An alternative approach may be to consider atherogenic dyslipidaemia as a continuous variable using the ratio log triglycerides)/HDL cholesterol or log triglycerides/HDL cholesterol), based on fasting measures.
Diagnosis of diabetes was based on fasting glucose test only.
However, IRS was defined based on fasting insulin.
The authors did not find any effect on fasting triglycerides.
Following exclusions, data on fasting triglycerides and high density lipoprotein (HDL) cholesterol were available for 914 and 901 participants respectively.
Sample size calculations were based on a minimum predicted 15% change in fasting triglyceride and total cholesterol levels between the intervention groups, with an expected standard deviation of 15%.
Fasting triglyceride levels were categorized as "normal", "borderline", "high", and "extremely high" based on the National Cholesterol Education Project (NCEP) thresholds [ 23].
Composite Z scores, based on baseline and follow-up measurements, were constructed as a mean of the available standardised selected risk factors (fasting triglyceride, insulin sensitivity (HOMA), sum of four skinfolds, systolic blood pressure, total cholesterol/HDL (high-density lipoprotein) ratio and inverse aerobic fitness).
The two groups had been matched for age, weight, fasting triglyceride levels, insulin concentrations, total cholesterol and other factors.
