Lipid and inflammatory cardiovascular risk worsens over 3 years in youth with type 2 diabetes: the TODAY clinical trial

Abstract

Objective: Type 2 diabetes increases cardiovascular risk. We examined lipid profiles and inflammatory markers in 699 youth with recent-onset type 2 diabetes in the TODAY clinical trial and compared changes across treatment groups: metformin alone (M), metformin plus rosiglitazone (M+R), and metformin plus intensive lifestyle program (M+L).

Research design and methods: Multiethnic youth with type 2 diabetes received M, M+R, or M+L. Statin drugs were begun for LDL cholesterol (LDL) ≥ 130 mg/dL or triglycerides ≥ 300 mg/dL. Lipids, apolipoprotein B (apoB), LDL particle size, high-sensitivity c-reactive protein (hsCRP), homocysteine, plasminogen activator inhibitor-1 (PAI-1), and HbA1c were measured over 36 months or until loss of glycemic control.

Results: LDL, apoB, triglycerides, and non-HDL cholesterol (HDL) rose over 12 months and then stabilized over the next 24 months. Participants with LDL ≥ 130 mg/dL or using LDL-lowering therapy increased from 4.5 to 10.7% over 36 months, while 55.9% remained at LDL goal (<100 mg/dL) over that time. Treatment group did not impact LDL, apoB, or non-HDL. Small dense LDL (particle size, ≤ 0.263 relative flotation rate) was most common in M. Triglycerides were lower in M+L than M, and M+L attenuated the negative effect of hyperglycemia on triglycerides and HDL in females. hsCRP, PAI-1, and homocysteine increased over time. However, hsCRP was lower in M+R compared with M or M+L.

Conclusions: Dyslipidemia and chronic inflammation were common in youth with type 2 diabetes and worsened over time. Diabetes treatment, despite some treatment group differences in lipid and inflammatory marker change over time, is generally inadequate to control this worsening risk.

Trial registration: ClinicalTrials.gov NCT00081328.

Figures

Figure 1

Means (mg/dL) by treatment group and annual visit from baseline to 36 months for treatment groups M, M+R, and M+L. Vertical axes are scaled to show differences. Tests were performed using the log transform to normalize the distributions of triglycerides and hsCRP, but means were calculated for the original scale.

Figure 2

Regression of HbA1c (%) on lipid outcomes (mg/dL) for treatment groups M, M+R, and M+L. Data were included at 12-, 24-, and 36-month visits prior to treatment failure. The analysis tested treatment group differences adjusted for BMI. Slopes indicate change in lipid value per unit increase in HbA1c. Tests were performed using the log transform to normalize the distribution of triglycerides, but estimates of slope were calculated for the original scale.