine with atro-pine premedication for sedation and in-tubation in septic shock. Ketamine is acentral NMDA receptor blocker, whichblocks nuclear factor-kappa B transcrip-tion and reduces systemic interleukin-6production while maintaining an intactadrenal axis, which in turn maintainscardiovascular stability (117–125). Ket-amine can also be used as a sedation/analgesia infusion to maintain cardiovas-cular stability during mechanicalventilation (126). Etomidate is popular as

    672 Crit Care Med 2009 Vol. 37, No. 2

  • an induction agent because it maintainscardiovascular stability through blockadeof the vascular K� channel; however,even one dose used for intubation is in-dependently associated with increasedmortality in both children and adultswith septic shock, possibly secondary toinhibition of adrenal corticosteroid bio-synthesis. Therefore, it is not recom-mended for this purpose (127–131). Onlyone member of the task force continuesto support use of etomidate in pediatricseptic shock with the caveat that stressdose hydrocortisone be administered. Lit-tle has been published on the use of ket-amine or etomidate in newborns withshock so we cannot make recommenda-tions for or against the use of these drugsin newborns. When intubation and venti-lation are required the use of neuromus-cular blocking agents should be consid-ered.

    Intravascular Catheters and Monitor-ing. Minimal invasive monitoring is neces-sary in children with fluid-responsiveshock; however, central venous access andarterial pressure monitoring are recom-mended in children with fluid-refractoryshock. Maintenance of perfusion pressure(MAP�central venous pressure), or(MAP�IAP) if the abdomen is tense sec-ondary to bowel edema or ascitic fluid, isconsidered necessary for organ perfusion(38). Echocardiography is considered anappropriate noninvasive tool to rule out thepresence of pericardial effusion, evaluatecontractility, and depending on the skills ofthe echocardiographer, check ventricularfilling. Doppler echocardiography can beused to measure CO and SVC flow. CO�3.3 L/min/m2 �6.0 L/min/m2 and SVCflow �40 mL/kg/min in newborns are as-sociated with improved survival and neuro-logic function. Goal-directed therapy toachieve an ScvO2 saturation �70% is asso-ciated with improved outcome (13). Togain accurate measures of ScvO2, the tip ofthe catheter must be at or close to theSVC-right atrial or inferior vena cava-rightatrial junction (132). A pulmonary arterycatheter, pulse index contour cardiac out-put catheter, or femoral artery thermodilu-tion catheter can be used to measure CO(133) in those who remain in shock despitetherapies directed to clinical signs of per-fusion, MAP-central venous pressure,ScvO2, and echocardiographic analyses(134–144). The pulmonary artery cathetermeasures the pulmonary artery occlusionpressure to help identify selective left ven-tricular dysfunction, and can be used todetermine the relative contribution of right

    and left ventricle work. The pulse indexcontour cardiac output catheter estimatesglobal end-diastolic volume in the heart(both chambers) and extravascular lungwater and can be used to assess whetherpreload is adequate. None of these tech-niques is possible in neonates and smallerinfants. Other noninvasive monitors under-going evaluation in newborns and childreninclude percutaneous venous oxygen satu-ration, aortic ultrasound, perfusion index(pulse-oximetry), near infrared spectros-copy, sublingual Pco2, and sublingual mi-crovascular orthogonal polarization spec-troscopy scanning. All show promise;however, none have been tested in goal-directed therapy trials (145–152).

    Cardiovascular Drug Therapy

    When considering the use of cardio-vascular agents in the management ofinfants and children with septic shock,several important points need emphasis.The first is that septic shock represents adynamic process so that the agents se-lected and their infusion dose may haveto be changed over time based on theneed to maintain adequate organ perfu-sion. It is also important to recognizethat the vasoactive agents are character-ized by varying effects on SVR and pul-monary vascular resistance (i.e., vasodi-lators or vasopressors), contractility (i.e.,inotropy) and HR (chronotropy). Thesepharmacologic effects are determined bythe pharmacokinetics of the agent andthe pharmacodynamics of the patient’sresponse to the agent. In critically ill sep-tic children, perfusion and function ofthe liver and kidney are often altered,leading to changes in drug pharmacoki-netics with higher concentrations ob-served than anticipated. Thus, the infu-sion doses quoted in many textbooks areapproximations of starting rates andshould be adjusted based on the patient’sresponse. The latter is also determined bythe pharmacodynamic response to theagent, which is commonly altered in sep-tic patients. For example, patients withsepsis have a well-recognized reduced re-sponse to alpha-adrenergic agonists thatis mediated by excess NO production aswell as alterations in the alpha-adrener-gic receptor system. Similarly, cardiacbeta-adrenergic responsiveness may bereduced by the effect of NO and otherinflammatory cytokines.


