Longitudinal study of heart disease in a Jamaican rural population

Bull. Org. mond. Sante 1972, 46, 429-442 Bull. Wld Hlth Org. Longitudinal study of heart disease in a Jamaican rural population * 1. Prevalence, with special reference to ECG findings W. E. MIALL,' E. DEL CAMPO,2 J. FODOR,2 J. R. NAVA RHODE,2 L. RUIZ,2 K. L. STANDARD,3 & A. V. SWAN4 A long-term epidemiological study of heart disease in a representative rural community in Jamaica was started in 1962-63 and the first follow-up survey was carried out in 1967-68. This report describes the prevalence ofseveral cardiovascular characteristics at each survey, and their associations with other measurements. The nature of the electrocardiographic abnormalities and their relationship with symptoms of effort pain and prolonged chest pain suggests that much of the disease seen in this population is ultimately ischaemic in origin despite evidence that classical myocardial infarction and severe coronary atheroma are relatively infrequent. Nevertheless both the symptoms and the electrocardiographic abnor- malities had features that were not completely typical of occlusive disease of extramural coronary arteries. These findings are discussed in terms of the four conditions-hyperten- sion, conventional coronary heart disease, small artery disease, and cardiomyopathy- that are believed to account for most cases of heart disease in this community, and it is con- cluded that the overall pattern of disease cannot be explained by any single disorder of overriding importance. The evidence suggests that all may be important contributors. In Jamaica, occlusive coronary artery disease has been shown to be relatively rare. The International Atherosclerosis Project, which included Jamaica as one of sixteen countries collaborating in a study of the geographical distribution of atheroma, reported Jamaicans to have lower rates for raised atheroma- tous lesions of the aorta and major coronary arteries than are found in North Americans of similar ethnic origin, and much lower rates than are present in North American whites or Europeans (McGill et al., 1968). Arterial hypertension, on the other hand, is relatively common in Jamaica (Miall et al., 1962). These findings agree with clinical experience and are supported by analyses of routine autopsy data in the * 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, Medical Research Council Epidemio- logy 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. 2 WHO Fellow. 3Epidemiologist, MRC Epidemiology Unit (Jamaica). I Statistician, MRC Epidemiology Unit (Jamaica). hospital of the University of the West Indies (Sum- merell et al., 1968). In 1962 the staff of the Medical Research Council's Epidemiology Unit (Jamaica) started a prospective study of cardiovascular disease in a random sample of the general population of a representative rural community to obtain a more complete picture of the pattern of cardiac disorders than can be gained from the study of hospital patients and autopsy series. The subjects in this sample were initially examined in 1962 and 1963, and the survivors were reexamined 5 years later. This paper describes the prevalence of cardiovas- cular disorders seen at the initial survey and at the first 5-year follow-up, and compares it with reported results of other similar investigations. Subsequent papers in the series (Miall et al., 1972a, 1972b) report the relationship between the initial cardio- vascular findings and mortality in the ensuing 5-year interval, the changes occurring in serial ECGs, and the quantitative characteristics ofECG measurements and their relationship with age, arterial pressure, heart rate, and body build. 2816 - 429 430 W. E. MIALL ET AL. POPULATION AND METHODS OF STUDY The population studied The people chosen for this study were the inhabi- tants of an agricultural area centred on the village of Lawrence Tavern in upper St Andrew, 24 km from the island's capital, Kingston; the majority were -farmers and their wives, for few alternative occu- pations are available in that district. The terrain at Lawrence Tavern is remarkable for its steep hill- sides, which are separated by deep gullies. Though the houses tend to be clustered along the ridges between gullies, the tracks connecting them with the few roads are steep. For this reason, and because most men and some women work on land that is too precipitous for mechanical methods of culti- vation to be used, almost all must lead physically active lives. The agriculture is based on the culti- vation of smallholdings of 8 000-12 000 m2 where a wide variety of fruit and vegetable crops are grown for home consumption or for sale in Kingston mar- kets. Many families have relatives overseas remit- ting earnings back to the district, and 'this, combined with the proximity to Kingston where remunerative work is more readily available, is resulting in notice- able improvements in housing, furnishing, and cloth- ing. Better communications, public services, water supplies, etc., are being provided from national resources. Though some people live at little above subsistence level, few are destitute. The Jamaican population is predominantly of African origin but shows evidence of its mixed ancestry; many people have European, Indian, or Chinese features though rural populations show less racial admixture than is seen in the larger towns. Both men and women are muscular and well built. In general the health standards are good and the dis- eases now prevalent are those of Western industria- lized populations living in temperate climates rather than of tropical underdeveloped areas. Infant