- Email: [email protected]
HDL levels in pacific islanders
251
Atherosclerosis, 40 (1981) 257-262 Elsevier/North-Holland Scientific Publishers, Ltd.
HDL LEVELS IN PACIFIC ISLANDERS
P.J. NESTEL and P. ZIMMET Baker Medical Research Institute, Alfred Hospital, and Department of Metabolic Medicine and Epidemiology, Royal Southern Memorial Hospital, Melbourne (Australia) (Received 27 October, 1980) (Revised, received 23 March, 1981) (Accepted 30 March, 1981)
Summary Although the concentration of high density lipoprotein (HDL) cholesterol is a sensitive index of coronary heart disease (CHD) risk in Caucasians there is little evidence for differences in HDL levels between populations with low and high prevalence of CHD. This study compares the concentrations of apolipoprotein AI, the major HDL protein, in South Pacific Islanders of different ethnic origin. In two rural populations on Ouvea, the mean AI levels were 94 and 91 mg/dl, respectively, for Melanesian and Polynesian men. These values are substantially lower than in a group of Australian men (120 rt 20 mg/dl). A further comparison between rural and urbanized Polynesian males in Western Samoa showed that urbanization led to higher AI levels (117 mg/dl in the town versus 94 mg/dl in the village). A survey of factors that influence HDL levels suggests that the higher AI levels were due to changes in diet, including fat and alcohol consumption. The prevalence of probable CHD in Western Samoan men aged 40-69, determined by electrocardiography, was about half that reported in an Australian survey. Thus the lower AI levels in the Western Samoan men appeared not to be related to the prevalence of CHD which was less than among Australians who have higher HDL concentrations. Key words:
Alcohol - Apolipoproteins - High density lipoprotein ease -Pacific islanders - Urbanization
- Zschaemic heart dis-
The study was supported by gmntd fkom the National Health and Medhl Reeeerch Coumil of iwetralia, and National In&t&es of Health (ROI AM 2644841). One of the author8 (P.2.) was short-kxn conmltmt for the World HePIth Organization (Weetern Samoa) and the South Pacific Commission (New Caledonia). The support of these two orzanhations is tcratefully acknowledged.
0021-9150/81/000~000/$02.50
0 1981 Elsevier/North-Holland
Scientific Publishers, Ltd.
258
Introduction In affluent populations, characterized by high rates of coronary heart disease (CHD), the high density lipoprotein (HDL) cholesterol concentration is currently the most sensitive lipid index of future clinical CHD in middle-aged and older individuals [ 11. The independent inverse correlation between HDL cholesterol and CHD has now been observed in every prospective study in which this has been examined [ 2,3]. There are as yet no conclusive comparisons of HDL cholesterol concentrations between populations with low and high rates of CHD, to provide the kind of evidence that established total serum cholesterol as a risk factor. The available information is limited and inconsistent with reports of HDL cholesterol that are both higher (in Jamaicans [4] and black South Africans [ 51) and lower (in Mexican Indians [ 61) than are found in average Caucasian populations. Furthermore, vegetarians who by virtue of their very low serum cholesterol levels have a theoretically lowered risk, nevertheless, have low levels of HDL cholesterol [7]. The present comparisons in South Pacific Islanders of ethnically diverse populations and of a rural and an urbanized community provided opportunities to explore several aspects of the regulation of HDL in populations with a presumed low incidence of CHD. Methods Subjects studied Epidemiological