Influence on outcome after cardiac arrest of time elapsed between call for help and start of bystander basic CPR

Resuscitation, 25 (1993) 227-234 0 1993 Elsevier Scientific Publishers Ireland Ltd. All rights reserved. 0300-9572/93/$06.00

227

Influence on outcome after cardiac arrest of time elapsed between call for help and start of bystander basic CPR P.R. Martensa, A. Mulliea, P. Calleb, R. Van Hoeyweghena and The Belgian Cerebral Resuscitation Study Group* “Department of Anaesthesia and Critical Care, Algemeen Ziekenhuis, Sint Jan Hospital, Ruadershove IO, 8000 Brugge and bDepartment of Emergency Medicine, Universitair Ziekenhuis. Gent (Belgium)

(Received August 12th, 1992; Accepted December 12th, 1992)

The exact impact of the ‘interval between cardiac arrest (CA) and the start of basic cardiopulmonary resuscitation (CPR) performed by bystanders’ on outcome is not fully established. We retrospectively evaluated data with regard to response intervals of 1195 out-of-hospital CA interventions where bystander CPR was performed and continued by the eight mobile intensive care units (MICUs) participating in the Belgian Cerebral Resuscitation Registry between 1982 and 1990. Partial correlations between time elapsed from CALL to CPR by lay public and outcome were determined when the effect of response times of 1st and 2nd tier were removed. The following groups were studied: ventricular tibrillation (VF), asystole and electromechanical dissociation (EMD), non-witnessed and witnessed. Good outcome was represented by initial restoration of spontaneous circulation (ROSC successes) and by prolonged survival (CPR successes) being 22.7 and 9.7%, respectively. The mean time f S.E.M. between CALL and CPR initiated by lay people for the studied population (n = 1195) was 2.5 f 0.1 min. The partial correlation coefftcient between prolonged survival and time passed between CALL and bystander CPR was negative for all types of CA, yet significance was reached only in the non-witnessed group. Using ROSC as the endpoint significance is achieved in all groups except the VF patients, where the intervention times were shorter. In our population, prolonged survival was independently and negatively influenced by a delay between CALL and any CPR in the non-witnessed CA group (n = 421). Key words: basic cardiopulmonary resuscitation (CPR); bystander; outcome; partial correlation

Correspondence to: P. Martens, Department of Anaesthesia and Critical Care, Algemeen Ziekenhuis, Sint Jan Hospital, Ruddershove 10, 8000 Brugge Belgium. *Members of The Cerebral Resuscitation Study Group of the Belgian Society for Intensive Care & Emergency Medicine: A. Mullie4 P. Martens’, W. Buylaertb, P. Calleb. H. Houbrechtsb, L. Con&. D. Lauwaert’, H. Deloozd, H. Verbuggend, R. De Cock’, M. Weeghma&. J. Menned, L. Bossaerts, R. Van Hoeyweghens, P. Lewih and A. Quetsh. aDepartment o f Cri tical ’ Care Medicine, Algemeen Ziekenhuis

Sint Jan, Brugge, bDepartment of Emergency Medicine, Universitair Ziekenhuis, Gent, ‘Department of Emergency Medicine, Akademisch Ziekenhuis, Vrije Universiteit, Brussel, dDepartment of Emergency Medicine, Universitair Ziekenhuis, Sint Rafael-Gasthuisberg, Leuven, Y)epartment of Emergency Medicine, Imeldaxiekenhuis, Bonheiden, ‘Department of Emergency Medicine, Jan Palfijn Ziekenhuis, Merksem, s Department of Intensive Care, UIA, Univertitair Ziekenhuis, Antwerpen and hDepartment of Information Sciences, Janssen Research Foundation, Beerse.

In previous reports variables determining survival in the sudden death syndrome have been outlined by the Belgian Cerebral Resuscitation Study Group’. It is obvious from these and other reports that both time to basic life support (BLS) either performed by a citizen or an ambulance man (1st tier) and advanced life support (ALS), including defibrillation performed by mobile intensive care units (MICUs) (2nd tier) should be as short as possible*. Some authors suggest that bystander BLS improves outcome especially when ALS is delayed, as seems to be the case in Belgium3. Other authors, however, also reported improved survival rates when bystander BLS was initiated on victims of cardiac arrest (CA) in an emergency medical service (EMS) system with short response times4*‘. On the other hand, bystander BLS has been identified as a variable predicting poor prognosi#. Furthermore, Stueven found no significant difference in the survival rate between CA patients that were witnessed and had bystander BLS, and other victims that were witnessed by a bystander who did not perform BLS until the arrival of the ALS system7. Likewise, bystander BLS was reported not to be a prognostic factor for long-term survival after CA by Clinton and Peterson’. The aim of our study was to determine the exact importance of the interval between CA and any CPR (duration of no-flow when CA is confirmed) or interval to any treatment performed by lay people who are not part of an organized emergency response system. MATERIALS AND METHODS

