This 2-year, double-masked, randomized, placebo-controlled trial was designed to evaluate the safety and efficacy in preventing bone loss in postmenopausal women of two doses of transdermal 17betaestradiol (estradiol) delivered by a matrix patch, compared with placebo. One hundred and sixty healthy, hysterectomized postmenopausal volunteers aged 40-60 years with serum estradiol levels < 20 pg/ml were started on treatment at four centers in The Netherlands. Every 6 months, bone mineral density (BMD) was measured by dual-energy X-ray absorptiometry (DXA) at the lumbar spine, non-dominant wrist and left hip, and markers of bone turnover were assessed in urine and serum. The treatment arms were: estradiol, 100 microg/day (E-100, n = 53), oestradiol, 50 microg/day (E-50, n = 54), placebo (P-100, placebo to E-100, n = 27 or P-50, placebo to E-50, n = 26). Treatment was continued for up to 2 years. After 24 months, BMD of the lumbar spine in the E-100 group differed by 7.7% [5.8-9.5%] (mean [95% confidence interval]) from the placebo group and showed a mean (s.e.m.) increase in BMD from baseline of 5.9% (0.69%). For the E-50 group the difference compared with placebo was 6.2% [4.4-8.0%] and the absolute increase was 4.5% (0.62%); in the placebo group, the absolute change was -2.3% (0.48%). In the total wrist, the changes were: E-100: difference compared with placebo 2.5% [1.5 3.6%], absolute increase 0.6% (0.3%); E-50: difference compared with placebo 2.9% [1.8-3.9%], absolute increase 0.7% (0.25%); and absolute change on placebo: -2.5% (0.35%). In the total hip, the changes were: E-100: difference compared with placebo 3.7% [2.2-5.2%], absolute increase 2.8% (0.5%); E-50: difference compared with placebo: 3.2% [1,8-4.7%], absolute change 2.4% (0.36%); and absolute change on placebo -1.4% (0.66%). Three markers of bone turnover--serum bone-specific alkaline phosphatase, serum osteocalcin and urinary CTX--fell significantly during the trial. Breast pain was reported by 8% of women on placebo, by 6% of women on E-50 and by 17% of women on E-100. Estradiol delivered by the E-50 matrix patch effectively reversed bone loss in hysterectomized postmenopausal women with few side-effects. The marginal additional gain in BMD with the higher dose may be offset by a more important side effect profile.