    Dopamine (5–9 �g/kg/min), dobut-amine, or epinephrine (0.05–0.3 �g/kg/min) can be used as first-line inotropicsupport. Dobutamine may be used whenthere is a low CO state with adequate orincreased SVR (29, 84, 153–165). Dobut-amine or mid-dosage dopamine can beused as the first line of inotropic supportif supported by clinical and objective data(e.g., assessment of contractility by echo-cardiogram) when one of the initial goalsis to increase cardiac contractility in pa-tients with normal blood pressure. How-ever, children �12 months may be lessresponsive (161). Recent adult data raisesthe concern of increased mortality withthe use of dopamine (166). There is not aclear explanation for these observations.Possible explanations include the actionof a dopamine infusion to reduce the re-lease of hormones from the anterior pi-tuitary gland, such as prolactin, throughstimulation of the DA2 receptor, whichcan have important immunoprotectiveeffects, and inhibition of thyrotropin re-leasing hormone release. Adult data fa-vors the use of norepinephrine as a firstline agent in fluid-refractory vasodilated(and often hypotensive) septic shock(167–170). Although the majority ofadults with fluid-refractory, dopamine-resistant shock have high CO and lowSVR, children with this condition pre-dominantly have low CO.

    Dobutamine- or dopamine-refractorylow CO shock may be reversed with epi-nephrine infusion (29, 171–174). Epi-nephrine is more commonly used in chil-dren than in adults. Some members ofthe committee recommended use of low-dose epinephrine as a first-line choice forcold hypodynamic shock. It is clear thatepinephrine has potent inotropic andchronotropic effects, but its effects onperipheral vascular resistance and the en-docrine stress response may result in ad-ditional problems. At lower infusiondoses (�0.3 �/kg/min) epinephrine hasgreater beta-2-adrenergic effects in theperipheral vasculature with little alpha-adrenergic effect so that SVR falls, partic-ularly in the skeletal musculature andskin. This may redirect blood flow awayfrom the splanchnic circulation eventhough blood pressure and CO increases.This effect of epinephrine likely explainsthe observation that epinephrine tran-siently reduces gastric intramucosal pHin adults and animals with hyperdynamicsepsis (175), but there are no data avail-

    673Crit Care Med 2009 Vol. 37, No. 2

  • able to evaluate whether gut injury doesor does not occur with epinephrine use inchildren. Epinephrine stimulates glu-coneogenesis and glycogenolysis, and in-hibits the action of insulin, leading toincreased blood glucose concentrations.In addition, as part of the stimulation ofgluconeogenesis, epinephrine increasesthe shuttle of lactate to the liver as asubstrate for glucose production (theCori cycle). Thus, patients on epineph-rine infusion have increased plasma lac-tate concentrations independent ofchanges in organ perfusion, making thisparameter somewhat more difficult to in-terpret in children with septic shock.

    Ideally, epinephrine should be admin-istered by a secure central venous route,but in an emergency it may be infusedthrough a peripheral IV route or throughan intraosseous needle while attainingcentral access. The American Heart Asso-ciation/PALS guidelines for children rec-ommends the initial use of epinephrineby peripheral IV or interosseous for car-diopulmonary resuscitation or post-cardiopulmonary resuscitation shock,and by the intramuscular route for ana-phylaxis (176). Even though a commonperception, there is no data clarifying ifthe peripheral infiltration of epinephrineproduces more local damage than ob-served with dopamine. The severity oflocal symptoms likely depends on theconcentration of the vasoactive drug in-fusion and the duration of the peripheralinfiltration before being discovered. If pe-ripheral infiltration occurs with any cat-echolamine, its adverse effects may beantagonized by local subcutaneous infil-tration with phentolamine, 1–5 mg di-luted in 5 mL of normal saline.


    When pediatric patients are normo-tensive with a low CO and high SVR,initial treatment of fluid-refractory pa-tients consists of the use of an inotropicagent such as epinephrine or dobutaminethat tends to lower SVR. In addition, ashort-acting vasodilator may be added,such as sodium nitroprusside or nitro-glycerin to recruit microcirculation(177–182) and reduce ventricular after-load resulting in better ventricular ejec-tion and global CO, particularly whenventricular function is impaired. Orthog-onal polarizing spectroscopy showed thataddition of systemic IV nitroglycerin todopamine/norepinephrine infusion re-stored tongue microvascular blood flow

    during adult septic shock (183). Nitrova-sodilators can be titrated to the desiredeffect, but use of nitroprusside is limitedif there is reduced renal function second-ary to the accumulation of sodium thio-cyanate; use of nitroglycerin may alsohave limited utility over time through thedepletion of tissue thiols that are impor-tant for its vasodilating effect. Other va-sodilators that have been used in childreninclude prostacyclin, pentoxifylline,dopexamine, and fenoldopam (184–189).

