Longitudinal study of heart disease in a Jamaican rural population

Bull. Org. mond. Sant} 1972, 46, 685-694 Bull. Wid Hlth Org. Longitudinal study of heart disease in a Jamaican rural population * 2. Factors infuencing mortality W. E. MIALL,1 E. DEL CAMPO,2 J. FODOR,2 J. R. NAVA RHODE,2 L. RUIZ,2 K. L. STANDARD,3 & A. V. SWAN 4 The relationship between cardiovascular characteristics and mortality was investigated in an epidemiological study of heart disease in a representative adult rural community in Jamaica. Of 449 men and 469 women followed up for 5 years, 36 men and 28 women died and the data concerning their status as regards arterial pressure, electrocardiographic abnormalities, and histories of effort pain at the intial survey have been analysed. Cardio- vascular disease, and heart disease in particular, was the major cause of death in this population. Blood pressure levels exceeding 160/95 mm Hg had been recorded in about one third of the men and half the women who died and a clear trend was found between overall mortality and arterial pressure. Symptoms of effort pain and ECG abnormalities compatible with myocardial ischaemia, both of which were unexpectedly common, appeared to have independent prognostic significance. The prognosis of each was worse when asso- ciated with hypertension; hypertension unaccompanied by either effort pain or ECG " ischaemic " abnormality, on the other hand, caused no excess mortality in either sex within the period offollow-up. Although classical myocardial infarction was confirmed to be relatively infrequent, myocardial disorders with many of the features of ischaemic heart disease are an important cause of death in rural Jamaicans. In a survey of the prevalence of cardiovascular abnormalities in a random sample of Jamaican rural adults we reported high rates for symptems and signs of myocardial disorder (Miall et al., 1972). These cardiovascular abnormalities had many fea- tures suggesting myocardial ischaemia, which was surprising in a population of African origin, com- * This study was supported by the Medical Research Council and Ministry of Overseas Development and by a grant from the World Health Organization. 1 Formerly Director, MRC Epidemiology Unit (Jamaica), University of the West Indies, Kingston, Jamaica. Present address: Epidemiology and Medical Care Unit, Medical Research Council and Department of Health and Social Security, Northwick Park Hospital, Harrow, Middlesex, England. 2Wko Fellow. ' Epidemiologist, MRC Epidemiology Unit (Jamaica). Present address: Professor & Head, Department of Social & Preventive Medicine, University of the West Indies, Kingston, Jamaica. 'Statistician, MRC Epidemiology Unit (Jamaica), Uni- versity of the West Indies, Kingston, Jamaica. prising agricultural workers and their womenfolk who lead physically active lives, who are not obese (Ashcroft et al., 1966), who eat a diet low in calories and animal fats (Cruickshank & Fox, unpublished data), and who do not smoke heavily. Furthermore, as Jamaicans were included in the International Atherosclerosis Project and as it was shown that they had less severe and extensive coronary athe- roma than black or white populations in the USA (McGill et al., 1968), the finding that the prevalence of electrocardiographic abnormalities was equal to those reported from Western industrial communities required some explanation in terms of other cardiac disorders more ccmmon in Jamaica than elsewhere. Hypertensive heart disease alone did not appear to be an adequate explanation. The population investigated and the survey tech- niques have been described in the previous paper in this series. A total of 1 067 subjects aged 35-64 years, i.e., 88.9% of the 1 200 invited to attend for clinical examination, were seen in surveys in 1962-63 and 2847 -685- W. E. MIALL ET AL. 1963-64 and the survivors were followed up exactly 5 years later. This paper reports the relationship between the different cardiovascular abnormalities found in the initial survey and the mortality that occurred in the ensuing 5 years. METHODS To determine the risks associated with the dif- ferent abnormalities, certain arbitrary criteria have been used in defining the groups at risk. The stan- dardized questionnaire for cardiac symptoms devised by Rose (1962) was used. All who responded posi- tively to the questions concerning effort pain were analysed as cases of possible angina though some responses at the time seemed doubtful and the symptom was atypical in that it was usually only