Earlier epidemiologic studies have suggested an inverse association between non-Hodgkin's lymphoma and exposure to histamine(2) (H(2)) blockers, nonsteroidal anti-inflammatory drugs, cholesterol-lowering drugs, and antibiotics. Data from the PHARMO database were used to conduct a nested, population-based case-control study that included 1985-1998 drug-dispensing records for 300,000 residents of six Dutch cities. Included were those subjects without a previous history of cancer who were aged >/=20 years and were registered with an incident primary discharge diagnosis of non-Hodgkin's lymphoma between 1991 and 1998. This paper includes data on 211 cases and 800 controls individually matched on sex, age, community pharmacy, calendar time, and duration of follow-up. Conditional logistic regression analysis was used to evaluate the association between non-Hodgkin's lymphoma and categories of cumulative drug use in days. In multivariate analyses, nonsignificant risk reductions were found for all drugs tested, and the negative association tended to increase with increasing duration of use. For women, the odds ratio for H(2) blockers was 0.29 (95% confidence interval: 0.12, 0.69) and for analgesics was 0.40 (95% confidence interval: 0.22, 0.71). Results support an inverse association between occurrence of non-Hodgkin's lymphoma and use of H(2) blockers and analgesics among women, and they warrant confirmation in larger studies.