Methods. The cases (n = 6558) were compared with cancer controls (n = 3296) and noncancer controls admitted for trauma or acute infection (n = 2925). Odds ratios were estimated using multivariate logistic regression models.

Results. For women who used NSAIDs regularly beginning at least 1 year before admission, the odds ratios (OR) were 0.8 (95% CI 0.7, 1.0) with cancer controls and 0.7 (95% CI 0.6, 0.9) with noncancer controls. With noncancer controls, there was a statistically significant decreasing trend in the odds ratios as duration of use increased, whereas with cancer controls there was not. The reduction in risk for regular use was accounted for largely by a reduced odds ratio for one study center (Boston), which contributed 9% of the cases.

Conclusions. The data are compatible with a small reduction in risk, associated with regular NSAID use. However, inconsistencies in the data detract from a causal interpretation. (C) 1999 American Health Foundation and Academic Press.