Translate this page into:
Metformin is associated with 90-day decreased risk of death and AKI after shoulder arthroplasty in diabetic patients
⁎Corresponding author: Nathan Lanham. nlanham@pennstatehealth.psu.edu
-
Received: ,
Accepted: ,
This article was originally published by Reed Elsevier India Pvt. Ltd. and was migrated to Scientific Scholar after the change of Publisher.
Abstract
Abstract
Metformin is a widely used antidiabetic medication with emerging evidence suggesting anti-inflammatory effects, reduced musculoskeletal pain, and lower complication rates following joint arthroplasty. However, its impact on outcomes after shoulder arthroplasty remains unclear. The purpose of this study was to evaluate the association between metformin use and postoperative complications in diabetic patients undergoing shoulder arthroplasty, with the hypothesis that metformin users would experience fewer perioperative complications and reduced postoperative analgesic requirements.
Using the TriNetX database, diabetic patients undergoing shoulder arthroplasty were identified and stratified into two cohorts based on metformin prescription within one year prior to surgery. 1:1 propensity score matching was performed to balance relevant baseline comorbidities. Ninety-day postoperative medical and surgical complications and 2-year implant-related outcomes were then compared between cohorts, with statistical significance set at p < 0.05.
After matching, each cohort consisted of 5,181 patients. Patients prescribed metformin had a significantly decreased 90-day risk of death (0.31% vs 0.66%, p = 0.011) and acute kidney injury (3.05% vs 3.69%, p = 0.027) compared to the control cohort. No significant difference was observed in 90-day infectious complications including prosthetic joint infection (PJI), revision surgery and number of opioid prescriptions. Similarly, at 2 years postoperatively, rates of PJI, loosening, periprosthetic fracture and revision were all comparable between both cohorts.
Metformin use was associated with lower 90-day risks of mortality and acute kidney injury in diabetic patients undergoing shoulder arthroplasty, without an increased risk of implant-related complications. These findings support the perioperative safety of metformin and may help inform surgical decision-making and patient counseling
III; retrospective study
Keywords
Metformin
Arthroplasty
Complications
Total shoulder arthroplasty
Antidiabetics
1 Introduction
Adverse outcomes following joint replacement are frequently associated with diabetes, one of the most common comorbidities in this population.1 The incidence of diabetes mellitus has steadily increased over the past decade, with developed countries projected to experience a 20% rise in prevalence by 2030.2 Additionally, diabetes has been established as a significant risk factor for postoperative complications following joint replacement surgery, especially as it has been associated with increased risk of surgical site and joint infections following shoulder arthroplasty (SA).3
Metformin is a first-line therapy for type 2 diabetes and is believed to have additional clinical benefits beyond glucose metabolism. Recent studies have shown that metformin may play a role in modulating inflammatory pathways, tempering oxidative stress, and improving cartilage health in degenerative joints.4 Most importantly, metformin has been associated with a reduced risk of total knee and hip arthroplasty,5 as well as decreased risk of postoperative complications, most notably prosthetic joint infection, deep vein thrombosis, and revision surgery following total knee arthroplasty.6
In the setting of increasing diabetes prevalence, and subsequent potential increased metformin use, it is important to understand metformin's effects on complication rates in diabetic patients undergoing SA, as this remains undefined. As such, we aim to compare the incidence of adverse outcomes following SA between patients who are on metformin therapy and those who are not. Given metformin's known anti-inflammatory and cardiometabolic effects, understanding its impact in the surgical setting may help clarify its role in modulating perioperative risk. The purpose of this study is to evaluate the association between preoperative metformin use and postoperative complication rates following TSA, with the goal of better understanding its potential impact on surgical outcomes.
2 Methods
This study utilized data from TriNetX, a research database which consolidates de-identified electronic health records from participating healthcare organizations. The database is de-identified in compliance with the Health Insurance Portability and Accountability Act (HIPAA) Privacy Rule through standardized methods prior to researcher access. Therefore, this study was exempt from institutional review board (IRB) approval. All data were queried using Current Procedural Terminology (CPT), International Classification of Diseases, 10th Revision (ICD-10) and Systematized Nomenclature of Medicine (SNOMED) codes. The summary of the query codes used can be found in [Supplementary Table 1].
