The effect of beta-carotene on sputum cytology in smokers: a preliminary study

The use of biomarkers is a promising approach to the study of human cancer risk. Bronchial metaplasia in sputum cytology may be a marker for potential premalignancy that can be used for population studies. We recently performed a randomized, controlled trial in smokers on the effect of 14 weeks beta-carotene (20mg/day) on markers for DNA damage. We now have evaluated the application of sputum cytology in this study and performed a preliminary evaluation of the effect of beta-carotene. Of the 150 potential participants in this trial 75 were not eligible because they failed to produce sputum samples (n = 29), or because samples were unsatisfactory (n = 46). The eligible group was older (41 vs 37 years) and had smoked longer (23 vs 19 years), but had similar cigarette consumption (mean 21/day) and plasma cotinine levels. Metaplasia was graded in seven categories. Only 11 subjects (15%) showed minor or mild atypia on study entry. Agreement within and between observers was 95% within the same or an adjacent category. We observed no significant correlation between before and after treatment final metaplasia scores in either the beta-carotene (Spearman R = 0.18, P = 0.3) or placebo group (Spearman R = 0.17, P = 0.3). Initial metaplasia scores were somewhat higher in the beta-carotene group (n = 33) than in the placebo group (n = 42) (P = 0.06). Final metaplasia scores were similar in both groups (P = 0.69), and there was no decrease in metaplasia scores in the beta-carotene group (P = 0.75). This study indicates that sputum cytology may not yet be a readily applicable marker in studies of a healthy asymptomatic population, because many smokers do not spontaneously produce sputum, more severe lesions are rare, and variation over time in the minor lesions in large. Therefore, the preliminary evidence that beta-carotene has no influence should be interpreted with care.