Resuscitation (2006) 68, 351—358

CLINICAL PAPER

In-hospital cardiac arrest—–An Utstein style reportof seven years experience from the SahlgrenskaUniversity Hospital�

Martin Fredriksson ∗, Solveig Aune, Ann-Britt Thoren, Johan Herlitz

Sahlgrenska University Hospital, Department of Cardiology, SE-413 45 Goteborg, Sweden

Received 4 March 2005; received in revised form 5 July 2005; accepted 15 July 2005

KEYWORDSIn-hospital;Cardiac arrest;Heart arrest;Utstein;Outcome;DNAR

SummaryBackground: In-hospital cardiac arrest is one of the most stressful situations in mod-ern medicine. Since 1997, there has been a uniform way of reporting — the Utsteinguidelines for in-hospital cardiac arrest reporting.Material and methods: We have studied all consecutive cardiac arrest in theSahlgrenska University Hospital (SU) between 1994 and 2001 for who the rescueteam was alerted in all 833 patients. The primary endpoint for this study was sur-vival to discharge.Results: Thirty-seven percent survived to hospital discharge. Among patients whowere discharged alive, 86% were alive 1 year later. The survivors have a good cerebraloutcome (94% among those who were discharged alive had cerebral performancecategory (CPC) score 1 or 2). The organization at SU is efficient; 80% of the cardiacarrest had CPR within 1 min. Time from cardiac arrest to first defibrillation is amedian of 2 min. Almost two-thirds of the patients were admitted for cardiac relateddiagnoses.Conclusion: The current study is the largest single-centre study of in hospital cardiacarrest reported according to the Utstein guidelines. We report a high survival for in-hospital cardiac arrest. We have pointed out that a functional chain of survival, shortintervals before the start of CPR and defibrillation are probably contributing factorsfor this.© 2005 Elsevier Ireland Ltd. All rights reserved.

� A Spanish translated version of the summary of this articleappears as Appendix in the online version at 10.1016/j.resuscitation.2005.07.011.

∗ Corresponding author. Tel.: +46 31 8279 73.E-mail address: marty.lisa@swipnet.se (M. Fredriksson).

Introduction

Cardiac arrest in the hospital environment is one ofthe most stress filled situations in medical practice.Every minute counts. This area of research has hadthe same problem in reporting as out-of-hospitalcardiac arrest. For the latter, there has been a uni-form way of reporting since 1990 — the Utstein

0300-9572/$ — see front matter © 2005 Elsevier Ireland Ltd. All rights reserved.doi:10.1016/j.resuscitation.2005.07.011

352 M. Fredriksson et al.

guidelines. The advantages of a uniform nomen-clature are obvious and the in-hospital Utsteinguidelines were created in 1997.1 Since then sev-eral studies have been published,2—8 although somewere based on small numbers2,7 or only patientsfound in ventricular tachycardia or ventricular fib-rillation (VT/VF).6 The field is in need of moreresearch to clarify epidemiology, outcome and fac-tors associated with survival. Therefore, have weconducted this study of all consecutive in-hospitalcardiac arrests for whom resuscitation was startedin Sahlgrenska University Hospital (SU), situatedin Gothenburg, Sweden between 1 November 1994and 31 December 2001.

The aim of the study is to describe the out-come of patients suffering in-hospital cardiac arrestaccording to the Utstein style. We have previouslyreported from this registry at an earlier stage butnot in the Utstein style of reporting.9—12

Methods — general

We have collected the data in a prospective way,recording various factors at resuscitation including

(ALS) is performed accordingly to the EuropeanResuscitation Council guidelines.

Resuscitation equipmentThe ICU and CCU are equipped with manual defibril-lators. Other areas in the hospital have automatedexternal defibrillators (AED) close by. As an extraprecaution there is also a member of the assis-tant medical staff who receives the alarm and runsto the scene with a portable AED. Every ward hastheir own standardised emergency equipment withdrugs, intubation kit, etc.