mor- tality is less than 35 per 1 000 live births; life-expec- tatioti at birth was 63 for males and 67 for females for the period 1959-61 and it has probably risen since. Malaria and yellow fever have been eradicated. Yaws still occurs in isolated pockets but is a rarity. Neither the vector nor the trypanosome (Trypano- soma cruzi) responsible for Chagas' disease, an im- portant cause of heart disease in South America, is found in Jamaica. Hookworm infestation, though still prevalent, rarely causes severe anaemia. On the other hand hypertension, diabetes mellitus, and peptic ulceration are common. Methods and techniques The investigations were carried out in samples of adults aged 35-64 years at the time of the initial surveys. Two samples, together comprising 200 sub- jects of each sex in each of the three decades, were randomly selected from a population of about 8 000 that had been enumerated in a private census early in 1962. These 1 200 people were asked to attend for cardiological investigations at a central clinic in two groups of 600 during the autumn and early spring of 1962-63 and 1963-64: 1 067 of them (88.9 %) did so, 526 men and 541 women. During the succeeding 5 years, 36 men and 28 wo- men died, and 33 men and 40 women left the district. In the follow-up surveys carried out at the same time of year in 1967-68 and 1968-69, 413 men and 441 women (90.4% and 93.2%, respectively, of those still resident in the area) were reexamined by the same techniques. A. further 35 (3.8 %) were reexam- ined electrocardiographically only (Table 1). At the initial surveys medical, obstetric, and per- sonal histories were obtained; these included the questionnaire for eliciting information about cardio- vascular symptoms designed by Rose (1962). Blood pressure measurements, recorded from the right arm after the subject had been sitting for at least 5 min- utes, were made with a random-zero sphygmomano- meter designed by Garrow (1963). In the follow-up surveys the improved model of this instrument1 described by Wright & Dore (1970) was used. A clinical examination of the heart and peripheral cir- culation was carried out and a 12-lead electro- cardiogram was taken. The results of haematological tests have been reported previously (Miall et al., 1967). Haemo- globulin electrophoresis was performed on each specimen after lysing with water and toluene, using a vertical paper method (Lehmann & Ager, 1960) and the tris-buffer system of Goldberg (1959). Serum cholesterol was measured by the method of Zlatkis, Zak & Boyle, using Maclntyre & Ralston's (1954) modification. Urine specimens were tested by paper strips (Uristix) for protein and sugar; clean catch specimens from women were cultured and subjected to quantitative bacteriological investigation using the methods of Kass (1960). At the follow-up surveys most of these investi- gations were repeated using the same techniques and the same ECG reader. The prevalence of ECG 1 This instrument is made and distributed by Hawksley & Sons Ltd., 12 Peter Road, Lancing, Sussex, England. HEART DISEASE IN JAMAICA. 1 431 Table 1. Original population sample, Lawrence Tavern, Jamaica, and number available 5 years ater Followed-up Age at Original Did Lf ra No. available Responsefirst survey Opion follow-up Complete ECG only Total % ofthose available Males 35-44 168 4 19 145 128 8 136 93.8 45-54 180 9 9 162 143 7 150 92.6 55-64 178 23 5 150 142 5 147 98.0 total 526 36 33 457 413 20 433 94.7 Females 35-44 181 3 23 155 140 6 146 94.2 45-54 181 5 6 170 157 6 163 95.9 55-64 179 20 11 148 144 3 147 99.3 total 541 28 40 473 441 15 456 96.4 abnormalities in the first sample at the original survey was reported by Fodor et al. (1964) using a different ECG reader and the original Minnesota code cri- teria (Blackburn et al., 1960). In this paper we re- port the prevalence of ECG abnormalities for all subjects at the original and the 5-year follow-up surveys according to the revised code (Rose & Blackburn, 1968), to which reference should be made for definitions and criteria. RESULTS Prevalence of electrocardiographic abnormalities at both surveys The prevalence of the more important ECG ab- normalities was slightly higher at the initial survey among the original population than in those who subsequently survived to be followed up, as would be expected. Among those not followed-up are those who died, those who refused reexamination, and those who left the district. The two former groups had higher rates of ECG abnormality than those reexamined but the migrants had lower rates.' 1 Tables showing the prevalence of ECG abnormalities in males and females at the initial survey and in the followed- up population, by type of abnormality and by age group, have been deposited in the WHO Library. Copies of the tables may be obtained on request to the Chief Librarian, World Health Organization, 1211 Geneva, Switzerland. The prevalence in the followed-up population, on the other hand, increased between the surveys. When these findings are interpreted a further bias needs consideration. Those who were followed-up were, at the time of the first survey, a selected group of people subsequently known to have survived 5 years, whereas no such selection applies to them at the second survey. As ECG abnormality is associated with increased mortality this selection contributes to, but does not fully account for, the differences between the findings of the two surveys. The ECG tracings for the two surveys were coded independently but when codes changed from one occasion to the next the two tracings were directly compared to ascertain that the recorded changes were