A comparison of serum total cholesterol levels and their determinants between the Federal Republic of Germany and the United States.

Mean serum total cholesterol levels appear to be higher in the Federal Republic of Germany (FRG) than in the United States (US) while coronary heart disease death rates are lower. The study examined possible factors for the difference including possible differences in laboratory methodology. Cross-sectional data from the first two waves of the German National Health Surveys (1984-1986 and 1987-1989; n = 9709) and from the Second National Health and Nutrition Examination Survey (1976-1980; n = 7832) were compared for men and women 25-69 years of age. The influence of age, body mass index, diet, cigarette smoking, education, income, use of oral contraceptives or antihypertensive agents, alcohol consumption and potential differences in laboratory measurement were explored using multiple regression techniques separately for men and women for ages 25-39, 40-59 and 60-69 years of age. Overall ages, unadjusted mean total cholesterol levels were higher in German than US men (6.02 vs. 5.64 mmol/l) and in German than US women (6.04 vs. 5.80 mmol/l) as were HDL cholesterol levels (men: 1.30 vs 1.14 mmol/l; women: 1.65 vs. 1.38 mmol/l). Adjusting for lifestyle factors explained, on the average, 40% of the differences in mean total cholesterol of which half or 20% was accounted for by adjusting for alcohol intake. Adjusting for possible laboratory differences explained, on the average, an additional 30% of the differences. Frequency of alcohol intake was the most important factor in explaining differences in mean HDL cholesterol levels. Adjustment for differences in alcohol intake had negligible effects on reducing the differences in mean non-HDL cholesterol.

Oral arginine does not stimulate basal or augment exercise-induced GH secretion in either young or old adults.

BACKGROUND: Growth hormone (GH) helps maintain body composition and metabolism in adults. However, basal and peak GH decline with age. Exercise produces a physiologic GH response that is subnormal in elderly people. Arginine (Arg) infusion can augment GH secretion, but the efficacy of oral Arg to improve GH response to exercise has not been explored. We investigated whether oral Arg increases GH secretion in young and old people at rest and during exercise. METHODS: Twenty young (Y: 22.1 +/- 0.9 y; SEM) and 8 old (O: 68.5 +/- 2.1 y) male and female subjects underwent three different trials following determination of their one-repetition maximum strength (1-RM); exercise only (EO; 3 sets, 8-10 reps at 85% of 1-RM; on 12 separate resistive lifts), Arg only (5.0 g), or Arg + exercise. Blood samples were collected between successive lifts, and GH (ng x ml(-1)) was determined via RIA. RESULTS: In Y vs O: Basal GH secreted (area under the curve) was 543.6 +/- 84.0 vs 211.5 +/- 63.0. During EO, values were 986.6 +/- 156.6 and 517.8 +/- 85.5. Both were significantly lower in the older individuals (p < .05). Oral Arg alone did not result in any increase in GH secretion at rest (310.8 +/- 73.2 vs 262.9 +/- 141.2). When Arg was coadministered during exercise, GH release was not affected in either the young or old and appeared to be blunted in the young compared to the exercise only trial in the young. CONCLUSION: Based upon these findings, we concluded that oral Arg does not stimulate GH secretion and may impair GH release during resistive exercise.

Omega-3 fatty acids in adipose tissue and risk of myocardial infarction: the EURAMIC study.

Omega-3 fatty acids have potential antiatherogenic, antithrombotic, and antiarrhythmic properties, but their role in coronary heart disease remains controversial. To evaluate the association of omega-3 fatty acids in adipose tissue with the risk of myocardial infarction in men, a case-control study was conducted in eight European countries and Israel. Cases (n=639) included patients with a first myocardial infarction admitted to coronary care units within 24 hours from the onset of symptoms. Controls (n=700) were selected to represent the populations originating the cases. Adipose tissue levels of fatty acids were determined by capillary gas chromatography. The mean (+/-SD) proportion of alpha-linolenic acid was 0.77% (+/-0.19) of fatty acids in cases and 0.80% (+/-0.19) of fatty acids in controls (P=0.01). The relative risk for the highest quintile of alpha-linolenic acid compared with the lowest was 0.42 (95% confidence interval [CI] 0.22 to 0.81, P-trend=0.02). After adjusting for classical risk factors, the relative risk for the highest quintile was 0.68 (95% CI 0.31 to 1.49, P-trend=0.38). The mean proportion of docosahexaenoic acid was 0.24% (+/-0.13) of fatty acids in cases and 0.25% (+/-0.13) of fatty acids in controls (P=0. 14), with no evidence of association with risk of myocardial infarction. In this large case-control study we could not detect a protective effect of docosahexaenoic acid on the risk of myocardial infarction. The protective effect of alpha-linolenic acid was attenuated after adjusting for classical risk factors (mainly smoking), but it deserves further research.