provoked by very considerable exertion. Likewise, those who gave a history of severe chest pain of more than 30 min duration were analysed as cases of possible infarction though they rarely described an episode with the intensity of symptoms associated with a major coronary occlusion, and such symp- toms were described by several subjects who had not been hospitalized at the time. The criteria for hypertension recommended by WHO-160 and/or 95 mm Hg-have been used to ismplify the description of the findings but analyses avoiding the use of arbitrary thresholds have also been carried out for blood pressure and other char- acteristics such as serum cholesterol and body weight whose distributions are unimodal. The criteria used for ECG abnormalities are those of the revised Minnesota code (Rose & Blackburn, 1968). Those with Q/QS items (11-3), S-T depression (41-3), T wave changes (5,3), or left bundle branch block (71) have been grouped as compatible with myocardial ischaemia. The risks associated with abnormal findings at the first survey have been assessed by mortality ratios calculated by the following method. The number of total deaths that occurred within each age and sex group being known, the number of "expected" deaths within subgroups was calcu- lated on a proportional basis. The ratio of the ob- served to the expected deaths was then expressed as a percentage. Where a single figure is used for all ages, it represents the ratio of the observed deaths to the sum of the expected deaths calculated for separate 5-year age groups, and is therefore age corrected. During the 5-year interval 36 men and 28 women died. The analyses are therefore based on small numbers at the present time. An analysis based on all the deaths that occurred before 1 Octo- ber 1970 is included in Appendix Table 1. RESULTS Altogether 413 (90.4%) of the 457 men available for follow-up and 441 (93.2%) of the 473 women still living in the area were fully reexamined. Includ- ing those who died, information is available for 91.1 % of men and 93.6% of women. Many of those who left the area emigrated overseas and it was uncertain whether they were still alive. For simplicity, data for those who left the area and for those still alive but not fully examined in the follow-up survey have been omitted from the analysis. Deaths were attributed to cardiovascular-renal causes in 24 of the 36 men: 16 of these deaths were from cardiac causes and 5 from cerebrovascular causes (Table 1). Among the cardiac deaths in men, 4 were sudden. Three of the 4 men had given a history of angina; one had also given a history of intermittent claudication. Three of the 4 men had ECGs showing Minnesota code 11.3, 4_3, or 51-3 items-changes compatible with myocardial ischae- mia. The 12 other cardiac deaths in men included 3 with angina, 2 with hypertension and 2 with diabetes. Only 3 of the 12 had shown ECG " ischae- mic" patterns. The 5 men who died of cerebro- vascular disease were all hypertensive; 3 had given histories of angina, 1 also with a history of possible infarction. Three of the 5 had 11.3, 46-3, or 51-3 ECG items. None of those dying of renal disease had hypertension, angina, or ECG changes. The 12 deaths not due to cardiovascular-renal causes included 1 man with angina and ECG changes indicating diffuse myocardial disease and 3 others with ECG changes compatible with ischaemia. Only 1 of the 12 was hypertensive. Of the 28 deaths in women, 17 resulted from cardiovascular-renal conditions; 9 of these were from cardiac causes and 5 from cerebrovascular causes. Only 1 of the 9 persons who died from cardiac conditions had given a history of angina but 7 had shown 11-3, 413, or 5,3 ECG items; 4 were hypertensive. Four of the 5 who died of cerebro- vascular causes were hypertensive, 2 had angina, and 3 had 11-8, 41-3, or 51-3 items in their ECGs. One was diabetic. The 3 women who died of renal disease were all hypertensive; one of those with angina and ECG changes had diabetes and died of the Kimmelsteil- 686 HEART DISEASE IN JAMAICA. 