Macrovascular Residual Risk Studies

COMMENT

To assess the impact of serum triglycerides levels on CHD risk, Sarwar et al. conducted two separate nested case-control comparisons in 2 different prospective, population-based cohorts, the Reykjavik1 and the European Prospective Investigation of Cancer (EPIC)-Norfolk2 studies. “To help put these new data in context,” the authors updated a previous meta-analysis of prospective studies carried out in Western populations, adding 12 more studies.

Nested case-control studies: both fasting and non-fasting triglycerides increase CHD risk

A major difference between the EPIC-Norfolk study and the Reykjavik study was that baseline triglycerides were measured in non-fasting individuals in Norfolk and in fasting individuals in Reykjavik.

Conducted in the UK, the EPIC-Norfolk study enrolled 25 668 men and women identified through general practice registers. Non-fasting venous blood samples were obtained during the screening visit. Between 1993 and 2003, 1138 incident cases of CHD (including 785 nonfatal CHD events) were recorded. For the nested case-control study, Sarwar et al selected 2276 controls free of cardiovascular disease. Controls were individually matched to cases by age, gender, and date of screening.

The Reykjavik study recruited 8888 men and 9681 women without a history of myocardial infarction between 1967 and 1991. Fasting venous blood samples were collected at baseline. A total of 2459 men and women suffered major CHD events between the beginning of follow-up and December 31, 1995. To perform the nested case-control study, 3969 controls were selected and matched to cases regarding calendar-year of recruitment, gender, and age who had survived to the end of the study period without myocardial infarction.

The use of non-fasting blood samples in one study and fasting ones in the other was reflected by a higher mean level of triglycerides in the EPIC-Norfolk study (1.90 ± 1.17 mmol/L) than in the Reykjavik study (1.03 ± 0.62 mmol/L).

Cases and controls were directly compared within each case-control study to avoid potential biases. Repeat measurements of triglycerides made in a subset of patients in both the original studies allowed for controlling analyses for a potential regression dilution bias (i.e. underestimation of impact of triglycerides levels due to intra-individual variability in measurements).

After adjustment for established cardiovascular risk factors at baseline, and for HDL cholesterol in the EPIC-Norfolk study, CHD risk was found to be increased by 57% in the EPIC-Norfolk participants that were in the upper tertile of non-fasting triglycerides compared with those in the lower tertile. An even more striking 76% increase in CHD risk was observed in the Reykjavik participants in the upper tertile as compared with the lower tertile of fasting triglycerides.

Using another mode of calculation, the authors reported that CHD risk was increased by 27% and by 19% for each 1.0-mmol/L increase in baseline fasting (Reykjavik) and non-fasting (EPIC-Norfolk ) triglycerides concentration, respectively.

A meta-analysis of 29 prospective studies confirms the increase in CHD risk associated with elevated triglycerides

In the updated analysis of 29 prospective studies, a total of 262 525 participants with 10 158 CHD cases could be analyzed. Results were in line with those of the nested case-control studies: calculation of the odds ratio for CHD after adjustment for established cardiovascular risk factors showed a 72% higher CHD risk in individuals in the upper tertile of triglycerides levels.

In contrast to other studies, there were no important differences in the strength of associations between triglycerides and CHD risk in studies in which fasting participants were included compared with studies of non-fasting participants. Furthermore, the impact of triglycerides on CHD risk appeared to be similar in men and women.

Although the analyzed studies were conducted in Western populations, the combined odds ratio yielded by the meta-analysis was very similar to the combined odds ratio previously reported in Asian and Pacific populations.4

These analyses represented the largest and most comprehensive epidemiological assessment of CHD risk associated with triglycerides concentrations ever performed in Western populations. Several limitations are inherent to the number and design of individuals studies. For example, variations in the methods of data adjustment used in each study didn’t make it possible to consistently adjust for possible confounding factors in the updated meta-analysis of 27 available prospective studies.

Despite these limitations, these studies provide consistent evidence that elevated triglycerides levels are associated with increased CHD risk, even after adjustment for established coronary risk factors, including HDL cholesterol.

Figure 1: Odds ratios for incident CHD events in individuals in the upper tertile of triglycerides concentrations compared with those in the lowest tertile in the EPIC-Norfolk study, the Reykjavik study, and in a meta-analysis of 29 prospective studies