Predictors of cessation in smokers suspected of TB: Secondary analysis of data from a cluster randomized controlled trial

Background: Smoking cessation services are rarely found within health services in low income countries. Given the interactions between Tuberculosis (TB) and tobacco, including cessation support within TB programs offers a promising cost-effective solution. We conducted secondary analysis of data from a cluster randomized controlled trial of smoking cessation in health centers in Pakistan to identify predictors of continuous and short-term abstinence in smokers suspected of TB using cigarettes or hookah. Methods: Predictor variables of those continuously abstinent at 5 and 25 weeks post quit-date (continuous abstinence) and those abstinent only at 5 weeks (short-term abstinence) were compared with those who continued smoking and with each other. Self-reported abstinence at both time points was confirmed biochemically. Results: Data obtained from 1955 trial participants were analyzed. The factors that predicted continued smoking when compared to continuous abstinence were: being older RR 0.97 (0.95 to 0.98), smoking higher quantities of tobacco RR 0.975 (0.97 to 0.98) and sharing a workplace with other smokers RR 0.88 (0.77 to 0.99). Those with a confirmed TB diagnosis were more likely to remain continuously abstinent than those without RR 1.27 (1.10-1.47). Conclusions: Those diagnosed with TB are more likely to be abstinent than those diagnosed with other respiratory conditions. Beyond this, predictors of continued smoking in Pakistan are similar to those in high income contexts. Taking advantage of the 'teachable moment' that a TB diagnosis provides is an efficient means for resource-poor TB programs in low income settings to increase tobacco cessation and improve health outcomes.