Assessing Glycosylated Hemoglobin Thresholds for Development of Cardiovascular Disease by Racial and Ethnic Groups

Abstract

Background Diabetes is the strongest risk factor for cardiovascular disease, and although glycosylated hemoglobin (HbA1c) levels are known to vary by race, no racial and ethnic–specific diagnostic thresholds exist for diabetes in prediction of cardiovascular disease events. The purpose of this study is to determine whether HbA1c thresholds for predicting major adverse cardiovascular events (MACEs) differ among racial and ethnic groups. Methods and Results This is a retrospective cohort study of Kaiser Permanente Northern California adult members (n=309 636) with no history of cardiovascular disease who had HbA1c values and race and ethnicity data available between 2014 and 2019. Multivariable logistic regression was used to evaluate the odds of MACEs by the following racial and ethnic groups: Filipino, South Asian, East Asian, Black, White, and Hispanic. A Youden index was used to calculate HbA1c thresholds for MACE prediction by each racial and ethnic group, stratified by sex. Among studied racial and ethnic groups, South Asian race was associated with the greatest odds of MACEs (1.641 [95% CI, 1.456–1.843]; P <0.0001). HbA1c was a positive predictor for MACEs, with an odds ratio of 1.024 (95% CI, 1.022–1.025) for each 0.1% increment increase in HbA1c. HbA1c values varied between 6.0% and 7.6% in MACE prediction by race and ethnicity and sex. White individuals, South Asian individuals, East Asian women, and Black men had HbA1c thresholds for MACE prediction in the prediabetic range, between 6.0% and 6.2%. Black women, Hispanic men, and East Asian men had HbA1c thresholds of 6.2% to 6.6%, less than the typical threshold of 7.0% that is used as a treatment goal. Conclusions Findings suggest that the use of race and ethnic– and sex‐specific HbA1c thresholds may need to be considered in treatment goals and cardiovascular disease risk estimation.