real. Minor quantitative changes, possibly artificial and due to technique, accounted for some reclassi- fication. A secular increase in prevalence seems un- likely in view of the short time interval between sur- veys. Q/QS items (11-3). Abnormal Q/QS patterns were coded in the ECGs of 17 men (3.2%) at the first survey; 7 showed QS patterns in the right precordial chest leads, V1-V4, suggestive of anteroseptal infarc- tion but compatible with left ventricle hypertrophy; 7 of the codable Q waves were restricted to the avL lead. Only 6 of the 17 men with Q/QS abnormalities were hypertensive (>160/95) at the time. Thirteen W. E. MIALL ET AL. tracings from women (2.4%) showed Q/QS abnor- malities at the first survey of which 5 were anterosep- tal QS patterns and 4 were abnormal Q waves in avL: 7 of the 13 were hypertensive. Among those subjects who were followed-up the prevalence of Q/QS items doubled from 3.2% to 6.7% in men and from 1.5 % to 3.7% in women. Much of the increase was in minor Q/QS patterns coded 13. Left axis deviation (21). At the initial survey 20 men (3.8 %) and 8 women (1.5 %) had ECGs showing left axis deviation of -30° or more: 5 of the 20 men and 6 of the 8 women were hypertensive. The pre- valence increased from 3.7% to 6.0% in men and from 1.5% to 3.1 % in women. High amplitude R waves (3J). Tall R waves were unusually common in men and showed no consistent trend with increasing age; in women the prevalence was lower but showed increased rates in the older age groups. In the followed-up population the pre- valence increased from 29.5 % to 34.6% in men and from 12.3% to 13.8% in women. S-T segment depression (41-3). At the original sur- vey 19 men (3.6 %) and 29 women (5.4 %) showed S-T depression according to the Minnesota code criteria. One of the men had aortic incompetence; 13 of the 19 men and 15 of the 29 women were hypertensive. The more severe grades of S-T depression (41 and 42) increased in prevalence whereas the lesser grade (43) decreased between the two surveys. The overall rates increased from 3.7% to 4.8 % in men and from 4.8% to 5.0% in women. T wave inversion (51-,). Nineteen men (3.6%) showed T wave inversion at the first survey and without exception this was present in lateral leads (I, II, avL, V5, or V6): 12 of the 19 had raised arterial pressures. T wave inversion was detected in 29 wo- men (5.4 %) but this occurred in lateral leads in only 12, of whom 10 were hypertensive. In 6 women, T wave inversion was found as an isolated abnormality in leads V2-4. Of the 29 with inverted T waves, 14 were hypertensive. The prevalence ofT wave inversion increased from 3.2 to 6.7% in men, and from 4.6 to 7.5% in women, the increase occurring largely in the oldest age groups. Flat T waves (53). Flat T waves, or diphasic or inverted T waves with the negative phase less than 1 mm in amplitude, were recorded in the ECGs of 28 men and 54 women at the first survey, of whom 13 men and 32 women were hypertensive. As with T wave inversion, flat T waves in anteroseptal leads were more common in women (13 cases) than in men (1 case). In lateral leads V5 and V6 they were present in 14 men, of whom 4 were hypertensive and 17 women, of whom 9 were hypertensive. The prevalence increased in the intersurvey inter- val in men, from 4.8 to 8.3%, again largely in the oldest age group. The rate remained about con- stant, 9.4-9.0%, in women. Conduction defects (6,3, 71-3). Atrioventricular conduction defects were remarkably common, and were detected in 27 men and 7 women at the first sur- vey, and in all but two cases the defects were of first degree (i.e., PR interval 0.22 seconds or greater). Only 1 of the 27 men with an A-V conduction defect had a heart rate below 50 beats per minute, but their mean rate of 65.4 beats per minute was slower than that for all other men (72.4 beats per minute) as would be expected. The prevalence remained constant in men, 4.6%, and increased from 1.8% to 2.0% in women. Left bundle branch block occurred in only 3 sub- jects and no new case developed in the intersurvey interval. Complete right bundle branch block was seen in 6 men, incomplete block in 6 men and 2 women. Disorders of rhythm. Ectopic beats, coded as 83 if forming 100% of all beats, were present in 7 men and 2 women at the first survey, and were unusual in that 5 of the 9 cases showed supraventricular pre- mature beats. At follow-up the prevalence increased from 1.2% to 3.0% in men but remained constant in women. Supraventricular ectopic beats predomi- nated in 11 of the 13 cases with this code at the second survey. Atrial fibrillation was encountered in only I per- son, a woman with mild hypertension who was thought to have an obscure cardiomyopathy. Sinus tachycardia (87 if occurring at a rate exceeding 100 beats per minute) was more common in women than in men and at the first than at the second sur- vey. Sinus bradycardia (88 if fewer than 50 beats per minute) was much more common in men than women and its prevalence increased at follow-up. These changes were presumed to result from increased familiarity with the procedure. Miscellaneous items. An interesting feature of this series of electrocardiograms was the frequency of evidence of interatrial delay. Morris's index of left atrial conduction delay, derived from the product of the width (in seconds) and the depth (in mm) of the terminal part of P in V1, is considered abnormal if it exceeds 0.03 mm -sec (Morris et al., 1964J. A 432 HEART DISEASE IN JAMAICA. 