Tissue stores of individual monounsaturated fatty acids and breast cancer: the EURAMIC study. European Community Multicenter Study on Antioxidants, Myocardial Infarction, and Breast Cancer.

The strongest evidence that monunsaturated fat may influence breast cancer risk comes from studies of southern European populations, in whom intake of oleic acid sources, particularly olive oil, appears protective. No previous study has examined the relation of adipose tissue fatty acid content to breast cancer in such a population. We used adipose biopsies with diverse fat intake patterns gathered in 5 European centers, including southern Europe (Malaga, Spain), to test the hypothesis that stores of oleic acid or other monounsaturates are inversely associated with breast cancer. Gluteal fat aspirates were obtained from 291 postmenopausal incident breast cancer patients and 351 control subjects, frequency-matched for age and catchment area. Logistic regression was used to model breast cancer by monounsaturates, with established risk factors controlled for. Oleic acid showed a strong inverse association with breast cancer in the Spanish center. The odds ratio for the difference between 75th and 25th percentiles was 0.40 (95% CI: 0.28, 0.58) in Malaga and 1.27 (0.88, 1.85) in all other centers pooled, with a peak at 2.36 (1.01, 5.50) for Zeist. Palmitoleic and myristoleic acids showed evidence of an inverse association outside Spain, and cis-vaccenic acid showed a positive association in 3 centers. These data do not support the hypothesis that increasing tissue stores of oleic acid are protective against breast cancer in non-Spanish populations. This finding implies that the strong protective associations reported for olive oil intake in dietary studies may be due to some other protective components of the oil and not to the direct effect of oleic acid uptake. Alternatively, high olive oil intake may indicate some other protective aspect of the lifestyle of these women.

Deficient dietary iron intakes among women and children in Russia: evidence from the Russian Longitudinal Monitoring Survey.

OBJECTIVES: This study evaluated the iron sufficiency of the Russian diet. METHODS: Data were obtained from 24-hour dietary recalls conducted in 4 rounds (1992 through 1994) of a nationally representative longitudinal survey of 10,548 women and children. Iron bioavailability was estimated via algorithms adjusting for enhancers (heme, vitamin C) and inhibitors (tannins in tea, phytates in grains) consumed at the same meal. RESULTS: Dietary iron intakes were deficient in the most vulnerable groups: young children and women of reproductive age. Poverty status was strongly associated with deficiency. After adjustment for enhancers and inhibitors, estimated bioavailable iron intakes at 3% to 4% of total iron were inadequate in all women and children. CONCLUSIONS: These dietary data suggest that Russian women and children are at high risk of iron deficiency. Grain products rich in phytates, which inhibit absorption, were the major food source of iron in Russia. High intakes of tea and low consumption of vitamin C also inhibited iron bioavailability. Since changes in eating behavior could potentially double iron bioavailability, educational programs should be explored as a strategy for improving iron nutriture.

Adipose tissue omega-3 and omega-6 fatty acid content and breast cancer in the EURAMIC study. European Community Multicenter Study on Antioxidants, Myocardial Infarction, and Breast Cancer.

The fatty acid content of adipose tissue in postmenopausal breast cancer cases and controls from five European countries in the European Community Multicenter Study on Antioxidants, Myocardial Infarction, and Cancer (EURAMIC) breast cancer study (1991-1992) was used to explore the hypothesis that fatty acids of the omega-3 family inhibit breast cancer and that the degree of inhibition depends on background levels of omega-6 polyunsaturates. Considered in isolation, the level of omega-3 or omega-6 fat in adipose tissue displayed little consistent association with breast cancer across study centers. The ratio of long-chain omega-3 fatty acids to total omega-6 fat showed an inverse association with breast cancer in four of five centers. In Malaga, Spain, the odds ratio for the highest tertile relative to the lowest reached 0.32 (95% confidence interval 0.13-0.82). In this center, total omega-6 fatty acid was strongly associated with breast cancer. With all centers pooled, the odds ratio for long-chain omega-3 to total omega-6 reached 0.80 for the second tertile and 0.65 for the third tertile, a downward trend bordering on statistical significance (p for trend = 0.055). While not definitive, these results provide evidence for the hypothesis that the balance between omega-3 and omega-6 fat may play a role in breast cancer.

