Validity of self-reports of caffeine use.

The relationship between self-reports of caffeine ingestion on two occasions and measured plasma concentrations of caffeine and its major metabolites was examined. A subject population [25 men and 25 women, age 20-45 years (mean: 28.7 yr)] that was enrolled in a benzodiazepine pharmacokinetic study underwent general medical screening on two occasions, each including detailed caffeine histories. Before beginning their scheduled study, plasma samples were obtained and evaluated by HPLC for caffeine, paraxanthine, theophylline, and theobromine. These values were compared with estimates of caffeine consumption in mg/day generated from both histories. There was no significant difference between plasma levels of caffeine, metabolites, or caffeine plus metabolites for categories corresponding to reports of low, intermediate or high caffeine use. A self-reported caffeine consumption of greater than 300 mg/day (high) did correlate, however, with a significant smoking history. The authors conclude that self-reports of caffeine ingestion do not accurately reflect acute exposure, and that if caffeine use is of importance in a given setting, reports should be confirmed by biochemical means.