Background Results from in vitro and animal studies suggest that cyclooxygenase (COX) inhibitors may reduce the risk of melanoma, but among humans, the evidence remains limited.

Objective In a pilot retrospective cohort, to determine the relationship between the use of COX inhibitors and the incidence, recurrence, and metastases of melanoma in high-risk patients.

Methods Reviewing computerized records at the Miami Veterans Affairs Medical Center, we retrospectively examined the association between COX inhibitor use and melanoma incidence, recurrence, and metastases in high-risk subjects: white subjects previously diagnosed with melanoma (1996–2003). We evaluated three potential outcomes: new melanoma diagnosis, recurrence of a previous melanoma, and melanoma metastasis.

Results Eighty-three subjects with melanoma were included. There was one metastasis among 28 subjects prescribed COX inhibitors, whereas four new melanomas (7.3%), two melanoma recurrences, and six metastases (10.9%) occurred among 55 patients not prescribed COX inhibitors. Although no individual outcomes measures reached statistical significance, combining the three measures, these were significantly lower in users of COX inhibitors compared with nonusers (1 vs 12; p = .05). After adjustment for age and tumor depth of invasion, COX inhibitor users had significantly lower rates of melanoma outcome measures (odds ratio 0.08; 95% confidence interval 0.01–0.77; p = .03).

Conclusion Potential exists for chemoprevention of melanoma among high-risk patients.