Abstract
BACKGROUND: Phosphodiesterase 5 (PDE5) inhibitors have been hypothesized to have chemoprotective effects in colorectal cancer. Current population-based epidemiologic evidence is, however, limited and inconsistent. METHODS: Among 18,123 men in the Health Professionals Follow-up Study who had at least one lower gastrointestinal endoscopy, we evaluated the association between PDE5 inhibitor use and risk of conventional adenoma and serrated lesion between 2000 and 2010, adjusted for repeated observations and multiple risk factors. We stratified by erectile dysfunction to account for potential "confounding by indication." RESULTS: We documented 2,595 conventional adenomas and 1,395 serrated lesion polyps during the follow-up period. Using people without erectile dysfunction as reference group, recent PDE5 inhibitor use at baseline was not associated with lower risk of conventional adenoma [erectile dysfunction with PDE5 inhibitors: OR = 1.08; 95% confidence interval (CI) = 0.92-1.26; erectile dysfunction without PDE5 inhibitors: OR = 0.95; 95% CI, 0.85-1.06], serrated lesions (erectile dysfunction with PDE5 inhibitors: OR = 1.19; 95% CI = 0.97-1.45; erectile dysfunction without PDE5 inhibitors: OR = 1.03; 95% CI = 0.89-1.19), or advanced conventional adenomas (erectile dysfunction with PDE5 inhibitors: OR = 1.20; 95% CI = 0.94-1.53; erectile dysfunction without PDE5 inhibitors: OR = 0.95; 95% CI = 0.79-1.14). No association was found for PDE5 inhibitor use ever before as well. CONCLUSIONS: We found no evidence of an association between PDE5 inhibitor use and risk of conventional and serrated precursors of colorectal cancer. IMPACT: We show that PDE5 inhibitor use is not associated with precursors of colorectal cancer adjusted for medical and lifestyle risk factors among a large population of men with 10 years of follow-up.