first year that cutaneous warts on the hands were assessed, were included in this analysis (Melski et al., 1977; Stern et al., 1979). The Mann-Whitney (Wilcoxon) twosample test for medians and Student t test for means were used to test for statistical significant differences between groups. Poisson regression was used to calculate the association of developing subsequent SCCs or BCCs with the presence of cutaneous hand warts in 1985e1986 and other risk factors. Age, sex, geographic location, and population incidence rates were used to calculate expected number of tumors (Scotto et al., 1974; Stern et al., 1979). We performed a univariate analysis of the association of cutaneous warts with previous history of SCC or BCC, level of exposure to PUVA up to the year 1985, high UVB exposure (300 treatments or more), high doses of methotrexate (if given for 36 or more months), and high doses of cyclosporine (if given for 2 or more years). Factors significant on univariate analysis were included in the multivariable model.