Patients who have poorly controlled diabetes mellitus are at increased risk of periprosthetic joint infection (PJI). Nevertheless, an optimal preoperative hemoglobin A1c (HbA1c) threshold has not been established. This study sought to identify preoperative HbA1c thresholds that were predictive of adverse outcomes for total hip (THA) and total knee arthroplasty (TKA) patients. A healthcare database was used to identify primary THAs or TKAs performed from 2016 to 2021 with a preoperative HbA1c value within 28 days of surgery. The primary outcome was PJI within 90 days postoperatively. Secondary outcomes included aggregate medical and surgical complications. Restricted cubic splines (RCSs) were generated using logistic regression to quantify the impact of HbA1c as a continuous variable on the risk of PJI. Between HbA1c values of 5% and 12%, relevant sensitivity and specificity measurements were calculated at intervals of 0.5%. A Youden's J statistic identified clinically relevant preoperative HbA1c thresholds. In total, 17,481 elective arthroplasty patients who had a preoperative HbA1c were identified. The mean preoperative HbA1c was 6.5%. For TKA, a PJI threshold of 9.7% was identified (sensitivity: 19.4%, specificity: 99.1%), while for THA, a PJI threshold of 7.8% was identified (sensitivity: 22.7%, specificity: 89.9%). The threshold for aggregate medical complications was 6.8% for TKA (sensitivity: 53.7%, specificity: 59.1%) and 6.5% for THA (sensitivity: 45.5%, specificity: 66.5%). No association was observed between HbA1c and aggregate surgical complications for either THA or TKA. This study identified PJI and medical complication HbA1c thresholds above which patients were at a significantly increased risk of early postoperative complications. While our findings suggest that HbA1c has limited predictive utility for postoperative complications, it remains an accessible biomarker that can aid in preoperative risk stratification. Future studies should explore other promising or complementary biomarkers that may be more effective for preoperative risk stratification.
Read full abstract