Chemoprevention with retinoids is currently an experimental approach to prevent local relapses and second primaries in treated head and neck cancer patients. We evaluated the effectiveness of vitamin A in preventing the above events in a randomised trial involving 106 head and neck cancer patients who had achieved complete regression of their disease with radiotherapy and/or surgery. They were randomised to receive retinyl palmitate (200,000 IU per week for 1 year) or placebo. 50 subjects on vitamin A and 43 on placebo completed 1 year supplementation; 49 in the former group and 42 in the latter could be evaluated over a 3 year period from the initiation of the study. One fifth (11/56) of patients in the vitamin A group and one tenth (5/50) in the placebo group had loco-regional recurrence. The frequency of recurrences in stage I patients in the vitamin A group was higher compared to the placebo group, although it was not statistically significant. No second primaries were observed in the vitamin A group; 2 patients in the placebo group had second primaries. No clinically obvious side effects were observed with vitamin A. The higher frequency of recurrences in the vitamin A group is of concern although it may be a chance finding given the small size of the trial. The effect on second primaries is consistent with other observations reported in literature.