Is tobacco use associated with risk of recurrence and mortality among people with tuberculosis? A systematic review and meta-analysis

Aishwarya Vidyasagaran, Anne Elizabeth Readshaw, Alexander Jarde, Faraz Siddiqui, Anna Marshall, Janita Akram, Jonathan E. Golub, Kamran Siddiqi, Omara Fatima Dogar*

*Corresponding author for this work

Research output: Contribution to journalArticlepeer-review


Associations between smoking and poor tuberculosis (TB) treatment outcomes are well documented. However, for important outcomes such as TB recurrence/relapse and mortality-during-treatment, as well as for associations with smokeless tobacco (ST), the evidence is not systematically summarised.
Research Question
Is tobacco use associated with risk of poor treatment outcomes among people with TB.
Study Design and Methods
MEDLINE, EMBASE and CINAHL were searched on November 22, 2021. Epidemiological studies reporting associations between tobacco use and at least one TB treatment outcome were eligible. Independent double screening, extractions and quality assessments were undertaken. Random-effects meta-analyses were conducted for the two primary review outcomes (TB recurrence/relapse and mortality-during-treatment), and heterogeneity was explored using subgroups. Other outcomes were narratively synthesised. PROSPERO registration: CRD42017060821.
Our searches identified 1,249 records, of which 28 were included in the meta-analyses. Based on 15 studies, higher risk of TB recurrence/relapse was found with ever-smoking vs. never-smoking (RR=1·78, 95%CI=1·31-2·43, I2=85%), current-smoking vs. non-smoking (RR=1·95, 95%CI=1·59-2·40, I2=72%), and past-smoking vs. never-smoking (RR=1·84, 95%CI=1·21-2·80, I2=4%); heterogeneity arose from differences in study quality, design and participant characteristics. Thirty-eight studies were identified for mortality, of which 13 reported mortality-during-treatment. Ever-smoking (RR=1·55, 95%CI=1·32-1·81, I2=0%), and current-smoking (RR=1·51, 95%CI=1·09-2·10, I2=87%) significantly increased the likelihood of mortality-during-treatment among people with TB compared to never and non-current smoking, respectively; heterogeneity was largely explained by differences in study design. Almost all studies in the meta-analyses scored ‘high’ or ‘moderate’ on quality assessments. Narrative synthesis showed that smoking was a risk factor for other unfavourable TB treatment outcomes, as previously documented. Evidence on ST was limited, but identified studies suggested an increased risk for poor outcomes with their use compared to non-use.
Smoking significantly increases the risk of TB recurrence/relapse and mortality-during-treatment among people with TB, highlighting the need to address smoking to improve TB outcomes.
Original languageEnglish
Pages (from-to)22-47
Number of pages26
Issue number1
Early online date29 Aug 2023
Publication statusPublished - 1 Jan 2024

Bibliographical note

© 2023 The Author(s).

Cite this