2 Table 1. Causes of death as certified at the 5-year follow-up, Lawrence Tavern, Jamaica Cardiovascular-renal Age group Other Not known Total (years) Cardiac Cerebro- Renal vascular Males 35-44 2 2 0 0 0 4 45-54 2 1 2 4 0 9 55-64 1 2 2 1 8 0 23 total 16 5 3 12 0 36 Females 35-44 1 0 0 2 0 3 45-54 1 1 1 2 0 5 55-64 7 4 2 6 1 20 total 9 5 3 10 1 28 Wilson syndrome. Three of the remaining 11 had ECG abnormalities including one woman who died of sickle-ell disease; one woman with diabetes had angina. Four of the 11 were hypertensive. Twelve autopsies were carried out at the Univer- sity of the West Indies Hospital on those considered to have heart disease when seen at the survey. Four had shown ECG signs of ischaemia and had had either angina or a history of infarction; one of these had small-artery disease and a heart weight of 610 g; two had severe occlusive atheroma with old infarc- tions and heart weights of 480 g and 640 g; one had mild coronary atheroma and minimal myocardial fibrosis but a heart weight of 625 g. Six subjects had had ECG signs of ischaemia but no angina; three of these had moderately severe coronary athe- roma and heart weights of 650 g, 540 g, and 480 g. Three with nonspecific T wave changes had minimal atheroma of the extramural arteries; one, a dia- betic, had a heart weight of 580 g; one, a woman with sickle-cell disease, had a heart weight of 505 g, and the third, a hypertensive dwarf, had a heart weight of 220 g. The remaining two subjects had had angina alone; one who had been hypertensive had healthy coronary vessels but a heart weight of 450 g, the other had multiple myelomatosis with severe anaemia (5.8 g haemoglobin/100 ml blood) and a heart weight of 340 g. Six of these twelve subjects had had severe hypertension. Although it is clear that there is reasonably close agreement between the survey findings, the certified causes of death, and the autopsy results we believe the causes of death to be insufficiently reliable to warrant any further analysis. The inclusion of only those deaths whose causes are known with some certainty and the exclusion of those in which there was doubt would, however, make comparison with other studies impossible. The following analyses are therefore based on all deaths, irrespective of cause. Influence of hypertension, certain ECG changes, and effort pain on prognosis The 449 men and 469 women who were followed up or who had died in the intersurvey period have been analysed in terms of their first survey status as regards hypertension, ECG changes, and histories of chest pain on exertion. The ECG changes include only those compatible with, but not necessarily due to, myocardial ischaemia, i.e., Q/QS items, S-T seg- ment depression, T wave inversion, T wave flatten- ing, and left bundle branch block. Fig. 1 summarizes the composition of the followed-up population. Those with hypertension, with angina, and with ECG " ischaemia " (11_3, 41-3, 51-3, 71) are indicated and the extent of overlap between these groups can be seen. 687 W. E. MIALL ET AL. MALES FEMALES L] Normotension Hypertension 1- WHO 20166 Fig. 1. Initial prevalence of hypertension (> 160 and/or 95 mm Hg), angina pectoris, and ECG "ischaemia" (Minnesota code items 11-3, 41-3, 51-3, and 71) in the followed-up population, by age and sex. The size of each area is proportional to the number of subjects in the group. Angina | Ischaemia Table 2 shows the numbers of subjects at risk, the numbers who died and the numbers of " expected deaths" for each category. The age-adjusted mor- tality ratios for each group are shown in Fig. 2. The mortality ratio for all hypertensives exceeded the expected (119 for men; 134 for women) but in both sexes the excess mortality was restricted to those with 11-3, 41-3, 51-3, or 71 items. Hypertensive sub- jects without ECG evidence of cardiac damage had mortality ratios of only 59 for men and 82 for women, whereas with ECG changes they showed ratios of 288 and 231, respectively. Normotensive men and women without ECG evidence of myocardial dam- age had mortality ratios of 77 and 65, respectively, but with these ECG items the ratios were 200 and 138. Where a positive response had been given to the angina questionnaire, mortality ratios were markedly increased for men but were not consistently increased in women. Hypertensive men with ECG changes had a mor- tality ratio of 484 if they also had angina but of 220 if they did not. Hypertensive men without ECG changes showed mortality ratios of 364 if they had given a history of angina and only 32 if they had given no such history. The pattern among normo- tensive men was equally regular. Those with ECG changes and angina had a mortality ratio of 294 and those with ECG changes without angina one of 177. Normotensive persons without ECG abnor- malities but with angina showed a ratio of 184, and those without angina one of 69 (see Appendix Table 1). 55 - 64 years 688 111111- 36 - 44 years HEART DISEASE IN JAMAICA. 