The Global Collaborative Network on TriNetX was queried for patients undergoing SA. Patients younger than 18 years of age and those without type 2 diabetes were excluded. The remaining cohort was stratified based on the presence of a preoperative prescription of metformin. The first cohort was comprised of patients prescribed metformin (within one year and up to 14 days before surgery) and the control cohort was comprised of patients not prescribed metformin.
Propensity score matching (PSM) was performed using the nearest-neighbor matching method using a tolerance threshold of 0.01, with matches constrained to propensity score differences of P ≤ 0.1. We accounted for multiple patient characteristics and relevant comorbidities including: age at surgery, sex, insulin use, hypertension, hyperlipidemia, ischemic heart disease, heart failure, chronic kidney disease, liver disease, and tobacco use, renal function (estimated glomerular filtration rate eGFR <15, 15–30, 30–45, 45–60, 60–90, and >90 mL/min/1.73 m2), glycemic control categorized by HbA1c (0–5.6%, 5.6–6.5%, 6.5–7.5%, 7.5–8.5%, 8.5–9%, and ≥9%), and body mass index (≤25, 25–30, 30–40, and ≥40 kg/m2).
Outcomes were evaluated between cohorts and included 90-day postoperative medical complications (urinary tract infection, pneumonia, sepsis, stroke, venous thromboembolism, acute kidney injury, readmission, and mortality), opioid prescriptions, and 2-year surgical outcomes (wound disruption, periprosthetic joint infection, mechanical loosening, periprosthetic fracture, and revision surgery).
For statistical analysis, we used the TriNetX software to calculate risk percentage, risk ratio (RR), and p-values. For categorical variables, the chi-squared test was performed, while the student's t-test was utilized for continuous variables. Significance was set for α < 0.05.
3 Results
15,438 adult diabetic patients undergoing SA were initially included in our study. 5,357 patients were prescribed metformin compared to 10,081 controls. After propensity score matching (PSM), each cohort was comprised of 5,181 patients. All baseline demographics were adequately matched between cohorts except eGFR with lower percentage of patients with eGFR 15-30 (7.01% vs 12.57%, standardized mean difference 0.188) and eGFR 30-45 (23.16% vs 28.51%, standardized mean difference 0.122) in the metformin cohort. The characteristics of both cohorts after PSM are shown in Table 1.
| Characteristic Name | Before matching | After matching | ||||
| Metformin (n = 5,357) | Control (n = 10,081) | Standardized Mean Difference | Metformin (n = 5,181) | Control (n = 5,181) | Standardized Mean Difference | |
| Age at Index | 68.98 ± 7.98 | 69.90 ± 8.51 | 0.111 | 69.11 ± 7.95 | 69.12 ± 8.29 | 0.001 |
| Female | 50.23% | 55.00% | 0.095 | 50.98% | 51.38% | 0.008 |
| Male | 47.30% | 41.98% | 0.107 | 46.52% | 46.21% | 0.006 |
| Insulin use | 54.23% | 48.43% | 0.116 | 53.72% | 53.52% | 0.004 |
| Hypertensive diseases | 92.46% | 91.61% | 0.031 | 92.34% | 92.63% | 0.011 |
