Risk factors for postoperative transfusion in diabetic patients following total shoulder arthroplasty.

Background: Diabetes has been reported as a risk factor for postoperative transfusion following total shoulder arthroplasty (TSA). However, the risk factors specific to diabetic patients that increase their likelihood of postoperative blood transfusion remains understudied. The purpose of the study was to investigate the risk factors that are associated with 30-day postoperative transfusion among diabetic patients who undergo TSA.

Methods: The American College of Surgeons National Surgical Quality Improvement Program database was queried for all patients who underwent TSA between 2015 and 2020. Both patients with and without diabetes were divided into cohorts based on 30-day postoperative transfusion requirement. Bivariate logistic regression was used to compare patient demographics and comorbidities. Multivariate logistic regression, adjusted for all significant patient demographics and comorbidities, was used to identify the characteristics independently associated with postoperative transfusion.

Results: A total of 4376 diabetic patients remained after exclusion criteria, with 4264 (97.4%) patients who did not require postoperative transfusion and 112 (2.6%) patients who did require postoperative transfusion. On multivariate analysis, female gender (odds ratio [OR] 2.43, 95% confidence interval [CI] 1.52-3.89;  < .001), American Society of Anesthesiologists ≥3 (OR 2.46, 95% CI 1.10-5.48;  = .028), bleeding disorder (OR 2.94, 95% CI 1.50-5.76;  = .002), transfusion prior to surgery (OR 12.19, 95% CI 4.25-35.00;  < .001), preoperative anemia (OR 8.76, 95% CI 5.47-14.03;  < .001), and operative duration ≥129 minutes (OR 4.05, 95% CI 2.58-6.36;  < .001) were found to be independent risk factors for postoperative transfusion among diabetic patients. Our nondiabetic cohort included 19,289 patients, with 341 (1.8%) requiring postoperative transfusion. On Multivariate analysis, we found similar risk factors for transfusion to our diabetic population, as well as age ≥75 (OR 1.80, 95% CI 1.37-2.35;  < .001) and dependent functional status (OR 2.16, 95% CI 1.40-3.32;  < .001) to be independent risk factors for postoperative transfusion among nondiabetic patients.

Conclusion: Female gender, American Society of Anesthesiologists ≥3, bleeding disorder, transfusion prior to surgery, preoperative anemia, and operative duration ≥129 minutes were independently associated with postoperative transfusion following TSA in diabetic patients. These findings encourage physicians to carefully assess patients with diabetes preoperatively to minimize adverse outcomes.