It is well established that serum total and high-density lipoprotein (HDL) cholesterol are important predictors of coronary heart disease mortality in middle-aged men. 1–4 Whether this relation holds in the elderly is not clear. 5–11 It is also important to know whether the relation between serum cholesterol and coronary heart disease mortality in elderly men differs between countries. After 25 years of follow-up, the Seven Countries Study showed that the strength of the relation between serum total cholesterol and coronary heart disease mortality is similar in middle-aged men across cultures. 4 The absolute risks, however, were markedly different. Similar information on HDL cholesterol is lacking because at the beginning of the Seven Countries Study, HDL cholesterol was not measured. Information on both total and HDL cholesterol was collected in the Finland, Italy and the Netherlands Elderly (FINE) Study. This study is an extension of the Seven Countries Study and has been carried out since 1984 in Finland, the Netherlands and Italy. The FINE Study provides the possibility to investigate the relations between serum total and HDL cholesterol and 10-year coronary heart disease mortality in elderly men, using time-dependent covariates.
Subjects and Methods
The FINE Study is an extension of the Seven Countries Study. 12 The survivors of the cohorts in Finland, the Netherlands and Italy were re-examined after 25 years. This examination was the baseline of the FINE Study, a prospective study in elderly men ages 65–84 years. The FINE Study includes the two Finnish cohorts (both rural) known as East Finland and West Finland, the Dutch cohort from the small town of Zutphen, and the two rural cohorts in Italy from the villages of Crevalcore and Montegiorgio (in northern and central Italy, respectively).
In Finland, 716 men were re-examined in 1984, and in rural Italy, 682 men in 1985. In the Netherlands, 380 survivors were re-examined in 1985, together with a new random sample consisting of 507 men ages 65–84 years in Zutphen who did not participate earlier in the Zutphen Study. The response rate was 94% in Finland, 74% in the Netherlands, and 76% in Italy.
Subjects with incomplete data for the main cardiovascular risk factors were excluded. The data of 668 Finnish men, 824 Dutch men, and 640 Italian men were used in the present analyses.
All men were examined according to the international protocol used in previous surveys of the Seven Countries Study. 12 Survivors were re-examined after 5 and 10 years of follow-up.
In Finland, fasting blood samples were taken, and nonfasting blood samples were taken in the Netherlands and Italy. In each country, total and HDL cholesterol determinations were carried out in lipid laboratories standardized according to the criteria of the World Health Organization Lipid Reference Laboratories in Prague, Czechoslovakia, or Atlanta, GA. 13 In all three laboratories, serum total cholesterol was determined enzymatically with the CHOD-PAP mono-test kit of Boehringer Mannheim (Mannheim, Germany). 14,15 HDL cholesterol was determined after precipitation of the apolipoprotein B containing lipoproteins with dextran-magnesium chloride in Finland, dextran-magnesium sulfate in the Netherlands, and magnesium phosphotungstate in Italy. 16,17 Hypercholesterolemia was defined as a serum total cholesterol concentration of 6.5 mmol/liter (251 mg/dL) or more. A low-HDL cholesterol level was defined as HDL cholesterol less than 0.9 mmol/liter (35 mg/dL). 18
Height and weight were measured with the subjects in light clothing without shoes, and body mass index was calculated (kg/m2). Blood pressure was measured twice on the right arm with the men in a supine position using a standard sphygmomanometer in Finland and Italy and a random-zero sphygmomanometer in the Netherlands. Systolic and diastolic blood pressure were recorded at the onset of the first and fifth Korotkoff phase, respectively.
Smoking classifications were never-smoker, ex-smoker, and current smoker. Alcohol intake was measured as number of drinks per day and subsequently converted to grams of alcohol by multiplying each glass of beer, wine, or spirit with its alcohol content. 13
The history of coronary heart disease was defined by the Rose questionnaire, 19 combined with information from reported clinical records and additional questions from the examining physician. Coronary heart disease was considered to be present when either myocardial infarction (definite) or angina pectoris (definite) was diagnosed.
Complete follow-up information after 10 years, obtained through official death certificates, was available for 99.7% of the population of the FINE Study. In Finland, only information on causes of death was available. In the Netherlands and Italy, causes of death were validated through review of clinical records. Final causes of death were adjudicated by a single reviewer using the 9th revision of the World Health Organization’s International Classification of Diseases,20 adopting a hierarchical order when multiple causes were given, as follows: violent causes, cancer, coronary heart disease, stroke, and other. Coronary heart disease as the primary cause of death was classified as 410–414 according to the manual.
