Annex 3 – Risk of acquiring tuberculosis infection, progression to active disease and the effect of treatment on infectiousness

This section summarizes a series of complementary systematic reviews aimed at describing the risk of developing tuberculosis (TB) infection or progressing to TB disease in specific at-risk populations. It also provides a brief description of the effect of treatment on infectiousness.

Although anyone in any setting can be at risk of acquiring TB infection and progressing to TB disease once exposed (depending on host susceptibility), the actual risk of transmission is influenced by many physiological, structural and social determinants, including the ways in which individuals live and interact, severity of the TB disease in a source case, proximity, duration and frequency of contact. This effect is more prominent among health workers, household contacts of TB patients and other persons in non-health care congregate settings (see Online annex 6).

The extent of the risks of acquiring TB infection and of progressing to active TB disease, as well as the infectiousness of TB patients while on treatment, have been reported in various studies. To better inform the development of targeted recommendations based on the extent of increase in risk for M. tuberculosis transmission, the WHO Guideline Steering Group outlined a series of background questions to assess such risks in the context of the current guideline. These background questions were designed to support the making of specific recommendations for at-risk groups – health workers, TB contacts and individuals exposed to active TB disease while in at-risk congregate settings.

The objectives of the three systematic reviews informing the recommendations given here were to determine:

  • the risk of TB infection or active TB disease among health workers (compared with the general population);
  • the risk of TB infection or active TB disease among individuals living in TB-affected households and among those spending time in congregate settings (compared with the general population); and
  • how the infectiousness of TB patients changes after starting on effective treatment (depending on susceptibility patterns).
Review for background question 1 – occurrence of latent TB infection and incidence of active disease in health workers

This systematic review identified 41 articles that reported data on at least one of TB infection prevalence, TB infection incidence and TB disease incidence, in health workers and in a suitable comparator population. Studies were conducted in low and high TB burden countries; few of these studies reported adjusted effect estimates. Overall, health workers had twice the odds of TB infection compared with the general population without exposure to a health care setting, with little difference when stratified by high or low TB burden countries. Heterogeneity across these studies was low. Three studies from high TB burden settings reported adjusted odds ratios (ORs) for TB infection prevalence; the summary OR indicated that health workers had a 60% (35–93%) increase in odds of TB infection compared with the general population.

Summary rate ratios (RRs) from the metaanalyses for TB disease incidence need to be interpreted with extreme caution because heterogeneity across studies was very high. For high TB burden countries (eight studies), the random and fixed effects summary RRs were 4.32 and 3.00, respectively. For low TB burden countries (12 studies), the summary RR using random effects was 1.28, with the 95% confidence interval (CI) crossing 1.0. Five of 12 unadjusted study-level RRs were below 1.0.

Online annexes provide access to the data analysis report.

Review for background question 2 – occurrence of latent TB infection and progression to active disease in TB contacts and other individuals in congregate settings

This systematic review identified 93 primary research articles that reported data on TB infection prevalence and incidence, or on TB disease incidence, in TB-affected households (72 articles) or congregate settings (21 articles) compared with a comparator population. Studies were conducted in low and high TB burden countries; few of these studies reported adjusted effect estimates.

For congregate settings, most data came from correctional facilities. Five studies reported TB disease incidence in correctional facilities; stratified by TB disease burden, the summary RRs indicated higher TB disease incidence rates compared with the general population (RR 6.5 for high TB burden countries and 7.3 for low TB burden countries). However I2s were large (>80%), indicating extremely high study heterogeneity, and the CIs were wide, making the summary estimate difficult to interpret. Six studies did not contribute data to the metaanalysis because the data were too sparse to calculate a CI for the effect estimate. Three studies conducted in nursing homes in low TB burden settings were included in a meta-analysis for TB disease incidence. The summary estimate indicated similar rates for nursing home residents and for the general population, although study heterogeneity was high.

For TB-affected households, the most common outcome was TB infection prevalence, measured using the tuberculin skin test (TST) or interferon-gamma release assay (IGRA). For the latter, QuantiFERON-TB Gold and T SPOT assays were used. Study heterogeneity was extremely high (>84%) for the outcome of TST positivity prevalence among children, adults and combined populations (i.e. both children and adults) living in high TB burden settings. Summary ORs, using random effects, were 3.9, 3.0 and 2.2, respectively, for children, adults and combined populations (the original article did not report data separately for children and adults), comparing TB-affected households with a comparator population.

Extreme caution must be taken in interpreting these summary estimates because high study heterogeneity means they are difficult to interpret. Fewer studies contributed to the IGRA outcome. Study heterogeneity was low for children and adult populations, separately, in high TB burden settings, and summary estimates suggested about a twofold to threefold increase in odds of TB infection for TB-affected households compared with comparator households. Case–control studies mainly contributed data for the outcome of TB disease prevalence, and most studies were conducted in a high TB burden setting. Among children and adults separately, the summary ORs were 5.3 and 2.9, respectively, and study heterogeneity was less than 60%. Meta-analyses were based on unadjusted ORs for unmatched studies and ORs from matched case–control studies, controlling for the matching variables only.

Study quality using the modified Newcastle Ottawa scale was low for all study designs.

Online annexes provide access to the data analysis report.

Review for background question 3 – the effect of effective anti-TB treatment on infectiousness

This systematic review was undertaken to determine smear and culture conversion times achieved by effective anti-TB therapy, and to capture the literature examining the effect of TB treatment on the infectiousness of TB patients, as measured using exposed laboratory animals.

After title and abstract sifting of 5290 unique records identified through literature search, and full text review of 180 articles, 47 papers were included for data extraction and analysis for time to smear and culture conversion. Considerable variation in smear and culture conversion times was seen across the studies.

For the six studies reporting a summary estimate of time to smear conversion, the median time to conversion was 20–27 days and the mean time to conversion was 29–61 days. The summary estimates across eight studies of the proportion of baseline smear-positive patients achieving smear conversion at the 1-, 2- and 3-month time points were 0.3 (95% CI: 0.27–0.34), 0.79 (95% CI: 0.70–0.87) and 0.95 (95% CI: 0.94– 0.95), respectively.

For the five studies reporting a summary estimate of time to solid culture conversion, the median time to conversion was 35–49 days. For the five studies reporting a summary estimate of time to culture conversion where culture method was not specified, the median time to conversion was 32–48 days. For the five studies reporting a summary estimate of time to liquid culture conversion, the median time to conversion was 28–125 days.

The proportions of patients achieving culture conversion after 2 months of appropriate treatment for proven drug-susceptible TB for solid culture, unspecified culture method and liquid culture were 0.84 (95% CI: 0.80–0.88), 0.78 (95% CI: 0.67–0.88) and 0.67 (95% CI: 0.54–0.81), respectively.

Factors affecting time to culture conversion, such as the proportion of the study population with heavily smear-positive disease, were rarely reported and are likely to have accounted for some of the heterogeneity.

Four relevant papers were identified in the additional systematic review of the effect of TB treatment on infectiousness of TB patients to experimental animals. All of these studies contained data suggesting that patients on TB treatment are less infectious to guineapigs than patients not receiving effective TB treatment. However, there were no data indicating the time it takes for a patient receiving effective treatment to become noninfectious to guinea-pigs.

Online annexes provide access to the data analysis report.

Book navigation