surveys were carried out in the Pacific countries of Western Samoa and New Caledonia to determine prevalence rates for ischaemic heart disease, diabetes mellitus, hypertension, obesity and hyperlipidaemia. Western Samoa was chosen because it contains the world’s largest population of Polynesians, 90% of the 150 000 inhabitants being full-blooded Polynesians. New Caledonia is ethnically more diverse, about 85% of the population being Melanesian, and the remainder Polynesian and of mixed origin. The New Caledonian survey was carried out on the island of Ouvea because it has retained traditional Melanesian living habits. The surveys were conducted during 1978. In each proposed survey area a house-to-house census established the age and sex distributions of residents 20 years and older. The purpose of the surveys was explained and appointments given to a representative subsample. Between 900 and 1000 people responded from each of the designated rural and urban areas; this represented only about 80% of those selected but the age and sex distribution of the responders was similar to that in the surveyed population and in a national census of those areas carried out two years previously. Subjects were asked to fast overnight and blood samples were obtained in the early morning. A 12-lead electrocardiogram was performed in subjects over the age of 30. Further details have been published [ 81. (1) Western Samoa: Comparison of rural and urban populations A subsample of males with age range 20-78 was selected for the measurements of plasma AI apolipoprotein from rural [ 911 and urban [ SS] inhabitants. This was approximately 10% of the total sample and was chosen so that it
NI SNOLLWLLN33N03
IV NI3&O?IdOdI’IOdV
Vl4SV’Id
aNV
SNOI,LV’IfldOd WIOJ 3HJs SA33pBLlS .!I0 S31&SIl3&3VZIVH3
aJO3a.Iaq$aM'[6] XOUaqU8DIJIU@S Ol!JaafqnSaq if8U.I S~~UOLU ~ChWS 203 3J)Z-98 uazoq pamy 8tuyd 30 saIdun_wy ~om~sa~oya ?aH 30 t~uauxamS8a~
C)lUyla
OM$
ay!J U~M!)XJ
&h?$T.U!ULW$~
%X09
W3q
S8tj WBq$
‘JiWMOH
?J%UO~
aq
0g maqfjpuno3ptmsuo~~8rrS~sap asarl) payaayr,umq aqq y?yb %~~o~(zaq~a~g -~*s~)~~@o~odorqy8 w mytati~od IO ysau8IaK m sail8m ay? pq8u@sap spuq8yu! p01 aqcj‘8aAng q vuoy48mo p3yotuazaa uo Qt.10 aA8ysaAIama~ &pm yunrp s! IO~ODIV -aAymdxa fiaA Byaq spoo3 pa?rodq - p.1ogp8.q -Fwopard SF $alp pu8 maw83 am~~sqns am SapuI pu81~ ay& ‘sqrpp~p8mya ?#J!aq PUN v@!aM PUN UownqwSFp a% u! PaIqmaSaJ Law YWM aldun= mo? at# 30 0~0~ ')noqo ‘~ualuaztmmu yqordodrlod8 IV 8umId aw 303 pqoalas amhi S38a6 gL-_oz pa&2 uatu u8jsauQod ~6 pm3 uysa~a~ aA~3-iQq2Q
suojyqndod
U?X.U ayl ‘8rdv y SqnD !JI.@UpU8 SIqOq ‘SaXOqSh3UI aq!)q aIq8pA8 ifIaa$ s! ~oyocyj~*suo!$8dnmo L.m$uapaS y a%Bua uam aqt)30 !)so&q*a+ pu8 .nBns anol~ ‘(salqqa8aA ‘14x3‘78am) spooB paumm q$ptAlaip nay? I)uaaraIddnsIlr@ -qg MOU SUBOUT&’ U.K+3aM U8qJU aI&T, ‘SUO~S8%0 ~UOUI&KW UO ~1~l.U l_fUIXQ SI ‘8~83 ‘a%aTaAaq~oyoc~p~puoy~ aq& *stroyytyd a* y moqq pnu8w uj a%B%ua Qxf8ur uaux au *loyoc~~q t-m‘paumu03 am spoo3 wa)sa~ atuos ~nq~uo~~~~~rmupuopazd ms t+ larp aw ‘Seaa p3.n~~ aq UI *aIdum a.uqua
a~JoS3~~38~2113IrI%!aqpurrq~adhlCpoq~~
uovqiwm
ah
aw
pa73aWal
69Z
260
