The basic techniques of cardiopulmonary resuscitation (CRP) have been established since 1960 when Kouwenhoven, Jude and Knickerbocker described closed-chest cardiac massage¹. Most hospitals worldwide now have cardiac arrest teams and there have been many studies which have

reported on outcome^2,3, factors which affect outcome^4,5, and whether one could predict outcome⁶. The aim of this study was to define the patient population in this hospital and, as an exercise in quality assurance, to see how our success rate compared with other studies. It was felt this audit would also serve to evaluate a 'Do Not Resuscitate' (DNR) policy. This audit was also used as a preliminary study to develop a scoring system to predict discharge from hospital following successful CPR.

The CPR team, consisting of medical, nursing and technical staff, has access to cardiac arrest trolleys located on each floor, outpatient departments, critical care areas and emergency room. Following a call-out, events are documented on a special form. All forms dating from 1.1.82 to 31.12.86 were analyzed.

Cardiopulmonary arrest was defined as the failure or cessation of circulation or respiration, resulting in a documented loss of consciousness and requiring initiation of CPR⁶. Any arrest not complying with this definition was excluded. Resuscitation was deemed successful if a stable circulation was established and the resuscitation team disbanded³. Patients were admitted to the Intensive Care Unit (ICU) or Coronary Care Unit (CCU) following successful CPR. If any patient required more than one episode of CPR during one stay in hospital, only the initial effort was examined. A subsequent arrest, if unsuccessfully treated, was recorded as a later death.

Patients were divided into 22 different disease categories. These categories were not mutually exclusive, i.e. a patient could be included in two or more different categories. We also recorded whether a doctor was present at the time of the arrest, whether the patient required intubation and the response time of the doctors. Finally patients were divided into seven mutually exclusive categories relating to the initial ECG findings.

The number of cardiac arrests per annum varied from 49 to 86, increasing each year as the work load of the hospital increased. Of the 472 forms which were returned for analysis 115 were deemed not to have had an arrest as per definition, 4 forms were inadequately filled in and 6 patients had a second cardiac arrest, leaving 347 forms to be analyzed. Table I summarizes the outcome of CPR in these patients.

347 CPR patients

119 die in hospital 61 leave hospital

(34% of total, 66% (18% of total, 34%

of survivors) of survivors)

(17% of survivors,

8.6% of total)

Diagnosis Total % Survive % Leave L/T% L/S%

Total arrests 347 100 180 52 61 15 34

Cardiac infarct 67 19 33 49 15 22 45

Cardiac valves/Rh.D. 44 13 25 57 11 25 44

Cardiac myopathy 13 4 7 54 6 46 86

Thrombo embolic 16 5 8 50 1 6 13

CCF-Pulm. edema 47 14 21 45 11 23 52

Renal failure 38 11 22 58 5 13 23

Stroke/CVA 19 5 8 42 1 5 13

Neoplasia-blood 9 3 3 33 1 11 33

Neoplasia GIT upper 9 3 5 56 0 0 0

Neoplasia GIT lower 2 1 0 0 0 0 0

Neoplasia metastases 4 1 2 50 0 0 0

Chest tumour 4 1 1 25 0 0 0

Neoplasia total 28 8 11 39 1 4 9

Infection (chest) 19 5 9 47 4 21 44

C.O.A.D. 14 4 10 71 4 29 40

Resp. failure 8 2 5 62 2 25 40

Diab. mellitus 39 11 21 54 5 13 24

Liver failure 14 4 7 50 0 0 0

Trauma 15 4 5 33 2 t3 40

Head injury 7 2 6 86 1 14 17

Septicemia 20 6 11 55 1 5 9

Post-op 46 13 28 61 8 17 29

Other diagnosis 78 22 40 51 11 14 28

L/T% percentage of total number of patients who left hospital.

L/S% percentage of initial survivors who left hospital.

Total % Survive % Leave L/T% L/S %

Total Arrests 347 100 180 52 61 18 34

Doctor Present140 40 68 49 20 14 29

Intubation 284 82 120 42 29 10 24

Total % Survive % Leave L/T% L/S%

Asystole 158 46 63 40 14 9 22

VF 50 14 31 62 16 32 52

VT 15 4 11 73 6 40 55

AF 4 1 3 75 2 50 67

Bradycardia71 20 42 59 12 17 29

SR 23 7 16 70 7 30 44

Other 26 7 14 54 4 15 29

Resuscitation was successful in 180 patients (52%). From these, 30 (17%) died within 24 hours, and 89 died later. Sixty one patients left hospital, representing 18% of the total or 34% of the initial survivors.

Table 2 shows the patients divided into 22 disease categories, with the number of initial survivors and the number who left hospital in each category. The only group, which was mutually exclusive, was the neoplasia group, where a patient with metastases would not be included in another neoplasia subgroup.

Table 3 shows the figures for the presence of a doctor and the need for tracheal intubation in all cases analyzed.