Resuscitation alertIf a patient encounters serious medical problems(e.g. cardiac arrest, airway difficulties) the person-nel calls a specific number, which gives them directaccess to the hospital switchboard. The telephon-ist at the switchboard then follows standardisedinstructions and alerts the resuscitation team byradiopager. The resuscitation-team reports in byphone to the switchboard and is directed to theward or area where the event is in progress. Thetelephonist also notes the personal identificationnumber of the patient (this is unique for all SwedishcebhtswIr

DSdrai

D

Tpowpnea

Att

the time when the CPR team was alerted. We thenconducted a retrospective evaluation of the med-ical records of patients’ previous history and finaloutcome.

Settings

Sahlgrenska University Hospital is a primary/secondary/tertiary hospital. The hospital is dividedover three different locations: Sahlgrenska (SU/S),Ostra (SU/O) and Molndal (SU/M). SU are servingboth a local population of about 600,000 as a pri-mary/secondary and tertiary hospital and a regionalpopulation of about 1,500,000 as a tertiary hospital.In this study we report the cardiac arrests at SU/S.Like all large hospital organizations, there has beenconstant change in the number of wards and avail-able beds, but during the study period there hasbeen an average of about 900 beds, 24 beds inthe intensive care unit (ICU), and 38 beds in thecoronary care unit (CCU) of which 14 have ICU capa-bility.

Resuscitation team

During the study period there was a designatedresuscitation team available around the clock. It iscomposed of three physicians: a specialist in inter-nal medicine, an anaesthesiologist and a cardiolo-gist. Cardiopulmonary resuscitation (CPR) includingbasic life support (BLS) and advanced life support

itizens), ward number, date and time on a specialmergency call list. These lists are later collectedy the ‘‘CPR-centre’’ for analysis and follow-up. Allospital staff are trained in BLS at a regular basis,he hospital has its own organization for this — theo-called ‘‘CPR-centre’’. Medical personnel at theards are also trained to use the AED. The nurses at

CU and CCU also receive training in manual defib-illation.

o not attempt resuscitation (DNAR)U has a policy that the senior physician at the wardecides about DNAR in consensus with the patient orelatives. This is stated in the hospital record, buts a dynamic, on going decision which may be mod-fied according to the patient’s medical progress.

ata acquisition

he data on who, what and where the event tooklace came from three different sources: (1) the listf all emergency calls to the telephone switchboardhich enabled us to access information about whichatient, date, time and location. (2) The wardurse completed a standardized protocol about thevent. (3) The cardiologist on scene also completeddifferent protocol about the event.All this information is later put into a Microsoft©

ccess© database. From this data we then retrievehe patient’s hospital records including death cer-ificate and autopsy report. Thereafter additional

In-hospital cardiac arrest—–An Utstein style 353

data about the patient and event is put into thedatabase via a formal protocol.

Mearsurement and recording of times andintervals

The responsible nurse at the scene measures theintervals. The timing is made by referring to oneclock only to avoid inter-clock differences. Thetimes are noted in the specific nurse cardiac arrestprotocol as mentioned under 2.3.

Cerebral performance category (CPC score)

Cerebral performance category is a way to mea-sure the neurological outcome. A CPC score of 1represents no neurological impairment, CPC score2 represents mild impairment with capacity of theactivities of daily living (ADL). CPC scores 3 and 4represents poor neurological outcome, dependenton others for ADL or living in a sheltered environ-ment.

R

T

W(wpo4p

Figure 1 CPC score for survivors (N = 310).

Cerebral performance category (CPC) score(Figure 1)

The CPC scores prior to the cardiac arrest for thewhole group (N = 833) was as follows: 87% had CPCscore of 1, 7% has CPC score 2. Corresponding fig-ures for survivors are shown in Figure 1.

Times and intervals (Table 1)

Out of the 531 cardiac arrests, for which we hadinformation regarding the interval from the col-lapse until start of CPR, 80% received CPR withinone minute.

Witnessing of cardiac arrest

The cardiac arrests were witnessed in 91% of cases.Cardiac arrest patients presenting with a initialrecorded rhythm of VT/VF were witnessed in 97%.Cases with other rhythms were witnessed in 87%.

esults

emplate

e report an overall survival to discharge of 34%310/910) of the arrests. There were 833 patientsho in all had 910 cardiac arrests. Out of the 833atients, 310 survived to discharge, a survival ratef 37% (310/833). Out of the 523 patients who died,35 (83%) died within 24 h after collapse. See Tem-late 1.