1 total of 53 men (10.1 %) and 27 women (5.0%) had an abnormal index at the first survey, of whom 20 men and 15 women were hypertensive, and 17 men and 6 women showed Minnesota code items 11-3, 4,j3, or 5-3. Distributions of arterial pressure at both surveys Means and standard deviations for the distribu- tions of systolic and diastolic pressures are shown in Table 2, together with the proportion of subjects with values of 160 and/or 95 mm Hg or above, for the original population and that followed-up, at the first survey, and for the followed-up population at the second survey. The distributions at the first survey were closely similar for the original population and those subse- quently reexamined. In the groups followed-up the mean systolic pressures rose somewhat less than expected and mean diastolic pressures showed a slight fall instead of an expected rise. For both sexes, distributions for the subjects fol- lowed-up were less skewed and less leptokurtic at the second survey. This mainly resulted from the finding of greater proportions of subjects with pres- sures in the lower ranges than were found initially. The natural trend for arterial pressure, particularly systolic, to increase was counterbalanced by a fall in pressure in many subjects, possibly explained by greater familiarity with the procedure. Effective therapy probably influenced these distributions also but differences in measurement technique may have contributed. The prevalence of hypertension, defined in terms of the thresholds 160 and/or 95 mm Hg, shows the expected increase. Prevalence ofsymptoms at both surveys The assessment of cardiac symptoms was difficult. All those who responded positively to the question- naire concerning effort pain are considered as cases of possible angina, though many responses at the time seemed doubtful and were recorded in the notes as such. Similarly those who gave a history of an episode of prolonged chest pain are all considered as possible cases ofmyocardial infarction, though few described as dramatic an illness as occurs with the occlusion of a major coronary artery. Table 3 shows the prevalence of possible angina and possible myocardial infarction at the initial sur- vey and the relationships between these symptoms and EGC abnormalities, hypertension, and diabetes. Positive responses to the angina questionnaire were received from 45 men (8.5 %) and 55 women (10.2 %). A history of prolonged chest pain suggestive of myocardial infarction was given by 15 men (2.9%) and 14 women (2.6 %), of whom 5 men and 4 women also had angina. These high prevalence rates were surprising in view of the suspected infrequency of ischaemic heart disease. Angina was rarely volun- teered as a significant symptom by those giving unstructured medical histories, but the terrain at Lawrence Tavern is a severe stimulus for provoking the symptom and we show in a subsequent paper that a positive response to the angina questionnaire was of considerable prognostic significance. ECGs with 11-3, 41-3, 51-3, or 71 Minnesota code items were present initially in 33% of men and in 34% of women with a history of either possible infarction or angina, compared with rates of 9.1 % and 14.9%, respectively, in those without such his- tories. Hypertension was present in 35% of men and 45% of women with symptoms and in 19.1 % and 29.4% of those without; diabetes was detected in 4% of men and 8% of women with symptoms and in 1.7% and 2.1 % of those without. In neither sex were ECG abnormalities significantly associated with high cholesterol levels. In women but not in men, cholesterol levels of those with angina were signi- ficantly (p<0.01) higher than in those without the symptom. Of the 45 men with possible angina, 28 were followed-up but only 12 (43%) retained this symp- tom; 13 men responded to the questionnaire posi- tively at the second occasion but not at the first and 3 of them had ECG changes suggesting ischaemia. Of the 55 women, 42 were followed-up and only 17 retained their symptoms (40 %); 23 women responded positively only at the second survey and 4 of them had ECG changes. Subjects who were subsequently followed-up had, at the first survey, lower rates for effort pain than those for the total population, as would be expected if effort pain were associated with higher mortality. At the second survey, however, the prevalence was lower than at the first-6.3% in men and 9.1 % in women-despite the 5-year increase in age. In the interval between surveys only 4 men were thought to have had possible myocardial infarctions. Two died suddenly and no autopsy examination was performed. Two occurred in men who have since died and in whom small artery disease, with macro- scopic evidence of myocardial scarring, was found at autopsy. Two women gave histories suggestive of myocardial infarction in the intersurvey period, one of which was confirmed electrocardiographically. 433 W. E. M1ALL ET AL. 00)DON0O C N 10(1 V-N _~ ON- 0 _- oO N CN) D00) N 10 N cda, N N L _- _6 e CO1 N c- tD O ui 0- C10(1 0D- I*. ( co L00 O) qN 00 - 10 ) 0 (1)C') Nc6 c6 N tD 0-0 am u(1N (1 DCNN ') (1) ) m N dCD m- M Ocow c C') 0O 0 cN 00 N- N - 'O - - - - 0 w- qt 0 CO - W- T- 00 o N 0) N - 0 - - - 0 0 0 40. *lo E uni ,E unW .co LO M Lo 0) O W- Co '-10 CAA NO4 O,Oei c-Jr.Z 99 10L NOO- r - CTN NN 9ON N (1) 010 N N _ON 10 N o N 0 E LI. 0 D6 v- N 10 I- 1 LoO UN 0 N u ai NN m C1 cIN* ood 010o00,00W 0) - co v)-CO e 0) - W1 , (o 0 _r- N cO o - LO 1 w- (1C 6LD ND r- it 0 - 0 - 00 N W N rN * 0 IN Lo 00 - N4- 0 0 Ca CD; CD 0 - O 0. 