We studied the effect of fenretinide [N-(4-hydroxyphenyl)retinamide (4-HPR)], a synthetic analogue of retinoic acid, on plasma insulin-like growth factor I (IGF-I) levels in a consecutive cohort of stage I breast cancer patients belonging to a randomized phase III trial of breast cancer chemoprevention. Thirty-two women receiving 4-HPR 200 mg/daily and 28 untreated controls entered the study. IGF-I levels were determined after acid-ethanol extraction, on plasma obtained at randomization and after a mean time of 10.8 +/- 0.3 months. At baseline, there was no difference in IGF-I levels between the two groups [152.9 +/- 9.4 versus 159.2 +/- 7.0 ng/ml in treated and control group (P = 0.59), respectively]. After follow-up time, while plasma IGF-I levels were unchanged in control patients (163.3 +/- 7.4 ng/ml; P = 0.5), they were significantly reduced to 134.6 +/- 8.1 ng/ml in the patients treated with 4-HPR (P = 0.003 and P = 0.011 versus baseline and control values, respectively). Multiple regression analysis showed that treatment was the only determinant of IGF-I decline. Moreover, the interaction between treatment and age was significant, in that the decrease of IGF-I levels induced by 4-HPR administration was much more pronounced in younger patients, while an age-related decline was observed in controls. We conclude that the synthetic retinoid 4-HPR lowers circulating IGF-I levels in early breast cancer patients. Although the importance of this observation for the clinical prevention of breast cancer remains to be established, it further substantiates the rationale of the combination of 4-HPR with tamoxifen, which is known to decrease IGF-I as well and to act synergistically with the retinoid in preclinical models.