We computed age-standardized 10-year mortality rates of coronary heart disease by weighting the mortality rates for 5-year age categories to the age distribution of the total study population. We used Cox proportional hazard (survival) analysis, pooled after stratification by cohort, to investigate the relation between serum total and HDL cholesterol and coronary heart disease mortality during 10 years of follow-up. We included serum total and HDL cholesterol and all of the covariables as time-dependent covariates. The baseline measurement and the measurement after 5 years of follow-up were both used. For those who died during the first 5 years of follow-up, only the baseline measurement was used. For those who died during the second 5 years of follow-up, the baseline measurement was used for the first 5-year period, and the measurement after 5 years of follow-up was used for the second 5-year period. Relative risks (RRs) are presented for each 1.00 mmol/liter increase in total cholesterol and 0.10 mmol/liter increase in HDL cholesterol, and for tertiles, the lowest tertile being the referent category. In all analyses, adjustment was made for age (years), body mass index (kg/m2), systolic blood pressure (mmHg), cigarette smoking (ex-smoker and never- or current smoker), and history of coronary heart disease. For HDL cholesterol, adjustment was also made for alcohol intake (0, 1–19, 20–39, and ≥40 gm/day). Analyses were carried out with and without respondents with a history of coronary heart disease, but because these results did not differ by much, the results reported in this article are based on all respondents. We evaluated interaction by comparing the risk in the group with both an unfavorable cholesterol and covariate level with the risk expected from the additive effects of cholesterol and the covariate alone. 21
Mean total cholesterol levels were similar in Finland and the Netherlands but lower in Italy in both men with and without coronary heart disease mortality (Table 1). The mean prevalence of hypercholesterolemia (total cholesterol ≥6.5 mmol/liter) was 37% in Finland, 36% in the Netherlands, and 26% in Italy. Mean HDL cholesterol levels were different among the three countries, with the lowest level in the Netherlands and the highest level in Italy in both men with and without coronary heart disease mortality. The mean prevalence of an HDL cholesterol level less than 0.9 mmol/liter was 14% in Finland, 23% in the Netherlands, and 8% in Italy.
The age-adjusted coronary heart disease mortality rates were much higher in Finland than in the Netherlands and Italy (Figure 1). The number of men who died from coronary heart disease after 10 years of follow-up was 140 in Finland (21%), 88 in the Netherlands (11%), and 48 in Italy (8%). For total cholesterol, the age-adjusted coronary heart disease mortality rate in the lowest tertile in Finland was more than 1.5 times higher than that in the highest tertile in the Netherlands and Italy. For HDL cholesterol, the age-adjusted coronary heart disease mortality rate in Finland and the Netherlands decreased with, respectively, 12 and 3 percentage points from the lowest to the highest tertile, in contrast to an increase in the coronary heart disease mortality rate in Italy, with about 4 percentage points.
The risk of mortality from coronary heart disease increased with each 1.00 mmol/liter increase of total cholesterol, with the highest RRs in the Netherlands (Table 2). In the Netherlands and in Italy, the RR in the highest total cholesterol tertile was about two times that in the lowest tertile, and there was a trend across the total cholesterol tertiles. In Finland there was no trend across the tertiles.
In Finland, the risk of mortality from coronary heart disease decreased with an increase in HDL cholesterol (Table 3). In the Netherlands and Italy, HDL cholesterol was not associated with mortality from coronary heart disease.
