Physician Clalit Health Services, and Technion Faculty of Medicine Haifa, Israel
Background: In Alzheimer's dementia low levels of cGMP have been described. PDE5 inhibitors (PDE5i) indicated to treat impotence increase cytoplasmic cGMP levels, and had been associated with a decrease in dementia risk. We studied the association between PDE5i and dementia risk in the real-world, utilizing demographic and clinical variables to reduce potential biases. Methods: Retrospective cohort study using computerized database of Clalit, Israel's largest health care provider. We assembled 3 cohorts: PDE5i users; patients with a diagnosis of impotence and no PDE5i; and a second control group of multivitamins users. We performed multivariable cox-dependent analysis using clinical variables identified at cohort entry date, for the association between PDE5i dispensing and new dementia diagnosis within patients with impotence. Second, PDE5i users were matched to multivitamin users, on their estimated propensity. A marginal model with robust sandwich estimate was used to model the time for a new diagnosis of dementia. Dementia diagnosed at least 2 years after cohort entry date was accounted for. In a sensitivity analysis 5-years lag time used. Results: In a mean follow-up of 7.91(SD 4.67) years on 133,336 patients with impotence, 8,631 received a diagnosis of dementia. In a multivariable cox-dependent analysis, dispensing of PDE5i was not associated with reduced dementia risk HR=0.95 95%CI (0.86-1.04), p=0.250. In an intention-to-treat analysis of 65,855 PDE5i new users matched 1:1 to multivitamins users, HR=0.93 95%CI (0.89, 0.97), p=0.002 for new dementia. Using 5 years lag time with 42,050 PDE5i matched 1:1 to multivitamins users HR=0.98 (0.93, 1.03), p=0.780 for new dementia diagnosis. Conclusion: PDE5i treatment is associated with very mild/no decrease in risk of dementia. Additional sensitivity analyses are underway.
.jpg "Naomi Gronich, MD (she/her/hers) photo")