    An alternative approach to improvecardiac contractility and lower SVR isbased on the use of type III phosphodies-terase inhibitors (PDEIs) (190–196). Thisclass of agents, which includes milrinoneand inamrinone (formerly amrinone, butthe name was changed to avoid confusionwith amiodarone), has a synergistic effectwith beta-adrenergic agonists since thelatter agents stimulate intracellular cy-clic adenosine monophosphate (cAMP)production, whereas the PDEIs increaseintracellular cAMP by blocking its hydro-lysis. Because the PDEIs do not dependon a receptor mechanism, they maintaintheir action even when the beta-adrener-gic receptors are down-regulated or havereduced functional responsiveness. Themain limitation of these agents is theirneed for normal renal function (for mil-rinone clearance) and liver function (forinamrinone clearance). Inamrinone andmilrinone are rarely used in adults withseptic shock because catecholamine re-fractory low CO and high vascular resis-tance is uncommon; however, this hemo-dynamic state represents a majorproportion of children with fluid-refrac-tory, dopamine-resistant shock. Fluid bo-luses are likely to be required if inamri-none or milrinone are administered withfull loading doses. Because milrinone andinamrinone have long half lifes (1�10hrs depending on organ function) it cantake 3–30 hrs to reach 90% of steady stateif no loading dose is used. Although rec-ommended in the literature some indi-viduals in the committee choose not touse boluses of inamrinone or milrinone.This group administers the drugs as acontinuous infusion only. Other mem-bers divide the bolus in five equal aliquotsadministering each aliquot over 10 minsif blood pressure remains within an ac-ceptable range. If blood pressure falls, itis typically because of the desired vasodi-lation and can be reversed by titrated(e.g., 5 mL/kg) boluses of isotonic crys-talloid or colloid. Because of the longelimination half-life, these drugs should

    be discontinued at the first sign of ar-rhythmia, or hypotension caused by ex-cessively diminished SVR. Hypotension-related toxicity can also be potentiallyovercome by beginning norepinephrineor vasopressin. Norepinephrine counter-acts the effects of increased cAMP in vas-cular tissue by stimulating the alpha re-ceptor resulting in vasoconstriction.Norepinephrine has little effect at the vas-cular �2 receptor.

    Rescue from refractory shock has beendescribed in case reports and series usingtwo medications with type III phosphodi-esterase activity. Levosimendan is apromising new medication that increasesCa��/actin/tropomyosin complex bind-ing sensitivity and also has some type IIIPDEI and adenosine triphosphate-sensi-tive K� channel activity. Because one ofthe pathogenic mechanisms of endotox-in-induced heart dysfunction is desensi-tization of Ca��/actin/tropomyosin com-plex binding (197–202), this drug allowstreatment at this fundamental level ofsignal transduction overcoming the lossof contractility that characterizes septicshock. Enoximone is a type III PDEI with10 times more �1 cAMP hydrolysis inhi-bition than �2 cAMP hydrolysis inhibition(203–205). Hence, it can be used to in-crease cardiac performance with less riskof undesired hypotension.

    Vasopressor Therapy

    Dopamine remains the first-line vaso-pressor for fluid-refractory hypotensiveshock in the setting of low SVR. However,there is some evidence that patientstreated with dopamine have a worse out-come than those treated without dopa-mine (206) and that norepinephrine,when used exclusively in this setting,leads to adequate outcomes (168). Thereis also literature demonstrating an age-specific insensitivity to dopamine (207–216). Dopamine causes vasoconstrictionby releasing norepinephrine from sympa-thetic vesicles as well as acting directlyon alpha-adrenergic receptors. Immatureanimals and young humans (�6 months)may not have developed their full compo-nent of sympathetic innervation so theyhave reduced releasable stores of norepi-nephrine. Dopamine-resistant shockcommonly responds to norepinephrine orhigh-dose epinephrine (29, 217–219).Some committee members advocate theuse of low-dose norepinephrine as a first-line agent for fluid-refractory hypotensivehyperdynamic shock. Based on experi-

    674 Crit Care Med 2009 Vol. 37, No. 2

  • mental and clinical data, norepinephrineis recommended as the first-line agent inadults with fluid-refractory shock. If thepatient’s clinical state is characterized bylow SVR (e.g., wide pulse pressure withDBP that is less than one-half the systolicpressure), norepinephrine is recom-mended alone. Other experts have recom-mended combining norepinephrine withdobutamine, recognizing that dobut-amine is a potent inotrope that has in-trinsic vasodilating action that may behelpful to counteract excessive vasocon-striction from norepinephrine. Improvedcapillary and gut blood flow were ob-served in animal and human studies withnorepinephrine plus dobutamine in com-parison with high-dose dopamine or epi-nephrine.

    Vasopressin has been shown to in-crease MAP, SVR, and urine output inpatients with vasodilatory septic shockand hyporesponsiveness to catecholamines(167, 220–229). Vasopressin’s action isindependent of catecholamine receptorstimulation, and therefore its efficacy isnot affected by alpha-adrenergic receptordown-regulation often seen in septicshock. Terlipressin, a long acting form ofvasopressin, has been reported to reversevasodilated shock as well (228, 230).