| Disorders of lipoprotein metabolism and other lipidemias | 87.83% | 82.99% | 0.137 | 87.47% | 87.17% | 0.009 |
| Ischemic heart diseases | 37.71% | 39.82% | 0.043 | 38.00% | 37.04% | 0.020 |
| Chronic kidney disease (CKD) | 22.55% | 30.12% | 0.172 | 23.22% | 23.20% | <0.001 |
| Diseases of liver | 20.40% | 20.15% | 0.006 | 20.40% | 20.25% | 0.004 |
| Nicotine dependence | 18.09% | 17.52% | 0.015 | 17.91% | 17.33% | 0.015 |
| Heart failure | 17.30% | 20.90% | 0.092 | 17.78% | 17.35% | 0.011 |
| Tobacco use | 5.17% | 5.63% | 0.021 | 5.17% | 6.04% | 0.038 |
| Estimated Glomerular filtration rate | ||||||
| At most 15 mL/min/{1.73_m2} | 5.58% | 6.33% | 0.032 | 5.69% | 5.62% | 0.003 |
| 15-30 mL/min/{1.73_m2} | 6.81% | 14.26% | 0.244 | 7.01% | 12.57% | 0.188 |
| 30-45 mL/min/{1.73_m2} | 22.76% | 31.39% | 0.195 | 23.16% | 28.51% | 0.122 |
| 45-60 mL/min/{1.73_m2} | 49.00% | 52.40% | 0.068 | 49.33% | 51.61% | 0.046 |
| 60-90 mL/min/{1.73_m2} | 82.57% | 77.32% | 0.131 | 82.36% | 80.74% | 0.042 |
| At least 90 mL/min/{1.73_m2} | 56.19% | 47.59% | 0.173 | 55.47% | 52.58% | 0.058 |
| Hemoglobin A1c | ||||||
| 0-5.6 % | 17.21% | 21.89% | 0.118 | 17.43% | 17.33% | 0.003 |
| 5.6-6.5 % | 62.78% | 57.39% | 0.110 | 62.19% | 61.42% | 0.016 |
| 6.5-7.5 % | 65.43% | 49.58% | 0.325 | 64.25% | 64.02% | 0.005 |
| 7.5-8.5 % | 41.35% | 26.91% | 0.308 | 39.47% | 39.63% | 0.003 |
| 8.5-9 % | 17.72% | 11.44% | 0.179 | 16.70% | 16.54% | 0.004 |
| At least 9 % | 22.53% | 15.89% | 0.169 | 21.60% | 21.46% | 0.003 |
| BMI | ||||||
| At most 25 kg/m2 | 16.20% | 18.35% | 0.057 | 16.60% | 16.23% | 0.010 |
| 25-30 kg/m2 | 38.49% | 41.17% | 0.055 | 39.01% | 39.45% | 0.009 |
| 30-40 kg/m2 | 62.37% | 62.02% | 0.007 | 62.44% | 62.75% | 0.006 |
| At least 40 kg/m2 | 26.90% | 27.66% | 0.017 | 27.18% | 26.13% | 0.024 |
3.1 90-Day postoperative medical complications
After propensity score matching, there were no significant differences in rates of urinary tract infection (3.42% vs 4.11%; RR 0.83, p = 0.063), pneumonia (2.08% vs 1.81%; RR 1.15, p = 0.320), sepsis (1.04% vs 1.27%; RR 0.82, p = 0.271), stroke (1.62% vs 1.45%; RR 1.12, p = 0.472), venous thromboembolism (1.31% vs 0.95%; RR 1.39, p = 0.077), or readmission (9.88% vs 9.57%; RR 1.03, p = 0.596). Furthermore, while significantly more patients in the metformin cohort received an opioid prescription postoperatively (80.25% vs 76.59%, RR 1.05, p < 0.001), the number of prescriptions received was similar between both cohorts (mean 3.019 ± 3.322 vs 2.966 ± 3.371, p = 0.471). However, patients prescribed metformin had significantly lower risks of acute kidney injury (3.05% vs 3.69%; RR 0.79, p = 0.027) and mortality (0.31% vs 0.66%; RR 0.47, p = 0.011).
3.2 90-Day postoperative surgical complications
Rates of wound disruption (0.60% vs 0.48%; RR 1.24, p = 0.421), periprosthetic joint infection (0.54% vs 0.33%; RR 1.65, p = 0.100) and revision surgery (1.58% vs 1.60%, RR 0.99, p = 0.937) were similar between cohorts. All 90-day outcomes are summarized in Table 2.
