A new retrospective cohort study from the Department of Organ Transplantation at The Second Affiliated Hospital of Guangzhou Medical University, published in BMC Infectious Diseases, addresses that question with unusually detailed infection phenotyping. The research team, led by Rongxin Chen and Jialin Wu, with Luhao Liu and Zheng Chen as corresponding authors, followed 228 SPK recipients and 98 diabetic recipients of kidney transplantation alone, or KTA, who underwent surgery between January 2019 and June 2025 at a single center. The primary endpoint was infection-related hospitalization within 12 months of transplant, a measure chosen because it captures infections severe enough to disrupt recovery, not merely laboratory findings. Secondary endpoints included bloodstream infection, infections at specific anatomical sites, reactivation of cytomegalovirus, Pneumocystis jirovecii pneumonia, and one-year patient and graft survival.

The headline finding is a statistically robust increase in infection-related hospitalizations after SPK. Recipients of the combined transplant experienced 0.54 hospitalization episodes for infection per person-year in the first 12 months, compared with 0.31 episodes per person-year among kidney-alone recipients, an incidence rate ratio of 1.76 with a 95 percent confidence interval of 1.22 to 2.57 and a p-value of 0.003. In plain terms, SPK patients were hospitalized for infection roughly three-quarters more often than their kidney-alone counterparts. When the analysis was refined with multivariable Cox regression, accounting for differences in baseline characteristics between the two groups, the association persisted: SPK performed with rabbit anti-thymocyte globulin induction remained significantly more likely to result in infection-related hospitalization than KTA performed with basiliximab induction, with an adjusted hazard ratio of 1.6 and a p-value of 0.04.

The pattern of infections differed strikingly between the two operations, and that pattern is where the biology becomes visible. Surgical-site and intra-abdominal infections dominated after SPK, affecting 18.9 percent of those recipients compared with just 3.1 percent of kidney-alone recipients, an odds ratio of 7.36. This is precisely what one would expect from a procedure that opens the peritoneal cavity, places an additional vascularized organ, and creates two sets of anastomoses where leaks and collections can occur. By contrast, pulmonary infections were more common after kidney transplantation alone, at 25.5 percent versus 7.9 percent after SPK. The authors also documented a dramatic difference in bloodstream infection: bacteremia occurred at a rate of 0.10 episodes per person-year after SPK versus 0.01 after KTA, an incidence rate ratio of 9.4, indicating that the abdominal surgical burden of the double transplant translates directly into a far higher risk of organisms entering the circulation.

Viral complications added a second layer of risk. Cytomegalovirus DNAemia, detected by polymerase chain reaction in the blood, developed in 53.5 percent of SPK recipients, meaning that more than half of the cohort experienced reactivation of this herpesvirus within the first year. Cytomegalovirus is a well-known driver of indirect effects in transplant recipients: it is immunomodulatory, impairing host defenses well beyond its own direct tissue damage, and prior literature has identified it as a dominant risk factor for Pneumocystis jirovecii pneumonia, a fungal infection that can be devastating in immunosuppressed patients. This study provided a striking confirmation of that relationship. In every affected SPK recipient, cytomegalovirus DNAemia preceded the development of Pneumocystis pneumonia, and the authors report that extending prophylaxis with trimethoprim-sulfamethoxazole after documented DNAemia was followed by no subsequent cases of the fungal infection in the cohort.

That final observation deserves emphasis because it points toward a practical intervention. The authors are careful to describe it as preliminary, since a single-center retrospective series cannot establish causation the way a randomized trial could, and the number of patients who progressed to Pneumocystis pneumonia is necessarily small. Even so, the temporal chain they observed, in which cytomegalovirus reactivation consistently preceded Pneumocystis infection and prolonged prophylaxis appeared to interrupt that chain, is exactly the kind of signal that prospective studies should test. If validated, monitoring for cytomegalovirus DNAemia could become a trigger for extending anti-pneumocystis coverage, a low-cost strategy that might spare high-risk patients a life-threatening complication.