In Italy, the risk for coronary heart disease mortality in men with both a low HDL cholesterol level and a high body mass index was lower than expected from the additive effects, which turned out to be dependent on the level of alcohol intake: there was an inverse association between HDL cholesterol and coronary heart disease mortality in lean men (body mass index <25 kg/m2) with no or moderate alcohol intake (<40 gm/day) [RR = 0.76; 95% confidence interval (CI) = 0.59–0.97] and a positive association in overweight men (body mass index ≥25 kg/m2) with a high alcohol intake (≥40 gm/day) (RR = 1.25; 95% CI = 1.07–1.46). The mean HDL concentration was 1.29 mmol/liter in both subgroups. The total cholesterol concentrations were 5.88 mmol/liter in lean men with no or moderate alcohol intake and 6.01 mmol/liter in overweight men with a high alcohol intake. In the other two subgroups, lean men with a high alcohol intake and overweight men with no or moderate alcohol intake, a weak association was found with coronary heart disease mortality [RR = 0.90 (95% CI = 0.72–1.13) and RR = 0.96 (95% CI = 0.81–1.14), respectively]. Also in Finland and the Netherlands, there was an inverse, but less strong, relation between HDL cholesterol and coronary heart disease mortality [RR = 0.93 (95% CI = 0.85–1.02) and RR = 0.99 (95% CI = 0.85–1.15), respectively] in lean men with no or moderate alcohol intake. In the total population of the FINE Study, the RR in lean men with no or moderate alcohol intake was 0.92 (95% CI = 0.86–0.99).
The results of the present study show that serum total cholesterol is positively associated with 10-year coronary heart disease mortality in elderly men in Finland, the Netherlands, and Italy. HDL cholesterol is inversely associated with 10-year coronary heart disease mortality in Finnish men only.
Nonresponse could have influenced the results of our study, especially in the Netherlands and Italy. In Zutphen, the Netherlands, however, there was no difference in average serum total cholesterol level in 1960 between respondents of the 25-year follow-up survey and nonrespondents, men who did not participate in the 25-year follow-up survey. Similar results were found in Montegiorgio, Italy. In Crevalcore, Italy, the respondents had in 1960 a lower total cholesterol level (0.16 mmol/liter) compared with the nonrespondents. The difference in average serum total cholesterol level in 1960 between the respondents in the three countries together and the nonrespondents was 0.08 mmol/liter. It is therefore unlikely that our results are strongly influenced by nonresponse.
Estimation of absolute risks is important, because decisions to treat high-risk persons are based on absolute risks, which increase with age. For total cholesterol, the mortality rates for coronary heart disease showed large differences among the countries, with the highest rates in Finland, intermediate rates in the Netherlands, and the lowest rates in Italy. In the middle total cholesterol tertile for example, the absolute risk in Finland is almost 1.5 times higher than that in the Netherlands and 2.5 times that in Italy. This finding is in accordance with previous reports from the same countries in middle-aged men. 4,22
We found a positive association in this study between serum total cholesterol and coronary heart disease mortality. The results using time-dependent covariates did not differ much from those using only the baseline cholesterol measurement for each respondent (results not shown). Usually the relation found using time-dependent covariates is stronger than that using the baseline measurement only, because with longer follow-up a risk factor loses its predictive value. 23 The more or less similar RRs found in the present study for the two methods could be explained by the relatively short follow-up period of 10 years.
Other epidemiologic studies among elderly men have so far given equivocal results concerning the relation between serum total cholesterol and coronary heart disease mortality. 5–11 A possible explanation for this result could be the age at which the cholesterol level of the respondents was measured. Our hypothesis is that in the studies that did not find a relation between total cholesterol and coronary heart disease mortality, the respondents were older at the time of the measurements (>70 years) compared with the studies that did find a positive relation (>60 or 65 years). Also, the present study provides some evidence for this hypothesis. Our data showed that the association between total cholesterol and coronary heart disease mortality in the total FINE population in the first 5 years of follow-up (men ages 65–84 years) was stronger [RR = 1.21 (95% CI = 1.06–1.38) per 1.00 mmol/liter increase] than in the second 5 years of follow-up (men ages 70–89 years) [RR = 1.13 (95% CI = 0.94–1.36) per 1.00 mmol/liter increase]. The Whitehall study also observed a weaker association between total cholesterol and coronary heart disease mortality for men whose cholesterol concentration was measured at older age. 24 In the Seven Countries Study, the overall estimate of the predictive power of serum total cholesterol decreased from 1.25 (95% CI = 1.19–1.35, per 1.00 mmol/liter increase) in middle-aged men to 1.17 (95% CI = 1.06–1.29) in the elderly in the present analyses. 4 The results of the present study, however, indicate that serum total cholesterol remains a predictor of coronary heart disease mortality in men ages 65–84 years in different countries.
