Hemoglobin A1c modifies the association between triglyceride and time in hypoglycemia determined by flash glucose monitoring in adults with type 1 diabetes: implications for individualized therapy and decision-making.

Background: We aimed to investigate the associations of flash glucose monitoring (FGM)-derived metrics with lipid profiles and identify potential modifiers of these associations among adults with type 1 diabetes (T1D).

Methods: A cross-sectional study was conducted among 108 Chinese adults with T1D who used FGM for 14 consecutive days. The relationship between FGM-derived metrics and lipid variables and potential modifiers were identified using interaction and subgroup analysis.

Results: Serum triglyceride level inversely correlated with time below range (glucose <3.9 mmol/L) and time in range (glucose 3.9-10.0 mmol/L) and positively correlated with time above range (glucose >10.0 mmol/L) (Spearman's r=-0.34, -0.25, 0.34, respectively, all P<0.01). Additionally, triglyceride levels had positive correlation with absolute measures of glycemic variability (GV) but not with the coefficient of variation for glucose (Spearman's r=0.12, P>0.05), a relative measure. Multivariate linear regression analysis adjusting for potential confounders including gender, age, disease duration, body mass index (BMI), daily insulin dose, fasting C-peptide, and dyslipidemia medication use showed that higher triglyceride level independently predicted decrease in time below range and time in range and increase in time above range (all P<0.01). Furthermore, interaction analysis found that the interaction between HbA1c and triglyceride was significant in the time below range (P for interaction =0.034). The association between triglyceride and time below range differed substantially after stratification by HbA1c, which was significant in those with HbA1c <7.0% whereas inconsequential among those with HbA1c ≥7.0%. In those with HbA1c <7.0% (n=44), the area under receiver operating characteristic curve of triglyceride predicting achievement of targets of time below range (<4%) was 0.856 (95% confidence interval 0.688-1.000, P=0.042) with an optimal cutoff value of 0.50 mmol/L (sensitivity 100%, specificity 66.7%, positive predictive value 94.4%).

Conclusions: In adults with T1D, HbA1c may be a potential modifier of the association between triglyceride and time below range, suggesting it might be necessary for those with HbA1c <7.0% accompanied by lower triglyceride levels to set a less intensive glycemic target to minimize risk of hypoglycemia.