Robotic-assisted urologic surgery is now the default for radical prostatectomy and most partial nephrectomies, and the obesity drugs that the same patient population is increasingly taking sit on a list that anesthesia and surgical societies tell those patients to pause before going under. The pause rule comes from gastroparesis and the aspiration concern that follows. The published evidence behind it sits mostly in endoscopy and bariatric series. Urology has had to extrapolate.
A single-center retrospective cohort just put numbers on the urology side of the question. Published May 29 in the Journal of Robotic Surgery ↗, the study pulled 361 adults with body mass index of 35 or higher who had undergone robotic-assisted radical prostatectomy, partial nephrectomy, or radical nephrectomy at one center between 2017 and 2024. The exposure rule was strict: established GLP-1 receptor agonist use meant at least three months of documented prescriptions both before and after surgery, so transient or recently discontinued users were not in the comparison. Propensity scores ran 2:1 nearest-neighbor matching against non-users. After matching the analytic set was 36 GLP-1 users against 59 controls, with mean absolute standardized mean difference of 0.036 across covariates and a maximum of 0.074. Balanced.
The primary outcome was any postoperative complication on the Clavien-Dindo scale, a five-grade classification of how severe a postoperative complication was based on what was needed to treat it. The odds ratio for GLP-1 users versus controls came in at 2.39, with a 95% confidence interval running from 0.63 to 9.10. That interval is what the rest of the readout looks like. Major complications (Clavien-Dindo III or higher) came back at OR 3.70 (95% CI 0.45 to 30.32). The comprehensive complication index, a continuous severity-weighted score on the same scale, showed a mean difference of 1.29 between groups with a p-value of 0.64. Each headline number runs the wrong direction (more complications in GLP-1 users) and each confidence band swallows the null by enough to make the point estimate uninformative.
The continuous secondary endpoints tilted the other way without crossing into significance. Estimated blood loss was 81 mL lower in GLP-1 users (95% CI -165.85 to +3.22, p=0.06). Length of stay was 0.78 days shorter (p=0.11). Peak postoperative pain scores and emergency department utilization at 90 days and one year did not differ between the groups.
For a question that has had to live on indirect evidence, the urology-specific number that matters is that there is no urology-specific number that crosses null. The authors are careful to note that with 36 exposed patients the power constraints preclude any conclusion of equivalence. What the data does support, they argue, is moving away from the reflexive perioperative discontinuation that has anchored the conversation since the American Society of Anesthesiologists consensus guidance ↗ was issued in mid-2023, and toward an individualized risk stratification that incorporates the patient's actual surgical and medication context.
A few things keep this from being the definitive answer. The data is single-center and retrospective. The 2:1 matching ratio chose statistical efficiency over sample symmetry, which contributes to the width of every confidence interval in the safety panel. Robotic urologic surgery is not a single thing; partial nephrectomy and radical prostatectomy have different aspiration risk profiles around the time the patient is positioned and intubated. And the established-use rule explicitly excludes the patient the anesthesia societies are most concerned about, which is the patient who started a GLP-1 receptor agonist ↗ recently or who is mid dose escalation when the surgery date arrives.
What this cohort does usefully add is a non-null marker on a map that has been mostly white space. The signal for urology is that there is no signal yet, and the recommendation should not lead the data.