The prevalence of metabolic syndrome is increasing along with breast cancer incidence worldwide. Because fenretinide improves insulin action and glucose tolerance in insulin-resistant obese mice and because tamoxifen has shown to regulate several markers involved in metabolic syndrome, we sought to investigate the effect of fenretinide or tamoxifen at low dose on features linked to insulin resistance in premenopausal women at risk for breast cancer. We randomized 235 women to low-dose tamoxifen (5 mg/daily), fenretinide (200 mg/daily), or their combination or placebo for 2 years. We used the homeostasis model assessment (HOMA; fasting insulin x glucose/22.5) to estimate insulin sensitivity. Women were considered to improve insulin sensitivity when they shifted from a HOMA >/=2.8 to <2.8. There was no effect of fenretinide or tamoxifen on HOMA overall, but overweight women (body mass index, >or=25 kg/m(2)) had a 7-fold greater probability to normalize HOMA after 2 years of fenretinide treatment [odds ratio (OR), 7.0; 95% confidence interval (95% CI), 1.2-40.5], with 25% of women improving their insulin sensitivity, whereas tamoxifen decreased insulin sensitivity by almost 7 times compared with subjects not taking tamoxifen (OR, 0.15; 95% CI, 0.03-0.88). In this group only, 5% improved their insulin sensitivity. Interestingly, women with intraepithelial or microinvasive neoplasia had higher HOMA (3.0) than unaffected subjects (2.8; P = 0.07). Fenretinide can positively balance the metabolic profile in overweight premenopausal women and this may favorably affect breast cancer risk. Furthermore, features of the metabolic syndrome should be taken into consideration before proposing tamoxifen for breast cancer prevention. The clinical implications of these results require further investigations.