chose to measure the plasma apolipoprotein AI concentrations as an index of HDL since values for AI correlate highly with HDL cholesterol. Apolipoprotein AI concentrations were quantified by electroimmunoassay [lo]. Results The mean values (*SD) for the 4 communities are shown in Table 1. The apolipoprotein-AI concentrations were similar in the Melanesians and Polynesians of New Caledonia and the rural population in Western Samoa. The urbanized community in Western Samoa showed a significantly higher mean AI concentration than was found in the rural populations (P < 0.001). In fact, the mean and range of plasma apolipoprotein concentrations in urbanized Samoans resembled that of healthy men of similar age in Melbourne (120 + 20) [ 111. Discussion The mean values for apolipoprotein-AI in the 3 rural populations were below the mean + 1 SD for healthy white men in Melbourne. Prevalence figures for ischaemic heart disease are not available on our Melbourne males. However, they are available for a comparable Australian population in the town of Busselton, Western Australia [ 121. The prevalence of probable CHD was determined in the Western Samoan males according to the Minnesota coding, the criteria used in Busselton [ 121 and Tecumseh, U.S.A. [ 131. The prevalence was clearly lower in the Western Samoans (unpublished): in men aged 40-69, the prevalence in Busselton was 9% whereas in Samoan men aged over 40 years the prevalence was 4.5%. The rate in Tecumseh is similar to that in Busselton. It is therefore highly probable that middle-aged Western Samoans have a much lower prevalence of CHD than do Australian men of similar age. Connor et al. [6] have also reported low HDL cholesterol levels in the Tarahumara Indians of Mexico in whom CHD is rare. This does not cast doubt on HDL cholesterol as an excellent index of future CHD in affluent urbanized societies, but draws attention to the large number of factors that may determine HDL levels, one of which may eventually be shown to be responsible for the link between HDL and CHD. The main finding was the rise in the apolipoprotein-AI concentration with urbanization in Western Samoa. In terms of factors known at present to influence HDL, differences in diet especially carbohydrates [ 141, physical exertion [ 151, alcohol consumption [ 161, body weight [ 171 and cigarette smoking [ 181 need to be considered. The higher consumption of carbohydrates [8,19] as vegetables and cereals (yams, sweet potato, taro and rice) and the lesser intake of alcohol [19] in the rural community favour the observed difference in apolipoprotein-AI levels; with urbanization the Samoans are known to move away from traditional food patterns to greater consumption of fat, meat, sugar and alcohol [8]. We have reported similar dietary trends in Micronesians in Nauru [20]. Smoking of cigarettes was equally common in rural and urban Samoans (Table 1). On the other hand, the differences in physical exertion [8] (less in the town, at least from a consideration of changes in occupation), and overweight (Ta-
261
ble 1: more common in the town), would favour higher HDL levels in the rural groups. This apparent inconsistency suggests either that the dietary differences, including alcohol, are of greater consequence or that additional unidentified factors are responsible. In favour of the former is the finding of similar low apolipoprotein-AI levels in urbanized Caucasians who are vegetarians and also abstain from alcohol (Nestel and Billington, unpublished). Finally, the similar apolipoprotein-AI concentrations in ruraI Melanesians and rural Polynesians in New Caledonia is interesting in view of the clear differences in the prevalence of diabetes among them, being more frequent among the Polynesians [ 211. Other reports in Caucasians have shown lower HDL cholesterol concentrations in diabetics [ 221. Acknowledgements The authors thank MS M. O’Connor and MS J. Ma for technical assistance and MS S. Whitehouse and Dr L. Jackson for data analysis. References 1 Gordon, T., Caste& W.P.. I-Ii&tland. M.C.. KannelI, W.B. and Dawber. T.R., High density lipoprotein as a protective factor against coronary heart disease, Amer. J. Med., 62 (1977) 707.
