Impact of Poorly Controlled Diabetes and Glycosylated Hemoglobin Values in Geriatric Hip Fracture Mortality Risk Assessment

Introduction The presence of poorly-controlled diabetes in the setting of geriatric hip fractures has been shown to increase all-cause mortality and worsen outcomes. This study aimed to assess whether the addition of a patient's glycated hemoglobin (A1c) value to a validated geriatric inpatient risk tool improves the predictive capacity of the risk tool. Methods A cohort of 2430 patients >55 years old treated for low-energy mechanism hip fractures between October 2014 to November 2021 were reviewed for demographics (including diabetes diagnoses and their respective hemoglobin A1c values at the time of admission), injury details, hospital quality measures, and mortality. As past work demonstrated a hemoglobin A1c value above 8% to be the tipping point for worse outcomes, the baseline Score for Trauma Triage in the Geriatric and Middle-Aged (STTGMA) tool for inpatient mortality in hip fractures (STTGMAHIP_FX_SCORE - Score for Trauma Triage in the Geriatric and Middle-Aged Hip Fracture Score) was modified to include a patient's hemoglobin A1c using an 8% cutoff (STTGMAHIP_8%A1c - Score for Trauma Triage in the Geriatric and Middle-Aged Hip 8% Hemoglobin A1c Cutoff Score). The new model's predictive ability (as measured by the area under the receiver operating curves (AUROCs)) for inpatient mortality was compared to the baseline tool using DeLong's test. Risk quartiles were generated for the new tool, and comparative analyses were conducted on hospital quality measures and outcomes. Results Five hundred and sixty-five patients (23%) were noted to have diabetes mellitus, and 76 patients had an A1c above 8%. Patients with a hemoglobin A1c above 8% had a higher rate of inpatient complications and mortality through one year. The STTGMAHIP_8%A1c score significantly improved the predictive capacity for inpatient mortality compared to STTGMAHIP_FX_SCORE (0.786 vs. 0.672, p=0.0456). Upon analysis of the risk quartiles, the highest risk cohort was found to have a longer length of stay (p<0.001), with higher rates of inpatient (p<0.001) and 30-day mortality (p<0.001) and need for admission to the intensive care unit (p<0.001) as compared to the minimal risk cohort. Patients in the lowest risk quartile were most likely to be discharged home (p<0.001). Conclusion Patients who present with a hemoglobin A1c above 8% experienced significantly worse outcomes than those below 8%. The inclusion of a patient's hemoglobin A1c as a cutoff score improves the STTGMAHIP_FX_SCORE tool to predict mortality and risk stratify patient outcomes. While diabetes presents another medical challenge to manage, providers may utilize this new variable to better highlight at-risk diabetic patients.