    Although angiotensin can also be usedto increase blood pressure in patientswho are refractory to norepinephrine, itsclinical role is not as well defined (231).Phenylephrine is another pure vasopres-sor with no beta-adrenergic activity(232). Its clinical role is also limited sinceit may improve blood pressure but reduceblood flow through its action to increaseSVR. Vasopressors can be titrated to endpoints of perfusion pressure (MAP-centralvenous pressure) or SVR that promoteoptimum urine output and creatinineclearance (69, 71, 217, 218), but excessivevasoconstriction compromising micro-circulatory flow should be avoided. NOinhibitors and methylene blue are consid-ered investigational therapies (233–235).Studies have shown an increased mortal-ity with nonselective NO synthase inhib-itors suggesting that simply increasingblood pressure through excessive vaso-constriction has adverse effects (138).Low-dose arginine vasopressin (in doses�0.04 units/kg/min) as an adjunctiveagent has short-term hemodynamic ben-efits in adults with vasodilatory shock. Itis not currently recommended for treat-ment of cardiogenic shock, hence itshould not be used without ScvO2/COmonitoring. The effect of low-dose argi-

    nine vasopressin on clinically importantoutcomes such as mortality remains un-certain. The Vasopressin and SepticShock Trial, a randomized controlledclinical trial that compared low-dose ar-ginine vasopressin with norepinephrinein patients with septic shock, showed nodifference between regimens in the 28-day mortality primary end point (236).The safety and efficacy of low-dose argi-nine vasopressin have yet to be demon-strated in children with septic shock, andawait the results of an ongoing random-ized controlled trial (237, 238).

    Glucose, Calcium, Thyroid, and Hy-drocortisone Replacement. It is impor-tant to maintain metabolic and hormonalhomeostasis in newborns and children.Hypoglycemia can cause neurologic dev-astation when missed. Therefore, hypo-glycemia must be rapidly diagnosed andpromptly treated. Required glucose infu-sion rates for normal humans are agespecific but can be met by delivering aD10%-containing solution at mainte-nance fluid rates (8 mg/kg/min glucose innewborns, 5 mg/kg/min glucose in chil-dren, and 2 mg/kg/min in adolescents).Patients with liver failure will requirehigher glucose infusion rates (up to 16mg/kg/min). Hyperglycemia is also a riskfactor for mortality. Lin and Carcillo(239) reported that children with septicshock, who had hyperglycemia (�140mg/dL) and an elevated anion gap,showed resolution of their anion gapwhen insulin was added to their glucoseregimen. This was associated with rever-sal of catabolism as measured by urinaryorganic acids. Infants with metabolic dis-ease are particularly vulnerable to cata-bolic failure and must be treated withappropriate glucose delivery, and whenneeded insulin to assure glucose uptake,during septic shock. It is important tonote that insulin requirements decreaseat approximately 18 hrs after the onset ofshock. Infusion of insulin and glucose arealso effective inotropes. Two members ofthe task force preferred using D5%-containing solution for patients with hy-perglycemia. Greater than 90% of thecommittee agreed with meeting glucoserequirements and treating hyperglycemiawith insulin. Hypocalcemia is a frequent,reversible contributor to cardiac dysfunc-tion as well (56, 240, 241). Calcium re-placement should be directed to normal-ize ionized calcium concentration. Onemember of the task force did not agreethat calcium replacement should begiven for hypocalcemia. All agreed that

    care should be taken to not overtreat ascalcium toxicity may occur with elevatedconcentrations.

    Replacement with thyroid and/or hy-drocortisone can also be lifesaving inchildren with thyroid and/or adrenal in-sufficiency and catecholamine-resistantshock (29, 55, 242–260). Infusion therapywith triiodothyronine has been beneficialin postoperative congenital heart diseasepatients but has yet to be studied in chil-dren with septic shock (253). Hypothy-roidism is relatively common in childrenwith trisomy 21 and children with centralnervous system pathology (e.g., pituitaryabnormality). Unlike adults, children aremore likely to have absolute adrenal in-sufficiency defined by a basal cortisol�18 �g/dL and a peak adrenocortico-tropic hormone (ACTH)-stimulated corti-sol concentration �18 �g/dL. Nonsurvi-vors have exceedingly high ACTH/cortisolratios within the first 8 hrs of meningo-coccal shock (206). Aldosterone levels arealso markedly depressed in meningococ-cemia (261). Patients at risk of inade-quate cortisol/aldosterone production inthe setting of shock include children withpurpura fulminans and Waterhouse-Friderichsen syndrome, children whopreviously received steroid therapies forchronic illness, and children with pitu-itary or adrenal abnormalities. Review ofthe pediatric literature found case series(251, 252) and randomized trials (242,243) that used “shock dose” hydrocorti-sone in children. The first randomizedcontrolled trial showed improved out-come with hydrocortisone therapy inchildren with dengue shock. The secondstudy was underpowered and showed noeffect of hydrocortisone therapy on out-come in children with dengue shock. Thereported shock dose of hydrocortisone is25 times higher than the stress dose (242,243, 247, 248, 250–252, 258, 259). Atpresent the committee makes no changesfrom the 2002 recommendation. Thecommittee only recommends hydrocorti-sone treatment for patients with absoluteadrenal insufficiency (peak cortisol con-centration attained after corticotropinstimulation �18 �g/dL) or adrenal-pituitary axis failure and catecholamine-resistant shock. Some support the use ofstress dose only whereas others supportthe use of shock dosage when needed. Inthe absence of any new studies sheddinglight on the subject since 2002, the dosecan be titrated to resolution of shockusing between 2 mg/kg and 50 mg/kg/dayas a continuous infusion or intermittent

    675Crit Care Med 2009 Vol. 37, No. 2

  • dosing if desired. The treatment shouldbe weaned off as tolerated to minimizepotential long-term toxicities.