354 M. Fredriksson et al.

Table 1 Time intervals

N Mean (min) Median (min) S.D. (min) Range (min)

CA to E.T.a alerted 581 1,46 1 2,24 0—16CA to CPR-start 531 1,09 0 2,50 0—25CA to 1st defibrillation 333 4,28 2 7,37 0—81

a E.T.: emergency team.

Where did the cardiac arrest occur?

Forty-seven percent of the cardiac arrests occurredin a monitored area, i.e. an area with ECG monitor-ing facilities.

Why were the patients hospitalised?(Table 2)

Almost one-third of the patients were admitteddue to acute myocardial infarction. Almost two-thirds of the patients were admitted due to cardiacrelated diagnoses.

Table 2 Reason for hospitalization (N = 831)

Reason of admission —diagnosis

N (%)

Acute myocardial infarction 252 (30.3%)Other diagnosis 83 (10%)Angina pectoris 75 (9.0%)Heart failure 57 (6.9%)

Trauma (surgical) 7 (0.8%)Other cardiac investigations 4 (0.5%)Diabetes mellitus 4 (0.5%)Rheumatic disease 4 (0.5%)Hypertension 2 (0.2%)Transitory cerebral

ischaemic attack1 (0.1%)

Table 3 Previous history on admission to hospital

Health history at admissionto hospital

N (%)

Angina pectoris 353 (43%)Hypertension 299 (36%)Myocardial infarction 284 (34%)Heart failure 238 (29%)Diabetes mellitus 161 (20%)Smoking 142 (19%)Valvuler heart disease 124 (15%)Stroke 108 (13%)Obstructive pulmonary disease 92 (11%)Malignancy 94 (11%)Renal disease 89 (11%)Previously treated malignancy 20 (10%)Peripheral artery disease 19 (9%)Alcoholism 41 (5%)Rheumatic disease 35 (4%)Previous cardiac arrest 31 (4%)Cardiomyopathy 5 (2%)

Patients characteristics — co-morbidity(Table 3)

Nearly half of the patients had a history of anginapectoris, one-third had a history of myocardialinfarction and one fifth had a history of diabetes.

Age and survival (Table 4)

Seventy percent of the resuscitation attempts werestarted on patients 65 years or older. Survival ratein the age group 75—84 was 29%.

Sex

Sixty-three percent of the studied patients weremale. Thirty-eight percent of the males and 37%of the females survived.

Patients with more than one cardiac arrestduring the hospital stay

There were 833 patients who had 910 cardiacarrests, 53 patients with more than one arrest. Inthis subset, 53% (28/53) survived to hospital dis-c

harge.

Cardiac arythmia 49 (5.9%)Infection 41 (4.9%)Cancer 41 (4.9%)CABG 30 (3.6%)Operation other than CABG 29 (3.5%)Stroke 26 (3.1%)PCI 23 (2.8%)Coronary angiogram 16 (1.9%)Cladicatio intermittens and

equivalents16 (1.9%)

Trauma (orthopedical) 14 (1.7%)Pulmonary embolism 12 (1.4%)Renal disease 12 (1.4%)Gastro-intestinal bleeding 11 (1.3%)Intoxication 8 (1.0%)Chest pain, other reasons 7 (0.8%)Cronical pulmonary disease 7 (0.8%)

In-hospital cardiac arrest—–An Utstein style 355

Table 4 Age, survival and rhythm for resuscitation attempts (N = 910)

Age N Survival to discharge (%) VT/VF (%) Other rhythms (%)

15—24 6 83 50 5025—34 11 22 73 2735—44 22 48 48 5245—54 84 55 65 3555—64 153 49 59 4165—74 262 41 55 4575—84 290 29 46 54>85 79 14 27 73

Length of hospital stay

The patients discharged alive had a median hospi-tal time of 2 days (range 0—54 days) prior to thearrest and 10 days (range 0—106) after the arrest.The day of the arrest is counted as day 0 after thearrest. The patients who did not survive to hospi-tal discharge had a median hospital time of 3 days(range 0—100) prior to the arrest.

Where were the patients discharged to?

