Hemoglobin A1c monitoring practices before lower extremity bypass in patients with diabetes vary broadly and do not predict outcomes

Abstract

ObjectiveHemoglobin A1c (HbA1c) is used as a marker of glycemic control, but the role of HbA1c before lower extremity bypass (LEB) in patients with diabetes remains unclear. We sought to characterize patients with diabetes undergoing LEB with and without HbA1c monitoring and to determine if HbA1c monitoring practices correlate with better outcomes. MethodsThe Vascular Quality Initiative was queried for all LEB in patients with diabetes (2010-2020). Patients with diabetes were characterized based on therapy: diet-controlled, noninsulin medication use, or insulin use. Glycemic control was characterized by preoperative HbA1c within 6 months of surgery: unknown control (no HbA1c), well-controlled (HbA1c <7%), poorly-controlled (HbA1c 7%-10%), and uncontrolled (HbA1c >10%). Centers with >5 LEB/y were stratified into terciles according to rate of HbA1c monitoring. The unadjusted associations between glycemic control and in-hospital major adverse limb events, major adverse cardiac events, and mortality were assessed with univariate methods. The independent association of center-level HbA1c monitoring with 5-year survival and 3-year amputation-free survival (AFS) was determined with Kaplan-Meier analyses and Cox regression modeling, adjusted for differences in patient characteristics and center volume. ResultsOf 16,092 patients with diabetes undergoing LEB, 4055 (25%) did not have a documented HbA1c. Insulin use was less common in no A1c (48%) and well-controlled diabetes (39%) compared with poorly controlled (67%) and uncontrolled diabetes (78%) (P < .01). In univariate analyses, glycemic control was not associated with differences for in-hospital major adverse limb events, major adverse cardiac events, or mortality. Of 162 centers, HbA1c monitoring practices varied widely (range: 12.5%-100% of LEB). The 3-year AFS and 5-year survival were worse in the highest monitoring tercile vs the lowest (73.6% vs 77.3%, P < .01, 72.1% vs 77.5%, P < .01, respectively). On multivariable analyses, centers in the highest tercile of monitoring had the greatest hazard of AFS (hazard ratio: 1.21, 95% confidence interval: 1.1-1.3, P < .001) and overall mortality (hazard ratio: 1.19, 95% confidence interval: 1.1-1.3, P < 0.001), compared with the centers in the lowest tercile of monitoring. ConclusionsPatients with diabetes and no preoperative HbA1c monitoring do not have worse LEB outcomes compared with those with HbA1c monitoring. Preoperative HbA1c monitoring varies widely, suggesting broad differences in practice and documentation. Centers with the highest rates of monitoring demonstrated inferior outcomes, likely due to other confounding unmeasured variables. These findings indicate that HbA1c monitoring before LEB, unto itself, should not be used as a measure of surgical quality.