Skeletal effects of cyclic recombinant human growth hormone and salmon calcitonin in osteopenic postmenopausal women.

The objectives of this study were to determine whether cyclic administration of recombinant human GH, with or without the antiresorptive agent, salmon calcitonin (CT), provides clinically meaningful increases in bone mineral density (BMD) at the lumbar spine and proximal femur in postmenopausal osteopenic women. The design of the study was a randomized clinical trial consisting of 12 56-day treatment cycles. Each cycle was initiated by a 12-day period of hormone administration, followed by 44 days of supplemental calcium only. Cycles of hormone administration consisted of 7 daily injections of recombinant GH (20 micrograms/kg.day) or its placebo, followed by 5 daily injections of salmon CT (100 U/day) or its placebo. The study was performed at the Palo Alto Veterans Affairs medical center. The patients were 84 healthy women with lumbar spine BMD more than 1 SD below the average value for a healthy 25-yr-old Caucasian woman. BMD was measured at the lumbar spine and proximal femur by dual energy x-ray absorptiometry. Biochemical markers of bone turnover and circulating insulin-like growth factor I were also measured. GH treatment increased insulin-like growth factor I concentrations from low values at baseline (112 +/- 56 ng/mL) to the young normal range (approximately 430 +/- 125 ng/mL). Groups receiving GH plus CT or GH plus placebo increased lumbar spine BMD at 2 yr by 2.70 +/- 0.81% (P < 0.01) and 1.72 +/- 0.74% (P < 0.05; intention to treat analysis). No significant change occurred in women receiving placebo plus CT or combined placebo. Significant increases in total hip BMD of 1-2% were observed for the GH plus placebo and placebo plus CT groups, with a nonsignificant trend in the GH plus CT group. For the femoral trochanter, significant increases were observed in the GH plus CT and placebo plus CT groups only. No significant change in femoral neck BMD was observed in any group. Women taking replacement estrogen had the same BMD response as those who were estrogen deficient. No significant increase in BMD was observed between 24 and 36 months in the 62 women who returned for a 3 yr measurement. In response to GH, short term increases in resorption and formation markers were observed, but these had decreased before the next treatment cycle. No long term changes in resorption markers were observed, but women in the GH groups showed a sustained rise in circulating osteocalcin over the entire 2-yr protocol. GH given cyclically with or without CT for 2 yr achieved statistically significant increases in BMD of the lumbar spine and selected areas of the hip in postmenopausal women. These gains were less marked than those achieved with estrogen or bisphosphonates and were associated with a relatively high incidence of adverse experiences. Therefore, it is unlikely that cyclic GH with or without CT will prove clinically useful in the treatment of postmenopausal women with osteoporosis.

Use of oral contraceptives and serum beta-carotene.

OBJECTIVE: Antioxidants, in particular carotenoids, may influence the risk for cardiovascular disease. This study investigates the influence of oral contraceptives (OC) on the serum concentration of beta-carotene, which may in turn affect the risk of cardiovascular diseases due to its antioxidative impact. DESIGN: Cross-sectional epidemiologic study. Examinations included a detailed questionnaire on medical history and lifestyle factors, a 7 day food record, and blood samples. SETTING: National health and nutrition survey among healthy people living in private homes in West Germany in 1987-1988. SUBJECTS: Nonpregnant and nonlactating women aged 18-44 (n = 610). RESULTS: Overall, the use of OC was negatively associated with serum beta-carotene concentration in bi- and multivariable analyses after adjustment for age, smoking, alcohol consumption, dietary intake of beta-carotene, use of vitamin supplements, body mass index, pregnancies, and serum concentrations of total triglyceride and cholesterol. A strong interaction between OC use and age on beta-carotene concentration was observed. While no relationship between OC use and serum beta-carotene was seen in the youngest age-group (18-24 y), there was a modest but significant negative association between OC use and beta-carotene levels among 25-34 y old women. The use of OC was associated with a strong decrease in beta-carotene levels among 35-44 y old women. The interaction between OC use and age could partly be explained by age dependent use of OC with higher estrogen content. CONCLUSIONS: OC use seems to be strongly related to serum beta-carotene levels, particularly among women above the age of 35. Further studies are needed to clarify the underlying mechanisms of this association and its implications for health risks of OC use.

