Tuberculosis remains one of the most consequential infectious diseases on earth, yet its true toll — stratified by HIV co-infection, drug resistance, geography, and modifiable risk factors — has been difficult to track with precision. This GBD 2023 analysis arrives at a critical moment: just as global health funding contractions threaten hard-won surveillance and treatment infrastructure, a rigorous three-decade baseline becomes essential for calibrating the WHO's 2035 elimination targets.
Drawing on vital registration data, verbal autopsy records, surveillance reports, and minimally invasive tissue sampling from 204 countries and territories spanning 1990 to 2023, the study deployed two complementary modeling platforms — the Cause of Death Ensemble model for mortality and DisMod-MR 2.1 for simultaneous incidence-prevalence-mortality estimation. Critically, a population attributable fraction framework was applied to disaggregate burden by HIV status and drug-resistance profile, including multidrug-resistant TB. Disability-adjusted life-years were calculated, and separate risk factor attributable fractions were derived for alcohol use, smoking, and elevated fasting plasma glucose — three metabolic and behavioral exposures rarely quantified together in TB burden analyses.
This represents one of the most methodologically comprehensive TB burden assessments to date, and its implications extend well beyond epidemiology. By pairing MDR-TB estimates with HIV stratification across granular age-sex subgroups, the analysis reveals where treatment gaps are likely deepest and where dual-burden populations — immunocompromised individuals facing drug-resistant strains — face compounded mortality risk. The WHO End TB Strategy's 95% mortality reduction target by 2035 is benchmarked against these estimates, providing the clearest picture yet of how far the world has come and how far it must travel. Given that this analysis predates anticipated funding disruptions, it also establishes the counterfactual baseline against which future setbacks will be measured. Its chief limitation is the variable quality of underlying data in high-burden, low-resource settings where TB deaths are most often miscoded or unregistered. Still, as a systematic anchor for policy, this is an unambiguously paradigm-informing contribution.