    Administration of prolonged hydro-cortisone and fludrocortisone (6 mg/kg/day cortisol equivalent � 7 days) hadbeen recommended for adults with do-pamine-resistant septic shock and rela-tive adrenal insufficiency (basal cortisol�18 �g/dL with cortisol increment aftercorticotropin stimulation �9 �g/dL)(260); however, adult guidelines now rec-ommend this therapy for any adult withdopamine-resistant septic shock. Thecontinuing debate on whether thisshould similarly be an adjunctive therapyfor pediatric sepsis will likely only be re-solved with yet-to-be done pediatric tri-als. Since 2002, a randomized trial of a7-day course of 3 mg/kg/day of intermit-tent hydrocortisone therapy for dopam-ine-treated septic shock in premature ba-bies was performed. These babies hadreduced dopamine requirements but noimprovement in mortality (58, 262, 264).Unlike dexamethasone, which was associ-ated with neurologic consequences inpremature babies (261), hydrocortisonedid not cause similar complications inpremature babies (263).

    Multiple pediatric studies conductedover the interval 1999–2006 provide con-sistent evidence that children who suc-cumbed from septic shock exhibitedlower cortisol levels than those who sur-vived, and that septic shock nonsurvivorshad lower random plasma cortisol con-centrations compared with septic shocksurvivors; the latter had lower randomplasma cortisol concentrations comparedwith sepsis survivors (254, 255, 265–267).This effect is not attributable to inade-quate ACTH adrenal stimulation; on thecontrary, an opposite trend prevails,namely septic shock nonsurvivors exhibithigh circulating ACTH concentrationscompared with septic shock survivors,who in turn have higher circulatingACTH concentrations compared with pa-tients with sepsis. One retrospective co-hort study using the Pediatric Health In-formation System database examinedfactors associated with outcome in chil-dren with severe sepsis as operationallyidentified by a combination of infectionplus need for a vasoactive infusion andmechanical ventilation (268). Among6693 children meeting the definition ofsevere sepsis, mortality was 30% for chil-dren who received steroids comparedwith 18% for those who did not (crudeodds ratio 1.9) (95% confidence interval

    1.7–2.2). An important liability of thisinvestigation relates to lack of illness se-verity data. Although steroids may havebeen given preferentially to more severelyill children, their use was associated withincreased mortality. Steroid use waslinked to disseminated candidiasis in acase report (269). The committee contin-ues to maintain equipoise on the ques-tion of adjunctive steroid therapy for pe-diatric sepsis (outside of classic adrenalor hypophyseal pituitary axis (HPA) axisinsufficiency), pending prospective ran-domized clinical trials.

    Persistent Pulmonary Artery Hyper-tension of the Newborn Therapy. InhaledNO therapy is the treatment of choice foruncomplicated PPHN (270, 271). However,metabolic alkalinization remains an impor-tant initial resuscitative strategy duringshock because PPHN can reverse when ac-idosis is corrected (272). For centers withaccess to inhaled NO, this is the only selec-tive pulmonary vasodilator reported to beeffective in reversal of PPHN (270, 271,273–278). Milrinone or inamrinone may beadded to improve heart function as toler-ated (279–281). ECMO remains the ther-apy of choice for patients with refractoryPPHN and sepsis (282–285). New investiga-tions support use of inhaled iloprost (syn-thetic analog of prostacyclin) or adenosineinfusion as modes of therapy for PPHN(286–291).

    Extracorporeal Therapies. ECMO isnot routinely used in adults (with thenotable exception of the University ofMichigan) (282). ECMO is a viable ther-apy for refractory septic shock in neo-nates (283) and children because neo-nates (approximately 80% survival) andchildren (approximately 50% survival)(292–295) have the same outcomeswhether the indication for ECMO is re-fractory respiratory failure or refractoryshock from sepsis or not. It is also effec-tive in adult hantavirus victims with lowCO/high SVR shock (296, 297). AlthoughECMO survival is similar in pediatric pa-tients with and without sepsis, throm-botic complications can be more com-mon in sepsis. Efforts are warranted toreduce ECMO-induced hemolysis becausefree heme scavenges NO, adenosine, anda disintegrin and metalloprotease withthrombospondin motifs-13 (ADAMTS-13;von Willebrand factor cleaving protease)leading to microvascular thrombosis, re-versal of portal blood flow and multipleorgan failure (298, 299). Nitroglycerin(NO donor), adenosine, and fresh frozenplasma (FFP) (ADAMTS-13) can be in-

    fused to attempt to neutralize these ef-fects. Hemolysis can be avoided, in part,by using the proper-sized cannula for ageand limiting ECMO total blood flow to nogreater than 110 mL/kg/min. AdditionalCO can be attained using inotrope/vaspodilator therapies.