Fifty-four percent were discharged home; 27% weredischarged to rehabilitation-facilities. The remain-ing 19% were discharged to other destinations.

Survival of patients with a witnessedcardiac arrest found in ventricularfibrillation of presumed cardiac aetiology

Of the total 833 patients with cardiac arrests, 204patients were a match for this subset. Fifty-ninepercent survived to discharge.

DNAR

In the period 1995—2001, there were 479,258patients admitted at the different wards ofSahlgrenska University Hospital. In the period7891 patients died during the hospital stay. Theresuscitation-team was alerted due to true cardiacarrest 910 times.

Discussion

General results — template

Ventricular fibrillation or tachycardia was recordedas the initial rhythm in almost 49% of the arrests.This is a high figure in comparison with otherreported VT/VF rates. Previous Utstein-reportshave found VT/VF to occur in between 8 and33%.2—8 We do not know why the rate VT/VF ishigher in Sahlgrenska, however, we can concludethat the intervals from collapse to defibrillation arecomparable with others.5 A possible reason can bea functional DNAR policy, and it is probable we havean even higher VT/VF rate since the patients defib-rillated once with rapid success in the CCU does notat

apcObforarivnh

lways require activation of a call for the resusci-ation team.

The overall survival rate for in-hospital cardiacrrest is 37%. This survival is very high in com-arison with other Utstein reports of in hospitalardiac arrest which range from 8 to 26%.2—5,7,8.ne of the reasons for this is most likely toe the high percentage of VT/VF, since patientsound in this rhythm have a more favourableutcome compared with patients found in otherhythms. There are also reasons to believe that

functional DNAR policy influences the survivalate.13 We have shown earlier that a selectionn whom to start resuscitation influences the sur-ival in out-of-hospital cardiac arrest,14 there iso reason why this also is not applicable in-ospital.

Time trends (Figure 2)

No clear trend is seen. We report an average sur-vival rate over these years of 37%.

Figure 2 Survival to discharge over time.

356 M. Fredriksson et al.

Outcome in terms of CPC score anddischarge location (Figure 1)

We observed that only a small proportion ofpatients had a higher CPC score (i.e. worse cere-bral outcome) after resuscitation from a cardiacarrest. This indicates that the vast majority of in-hospital cardiac arrest victims come out very well.The patients were discharged home or to rehabilita-tion facilities to a large extent, again showing thatin most patients the outcome is good. Other stud-ies have also shown good cerebral outcome after inhospital cardiac arrest. Skogvoll et al.8 found no dif-ference in CPC score at admission and at discharge,Huang et al.2 reported 75% of cardiac arrest caseshaving a good cerebral outcome.

Location of arrest

Forty-seven percent of the cardiac arrests wherethe emergency team was called for occurredwithin a ward with ECG monitoring facilities. Thisis in line with earlier published Utstein reports(47—71%).2,4,15 However, we have no informationhow many patients were actually ECG monitored at

Although DNAR strongly influences the survivalrate, this area is not well investigated. Only a fewstudies of in-hospital cardiac arrest report figuresfor DNAR, and some of them are quite old, so we donot really know the current situation. For examplethe large American based NRCPR study5 from 2003does not report how many resuscitation that werecommenced and the total number of patients whodied during the study period. This is unfortunatebecause such a large study could have shed somelight over the current situation.

We have shown earlier that a selection of thosein whom resuscitation should be started influencesthe survival from out-of-hospital cardiac arrest,14

i.e. there is a peak-value, where one have made theselection for maximum survival. It is possible thatthe same pattern is applies to in-hospital cardiacarrest.

Time intervals (Table 1)

The time between collapse and first defibrillationwas 2 min (median value). Peberdy et al.5 reportsa median value of less than 1 min. The associationbetween the interval from collapse to defibrillationapslfi

W(

Trmae

P(

Tpa

T

SNersp

the time of arrest.

Age (Table 4)

Increasing age has been proven to diminish thechances of survival in out of hospital cardiacarrest.16,17 We found that 70% of the resuscita-tions were started on patients 65 years or older.Resuscitation should not be considered futile juston the basis of old age, since 29% in the age group75—84 years were discharged alive. The patients’co-morbidity and status before the arrest must betaken into consideration. In the 25—34 years group,the survival was low, although the VT/VF rate washigh. This must be interpreted with caution, sincethe group only contains 11 individuals, too small anumber to draw any real conclusions from.

