To investigate the relations between caffeine-derived metabolites (methylxanthines) and plasma lipids by use of population-based data from 2 European countries.
Families were randomly selected from the general population of northern Belgium (FLEMENGHO), from August 12, 1985, until November 22, 1990, and 3 Swiss cities (SKIPOGH), from November 25, 2009, through April 4, 2013. We measured plasma concentrations (FLEMENGHO, SKIPOGH) and 24-hour urinary excretions (SKIPOGH) of 4 methylxanthines-caffeine, paraxanthine, theobromine, and theophylline-using ultra-high-performance liquid chromatography-tandem mass spectrometry. We used enzymatic methods to estimate total cholesterol, high-density lipoprotein cholesterol, and triglyceride levels and the Friedewald equation for low-density lipoprotein cholesterol levels in plasma. We applied sex-specific mixed models to investigate associations between methylxanthines and plasma lipids, adjusting for major confounders.
In both FLEMENGHO (N=1987; 1055 [53%] female participants) and SKIPOGH (N=990; 523 [53%] female participants), total cholesterol, low-density lipoprotein cholesterol, and triglyceride levels increased across quartiles of plasma caffeine, paraxanthine, and theophylline (total cholesterol levels by caffeine quartiles in FLEMENGHO, male participants: 5.01±0.06 mmol/L, 5.05±0.06 mmol/L, 5.27±0.06 mmol/L, 5.62±0.06 mmol/L; female participants: 5.24±0.06 mmol/L, 5.15±0.05 mmol/L, 5.25±0.05 mmol/L, 5.42±0.05 mmol/L). Similar results were observed using urinary methylxanthines in SKIPOGH (total cholesterol levels by caffeine quartiles, male participants: 4.54±0.08 mmol/L, 4.94±0.08 mmol/L, 4.87±0.08 mmol/L, 5.27±0.09 mmol/L; female participants: 5.12±0.07 mmol/L, 5.21±0.07 mmol/L, 5.28±0.05 mmol/L, 5.28±0.07 mmol/L). Furthermore, urinary caffeine and theophylline were positively associated with high-density lipoprotein cholesterol in SKIPOGH male participants.
Plasma and urinary caffeine, paraxanthine, and theophylline were positively associated with plasma lipids, whereas the associations involving theobromine were less clear. We postulate that the positive association between caffeine intake and plasma lipids may be related to the sympathomimetic function of methylxanthines, mitigating the overall health-beneficial effect of caffeine intake.