In 36 cases, the response time of doctors was not recorded. Out of the remaining 311 forms analyzed a doctor was present in 130 cases. In the others, the mean response time was 2.3 minutes. There were 29 cases with a response time of over 5 minutes. Of these, 22 were between midnight and 7 am. A total of 49 patients were resuscitated between midnight and 7 am, and of these, 11 were resuscitated successfully; out of these 11, 3 had a response time of more than 5 minutes.

Eighteen per cent of the patients who underwent CPR survived to be discharged from hospital. This result is similar to those obtained in other series where the success rate has ranged from 2.1-21 %^2-9 . The results shown in table 1 are very similar to those reported by Bedell et al⁶ where 44% of

patients immediately survived and of these 24% died within 24 hours and 32% left hospital.

An interesting feature of the present series was a lower incidence of myocardial infarction than reported elsewhere^3-6. This may possibly be due to patients having less coronary artery disease in Saudi Arabia than in a Western society, although a recent report from Dammam showed that 40% of their cardiac arrests were due to cardiovascular disease⁹. However, if the patients with congestive cardiac failure, rheumatic heart disease and cardiomyopathies who may have had a myocardial infarct as the predisposing cause of their arrest were included then the incidence in this study would be similar to other reports.

In keeping with other studies^4,6, the patients with thromboembolic disease, renal failure, stroke, neoplasia, head injury and septicemia did badly. Surprisingly, and to our knowledge not previously reported, no patients with liver failure who arrested left hospital. Contrary to the experience of Bedell et al⁶, who reported 100% mortality in 58 patients with pneumonia, 21% of the 19 pneumonia patients left hospital. Sowden et al⁴ reported that 50% of patients with pneumonia who arrested left hospital so the 100% mortality reported by Bedell et al⁶ remains unexplained.

The presence of a doctor at the time of the arrest was not associated with any significant change in mortality while, as expected, the need for intubation was associated with an increased mortality. In view of the small numbers involved no statistical significance could be derived for any of the nonintubated subgroups.

Patients who developed ventricular fibrillation or tachycardia were more likely to survive than patients who developed asvstole. The 9% incidence of discharged patients in the asystole group is similar to the 6% reported by Stiles et al¹⁰. Patients with severe bradycardia made up 20% of the total initial ECG findings. No other study reports a similar occurrence. It may be argued that these patients in fact did not arrest. However, in all cases there was a documented loss of consciousness and all required CPR. Furthermore, there is no significant difference in success rate from the remainder of the study population.

Our response time of 84% within 5 minutes compares favorably with the 89% reported by Bedell et al⁶. We should point out that our figure does not include cases where the doctor was present at the arrest.

During the 5 years of this study 1728 patients died in the hospital. Therefore there appears to be a cardiac call put out in only 20% of cases. This is lower than the 30% reported by Bedell et al⁶ and may be related to different legal pressures in the two countries. Of the 135 patients with neoplasia who died during the 5 years of the study, CPR was undertaken on 21%. However, only one of these patients eventually left hospital. It appears therefore that the DNR orders are proving effective. However, reappraisal of these orders, especially in certain conditions such as liver failure and neoplasia, could alleviate further unnecessary prolongation of death.

The importance of the site of the arrest on outcome is unclear. Some papers quote the incidence of low survival rates in cardiac arrests on general wards³. Others, however, quote low survival rates in the Intensive Care Unit¹¹. Others do not refer to the site of resuscitation at all^4,5,6. Our aim was to review our system as a whole. We did not feel, therefore, that the site of resuscitation was relevant to our study.

The majorltv of patients will die despite what appears to be a successful CPR. This figure varies from 54%⁴ to 83%⁹. Similarly, a large proportion die within 24 hours of the arrest. This study shows an incidence of 17% while others show an incidence of 54%³ and 24%⁶. The lower figures reported here may reflect the policy that all successfully resuscitated patients are admitted to the ICU or CCU and are therefore being kept alive for longer than if they had been left on the ward.

The question that needs to be asked is whether, at an early stage, patients can be predicted to survive or not. Quite clearly a patient who, very soon following resuscitation, is awake and maintaining his own blood pressure and ventilation should have a better prognosis than another patient who is being ventilated, requiring cardiac support and who is unconscious. If one could predict mortality following initial successful CPR, it would help avoid unnecessary treatment and prolonged suffering in hopeless cases. It would allow a dignified death and also a rational allocation of resources. Such criteria have been derived and have been submitted for publication.

There were 347 cardiac arrests analyzed over a 5 year period. 180 patients were successfully resuscitated and 61 eventually left hospital. Patients with thromboembolic disease, renal failure, stroke, neoplasia, head injury and septicemia did badly. No patient with liver failure, who arrested left hospital. The need for intubation at the arrest was associated with an increased mortality. Patients who developed ventricular fibrillation or tachycardia were more likely to survive than patients who developed systole. Our discharge outcome, of 18% compares favorably, with all previous studies.