Dynamics Of Sputum Conversion During Effective Tuberculosis Treatment: A Systematic Review And Meta-Analysis

Background Two weeks' isolation is widely recommended for people commencing treatment for pulmonary tuberculosis (TB). The evidence that this corresponds to clearance of potentially infectious tuberculous mycobacteria in sputum is not well established. This World Health Organization-commissioned review investigated sputum sterilisation dynamics during TB treatment.Methods and findings For the main analysis, 2 systematic literature searches of OvidSP MEDLINE, Embase, and Global Health, and EBSCO CINAHL Plus were conducted to identify studies with data on TB infectiousness (all studies to search date, 1 December 2017) and all randomised controlled trials (RCTs) for drug-susceptible TB (from 1 January 1990 to search date, 20 February 2018). Included articles reported on patients receiving effective treatment for culture-confirmed drug-susceptible pulmonary TB. The outcome of interest was sputum bacteriological conversion: the proportion of patients having converted by a defined time point or a summary measure of time to conversion, assessed by smear or culture. Any study design where more than 10 participants were included was considered. Record sifting and data extraction were performed in duplicate. Random effects meta-analyses were performed. A narrative summary additionally describes the results of a systematic search for data evaluating infectiousness from humans to experimental animals (PubMed, all studies to 27 March 2018). Other evidence on duration of infectiousness-including studies reporting on cough dynamics, human tuberculin skin test conversion, or early bactericidal activity of TB treatments-was outside the scope of this review. The literature search was repeated on 22 November 2020, at the request of the editors, to identify studies published after the previous censor date. Four small studies reporting 3 different outcome measures were identified, which included no data that would alter the findings of the review; they are not included in the meta-analyses. Of 5,290 identified records, 44 were included. Twenty-seven (61%) were RCTs and 17 (39%) were cohort studies. Thirteen studies (30%) reported data from Africa, 12 (27%) from Asia, 6 (14%) from South America, 5 (11%) from North America, and 4 (9%) from Europe. Four studies reported data from multiple continents. Summary estimates suggested smear conversion in 9% of patients at 2 weeks (95% CI 3%-24%, 1 single study [N = 1]), and 82% of patients at 2 months of treatment (95% CI 78%-86%, N = 10). Among baseline smear-positive patients, solid culture conversion occurred by 2 weeks in 5% (95% CI 0%-14%, N = 2), increasing to 88% at 2 months (95% CI 84%-92%, N = 20). At equivalent time points, liquid culture conversion was achieved in 3% (95% CI 1%-16%, N = 1) and 59% (95% CI 47%-70%, N = 8). Significant heterogeneity was observed. Further interrogation of the data to explain this heterogeneity was limited by the lack of disaggregation of results, including by factors such as HIV status, baseline smear status, and the presence or absence of lung cavitation.Conclusions This systematic review found that most patients remained culture positive at 2 weeks of TB treatment, challenging the view that individuals are not infectious after this interval. Culture positivity is, however, only 1 component of infectiousness, with reduced cough frequency and aerosol generation after TB treatment initiation likely to also be important. Studies that integrate our findings with data on cough dynamics could provide a more complete perspective on potential transmission of Mycobacterium tuberculosis by individuals on treatment.Trial registration Systematic review registration: PROSPERO 85226.Author summaryWhy was this study done?Though cited in many countries' national guidance, the evidence that individuals with pulmonary tuberculosis (TB) are rendered non-infectious by 2 weeks of effective TB treatment is challenged. This systematic review was commissioned by the World Health Organization to provide evidence to inform TB infection prevention and control guidelines. We sought to synthesise the available data on the clearance of potentially infectious TB bacteria from the sputum of patients after starting effective treatment.What did the researchers do and find?We performed systematic searches of literature databases to identify relevant articles, using predetermined inclusion criteria. Extracted data were synthesised using narrative summaries and meta-analyses. A minority of patients had clearance of TB bacteria from sputum at 2 weeks of effective treatment, as assessed by either sputum smear or culture. As expected, the proportion having cleared TB bacteria from sputum increased over time; however, at 2 months of treatment 12% and 41% of patients still had viable TB bacteria present, as assessed by solid and liquid culture, respectively.What do these findings mean?The presence of viable TB bacteria in the sputum of pulmonary TB patients beyond 2 weeks of effective treatment suggests individuals may be infectious for longer than this interval. TB transmission requires the presence of viable mycobacteria in sputum and a mechanism for aerosol or droplet spread. Understanding how other factors, such as the presence of cough, change during treatment is also important for TB infection prevention and control.