Non-Linear U-Shaped Trend Associations between HbA1c and All-Cause and Cardiovascular Mortality

Several well-designed prospective studies have assessed the association between glycated hemoglobin (HbA1c) and mortality risk in both the general population and diabetic patients. However, the results have been inconsistent.

We conducted a prospective study using data from NHANES 2003 - 2018, which included 38,072 adults. Cox proportional hazards regression and restricted cubic spline models were used to evaluate the linear and non-linear associations between HbA1c and both all-cause mortality and cause-specific mortality. To better detect non-linear relationships, we converted HbA1c from a continuous to a categorical variable. Kaplan-Meier survival cur-ves and subgroup analyses were also performed.

During an average follow-up period of 99.7 ± 53.6 months, 4,723 participants (12.4%) died. We observed a significant non-linear U-shaped association between baseline HbA1c levels and mortality. Compared with a moderate HbA1c level of 5.5%, the risk of all-cause and cause-specific mortality decreased for HbA1c values below 5.5%. However, mortality risk significantly increased for HbA1c levels above 5.5%. This indicates that individuals with either low or high HbA1c levels are at a higher risk of all-cause mortality. Additionally, we identified an L-shaped trend in cancer mortality.

We demonstrated a U-shaped association between baseline HbA1c levels and all-cause mortality in the general population. Individuals with both low and high HbA1c levels are at a greater risk of all-cause mortality. Future clinical trials on diabetes management should consider the consequences of lower HbA1c levels, especially across different populations.