Total physical activity might not be a good measure in the relationship with HDL cholesterol and triglycerides in a...

Evidence suggests that physical activity (PA) has a beneficial effect on high-density lipoprotein cholesterol (HDL) and triglycerides. However, observational studies show contrasting results for this association between different ethnic groups. It is unclear whether this is due to differences in the PA composition. The aim of this study was to assess the relationship of the total PA, along with its intensity and duration, with HDL and triglycerides in a multi-ethnic population.

The study population was sampled from the SUNSET study and included: 502 European- Dutch, 338 Hindustani-Surinamese, and 596 African-Surinamese participants living in Amsterdam, the Netherlands. We assessed PA with the SQUASH questionnaire. We calculated age-sex-adjusted betas, geometric mean ratios (GMRs), and prevalence ratios (PRs) to assess the relationship of PA with HDL and triglycerides.

In the adjusted models, the highest total PA tertile compared to the lowest tertile was beneficially associated with HDL (beta: 0.08, 95% CI: 0.00, 0.16 and PR low HDL 0.59, 95% CI: 0.39, 0.88) and triglycerides (GMR: 0.93, 95% CI: 0.83, 1.03 and PR: 0.56, 95% CI: 0.29, 1.08) for the African-Surinamese. No statistically significant associations appeared for total PA among the European-Dutch and Hindustani-Surinamese. The adjusted models with the intensity score and HDL showed beneficial associations for the European-Dutch (beta: 0.06, 95% CI: 0.03, 0.10) and African-Surinamese (beta: 0.06, 0.02, 0.10), for log triglycerides for the European-Dutch (beta: -0.08, 95% CI: -0.12, 0.03), Hindustani-Surinamese (beta: -0.06, 95% CI: -0.16, 0.03), and African-Surinamese (beta: -0.04, 95% CI: -0.10, 0.01). Excepting HDL in African-Surinamese, the duration score was unrelated to HDL and triglycerides in any group.

Activity intensity related beneficially to blood lipids in almost every ethnic group. The activity duration was unrelated to blood lipids, while the total PA 'summary score' was associated only with blood lipids for African-Surinamese. The difference in total PA composition is the most probable explanation for ethnic differences in the total PA association with blood lipids. Multi-ethnic observational studies should include not only a measure of the total PA, but other measures of PA as well, particularly the intensity of activity.