Protective effects of SGLT2 inhibitors and GLP-1 receptor agonists on infection-related implant failure following ankle fracture ORIF: A global big data analysis

Abstract

Background

Patients with type 2 diabetes (T2DM) undergoing ankle fracture open reduction and internal fixation (ORIF) face high risks of surgical site infection (SSI) and implant failure. This study evaluated whether preoperative use of sodium–glucose cotransporter–2 inhibitors (SGLT2is) or glucagon-like peptide–1 receptor agonists (GLP-1 RAs) improves these postoperative outcomes.

Methods

Using the TriNetX global database (2005–2025), we identified T2DM patients undergoing ankle fracture ORIF. Users of SGLT2is or GLP-1 RAs within 180 days before surgery were compared with non-users using 1:1 propensity score matching (PSM). The primary outcome was a 90-day composite of SSI or infection-related implant removal. Secondary outcomes included the 1-year composite and all-cause mortality.

Results

After PSM, 1289 users were compared with 1289 non-users. At 90 days, the composite infection-related outcome occurred in 1.45% of users versus 6.26% of non-users (HR 0.34, 95% CI 0.19–0.62; p < 0.001). At 1 year, the risk remained significantly lower in users (1.79% vs 11.92%; HR 0.42, 95% CI 0.25–0.72; p < 0.001). While 1-year mortality showed a downward trend in the primary cohort (HR 0.61, p = 0.1393), sensitivity analyses excluding dual-therapy users demonstrated a statistically significant survival advantage for single-agent users (HR 0.41, 95% CI 0.21–0.80; p = 0.007). Further analysis restricted to active users (30-day exposure) showed no observed instances of mortality or 1-year infection-related complications (both p < 0.001), although these findings should be interpreted cautiously given the small number of events.

Conclusions

Preoperative use of SGLT2is or GLP-1 RAs was associated with a substantial reduction in infection-related implant failure and a lower risk of mortality, particularly when used as monotherapy or in closer proximity to surgery. These findings suggest that modern cardiometabolic agents may represent valuable components of preoperative optimization strategies in high-risk diabetic patients, although confirmation in prospective studies is warranted.

Keywords

GLP-1 receptor agonist SGLT2 inhibitor ankle fracture surgical site infection type 2 diabetes open reduction internal fixation

Introduction

Patients with type 2 diabetes mellitus represent a substantial proportion of those requiring surgical intervention and experience higher rates of postoperative complications compared with non-diabetic patients.¹ Ankle fracture open reduction and internal fixation is a common orthopedic procedure, particularly in older populations with multiple comorbidities.² Diabetic patients undergoing ankle fracture ORIF face elevated risks of surgical site infection, implant failure, nonunion, and perioperative mortality.³ Recent literature has identified several modifiable risk factors for post-ORIF infection in diabetic patients, including poor glycemic control, obesity, smoking status, chronic kidney disease, and open fracture pattern.^4–6

Recent observational studies in arthroplasty and general surgery have suggested that modern cardiometabolic agents—including sodium–glucose cotransporter–2 inhibitors and glucagon-like peptide–1 receptor agonists—may reduce perioperative infection risk and mortality through improved glucose control, weight reduction, and anti-inflammatory effects.^7–10 However, evidence in ankle fracture surgery remains sparse and inconsistent. While some studies have reported protection against infection and wound complications following GLP-1 RA use,¹¹ others have documented increased risks of surgical site complications and nonunion.¹² To our knowledge, no large-scale multicenter study has examined the combined effect of SGLT2is and GLP-1 RAs in ankle fracture ORIF, nor have SGLT2 inhibitors been directly studied in this surgical population.

We hypothesized that preoperative use of SGLT2is or GLP-1 RAs would be associated with lower risks of infection-related implant failure and reduced all-cause mortality after ankle fracture ORIF in patients with T2DM compared with non-use. To test this hypothesis, we conducted a large, multicenter retrospective cohort study using the TriNetX Research Network, which aggregates de-identified electronic health record data from multiple healthcare organizations.

Materials and methods

Study design and data source

We conducted a retrospective propensity score–matched cohort study using de-identified electronic health record data from the TriNetX Research Network (TriNetX LLC, Cambridge, MA, USA). The network aggregates standardized data from multiple healthcare organizations, including diagnoses, procedures, medications, and laboratory results, encoded using common clinical terminologies such as ICD-10-CM, CPT, and LOINC. The study involved secondary analysis of existing, de-identified data without direct patient contact or intervention. The study protocol was reviewed by the Institutional Review Board of Shin Kong Wu Ho-Su Memorial Hospital and was determined to meet the criteria for exemption from formal ethical review (IRB No. 20260113R).