0 1> ._ o,-1c 0 C aA 434 UZ co 0 E co I- c .2 0> CL 0X 3 0 *I C C0 0 _0O Co 0 0 0.0 CD 00 E 5 00 Q 0 o .C -o 0. o-0 =0@ o) c OC - o 0., .0 I- CI) 0o 0. *o>. LO. -a-2 2 ocoS - 0... *- 0. =C. 0>( 3,- >n: o XEL .1, Ln 0 CJ) =wa- 0> o4..,Q'- >a L I 0 ... itOo - . C- 0st 0 t 'O a r ( U- 0 m CO) 4) co co W -a 0 W W CD ,it CD LO LC) (A co 4) le LC) LC) lqt U) co 4) 1* It LO CV) HEART DISEASE IN JAMAICA. I Table 3. Relationship between symptoms, ECG changes, hypertension, and diabetes, Lawrence Tavern, Jamaica Males Females Agrup Symptoms No. with No. with No. with No. with(yeroup Symptom Total ECG B.P. No. with Total ECG B.P. No. with(years) no. 11-3. 41-3. >160/95 diabetes no. 11-3, 41-3, >160/95 diabetes 51-3, or 7i (mm Hg) 51-3, or 7i (mm Hg) possible Ml 1 - - - 3 - 1 35-44 possible angina 15 4 3 - 12 1 2 both 1 - - - - - - neither 152 7 16 - 166 14 22 2 possible Ml 4 2 2 - 5 - - 1 possible angina 8 1 4 1 18 9 12 - both 2 - - - 3 2 2 1 neither 166 14 29 2 155 24 48 2 possible Ml 5 2 3 1 2 2 - 1 possible angina 17 7 5 - 21 8 12 1 both 2 2 2 - 1 - - 1 neither 154 22 45 6 155 33 70 6 possible Ml 10 4 5 1 10 2 1 2 total possible angina 40 12 i33% 12j35% 1 4% 51 18 34% 26 45% 1 | %both 5 -J % 4 ) 2 2 neither 472 43 (9.1 %) 90 (19.1 %) 8 (1.7%) 476 71 (14.9%) 140 (29.4%) 10 (2.1 %) Intermittent claudication developed in two men, one with angina and one with a past history of pos- sible infarction with ECG evidence of ischaemia. Relationship between ECG abnormalities and arterial pressure Table 4 shows the relationship between arterial pressure and ECG abnormalities at the first survey, by age and sex. The rates for each type of abnorma- lity, Q/QS items with or without other changes, S-T depression and T wave changes alone, T wave in- version alone, T wave flattening alone, and LBBB, were almost uniformly higher in hypertensive than in normotensive subjects in each age group and in both sexes, the overall differences being highly significant as would be expected, but 32 ofthe 61 men (52 %) and 42 of the 93 women (45 %) were not hypertensive at the time. Table 4. Relationship between ECG abnormalities and hypertension at the first survey, Lawrence Tavern, Jamaica Males Females Age Blood pressure With ECG changes With ECG changes (years) (mm Hg) Total 11-3, 41-3, 51-3, Total 1 a-3d4i73 5i 3 no. and 71 no. and_____71 _ No. % No. % 35-44 > 1 60/95 1 9 5 26.3 25 7 28.0 <1 60/95 149 6 4.0 156 8 5.1 45-54 > 1 60/95 35 1 0 28.6 62 1 7 27.4 <160/95 145 7 4.8 119 18 12.7 55-64 > 1 60/95 55 1 4 25.5 82 27 32.9 <160/95 123 19 15.4 97 16 16.5 all ages >160/95 109 29 29.3 169 51 30.2 < 1 60/95 427 32 7.5 372 42 11.3 435 436 W. E. MIALL ET. AL. Table 5. Prevalence of 11-3, 41-3, 51-3, or 71 items according to blood pressure status and the presence or absence of high amplitude R waves (31) by age and sex, Lawrence Tavern, Jamaica: first survey, Prevalence among Prevalence among Prevalence among Prevalence among Age at hypertensives with 3i hypertensives without 31 normotensives with 3i normotensives without 31 first survey Total No. Total No. Total No. Total No. no. positive no. positive no. positive no. positive Males 35-44 7 2 29 1 2 3 25 43 3 7 106 3 2.8 45-54 1 6 5 31 1 9 5 26 37 4 1 1 108 3 2.8 55-64 18. 7 39 37 7 1 9 33 9 27 90 1 0 11.1 total 41 14 34 68 15 22 113 16 14.2 304 16 5.3 Females 35-44 4 2 50 21 5 24 1 0 0 - 146 8 5.5 45-54 11 4 36 51 13 25 5 0 - 114 18 15.8 55-64 26 10 38 56 17 30 11 3 27 86 13 15.0 total 41 16 39 128 35 27.3 26 3 11 346 39 11.3 Relationship between ECG abnormalities and left ventricular hypertrophy The Minnesota code voltage criteria for high amplitude R waves (R amplitude greater than 26 mm in V5 or V6; greater than 20 mm in I, II, III, or avF, or greater than 12 mm in avL) are based on normal values for Americans of European origin. Ethnic differences in R wave amplitudes occur and the Minnesota code criteria (31) may not be specific for left ventricular hypertrophy in other races. Subjects with values exceeding the criteria probably do not all have ventricular hypertrophy but they should include the great majority of those that do have it. For hypertensive and normotensive men and hy- pertensive women, the prevalence of 1,3, 41-3, 51-3, and 7, abnormalities was greater in those with left ventricular hypertrophy (Table 5). Comparison with other studies A survey carried out by this Unit in Guyana pro- vides the only comparable data from the Caribbean area. Ashcroft et al. (1970) examined adults aged 35-54 years from defined populations of Guyanese of African and Indian origin, using the same tech- niques and the same ECG reader as were used in Jamaica. The prevalence of major Q/QS patterns, left axis deviation, S-T depression, and T wave changes did not differ significantly in males or females in the three populations. Minor Q/QS abnormalities were slightly more common in Guyanese men of Indian origin. High amplitude R waves were significantly 1 more common in African than in Indian men in Guyana and significantly more frequent in Jamaican men than in Guyanese men of African origin. Atrio- ventricular conduction defects in men were also significantly more common in Jamaican than in Guyanese men. None of the differences in the pre- valence of high amplitude R waves were statistically significant in women. Table 6 shows comparisons for certain major ECG items between Lawrence Tavern males aged 40-49 and data from Framingham (Higgins et al., 1965) and Tecumseh (Ostrander, 1966) in the USA and from four European studies reported by Rose et al. (1968). The rates for Q/QS abnormalities and left axis deviation are similar in Lawrence Tavern and in the industrialized communities in the USA; the pre- valences of high amplitude R waves (significantly greater in the Jamaican survey than in all the other surveys shown), of S-T depression (significantly greater in Jamaica than in The Hague), and of T wave changes (significantly greater in Jamaica than 'Using the 5 % significance level. HEART DISEASE IN JAMAICA. 