A disaggregated dynamic model of pulmonary and extrapulmonary tuberculosis for vaccine and booster-schedule evaluation: VPM1002 in India as a case study
- 1Aseesa Inc., and UCSF
Abstract
IntroductionExtrapulmonary tuberculosis (EPTB) is rising in India and carries higher case fatality than pulmonary tuberculosis (PTB). A Phase III trial of VPM1002 missed its primary composite endpoint, but secondary endpoints showed a stronger EPTB signal (vaccine efficacy 42{middle dot}3%, 95% CI -9{middle dot}1 to 69{middle dot}4). We identified no published model resolving tuberculosis by disease type, nutritional and socioeconomic status, and dose event against a dynamic baseline.
MethodsWe built a deterministic model with 1,152 population compartments (72 strata by age, HIV status, body-mass index and socioeconomic status), calibrated to WHO India incidence and mortality for 2015-2024. Separate PTB and EPTB effectiveness posteriors, derived by Bayesian evidence synthesis of the trials per-protocol and modified intention-to-treat estimates, were propagated through 1,000 Monte Carlo iterations. Three doses were modeled at 10-year intervals assuming 3 or 10 years of protection. Economic analysis used societal and health-system perspectives with 3% discounting.
ResultsPosterior median effectiveness was higher against EPTB (42{middle dot}9%, 95% credible interval [CrI] -5{middle dot}6 to 69{middle dot}1) than PTB (14{middle dot}0%, -19{middle dot}0 to 37{middle dot}9). Per 1,000,000 targeted under 10-year protection, three doses averted 5,218 cases (95% uncertainty interval 153-10,157), 682 deaths (24-1,255) and 9,463 disability-adjusted life-years (DALYs; 328-17,404); under 3-year protection, 1,961 cases and 3,708 DALYs. EPTB accounted for about two-thirds of averted deaths and DALYs but about half of averted cases. Under 10-year protection, the first booster added 46% more cases and 34% more DALYs at an incremental benefit-cost ratio of 19{middle dot}1 (initial dose 20{middle dot}7); the second booster added 8% more cases and 4{middle dot}5% more DALYs at 6{middle dot}8. Median gross incremental cost-effectiveness ratios were US$329 (10-year) and US$832 (3-year) per DALY averted.
ConclusionComposite-endpoint models cannot show this dissociation. A booster a decade later was nearly as economically efficient but averted less disease. This is an adolescent and adult booster program, not a replacement for infant BCG.
What is already known on this topicO_LITB vaccine modeling studies for India and other high burden settings have relied on aggregate single efficacy estimates that combine pulmonary and extrapulmonary disease into a single endpoint. C_LIO_LIPublished models have evaluated vaccines against static programmatic baselines and a single course intervention, so the marginal value of revaccination has rarely been quantified. C_LI
What this study addsO_LIThis 1,152 population compartment model tracks pulmonary and extrapulmonary TB as separate compartments with independently calibrated natural history and separately propagated Bayesian effectiveness posteriors. C_LIO_LIVaccination history is preserved through disease treatment and recovery, so people infected after vaccination remain correctly identified as previously vaccinated and eligible for later doses; this becomes relevant only once boosting is modeled C_LIO_LIExtrapulmonary disease accounts for approximately 2/3 of averted deaths and DALYs against roughly half of averted cases, a dissociation that a composite endpoint model cannot provide. C_LIO_LISeed-matched paired Monte Carlo makes each dose events marginal contribution directly estimable rather than inferential. C_LI
How this study might affect research, practice or policyO_LIPolicy assessments anchored to composite endpoints, or assuming a single lifetime dose, may underestimate this vaccines population value. C_LIO_LIA single booster a decade after the initial dose is close to as economically efficient as the initial dose; a second booster is materially less so within a horizon to 2050. C_LIO_LIThe framework is country- and candidate-agnostic and can be recalibrated for other high-burden settings and other vaccine candidates, including those evaluated in other Phase III trials. C_LIO_LIThe schedule evaluated here is an adolescent and adult booster program and does not bear on infant BCG policy. C_LI
Related
- Clinical Characteristics of Patients with Diabetes and Pulmonary Tuberculosis with Additional Extrapulmonary Involvement: A Retrospective Comparative Analysis.
- Finding a needle in a haystack: Evaluation of Cartridge-Based Nucleic Acid Amplification Test in comparison with microscopy and culture for diagnosis of Extrapulmonary Tuberculosis - A study from a tertiary care centre, Kerala.