PIP: Beta-carotene, a provitamin with antioxidant effects, may substantially reduce the risk of coronary artery disease and acute myocardial infarction. A cross-sectional epidemiologic study involving 610 West German women 18-44 years of age indicates that oral contraceptive (OC) use has a negative impact on serum levels of beta-carotene. 195 respondents (32%) were current OC users, 322 (53%) were past users, and 91 (15%) had never used OCs. Median serum beta-carotene levels were significantly lower in current OC users (25.1 mcg/dl) than in past (32.5 mcg/dl) and never users (31.2 mcg/dl). The percentage of women with beta-carotene levels below the desirable value of 21.5 mcg/dl was significantly higher in the current OC use group (34%) than in the 2 other groups combined (21%), yielding an overall odds ratio (OR) of 1.9 (95% confidence interval, 1.3-2.8). The strength of the association between OC use and decreased beta-carotene increased with age (OR of 1.5 for women 18-24 years, 1.9 for those 25-34 years, and 3.4 for 35-44-year old women). In addition, the decrease of beta-carotene was larger for OCs containing 50 mcg of ethinyl estradiol than for low-dose formulations. Also observed were significant associations between serum beta-carotene levels and smoking, alcohol intake, body mass index, triglycerides, nutrient intake, and total cholesterol levels. Although further studies are required to identify the mechanisms underlying the OC-beta-carotene association and define its implications for women's health, OC users should be advised to consume vegetables rich in beta-carotene.

Association between toenail selenium and risk of acute myocardial infarction in European men. The EURAMIC Study. European Antioxidant Myocardial Infarction and Breast Cancer.

The association between selenium status and risk of acute myocardial infarction was examined in a multicenter case-control study in 10 centers from Europe and Israel in 1991-1992. Selenium in toenails was assessed for 683 nonfatal male cases with first acute myocardial infarction and 729 controls less than 70 years of age. Median toenail selenium content was 0.553 microgram/g for cases and 0.590 microgram/g for controls. After adjustment for age, center, and smoking, the odds ratio for myocardial infarction in the highest quintile of selenium as compared with the lowest was 0.63 (95 percent confidence interval 0.37-1.07, p for trend = 0.08). The observed inverse trend was somewhat stronger when the authors adjusted for vitamin E status (p = 0.05). Analysis stratified for smoking habits showed an inverse association in former smokers (odds ratio for the 75th-25th percentile contrast = 0.63 (95 percent confidence interval 0.43-0.94)), but not in current smokers (odds ratio = 0.97 (0.71-1.32)) or in those who had never smoked (odds ratio = 1.55 (0.87-2.76)). Analysis stratified by center showed a significant inverse association between selenium levels and risk of myocardial infarction for Germany (Berlin) only (75th to 25th percentile odds ratio = 0.62 (95 percent confidence interval 0.42-0.91)), which was the center with the lowest selenium levels. It appears that the increased risk of acute myocardial infarction at low levels of selenium intake is largely explained by cigarette smoking; selenium status does not appear to be an important determinant of risk of myocardial infarction at the levels observed in a large part of Europe.

Tea and cancer prevention: an evaluation of the epidemiologic literature.

Animal and in vitro studies provide evidence of an anticarcinogenic potential of active ingredients in teas. This review encompasses epidemiologic studies of stomach, colon, and lung cancer as well as the evidence of a relationship between tea drinking and cancer at large in humans. Cohort studies do not suggest a protective role for tea drinking in the total risk of cancer. Site-specific studies reveal a more complex picture. The epidemiologic studies on tea drinking and stomach cancer do not justify claims of a cancer-protective effect. A protective effect of green tea on the development of colon cancer is suggested. The evidence regarding black tea is less clear, with some indication of a risk of colon or rectal cancer associated with regular use of black tea. The studies on tea and lung cancer also suggest an increased risk with increased tea consumption. The range and crude categorization of tea consumption, choice of control groups, and inadequate control for confounding might have obscured possible relationships. From the limited studies that suggest a favorable effect from tea, it is likely that benefits are restricted to high intakes in high-risk populations.

Tissue antioxidants and postmenopausal breast cancer: the European Community Multicentre Study on Antioxidants, Myocardial Infarction, and Cancer of the Breast (EURAMIC).