2 Table 2. Observed and expected deaths in hypertensive a and normotensive subjects according to their ECG and angina status at the initial surveys: 5-year follow-up, Lawrence Tavern, Jamaica Hypertensive at initial survey Normotensive at initial survey With Without With Without ECG ischaemia b ECG ischaemia b ECG ischaemiab ECG ischaemia b Total With Without With Without With Without With Without angina angina angina angina angina angina angina angina Males at risk 6 20 7 64 7 28 20 297 449 observed deaths 3 4 2 2 2 5 3 15 36 expected deaths C 0.62 1.81 0.55 6.28 0.68 2.82 1.59 21.65 36.00 Females at risk 13 38 14 83 4 43 16 258 469 observed deaths 4 5 0 6 0 4 1 8 28 expected deaths c 0.95 2.95 0.97 6.33 0.23 2.66 1.17 12.73 27.99 a Hypertension-WHO criteria >160 and/or 95 mm Hg. b ECG ischaemia-Minnesota Code (revised) 11, 12, 13, 4i, 42, 43, 51, 52, 53, or 71. c Expected deaths are the sum of the calculated expected deaths in 3 decades, 35-44, 45-54, 55-64 years. Males followed up Females followed up (449) (469) 100 352 WRN 31 1 100~ ~ ~ ~~~~~ ~~~~~~~~~~~~ WITHEcfl4 300 .0 | S ~~~~~~~~~~~~~~~~~~~1 7. wo- 35 f 1200 - x *71 I rHHUT]I~G ] Fig. 2. Mortality ratios in subjects initially with and without hypertension (>160 and/or 95 mm Hg), ECG " ischaemia n (Minnesota code items 11-3, 41-3, 51-3, and 71), and angina pectoris, in the followed-up population: 5-year follow-up survey, Lawrence Tavern, Jamaica. The height of each column represents the ratio of observed to " expected " deaths times 100. The circled figures represent the number of deaths observed. 689 W. E. MIALL ET AL. Table 3. Means and standard deviations of arterial pressure, serum cholesterol, and weight at the initial survey, Lawrence Tavern, Jamaica Males Females Characteristic 35-44 45-54 55-64 35-44 45-64 55-64(years) (years) (years) (years) (years) (years) Mean S.D. Mean S.D. Mean S.D. Mean S.D. Mean S.D. Mean S.D. systolic pressure (mm Hg) 128.2 18.4 134.7 22.5 142.3 27.8 128.0 20.0 145.4 29.0 156.1 28.5 diastolic pressure (mm Hg) 80.6 12.8 84.5 13.4 85.7 14.6 82.6 12.4 90.2 16.1 90.9 15.5 serum cholesterol (mg/100 ml) 203.4 37.8 212.2 41.6 213.9 43.7 219.1 45.0 246.3 48.4 247.3 51.0 weight (kg) 63.0 8.0 63.0 8.1 59.8 7.2 61.1 11.3 60.7 12.7 55.9 10.4 haemoglobin (g/100 ml) 13.9 1.3 13.4 1.8 13.4 1.4 12.5 1.2 12.6 1.6 12.6 1.4 packed cell volume (%) 43.1 3.3 42.2 4.6 41.9 3.6 39.1 3.1 39.5 3.9 39.7 3.6 The numbers available are insufficient for a pre- cise analysis of mortality that allows for differences in age distributions. Since these differences are small, all age groups are considered together. The relationship between angina (disregarding ECG " ischaemia ") and mortality is significant in hyper- tensive men (P< 0.01) and in normotensive men (P<0.05). Likewise the relationship between ECG " ischaemia " (disregarding angina) and mortality is significant in both hypertensive men (P<0.02) and normotensive men (P<0.01) when age groups are pooled. In women, angina (disregarding ECG " ischaemia ") and mortality are not significantly related in either hypertensive or normotensive sub- jects, and the relationship with ECG " ischaemia " (disregarding angina) and mortality is not quite sig- nificant (0.10>P>0.05) in hypertensive and not significant in normotensive women. Subjects with either angina or ischaemia have a higher mortality than those with neither in all sex and blood pressure groups, while subjects with both, except for normotensive women, have a still higher mortality. This progressive increase in mor- tality is significant for both normotensive and hypertensive men (P 0.01) and for hypertensive women (P< 0.05). Larger numbers would be needed to test how much of this effect is due to angina and how much to " ischaemia ". However, it is reasonable to infer from these data that the observed increase in mortality is not simply due to one factor alone. Relationship between overall mortality and arterial pressure, serum cholesterol, and body weight Table 3 shows the means and standard deviations of systolic and diastolic pressure, serum cholesterol, body weight, haemoglobin, and packed-cell-volume values for men and women at the initial survey. For the first four of these characteristics, mortality rates have been calculated for all subjects according to their deviation from the mean for their age and sex group. If standard deviation units are used no account need be taken of the different mortality rates in different age groups or of the changes in mean values with age. The mortality rates per 1 000 at risk for those whose characteristics