Incremental Yield, Safety and Positivity Predictors of Induced Sputum for Adult Pulmonary Tuberculosis Diagnosis: A Cross Sectional Study

ABSTRACTBackground. The early diagnosis of pulmonary tuberculosis is one of the most effective intervention in tuberculosis (TB) control strategy, but it is still facing a high proportion of smear negative sputum. We aimed to assess the yield and safety of sputum induction (SI) for pulmonary tuberculosis (PT) diagnosis, and to seek its positivity predictors, in adults with negative analysis of spontaneous sputum. Methods. Adults suspected to have pulmonary tuberculosis, and who had 2 consecutive negative acid fast bacilli (AFB) and Loop-mediated isothermal amplification (TB-LAMP), were invited to undergo a unique 10% hypertonic saline inhalation. Only consenting patients who were able to undergo the inhalation procedure safely were ultimately retained. The induced sputum was analyzed through auramine AFB smear and TB-LAMP test. Results. A total of 110 patients were selected. The AFB smear and TB-LAMP were positive for 12 patients (10.9%). Limiting the sample to patients with salivary spontaneous sputum, salivary sputum + nodules, and salivary sputum + cavities improved diagnosis yield to 18.9%, 32.0% and 62.5%, respectively. Predictors for a positive IS analysis [adjusted odds ratio (95% confident interval), P-value] were nodules [15.2 (1.8, 132.2), 0.002] and cavities [32.7 (3.2, 336.3), < 0.001]. The procedure induced 0.9 – 1.8% minor adverse events. Conclusion. Our results showed a low diagnosis yield and a good safety in this study. Selection of patients based on the type of spontaneous sputum and CXR significantly improved the diagnosis yield.RESUMEIntroduction. Le diagnostic precoce de la tuberculose pulmonaire (TP) est une intervention majeure de controle de la tuberculose (TB). Cependant, il reste compromis par la proportion elevee d’analyse negative de l’expectoration. Cette etude evaluait le rendement et l’innocuite de l’expectoration induite (EI) pour le diagnostic de la TP, et recherchait les facteurs associes a sa positivite, chez des adultes presentant une analyse negative de l’expectoration spontanee. Methodes. Une recherche de tuberculose etait realisee sur 2 expectorats par recherche de bacilles acido-alcoolo resistants (BAAR) et Loop-mediated isothermal amplification (TB-LAMP). Ceux presentant une recherche negative etaient invites a produire une expectoration induite par la nebulisation de chlorure de sodium a 10%. Seuls les patients consentants de participer et capables d’effectuer l’examen etaient retenus. Les analyses comportaient la microscopie a fluorescence (coloration a l’auramine) et le TB-LAMP. Resultats. Au total 110 patients ont ete retenus. L’analyse de l’EI etait positive pour 12 (10,9%) patients. L’expectorat spontane salivaire, et son association a des nodules pulmonaires ou des lesions cavitaires ameliorait le rendement diagnostic a 18,9%, 32,0% et 62,5% respectivement. Les predicteurs de positivite [rapport de cotes ajuste (intervalle de confiance a 95%), P-valeur] etaient la presence de lesions nodulaires [15,2 (1,8; 132,2), 0,002] et cavitaires [32,7 (3,2; 336,3), < 0,001]. La procedure a provoque 0,9 a 1,8% d’effets secondaires mineurs. Conclusion. Le rendement de l’EI etait faible et l’innocuite bonne chez les patients tout venant. L’expectoration salivaire et les lesions nodulaires ou cavitaires amelioreraient significativement ce rendement.