Dementia and PDE5 inhibitors: A retrospective cohort study of veterans

Technical Summary: Dementia and PDE5 Inhibitors in a Veteran Cohort

Problem Statement
Dementia represents a critical global public health crisis, with over 55 million individuals currently affected and projections exceeding 130 million by 2050. The condition imposes severe cognitive decline, loss of independence, and substantial economic burdens, with annual global costs estimated to exceed one trillion U.S. dollars. Currently, no cure exists, and therapeutic options are limited to symptomatic relief. Consequently, strategies capable of delaying the onset of dementia, even by a few years, are of significant clinical and societal value. While preclinical studies suggest that Phosphodiesterase type 5 (PDE5) inhibitors—originally developed for erectile dysfunction (ED) and pulmonary arterial hypertension—may offer neuroprotective benefits via nitric oxide–cGMP signaling, synaptic plasticity enhancement, and improved cerebral blood flow, large-scale human evidence regarding their impact on dementia incidence remains limited.

Methodology
This study employed a retrospective cohort design utilizing national data from the U.S. Department of Veterans Affairs (VA) Corporate Data Warehouse (CDW). The study population consisted of approximately 2.5 million male veterans who received care for erectile dysfunction (ED) between January 1998 and July 2024.

• Exposure Definition: The exposed cohort comprised veterans prescribed a PDE5 inhibitor (sildenafil, tadalafil, or vardenafil), identified via outpatient pharmacy dispensing records. The index date was the first dispensing event.
• Control Group: A control cohort of veterans with ED but no PDE5 inhibitor exposure was established. To ensure temporal alignment and reduce calendar-time bias, index dates for controls were randomly selected from the distribution of PDE5 inhibitor index dates.
• Matching: Propensity score matching (1:1 greedy nearest neighbor) was performed using SAS to balance the cohorts on baseline covariates, including age, BMI, follow-up duration, and comorbidities (cardiovascular and mental health conditions). This resulted in 439,120 individuals per group.
• Outcomes: The primary outcomes were new incident diagnoses of Alzheimer's disease (AD), vascular dementia (VaD), frontotemporal dementia, and senile dementia, identified via ICD-9 and ICD-10 codes post-index date.
• Analysis: The study utilized an intention-to-treat framework due to a lack of adherence data. Statistical analyses included odds ratios (OR) with 95% confidence intervals for incidence, expanded means tests for time-to-diagnosis, and Kaplan–Meier survival analyses to assess cumulative incidence and hazard ratios.

Key Results

• Baseline Characteristics: After propensity score matching, baseline characteristics were well-balanced between the PDE5 inhibitor users and the control group. Baseline dementia prevalence was similarly low in both groups (approximately 0.17% for AD, 0.19% for VaD, and 0.014% for frontotemporal dementia), reflecting the exclusion of prevalent cases and the functional nature of the ED cohort.
• Dementia Incidence: During longitudinal follow-up, PDE5 inhibitor users demonstrated significantly lower rates of new dementia diagnoses compared to controls:
  • Alzheimer's Disease: OR 0.91 (95% CI, 0.85–0.96).
  • Frontotemporal Dementia: OR 0.83 (95% CI, 0.69–1.00).
  • Vascular Dementia: OR 0.84 (95% CI, 0.79–0.89).
  • Senile Dementia: No significant difference was observed between groups.
• Time-to-Diagnosis: Kaplan–Meier analyses revealed a sustained delay in the onset of dementia for the PDE5 inhibitor group. The time to diagnosis was delayed by approximately two to three years across several subtypes. The separation in survival curves was gradual and progressive, suggesting a modulation of neurovascular or degenerative processes rather than isolated short-term effects.
• Mortality: While significant mortality was observed in the overall population, death rates did not differ significantly between the two groups, indicating that the observed differences in dementia outcomes were not confounded by differential mortality.

Significance and Claims
The paper claims that PDE5 inhibitor exposure is associated with a reduced incidence of specific dementia subtypes and a clinically meaningful delay in the time to diagnosis among male veterans with ED. The authors suggest that these findings support the hypothesis that PDE5 inhibitors may exert protective effects through vascular and neuroprotective mechanisms, particularly relevant given the vascular component of Alzheimer's and the endothelial dysfunction associated with ED.

The study highlights the potential for repurposing established, well-tolerated medications to mitigate dementia risk. The authors emphasize that even modest delays in disease onset can lead to substantial improvements in quality of life, caregiver burden, and healthcare utilization. However, the paper maintains a modest tone regarding causality, acknowledging limitations such as the observational design, reliance on administrative coding, lack of adherence data, and the restriction to a male veteran population with ED, which may limit generalizability to women or non-veteran populations. The authors conclude that these findings warrant further prospective research and clinical trials to explore PDE5 inhibitors as potential disease-modifying therapies.

A scientific accuracy reviewer checked the draft against the paper and flagged these problems:

• Summary claims no significant association between PDE5 use and incident dementia, contradicting the paper's finding of significant reductions. (the paper says: "There is no significant association between PDE5 inhibitor use and the risk of incident dementia.")

Correction Note: The sentence "There is no significant association between PDE5 inhibitor use and the risk of incident dementia" appears in the Discussion section of the provided text, but it directly contradicts the Abstract, Results, and the specific statistical findings (ORs < 1 with significant CIs) presented throughout the paper. The Abstract explicitly states: "new diagnoses of Alzheimer, frontotemporal, and vascular dementia were significantly lower among PDE5 inhibitor users." The Results section confirms: "New cases of Alzheimer's disease occurred less frequently in the PDE5 inhibitor-treated group compared to the control group." The flagged sentence in the Discussion appears to be a textual error or a misstatement within the source document itself, as it is inconsistent with the study's own data and conclusions. The corrected summary above reflects the actual findings (significant reductions) as reported in the Abstract and Results, rather than the contradictory sentence found in the Discussion.