Inverse correlation of free triiodothyronine with glycated albumin and the glycated albumin/glycated hemoglobin ratio in hemodialysis patients: a cross-sectional study

Abstract

BackgroundThat the prevalence of low triiodothyronine (T3) syndrome is high among hemodialysis (HD) patients has been previously established. Herein, we examined the association of glycated albumin (GA) and the GA to glycated hemoglobin (HbA1c) ratio (GA/HbA1c) with free triiodothyronine (FT3) in HD patients.MethodsWe conducted a cross-sectional study on 134 patients (68 patients with diabetes mellitus [DM group] and 66 patients without diabetes mellitus [non-DM group]) who received maintenance HD at our dialysis clinic between 2014 and 2018. Univariate linear regression analyses of GA, GA/HbA1c, or HbA1c with several clinical variables were primarily conducted. Multiple regression analyses with GA (or GA/HbA1c) as the objective variable were conducted with explanatory variable FT3 adjusted for age, sex, Hb, Alb, and average plasma glucose (Av-PG) (or HbA1c).ResultsIn the DM and non-DM groups, GA tended to be inversely correlated with FT3, although significantly so only in the non-DM group. GA/HbA1c also showed a strong significant inverse correlation with FT3 in the DM group and the non-DM group. FT3 and GA/HbA1c were also significantly correlated with the Geriatric Nutritional Risk Index in the DM group and non-DM group. In the multivariate analysis, which was adjusted for age, sex, Hb, Alb, and HbA1c, FT3 was a significant and independent factor associated with GA in the DM group (β = − 0.334, p < 0.001) and in the non-DM group (β = − 0.412, p < 0.001). The regression equations obtained by stepwise multiple regression analyses using all of these variables as independent variables were GA = 3.3HbA1c − 4.4FT3 + 1.9sex + 8.8 for the DM group and GA = − 2.4FT3 + 0.04Age − 0.5Hb + 25.2 for the non-DM group. These contribution rates (i.e., coefficient of determination) were R2 = 0.708 in the DM group and R2 = 0.347 in the non-DM group, In the DM group, the estimation formulas, based on the regression equation [GA (men) = 3.3HbA1c − 4.4FT3 + 10.7 and GA (women) = 3.3HbA1c − 4.4FT3 + 8.8], showed very high contribution rates (i.e., coefficient of determination R2 = 0.674 for men and 0.761 for women) for the GA measured values.ConclusionsGA and GA/HbA1c have a close relationship with FT3 in HD patients. The estimation formulas of GA could be obtained. In particular, the estimation formulas in the DM group are believed to be useful in considering HbA1c and FT3 simultaneously when evaluating GA.