A systematic review of the literature was conducted to assess the effectiveness of calcium supplements and/or dietary calcium for the prevention of osteoporotic fractures in postmenopausal women. Studies were identified by conducting a Medline search using the text words "fracture" and "calcium" for the period 1966 to March 1997 and by reviewing articles known to the authors. Only studies with fracture outcomes were eligible. There were 14 studies of calcium supplements (including 4 randomized trials), 18 studies of dietary calcium and hip fracture (no randomized trials), and 5 studies of dietary calcium and other fracture sites (no randomized trials). The 4 randomized trials of calcium supplements (mean calcium dose: 1050 mg) found relative risk (RR) reductions between 25% and 70%. Meta-analytic techniques for dose-response data were used to investigate and pool the findings of 16 observational studies of dietary calcium and hip fracture. These hip fracture studies were not consistent and heterogeneity of study findings (p = 0.02) was not easily explained by subject characteristics or study design. Pooling study results gave an odds ratio (OR) of 0.96 (95% confidence interval, (CI) 0.93-0.99) per 300 mg/day increase in calcium intake (the equivalent of one glass of milk). This is likely to be an underestimate of calcium's true effect because of inaccurate measurement of dietary calcium in observational studies. This review supports the current clinical and public health policy of recommending increased calcium intake among older women for fracture prevention.