1 Table 6. Prevalence of certain ECG abnormalities according to Minnesota Code criteria in Lawrence Tavern, Jamaica and selected US and European surveys: males 40-49 years Jamaica USA Europe ECG items Code Tawven Framingham a Tecumseh b Brussels c The Hague c Naples Moscow cECGitems (177) (779) (542) (631) (377) (412) (169) No. % No. % No. % No. % No. % No. % No. % major Q/QS items i i 1 1 3 2 1 5 -12 2 1.7 8 1.2 2 09 5 1.1 2 0.8 3 1.9 3 1.8 minor Q/QS items 13 1 0.6 1 0.1 6 1.1 8 1.3 8 2.1 8 1.9 8 4.7 left axis deviation 21 6 3.4 23 3.0 22 4.1 high amplitude R waves 3i 53 29.9 51 6.5 9 1.7 10 1.6 18 4.8 14 3.4 29 17.2 ST depression 4i 1 5 2 1 - 1 - 42 5 4.5 7 2.2 4 1.7 6 1.3 3 1.3 9 2.7 1 0.6 43 2 5 3 - 2 1 - Twave inversion 52 1 10.2 8.1 7.7 2.1 1.6 4.6 _ 2.4 52 6 1 1 8 6 - 7 T wave flattening 53 11 52 34 6 6 12 4 LBBBd 7i - - 3 0.4 1 0.2 - - - - 1 0.2 - - a Higgins et al. (1965). b Ostrander (1966). c Rose et al. (1968). d Left bundle branch block. in Brussels or The Hague), are highest in the Jamai- can survey. In comparing the prevalence of some minor ab- normalities in different ethnic groups a possible coding artefact should be considered. For example, a Q wave in avL is codable provided its duration meets certain criteria and the amplitude of the R wave exceeds 3 mm. Similarly T wave inversion (52) and flattening (53) can be coded as abnormal in avL only if theR wave exceeds 5 mm. In a population where QRS amplitudes are unusually large, more abnormalities in avL will be coded merely because the R wave amplitudes meet these criteria, whereas the same Q and T abnormalities might not be cod- able in a population with QRS complexes of smaller amplitude. The variability of responses to the angina question- naire can be compared with that reported by Rose (1968) for a London industrial population of males initially aged 35-59 years. In Rose's study approxi- mately 4p% of those assessed as " angina positive " at the initial examination responded positively at an- nual follow-up surveys. In this Jamaican study the rates were 43% for men and 40% for women, and in Jamaica, as was noted by Rose (1968) for Lon- don, " there was no obvious coincidence in time between the appearance of symptoms and of ECG changes; the two manifestations appeared to come and go independently of one another ". The relationship between ECG changes and hyper- tension can be compared with that found in the Framingham study. Data from the initial examin- ation of the population followed-up for 8 years and described by Higgins et al. (1965) were kindly pro- vided by Dr D. Shurtleff broken down in terms of ECG items 13.3, 41-3, 5,-3, and 71 and blood pressure levels. The proportions of the Framingham and Lawrence Tavern populations with and without hyperten- sion (> 160/95 mm Hg) and with and without any of the above ECG codes are shown, by age, in Fig. 1. Among men the percentage with the specified ECG changes and their apportionment among hypertensive and normotensive subjects was almost identical in the 437 W. E. MLALL ET AL. FRAMINGHAM MalesI INormotensive Hypertensive (>160/95) IE LAWIENCE TAVERN 35-44 years 45-54 55-64 35- 44 years 45-54 55- 64 years 55- 64 Hypertensive (>160/95i)E r I 35- 44 years 45- 54 55- 64 100 80 60 40 20 Percentage 20 40 60 80 100 WHO 20165 Fig. 1. Percentage of persons with and without hypertension and percentage of persons with ECG items 11-3,41-3, 51-3, and 7i (diagonal shading) in Framingham, USA, and Lawrence Tavern, Jamaica, by age and sex. (The oldest age group in Framingham includes subjects aged 55-62 years not 55-64 years.) two studies, though hypertension was somewhat commoner in each age group at Framingham. In women the proportions with ECG changes were higher in Lawrence Tavern in the older age groups, though the prevalence of hypertension was almost identical in the two surveys. In general the relationship between ECG changes and hypertension was closely similar in these two very different populations. DISCUSSION This survey showed that in a representative sample of rural Jamaican adults the prevalence of electro- cardiographic abnormalities suggesting ischaemia was similar to that found in economically more privileged communities where the extent and seve- rity of coronary atheroma, and the incidence of myocardial infarction, are believed to be much higher (Tejada et al., 1968). It also demonstrated signi- ficantly higher rates for the prevalence of A-V con- duction defects and high amplitude R waves, and some evidence that interatrial conduction delay was unusually common and that supraventricular ectopic beats were comparatively frequent. Comparison of the results of such surveys have been somewhat disappointing in relating the pre- valence of electrocardiographic abnormalities to in- ternational differences in the prevalence of coronary artery disease (Higgins et al., 1965; Rose et al., 1968) and it has been suggested that the Minnesota code might be found to be more useful if abnor- malities were described not only by their type but also by their location (Ashcroft et al., 1970). The frequency of ECG abnormalities in different geo- graphical areas does not necessarily reflect accurately the frequency of coronary artery pathology. Damage to the myocardium is revealed by a limited range of abnormal electrocardiographic patterns and it is well known, for example, that unexplained cardiomyo- pathies may present with ECG signs of infarction. This is true of the cardiomyopathy described from &&XXMA 1-xSyeAzes 438 ___1 I I