Tuberculosis remains one of the most preventable mass killers on earth, and a granular 33-year accounting of its toll now arrives at a moment when global health funding cuts risk reversing hard-won gains. Understanding exactly where progress has — and has not — occurred is foundational to any rational response strategy as the WHO's 2035 elimination targets approach their midpoint reckoning.
This GBD 2023 systematic analysis quantified TB mortality, morbidity, and disability-adjusted life-years (DALYs) across 204 countries and territories from 1990 through 2023, stratifying results by HIV co-infection status and drug-resistance profiles including multidrug-resistant TB (MDR-TB). Using an ensemble cause-of-death modelling platform fed by vital registration, surveillance, verbal autopsy, and minimally invasive tissue sampling data — alongside the DisMod-MR 2.1 simultaneous incidence-prevalence-mortality estimator — the study constructed one of the most methodologically rigorous long-term TB burden assessments to date. A comparative risk assessment framework further attributed burden fractions to alcohol use, smoking, and elevated fasting plasma glucose. Global TB mortality declined roughly 37% over the study period, yet progress toward the WHO End TB targets of 95% mortality reduction and 90% incidence reduction by 2035 remains substantially off-track, with MDR-TB and HIV-associated TB representing the steepest unfinished challenges.
For context, global TB deaths peaked at approximately 1.9 million annually in the mid-1990s, driven substantially by the HIV epidemic; antiretroviral scale-up thereafter produced the most dramatic mortality declines in high-burden regions. The current analysis reinforces that HIV-TB co-infection and MDR-TB are not simply clinical complications — they are structurally distinct epidemics requiring separate strategic targeting. The modifiable risk factors examined here — particularly smoking and dysglycemia — underscore that TB vulnerability is partly lifestyle-mediated, a point often lost in purely infectious-disease framing. The primary limitation is that ensemble modelling, while powerful, introduces uncertainty bands that widen considerably in low-resource settings with incomplete vital registration. This is a large, confirmatory-yet-alarming study: it consolidates existing knowledge but delivers a clear signal that current trajectories will miss elimination targets by a substantial margin, making its timing politically and programmatically consequential.