The most reassuring part of the study is what the extra infections did not do. One-year patient survival was statistically indistinguishable between the groups, at 96.3 percent for SPK recipients and 95.0 percent for kidney-alone recipients, with a p-value of 0.32. This suggests that although the combined operation generates more infectious complications, those complications were largely treatable and did not, within the first year, claim additional lives. Better still, death-censored kidney graft survival was actually higher after SPK, at 98.8 percent versus 93.0 percent, a difference that reached statistical significance with a p-value of 0.04. Pancreas graft survival at one year was 94.2 percent. The combined transplant, in other words, delivered its expected metabolic and renal benefits while weathering its own infectious storm.

The methodological architecture of the study strengthens confidence in these findings. The investigators used Kaplan-Meier estimation to characterize time-to-event patterns, multivariable Cox regression to adjust for confounders, and, critically, a Fine-Gray competing-risk sensitivity analysis. The latter is a technical detail that matters: when patients die, they can no longer experience infection, and naive analyses can distort infection rates by treating death as ordinary censoring. The Fine-Gray model treats death as a competing event and confirms that the infection signal is not an artifact of different mortality patterns between the groups. The use of infection-related hospitalization as the primary endpoint also avoids the common pitfall of counting asymptomatic viral detections as infections, anchoring the analysis in events that carry real clinical and economic weight.

The study’s limitations follow naturally from its design. It was conducted at a single center in Guangzhou, China, so induction and maintenance immunosuppression regimens, antimicrobial prophylaxis protocols, and local epidemiology of multidrug-resistant organisms may not generalize everywhere. The retrospective design means the two groups were not randomized, and although multivariable adjustment mitigated measurable imbalances, unmeasured differences between patients selected for a major double operation and those receiving a kidney alone may persist. Importantly, the comparison of induction agents, rabbit anti-thymocyte globulin in the SPK group versus basiliximab in the kidney-alone group, means that part of the adjusted hazard ratio reflects differences in induction strategy as well as the operation itself, a confounding that the authors acknowledge implicitly through their phrasing of the adjusted comparison. Rabbit anti-thymocyte globulin is a lymphocyte-depleting agent associated with deeper and more prolonged immunosuppression, which plausibly contributes to the infection excess.

For clinicians and patients weighing these two operations, the message is nuanced rather than alarming. The double transplant carries a real and quantifiable early infectious price, concentrated in surgical-site and intra-abdominal infections, bacteremia, and cytomegalovirus reactivation, roughly 76 percent more infection-related hospitalizations in the first year. But that price did not purchase worse outcomes; survival was equivalent and kidney graft survival was better, while the pancreas graft offered insulin independence for the vast majority of recipients. The practical implications are twofold: centers performing SPK should expect and surveil for the specific infection profile documented here, and the apparent protective effect of extending Pneumocystis prophylaxis after cytomegalovirus DNAemia is a hypothesis worth testing prospectively. As SPK volumes grow worldwide and donor organs remain scarce, studies like this one, which quantify the trade-offs rather than assume them, are what allow patients and their surgeons to make that choice with open eyes.

Subject of Research: Infection outcomes in the first year after simultaneous pancreas-kidney transplantation compared with kidney transplantation alone in diabetic recipients

Article Title: One-year infection outcomes in simultaneous pancreas-kidney versus kidney transplantation alone in diabetic recipients: a retrospective cohort study

Article References: One-year infection outcomes in simultaneous pancreas-kidney versus kidney transplantation alone in diabetic recipients: a retrospective cohort study. (n.d.). https://doi.org/10.1186/s12879-026-14532-8

Image Credits: AI Generated

DOI: 10.1186/s12879-026-14532-8

Keywords: simultaneous pancreas-kidney transplantation, kidney transplantation, diabetes, organ transplantation, infection-related hospitalization, bacteremia, surgical site infection, cytomegalovirus, Pneumocystis jirovecii pneumonia, immunosuppression, graft survival, retrospective cohort study

The commentary, authored by Syeda Bareera of Khyber Girls Medical College in Peshawar, Syed Hatim Hussain and Syed Muhammad Momin of Nowshera Medical College, Syed Muhammad Ahmad of Khyber Medical College, and Muhammad Fayyaz of Sphinghar University in Afghanistan, does not dismiss the original findings. Instead, the authors contend that the apparent link between GLP-1 receptor agonists and wound dehiscence may depend heavily on how drug exposure is defined, how rare outcomes are handled statistically, and whether wound risk is evaluated at the level of specific procedures rather than broad surgical categories. These three methodological distinctions, they argue, deserve far closer consideration before surgeons change practice or patients become alarmed. The commentary is a masterclass in the kind of critical appraisal that modern evidence-based medicine demands, particularly when a topic touches one of the most widely discussed drug classes in the world.