2 Rhoads, G.G.. GuIbrandsen. C.L. and Kagan, A., Serum lipoproteins and coronary heart disease in a popidation study of Hawaii Japanese men, N. Eng. J. Med., 294 (1976) 293. 3 Miller. N.E., TheUe. D.S.. Forde, O.H. and Mjds. O.D., The TromsG Heart Study -High density Iipoprotein and coronary heart disease, Lance& i (1977) 966. 4 MiRer, G.J.. Miller. N.E. and Ashcroft. MT., Inverse relationship in Jamaica between plasma high density Iipoprotein cholesterol concentration and coronary disease risk as predicted by multiple risk factor status, CIin. Sci. Mol. Med., 61 (1976) 475. 5 Walker. A.R.P. and Walker, B.F., High-density-Itpoprotein cholesterol in African children and adults in a population free of coronary heart disease, Brit. Med. J.. II (1978) 1336. 6 Connor, W.E.. Cerqueira. M.T., Connor. R.W., WaBace. R.B.. Malinow. MR. and Casdorph. H.R.. The plasma Iipids, Iipoproteins. and diet of the Tarahumara Indians of Mexico, Amer. J. Clin. Nutr., 31 (1978) 1131. 7 Sacks, F.M., Caste& W.P., Dormer. A. and Kass, E.H., Plasma lipids and lipoproteins in vegetarians and controls. N. EngI. J. Med., 292 (1976) 1148. 8 Zhnmet, P.. Faaivso, 8. Ainvu, J.. Whitehouse. S.. Mihre. B. and De Boer, W., The prevalence of diabe-
tes in the rural and urban Polynesian population of Western Samoa, Diabetes, 30 (1981) 45. 9 Report of the High Density Lipoprotein Methodology Workshop, (N.I.H. Publication. NO. 79). U.S. Dept. H.E.W.. 1979.~. 1661. 10 Curry, M.D., Alaupovic, P. and Suer&m, C.A.. Determination of apolfpoprotein A and its constitutive AI and AII polypeptides by separate electroimmunoassay, Ciin. Chem., 22 (1976) 316. 11 Fidge. N.H., Nestel. P.J.. Ishikawa. T.. Reardon. M. and BiBington. T., Turnover of apoproteins AI and AH of high density lipoprotein and the re.Iationahip to other lipoproteins in normal and hyperlipidemic individuals, Metabolism, 29 (1980) 643. 12 Welbom. T.A.. Cumpston. G.N., CuBen. K.J.. Cumow. D.H.. McCall. M.G. and Stenhouae, N.S.. The prevaIence of coronary heart disease and associated factors in an AustraIian rural community, Amer. J. Epidem., 80 (1969) 521. 13 Epstein, F.H.. Ostrander. Jr., L.D.. Johnson, B.C., Payne. M.W.. Hayner, N.S.. Keller, J.B. and Francis, T., Epidemiologicical studies of cardiovascular diseases in a total community - Tecumseh. Michigan. AM. Intern. Med., 62 (1966) 1170. 14 Blum. C.B.. Levy, R.I.. Eisenberg, 5.. HalI. M.. Goebel. R.H. and Berman, M., High density lipoprotein metabolism in man. J. CIin. Invest., 60 (1977) 796. 15 Wood, P.D., HaskeB, W., Mem. H.. Lewis, 5.. Etem. M.P. and Farquhar. J.W.. The distribution of plasma Iipoproteins in middleage maie runners, Metabolism, 25 (1976) 1249. 16 CasteUi. W.P.. Doyle, J.T.. Gordon, T.. Hames, C.G.. IiidrtIand. M.C.. HuBey, S.B.. Kagan. A. and Zukel, W.J.. Alcohol and biood Iipids. Lance& ii (1977) 153. 17 Cadson. L.A. and Ericsson. M.. Quantitative and qualitative serum lipoprotein anaI~si% Part 1 (8tudies in healthy men and women). Atherosclerosis, 21 (1976) 417.
13 Garrison. R.J.. Kannel. W.B., Feinleib. M.. Cartel& W.P.. McNamara. P.M. and Padgett. S.J., Cigarette smoking and HDL cholesterol, Atherosclerode. 30 (1978) 17. 19 Zbnmet, P., Cardiovascular and metabolic diseases in Western Samoa. World Health Organization, Western Pacific Regional Office, Manila, 1930. 20 Ringrose. H. and Zimmet. P., Nutrient intakes in an urbanized Micronesian population with a high diabetes prevalence, Amer. J. Clin. Nutr., 32 (1979) 1334. 21 Zimmet, P., Canteloube, D.. LeGonidec. G., Couzigou. P.. Pegbini. M., Bennett, P. and Kuberski, T., Diabetes. 29 (1930) Supplement 2. Abstract 266. 22 Lopes-Virella, M.F.L., Stone, P. and Colwell, J.A., Serum high density lipoprotein in diabetic patients. Diabetologia. 13 (1977) 286.