Study population

We identified adult patients (aged 18–99 years) with type 2 diabetes mellitus (T2DM) who underwent open reduction and internal fixation (ORIF) for ankle fractures between 2005 and 2025. Type 2 diabetes mellitus was defined by the presence of ICD-10-CM code E11 recorded at any time before or on the index date. The index event was defined as the first inpatient encounter with a procedure code corresponding to ankle ORIF, including CPT codes 27814 (bimalleolar ORIF), 27822 (trimalleolar ORIF without posterior lip fixation), 27823 (trimalleolar ORIF with posterior lip fixation), or 27792 (lateral malleolus ORIF). Inclusion criteria required patients to be 18–99 years of age with complete demographic data; patients with missing age or invalid demographic information were excluded.

Exposure definition

Due to limited event counts in the ankle fracture ORIF population, we combined SGLT2 inhibitors and GLP-1 receptor agonists into a single exposure category to preserve statistical power and reflect contemporary clinical practice in which these agents are prescribed as core components of diabetes and cardio-renal protective strategies and are often used sequentially or in combination in high-risk patients with T2DM.^13,14 A small subset of patients likely received both an SGLT2 inhibitor and a GLP-1 receptor agonist within the 180-day preoperative window; however, prescription-level details (treatment sequence and overlap) and event counts were insufficient to support a stable, separate analysis of dual-agent users. As a result, the present study cannot provide valid class-specific estimates or compare single-agent with combination therapy.

The SGLT2i/GLP-1 RA user cohort included patients with at least one documented prescription or administration record for any SGLT2 inhibitor or GLP-1 receptor agonist within 180 days prior to the index ORIF. This window was selected to enrich for patients receiving ongoing cardiometabolic therapy while retaining sufficient sample size and event counts for propensity score matching. Because prescription start and stop dates, refills, and adherence data are incompletely captured across the underlying electronic health records, we were unable to reliably construct a narrower, high-fidelity exposure window that ensured active use at the exact time of surgery. The non-user cohort comprised patients without any recorded prescription or administration of SGLT2 inhibitors or GLP-1 RAs during the same 180-day preoperative window.

Outcomes

The primary outcome was a composite of surgical site infection or infection-related implant removal occurring within 90 days after the index ORIF. Secondary outcomes included a composite of surgical site infection or infection-related implant removal within 1 year, as well as all-cause mortality within 1 year. Surgical site infection was identified using ICD-10-CM diagnosis codes for infection following a procedure (T81.4) and sepsis following a procedure (T81.44). Infection-related implant removal was identified using procedure codes for debridement of bone or soft tissue around the ankle or hindfoot, or irrigation and debridement, in the presence of infection-related diagnosis codes. Nonunion and delayed union were recorded when available but were rare events in this cohort. Because event counts were too low for reliable inferential analysis, these endpoints were not included among the primary or secondary outcomes. For each endpoint, we excluded patients with evidence of that same outcome recorded during the baseline period before the start of the respective follow-up window to ensure that only incident postoperative events were analyzed. Thus, patients with SSI or infection-related implant removal documented before the 90-day or 1-year risk windows were excluded from the corresponding analyses, and patients with any record of death before the start of the 1-year mortality window were excluded from the mortality analysis. The numbers of excluded patients for each endpoint are reported in the footnote to Table 2.

Table 2.

Outcomes after ankle fracture ORIF in patients with type 2 diabetes receiving SGLT2 inhibitors or GLP-1 receptor agonists versus non-users.

	Cohort	Patients, n	Events, n	Risk (%)	Risk ratio (95% CI)	Hazard ratio (95% CI)	Log-rank p
90-day SSI or infection-related implant removal	SGLT2 or GLP-1 user	894	13	1.45	0.23 (0.13- 0.42)	0.34 (0.19- 0.62)	0.0002
	Non-user	1055	66	6.26	-	-	-
1-year SSI or infection-related implant removal	SGLT2 or GLP-1 user	892	16	1.79	0.15 (0.09-0.25)	0.42 (0.25- 0.72)	0.0013
	Non-user	1057	126	11.92	-	-	-
1-year all-cause mortality	SGLT2 or GLP-1 user	1281	12	0.94	0.15 (0.08-0.27)	0.61 (0.31- 1.18)	0.1393
	Non-user	1284	80	6.23	-	-	-

*Patients who experienced the outcome before the start of the respective time window were excluded from each analysis as follows: for 90-day SSI or infection-related implant removal, 395 SGLT2/GLP-1 users and 234 non-users; for 1-year SSI or infection-related implant removal, 395 users and 230 non-users; and for 1-year all-cause mortality, 10 users and 10 non-users.

SSI, surgical site infection; ORIF: open reduction and internal fixation; SGLT2: sodium-glucose co-transporter 2 inhibitor; GLP-1: Glucagon-like peptide-1; CI: Confidence Interval.