- Assessing early therapeutic response in extrapulmonary tuberculosis using [18F]FDG-PET/CT imaging
- High burden of extrapulmonary tuberculosis and its clinical determinants: a 6-year retrospective analysis of patients at Hargeisa TB Hospital, Somaliland
- Clinical and laboratory determinants of treatment response and rifampicin resistance among Xpert MTB/RIF-confirmed extrapulmonary tuberculosis patients: A prospective study from Central India
BibTeX
@misc{revathy2026disaggregated,
title = {{A disaggregated dynamic model of pulmonary and extrapulmonary tuberculosis for vaccine and booster-schedule evaluation: VPM1002 in India as a case study}},
author = {M. Revathy and V. Niranjan and A. Uttarkar and V. Swaminathan and A. Carrese and A. Bhargava},
howpublished = {medRxiv},
year = {2026},
month = jul,
doi = {10.64898/2026.07.31.26359434},
url = {https://www.medrxiv.org/content/10.64898/2026.07.31.26359434v2},
note = {CC BY-NC-ND},
}
RIS
TY - UNPB
TI - A disaggregated dynamic model of pulmonary and extrapulmonary tuberculosis for vaccine and booster-schedule evaluation: VPM1002 in India as a case study
AU - M. Revathy
AU - V. Niranjan
AU - A. Uttarkar
AU - V. Swaminathan
AU - A. Carrese
AU - A. Bhargava
PY - 2026
DA - 2026/07/31
PB - medRxiv
DO - 10.64898/2026.07.31.26359434
UR - https://www.medrxiv.org/content/10.64898/2026.07.31.26359434v2
AB - IntroductionExtrapulmonary tuberculosis (EPTB) is rising in India and carries higher case fatality than pulmonary tuberculosis (PTB). A Phase III trial of VPM1002 missed its primary composite endpoint, but secondary endpoints showed a stronger EPTB signal (vaccine efficacy 42{middle dot}3%, 95% CI -9{middle dot}1 to 69{middle dot}4). We identified no published model resolving tuberculosis by disease type, nutritional and socioeconomic status, and dose event against a dynamic baseline. MethodsWe built a deterministic model with 1,152 population compartments (72 strata by age, HIV status, body-mass index and socioeconomic status), calibrated to WHO India incidence and mortality for 2015-2024. Separate PTB and EPTB effectiveness posteriors, derived by Bayesian evidence synthesis of the trials per-protocol and modified intention-to-treat estimates, were propagated through 1,000 Monte Carlo iterations. Three doses were modeled at 10-year intervals assuming 3 or 10 years of protection. Economic analysis used societal and health-system perspectives with 3% discounting. ResultsPosterior median effectiveness was higher against EPTB (42{middle dot}9%, 95% credible interval [CrI] -5{middle dot}6 to 69{middle dot}1) than PTB (14{middle dot}0%, -19{middle dot}0 to 37{middle dot}9). Per 1,000,000 targeted under 10-year protection, three doses averted 5,218 cases (95% uncertainty interval 153-10,157), 682 deaths (24-1,255) and 9,463 disability-adjusted life-years (DALYs; 328-17,404); under 3-year protection, 1,961 cases and 3,708 DALYs. EPTB accounted for about two-thirds of averted deaths and DALYs but about half of averted cases. Under 10-year protection, the first booster added 46% more cases and 34% more DALYs at an incremental benefit-cost ratio of 19{middle dot}1 (initial dose 20{middle dot}7); the second booster added 8% more cases and 4{middle dot}5% more DALYs at 6{middle dot}8. Median gross incremental cost-effectiveness ratios were US$329 (10-year) and US$832 (3-year) per DALY averted. ConclusionComposite-endpoint models cannot show this dissociation. A booster a decade later was nearly as economically efficient but averted less disease. This is an adolescent and adult booster program, not a replacement for infant BCG. What is already known on this topicO_LITB vaccine modeling studies for India and other high burden settings have relied on aggregate single efficacy estimates that combine pulmonary and extrapulmonary disease into a single endpoint. C_LIO_LIPublished models have evaluated vaccines against static programmatic baselines and a single course intervention, so the marginal value of revaccination has rarely been quantified. C_LI What this study addsO_LIThis 1,152 population compartment model tracks pulmonary and extrapulmonary TB as separate compartments with independently calibrated natural history and separately propagated Bayesian effectiveness posteriors. C_LIO_LIVaccination history is preserved through disease treatment and recovery, so people infected after vaccination remain correctly identified as previously vaccinated and eligible for later doses; this becomes relevant only once boosting is modeled C_LIO_LIExtrapulmonary disease accounts for approximately 2/3 of averted deaths and DALYs against roughly half of averted cases, a dissociation that a composite endpoint model cannot provide. C_LIO_LISeed-matched paired Monte Carlo makes each dose events marginal contribution directly estimable rather than inferential. C_LI How this study might affect research, practice or policyO_LIPolicy assessments anchored to composite endpoints, or assuming a single lifetime dose, may underestimate this vaccines population value. C_LIO_LIA single booster a decade after the initial dose is close to as economically efficient as the initial dose; a second booster is materially less so within a horizon to 2050. C_LIO_LIThe framework is country- and candidate-agnostic and can be recalibrated for other high-burden settings and other vaccine candidates, including those evaluated in other Phase III trials. C_LIO_LIThe schedule evaluated here is an adolescent and adult booster program and does not bear on infant BCG policy. C_LI
ER -