WHO Multicenter, Multi-factor Coronary Rehabilitation Trial

Background:

The application of principles of rehabilitation to coronary disease – especially myocardial infarction – already applied to other physical disorders – was first considered in the 1950s. In 1964 the Report of a WHO Expert Committee on Rehabilitation of Patients with Cardiovascular Diseases (1) reviewed the situation and stressed that this field required further research, in particular with regard to positive and negative effects of physical rehabilitation. This idea was further developed in 1967 in the Report of a Seminar on Rehabilitation of Patients with Cardiovascular Diseases (2), arranged in Noorwijk aan Zee, Netherlands, by the WHO Regional Office for Europe. The report suggested that efforts should be made to promote coordinated studies in different parts of Europe, to evaluate factors that may affect rehabilitation, and to measure the impact of proposed programmes on morbidity and mortality after myocardial infarction. Addressing this problem, a study project based on coordinated studies was established in Copenhagen in 1970 (3). The WHO Regional Office for Europe agreed to act as a coordinator of the study.

Design and organization of the study:

The common protocol provided guidelines for standard procedures, quality control, central monitoring, and processing and analysis of data through the WHO Regional Office for Europe, with policy direction by a working group of the principal investigators. The enrolment of patients in the local studies began in June 1973. Only men below 65 years were enrolled Patients were randomized at admission to the study (at discharge from hospital) in two groups – a rehabilitation group (R) and a control group (C). As individual randomization was not feasible or acceptable to some centres, alternative procedures were allowed – randomization of areas or hospitals to R and C. The comprehensive intervention in R patients comprised the following elements: treatment of heart failure, arterial hypertension, dysrhythmias, diabetes mellitus etc; stopping smoking, reducing alcohol consumption; normalizing or lowering of serum lipid levels; losing weight; and improving physical working capacity. The planned duration of the study was three years and the patients were re-examined at 12, 24 and 36 months. All data at initial examination and at follow-up examination and re-examinations were entered into uniform record forms and mailed to the WHO in Geneva. A special death record was used to report on patients who died during the study.

It was calculated originally that, to permit detection of 25 % reduction in subsequent reinfarction and death at the 0.05 level of statistical significance, approximately 6800 patient-years of observation would be needed. It was estimated that 2200 patients followed up for three years would be adequate to provide the desired answer. Allowance was made for a dropout rate of 12 – 15%. Originally 24 centers volunteered to participate in the study – 23 from Europe and one from Israel. The recruitment period was protracted and ended in October 1975 when the preset number of 3000 patients was attained

Results:

For various reasons, 17 centers from 12 countries took part in the final evaluation of the results of the study: Belgium (2 centers), Czechoslovakia. Finland (2 centers), German Democratic Republic (3 centers), Hungary (2 centers), Israel, Italy, Lithuania, Poland, Roumania, Soviet Union, and Spain. Thus, the final number of patients in the R group was 1360 and in the C group 1247. Due to problems in the randomization in six of the 17 centers the C group became substantially smaller than the R group. Only six centres (Ghent, Helsinki, Turku, Kaunas and Warsaw) had recruited R and C groups that had comparable baseline characteristics. Although the original sample size calculations had been performed with the idea of pooling the results from all centres, such analyses were not done due to “the imbalance between the R and C groups at baseline”. Unfortunately the tables showing the results do not give the numbers of patients with different endpoints (CVD death, death from MI, sudden death, all-cause death, non-fatal MI, fatal MI) in the R and C group at each center, but instead the results are given as cumulative percentages for 3-month, 6-month, 1-year, 2-year and 3-year follow-up periods. Since the number of patients recruited at individuals centers was in most instances small (median number 128 patients (range from 54 to 405), it is clear that the statistical power of analyses by centers was very poor. In three centers there was statistically or almost statistically significant (p = or < 0.05) difference in all-cause mortality, CVD mortality and mortality from MI between the R and C groups: in Turku and Moscow in a favourable direction and in Ghent in an unfavourable direction. A recalculated table is as follows: 

Endpoint	R group (n = 1360)	C group  (n = 1247)
Death from all causes	195 (14.3%)	211 (16.9%)
Death from CVD	178 (13.3%)	197 (%15.8)
Death from MI	153 (11.2%)	183 (14.7%)
Sudden death	115 (8.5%)	109 (8.7%)
Non-fatal MI	129 (9.4%)	112 (9.0%)

 

Thus, with all the precautions needed for pooled data there was a trend in a favourable direction in all-cause mortality (relative risk reduction by 15%), in CVD mortality (relative risk reduction by 16%) and in mortality from MI (relative risk reduction by 24%).

Conclusions:

The report concludes that the study failed to provide an answer to the question of whether comprehensive programmes, as applied in separate centers, could reduce mortality and morbidity after MI in the long term. A number of problems in the conduct of the study were listed: 1) The number of patients and the duration of observations were inadequate for detecting differences between R and C groups at the level of centers; 2) Most centers showed lower mortality and morbidity than would be expected from data from population based studies on MI patients, e.g. results from MI register studies conducted in Europe in the beginning of the 1970s; 3)The programmes, although comprehensive in their design and potentially efficient, were not always adequately implemented in the R group. This seems likely looking at the results concerning changes in risk factors and behaviour between R and C groups; and 4) The management of the trial and surveillance of its conduct as a whole was inefficient, resulting in the deviations from the original study design in many centres leading to the loss or exclusion of seven of the original 24 participating centers from final data analyses.

Finnish Centers:

The two Finnish centers participating in individual randomization in the WHO study – Helsinki and Turku – have published their pooled results for 3-year, 10-year and 15-year follow-up periods. The total number of male patients who had MI below age 65 was 375. After three years’ follow-up (5) the cumulative coronary mortality was significantly smaller in the R group than in the C group (18.6 % vs. 29.4%; p = 0.02). This difference was mainly due to the reduction of sudden deaths in the R group (5.8% vs. 14.4%; p<0.01). After 10 years’ follow-up (6) the lower coronary mortality in the R group persisted (35.1% vs. 47.1%; p=0.02) and this also applied to sudden deaths (12.8% vs. 23.0%; p=0.01). During 15 years’ follow-up (7) these differences still persisted (coronary mortality 47.9% vs. 58.5%: p=0.04, and sudden death 16.5% vs. 28.9%; p=0.006). (KP)