Atherosclerosis, 40 (1981) 257-262 Elsevier/North-Holland Scientific Publishers, Ltd.
HDL LEVELS IN PACIFIC ISLANDERS
P.J. NESTEL and P. ZIMMET Baker Medical Research Institute, Alfred Hospital, and Department of Metabolic Medicine and Epidemiology, Royal Southern Memorial Hospital, Melbourne (Australia) (Received 27 October, 1980) (Revised, received 23 March, 1981) (Accepted 30 March, 1981)
Summary Although the concentration of high density lipoprotein (HDL) cholesterol is a sensitive index of coronary heart disease (CHD) risk in Caucasians there is little evidence for differences in HDL levels between populations with low and high prevalence of CHD. This study compares the concentrations of apolipoprotein AI, the major HDL protein, in South Pacific Islanders of different ethnic origin. In two rural populations on Ouvea, the mean AI levels were 94 and 91 mg/dl, respectively, for Melanesian and Polynesian men. These values are substantially lower than in a group of Australian men (120 rt 20 mg/dl). A further comparison between rural and urbanized Polynesian males in Western Samoa showed that urbanization led to higher AI levels (117 mg/dl in the town versus 94 mg/dl in the village). A survey of factors that influence HDL levels suggests that the higher AI levels were due to changes in diet, including fat and alcohol consumption. The prevalence of probable CHD in Western Samoan men aged 40-69, determined by electrocardiography, was about half that reported in an Australian survey. Thus the lower AI levels in the Western Samoan men appeared not to be related to the prevalence of CHD which was less than among Australians who have higher HDL concentrations. Key words:
Alcohol - Apolipoproteins - High density lipoprotein ease -Pacific islanders - Urbanization
- Zschaemic heart dis-
The study was supported by gmntd fkom the National Health and Medhl Reeeerch Coumil of iwetralia, and National In&t&es of Health (ROI AM 2644841). One of the author8 (P.2.) was short-kxn conmltmt for the World HePIth Organization (Weetern Samoa) and the South Pacific Commission (New Caledonia). The support of these two orzanhations is tcratefully acknowledged.
0021-9150/81/000~000/$02.50
0 1981 Elsevier/North-Holland
Scientific Publishers, Ltd.
258
Introduction In affluent populations, characterized by high rates of coronary heart disease (CHD), the high density lipoprotein (HDL) cholesterol concentration is currently the most sensitive lipid index of future clinical CHD in middle-aged and older individuals [ 11. The independent inverse correlation between HDL cholesterol and CHD has now been observed in every prospective study in which this has been examined [ 2,3]. There are as yet no conclusive comparisons of HDL cholesterol concentrations between populations with low and high rates of CHD, to provide the kind of evidence that established total serum cholesterol as a risk factor. The available information is limited and inconsistent with reports of HDL cholesterol that are both higher (in Jamaicans [4] and black South Africans [ 51) and lower (in Mexican Indians [ 61) than are found in average Caucasian populations. Furthermore, vegetarians who by virtue of their very low serum cholesterol levels have a theoretically lowered risk, nevertheless, have low levels of HDL cholesterol [7]. The present comparisons in South Pacific Islanders of ethnically diverse populations and of a rural and an urbanized community provided opportunities to explore several aspects of the regulation of HDL in populations with a presumed low incidence of CHD. Methods Subjects studied Epidemiological surveys were carried out in the Pacific