    Investigators also reported that theuse of high flux CRRT (�35 mL/kg/hfiltration-dialysis flux), with concomitantFFP or antithrombotic protein C infusionto reverse prolonged INR without causingfluid overload, reduced inotrope/vaso-pressor requirements in children with re-fractory septic shock and purpura (7, 15,300–305). The basis of this beneficial ef-fect remains unknown. It could resultfrom prevention of fluid overload, clear-ance of lactate and organic acids, bindingof inflammatory mediators, reversal ofcoagulopathy, or some combination ofthese actions.


    Pediatric Septic Shock

    Diagnosis: The inflammatory triad offever, tachycardia, and vasodilation arecommon in children with benign infec-tions (Fig. 1). Septic shock is suspectedwhen children with this triad have achange in mental status manifested asirritability, inappropriate crying, drowsi-ness, confusion, poor interaction withparents, lethargy, or becoming unarous-able. The clinical diagnosis of septicshock is made in children who 1) have asuspected infection manifested by hypo-thermia or hyperthermia, and 2) haveclinical signs of inadequate tissue perfu-sion including any of the following; de-creased or altered mental status, pro-longed capillary refill �2 secs (coldshock), diminished pulses (cold shock)mottled cool extremities (cold shock) orflash capillary refill (warm shock),bounding peripheral pulses, and widepulse pressure (warm shock) or decreasedurine output �1 ml/kg/h. Hypotension isnot necessary for the clinical diagnosis ofseptic shock; however, its presence in achild with clinical suspicion of infectionis confirmatory.

    ABCs: The First Hour ofResuscitation (EmergencyRoom Resuscitation)

    Goals: (Level III). Maintain or restoreairway, oxygenation, and ventilation (Table1); maintain or restore circulation, defined

    676 Crit Care Med 2009 Vol. 37, No. 2

  • as normal perfusion and blood pressure;maintain or restore threshold HR.

    Therapeutic End Points (Level III).Capillary refill �2 secs, normal pulseswith no differential between the quality ofperipheral and central pulses, warm ex-tremities, urine output �1 mL/kg/h, nor-mal mental status, normal blood pressurefor age (noninvasive blood pressure onlyreliable when pulses palpable), normalglucose concentration, normal ionizedcalcium concentration.

    Monitoring (Level III). Pulse oximeter,continuous electrocardiography, bloodpressure and pulse pressure. Note pulse

    pressure and diastolic pressure to helpdistinguish between low SVR (wide pulsepressure due to low DBP) and high SVR(narrow pulse pressure). Temperature,urine output, glucose, ionized calcium.

    Airway and Breathing (Level III). Air-way and breathing should be rigorouslymonitored and maintained. Lung compli-ance and work of breathing may changeprecipitously. In early sepsis, patients of-ten have a respiratory alkalosis from cen-trally mediated hyperventilation. As sepsisprogresses, patients may have hypoxemia aswell as metabolic acidosis and are at highrisk to develop respiratory acidosis sec-

    ondary to a combination of parenchymallung disease and/or inadequate respira-tory effort due to altered mental status.The decision to intubate and ventilate isbased on clinical assessment of increasedwork of breathing, hypoventilation, orimpaired mental status. Waiting for con-firmatory laboratory tests is discouraged.Up to 40% of CO is used for work ofbreathing. Therefore, intubation and me-chanical ventilation can reverse shock. Ifpossible, volume loading and peripheralor central inotropic/vasoactive drug sup-port is recommended before and duringintubation because of relative or absolutehypovolemia, cardiac dysfunction, andthe risk of suppressing endogenous stresshormone response with agents that facil-itate intubation. Etomidate is not recom-mended. Ketamine with atropine pretreat-ment and benzodiazepine postintubation canbe used as a sedative/induction regimenof choice to promote cardiovascular in-tegrity. A short-acting neuromuscularblocker can facilitate intubation if theprovider is confident she/he can maintainairway patency.