were 1 S.D. or more below the mean, within 1 S.D. of the mean, between 1 and 2 S.D. above the mean, and 2 or more S.D. above the mean are shown in Table 4. With systolic and diastolic pressures the higher the pressure the greater the mortality in both sexes. Mortality trends for serum cholesterol and for body weight were, some- what surprisingly, in the opposite direction. For both sexes, and for each of these characteristics, mortality was highest in those with low values and lowest in those with high values. The data from which the calculations were made for cholesterol levels show that among men 4 deaths occurred in the 66 with values exceeding 250 mg/100 ml (6%) whereas 18 occurred in the 187 with values below 200 mg/100 ml (10%). Among women the com- parable figures were 8 of 150 with values over 250 mg/100 ml (5%Y.) and 12 of 109 with values below 200 mg/100 ml (11 %). In neither case are these differences statistically significant but there is no suggestion that high serum cholesterol levels are associated with excess mortality in this popula- tion. This finding supports other evidence indica- ting that causes of death associated with high chol- esterol levels, such as those due to coronary athe- roma, are relatively uncommon. 690 HEART DISEASE IN JAMAICA. 2 Table 4. Five-year mortality rates (per 1 000) for males and females, according to their deviation from normal values of four characteristics-systolic pressure, diastolic pressure, serum cholesterol and weight: Lawrence Tavern, Jamaica 1 S.D. or more Within 1 S.D. 1-2 S.D. above 2 S.D. or more Characteristic below mean value of mean value mean value above mean value for age for age for age for age Males systolic pressure 74 77 94 i11 diastolic pressure 79 76 83 95 serum cholesterol 109 85 63 0 body weight 153 72 67 0 Females systolic pressure 18 48 111 217 diastolic pressure 40 55 77 143 serum cholesterol 132 47 44 0 body weight 83 61 33 0 Serum cholesterol levels in Lawrence Tavern are weakly but in general positively correlated with measures of body bulk; the lower mortality of those with higher cholesterol levels may merely reflect their better nutritional state. The inclusion of sub- jects who at the time of the first survey were already suffering from, and subsequently died of, chronic wasting disease might therefore have explained the finding. An analysis restricted to those who died and at the first survey had shown some cardio- vascular abnormality excluded most deaths due to noncardiovascular causes but left the relationship between high mortality and low serum cholesterol and body weight almost unchanged. Serum chol- esterol levels have also been shown to be lower in anaemic subjects (Elwood et al., 1970) but the weak relationships between cholesterol and haemoglobin values in this study suggest that anaemia is unlikely to play an important role in causing high mortality in those with low chol- esterol values. Influence of glycosuria on mortality The 9 males and 13 females with either glycosuria or known diabetes who were followed up included 30% more hypertensive subjects, 67% more cases of angina, and 95% more with ECG code items 11.3, 41-3, 51-3, and 71 than would be expected on the basis of their number, age, and sex. Six of the 22 died though the " expected " number was 2.23. This small group confirmed that diabetics are sub- ject to the same cardiovascular complications in rural Jamaica as elsewhere. Comparison with other studies Few other surveys have reported the prognostic significance of cardiovascular abnormalities in a manner that allows comparison with our findings. Comparison with the 8-year follow-up of the Fram- ingham project (Higgins et al., 1965) (data kindly provided by Dr D. Shurtleff) shows a close similarity in mortality ratios for hypertensive and normoten- sive subjects with and without 11-8, 41-4, 51-3, or 71 items (Table 5). The similarity in the prognostic significance of various grouped ECG items (Q/QS patterns with or without S-T and T changes, S-T depression and T wave changes, and T wave changes alone) in Framingham and Lawrence Tavern has been described (Miall, 1970). Keys and his colleagues have recently reported the 5-year follow-up of their cardiovascular surveys in 7 countries (Keys, 1970) but comparison using identical criteria is impossible and their analysis of the use of electrocardiograms in prediction relates solely to coronary disease incidence (Blackburn et al., 1970). DISCUSSION The mortality study reported here adds significance to the finding of