What is the association between insulin dependence and the incidence of postoperative complications, perioperative glucose control, and five-year reoperation rates after total joint replacement?

Abstract

Diabetes mellitus (DM) is associated with adverse outcomes following total hip arthroplasty (THA) and total knee arthroplasty (TKA), yet optimal perioperative risk stratification remains unclear. This study evaluated the association of insulin dependence with: 1) 90-day complications; 2) perioperative glucose control; and 3) mid-term reoperations in diabetic patients undergoing primary THA and TKA. We performed a retrospective, single-centre study of patients with DM undergoing primary THA or TKA between 2015 and 2022. Patients with preoperative glycated haemoglobin (HbA1c) ≥ 6.5 were categorized as insulin-dependent DM (IDDM) or insulin-independent DM (NIDDM). Demographic and clinical variables including preoperative haemoglobin A1C and oral DM medication usage were recorded. The primary outcome was 90-day complications. Secondary outcomes included mean and maximum perioperative glucose levels and reoperations. Multivariable logistic and linear regression analyses were used to evaluate association between insulin dependence and outcomes while adjusting for relevant confounders. A total of 830 patients were included: 351 (42.3%) with IDDM and 479 (57.7%) with NIDDM. Overall, 90-day complications were more common in the IDDM cohort (35.9% vs 16.3%, p < 0.01). Insulin dependence was associated with increased odds of 90-day complications (odds ratio 2.81, 95% CI 1.98 to 3.98). Male sex, metformin use, and glucagon-like peptide-1 receptor agonist use were associated with lower complication risk. Preoperative HbA1c was not independently associated with complications. Insulin dependence and preoperative haemoglobin A1C were associated with worse perioperative glucose control. Reoperation rates were similar between groups. Insulin dependence is an independent risk factor for 90-day complications and poorer perioperative glucose control, but not reoperation, following primary THA and TKA. These findings suggest insulin dependence is a clinically meaningful risk marker beyond glycaemic indices. Further investigation into optimization strategies to mitigate this perioperative risk is warranted.