The purpose of this study was to determine whether caffeine consumption confounds the relationship among adrenergic tone, as measured by urinary norepinephrine (NE), blood pressure (BP) and obstructive sleep apnoea (OSA). Data were analysed using correlation and regression analysis, analysis of covariance and t-tests. Subjects included normotensives and hypertensives with and without OSA: 38 men, 23 women, aged 30-60 y; 100-150% of ideal body weight; without other major illness. Patients were studied using polysomnography, caffeine consumption was assessed, 24-h urinary NE levels were examined and ambulatory BP was recorded. Patients with OSA (N=27) reported significantly greater caffeine consumption than those without OSA (N=34) (295 vs. 103 mg, P=0.010), but caffeine was not significantly correlated with their ambulatory BP. In contrast, NE excretion correlated with caffeine consumption (r=0.24, P=0.041), apnoea severity (r=0.65, P < 0.001) and BP (r=0.34, P < 0.005). Significant OSA-NE and BP-NE relationships remained even after controlling for caffeine consumption. Patients with OSA consumed nearly three times the amount of caffeine as patients without OSA. While caffeine partially explains the increased adrenergic tone in patients with OSA and the relationship between BP and NE, it does not appear to contribute significantly to the relationship between OSA and elevated BP.