Dietary trials with subjects on a freely selected diet may be affected by unwanted behavioural changes. Few studies, if any, have examined changes in coffee consumption and possible concomitant changes in diet and health-related habits. The aim of the present study was to examine whether induced changes in coffee consumption lead to changes in food habits and leisure-time physical activity. Healthy, non-smoking coffee-drinkers (n 214) were asked to change their coffee habits in a controlled clinical trial on the metabolic effects of coffee. The participants were asked to maintain their usual dietary habits. Self-perceived changes in diet and physical activity during the 6-week intervention period were assessed at the end. In the analyses, the participants were rearranged into groups reflecting the difference in coffee intake during the trial as compared with habitual intake. Associations with changes in food intake or physical activity were analysed by Spearman rank correlation. Changes in intake of 'chocolate, sweets' (r 0.179, P<0.05), 'cakes, sweet biscuits, pastry' (r 0.306, P<0.001), and 'jam' r 0.198, P<0.05) showed positive associations with change in coffee intake during the trial. Negative associations were found for 'dishes with fish' (r -0.204, P<0.01) and many of the drinks as well as with physical activity (r -0.164, P<0.05). Induced changes in coffee intake seem to alter ad libitum intake of several foods. The recognized associations between health behaviours may have physiological explanations.