However, a different Australian randomized study (the Nambour Skin Cancer Prevention Trial) showed that, after 4.5 years of follow-up, there was no statistically significant difference in the incidence of BCCs or SCCs with regular SPF 16 sunscreen use. This study did not include a sunscreen placebo. Although a secondary subset analysis of the overall number of tumors showed a reduction in the frequency of SCCs on the sites of daily sunscreen application, the validity of the finding is questionable because of the possible effects of extensive multiple statistical testing. An 8-year post-trial observational follow-up demonstrated statistically significant reductions in both the frequency and the overall incidence of SCCs in the regular sunscreen-use arm, but the reliability of these findings is uncertain given their occurrence outside of the controlled-trial environment.
In the Physicians' Health Study, 21,884 male physicians with no reported history of BCC or SCC were randomly assigned to take 50 mg doses of daily oral beta carotene versus placebo in a 2 � 2 factorial trial of beta carotene and aspirin. Incidence of NMSCs was a secondary endpoint in the trial. After 12 years of beta carotene or placebo administration, there was no difference in incidence of either BCC or SCC. RCTs of long-term treatment with beta carotene in individuals previously treated for NMSC also showed no benefit in preventing the occurrence of new NMSCs.[16,19]
High-dose isotretinoin was found to prevent new skin cancers in individuals with xeroderma pigmentosum. However, a RCT of long-term treatment with isotretinoin in individuals previously treated for BCC showed that this agent did not prevent the occurrence of new BCCs but did produce side effects characteristic of isotretinoin treatment.[21,22]
A multicenter, double-blind, randomized, placebo-controlled trial of 1,312 patients with a history of BCC or SCC and a mean follow-up of 6.4 years showed that 200 �g of selenium (in brewer's yeast tablets) did not have a statistically significant effect on the primary endpoint of BCC development and may increase the risk of SCC and total NMSC.[23,24] The cumulative incidence of NMSC was 0.20 versus 0.16 per person year of follow-up in the selenium and placebo groups, respectively (unadjusted relative risk = 1.27; 95% CI, 1.11-1.45).
The use of celecoxib as a chemopreventive agent for actinic keratosis was assessed in a double-blind, randomized, placebo-controlled trial. Two hundred forty high-risk men and women (each with 10-40 actinic keratoses and a history of previous skin cancer) received 200 mg doses of celecoxib twice daily or a placebo for 9 months with an additional 2-month follow-up. No difference was found in the incidence of actinic keratosis, but a post-hoc analysis revealed a statistically significant difference in the mean number of NMSCs per patient (rate ratio = 0.43; 95% CI, 0.24-0.75; absolute difference, 0.2 lesions per patient). The ultimate utility of celecoxib in preventing NMSCs remains unclear, given the exploratory nature of the analysis, the challenge of interpreting benefits in fractions of lesions, and the potential for serious adverse cardiovascular effects associated with long-term use of nonsteroidal anti-inflammatory drugs. However, the unexpected finding of the lack of effect of celecoxib on actinic keratosis but apparent effect on SCC and BCC incidence raises questions about the use of actinic keratosis as an intermediate endpoint for SCC and BCC and our understanding of the natural history of NMSCs.