Head-to-head Comparison of the Safety and Efficacy Profiles of Three Phosphodiesterase Type 5 Inhibitors through Patient-reported Outcomes of 130 000 Patients from a Direct-to-consumer Platform Database

Sildenafil scored higher than tadalafil on satisfaction and hardness in 132,100 real-world patients; vardenafil had the worst side-effect profile (47.4% vs ~34%).

Journal: European Urology Focus | Published: 2025-11-19 | Type: Retrospective cohort, comparative study | PMID: 41261009 Authors: Rodler S et al. — University Hospital Schleswig-Holstein, University of Freiburg, University of Marburg, Asklepios Klinik Altona, and Wellster Healthtech Group (Munich) Funding/COI: Not listed — but two authors (Schröder F, Garrahy E) are employed by Wellster Healthtech Group, the DTC telemedical platform whose proprietary database provided all study data.

Summary

This retrospective study mined 132,100 ED patients from a European direct-to-consumer telemedicine platform to compare sildenafil, tadalafil, and vardenafil on efficacy and adverse events via digital questionnaires. Sildenafil at medium/high doses outperformed equivalent tadalafil doses on patient satisfaction and erection quality scores, while vardenafil showed substantially higher adverse event rates than the other two. The findings are real-world in scale but compromised by the fact that the company supplying the data also co-authored the paper.

Claims

• Sildenafil 50/100 mg produced larger mean improvements in patient satisfaction (2.534 ± 1.3) than tadalafil 10/20 mg (2.235 ± 1.4), p < 0.0001
• Sildenafil also outperformed tadalafil on erection hardness (mean change 2.245 vs 2.105, p < 0.0001) and maintenance (2.264 vs 2.120, p < 0.0001)
• Adverse event rates: sildenafil 35.0% (3,572/10,201), tadalafil 33.9% (2,225/6,564) — not statistically different (p = 0.14)
• Vardenafil adverse event rate 47.4% (268/565) — significantly higher than sildenafil (p < 0.0001)
• Nasal congestion was the most common adverse event across all three drugs at medium/high doses
• Sildenafil less likely to cause reflux: 5.3% vs tadalafil 8.0% (p < 0.0001)
• Tadalafil less likely to cause flushing (absolute difference not fully reported in abstract)

Study Quality

The sample size of 132,100 is the clearest strength here — it dwarfs any RCT in this space, and real-world adverse event rates are almost certainly closer to clinical reality than trial populations. Patient-reported outcomes via digital questionnaire are a validated methodology for ED research, though the instruments used aren't identified in the abstract, making reproducibility uncertain.

The critical methodological problem is dose matching. The study compares sildenafil 50/100 mg against tadalafil 10/20 mg — but these aren't pharmacologically equivalent doses. Tadalafil's mechanism and half-life (~17.5 hours vs ~4 hours for sildenafil) differ substantially, and comparing them on onset or satisfaction without accounting for dosing context (on-demand vs. daily use) is a significant analytical shortcut. The retrospective, uncontrolled design means confounders — age, comorbidities, prior PDE5i experience — are uncontrolled.

Red Flags

• Undisclosed conflict of interest: Two authors are employed by Wellster Healthtech Group, whose platform generated all 132,100 patient records. The paper lists no COI.
• Dose equivalence problem: Sildenafil 50/100 mg vs tadalafil 10/20 mg are not equivalent comparators. Sildenafil's higher satisfaction scores may partially reflect dose-effect rather than drug superiority.
• Self-acknowledged healthy population bias: DTC telemedical patients skew younger and healthier than clinic populations, limiting generalizability to men with significant comorbid cardiovascular or metabolic disease.
• No blinding, no randomization: Treatment assignment was not randomized; men who chose sildenafil may differ systematically from those who chose tadalafil.
• Questionnaire instruments not specified: Without knowing which validated PRO tools were used, the magnitude of score changes (2.534 vs 2.235) can't be interpreted clinically.
• Vardenafil group dramatically smaller (n=565) compared to sildenafil (n=10,201) and tadalafil (n=6,564) — the higher adverse event rate may reflect selection bias.

Strengths

• Largest head-to-head PDE5i comparison in the literature by an order of magnitude
• Real-world adverse event rates more reflective of clinical practice than controlled trials
• Captures nasal congestion, reflux, and flushing rates that are often underreported in manufacturer-funded RCTs
• Demonstrates that DTC platform data can generate research-grade datasets, a methodologically important proof of concept

Verdict

The scale is genuinely impressive, and real-world adverse event data at this volume fills a gap that small RCTs can't. But the undisclosed conflict of interest — company employees publishing findings from their own platform's proprietary data, with no listed COI — is a serious transparency failure that should have been flagged in peer review. The dose-comparison design also makes the efficacy headline (sildenafil beats tadalafil) harder to trust than it looks. Treat the adverse event frequencies as useful approximations of real-world rates; treat the efficacy comparison with skepticism until replicated with equivalent dosing.