Clinical usefulness of different lipid measures for prediction of coronary heart disease in type 2 diabetes: a report from the Swedish National Diabetes Register

Abstract

Objective: We assessed the association between different blood lipid measures and risk of fatal/nonfatal coronary heart disease (CHD).

Research design and methods: We conducted an observational study of patients with type 2 diabetes from the Swedish National Diabetes Register. Baseline LDL cholesterol, non-HDL cholesterol, ratio of non-HDL to HDL cholesterol (non-HDL:HDL), and ratio of triacylglycerol to HDL cholesterol (TG:HDL) was measured in 18,673 patients aged 30-70 years, followed for a mean of 4.8 years from 2003 to 2007.

Results: Hazard ratios (HRs) for CHD per 1-SD increment in lipid measures were 1.23 with non-HDL:HDL, 1.20 with non-HDL cholesterol, 1.17 with LDL cholesterol, and 1.15 with TG:HDL (all P < 0.001 when adjusted for clinical characteristics and nonlipid risk factors). The best global model fit was found with non-HDL:HDL. When patients within the lowest tertile of a lipid measure were compared with those with all lipid measures within the highest tertile, the adjusted HR for CHD was 0.62 with non-HDL:HDL <3.5 mmol/L, 0.65 with non-HDL cholesterol <3.3 mmol/L, and 0.70 with LDL cholesterol <2.5 mmol/L (all P < 0.001). The lowest tertile of LDL and non-HDL cholesterol corresponded with treatment targets according to U.S. and European guidelines. HRs for CHD were 0.52, 0.62, and 0.66 with the lowest deciles of non-HDL:HDL, non-HDL cholesterol, and LDL cholesterol ≤1.8 mmol/L (all P < 0.001). Mean TG:HDL was considerably lower in patients within the lowest tertile of non-HDL:HDL, 0.82 ± 0.47, than in those within the lowest tertile of LDL cholesterol (<2.5 mmol/L), 1.49 ± 1.03.

Conclusions: Non-HDL:HDL had a stronger effect on CHD risk than LDL cholesterol, and low TG:HDL values were more often seen within the lowest non-HDL:HDL tertile than within the lowest LDL cholesterol tertile. LDL cholesterol was not the best predictor of CHD risk in type 2 diabetes.

Figure 1

Five-year rates of CHD by lipid measures per one-unit increase in a Cox proportional hazards model of 18,673 patients. In each figure, the spline represents event rates (solid line) with 95% CIs (dashed lines) of fatal/nonfatal CHD as a cubic function of the lipid measure.