| 90 days outcomes | Metformin (%) | Control (%) | Risk Ratio | p-value |
| Wound disruption | 0.60% | 0.48% | 1.24 | 0.421 |
| UTI | 3.42% | 4.11% | 0.83 | 0.063 |
| PNA | 2.08% | 1.81% | 1.15 | 0.320 |
| Sepsis | 1.04% | 1.27% | 0.82 | 0.271 |
| Stroke | 1.62% | 1.45% | 1.12 | 0.472 |
| VTE | 1.31% | 0.95% | 1.39 | 0.077 |
| AKI | 3.05% | 3.69% | 0.79 | 0.027 |
| Readmission | 9.88% | 9.57% | 1.03 | 0.596 |
| Death | 0.31% | 0.66% | 0.47 | 0.011 |
| PJI | 0.54% | 0.33% | 1.65 | 0.1 |
| Revision | 1.58% | 1.60% | 0.99 | 0.937 |
| Opioid | 80.25% | 76.59% | 1.05 | <0.001 |
| Number of opioid prescription (mean±standard deviation) | 3.019 ± 3.322 | 2.966 ± 3.371 | - | 0.471 |
3.3 2-Year surgical outcomes
There were no significant differences in 2-year implant-related complications, including periprosthetic joint infection (1.20% vs 1.00%; RR 1.19, p = 0.346), mechanical loosening (0.95% vs 0.81%; RR 1.17, p = 0.461), periprosthetic fracture (0.75% vs 0.95%; RR 0.80, p = 0.284), or revision surgery (3.59% vs 3.44%; RR 1.05, p = 0.669). All 2-year outcomes are summarized in Table 3.
| 2 years outcomes | Metformin (%) | Control (%) | Risk Ratio | p-value |
| PJI | 1.20% | 1.00% | 1.19 | 0.346 |
| Mechanical loosening | 0.95% | 0.81% | 1.17 | 0.461 |
| Periprosthetic fracture | 0.75% | 0.95% | 0.80 | 0.284 |
| Revision | 3.59% | 3.44% | 1.05 | 0.669 |
4 Discussion
The most notable findings of the present study were that patients prescribed metformin had a significantly reduced risk of both acute kidney injury and death following shoulder arthroplasty. Interestingly, patients prescribed metformin had significantly more opioid prescriptions, however, the number of prescriptions among patients prescribed opiods was similar. Importantly, there was no increased risk of implant related complications at 2 years postoperatively. To the best of our knowledge, this is the first study that evaluated the effect of preoperative metformin intake on the complication risks following SA in diabetic patients.
The findings of the present study support the existing literature suggesting that preoperative metformin use is safe in the context of surgery. Overall, patients prescribed metformin undergoing shoulder arthroplasty did not experience increased perioperative risks, consistent with prior studies examining metformin in total knee arthroplasty and its favorable safety profile during surgical stress.6,7 In a large database study of 64,372 patients, Wilson et al. reported that patients not on metformin were more likely to have periprosthetic joint infections, DVT, AKI and revisions compared to their metformin-using counterparts following total knee arthroplasty.6 Similarly, Lin et al. reported decreased risk of acute renal failure following major surgery in patients on metformin compared to controls.7 Lastly, Peng et al. showed that metformin was associated with a decreased risk of AKI in patients undergoing coronary artery bypass graft surgery.8 The consistency of these findings across multiple surgical procedures reinforces the safety of metformin use in patients undergoing surgical interventions.
In addition, despite propensity score matching, there remained a significant difference in the proportion of patients with eGFR 15-30 and 30-45 between both cohorts with lower percentage in the metformin cohort. This can also explain the decreased risk of acute kidney injury observed in the metformin cohort, as well as the decreased risk of death. Postoperative acute kidney injury was found to be associated with increased risk of mortality, as reported by various studies across different surgical interventions 9–11. Furthermore, more patients prescribed metformin had opioid prescriptions compared to the control cohort. However, among those, the number of prescriptions was similar between cohorts. The increased number of patients receiving opioid prescriptions could reflect surgeon preference and different postoperative pain protocols rather than patient needs, as reflected by the similar number of prescriptions.
Importantly, there was no increased risk of implant related complications such as infection, loosening, periprosthetic fracture as well as revision surgery. Wilson et al. reported a decreased risk of prosthetic joint infection and revision following total knee arthroplasty with metformin use.6 Together, these findings confirm that metformin does not increase the risk of prosthesis complications, supporting its perioperative safety in diabetic patients undergoing shoulder arthroplasty.
