Background Smoking is an independent risk factor for tuberculosis infection, disease and mortality. It is therefore important that clinicians can identify patients who are actively smoking in order to provide optimal care. The reliability of self-reported smoking habits is often questioned but the development of biochemical assays that measure the nicotine metabolite cotinine, allow for an objective assessment of smoking status. Here we assess the accuracy of self-reported smoking habits in patients with tuberculosis by comparing these with urinary cotinine levels.

Methods Patients were recruited from two London tuberculosis clinics. Self-reported smoking habits were obtained from an interviewer administered questionnaire. Urinary cotinine levels were measured using a SmokeScreen ^® test kit (GFC Diagnostics, Oxfordshire, UK), a semi-quantitative colorimetric assay.

Results One hundred patients attending clinic for treatment of active tuberculosis or latent tuberculosis infection completed the study (Table 1). Nineteen reported using tobacco and had either smoked or chewed tobacco the day before testing, the quantity of which was representative of a standard day’s consumption. Duration of smoking ranged from 3 to 40 years with a mean of 13.4+10.5 and a median of 10.0. Eight further patients recorded stopping smoking between a few days and 6 years prior to testing, with a mean cessation period of 2.0+2.1 years. Two of these had later started chewing tobacco.

Although nineteen patients reported either smoking or chewing tobacco, a further six patients’ urine tested positive for cotinine, i.e. 24% of tobacco users failed to report its use (Table 1). Furthermore, nine patients produced the colour change indicating heavy smoking (11–15/day) and ten patients produced a result indicating very heavy smoking (>16/day), although only two reported smoking 11–15 cigarettes the previous day and a further two must have chewed an equivalent amount of tobacco. This corresponds to 79% of heavy to very heavy tobacco users under-reporting its use.

Conclusion Discrepancies exist between objective and subjective assessments of smoking habits amongst patients with TB which may be attributable to an under-reporting of tobacco use, suggesting further avenues of research in this patient cohort.