We retrospectively evaluated data of 1195 out-of-hospital cardiac arrest interventions where BLS was performed by bystanders and followed by ALS performed by the eight MICUs, participating in the Belgian Cerebral Resuscitation Registry between 1982 and 1990. The overall profile of these CA patients, the EMS organisation, as well as the definitions used to describe CPR events before the arrival of the MICU and the overall outcome results have been published previously9310. The following sequence of time intervals (tiered CPR) has been registered: duration of complete CA (CA-ANY CPR), separated into EMS activation time (CA-CALL) and response time of bystander BLS (CALL-ANY CPR), duration of bystander BLS to CPR by ambulancemen (ANY-AMBU CPR), duration of AMBU-CPR to ALS (AMBU-MICU) and duration of ALS. In this study only patients who received bystander CPR and where the interval between any CPR and CPR by ambulance personnel was available and did not equal zero were selected. Bystanders who initiated BLS were identified as either genuine lay people or health care workers not belonging to the EMS. Time of onset or recognition of CA and time of start of BLS by lay public were retrospectively estimated by interviewing family members and witnesses. Time of call, defined as the arrival of the call in the EMS dispatching centers, was more accurately determined from radiophonic recordings in the dispatching centers. Arrival times of the 1st and 2nd tier (MICU) at the scene (vehicle stops) were also recorded chronologically from the EMS dispatching tape recorders (Fig. 1). Despite the im-

229

COLLAPSE

/ RECOGNITION I

TIME 0 : 00

1'

CALL

LAY PUBLIC OR BYSTANDER CPR

- BLS

CPR BY AMBULANCE OR EMS PERSONNEL

MICU TIME

X

: XX

CPR

2nd TIER

SERVICE 1st TIER

AT ARRIVAL - ALS

1

Fig. 1. Events associated with out-of-hospital cardiac arrest resuscitation attempts. Interval I: collapse to CALL, Interval 2: CALL to BYSTANDER CPR, Interval 3: CALL to arrival 1st tier (ambulancemen), Interval 4: CALL to arrival 2nd tier (MICU).

portance of the exact recording of the time of collapse to understand the ischaemic interval, imprecision surrounds the estimated time of collapse. Therefore correlation coefficients (RXY) have been tabulated with outcome as a function of the interval between CALL and BLS by lay public, and the interval between CALL and arrival of 1st and 2nd tier. Because intervention factors such as the interval between CALL and BLS by lay public (interval 2), the interval between CALL and BLS by ambulance service (interval 3), and the interval between CALL and arrival of MICU (interval 4) are not only interrelated but also affect the outcome (inherent confounding variables), the effect of time elapsed between CALL and BLS by lay public on outcome has been studied using the partial correlation technique (Fig. 1). These correlation coefficients have been calculated for all patients and for the following subgroups: ventricular librillation (VF) on arrival of the MICU and asystole + electromechanical dissociation (EMD) on arrival of the MICU;.non-witnessed and witnessed events. The formula for the partial correlation coefficient between variables 1 and 2 computed after eliminating the effects of the other variables, 3 and 4 is as follows”: R12.4 - h3.4R23.4 R,2.34 = 'b

- R:3.4)U - R223.4)

A level of significance calculated by Student’s t-test is reached when the P-value is less than 0.05. The actual intervals are expressed as means together with the standard error of the mean (S.E.M.) for all patients and as means for the subgroups VF, asystole + EMD, non-witnessed and witnessed (Table I). In an attempt to address the question of adequate neurologic outcome, survival

Table I. ’Mean value of intervals 2,3 and 4 with standard error of the mean (min) for all studied patients and for the subgroups VF, asystole (A) + EMD, non-witnessed (NW) and witnessed (W). The 1195 outof-hospital cardiac arrest patients where bystander CPR had been performed. CPCR registry 1982-1990. Variable

Mean

SE.