The first and perhaps most consequential point concerns the definition of exposure itself. The commentary authors emphasize that GLP-1 receptor agonist use should never be treated as a single, uniform perioperative exposure. There is a profound clinical difference between a patient who is actively taking a weekly injection in the days surrounding surgery, a patient who temporarily withheld the medication shortly before the operation, and a patient who discontinued therapy months earlier and has since achieved stable weight and nutritional status. Each of these scenarios carries a distinct physiological profile, and each could plausibly exert a different influence on wound healing. Yet in the studies included in the original meta-analysis, perioperative exposure timing was inconsistently reported at best. In some cohorts, cessation of the drug ranged from just one week before surgery to an average of 9.6 weeks, while in other cohorts the timing of discontinuation was not reported at all. This heterogeneity makes it genuinely difficult to know what the pooled estimates actually represent.

This ambiguity matters because clinical guidance on perioperative GLP-1 receptor agonist management has itself evolved rapidly. The initial recommendation from the American Society of Anesthesiologists advised holding weekly agents for one week before surgery, largely out of concern for residual gastric contents and aspiration risk during anesthesia. More recent multi-society guidance has shifted toward individualized assessment, weighing the timing of the last dose, the presence of gastrointestinal symptoms such as nausea and delayed gastric emptying, recent dose escalation, and patient-level risk factors. Against this moving target, the commentary authors argue that future studies must categorize exposure according to clinically meaningful perioperative windows: continued therapy, short-term withholding, prolonged withholding, or discontinuation after weight stabilization. They also suggest that investigators apply target trial emulation principles, a framework for causal inference from observational data championed by epidemiologist Miguel Hernán and colleagues, which aligns exposure assignment, follow-up, and outcome assessment with the causal question a randomized trial would ask. Only then, they contend, can observational analyses begin to disentangle the true effect of the drugs from the effects of weight change, nutrition, and timing.

The second methodological concern involves the statistical handling of rare events, a notoriously thorny problem in meta-analysis. In one of the included studies, led by Liang and colleagues, researchers reported zero surgical site infection events among GLP-1 receptor agonist users compared with seven events among controls, creating what statisticians call a single one-arm zero-event cell. Zero cells are poison for conventional risk ratio calculations, because dividing by zero is mathematically undefined and the choice of a continuity correction, a small artificial adjustment added to every cell, can materially change the pooled estimate. The original meta-analysis used Mantel-Haenszel random-effects risk ratios for binary outcomes, but the commentary authors note that it remains unclear whether a continuity correction was applied. They argue that a brief clarification of this analytic choice would help readers judge the robustness of the nonsignificant surgical site infection finding, and they cite the work of Xu and colleagues, who have proposed that meta-analysts select analytic methods according to the specific zero-event structure of their data rather than defaulting to a single approach.

Beyond clarification, the commentary authors recommend sensitivity analyses to test whether the nonsignificant infection finding survives alternative analytic conventions. These could include the Mantel-Haenszel risk difference, appropriate one-stage models, or Peto odds ratios, though the latter are only appropriate when groups are reasonably balanced and treatment effects are small. The point is not academic pedantry. When outcomes are uncommon and event counts are sparse, the apparent absence of a statistical signal can be an artifact of the analytic method rather than a genuine reflection of safety. For clinicians deciding whether to delay surgery or alter medication management, understanding the degree of uncertainty around nonsignificant findings is essential. The commentary authors argue that transparent reporting of these conventions would allow readers to interpret uncommon outcome estimates with appropriate caution, rather than mistaking statistical silence for evidence of no effect.

