What is the impact of anxiety and depression on penile prosthesis outcomes? A population-level assessment.

Men with anxiety or depression had 79% higher 90-day infection rates after penile implant surgery; antidepressants didn't close the gap.

Journal: The Journal of Sexual Medicine | Published: 2026-04-09 | Type: Retrospective cohort study | PMID: 41990124 Authors: Ryan Davis, Jeffrey C. Loh-Doyle, Kian Asanad — Keck School of Medicine, USC Department of Urology Funding/COI: Not disclosed

Summary

This study pulled 20 years of U.S. claims data on 15,803 men who received inflatable penile prostheses (IPP), finding that roughly one in five had a prior diagnosis of anxiety or depression (AD). After matching on 18 variables, men with AD showed substantially higher rates of infection, reoperation, and mechanical failure. Notably, patients already on AD medications at the time of surgery saw no reduction in these risks.

Claims

• Of 15,803 IPP patients, 3,386 (21.4%) had a history of AD; 3,132 were retained per cohort after propensity matching
• 90-day device infection: RR = 1.79 (95% CI 1.26–2.55) — a 79% higher risk
• 90-day reoperation: RR = 1.42 (95% CI 1.05–1.93) — a 42% higher risk
• All-time reoperation: RR = 1.17 (95% CI 1.02–1.35)
• All-time mechanical failure: RR = 1.33 (95% CI 1.08–1.65)
• Faster time to mechanical failure: HR = 1.27 (95% CI 1.02–1.58)
• AD medications at the time of surgery did not reduce any of these elevated risks
• AD patients also had significantly increased rates of medical complications and healthcare utilization post-surgery

Study Quality

This is the first population-level examination of this question, moving beyond the prior single-institution, single-surgeon data that limited generalizability. The TriNetX database spans 2005–2025 and provides genuine statistical power at n ≈ 3,100 per cohort. Matching on 18 variables — including demographics, comorbidities, and concurrent medications — is more rigorous than most database studies in this space, reducing (though not eliminating) confounding.

The retrospective design and reliance on ICD/CPT administrative codes introduce measurement error throughout: AD diagnoses, reoperation indications, and infection events are all coded by billing staff, not researchers. The study cannot tell us whether the AD was actively symptomatic at the time of surgery, how severe it was, whether patients were actually adherent to their medications, or what specific behaviors or biological mechanisms drove worse outcomes.

Red Flags

• Administrative codes define all study variables — diagnosis accuracy and severity are unverifiable
• Reoperation is counted but the indication (infection vs. mechanical failure vs. patient preference) cannot be distinguished
• Several confidence intervals barely clear 1.0, particularly for all-time reoperation (1.02–1.35) — these findings are statistically fragile
• Funding and conflicts of interest are not disclosed; industry ties to IPP manufacturers would be material here
• Causal inference is impossible: patients with AD differ in unmeasured ways (lifestyle, follow-up behavior, pain tolerance, support systems) that 18 matched variables won't capture
• No data on AD severity, duration, or treatment adherence
• The finding that AD medications "didn't help" is hard to interpret — we don't know which medications, at what doses, for how long, or with what efficacy

Strengths

• Largest sample to date for this specific clinical question
• Two-decade data window (2005–2025) improves temporal generalizability
• Population-level approach avoids single-center surgeon selection bias
• Pre-specified subgroup analysis on preoperative AD medications
• Multiple outcome domains assessed (infection, reoperation, mechanical failure, healthcare utilization)
• Results are directionally consistent across nearly all outcomes, suggesting a real signal rather than noise

Verdict

The signal is real and clinically plausible — anxiety and depression are known to impair wound healing, pain regulation, and care compliance, and the association here is statistically robust for the primary outcomes. But this paper proves association, not mechanism, and the administrative data limitation means every finding should be interpreted with a wide margin. The null result on AD medications is intriguing but uninterpretable without dose, adherence, and severity data. Worth reading if you're a prosthetic urologist; worth citing carefully if you're building a preoperative screening protocol — this study tells you that a problem exists more than why or how to fix it.