countries of Western Samoa and New Caledonia to determine prevalence rates for ischaemic heart disease, diabetes mellitus, hypertension, obesity and hyperlipidaemia. Western Samoa was chosen because it contains the world’s largest population of Polynesians, 90% of the 150 000 inhabitants being full-blooded Polynesians. New Caledonia is ethnically more diverse, about 85% of the population being Melanesian, and the remainder Polynesian and of mixed origin. The New Caledonian survey was carried out on the island of Ouvea because it has retained traditional Melanesian living habits. The surveys were conducted during 1978. In each proposed survey area a house-to-house census established the age and sex distributions of residents 20 years and older. The purpose of the surveys was explained and appointments given to a representative subsample. Between 900 and 1000 people responded from each of the designated rural and urban areas; this represented only about 80% of those selected but the age and sex distribution of the responders was similar to that in the surveyed population and in a national census of those areas carried out two years previously. Subjects were asked to fast overnight and blood samples were obtained in the early morning. A 12-lead electrocardiogram was performed in subjects over the age of 30. Further details have been published [ 81. (1) Western Samoa: Comparison of rural and urban populations A subsample of males with age range 20-78 was selected for the measurements of plasma AI apolipoprotein from rural [ 911 and urban [ SS] inhabitants. This was approximately 10% of the total sample and was chosen so that it
NI SNOLLWLLN33N03
IV NI3&O?IdOdI’IOdV
Vl4SV’Id
aNV
SNOI,LV’IfldOd WIOJ 3HJs SA33pBLlS .!I0 S31&SIl3&3VZIVH3
aJO3a.Iaq$aM'[6] XOUaqU8DIJIU@S Ol!JaafqnSaq if8U.I S~~UOLU ~ChWS 203 3J)Z-98 uazoq pamy 8tuyd 30 saIdun_wy ~om~sa~oya ?aH 30 t~uauxamS8a~
C)lUyla
OM$
ay!J U~M!)XJ
&h?$T.U!ULW$~
%X09
W3q
S8tj WBq$
‘JiWMOH
?J%UO~
aq
0g maqfjpuno3ptmsuo~~8rrS~sap asarl) payaayr,umq aqq y?yb %~~o~(zaq~a~g -~*s~)~~@o~odorqy8 w mytati~od IO ysau8IaK m sail8m ay? pq8u@sap spuq8yu! p01 aqcj‘8aAng q vuoy48mo p3yotuazaa uo Qt.10 aA8ysaAIama~ &pm yunrp s! IO~ODIV -aAymdxa fiaA Byaq spoo3 pa?rodq - p.1ogp8.q -Fwopard SF $alp pu8 maw83 am~~sqns am SapuI pu81~ ay& ‘sqrpp~p8mya ?#J!aq PUN v@!aM PUN UownqwSFp a% u! PaIqmaSaJ Law YWM aldun= mo? at# 30 0~0~ ')noqo ‘~ualuaztmmu yqordodrlod8 IV 8umId aw 303 pqoalas amhi S38a6 gL-_oz pa&2 uatu u8jsauQod ~6 pm3 uysa~a~ aA~3-iQq2Q
suojyqndod
U?X.U ayl ‘8rdv y SqnD !JI.@UpU8 SIqOq ‘SaXOqSh3UI aq!)q aIq8pA8 ifIaa$ s! ~oyocyj~*suo!$8dnmo L.m$uapaS y a%Bua uam aqt)30 !)so&q*a+ pu8 .nBns anol~ ‘(salqqa8aA ‘14x3‘78am) spooB paumm q$ptAlaip nay? I)uaaraIddnsIlr@ -qg MOU SUBOUT&’ U.K+3aM U8qJU aI&T, ‘SUO~S8%0 ~UOUI&KW UO ~1~l.U l_fUIXQ SI ‘8~83 ‘a%aTaAaq~oyoc~p~puoy~ aq& *stroyytyd a* y moqq pnu8w uj a%B%ua Qxf8ur uaux au *loyoc~~q t-m‘paumu03 am spoo3 wa)sa~ atuos ~nq~uo~~~~~rmupuopazd ms t+ larp aw ‘Seaa p3.n~~ aq UI *aIdum a.uqua
a~JoS3~~38~2113IrI%!aqpurrq~adhlCpoq~~
uovqiwm
ah
aw
pa73aWal
69Z
260
chose to measure the plasma apolipoprotein AI concentrations as an index of HDL since values for AI correlate highly with HDL cholesterol. Apolipoprotein AI concentrations were quantified by electroimmunoassay [lo]. Results The mean values (*SD) for the 4 communities are shown in Table 1. The apolipoprotein-AI concentrations were similar in the Melanesians and Polynesians of New Caledonia and the rural population in Western Samoa. The urbanized community in Western Samoa showed a significantly higher mean AI concentration than was found in the rural populations (P < 0.001). In fact, the mean and range of plasma apolipoprotein concentrations in urbanized Samoans resembled that of healthy men of similar age in Melbourne (120 + 20) [ 111. Discussion The mean values for apolipoprotein-AI in the 3 rural populations were below the mean + 1 SD for healthy white men in Melbourne. Prevalence figures for ischaemic heart disease are not available on our Melbourne males. However, they are available for a comparable Australian population in the town of Busselton, Western Australia [ 121. The prevalence of probable CHD was determined in the Western Samoan males according to the Minnesota coding, the criteria used in Busselton [ 121 and Tecumseh, U.S.A. [ 131. The prevalence was clearly lower in the Western Samoans (unpublished): in men aged 40-69, the prevalence in Busselton was 9% whereas in Samoan men aged over 40 years the prevalence was 4.5%. The rate in Tecumseh is similar to that in Busselton. It is therefore highly probable that middle-aged Western Samoans have a much lower prevalence of CHD than do Australian men of similar age. Connor et al. [6] have also reported low HDL cholesterol levels in the Tarahumara Indians of Mexico in whom CHD is rare. This does not cast doubt on HDL cholesterol as an excellent index of future CHD in affluent urbanized societies, but draws attention to the large number of factors that may determine HDL levels, one of which may eventually be shown to be responsible for the link between HDL and CHD. The main finding was the rise in the apolipoprotein-AI concentration with urbanization in Western Samoa. In terms of factors known at present to influence HDL, differences in diet especially carbohydrates [ 141, physical exertion [ 151, alcohol consumption [ 161, body weight [ 171 and cigarette smoking [ 181 need to be considered. The higher consumption of carbohydrates [8,19] as vegetables and cereals (yams, sweet potato, taro and rice) and the lesser intake of alcohol [19] in the rural community favour the observed difference in apolipoprotein-AI levels; with urbanization the Samoans are known to move away from traditional food patterns to greater consumption of fat, meat, sugar and alcohol [8]. We have reported similar dietary trends in Micronesians in Nauru [20]. Smoking of cigarettes was equally common in rural and urban Samoans (Table 1). On the other hand, the differences in physical exertion [8] (less in the town, at least from a consideration of changes in occupation), and overweight (Ta-
261
ble 1: more common in the town), would favour higher HDL levels in the rural groups. This apparent inconsistency suggests either that the dietary differences, including alcohol, are of greater consequence or that additional unidentified factors are responsible. In favour of the former is the finding of similar low apolipoprotein-AI levels in urbanized Caucasians who are vegetarians and also abstain from alcohol (Nestel and Billington, unpublished). Finally, the similar apolipoprotein-AI concentrations in ruraI Melanesians and rural Polynesians in New Caledonia is interesting in view of the clear differences in the prevalence of diabetes among them, being more frequent among the Polynesians [ 211. Other reports in Caucasians have shown lower HDL cholesterol concentrations in diabetics [ 221. Acknowledgements The authors thank MS M. O’Connor and MS J. Ma for technical assistance and MS S. Whitehouse and Dr L. Jackson for data analysis. References 1 Gordon, T., Caste& W.P.. I-Ii&tland. M.C.. KannelI, W.B. and Dawber. T.R., High density lipoprotein as a protective factor against coronary heart disease, Amer. J. Med., 62 (1977) 707.