HEART DISEASE IN JAMAICA. I Jamaica by Stuart & Hayes (1963). Since it is un- likely that the similar prevalence ofECG abnormali- ties in this rural population in Jamaica and urban populations in America and Europe indicate equal severities of atheroma of the main coronary arteries, other causes of these ECG changes must contribute to the prevalence in Jamaica and these must be more common than in the communities for which the sur- vey findings have been compared. Valvular heart disease was rarely encountered; seven cases of aortic stenosis or incompetence and two cases of severe mitral regurgitation were detected at the original survey. Two men and two women had an " ischae- mic" type of ECG pattern, which could have re- sulted from severe anaemia. Hypertension is clearly an important cause of abnormal ECG findings in Jamaica as elsewhere, particularly among women. Hypertension is com- mon in this population (Miall et al., 1962) but less common than has been reported in other surveys in the West Indies (Johnson & Remington, 1961; Schneckloth et al., 1962) and considerably less com- mon than reported for negroes of low-income farm- ing families in the USA (Gover, 1948) and for the negro population of Evans County, Ga., USA (McDonough et al., 1964). Hypertension in terms of pressures equalling or exceeding 160/95 mm Hg is not more frequent in Lawrence Tavern than in Framingham (Fig. 1) and its association with ECG changes suggestive of myocardial ischaemia is simi- lar in both populations. A clear relationship was found between hyper- tension and ECG abnormalities but almost half those with each type of " ischaemic " abnormality were not hypertensive. Without employing coronary angiographic techniques it is impossible to differen- tiate between hypertensive heart disease without coronary involvement and coronary heart disease with hypertension. The association of ECG signs with symptoms of effort pain suggests that much of the disease seen is ultimately ischaemic in origin, but the eliciting ofinformation about symptoms from Jamaican coun- try folk is difficult and unreliable. Most of the " angi- na-positive" subjects did not volunteer the infor- mation spontaneously but did so in response to direct questioning. As happens frequently, the symp- toms found in population surveys appear to be rela- tively mild compared with those found in hospital practice and this was true of this type of effort pain, which was rarely classical angina pectoris and was usually only provoked by considerable exertion. Sub- sequent exercise tests on a bicycle ergometer failed to induce angina in several subjects who had pre- viously responded positively to the questionnaire. Those who had given a history of severe chest pain of more than 30 min duration had rarely been hospi- talized; it was unusual for the history to suggest an episode of the severity of a classical myocardial in- farction. Nevertheless the association of prolonged chest pain with discomfort on effort and with ECG changes confirms that these symptoms were also of cardiac origin in many subjects. Though the evidence for their cardiac origin is strong these symptoms show other features that cast doubt on their relationship with occlusive coronary artery disease in the usual sense of the term. The lack of any clear-cut relation- ship between effort pain and age in men (Table 3) was unexpected, as was the lack of association with hypercholesterolaemia. In a subsequent paper we show the relationship between symptoms and signs discovered during the original survey and the subsequent mortality, much of which was the result of cardiovascular disease. Though several subjects who died in the intersurvey period showed moderately advanced atheroma of the main branches of the coronary vessels at autop- sy, there were others whose pathology was largely restricted to the small coronary arteries (Miall et al., 1972a). The possible role of small-artery disease in contributing to the overall prevalence of myocardial disorders is also discussed in a subsequent paper (Miall et al., 1972b). One other cardiac disorder that may contribute to the overall prevalence in Jamaica requires discussion. Stuart & Hayes (1963) described from Jamaica a type of heart disease of unknown etiology with un- explained ventricular hypertrophy, different degrees of interstitial fibrosis, and round cell infiltration. This type of idiopathic cardiomegaly is diagnosed by exclusion of other organic lesions. Macroscopically the enlargement of the heart and the increased car- diac weight are due to hypertrophy and dilatation of both ventricles, with increase in the measurements of both inflow and outflow tracts and usually of both mitral and tricuspid valve rings. Microscopically there is irregular hypertrophy of muscle fibres, which often show bizarre nuclei and a diffuse increase of connective tissue. Focal scarring is nearly always present and often extensive but endocardial changes are not a prominent feature, and the main coronary vessels are usually normal though changes in the smaller vessels have been noted (Bull. Wld Hlth Org., 1968). The pathology of the condition seen in 439 440 W. E. MIALL ET AL. Jamaica has recently been reviewed by Campbell et al. (1971). Cardiomyopathies of this type are now believed to be more widespread than was previously suspected. They are increasingly recognized in temperate as well as in tropical areas as a cause of congestive failure easily confused with that due to coronary athero- sclerosis, but their prevalence is difficult to measure (Brandfonbrener, 1968). Eight men and one woman in the present study were considered