The effect of health counselling on the smoking habits of 60 diabetic patients (aged less than 40 years) was assessed. Measurement of breath carbon monoxide (CO) and urinary cotinine, a metabolite of nicotine, were used as objective markers of smoking. All patients wished to cease smoking and the impact of health counselling was reviewed in a 'Stop Smoking' clinic. In addition to routine advice on the health hazards of smoking, half the patients and their families also received further counselling during a home visit by a health visitor. After 6 months many of the 60 patients claimed to have reduced their cigarette consumption. However, the urinary cotinine concentrations did not confirm this. Only one patient actually stopped smoking and he had sustained a myocardial infarction during the study. There was a small but significant reduction of breath CO in the patients seen at home by the health visitor but the urinary cotinine concentrations were unchanged. This suggests that these patients abstained from smoking for only a few hours before attending the 'Stop Smoking' clinic.