2 Rhoads, G.G.. GuIbrandsen. C.L. and Kagan, A., Serum lipoproteins and coronary heart disease in a popidation study of Hawaii Japanese men, N. Eng. J. Med., 294 (1976) 293. 3 Miller. N.E., TheUe. D.S.. Forde, O.H. and Mjds. O.D., The TromsG Heart Study -High density Iipoprotein and coronary heart disease, Lance& i (1977) 966. 4 MiRer, G.J.. Miller. N.E. and Ashcroft. MT., Inverse relationship in Jamaica between plasma high density Iipoprotein cholesterol concentration and coronary disease risk as predicted by multiple risk factor status, CIin. Sci. Mol. Med., 61 (1976) 475. 5 Walker. A.R.P. and Walker, B.F., High-density-Itpoprotein cholesterol in African children and adults in a population free of coronary heart disease, Brit. Med. J.. II (1978) 1336. 6 Connor, W.E.. Cerqueira. M.T., Connor. R.W., WaBace. R.B.. Malinow. MR. and Casdorph. H.R.. The plasma Iipids, Iipoproteins. and diet of the Tarahumara Indians of Mexico, Amer. J. Clin. Nutr., 31 (1978) 1131. 7 Sacks, F.M., Caste& W.P., Dormer. A. and Kass, E.H., Plasma lipids and lipoproteins in vegetarians and controls. N. EngI. J. Med., 292 (1976) 1148. 8 Zhnmet, P.. Faaivso, 8. Ainvu, J.. Whitehouse. S.. Mihre. B. and De Boer, W., The prevalence of diabe-
tes in the rural and urban Polynesian population of Western Samoa, Diabetes, 30 (1981) 45. 9 Report of the High Density Lipoprotein Methodology Workshop, (N.I.H. Publication. NO. 79). U.S. Dept. H.E.W.. 1979.~. 1661. 10 Curry, M.D., Alaupovic, P. and Suer&m, C.A.. Determination of apolfpoprotein A and its constitutive AI and AII polypeptides by separate electroimmunoassay, Ciin. Chem., 22 (1976) 316. 11 Fidge. N.H., Nestel. P.J.. Ishikawa. T.. Reardon. M. and BiBington. T., Turnover of apoproteins AI and AH of high density lipoprotein and the re.Iationahip to other lipoproteins in normal and hyperlipidemic individuals, Metabolism, 29 (1980) 643. 12 Welbom. T.A.. Cumpston. G.N., CuBen. K.J.. Cumow. D.H.. McCall. M.G. and Stenhouae, N.S.. The prevaIence of coronary heart disease and associated factors in an AustraIian rural community, Amer. J. Epidem., 80 (1969) 521. 13 Epstein, F.H.. Ostrander. Jr., L.D.. Johnson, B.C., Payne. M.W.. Hayner, N.S.. Keller, J.B. and Francis, T., Epidemiologicical studies of cardiovascular diseases in a total community - Tecumseh. Michigan. AM. Intern. Med., 62 (1966) 1170. 14 Blum. C.B.. Levy, R.I.. Eisenberg, 5.. HalI. M.. Goebel. R.H. and Berman, M., High density lipoprotein metabolism in man. J. CIin. Invest., 60 (1977) 796. 15 Wood, P.D., HaskeB, W., Mem. H.. Lewis, 5.. Etem. M.P. and Farquhar. J.W.. The distribution of plasma Iipoproteins in middleage maie runners, Metabolism, 25 (1976) 1249. 16 CasteUi. W.P.. Doyle, J.T.. Gordon, T.. Hames, C.G.. IiidrtIand. M.C.. HuBey, S.B.. Kagan. A. and Zukel, W.J.. Alcohol and biood Iipids. Lance& ii (1977) 153. 17 Cadson. L.A. and Ericsson. M.. Quantitative and qualitative serum lipoprotein anaI~si% Part 1 (8tudies in healthy men and women). Atherosclerosis, 21 (1976) 417.
13 Garrison. R.J.. Kannel. W.B., Feinleib. M.. Cartel& W.P.. McNamara. P.M. and Padgett. S.J., Cigarette smoking and HDL cholesterol, Atherosclerode. 30 (1978) 17. 19 Zbnmet, P., Cardiovascular and metabolic diseases in Western Samoa. World Health Organization, Western Pacific Regional Office, Manila, 1930. 20 Ringrose. H. and Zimmet. P., Nutrient intakes in an urbanized Micronesian population with a high diabetes prevalence, Amer. J. Clin. Nutr., 32 (1979) 1334. 21 Zimmet, P., Canteloube, D.. LeGonidec. G., Couzigou. P.. Pegbini. M., Bennett, P. and Kuberski, T., Diabetes. 29 (1930) Supplement 2. Abstract 266. 22 Lopes-Virella, M.F.L., Stone, P. and Colwell, J.A., Serum high density lipoprotein in diabetic patients. Diabetologia. 13 (1977) 286.