Baseline covariates and propensity score matching

Baseline covariates were assessed using all available data from each patient’s record during the 1-year period before the index ORIF. The following baseline variables were included as covariates: demographics (age at index and sex), comorbidities [overweight or obesity (ICD-10-CM E66), ischemic heart disease (I20–I25), chronic kidney disease (N18), heart failure (I50), smoking-related diagnoses including personal history of nicotine dependence (Z87.891), nicotine dependence (F17), and tobacco use (Z72.0)], and glycemic control as measured by hemoglobin A1c. Fracture-related characteristics included fracture pattern [trimalleolar (S82.85), bimalleolar (S82.84), and lateral malleolar (S82.6 or S82.83)], closed fractures (S82.891 A or S82.892 A), and an open injury composite defined as the presence of any open fracture (S82.891 B, S82.892 B, or S82.892 C) or open wound of the ankle, foot, or toes (S91.x).

Propensity score matching was performed using a 1:1 greedy nearest-neighbor algorithm without replacement to balance covariates between users and non-users. The propensity score model included all baseline variables listed above. A caliper width of 0.1 standard deviations of the logit of the propensity score was applied. Covariate balance was assessed using standardized mean differences (SMDs), with SMD values less than 0.1 considered indicative of acceptable balance. Baseline characteristics before and after matching are presented in Table 1.

Table 1.

Baseline characteristics before and after propensity score matching in patients with type 2 diabetes undergoing ankle fracture open reduction and internal fixation.

	Before propensity score matching			After propensity score matching
	SGLT2 or GLP-1 user (n = 1295)	Non-user (n = 17,516)	SMD	SGLT2 or GLP-1 user (n = 1289)	Non-user (n = 1289)	SMD
Demographics
Age at index, years, mean ± SD	62.7 ± 12.7	61.5 ± 14.2	0.09	62.7 ± 12.8	63.0 ± 13.0	0.02
Female, n (%)	844 (65.2)	11,279 (64.4)	0.02	840 (65.2)	876 (68.0)	0.06
Male, n (%)	451 (34.8)	6235 (35.6)	0.02	449 (34.8)	413 (32.0)	0.06
Comorbidities
Overweight or obesity, n (%)	625 (48.3)	5415 (30.9)	0.36	619 (48.0)	613 (47.6)	0.01
Ischemic heart disease, n (%)	410 (31.7)	3400 (19.4)	0.28	404 (31.3)	398 (30.9)	0.01
Chronic kidney disease, n (%)	392 (30.3)	3188 (18.2)	0.28	387 (30.0)	387 (30.0)	0.00
Heart failure, n (%)	329 (25.4)	2241 (12.8)	0.33	323 (25.1)	314 (24.4)	0.02
Hemoglobin A1c,%, mean ± SD	7.18 ± 1.66	7.16 ± 1.81	0.01	7.18 ± 1.65	7.09 ± 1.83	0.05
Personal history of NicotineDependence(Z87.891), n (%)	272 (21.0)	3074 (17.6)	0.09	271 (21.0)	248 (19.2)	0.04
Nicotine dependence (F17), n (%)	190 (14.7)	2570 (14.7)	0.00	188 (14.6)	187 (14.5)	0.00
Tobacco use (Z72.0), n (%)	60 (4.6)	574 (3.3)	0.07	59 (4.6)	54 (4.2)	0.02
Fracture pattern
Trimalleolar fracture, n (%)	165 (12.7)	5668 (32.4)	0.48	165 (12.8)	163 (12.6)	0.00
Bimalleolar fracture, n (%)	156 (12.0)	5580 (31.9)	0.49	156 (12.1)	142 (11.0)	0.03
Lateral malleolar fracture, n (%)	229 (17.7)	6788 (38.8)	0.34	229 (17.8)	232 (18.0)	0.01
Closed fracture, n (%)	161 (12.4)	5198 (29.7)	0.29	161 (12.5)	149 (11.6)	0.03
Open injury composite, n (%)	70 (5.4)	940 (5.4)	0.00	67 (5.2)	66 (5.1)	0.01

SGLT2: sodium-glucose co-transporter 2 inhibitor; GLP-1: Glucagon-like peptide-1; SMD: Standardized mean difference; SD: standared deviation.

Statistical analysis

In the matched cohorts, we calculated event counts, event rates (%), and risk ratios (RRs) with 95% confidence intervals (CIs) for each outcome. Time-to-event analyses included Kaplan–Meier survival curves and Cox proportional hazards models to estimate hazard ratios (HRs) with 95% CIs. Differences between survival curves were assessed using the log-rank test. The proportional hazards assumption was evaluated using standard diagnostic tests available in the TriNetX platform. All statistical tests were two-sided with α = 0.05. Analyses were performed using the TriNetX analytical platform with data standardized using ICD-10-CM diagnosis codes and CPT procedure codes.