Do not attempt resuscitation (DNAR)

Our DNAR policy results in resuscitation attempts in10% of the patients who die during the hospital stay.Other reports a wide range from 6 to 53%.8,18,19

Skrifvars et al.20 reports resuscitation attemptsin 12% of all in-hospital cardiac arrests, our corre-sponding figure is 13%.

Niemann and Stratton13 points out that DNAR-decisions affect the survival rates, when theyexcluded the DNAR patients from their study thisled to a 15% increase in survival to hospital dis-charge.

nd survival seems to be strong also after in hos-ital cardiac arrest, although only reported frommaller series.6,7 A possible reason for our excel-ent survival is that our patients have VT/VF as therst rhythm to a greater degree.

hy were the patients hospitalised?Table 2)

he patients were hospitalized more due to cardiacelated diagnoses than in other reports.5 Again thisight be explained by the DNAR policy, excludinglarge population of patients with cancer disease,tc.

atients characteristics — co-morbidityTable 3)

he population health history is in line with otheratient groups with ischaemic heart disease, e.g.cute coronary syndrome.21

ime trends (Figure 2)

U implemented AEDs at strategic locations inovember 1996, including a massive educationffort on AED handling. Unfortunately this does noteflect itself in the survival statistics. No statisticalignificance is seen, only a clear drop in the survivaleak in 1999 and then recovering. We have no clear

In-hospital cardiac arrest—–An Utstein style 357

explanation for this outcome. We can only specu-late that, as in all major changes in structural work,new strategies will take some time before showingresults. It is also possible that year-to-year changesin rhythms affects survival rates on a year-to-yearbasis.

Limitations

We have studied the total duration of the hospitalstay, including the time at another hospital if thepatients were transferred to another hospital afterthe cardiac arrest. This may influence the resultson the percentage of patients discharged alive in anegative way, but we feel that this is a more appro-priate way of reporting.

Patients with arrest in the emergency room(ER) were not included, which also may influencesurvival. This subgroup has not been targetedas an endpoint in a study yet, so we cannotwith certainty say what influence this exclusionhas.

Conclusion

TsoawipAiVctw

Csms

jaac

A

TSL

References1. Cummins RO, Chamberlain D, Hazinski MF, et al. Rec-

ommended guidelines for reviewing, reporting, andconducting research on in-hospital resuscitation: thein-hospital ‘‘Utstein-style’’. Resuscitation 1997;34:151—83.

2. Huang CH, Chen WJ, Ma MH, Chang WT, Lai CL, Lee YT. Fac-tors influencing the outcomes after in-hospital resuscitationin Taiwan. Resuscitation 2002;53(3):265—70.

3. Granja C, Cabral G, Vieira A. Outcome of cardiac arrests ina Portuguese hospital—–evaluation of a hospital cardiopul-monar resuscitation program at one year. Rev Port Cardiol2001;20(10):943—56.

4. Patrick A, Rankin N. The in-hospital Utstein style: use inreporting outcome from cardiac arrest in Middlemore Hos-pital 1995—1996. Resuscitation 1998;36(2):91—4.

5. Peberdy MA, Kaye W, Ornato JP, et al. Cardiopulmonaryresuscitation of adults in the hospital: a report of14720 cardiac arrests from the National Registry of Car-diopulmonary Resuscitation. Resuscitation 2003;58(3):297—308.

6. Spearpoint KG, McLean CP, Zideman DA. Early defibrillationand the chain of survival in ’in-hospital’ adult cardiac arrest;minutes count. Resuscitation 2000;44(3):165—9.

7. Skrifvars MB, Rosenberg PH, Finne P, et al. Evalua-tion of the in-hospital Utstein template in cardiopul-monary resuscitation in secondary hospitals. Resuscitation2003;56(3):275—82.