VF

A+EMD

NW

W

CALL-BYSTANDER CPR CALL-AMBU CALL-MICU

2.5 10.6 17.6

0.1 0.5 1

1.8 8.1 14.3

2.8 11.6 19.3

3.3 11.8 18.7

2 10 17

is presented in two end-points: firstly prolonged survival (1’) including only CPR successes, regardless of their functional state: normal (good recovery), moderately disabled (disabled but independent) or severely disabled (dependent on others) (Table II) and secondly, restoration of spontaneous circulation (ROSC = 1”) as shown in Table III, including all initially successfully resuscitated patients, both CPR successes (alive and awake at day 14 after CA) and CPR failures (dead or not awake at day 14). RESULTS

The mean interval f S.E.M. between CALL and basic CPR by lay public (all patients, n = 1195) was 2.5 * 0.100 min with a median of 1 min. Intervals from CALL to arrival of 1st and 2nd tier were, respectively 10.6 and 17.6 min. with medians of 9 and 13 min. For the subgroup of VF patients the means were 1.8, 8.7 and 14.3. (Table I, 3rd column) 39% of the non-EMS bystanders were genuine lay bystanders whereas 61% were identified as non-EMS health care workers. Of the total non-EMS bystander population, 22.7% (n = 271) achieved ROSC and 9.7% (n = 116) achieved longterm survival. First ECG showed VF in 34%, and 63.4% of the cases were witnessed.

Table II. Correlations between prolonged survival (CPR successes = 1’) and intervention times for all patients and the subgroups VF, asystole (A) + EMD, non-witnessed (NW) and witnessed (W) patients. 1‘, Prolonged survival or CPR successess; 2, interval 2; 3, interval 3; 4, interval 4. See text and figure for abbreviations.

RI,, R1,3 RI ‘4 R23 R24 R34

RI ‘2.34

P-value

All (n = 1195)

VF (n = 407)

A+EMD (n = 788)

:

-0.10 -0.13 -0.16 0.49 0.33 0.67

-0.10 -0.13 -0.16 0.40 0.29 0.64

-0.07 -0.05 -0.09 0.53 0.34 0.68

-0.17 -0.13 -0.15 0.57 0.36 0.68

-0.06 -0.11 -0.16 0.44 0.30 0.65

-0.05 NS

-0.05 NS

-0.06 NS

-0.12 co.05

-0.009 NS

421)

: = 755)

231 Table III. Correlations between ROSC (initial restoration of spontaneous circulation = 1”) and intervention times for all patients and the subgroups VF, asystole (A) + EMD, non-witnessed (NW) and witnessed (W) patients. lx, ROSC; 2, interval 2; 3, interval 3; 4, interval 4. See text and figure for abbreviations. All (n = 1195)

F=

h-4

-0.17 -0.18 -0.18

R1~2.34

-0.09

P-value


R1.2 JC.3

A+EMD (n = 788)

NW (n = 421)

W (n = 755)

-0.14 -0.16 -0.17

-0.15 -0.13 -0.13

-0.19 -0.19 -0.14

-0.13 -0.15 -0.19

-0.08 NS

-0.10 co.01

-0.10 < 0.05

-0.07 <0.05

407)

There is a direct correlation between outcome and intervals (Rt ,2, R,p3and R,,J, but this correlation is, as we expected, negative in all categories (Table II and III). These numbers become increasingly negative for the 1st (ambulance personnel) and 2nd tier (MICU), respectively with the exception of the asystole + EMD and nonwitnessed groups. The partial correlation coefficient between prolonged survival and time passed between CALL and bystander CPR appears to be negative for all types of CA. Significance is only reached in the non-witnessed group whereas the absolute value is extremely small in the witnessed group (R112.34 in Table II). Considering ROSC, this partial correlation coefftcient turns out to be significantly negative for all patients and also for the subgroups with the exception of the VF patients (Rln2.M in Table III). The absolute values of the partial correlation coefficients are always smaller than R12. On the contrary, we found a positive correlation between the arrest interval (CALL to bystander basic CPR) and bystander basic CPR interval as reflected in interval from CALL to basic CPR by the ambulance service (R&. This correlation is less positive in the VF and witnessed groups, because initiation of bystander basic CPR may coincide with or even precede the CALL, so the interval from CALL to CPR by bystanders (genuine lay + non EMS health care workers) may become zero. Because intervals 2 and 3 are included in interval 4 (Fig. l), the correlations Rz4and Rj4are obviously positive. In this respect R34almost doubles Rz4.These positive correlations are most pronounced in the asystole + EMD and non-witnessed groups. DISCUSSION

In our population good outcome was slightly higher than the 7% reported for all out-of-hospital CAs by the CPCR group lo. In addition, although the mean response intervals of 1st and 2nd tier are not shorter than in the total out-of-hospital population, the percentages of patients in VF (34%) and of witnessed cases (63%) are higher than the reported 25 and 53%, respectively lo. Therefore bystander basic CPR might produce its beneficial effect on survival by maintaining the heart in VF’*. Considering longterm survival R 112.34 is only reaching the level of significance