Antioxidants may protect against free radical mediated carcinogenesis. Epidemiological studies have not confirmed this hypothesis for breast cancer, possibly because of methodological limitations. Time-integrated exposure of alpha-tocopherol and beta-carotene in adipose tissue, and selenium in toenails was investigated in a case-control study among postmenopausal women, ages 50-74 years, from five European countries. The study group comprised 347 incident breast cancer cases and 374 controls. Mean antioxidant levels, adjusted for age and center, did not significantly differ for alpha-tocopherol (cases were 4.5% higher than controls), beta-carotene (3.0% lower), or selenium (1.8% lower). Odds ratios for highest versus lowest tertiles of exposure, adjusted for potential confounders, were 1.15 (95% confidence interval, 0.75-1.77), 0.74 (0.45-1.23), and 0.96 (0.63-1.47), respectively, without evidence for a decreasing trend. No statistically significant interactions were observed. Moreover, a provisional antioxidant score, indicating whether concentrations were above the median for zero, one, two, or all three antioxidants, yielded odds ratios of 1.00 (reference; all below median), 1.58, 1.58, and 1.21, respectively (chi2 for association = 4.00; P = 0.26). These results do not support the hypothesis that antioxidants are important determinants of this hormone-related malignancy among postmenopausal women.

Association between beta-carotene and acute myocardial infarction depends on polyunsaturated fatty acid status. The EURAMIC Study. European Study on Antioxidants, Myocardial Infarction, and Cancer of the Breast.

Because antioxidants may play a role in the prevention of coronary heart disease by inhibiting the peroxidation of polyunsaturated fatty acids (PUFAs), the combined association of diet-derived antioxidants and PUFAs with acute myocardial infarction (MI) was investigated. This multicenter case-control study included 674 patients and 725 control subjects in eight European countries and Israel. Fatty acid composition and alpha-tocopherol and beta-carotene levels were determined in adipose tissue; selenium level was determined in toenails. For alpha-tocopherol no association with MI was observed at any PUFA level. The overall multivariate odds ratio (OR) for low (10th percentile) versus high (90th percentile) beta-carotene was 1.98 (95% confidence interval [CI], 1.39 to 2.82). The strength of this inverse association with MI was dependent on PUFA levels (in tertiles): for low PUFA, the OR for low versus high beta-carotene was 1.79 (95% CI, 0.98 to 3.25), for medium PUFA the OR was 1.76 (95% CI, 1.00 to 3.11), and for high PUFA 3.47 (95% CI, 1.93 to 6.24). For selenium increased risk was observed only at the lowest PUFA tertile (OR, 2.49; 95% CI, 1.22 to 5.09). This interaction between selenium and PUFAs was not significant and may at least partly be explained by a higher proportion of smokers at the low PUFA level. These findings support the hypothesis that beta-carotene plays a role in the protection of PUFAs against oxidation and subsequently in the protection against MI. No evidence was found that alpha-tocopherol or selenium may protect against MI at any level of PUFA intake.

Antioxidants in adipose tissue and risk of myocardial infarction: the EURAMIC Study.

Laboratory and epidemiological studies suggest that the antioxidants, vitamin E and beta-carotene, protect against coronary heart disease. In a European multicentre case-control study alpha-tocopherol and beta-carotene concentrations were measured in adipose-tissue samples collected in 1991-92 from 683 people with acute myocardial infarction and 727 controls. Mean adipose-tissue beta-carotene concentration was 0.35 microgram/g in cases and 0.42 in controls, with age-adjusted and centre-adjusted mean difference 0.07 microgram/g (95% confidence interval [CI] 0.04-0.10). Mean alpha-tocopherol concentrations were 193 micrograms/g and 192 micrograms/g for cases and controls, respectively. The age-adjusted and centre-adjusted odds ratio for risk of myocardial infarction in the lowest quintile of beta-carotene as compared with the highest was 2.62 (95% CI 1.79-3.83). Additional control for body-mass index and smoking reduced the odds ratio to 1.78 (95% CI 1.17-2.71); other established risk factors did not substantially alter this ratio. The increased risk was mainly confined to current smokers: the multivariate odds ratio in the lowest beta-carotene quintile in smokers was 2.39 (95% CI 1.35-4.25), whereas it was 1.07 for people who had never smoked. A low alpha-tocopherol concentration was not associated with risk of myocardial infarction. Our results support the hypothesis that high beta-carotene concentrations within the normal range reduce the risk of a first myocardial infarction. The findings for alpha-tocopherol are compatible with previous observations of reduced risk among vitamin E supplement users only. The consumption of beta-carotene-rich foods such as carrots and green-leaf vegetables may reduce the risk of myocardial infarction.

