How accurate is the smoking history in newly diagnosed diabetic patients?

A smoking history was obtained from 94 consecutive newly diagnosed diabetic patients referred to an adult diabetic clinic. The smoking load was measured using urinary cotinine/creatinine ratios (COT/Cr). Fifty-six patients (60%) claimed to be non-smokers, but COT/Cr suggested active smoking in five of these. The patients who admitted to smoking were given standardised anti-smoking advice. At 3 months, 32 smoking patients were reviewed and 21 (66%) claimed to have reduced or stopped smoking. However, the median COT/Cr in the 32 patients showed no significant reduction (11.15 vs. 9.30 micrograms/mg). Urinary COT/Cr indicated that 6 patients had stopped smoking (median COT/Cr 6.98 fell to 0.97 micrograms/mg), but several patients had a marked rise in COT/Cr, demonstrating that their smoking habit had increased considerably. Therefore the smoking history obtained from new diabetic patients can be very misleading. An objective measure of smoking habits in the initial assessment and follow-up of diabetes may be worthwhile. Anti-smoking counselling at diagnosis of diabetes may persuade some smokers to stop.