232

in the non-witnessed group (Table II). The fact that the most important variables being intervals between arrival of 1st and 2nd tier are relatively long in the total population and particularly in the non-witnessed subgroup may explain why the effect of a short interval to any basic CPR is partially ‘ironed out’ when considering the correlation independently of these variables: RIj2,34 = -0.12vs. RI'2= -0.17 (non witnessed in Table II). These results are clearly in contradiction to the hypothesis that unwitnessed CAs are unlikely to benefit from any intervention and this was already observed before by Bossaert et al. l2. Considering ROSC the partial correlation coefficients become more negative with the exception of the nonwitnessed group and R1.2,34 remains non significant in the VF group where the mean time to any CPR is only 1.8 min (Table I). The outcome in the VF group where intervention times are shorter is probably determined far more by early defibrillation than by public CPR. This stresses the recommendation of activation of the EMS system as soon as possible before anything else. A more negative and signiticant partial correlation coefficient for the asystole + EMD group when considering ROSC (Table III) rather than CPR success (Table II) does not support the assumption of Thompson et al. who stated that bystander intervention had less effect on the immediate results of resuscitation than upon the mortality during hospitalizationi3. We anticipated that in order to detect a positive impact of a short interval between CALL and basic CPR by lay public on outcome, the interval between any basic CPR to basic CPR by the ambulance service should be long enough. This is obviously the case in our population (Table I). In contrast, a short arrest interval (CALL-ANY basic CPR interval) in combination with a short bystander basic CPR interval can be considered as a reflection of a sub-group that responds to rapid resuscitation and, consequently manifests a higher likelihood of survival anyway. When comparing Pvalues between all types of CA and the subgroups: VF, asystole + EMD, nonwitnessed and witnessed, one should keep in mind that the number of patients in the subgroups are considerably (30-500/) smaller. Positive associations between intervention intervals, particularly in the asystole + EMD and non-witnessed groups probably reflect a longer time to get from the rescue unit to the patient’s side. Therefore in cases of long bystander basic CPR it might have taken longer for the ambulancemen and the MICU-team to access these victims due to distance. Methodologically the study of the influence of bystander response time on outcome is difficult and frustrating for the following reasons. Firstly, a controlled prospective clinical study with randomization is impossible. Secondly, the value of the interval estimation between the unmoitored collapse and the onset of bystander CPR is debatable because one has to rely on persons often family of the victim, who are in an exceptionally emotional situation. Moreover this variable, being the shortest of all intervals within the CPR-time frame, is less likely to influence the outcome significantly. Thirdly, collapse does not always imply no flow. Indeed, ventricular tachycardia with limited blood flow and loss of consciousness, may precede VF. Fourthly, factors such as elevators, stairs, closed doors or other barriers are not evident in our response time data, since only times of arrival at the scene were recorded and not the time to the patient’s side. Therefore, underestimations of real times of onset of CPR by 1st and 2nd tier may have occurred. Fifthly, the heterogenity of ‘lay-people’ should also be taken into account when interpreting the results12. Indeed, medically trained bystanders such as physicians, nurses and paramedics, being

233

the case in the majority (61%) of our bystander CPR events, are according to the Utstein Style by definition considered as lay14. We tried to adhere to the Utstein recommendations as much as possible within the limitations of a retrospective analysis. Sixthly, possibly a small number of patients who had resuscitation attempts by lay rescueres and who were observed to have a pulse upon arrival of the emergency personnel may be false-positive cardiac arrests, as they might have been only respiratory arrests. Closed-chest basic CPR evolved about 30 years ago and was almost universally implemented not only by professional providers but also by lay bystanders because of its technical simplicity and noninvasiveness. In an animal model it has been shown that during ALS (i.e. after epinephrine administration) there are no statistical differences in myocardial blood flow and cerebral blood flow between animals pretreated with bystander basic CPR and others that received no-bystander basic CPR15. Sporadically lay basic CPR has been reported as an act not without hazard and patients with a history of lay basic CPR should be suspected of having damage to intrathoracic and - abdominal structures’6,17. In spite of numerous studies evaluating parameters associated with outcome, little attention has been paid to the impact of the amount of on-scene time spent by nonEMS personnel on the outcome of non-trauma-related cardiac arrest. A retrospective screening of 19 patients admitted to a coronary care unit after prehospital collapse and CPR showed that patients whose resuscitation was initiated by bystanders within 5 min fared much better than those whose resuscitation was delayed beyond 5 min pending the arrival of a trained rescue team18. Particularly in the one rescuer situation the bystander gets into the dilemma as to whether to initiate CPR and as a consequence delay early defibrillation or to call for professional help first, the latter being more likely when the level of the attendant person’s complete CPR training is poor. The interval between call and lay-CPR in our population is still short indicating that people are not queueing for more than 15 s to be answered, as appears to be the case in big centers. CONCLUSION