To assess the robustness of our findings, we additionally performed two sensitivity analyses. First, we repeated the analyses after restricting the exposure group to single-agent users (patients prescribed either an SGLT2 inhibitor or a GLP-1 receptor agonist, but not both). Second, we conducted an exploratory analysis defining active users as patients with documented SGLT2i or GLP-1 RA prescriptions within 30 days before ankle ORIF, and compared them with non-users using the same analytical framework where feasible.

Results

Study population and baseline characteristics

Before propensity score matching, the SGLT2i/GLP-1 RA user cohort comprised 1295 patients, and the non-user cohort comprised 17,516 patients, for a total of 18,811 eligible patients. Baseline characteristics differed significantly between groups before matching, particularly regarding comorbidity burden (obesity, ischemic heart disease, chronic kidney disease, heart failure) and fracture severity.

After 1:1 propensity score matching, 1289 users were paired with 1289 non-users. Baseline characteristics were well balanced between matched groups (Table 1), with all standardized mean differences <0.1. The matched cohort had a mean age of approximately 63 years, with approximately 65–68% female participants.

Approximately 48% had obesity, 31% had ischemic heart disease, 30% had chronic kidney disease, and 25% had heart failure. Smoking-related diagnoses were documented in approximately 14–21% of participants. Fracture patterns—including trimalleolar (12.6–12.8%), bimalleolar (11.0–12.1%), and lateral malleolar fractures (17.8–18.0%)—were balanced between groups. The open injury composite was present in approximately 5% of each group.

90-Day SSI or infection-related implant removal

The 90-day composite outcome of SSI or infection-related implant removal occurred in 13 of 894 SGLT2i/GLP-1 RA users (1.45%) compared with 66 of 1055 non-users (6.26%) (Table 2). The risk ratio was 0.23 (95% CI 0.13–0.42), representing a 77% relative risk reduction.

Kaplan–Meier analysis demonstrated early and sustained divergence of survival curves, with the user group maintaining higher event-free survival throughout the 90-day observation period. The hazard ratio was 0.34 (95% CI 0.19–0.62), and the log-rank test was highly significant (p = 0.0002). These findings indicate a substantial protective association of SGLT2i or GLP-1 RA use with infection-related implant failure in the early postoperative period.

1-Year SSI or infection-related implant removal

Over the 1-year follow-up period, SSI or infection-related implant removal occurred in 16 of 892 SGLT2i/GLP-1 RA users (1.79%) and 126 of 1057 non-users (11.92%) (Table 2). The risk ratio was 0.15 (95% CI 0.09–0.25), representing an 85% relative risk reduction. The hazard ratio from the Cox model was 0.42 (95% CI 0.25–0.72), with log-rank p = 0.0013, indicating sustained protection throughout the 1-year period. The magnitude and consistency of the protective effect across both time windows, together with the relatively narrow confidence intervals, support an association between cardiometabolic agent use and reduced infection-related implant failure in this high-risk population.

1-Year all-cause mortality

One-year all-cause mortality was 0.94% (12 of 1281 SGLT2i/GLP-1 RA users) compared with 6.23% (80 of 1284 non-users) (Table 2). The risk ratio was 0.15 (95% CI 0.08–0.27), indicating an 85% relative risk reduction in mortality.

Time-to-event analysis using the Cox proportional hazards model yielded a hazard ratio of 0.61 (95% CI 0.31–1.18). The log-rank test did not reach conventional statistical significance (p = 0.1393), although the point estimate and Kaplan–Meier curves demonstrated a trend toward lower mortality in SGLT2i/GLP-1 RA users. The discrepancy between the risk-ratio and time-to-event analyses likely reflects the relatively small absolute number of deaths (92 total) and limited statistical power, rather than a true absence of protective effect.

Sensitivity analyses

In a sensitivity analysis excluding dual-agent users, single-agent SGLT2i/GLP-1 RA users continued to demonstrate markedly lower 1-year risk of SSI or infection-related implant removal compared with non-users (1.42% vs 14.04%; hazard ratio 0.26, 95% CI 0.14–0.48). One-year all-cause mortality was also reduced (0.91% vs 6.61%; hazard ratio 0.41, 95% CI 0.21–0.80). Full numeric results for this sensitivity analysis are presented in Supplemental Table 1.

In an exploratory 30-day active-use analysis, patients with documented SGLT2i/GLP-1 RA prescriptions within 30 days before surgery experienced no observed events of 1-year SSI or infection-related implant removal (0/133) or all-cause mortality (0/194), whereas non-users had event rates of 14.8% (23/155) and 8.7% (17/196), respectively (both p < 0.001). Because of zero events in the active-user group, hazard ratios could not be estimated reliably; nevertheless, these findings are directionally consistent with the primary results and suggest that more proximate medication use may confer at least comparable protection.

