The potential impact of preventive therapy against tuberculosis in the WHO South-East Asian Region: a modelling approach

Abstract

Background: The prevention of tuberculosis (TB) is key for accelerating current, slow declines in TB burden. The 2018 World Health Organization (WHO) guidelines on eligibility for preventive therapy to treat latent TB infection (LTBI) include people living with human immunodeficiency virus (PLHIV), household contacts of TB patients including children, and those with clinical conditions including silicosis, dialysis, transplantation, etc. and other country-specific groups. We aimed to estimate the potential impact of full implementation of these guidelines in the WHO South-East Asian (SEA) Region, which bears the largest burden of TB and LTBI amongst the WHO regions.

Methods: We developed mathematical models of TB transmission dynamics, calibrated individually to each of the 11 countries in the region. We modelled preventive therapy in the absence of other TB interventions. As an alternative comparator, reflecting ongoing developments in TB control in the region, we also simulated improvements in the treatment cascade for active TB, including private sector engagement and intensified case-finding. Relative to both scenarios, for each country in the region, we projected TB cases and deaths averted between 2020 and 2030, by full uptake of preventive therapy, defined as comprehensive coverage amongst eligible populations as per WHO guidelines, and assuming outcomes consistent with clinical trials. We also performed sensitivity analysis to illustrate impact under less-than-optimal conditions.

Results: At the regional level, full uptake of preventive therapy amongst identified risk groups would reduce annual incidence rates in 2030 by 8.30% (95% CrI 6.48-10.83) relative to 2015, in the absence of any additional interventions. If implemented against a backdrop of improved TB treatment cascades, preventive therapy would achieve an incremental 6.93 percentage points (95% CrI 5.81-8.51) of reduction in annual incidence rates, compared to 2015. At the regional level, the numbers of individuals with latent TB infection that need to be treated to avert 1 TB case is 64 (95% CrI 55-74). Sensitivity analysis illustrates that results for impact are roughly proportional to 'effective coverage' (the product of actual coverage and effectiveness of the regimen).

Conclusions: Full implementation of WHO guidelines is important for ending TB in the SEA Region. Although future strategies will need to be expanded to the population level, to achieve large declines in TB incidence, the uptake of current tools can offer a valuable step in this direction.

Keywords: Epidemiology; Modelling; Preventive therapy; South East Asia; Tuberculosis.

Fig. 1

Illustration of the model structure. Panel a shows the compartmental model framework representing TB natural history and the uptake of TB services. Abbreviations: ‘Dx’ denotes diagnosis; ‘Tx’ denotes ‘treatment’. The structure shown here is further stratified by HIV status, according to the categories shown in panel b and by rifampicin resistance (not shown here for clarity). Further technical details are provided in the supporting information

Fig. 2

Model projections for incidence impact in SEAR, under adoption of WHO guidelines for management of LTBI in the region. See Table 1 for impact by country, as well as in terms of incidence and mortality. Shaded intervals show 95% Bayesian credible intervals. As described in the main text, the ‘status quo comparator represents current TB services continuing indefinitely without change, while the ‘Improved cascade’ comparator incorporates background improvements in TB care, including comprehensive engagement with the private healthcare sector, and intensified case-finding, throughout the region

Fig. 3

Numbers-needed-to-treat with preventive therapy, to prevent 1 TB case. Figure shows estimates stratified by the 11 countries in the region, as a simple proxy for the effort required to achieve the incidence declines shown in Fig. 2. Error bars show 95% Bayesian credible intervals. In the second panel, numbers-needed-to-treat are disproportionately high for the Maldives because of a low incidence (33 per 100 k population), as well as a low reported TB mortality rate (0.15 per 100 k population)