The association between glycated hemoglobin levels and mortality in non-diabetic elderly subjects.

Abstract

BACKGROUND: Glycated hemoglobin (HgbA1c) is extensively used as a screening tool for the diagnosis of diabetes mellitus (DM). Although its prognostic value among diabetics has been extensively studied, similar data among non-diabetics is scarce. The aim of this study was to evaluate the association between HgbA1c and mortality in non-diabetic elderly individuals. METHODS: A review of a computerized database of a large health care organization was performed. Non-diabetic individuals≥65years old with HgbA1c levels<6.5% evaluated during the years 2002-2012 were included in the analysis. Subjects were stratified into quintiles based on HgbA1c levels. Hazard ratios (HR) for mortality were compared between the quintiles. RESULTS: The cohort included 12,937 non-diabetic individuals≥65years who did not develop DM during the follow-up period. The lowest mortality rate was observed in the fourth HgbA1c quintile [HgbA1c 5.9-6.1% (41-43mmol/mol)]. Compared with the mortality rate in the fourth quintile, mortality rate was increased in the lowest (HgbA1c<5.39%, 35mmol/mol, HR 1.21 CI 1.09-1.35) and highest (HgbA1c>6.11%, 43mmol/mol, HR 1.17 CI 1.04-1.32) HgbA1c quintiles. This association persisted after adjustment for age, gender and co-morbidities. CONCLUSIONS: The association between HgbA1c and all-cause mortality in non-diabetic elderly individuals has a U-shape form with the highest mortality in those with the lowest and highest HgbA1c levels.