a high prevalence of cardiac symp- toms and signs described previously for this popula- tion (Miall et al., 1972). It is apparent that cardio- 691 692 W. E. MIALL ET AL. Table 5. Ratio of observed to expected deaths (xlOO) in the 8-year follow-up of Framingham, USA, and in the 5-year follow-up of Lawrence Tavern, Jamaica, according to blood pressure and ECG status at the original survey Males Females Blood pressure and ECG status Lawrence Lawrence Framingham a Tavern b Framingham a Tavern b BP > 160 and/or 95 mm Hg with 11-3, 41-3, 51-3, or 71 268 288 203 231 without these ECG items 147 59 109 82 BP < 160 and/or 95 mm Hg with 11-3, 41-3, 51-3, or 71 146 200 155 138 without these ECG items 62 78 76 65 a Framingham population aged 30-62 years at the first survey. b Lawrence Tavern population aged 35-64 years at the first survey. vascular disease, particularly heart disease, is the main cause of death in these rural Jamaican adults and though the number of deaths is still small it is clear that symptoms of effort pain and electro- cardiographic abnormalities have independent prog- nostic significance. Their significance in terms of cardiovascular mortality alone would presumably be greater than that indicated in this analysis based on all deaths. The influence on prognosis of effort pain or ECG "ischaemia " appears greater when both are present, and is greater for either if accompanied by hyper- tension, as would be expected. It is of interest that in this population hypertension, if unaccompanied by either angina or ECG " ischaemia ", appears to have little adverse influence on mortality within 5 years-or indeed within 7-8 years, as is shown in Appendix Table 1. This finding if confirmed on a larger scale may have relevance in screening those hypertensive subjects requiring therapy.1 In the preceding paper it was mentioned that both the symptoms and ECG signs encountered in this survey suggest that much of the disease seen was ultimately ischaemic in origin but that both of these characteristics had unusual features (Miall et al., 1972). Angina that was usually only pre- cipitated by severe exertion, prolonged chest pain that caused distress but not hospitalization, and 1 A recent review of all deaths known to have occurred before 1 October 1970 fully supports the 5-year analysis. Current mortality ratios based on 78 deaths (47 in men, 31 in women) are shown in Appendix Table 1. ECG changes that, when they suggested possible infarction, were often restricted to the right precor- dial leads or avL, are all atypical of coronary heart disease in the usual sense of the term. Though coronary atheroma and hypertensive heart disease contribute to the overall prevalence of car- diac disease in this community, small-artery involve- ment may be an additional factor that could under- lie the high prevalence of an ischaemic type of heart disease in a population that was expected on the basis of what is known of extramural coronary artery atheroma in Jamaicans (Robertson, 1959; Tejada et al., 1968; Summerell et al., 1968) to have a low prevalence. Campbell et al. (1971), in a recent analysis of another 21 cases of idiopathic cardiomegaly of the type described from Jamaica by Stuart & Hayes (1963), report pathological changes in the intramural vessels in this condition. The types of small vessel disease recognized by James and his colleagues (James, 1964, 1967; James & Fisch, 1963) include hereditary medial necrosis of small vessels and dis- orders associated with Friedreich's ataxia and other neurological and musculoskeletal disorders. No obvi- ous evidence of a hereditary element, or of associated extracardiac disease, was revealed by the survey. In the following paper in this series it is shown that factors influencing the progression and incidence of electrocardiographic abnormalities are those ex- pected for an ischaemic heart disorder, and the pos- sible contributory role of pathology of the small arteries in this population is discussed in more detail. HEART DISEASE IN JAMAICA. 