Discussion

This multicenter real-world evidence study demonstrates that preoperative use of SGLT2 inhibitors or GLP-1 receptor agonists is associated with substantially reduced risk of infection-related implant failure in patients with type 2 diabetes undergoing ankle fracture ORIF. The magnitude of effect is clinically significant: a 77% relative risk reduction at 90 days and an 85% relative risk reduction at 1 year. Regarding all-cause mortality, it is noteworthy that while the reduction did not reach significance in the primary matched cohort (HR 0.61; p = 0.1393), this specific analysis may have been limited by insufficient statistical power, as indicated by the wide confidence intervals. However, a clear and robust survival benefit emerged in our more stringent sensitivity analyses. Specifically, mortality was significantly lower in single-agent users (p = 0.007) and dropped to 0% in the 30-day active-user group (p < 0.001). This transition from a non-significant trend to robust statistical significance suggests that the systemic protection of SGLT2is and GLP-1 RAs becomes most evident when confounding factors, such as dual-agent interactions or remote medication exposure, are minimized.

These findings represent important new evidence in several respects. While prior studies have examined GLP-1 RA use in ankle fracture surgery with conflicting results—some showing infection protection and others demonstrating increased wound complications^11,12—SGLT2 inhibitor use in ankle fracture surgery has been virtually unstudied. This is the first large-scale database study to evaluate a combined SGLT2i/GLP-1 RA exposure in the specific context of ankle fracture ORIF, approximating a broader cardiometabolic optimization strategy.¹⁵

The protective associations observed may operate through several biological mechanisms. Both SGLT2 inhibitors and GLP-1 receptor agonists improve glycemic control, which correlates with better perioperative outcomes and reduced infection risk.¹⁶ GLP-1 RAs promote weight loss,¹⁷ while SGLT2 inhibitors have neutral to modest effects on weight^18,19; obesity is a well-established risk factor for post-operative SSI and implant-related infection.^20,21

Both drug classes exert anti-inflammatory and immunomodulatory effects beyond glucose lowering, including reduced pro-inflammatory cytokine production and improved innate immune responses.^22,23 These properties may enhance the host’s ability to combat perioperative bacterial colonization. Additionally, SGLT2 inhibitors provide cardio-renal protective effects through blood pressure reduction and improved renal perfusion,²⁴ while GLP-1 RAs improve lipid profiles and reduce cardiovascular risk,²⁵ all of which may indirectly reduce perioperative complications and mortality.

This study employed a combined SGLT2i/GLP-1 RA exposure category rather than analyzing agents separately. This decision was driven by insufficient event numbers within individual drug classes, which precluded robust statistical inference for separate cohorts and prevented meaningful head-to-head comparison. This approach is consistent with contemporary diabetes care, where SGLT2 inhibitors and GLP-1 receptor agonists are increasingly used sequentially or in combination to intensify glycemic control and reduce cardiovascular and renal risk in high-risk patients.²⁶ However, this design limitation precludes determination of which agent (or combination) drives the observed protective effect. Future prospective studies with larger populations should assess SGLT2is and GLP-1 RAs separately and examine potential differential effects on fracture-related outcomes.

Our findings align with observations from orthopedic and general surgical studies. Retrospective analyses of GLP-1 RA use in total knee and hip arthroplasty have reported approximately 30–40% relative reductions in periprosthetic joint infection risk,^27–29 which is broadly consistent with the magnitude of risk reduction observed in our study. However, discordant results in ankle fracture surgery—with some studies reporting increased wound complications^11,12—may reflect anatomical or biological differences between arthroplasty and fracture repair, or differences in study populations and confounding adjustment.

Our propensity score–matched design with comprehensive baseline covariate adjustment provides a methodologically stronger foundation for causal inference than many prior observational studies. Nonetheless, causality cannot be definitively established from observational data.

This study has several strengths. It leverages a large, multicenter cohort drawn from more than 170 healthcare organizations worldwide and applies rigorous 1:1 propensity score matching with excellent covariate balance, as indicated by standardized mean differences below 0.1 across all variables. Fracture severity is characterized in detail, distinguishing open versus closed injuries and specific fracture patterns, and the analysis focuses on clinically meaningful outcomes—surgical site infection, infection-related reoperation, and mortality—defined in a standardized manner. Confounding is addressed comprehensively through adjustment for key factors such as smoking status, chronic kidney disease, cardiac comorbidities, and obesity, and the direction and magnitude of effect estimates are consistent across both 90-day and 1-year follow-up windows.

