Implications of progressive lung damage and post-tuberculosis sequelae for the health benefits of prompt tuberculosis treatment in high HIV prevalence settings: a mathematical modelling analysis

Abstract

Background: Untreated pulmonary tuberculosis causes ongoing lung damage, which can persist after treatment. Conventional modelling approaches for assessing tuberculosis health effects might not fully capture these mechanisms. We evaluated how tuberculosis-associated lung damage and post-tuberculosis sequelae affect the lifetime health consequences of tuberculosis in high HIV prevalence settings.

Methods: We developed a microsimulation model (computer simulations that reproduce disease natural history and intervention effects for sampled individuals) representing dynamic changes in lung function for individuals evaluated for tuberculosis in routine clinical settings. We parametrised the model with data (from a previously published study) for three African countries with a high burden of tuberculosis and HIV: Uganda, Kenya, and South Africa, and estimated lifetime health outcomes under prompt, delayed, and no tuberculosis treatment scenarios. We compared results to earlier modelling approaches that omit progressive lung damage and post-tuberculosis sequelae.

Findings: We estimated a 5·1 years (95% uncertainty interval 3·8-6·4) reduction in life expectancy due to tuberculosis with prompt treatment, 7·7 years (5·5-10·1) with delayed treatment, and 18·5 years (15·5-20·6) with no treatment. Estimated per-person disability-adjusted life-years (DALYs) from tuberculosis were 11·4 years (8·9-14·2) with prompt treatment, 17·1 years (13·1-22·1) with delayed treatment, and 37·7 years (34·3-40·3) with no treatment. Compared with individuals without HIV, individuals with HIV had a greater proportion of tuberculosis-attributable deaths, but fewer life-years lost to tuberculosis. Post-tuberculosis DALYs represented 52·5% of total DALYs with prompt treatment, 42·7% with delayed treatment, and 9·1% with no treatment. Modelling approaches that omit progressive lung damage and post-tuberculosis sequelae underestimated lifetime health losses of tuberculosis by 48-57% and underestimated the benefits of prompt treatment by 45-64%.

Interpretation: Delayed initiation of tuberculosis treatment causes greater lung damage and higher mortality risks during and after the disease episode than prompt treatment. In settings with coprevalent tuberculosis and HIV, accounting for these factors substantially increased estimates of the lifetime disease burden and life expectancy loss caused by tuberculosis. These findings imply greater health effects and cost-effectiveness for interventions to prevent tuberculosis and achieve earlier treatment initiation than indicated in previous analytical approaches.

Funding: US National Institutes of Health.

Figure 1:. Survival curves for individuals with symptomatic tuberculosis under prompt, delayed, and no treatment scenarios

Figure 2:. DALYs attributable to tuberculosis disease for each scenario by HIV status

DALYs are stratified by whether DALYs resulted from reduced quality of life or premature mortality and by whether DALYs occurred during the tuberculosis episode or after tuberculosis cure. Bars represent 95% uncertainty intervals. DALYs were calculated by comparing each scenario to a no tuberculosis counterfactual scenario. Numerical results for plotted values are shown in the appendix (p 8). DALYs=disability-adjusted life-years.