2 Appendix Table 1. The observed and expected deaths in hypertensive and normotensive subjects at the 7-8-year follow-up (age-corrected) according to their ECG and angina status at the initial surveys, Lawrence Tavern, Jamaica Males Females Cardiovascular status at first surveys Observed Expected Mortality Observed Expected Mortality deaths deaths ratio (x 100) deaths deaths ratio (x 100) Hypertensive at first survey with 1 1-3, 41-3, 51-3, or 7i items with angina 5 0.83 602 4 1.05 381 without angina 5 2.37 211 6 3.28 183 without 11-3, 41-3, 51-3, or 7i items with angina 2 0.71 282 0 1.07 without angina 5 8.29 60 6 7.04 85 Normotensive at first survey with 11-3, 41-3, 51-3, or, 71 items with angina 2 0.90 222 1 0.37 270 without angina 5 3.73 134 5 3.07 163 without 11-3, 41-3, 51-3, or 7i items with angina 3 2.09 144 1 1.30 77 without angina 20 28.07 71 8 13.80 58 total 47 46.99 100 31 30.98 100 R12SUM1E tTUDE LONGITUDINALE DES CARDIOPATHIES DANS UNE POPULATION RURALE DE LA JAMAYQUE: 2. FACTEURS INFLUENQANT LA MORTALITE Dans ce deuxieme article consacre aux cardiopathies a la Jamaique, on examine les relations entre les divers troubles cardio-vasculaires deceles lors de l'enquete initiale et la mortalite enregistree au sein du groupe de population etudi6 pendant les cinq annees suivantes. Parmi les 449 hommes suivis durant cette periode, on a enregistr6 36 dec's; 24 ont ete attribues a une affection cardiaque ou renale, dont 16 'a une cause car- diaque. Sur les 28 dec&s survenus chez les 469 femmes, 16 ont ete attribu6s a une affection cardiaque ou r6nale, dont 8 'a une cause cardiaque. En depit d'une concordance satisfaisante entre les donnees de l'enquete, la cause indiquee sur le certificat de deces, et les resultats d'autop- sie, on a consider6 les mentions portees sur les certificats comme insuffisamment sfires et on a base l'analyse sur la mortalite globale. Les dossiers des 449 hommes et des 469 femmes qui ont ete reexamines ou qui sont morts au cours des cinq annees ont e analyses en fonction des observations faites lors de la premiere enquete portant sur l'hyper- tension arterielle (pression sanguine egale ou superieure it 160/95 mm Hg), les alterations du trac6 electrocardio- graphique (anomalies des ondes Q/QS, abaissement de S-T, anomalies de l'onde T, bloc de branche gauche complet) compatibles avec une ischemie myocardique, et l'existence de douleurs thoraciques a l'effort. On a constate que les douleurs a l'effort et les modi- fications du trac6 qevoquant une isch6mie avaient leur signification propre au regard du pronostic, ce dernier etant aggrave par leur association et une hypertension concomitante apportant un element de risque supple- mentaire. Quant a l'hypertension isolee, sans douleurs thoraciques et sans signes electrocardiographiques d'ischemie, elle n'a eu qu'une faible influence sur la mortalite pendant la periode consideree. Ces faits confirment les observations anterieures montrant la forte pr6valence des sympt6mes et des signes de troubles myocardiques compatibles avec une 693 694 W. E. MLALL ET AL. ischemie. L'autopsie de 12 sujets reconnus atteints d'une affection cardiaque a 1'examen initial a fait d6couvrir chez un petit nombre d'entre eux des l6sions ath6ro- mateuses peu importantes des arteres coronaires extra- murales et aussi des myocardes de poids anormalement et inexplicablement 6leve. I1 est probable que des lesions des petits vaisseaux coronaires intramuraux sont a l'ori- gine de certains de ces cas et jouent un r6le dans la pr6valence globale des cardiopathies dans la collectivit6 6tudiee. REFERENCES Ashcroft, M. T. et al. (1966) Brit. J. prev. soc. Med., 20, 22-26 Blackburn, H. et al. (1970) In: Keys, A., ed., Coronary heart disease in seven countries, New York, American Heart Association Inc., pp. 154-161 (American Heart Association, Monograph No. 29) Campbell, M. et al. (1971) Brit. Heart J., 33, 193-202 Elwood, P. C. et al. (1970) Lancet, 1, 589-591 Higgins, I. T. T. et al. (1965) Brit. J. prev. soc. Med., 19, 53-68 James, T. N. (1964) Progr. cardiovasc. Dis., 7, 43-64 James, T. N. (1967) Amer. J. Cardiol., 20, 679-691 James, T. N. & Fisch, C. (1963) Amer. Heart J., 66, 164-175 Keys, A. (1970) Coronary heart disease in seven countries, New York, American Heart Association Inc. (American Heart Association Monograph No. 29) McGill, H. C. et al. (1968) Lab. Invest., 18, 498-502 Miall, W. E. (1970) Current developments in South and Central America and the Caribbean. In: Proceedings of the Second International Symposium on Atherosclerosis Chicago, 1969, Berlin, Springer-Verlag Miall, W. E. et al. (1972) Bull. WldHlth Org., 46, 429 Robertson, W. B. (1959) Lancet, 1, 444-446 Rose, G. A. (1962) Bull. Wld Hlth Org., 27, 645-658 Rose, G. A. & Blackburn, H. (1968) Cardiovascular survey methods, Geneva (World Health Organization: Monograph Series, No. 56) Stuart, K. L. & Hayes, J. (1963) Quart. J. Med., 32, 99-114 Summerell, J. M. et al. (1968) Trop. geog. Med., 20 127-132 Tejada, C. et al. (1968) Lab. Invest., 18, 509-526