8. Skogvoll E, Isern E, Sangolt GK, Gisvold SE. In-hospital car-

1

1

1

1

1

1

1

1

his single-centre study is by far the largest Utsteintyle report for cardiac arrest in hospital. Theverall survival rate for in-hospital cardiac arrestmong patients for whom the resuscitation teamas called was 37%. This rate of survival is very high

n comparison with other Utstein reports of in hos-ital cardiac arrest ranging from 8 to 26%.2—4,8,15

combination of factors can contribute to this,ncluding the high percentage of patients found inT/VF, a functional DNAR policy, an efficient andompetent resuscitation team, well-trained hospi-al staff in CPR and accessible defibrillators in allards.Only a small proportion of patients had a high

PC score (i.e. worse cerebral outcome) afteruccessful resuscitation, indicating that the vastajority of in-hospital cardiac arrest victims who

urvived had a good cerebral outcome.Resuscitation should not be considered futile

ust on basis of old age, since 29% survived theirrrest in the age group 75—84 and were dischargedlive. In 80% was CPR started within 1 min afterollapse.

cknowledgements

his study was made possible by grants from thewedish Heart and Lung Foundation and Goteborgsakaresallskap.

diopulmonary resuscitation. 5 years’ incidence and survivalaccording to the Utstein template. Acta Anaesthesiol Scand1999;43(2):177—84.

9. Herlitz J, Bang A, Aune S, Ekstrom L, Lundstrom G, HolmbergS. Characteristics and outcome among patients sufferingin-hospital cardiac arrest in monitored and non-monitoredareas. Resuscitation 2001;48(2):125—35.

0. Herlitz J, Rundqvist S, Bang A, et al. Is there a dif-ference between women and men in characteristics andoutcome after in hospital cardiac arrest? Resuscitation2001;49(1):15—23.

1. Herlitz J, Bang A, Alsen B, Aune S. Characteristics andoutcome among patients suffering from in hospital cardiacarrest in relation to the interval between collapse and startof CPR. Resuscitation 2002;53(1):21—7.

2. Herlitz J, Bang A, Alsen B, Aune S. Characteristics andoutcome among patients suffering from in hospital cardiacarrest in relation to whether the arrest took place duringoffice hours. Resuscitation 2002;53(2):127—33.

3. Niemann JT, Stratton SJ. The Utstein template and the effectof in-hospital decisions: the impact of do-not-attempt resus-citation status on survival to discharge statistics. Resuscita-tion 2001;51(3):233—7.

4. Fredriksson M, Herlitz J, Engdahl J. Nineteen years’ expe-rience of out-of-hospital cardiac arrest in Gothenburg—–reported in Utstein style. Resuscitation 2003;58(1):37—47.

5. Skrifvars MB, Castren M, Kurola J, Rosenberg PH. In-hospitalcardiopulmonary resuscitation: organization, managementand training in hospitals of different levels of care. ActaAnaesthesiol Scand 2002;46(4):458—63.

6. Herlitz J, Eek M, Engdahl J, Holmberg M, Holmberg S. Factorsat resuscitation and outcome among patients suffering fromout of hospital cardiac arrest in relation to age. Resuscitation2003;58(3):309—17.

7. Bunch TJWC, Packer DL, Panutich MS, White RD. Admis-sion predictors of in-hospital mortality and subsequent long-term outcome in survivors of ventricular fibrillation out-of-

358 M. Fredriksson et al.

hospital cardiac arrest: a population-based study. Cardiology2004;102(1):41—7.

18. Hodgetts TJ, Kenward G, Vlackonikolis I, et al. Inci-dence, location and reasons for avoidable in-hospital car-diac arrest in a district general hospital. Resuscitation2002;54(2):115—23.

19. Schultz SC, Cullinane DC, Pasquale MD, Magnant C, EvansSR. Predicting in-hospital mortality during cardiopulmonaryresuscitation. Resuscitation 1996;33(1):13—7.

20. Skrifvars MB, Hilden HM, Finne P, Rosenberg PH, CastrenM. Prevalence of ’do not attempt resuscitation’ orders andliving wills among patients suffering cardiac arrest in foursecondary hospitals. Resuscitation 2003;58(1):65—71.

21. Steg PG, Goldberg RJ, Gore JM, et al. Baseline character-istics, management practices, and in-hospital outcomes ofpatients hospitalized with acute coronary syndromes in theGlobal Registry of Acute Coronary Events (GRACE). Am J Car-diol 2002;90(4):358—63.