We investigated the relationship between antihypertensive drug treatment of hypertensives and their mean serum lipid concentrations in population based studies in Germany. Data from three surveys (Luebeck Blood Pressure Study (LBS) of 1984, MONICA Augsburg Survey I of 1984/85, MONICA Augsburg Survey II of 1989/90), obtained on random samples of the population aged 25-64 years, were used for cross-sectional analyses. Moreover, prospective analyses were carried out on participants of the MONICA Augsburg cohort study of 1987/88 (3-year-follow-up of the MONICA Survey I). Blood pressure, non-fasting serum total cholesterol and HDL-cholesterol, and body height and weight were measured under strictly standardized conditions. Interview data were available on medical history including medication use, and on smoking and alcohol consumption. In cross-sectional and prospective analyses treated male and female hypertensives in each population had significantly lower crude mean HDL-C concentrations than untreated hypertensives, borderliners, or normotensives. Differences in mean HDL-C between untreated and treated hypertensives were attenuated but still significant after control of confounders and ranged from 1.8 to 6.1 mg/dl (i.e. in relative terms, -3.4 to -12.9%) in men and from 3.6 to 9.4 mg/dl (-5.7 to -14.9%) in women. By contrast, crude and multivariate associations of antihypertensive treatment with non-HDL-C (total minus HDL-C) levels were inconsistent and not significant. The inverse association of drug therapy with HDL-C was confirmed by prospective analyses in the MONICA cohort study supporting a causal relationship. Treatment patterns in a community (prevalence of prescribed drug classes) correlated with the magnitude and significance of HDL-C effects.(ABSTRACT TRUNCATED AT 250 WORDS)