Early Anesthesia-Period Stress Hyperglycemia Ratio and In-Hospital Mortality After Noncardiac Surgery in Patients with HbA1c and Glucose Measurements

Background/Objectives: We examined the association between early anesthesia-period relative hyperglycemia and in-hospital mortality after noncardiac surgery. Methods: This retrospective INSPIRE cohort included admissions with pre-anesthesia HbA1c and glucose measured within 60 min after anesthesia start. Early iSHR was first glucose divided by HbA1c-derived estimated average glucose; whole-anesthesia peak iSHR was secondary. Follow-up extended from anesthesia end until death or same-admission discharge, with variable duration. Logistic spline and linear models used patient-clustered standard errors. Results: Of 117,429 source admissions, 10,406 were included, with 101 deaths. The core-adjusted OR per 0.10 increase in early iSHR was 1.134 (95% CI 1.063–1.210). Further adjustment for measurement count, renal function, malignancy, hypotension, and blood-product records yielded an OR of 1.089 (1.014–1.170) in 10,352 admissions with 97 deaths. For whole-anesthesia peak iSHR (11,374 admissions; 104 deaths), otherwise matched models without and with low-MAP proportion and blood-product records yielded ORs of 1.126 (1.063–1.193) and 1.064 (0.999–1.133; p = 0.052), respectively. A separate TWA/count model yielded an OR of 1.039 (0.973–1.109) in 11,312 admissions. Internal comparisons showed no clear discrimination advantage over absolute early glucose. Conclusions: In clinically selected patients with pre-anesthesia HbA1c and early glucose measurements, early iSHR was associated with in-hospital mortality in core and further-adjusted models, although creatinine-complete and influence analyses indicated limited robustness. The whole-anesthesia peak-iSHR estimate attenuated and showed no clear association when hypotension and blood-product records were included; the TWA/count model also showed no clear association. Internal comparison did not establish a predictive advantage over absolute early glucose; causality cannot be inferred.