Relationship between HbA1c and mortality in a Japanese population

Abstract

HbA(1)c concentrations are known to be associated with all-cause excess mortality risk in Caucasians. However, the relationship has not been clarified well in the Japanese. In addition, studies of the relationship between HbA(1)c and mortality from malignant neoplasms are scarce. HbA(1)c was measured for 3,710 people of a cohort composed of A-bomb survivors and controls. At baseline they were divided into five groups: a normal HbA(1)c group of 1,143 individuals with HbA(1)c of <5.5%, a slightly high but normal HbA(1)c group of 1,341 individuals with HbA(1)c > or =5.5% to <6.0%, a slightly high HbA(1)c group of 589 individuals with HbA(1)c > or =6.0% to <6.5%, a high HbA(1)c group of 259 individuals with HbA(1)c > or =6.5%, and a group of 378 individuals known to have type 2 diabetes. Using a Cox proportional hazards model, hazard ratios based on comparisons with the normal HbA(1)c group were obtained. During the observation period there were 754 deaths. For all-cause and cardiovascular disease mortality, a significant increase of the hazard ratio was observed for the slightly high HbA(1)c group. A similar increase in malignant neoplasm-related mortality was observed for both the high HbA(1)c group and the diabetes group. Our results suggest that individuals in the Japanese population with HbA(1)c levels of 6% or more might have increased mortality risk. The results indicate that HbA(1)c measurements should be sought even for people who have not been diagnosed with diabetes.