The third concern is arguably the most intuitive: the composite category of body contouring surgery may simply be too broad for an outcome as mechanically dependent as wound dehiscence. Panniculectomy, abdominoplasty, lipoabdominoplasty, brachioplasty, thighplasty, and mastopexy differ dramatically in incision length, flap tension, extent of undermining, dead space, drain use, and postoperative mechanical stress on the closure. High-tension abdominal procedures such as panniculectomy and abdominoplasty may involve broad dissection and challenging closure-tension dynamics, whereas mastopexy or brachioplasty involve entirely different anatomic sites, incision geometry, and postoperative forces. These factors are directly related to dehiscence risk and are not interchangeable across procedure types. Pooling them into a single body contouring category risks diluting or distorting any true procedure-specific signal, and the commentary authors argue that future work should stratify analyses by procedure type wherever sample size permits, particularly separating high-tension abdominal operations from other contouring procedures.

The commentary extends this procedural critique to the outcome definitions themselves. Surgical site occurrences, as commonly reported in the plastic surgery literature, combine events with fundamentally different mechanisms: infection, seroma, hematoma, and dehiscence. The Centers for Disease Control and Prevention’s National Healthcare Safety Network framework links surgical site infection surveillance to operative procedure category and tissue depth, while leading plastic surgery researchers such as Gabriel, Gupta, and Orgill have emphasized the clinical importance of distinguishing among these wound events rather than treating them as interchangeable. A pooled estimate that blends a seroma with a wound separation tells clinicians very little about either. The commentary authors call for predefined wound-event definitions, standardized follow-up intervals, and procedure-stratified analyses in future research, arguing that only this level of granularity will reveal which patients, which operations, and which perioperative management strategies genuinely warrant additional caution.

Crucially, the authors are careful to frame their critique as a call for cautious interpretation rather than a rejection of the original work. They acknowledge that the finding of a possible association between GLP-1 receptor agonist use and wound dehiscence is clinically important, particularly as plastic surgeons increasingly encounter patients arriving after pharmacologic weight loss, often with altered nutritional status, rapid changes in tissue quality, and complex medication histories. The current evidence, they write, is clinically valuable but limited by heterogeneity and variable reporting across the underlying primary studies. Their goal is to ensure that the dehiscence signal stimulates better-designed research rather than premature clinical dogma. In an era when a single headline about a blockbuster drug class can reshape patient behavior overnight, the distinction between a preliminary association and an established risk is not a technicality; it is the difference between informed decision-making and unnecessary alarm.

The broader lesson of this scholarly exchange extends well beyond GLP-1 receptor agonists and the operating room. It illustrates how the quality of a meta-analysis is ultimately bounded by the quality and granularity of the primary studies it pools, and how seemingly technical choices, the definition of an exposure window, the treatment of a zero cell, the boundaries of a surgical category, can shape conclusions that reach millions of patients. As the obesity medication revolution continues and the wave of post-weight-loss body contouring surgery swells, the surgical community now has both a warning and a roadmap: take the wound-healing question seriously, but answer it with the methodological rigor it demands. Patients contemplating surgery while on these medications should discuss timing and individual risk with their surgical and prescribing teams, guided by evolving multi-society recommendations rather than by any single pooled estimate.

Subject of Research: Perioperative safety of GLP-1 receptor agonists in body contouring surgery

Article Title: Comment on: Safety of GLP-1 receptor agonists in body contouring surgery: a systematic review and meta-analysis

Article References: Bareera, S., Hussain, S. H., Momin, S. M., Ahmad, S. M., & Fayyaz, M. (2026). Comment on: Safety of GLP-1 receptor agonists in body contouring surgery: a systematic review and meta-analysis. BMC Plastic and Reconstructive Surgery, 2(1), Article 20. https://doi.org/10.1186/s44452-026-00033-9

Image Credits: AI Generated

DOI: 10.1186/s44452-026-00033-9

Keywords: GLP-1 receptor agonists, body contouring surgery, wound dehiscence, meta-analysis, systematic review, surgical site infection, perioperative management, plastic surgery, weight loss medications, zero-event studies, surgical complications, target trial emulation