Background: Laboratory evidence suggests that reduced phosphodiesterase type 5 (PDE5) expression increases the invasiveness of melanoma cells; hence, pharmacological inhibition of PDE5 could affect melanoma risk. Two major epidemiological studies have investigated this and come to differing conclusions. We therefore aimed to investigate whether PDE5 inhibitor use is associated with an increased risk of malignant melanoma, and whether any increase in risk is likely to represent a causal relationship.
Methods and findings: We conducted a matched cohort study using primary care data from the UK Clinical Practice Research Datalink. All men initiating a PDE5 inhibitor and with no prior cancer diagnosis were identified and matched on age, diabetes status, and general practice to up to four unexposed controls. Ever use of a PDE5 inhibitor and time-updated cumulative number of PDE5 inhibitor prescriptions were investigated as exposures, and the primary outcome was malignant melanoma. Basal cell carcinoma, solar keratosis, and colorectal cancer were investigated as negative control outcomes to exclude bias. Hazard ratios (HRs) were estimated from Cox models stratified by matched set and adjusted for potential confounders. 145,104 men with ≥1 PDE5 inhibitor prescription, and 560,933 unexposed matched controls were included. In total, 1,315 incident malignant melanoma diagnoses were observed during 3.44 million person-years of follow-up (mean 4.9 y per person). After adjusting for potential confounders, there was weak evidence of a small positive association between PDE5 inhibitor use and melanoma risk (HR = 1.14, 95% CI 1.01-1.29, p = 0.04). A similar increase in risk was seen for the two negative control outcomes related to sun exposure (HR = 1.15, 95% CI 1.11-1.19, p < 0.001, for basal cell carcinoma; HR = 1.21, 95% CI 1.17-1.25, p < 0.001, for solar keratosis), but there was no increased risk for colorectal cancer (HR = 0.91, 95% CI 0.85-0.98, p = 0.01). There was no evidence that risk increased with number of prescriptions received (p-trend = 0.83). In a post hoc analysis, there was strong evidence that solar keratosis was associated with future PDE5 inhibitor use (odds ratio = 1.28, 95% CI 1.23-1.34, p < 0.001), suggesting that men with higher sun exposure were more likely to become PDE5 inhibitor users. However, a limitation of our study was that we did not have individual-level data on sun exposure, so we could not directly control for this in the primary analysis.
Conclusions: Our results were not consistent with PDE5 inhibitors being causally associated with melanoma risk, and strongly suggest that observed risk increases are driven by greater sun exposure among patients exposed to a PDE5 inhibitor.