EURAMIC Study: antioxidants, myocardial infarction and breast cancer. Design and main hypotheses.

Epidemiological studies have not given sufficient evidence yet for the role of antioxidant nutrients in the prevention of cardiovascular disease. As regards cancer, an inverse association between beta-carotene intake and specific types of cancer, especially lung cancer, has been shown. For other cancer sites and other antioxidants, the association is less clear. The EURAMIC Study, an EC Concerted Action, is a case-control study conducted in 11 countries, in which the combined effect of vitamin E, beta-carotene and selenium, in relation to fatty acid intake, will be examined. The disease endpoints are acute myocardial infarction and early-stage breast cancer. The broad range of antioxidant intake, the use of biomarkers of exposure, and the analysis of pooled data will allow an estimate of the strength of the putative beneficial effect. In this paper the background and design of the study will be introduced.

Lipoprotein metabolism and coffee intake--who is at risk?

Data from a representative health and nutrition survey of German adults (sample of 1073 women and 806 men) were used to investigate the relationship between coffee consumption and the concentration of cholesterol in serum as well as other lipoprotein constituents. For these outcome variables multivariate analyses were conducted separately for men and women. Differences in age, body mass index, smoking habits, use of oral contraceptives, physical activity, alcohol, fish, fat, milk and tea consumption were controlled for in the models. Interactions between coffee drinking behavior and smoking habits as well as between coffee and the use of oral contraceptives in their relationship with serum cholesterol were of special interest in the analyses. Higher coffee intake (> 400 ml/d) showed higher total cholesterol, LDL cholesterol and lower triglyceride rich lipoprotein (TRL) and triglyceride concentrations in serum compared to lower intake (< 200 ml/d). Smoking appeared to be an aggravating factor in these relationships. Results of the linear regression analysis demonstrated an increase of 1.66 mg/dL LDL-C per cup of coffee daily consumed for men and of 1.58 mg/dL for women. The combination of high coffee intake, smoking and no oral contraceptive use ever was associated with the highest total and LDL-C and lowest TRL concentrations in this population. The observed differences may be explained by an increase of lipoprotein lipase activity due to coffee consumption.

The relationship between coffee consumption and serum cholesterol under consideration of smoking history.

A large German population sample of 6820 men and 7258 women was used to investigate the relationship between coffee consumption, total serum cholesterol and HDL cholesterol. Analyses were conducted on men and women separately. Differences in age, body mass index, diastolic blood pressure, smoking habits, alcohol, fish, milk and tea consumption, physical activity and medication use were controlled for in the analyses. Interactions between coffee consumption and smoking habits in their relationship with serum cholesterol were part of the analyses. For men, a positive relationship between coffee consumption and total serum cholesterol was found among smokers and life-long abstainers but not in the group of ex-smokers. In women a relationship between coffee consumption and total serum cholesterol was also present, but very weak (only statistical significant in covariance analyses). Levels of HDL cholesterol did not correlate with coffee consumption in either men and women. Hypotheses concerning the peculiarities of the group of ex-smokers are developed, and supporting empirical evidence is given. It is suggested that the group of ex-smokers should always be analysed separately.

The relationship between coffee consumption and lipid levels in young and older people in the Heidelberg-Michelstadt-Berlin study.

The relationship between coffee consumption and serum lipid levels was studied in a randomly selected sample of 395 young and 385 elderly adults in the Federal Republic of Germany. Analyses were done separately for men and women and for young and older people. After adjustment for body mass index, activity level, smoking, total energy intake, dietary fat, fish, milk, tea and alcohol consumption and oral contraceptive use a statistically significant increase in total serum- and LDL-cholesterol levels with increased coffee consumption was observed in the group of young men only. In this subgroup a difference of 0.11 mmol.l-1 serum cholesterol and of 0.10 mmol.l-1 LDL cholesterol for each additional cup of coffee consumed daily was calculated. This is one of the few studies that includes elderly people. No significant trend was seen between coffee consumption and serum- or LDL-cholesterol levels in the elderly.