Importance: Predicting the risk of type 2 diabetes (T2DM) accurately allow allocation of resources to preventive strategies. Besides fasting blood triglycerides (TG) are highly predictive for T2DM, TG remnant lipoproteins (TRL) translate more accurately physiopathological events that underlie progression to T2DM such as pancreatic steatosis and inflammation.
Objective: We hypothesized TRL-related factors could improve risk prediction for T2DM.
Methods: Regression models were used to predict 4-year risk of incident T2DM, starting with clinical characteristics and metabolic syndrome traits (age, sex, parental history of diabetes, hypertension, waist circumference, HbA1c, blood TG, hsCRP and HDL-C), adding TRL-related measurements, including TRL plasma concentration, particle size, cholesterol and TG content. TRL features were derived from NMR spectroscopy. Discrimination was assessed with area under the ROC curves (AUROCs).
Results: Among 13,628 individuals (1,132 new T2DM cases during follow-up) we generated a 10-variable model with AUROC 0.890 (95%CI: 0.868-0.911). This was reproduced in a subset cohort of 4,466 individuals (353 new cases of T2DM) who had available sNMR, with AUROC 0.891 (95%CI: 0.870-0.913). TRL-related markers in general did not improve predictive capacity for T2DM, but only TRL particle size (TRLZ) increased AUROC particularly in individuals with Hba1c<5.8% at baseline. In this subgroup, AUROC increased from 0.792 (95% CI: 0.764-0.813 - model without TRLZ) to 0.844 (95% CI: 0.802-0.869 - model with TRLZ) (p-value for the diff in AUCs = 0.00013). Further analyses showed that, in prediabetic individuals at baseline, TRLZ is highly correlated with obesity, insulin resistance and inflammatory activity, and less important in individuals with Hba1c<5.8%.
Conclusions: TRL particle size improve predictive capacity for T2DM, particularly in non-prediabetic individuals.
L.F. Carvalho: None. I.M. Bensenor: None. P.A. Lotufo: None. A.C. Sposito: None.
Brazilian Ministry of Health; Brazilian Ministry of Science, Technology and Innovation; Brazilian National Council for Scientific and Technological Development