Comparative prognostic performance of definitions of prediabetes: a prospective cohort analysis of the Atherosclerosis Risk in Communities (ARIC) study

Abstract

Background There is a lack of consensus across international organizations regarding definitions of prediabetes. Associations with complications can inform the comparative value of different prediabetes definitions. Methods We conducted a prospective cohort study of 10,844 Atherosclerosis Risk in Communities (ARIC) study participants without diagnosed diabetes who attended visit 2 (1990–92) and 7,194 who attended visit 4 (1996–98). Fasting glucose and HbA1c were measured at visit 2 and fasting glucose and 2-hour glucose were measured at visit 4. We compared prediabetes definitions based on fasting glucose (American Diabetes Association [ADA] 5.6–6.9 mmol/L and World Health Organization [WHO] 6.1–6.9 mmol/L), HbA1c (ADA 39–46 mmol/mol and International Expert Committee [IEC] 42–46 mmol/mol), and 2-hour glucose (ADA/WHO 7.8–11.0 mmol/L). Findings ADA fasting glucose-defined prediabetes (prevalence 37.9%) was the most sensitive for major clinical outcomes, while ADA and IEC HbA1c and WHO fasting glucose-based definitions (prevalence 18.7%, 9.0%, 11.2%, respectively) were more specific. After demographic adjustment, HbA1c-based definitions of prediabetes had higher hazard ratios and demonstrated better risk discrimination for chronic kidney disease, cardiovascular disease, peripheral arterial disease, and all-cause mortality compared to fasting glucose (modestly larger C-statistics, all p<0.05). For example, the C-statistic for incident chronic kidney disease was 0.636 for ADA fasting glucose categories and 0.640 for ADA HbA1c clinical categories (difference −0.005, 95%CI −0.008, −0.001). Additionally, ADA HbA1c-defined prediabetes also demonstrated significant overall improvement in the net reclassification index for cardiovascular outcomes and death compared to glucose-based definitions. Comparing ADA and WHO fasting glucose and ADA/WHO 2-hour did not reveal statistically significant differences in risk discrimination for chronic kidney disease, cardiovascular, or mortality outcomes. Interpretation Our results suggest that HbA1c-defined prediabetes definitions were more specific and provided modest improvements in risk discrimination for clinical complications. ADA fasting glucose was a more sensitive definition overall.