There are multiple limitations of the present study. First, the study was retrospective and used a large de-identified database which did not enable more granular evaluation of the data. Consequently, some results may not reflect the exact circumstances of the patients studied. For example, medication adherence could not be confirmed, and we were unable to determine the exact duration and dose patients were prescribed peri-operatively. Second, the present study used data extracted with ICD-10 and CPT codes which could lead to inclusion or exclusion of patients due to miscoding. Additionally, patients undergoing anatomic and reverse total shoulder arthroplasty were analyzed as a single group, which may limit our ability to detect procedure-specific associations with perioperative metformin use. However, this study benefits from a large cohort size, rigorous propensity score matching as well as being the first study to investigate metformin in the setting of SA.
5 Conclusion
While much of the existing literature around metformin has focused on lower-extremity arthroplasty, the present study extended these investigations to shoulder arthroplasty. It contributes to a growing body of evidence suggesting metformin is safe and may have a protective role in postoperative outcomes across a variety of orthopedic procedures. These findings help surgeons in perioperative patient counseling and surgical decision making. Future research should aim to elucidate the biological mechanisms through which metformin influences surgical outcomes in shoulder arthroplasty patients. Prospective clinical studies evaluating metformin dosage, timing, and duration would aid in clarifying its role in perioperative optimization. Further investigation into long-term functional outcomes, pain scores and patient-reported metrics would also provide a more holistic understanding of metformin's potential impact on recovery and quality of life following joint replacement.
Guardian/patient's consent
We only used de-identified data provided by TrinetX. As such, no consent or IRB approval was needed.
Ethical statement
We only used de-identified data as provided by the TriNetX research database.
Author contributions
Conceptualization: THS, NL; Data curation: THS; Formal analysis: THS; Investigation: THS; Methodology: THS; Project administration: THS, NL; Resources: THS; Software: THS; Supervision: THS; Validation: THS; Visualization: THS; Roles/Writing - original draft: THS, JC, RS, GU, AA, NL; and Writing - review & editing: THS, JC, RS, GU, AA, NL.
Funding statement
This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.
References
- The impact of diabetes on perioperative patient outcomes after total hip and total knee arthroplasty in the United States. J Arthroplast. Sep 2008;23(6 Suppl 1):92-98.
- [Google Scholar]
- Global estimates of the prevalence of diabetes for 2010 and 2030. Diabetes Res Clin Pract. Jan 2010;87(1):4-14.
- [Google Scholar]
- Diabetes increases risk for readmission and infection after shoulder arthroplasty: a national readmissions study of 113,713 patients. J Orthop. Apr 2023;38:25-29.
- [Google Scholar]
- The therapeutic potential of metformin in neurodegenerative diseases. Front Endocrinol. 2018;9:400.
- [Google Scholar]
- Metformin use and associated risk of total joint replacement in patients with type 2 diabetes: a population-based matched cohort study. CMAJ (Can Med Assoc J). Dec 19 2022;194(49):E1672-e1684.
- [Google Scholar]
- Metformin use is associated with fewer complications in patients with Type-2 diabetes undergoing total knee arthroplasty: a propensity score-matched analysis. J Bone Joint Surg Am. Apr 7 2021;103(7):601-608.
- [Google Scholar]
- Outcomes after surgery in patients with diabetes who used metformin: a retrospective cohort study based on a real-world database. BMJ Open Diabetes Res Care. Nov 2020;8(2)
- [Google Scholar]
- Metformin use is associated with reduced acute kidney injury following coronary artery bypass grafting in patients with type 2 diabetes: an inverse probability of treatment weighting analysis. Pharmacotherapy. Aug 2023;43(8):778-786.
- [Google Scholar]
- Post-operative acute kidney injury and five-year risk of death, myocardial infarction, and stroke among elective cardiac surgical patients: a cohort study. Crit Care. Dec 12 2013;17(6)
- [Google Scholar]
- Postoperative acute kidney injury in adult non-cardiac surgery: joint consensus report of the Acute Disease Quality Initiative and PeriOperative Quality Initiative. Nat Rev Nephrol. Sep 2021;17(9):605-618.
- [Google Scholar]
- Association between postoperative acute kidney injury and mortality after plastic and reconstructive surgery. Sci Rep. Nov 21 2022;12(1)
- [Google Scholar]