For HDL cholesterol, we found a slightly different relation with coronary heart disease mortality among the three countries. In Finland, we found an inverse association, and in the Netherlands and Italy, no association. For HDL cholesterol, we found somewhat lower RRs using time-dependent covariates compared with using only the baseline measurement (results not shown). Also in other studies, inconsistent results were noted for the association between HDL cholesterol and coronary heart disease mortality. 7,9,11
Overall, we found little relation between HDL cholesterol and coronary heart disease mortality in Italy. Exploring these differences, we found that there was an inverse association in Italy in lean men (body mass index <25 kg/m2) with no or moderate alcohol intake (<40 gm/day), but a positive association in overweight men (≥25 kg/m2) with a high alcohol intake (≥40 gm/day). These findings could be a chance result owing to the small numbers in the subgroups.
A biological explanation for the interaction among HDL cholesterol, body mass index, and alcohol might be that a potentially favorable effect of high-HDL cholesterol on coronary heart disease mortality may be counterbalanced by a high body mass index in combination with a high alcohol intake. Being overweight reduces HDL cholesterol through a high concentration of triglycerides and an increased catabolism of HDL by an excess of adipose tissue. 25 This effect could counterbalance the mean HDL cholesterol-increasing effect of alcohol. This hypothesis is supported by our data. The mean HDL cholesterol concentration in overweight men with a high alcohol intake (1.29 mmol/liter) was comparable with that in lean men with no or moderate alcohol intake (1.29 mmol/liter). Nevertheless, the total cholesterol concentrations were different between both groups. This finding indicates a disturbance of the lipid metabolism in overweight men with a high alcohol intake.
Using the combined data of Finland, the Netherlands, and Italy, we observed an inverse relation between HDL cholesterol and coronary heart disease mortality in lean men with no or moderate alcohol intake. Thus, in a particular subgroup of the elderly population, there may be an inverse association between HDL cholesterol and coronary heart disease mortality. We could not investigate this association in overweight men with a high alcohol intake in Finland and the Netherlands, because in both Finland and the Netherlands high alcohol intake (≥40 gm/day) is rare. We know of no other study that has reported on this interaction. A prospective follow-up study in middle-aged Russian men also found no relation between HDL cholesterol and coronary heart disease mortality, a result that could be due to the high intake of alcohol in these men. 26
We are indebted to the many people involved in this longitudinal study, including all participants and the fieldwork teams in Finland, Italy, and the Netherlands.
1. Anderson KM, Castelli WP, Levy D. Cholesterol and mortality: 30 years of follow-up from the Framingham study. JAMA 1987; 257:2176–2180.
2. Gordon DJ, Probstfield JL, Garrison RJ, Neaton JD, Castelli WP, Knoke JD, Jacobs DR, Bangdiwala S, Tyroler A. High-density lipoprotein cholesterol and cardiovascular disease: four prospective American studies. Circulation 1989; 79:8–15.
3. Neaton JD, Blackburn H, Jacobs D, Kuller L, Lee D-J, Sherwin R, Shih J, Stamler J, Wentworth D. Serum cholesterol level and mortality findings for men screened in the Multiple Risk Factor Intervention Trial. Arch Intern Med 1992; 152:1490–1500.
4. Verschuren WMM, Jacobs DR, Bloemberg BPM, Kromhout D, Menotti A, Aravanis C, Blackburn H, Buzina R, Dontas AS, Fidanza F, Karvonen MJ, Nedeljkovic S, Nissinen A, Toshima H. Serum total cholesterol and long-term coronary heart disease mortality in different cultures: twenty-five-year follow-up of the Seven Countries Study. JAMA 1995; 274:131–136.
5. Benfante R, Reed D. Is elevated serum cholesterol level a risk factor for coronary heart disease in the elderly? JAMA 1990; 263:393–396.
6. Rubin SM, Sidney S, Black DM, Browner WS, Hulley SB, Cummings SR. High blood cholesterol in elderly men and the excess risk for coronary heart disease. Ann Intern Med 1990; 113:916–920.
7. Manolio TA, Pearson TA, Wenger NK, Barrett-Connor E, Payne GH, Harlan WR. Cholesterol and heart disease in older persons and women: review of an NHLBI workshop. Ann Epidemiol 1992; 2:161–176.