However, several limitations should be acknowledged. First, we employed a combined SGLT2i/GLP-1 RA exposure category rather than analyzing agents separately. This approach was driven by insufficient event numbers within individual drug classes and among dual-agent users, which precluded robust statistical inference for separate cohorts and prevented meaningful head-to-head comparison. Consequently, the present study cannot determine whether one class, both classes, or their combination drives the observed protective effect. The combined exposure should therefore be interpreted as a proxy for a broader cardiometabolic optimization strategy, rather than as evidence that any specific drug class alone is responsible for the risk reduction. The single-agent sensitivity analysis, which excluded dual-agent users, yielded effect estimates that were consistent in direction and magnitude with the primary analysis, supporting the robustness of our findings despite this design limitation.

Second, exposure to SGLT2 inhibitors or GLP-1 receptor agonists was defined as any recorded use within 180 days before surgery, which does not guarantee that all classified “users” were actively taking the medication at the time of ORIF. Discontinuation, non-adherence, or temporary use may have led to non-differential misclassification of exposure status. Such misclassification would be expected to bias the associations toward the null, suggesting that the true protective effect, if causal, could be at least as large as the observed estimates. Our exploratory 30-day active-use analysis, in which no infection-related implant failures or deaths were observed among active users, was directionally consistent with the main analysis and supports the hypothesis that more proximate medication use may be associated with greater protection. However, these findings should be interpreted cautiously due to small sample size and zero events in the exposed group.

Third, as a retrospective observational study, our analysis remains vulnerable to residual confounding from unmeasured or incompletely measured factors such as surgical technique, intraoperative antibiotic regimens, soft-tissue management, nutritional status, and medication adherence. These variables may influence both the likelihood of receiving SGLT2 inhibitors or GLP-1 receptor agonists and the risk of postoperative infection or mortality, and cannot be fully accounted for in our propensity score model.

Fourth, outcome ascertainment relied on ICD-10-CM diagnosis and procedure codes for surgical site infection, sepsis, and infection-related implant removal, which are subject to documentation practices and coding variability across institutions. Misclassification of outcomes is therefore possible, and differential coding between exposure groups cannot be completely excluded.

Fifth, detailed fracture grading using systems such as the Gustilo–Anderson classification is not available in structured electronic data within the network. We used diagnostic codes as proxies to distinguish open versus closed injuries and to capture fracture pattern, but these proxies may not fully reflect injury severity or soft-tissue compromise, which are important determinants of infection risk and implant failure.

Finally, the relatively small absolute number of deaths limits the precision of the mortality estimates, and the apparent trend toward reduced mortality should be interpreted as exploratory.

Our findings suggest that SGLT2 inhibitors and GLP-1 receptor agonists may represent reasonable components of comprehensive preoperative optimization for patients with type 2 diabetes undergoing ankle fracture ORIF, particularly those at high risk for postoperative infection. These agents should be integrated thoughtfully with established infection-prevention measures, glycemic control strategies, smoking cessation support, and nutritional interventions, and prospective studies with multidisciplinary input are needed before routine perioperative initiation can be recommended.^30,31 We found no signal of increased risk associated with preoperative SGLT2 inhibitor or GLP-1 receptor agonist use in terms of surgical site infection, infection-related implant removal, or all-cause mortality. Recent concerns raised in some ankle fracture series about GLP-1 RA–associated nonunion or delayed wound healing were not corroborated in our primary outcomes; however, our study was underpowered to robustly evaluate nonunion given the low event counts. Given the inherent limitations of observational data, prospective randomized or well-controlled observational studies are needed to confirm causality, define optimal patient populations, and clarify underlying mechanisms.

Future research should include prospective studies with standardized outcome definitions and detailed collection of perioperative and behavioral variables, as well as well-controlled randomized trials in high-risk subgroups. In addition, larger datasets are needed to separately evaluate SGLT2 inhibitors and GLP-1 receptor agonists and to examine more granular fracture-healing outcomes, including nonunion and delayed union.

Conclusions

Among patients with type 2 diabetes undergoing ankle fracture ORIF, preoperative use of SGLT2 inhibitors or GLP-1 receptor agonists was associated with significantly lower 90-day and 1-year risks of infection-related implant failure compared with non-use. Additionally, a consistent signal toward reduced 1-year all-cause mortality was observed, although statistical power was limited. These findings support including modern cardiometabolic agents as part of perioperative optimization strategies in high-risk ankle fracture patients and highlight the need for prospective studies to confirm causality and clarify mechanisms.

Supplemental material

Supplemental Material - Protective effects of SGLT2 inhibitors and GLP-1 receptor agonists on infection-related implant failure following ankle fracture ORIF: A global big data analysis

Supplemental Material for Protective effects of SGLT2 inhibitors and GLP-1 receptor agonists on infection-related implant failure following ankle fracture ORIF: A global big data analysis by Mao Yi Yang, Li Wei Hung in Journal of Orthopaedic Surgery

Footnotes

Ethical considerations

The Institutional Review Board of our institution (No. 20260113R) waived the requirement for informed consent because only de-identified data from a global federated research network were analyzed.