    Circulation (Level II). Vascular accessshould be rapidly attained. Establish in-traosseous access if reliable venous accesscannot be attained in minutes. Fluid re-suscitation should commence immedi-ately unless hepatomegaly/rales arepresent. Recall that rales may be heard inchildren with pneumonia as a cause ofsepsis, so it does not always imply thatthe patient is fluid overloaded. If pneu-monia is suspected or confirmed, fluidresuscitation should proceed with carefulmonitoring of the child’s work of breath-ing and oxygen saturation. In the fluid-refractory patient, begin a peripheral ino-trope (low-dose dopamine or epinephrine)if a second peripheral IV/intraosseuscatheter is in place, while establishing acentral venous line. When administeredthrough a peripheral IV/intraosseus cath-eter, the inotrope should be infused ei-ther as a dilute solution or with a secondcarrier solution running at a flow rate toassure that it reaches the heart in atimely fashion. Care must be taken toreduce dosage if evidence of peripheralinfiltration/ischemia occurs as alpha-adrenergic receptor-mediated effects oc-cur at higher concentrations for epineph-rine and dopamine. Central dopamine,epinephrine, or norepinephrine can beadministered as a first line drug as indi-cated by hemodynamic state when a cen-tral line is in place. It is generally appro-priate to begin central venous infusion

    Figure 1. Algorithm for time sensitive, goal-directed stepwise management of hemodynamic supportin infants and children. Proceed to next step if shock persists. 1) First hour goals—Restore andmaintain heart rate thresholds, capillary refill �2 sec, and normal blood pressure in the firsthour/emergency department. Support oxygenation and ventilation as appropriate. 2) Subsequentintensive care unit goals—If shock is not reversed, intervene to restore and maintain normal perfusionpressure (mean arterial pressure [MAP]-central venous pressure [CVP]) for age, central venous O2saturation �70%, and CI �3.3, �6.0 L/min/m2 in pediatric intensive care unit (PICU). Hgb, hemo-globin; PICCO, pulse contour cardiac output; FATD, femoral arterial thermodilution; ECMO, extra-corporeal membrane oxygenation; CI, cardiac index; CRRT, continuous renal replacement therapy; IV,intravenous; IO, interosseous; IM, intramuscular.

    677Crit Care Med 2009 Vol. 37, No. 2

  • and wait until a pharmacologic effect isobserved before stopping the peripheralinfusion.

    Fluid Resuscitation (Level II). Rapidfluid boluses of 20 mL/kg (isotonic crys-talloid or 5% albumin) can be adminis-tered by push or rapid infusion device(pressure bag) while observing for signsof fluid overload (i.e., the development ofincreased work of breathing, rales, galloprhythm, or hepatomegaly). In the ab-sence of these clinical findings, repeatedfluid boluses can be administered to asmuch as 200 mL/kg in the first hour.Children commonly require 40 – 60mL/kg in the first hour. Fluid can bepushed with the goal of attaining normalperfusion and blood pressure. Hypoglyce-mia and hypocalcemia should be cor-rected. A D10%-containing isotonic IVsolution can be run at maintenance IVfluid rates to provide age appropriate glu-cose delivery and to prevent hypoglyce-mia.

    Hemodynamic Support (Level II).Central dopamine may be titratedthrough central venous access. If thechild has fluid refractory/dopamine resis-tant shock, then central epinephrine canbe started for cold shock (0.05–0.3 �g/kg/min) or norepinephrine can be ti-trated for warm shock to restore normalperfusion and blood pressure.

    Hydrocortisone Therapy (Level III). Ifa child is at risk of absolute adrenal in-sufficiency or adrenal pituitary axis fail-ure (e.g., purpura fulminans, congenitaladrenal hyperplasia, prior recent steroidexposure, hypothalamic/pituitary abnor-mality) and remains in shock despite epi-nephrine or norepinephrine infusion,then hydrocortisone can be administeredideally after attaining a blood sample forsubsequent determination of baselinecortisol concentration. Hydrocortisonemay be administered as an intermittentor continuous infusion at a dosage whichmay range from 1–2 mg/kg/day for stresscoverage to 50 mg/kg/day titrated to re-versal of shock.

    Stabilization: Beyond the FirstHour (Pediatric Intensive CareUnit Hemodynamic Support)

    Goals: (Level III). Normal perfusion;capillary refill �2 secs, threshold HRs.perfusion pressure (MAP�central venouspressure, or MAP�IAP) appropriate forage. ScvO2 �70%; CI �3.3 L/min/m

    2 and�6.0 L/min/m2.

    Therapeutic End Points: (Level III).Capillary refill �2 secs, threshold HRs,normal pulses with no differential be-tween the quality of the peripheral andcentral pulses, warm extremities, urineoutput �1 mL/kg/h, normal mental sta-tus, CI �3.3 and �6.0 L/min/m2 withnormal perfusion pressure (MAP�centralvenous pressure, or MAP�IAP) for age,ScvO2 �70%; maximize preload to maxi-mize CI, MAP � central venous pressure;normal INR, anion gap, and lactate.

    Monitoring (Level III). Pulse oximetry,continuous electrocardiogram, continuousintra-arterial blood pressure, temperature(core), urine output, central venous pres-sure/O2 saturation and/or pulmonary arterypressure/O2 saturation, CO, glucose andcalcium, INR, lactate, and anion gap.