to have had this kind of cardiac disorder. The overall pattern of heart disease detected in this Jamaican community cannot be explained by one disorder of overriding importance. The present evi- dence suggests that hypertension, occlusive coronary artery disease, small-artery disease, and unex- plained cardiomyopathy may all be important contri- butors. R1SUMt tTUDE LONGITUDINALE DES CARDIOPATHIES DANS UNE POPULATION RURALE DE LA JAMAYQUE: 1. PREVALENCE ET, EN PARTICULIER, OBSERVATIONS ELECTROCARDIOGRAPHIQUES En 1962/63, on a commence une etude prospective des maladies cardio-vasculaires dans un echantillon repre- sentatif d'une population rurale vivant dans une region montagneuse de la Jama-que. Deux cents personnes de chaque sexe, dans chacune des trois tranches d'age de 10 ans entre 35 et 64 ans, ont ete priees de se soumettre a des investigations cardiologiques dans un h6pital cen- tral; 1 067 d'entre elles (88,9%) ont accepte. Cinq ans plus tard, on a reexamine par les memes techniques 854 de ces sujets representant 90% environ du groupe initial. Les donnees relatives a la prevalence des maladies cardio- vasculaires recueillies au cours de ces enquetes sont ex- posees et comparees aux observations similaires faites ailleurs par d'autres chercheurs. Des etudes anterieures ont montre que l'atherome avec obliteration des arteres coronaires principales est rela- tivement rare chez les Jamaiquains, alors que l'hyper- tension artdrielle est frequente. Dans la presente etude, les traces electrocardiographiques, interpretes selon le Code du Minnesota, ont revele une prevalence des ano- malies des ondes Q/QS similaire a celle existant dans les socidtes occidentales industrialisees, et des taux plus eleves chez les Jamaiquains d'ondes R de grande ampli- tude, d'abaissement de S-T et d'anomalies de l'onde T. Des aspects electrocardiographiques generalement attri- bues a l'ischdmie du myocarde ont ete rencontres aussi souvent que dans les populations oiu les affections coro- nariennes avec obliteration en general, et l'infarctus du myocarde en particulier, sont frequentes. Ces anomalies des traces etaient associees a une forme legere de dou- leur thoracique a l'effort dans une proportion similaire ia celle observee dans les populations predispos6es aux accidents coronariens. Les douleurs thoraciques pro- longees, evoquant la possibilite d'un infarctus mais imposant rarement l'hospitalisation, n'ont pas et excep- tionnelles et se sont aussi accompagnees d'anomalies electrocardiographiques compatibles avec des troubles ischemiques. I1 semble peu probable que ces observations soient le reflet d'une prevalence egale des lesions atheromateuses des principales arteres coronaires chez les Jamaiquains de milieu rural et dans les populations urbaines d'Europe et d'Amerique. On a pu etablir une relation nette entre l'hypertension et les traces caracteristiques de l'ischemie, mais pres de la moitie des cas qui presentaient de tels traces n'etaient pas atteints d'hypertension (pression san- guine egale ou superieure au seuil arbitrairement choisi de 160/95 mm Hg). En outre, la prevalence de l'hyper- tension et la frequence de son association avec les anoma- lies electrocardiographiques etaient tres proches de celles enregistrees lors de l'enquete de Framingham (Etats-Unis d'Amerique). Ces anomalies etaient cependant associees a des signes d'hypertrophie ventriculaire gauche chez des sujets 'a tension normale ou hypertendus et on comptait parmi les cardiopathies decouvertes un certain nombre de cas de cardiomegalie d'origine indeterminee. Les 1lsions valvulaires etaient rares. Ces observations donnent a penser que l'hypertension, l'atheromatose des arteres coronaires, la cardiomyopathie d'origine indeterminee ainsi que peut-etre la (, maladie des petites arteres contribuent toutes a composer le tableau de la morbidit6 cardio-vasculaire a la Jamaique. REFERENCES Ashcroft, M. T. et al. (1970) Bull. Wld Hith Org., 42, 205-223 Blackburn, H. et al. (1960) Circulation, 21, 1160-1175 Brandfonbrener, M. (1968) J. chron. Dis., 21, 463-465 Bull. Wld Hlth Org., 1968, 38, 979-992 Campbell, M.et al. (1971) Brit. Heart J., 33, 193-202 Fodor, J. et al. (1964) Bull. Wld Hlth Org., 31, 321-335 Garrow, J. S. (1963) Lancet, 2, 1205 Goldberg, C. A. (1959) Clin. Chem., 5, 446 Gover, M. (1948) Publ. Hlth Rep. (Wash.), 63, 1083 HEART DISEASE IN JAMAICA. 1 441 Higgins, 1. T. T. et al. (1965) Brit. J. prev. soc. Med., 19, 53-68 Johnson, B. C. & Remington, R. D. (1961) J. chron. Dis., 13, 39-51 Kass, E. H. (1960) Arch. intern. Med., 105, 194-198 Lehmann, H. & Ager, J. A. M. (1954) Assoc. Clin. Path. Broadsheet 27 (New Series) 6 Maclntyre, I. & Ralston, M. (1954) Biochem. J., 56, p. x1iii McDonough, J. R. et al. (1964) Ann. intern. Med., 61, 208-228 McGill, H. C. et al. (1968) Lab. Invest., 18, 498-502 Miall, W. E. et al. (1962) Brit. med. J., 2, 497-506 Miall, W. E. et al. (1967) Brit. J. prev. soc. Med., 21, 45-55 Miall, W. E. et al. (1972a) Bull. Wid Hlth Org., 46, (in press) Miall,W. E. et al. (1972b) Bull. WldHlth Org., 46, (in press) Morris, J. J. et al. (1964) Circulation, 29, 242-251 Ostrander, L. D. (1966) Circulation, 34, 1069-1080 Rose, G. A. (1962) Bull. Wld Hlth Org., 27, 645-658 Rose, G. A. (1968) Brit. J. prev. soc. Med., 22, 12-15 Rose, G. A. & Blackburn, H. (1968) Cardiovascular survey methods, Geneva (World Health Organization: Mono- graph Series, No. 56) Rose, G. A. et al. (1968) Bull. Wid Hith Org., 38, 885-895 Schneckloth, R. E. et al. (1962) Amer. Heart J., 63, 607-628 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 Wright, B. M. & Dore, C. F. (1970) Lancet, 1, 337-338 2