8. Kronmal RA, Cain KC, Ye Z, Omenn GS. Total serum cholesterol levels and mortality risk as a function of age: a report based on the Framingham data. Arch Intern Med 1993; 153:1065–1073.
9. Corti MC, Guralnik JM, Salive ME, Harris T, Field TS, Wallace RB, Berkman LF, Seeman TE, Glynn RJ, Hennekens CH, Havlik RJ. HDL cholesterol predicts coronary heart disease mortality in older persons. JAMA 1995; 274:539–544.
10. Tervahauta M, Pekkanen J, Nissinen A. Risk factors of coronary heart disease and total mortality among elderly men with and without preexisting coronary heart disease: Finnish cohorts of the Seven Countries Study. J Am Coll Cardiol 1995; 26:1623–1629.
11. Weijenberg MP, Feskens EJM, Kromhout D. Total and high density lipoprotein cholesterol as risk factors for coronary heart disease in elderly men during 5 years of follow-up: the Zutphen Elderly Study. Am J Epidemiol 1996; 143:151–158.
12. Keys A, Aravanis C, Blackburn HW, Van Buchem FSP, Buzina R, Djordjevic BS, Dontas AS, Fidanza F, Karvonen MJ, Kimura N, Lekos D, Monti M, Puddu V, Taylor HL. Epidemiological studies related to coronary heart disease: characteristics of men aged 40–59 in seven countries. Acta Med Scand 1967; 460suppl:1–392.
13. Kromhout D, Nissinen A, Menotti A, Bloemberg BPM, Pekkanen J, Giampaoli S. Total and HDL cholesterol and their correlates in elderly men in Finland, Italy, and The Netherlands. Am J Epidemiol 1990; 131:855–863.
14. Siedel J, Schlumberger H, Klose S, Ziegenhorn J, Wahlefeld AW. Improved reagent for the enzymatic determination of serum cholesterol. J Clin Chem Clin Biochem 1981; 19:838–839.
15. Stähler F, Gruber W, Stinshoff K, Röschlau P. Eine praxisgerichte enzymatische cholesterinbestimmung. Med Lab 1977; 30:29–37.
16. Warnick GR, Benderson J, Albers JJ. Dextran sulfate-Mg2+
precipitation procedure for quantitation of high density lipoprotein cholesterol. Clin Chem 1982; 28:1379–1388.
17. Morisi G, Macchia T, Angelico F, Pacioni F, Zucca A. Determinazione automatica di trigliceridi, colesterolo, glucosio, acido urico: prospettive d’impiego in screening di medicina preventiva. Ann Ist Super Sanita 1979; 15:239–261.
18. European Atherosclerosis Society. Strategies for the prevention of coronary heart disease: a policy statement of the European Atherosclerosis Society. Eur Heart J 1987; 8:77–88.
19. Rose GA, Blackburn H. Cardiovascular Survey Methods. Geneva: World Health Organization, 1968.
20. World Health Organization. International Classification of Diseases, 9th revision. Geneva: World Health Organization, 1975.
21. Rothman KJ, Greenland S. Modern Epidemiology. 2nd ed. Philadelphia: Lippincott-Raven, 1998.
22. Keys A. Coronary heart disease in seven countries. Circulation 1970; 419(suppl I):I1–I211.
23. Cupples LA, D’Agostino RB, Anderson K, Kannel WB. Comparison of baseline and repeated measure covariate techniques in the Framingham Heart Study. Stat Med 1988; 7:205–218.
24. Shipley MJ, Pocock SJ, Marmot MG. Does plasma cholesterol concentration predict mortality from coronary heart disease in elderly people? 18 year follow up in Whitehall study. BMJ 1991; 303:89–92.
25. Grundy SM. Cholesterol and Atherosclerosis: Diagnosis and Treatment. Philadelphia: JB Lippincott, 1990.
26. Perova NV, Oganov RG, Williams DH, Irving SH, Abernathy JR, Deev AD, Shestov DB, Zhukovsky GS, Davis CE, Tyroler HA. Association of high-density-lipoprotein cholesterol with mortality and other risk factors for major chronic noncommunicable diseases in samples of US and Russian men. Ann Epidemiol 1995; 5:179–185.
Keywords:© 2000 Lippincott Williams & Wilkins, Inc.
cholesterol; coronary heart disease; elderly; cohort study; body mass index; alcohol; international comparisons