    Fluid Resuscitation (Level II). Fluidlosses and persistent hypovolemia sec-ondary to diffuse capillary leak can con-tinue for days. Ongoing fluid replacementshould be directed at clinical end pointsincluding perfusion, central venous pres-sure, echocardiographic determination ofend-diastolic volume, pulmonary capil-lary wedge pressure/end-diastolic volume(when available), and CO. Crystalloid isthe fluid of choice in patients with hemo-globin �10 g/dL. Red blood cell transfu-sion can be given to children with hemo-globin �10 g/dL. FFP is recommendedfor patients with prolonged INR but as aninfusion, not a bolus. After shock resus-citation, diuretics/peritoneal dialysis/high flux CRRT can be used to removefluid in patients who are 10% fluid over-loaded and unable to maintain fluid bal-ance with native urine output/extrarenallosses.

    Elevated lactate concentration and an-ion gap measurement can be treated byassuring both adequate oxygen deliveryand glucose utilization. Adequate oxygendelivery (indicated by a ScvO2 � 70%) canbe achieved by attaining hemoglobin �10g/dL and CO �3.3 L/min/m2 using ade-quate volume loading and inotrope/vasodilator support when needed (as de-scribed below). Appropriate glucosedelivery can be attained by giving a D10%containing isotonic IV solution at fluidmaintenance rate. Appropriate glucoseuptake can be attained in subsequentlyhyperglycemic patients by titrating an in-sulin infusion to reverse hyperglycemia(keep glucose concentration �150 mg/dL) while carefully monitoring to avoidhypoglycemia (keep glucose concentra-tion �80 mg/dL). The use of lesser glu-cose infusion rates (e.g., D5% or lower

    volumes of D10%) will not provide glu-cose delivery requirements.

    Hemodynamic support (Level II). He-modynamic support can be required fordays in children with fluid-refractory do-pamine resistant shock. Children withcatecholamine resistant shock canpresent with low CO/high SVR, high CO/low SVR, or low CO/low SVR shock. Al-though children with persistent shockcommonly have worsening cardiac fail-ure, hemodynamic states may completelychange with time. A pulmonary artery,pulse index contour cardiac output, orfemoral artery thermodilution catheter,or Doppler ultrasound should be usedwhen poor perfusion, including reducedurine output, acidosis, or hypotensionpersists despite use of hemodynamictherapies guided by clinical examination,blood pressure analysis, and arterial andScvO2 analysis. Children with cate-cholamine-resistant shock can respond toa change in hemodynamic therapeuticregimen with resolution of shock. Ther-apies should be directed to maintainmixed venous/ScvO2 �70%, CI �3.3L/min/m2 �6.0 L/min/m2, and a normalperfusion pressure for age (MAP-centralvenous pressure).

    Shock with Low CI, Normal BloodPressure, and High Systemic VascularResistance (Level II). This clinical state issimilar to that seen in a child with car-diogenic shock in whom afterload reduc-tion is a mainstay of therapy designed toimprove blood flow by reducing ventric-ular afterload and thus increasing ven-tricular emptying. Thus, nitroprusside ornitroglycerin is first line vasodilators inpatients with epinephrine-resistant septicshock and normal blood pressure. If cya-nide or isothiocyanate toxicity developsfrom nitroprusside, or methemoglobintoxicity develops from nitroglycerin, orthere is a continued low CO state, thenthe clinician should substitute milrinoneor inamrinone. As noted above, the longelimination half-life of these drugs canlead to slowly reversible toxicities (hypo-tension, tachyarrhythmias, or both) par-ticularly if abnormal renal or liver func-tion exists. Such toxicities can bereversed, in part, with norepinephrine orvasopressin infusion. Additional volumeloading may be necessary to prevent hy-potension when loading doses of milri-none or inamrinone are used. Levosi-mendan and enoximone may have a rolein recalcitrant low CO syndrome. Thyroidreplacement with triiodothyronine iswarranted for thyroid insufficiency, and

    678 Crit Care Med 2009 Vol. 37, No. 2

  • hydrocortisone replacement can be war-ranted for adrenal or hypothalamic-pituitary-adrenal axis insufficiency.

    Shock with Low CI, Low Blood Pres-sure, and Low Systemic Vascular Resis-tance (Level II). Norepinephrine can beadded to epinephrine to increase DBP andSVR. Once an adequate blood pressure isachieved, dobutamine, type III PDEI (par-ticularly enoximone, which has little va-sodilatory properties), or levosimendancan be added to norepinephrine to im-prove CI and ScvO2. Thyroid replacementwith triiodothyronine is warranted forthyroid insufficiency, and hydrocortisonereplacement is warranted for adrenal orhypothalamic-pituitary-adrenal axis in-sufficiency.

    Shock with High CI and Low SystemicVascular Resistance (Level II). When ti-tration of norepinephrine and fluid doesnot resolve hypotension, then low dosevasopressin, angiotensin, or terlipressincan be helpful in restoring blood pres-sure; however, these potent vasocon-strictors can reduce CO, therefore it isrecommended that these drugs are usedwith CO/ScvO2 monitoring. In this sit-uation, additional inotropic therapiesmay be required, such as low-dose epi-nephrine or dobutamine, or the vaso-pressor infusion may be reduced. Terli-pressin is a