Funding

Financial support from the Ministry of Science and Technology, Taiwan (ROC) (NSTC113-2221-E-341-002) and the Shin Kong Wu Ho-Su Memorial Hospital, Tawan (2021SKHADR017) are greatly appreciated.

Declaration of conflicting interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.*

Supplemental material

Supplemental material for this article is available online.

Appendix

References

Rizvi

Chillag

. Perioperative management of diabetes and hyperglycemia in patients undergoing orthopaedic surgery. J Am Acad Orthop Surg 2010; 18(7): 426–435. https://doi.org/10.5435/00124635-201007000-00005

Danilkowicz

Grimm

Kim

, et al. Increasing age and modifiable comorbidities are associated with short-term complications after open reduction and internal fixation of ankle fractures. Eur J Orthop Surg Traumatol 2022; 32(1): 113–119. https://doi.org/10.1007/s00590-021-02927-z

Gougoulias

Oshba

Dimitroulias

, et al. Ankle fractures in diabetic patients. EFORT Open Rev 2020; 5(8): 457–463. https://doi.org/10.1302/2058-5241.5.200025

Ahmad

Azmat Swati

Khan

, et al. Impact of smoking as a risk factor for surgical site infections after open reduction and internal fixation (ORIF) of distal radius fractures. Cureus 2025; 17(10): e94649. https://doi.org/10.7759/cureus.94649

Luan

Zheng

, et al. Evaluating risk factors for surgical site infections following open reduction and internal fixation surgery for ankle fractures: a systematic review and meta-analysis. Front Public Health 2025; 13: 1558994. https://doi.org/10.3389/fpubh.2025.1558994

Mastrokostas

Monas

Rodriguez

, et al. Patient demographics and risk factors for surgical site infections after open reduction and internal fixation for bimalleolar ankle fractures. J Am Acad Orthop Surg. 2025; Published online August 28, 2025. https://doi.org/10.5435/JAAOS-D-24-01181

García-Vega

Cinza-Sanjurjo

Tilves-Bellas

, et al. Sodium-glucose cotransporter-2 inhibitors and glucagon-like peptide 1 receptor agonists and cancer mortality. A real-world registry. Rev Esp Cardiol 2025; 78(3): 218–228. https://doi.org/10.1016/j.rec.2024.07.003

Steven

Kuntic

Münzel

, et al. Modern antidiabetic therapy by sodium-glucose cotransporter 2 inhibitors, glucagon-like peptide 1 receptor agonists, and dipeptidyl peptidase 4 inhibitors against cardiovascular diseases. Pharmacol Rev 2025; 77(5): 100082. https://doi.org/10.1016/j.pharmr.2025.100082

Hsieh

Lai

ECC

, et al. Risk of rotator cuff tear and rotator cuff repair surgery comparison between sodium-glucose cotransporter 2 inhibitors and glucagon like peptide-1 receptor agonists: a real-world study. Diabetes Metabol 2024; 50(2): 101522. https://doi.org/10.1016/j.diabet.2024.101522

10.

Chan

Chuang

Kuo

, et al. The association between Glucagon-Like Peptide-1 receptor agonists and postoperative complications after arthroplasty: a systematic review and meta-analysis. J Arthroplast 2025; 40(12): 3080–3088.e7. https://doi.org/10.1016/j.arth.2025.06.083

11.

Morningstar

Lauck

Bank

, et al. Impact of glucagon-like peptide-1 receptor agonists on postoperative outcomes following ankle fracture open reduction and internal fixation. J Foot Ankle Surg 2025; 64(5): 644–648. https://doi.org/10.1053/j.jfas.2025.05.002

12.

Verhey

Kile

. GLP-1 agonists increase risk of surgical site infection and nonunion in patients undergoing operative fixation of ankle fractures. Foot Ankle Orthop 2025; 10(4): 2473011425S00268. https://doi.org/10.1177/2473011425S00268

13.

Bae

. SGLT2 inhibitors and GLP-1 receptor agonists in diabetic kidney disease: evolving evidence and clinical application. Diabetes Metab J 2025; 49(3): 386–402. https://doi.org/10.4093/dmj.2025.0220

14.

Yepes-Cortés

Cardenas-Moreno

Daza-Arnedo

, et al. Combining GLP-1 receptor agonists and SGLT2 inhibitors in type 2 diabetes mellitus: a scoping review and expert insights for clinical practice utilizing the nominal group technique. Diabetes Ther 2025; 16(5): 813–849. https://doi.org/10.1007/s13300-025-01722-x

15.