Our data do not support the hypothesis that outcome is better in persons suffering from ventricular fibrillation who receive early bystander CPR (within a mean of 1.8 min. after calling) than in patients with witnessed VF who receive late bystander CPR. In this particular group the effect of early calling seems to have its importance by primarily resulting in early ambulance CPR and ALS (Table I, 3rd column). Bystander basic CPR might increase the number of patients subsequently found in VF by the MICU. Prolonged survival seems to be negatively influenced by a delay between call and any basic CPR in the non-witnessed cardiac arrest (Table II, 4th column). REFERENCES 1

Lewi PJ, Mullie A, Quets A, The Cerebral Resuscitation Study Group. Relevance and significance of pre-CPR conditions in Cardio-Pulmonary-Cerebral resuscitation. A graphic analysis by means of Spectramap. Resuscitation 1989; 17 Suppl: S35-S44.

234 2 3 4

5 6 7 8 9 10 11 12

13 14

15 16 17 18

Eisenberg MS, Horwood BT, Cummins RO, Reynolds-Haertle R, Hearne TR. Cardiac arrest and resuscitation: a tale of 29 cities. Ann Emerg Med 1990, 19: 179-186. Mullie A, Van Hoeyweghen R, Quets A, The Cerebral Resuscitation Study Group. Influence of time intervals on outcome of CPR. Resuscitation 1989; 17 Suppl: S23-S33. Spa&e DW, Hanlon T, Criss EA, Valenxuela TD et al. Prehospital cardiac arrest: the impact of witnessed collapse and bystander CPR in a metropolitan EMS system with short response times. Ann Emerg Med 1990; 19: 1264-1269. Tarp-Pederson C, Birk-Madsen E, Pederson A. The time factor in resuscitation initiated by ambulance drivers. Eur Heart J 1989; 10: 555-557. Kentsch M, Stendel M, Berkel H, Mueller-Esch G. Early prediction of prognosis in out-of-hospital cardiac arrest. Intensive Care Med 1990; 16: 378-383. Stueven HA. Prehospital CPR: a review, in perspective. Resuscitation 1989; 17 Suppl: S71-S77. Clinton JE, Peterson G. Nontraumatic prehospital cardiac arrest ages 1 to 39 years. Am J Emerg Med 1990; 8: 87-91. The Cerebral Resuscitation Study Group, The Belgian Cardio Pulmonary-Cerebral resuscitation registry. Form protocol. Resuscitation 1989; 17 Suppl: S5-SlO. Mullie A, Lewi P, Van Hoeyweghen R, The Cerebral Resuscitation Study Group. Pre-CPR conditions and the final outcome of CPR. Resuscitation 1989; 17 Suppl: Sll-S21. Snedecor GW, Cochran WG. Statistical methods, 7th edn. Ames, Iowa: Iowa State University Press, 1980. Bossaert L, Van Hoeyweghen R, The Cerebral Resuscitation Study Group. Bystander cardiopulmonary resuscitation (CPR) in out-of-hospital cardiac arrest. Resuscitation 1989; 17 Suppl: S55-S69. Thompson RG, Hallstrom AP, Cobb LA. Bystander-initiated cardiopulmonary resuscitation in the management of ventricular fibrillation. Ann Internal Med 1979; 90: 737-740. Cummins RO, Chamberlain DA, Abramson NS, Allen M, Baskett PJ et al. Recommended guidelines for uniform reporting of data from out-of-hospital cardiac arrest: the Utstein style, Circulation 1991; 84: 960-975. Hoekstra JW, Rinnert K, Van Ligten P, Neumar R, Werman HA, Brown CG. The effectiveness of bystander CPR in an animal model. Ann Emerg Med 1990; 19: 881-886. Reardon MJ, Gross DM, Vallone AM, Weiland AP, Walker WE. Atrial rupture in a child from cardiac massage by his parent. Ann Thorac Surg 1987; 43: 557-558. Bedell SE, Fulton EJ. Unexpected findings and complications at autopsy after cardiopulmonary resuscitation (CPR). Arch Intern Med 1986; 146: 1725-1728. Copley DP, Mantle JA, Rogers WJ, Russell RO, Rackley CE. Improved outcome for prehospital cardiopulmonary collapse with resuscitation by bystanders. Circulation 1977; 56: 901-905.