Rolek

Haber

Gajewska

, et al. SGLT2 inhibitors vs. GLP-1 agonists to treat the heart, the kidneys and the brain. J Cardiovasc Dev Dis 2023; 10(8): 322. https://doi.org/10.3390/jcdd10080322

16.

Zeng

Chen

Hsu

, et al. The effect of GLP-1 receptor agonist and SGLT2 inhibitor on infection risk: network meta-analysis. Clin Microbiol Infect 2026; 27: S1198–743X(26), 00019-4. https://doi.org/10.1016/j.cmi.2026.01.014

17.

Moiz

Filion

Tsoukas

, et al. Mechanisms of GLP-1 receptor agonist-induced weight loss: a review of central and peripheral pathways in appetite and energy regulation. Am J Med 2025; 138(6): 934–940. https://doi.org/10.1016/j.amjmed.2025.01.021

18.

Osto

Bonacina

Pirillo

, et al. Neutral effect of SGLT2 inhibitors on lipoprotein metabolism: from clinical evidence to molecular mechanisms. Pharmacol Res 2023; 188: 106667. https://doi.org/10.1016/j.phrs.2023.106667

19.

Janež

Fioretto

. SGLT2 inhibitors and the clinical implications of associated weight loss in type 2 diabetes: a narrative review. Diabetes Ther 2021; 12(8): 2249–2261. https://doi.org/10.1007/s13300-021-01104-z

20.

Hourigan

. Impact of obesity on surgical site infection in Colon and rectal surgery. Clin Colon Rectal Surg 2011; 24(4): 283–290. https://doi.org/10.1055/s-0031-1295691

21.

Meijs

Koek

MBG

Vos

, et al. The effect of body mass index on the risk of surgical site infection. Infect Control Hosp Epidemiol 2019; 40(9): 991–996. https://doi.org/10.1017/ice.2019.165

22.

Lee

Riella

. Narrative review of immunomodulatory and anti-inflammatory effects of sodium-glucose cotransporter 2 inhibitors: unveiling novel therapeutic frontiers. Kidney Int Rep 2024; 9(6): 1601–1613. https://doi.org/10.1016/j.ekir.2024.02.1435

23.

Ullah

Shen

. Immunomodulatory effects of anti-diabetic therapies: cytokine and chemokine modulation by metformin, sodium-glucose cotransporter 2 inhibitors, and glucagon-like peptide-1 receptor agonists (2013–2025). Eur J Med Chem 2025; 299: 118065. https://doi.org/10.1016/j.ejmech.2025.118065

24.

Zhu

Yang

, et al. Chronic cardiorenal syndrome: cardio-renal protective effect of SGLT2i. Ren Fail 2025; 47(1): 2575921. https://doi.org/10.1080/0886022X

25.

Westermeier

Fisman

. Glucagon-like peptide-1 receptor agonists (GLP-1RAs) and cardiometabolic protection: historical development and future challenges. Cardiovasc Diabetol 2025; 24: 44. https://doi.org/10.1186/s12933-025-02608-9

26.

Liakos

Karagiannis

Avgerinos

, et al. SGLT-2 inhibitors and GLP-1 receptor agonists as combination therapy in type 2 diabetes. Curr Diab Rep 2026; 26(1): 1. https://doi.org/10.1007/s11892-025-01616-z

27.

Impact of glucagon-like Peptide-1 receptor agonists on postoperative outcomes in arthroplasty: a systematic review ClinicalKey 臨床醫學資料庫

[cited 2026 Feb 4]. Available from: https://www.clinicalkey.com/#!/content/playContent/1-s2.0-S0883540325008782?returnurl=https:%2F%2Flinkinghub.elsevier.com%2Fretrieve%2Fpii%2FS0883540325008782%3Fshowall%3Dtrue&referrer=https:%2F%2Fpubmed.ncbi.nlm.nih.gov%2F

28.

Chan

Chuang

Kuo

29.

Buddhiraju

Kagabo

Khanuja

, et al. Decreased risk of readmission and complications with preoperative GLP-1 analog use in patients undergoing primary total joint arthroplasty. J Arthroplast 2024; 39(12): 2911–2915.e1. https://doi.org/10.1016/j.arth.2024.05.079

30.

Safi

Alyahya

Al-sheyab

, et al. The impact of Multidisciplinary Preoperative optimization program on postoperative outcomes among surgical oncology patients. J Clin Med 2025; 14(21): 7820. https://doi.org/10.3390/jcm14217820

31.

Lussiez

. Smoking cessation for preoperative optimization. Clin Colon Rectal Surg 2023; 36(3): 175–183. https://doi.org/10.1055/s-0043-1760870

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.37 MB