Practical considerations for a TB controlled human infection model (TB-CHIM); the case for TB-CHIM in Africa, a systematic review of the literature and report of 2 workshop discussions in UK and Malawi

**Background:** Tuberculosis killed more than 1.5 million people in 2021, with 214,000 among people living with HIV. The only licensed TB vaccine, BCG, shows highly variable efficacy (0–80%) in adults and is least effective in tropical regions. Multiple new vaccine candidates are in development, but phase 2b/3 trials are extremely expensive—the M72/AS01E candidate, which showed 49.7% efficacy at 3 years, would require an estimated GBP 400 million for full evaluation. Controlled human infection models (CHIM) could accelerate vaccine development by allowing down-selection of candidates before large efficacy trials. However, a TB-CHIM using wild-type M.tb faces major safety challenges: infection cannot be reliably eradicated, treatment requires 6 months of toxic therapy, and there is a ~12% risk of recurrent infection after treatment. BCG offers a safer alternative as a surrogate challenge agent.

**Methods:** The authors conducted a PROSPERO-registered systematic review (CRD42022302785) searching MEDLINE (1946–February 2022) and EBSCO CINAHL (1984–February 2022). Inclusion criteria were adult humans ≥18 years receiving live M.tb (wild-type or GMO) or BCG in interventional trials or prospective cohort studies with immune and/or microbiological endpoints. Two independent reviewers assessed titles, abstracts, and full manuscripts; discrepancies were resolved by a third reviewer. Risk of bias was assessed using the Cochrane Collaboration tool for RCTs and the Newcastle-Ottawa Scale for non-randomised studies. Additionally, two workshops were held—a 2-day residential workshop in the UK (September 2019) and a hybrid meeting in Malawi (June 2022)—to discuss practical steps for TB-CHIM development.

**Key Results:** Twenty-seven studies met inclusion criteria: 15 RCTs and 12 non-randomised interventional studies. Only three studies were conducted in high-burden settings (South Africa, Brazil). No studies used wild-type M.tb as a challenge agent. Most studies used intradermal BCG administration. Key findings include: (1) Intradermal BCG at doses of 6×10⁵–2.4×10⁶ CFU optimised bacterial detection from day 14 skin punch biopsies without increasing adverse effects. (2) BCG was recovered from all 40/40 punch biopsy specimens by both culture and qPCR in one optimised study, with strong correlation between techniques (r=0.664) and between dose and CFU count (r=0.749). (3) qPCR detected 1–2 logs higher copy numbers than culture. (4) For bronchoscopically instilled BCG, the minimal immunogenic dose was 1×10⁴ CFU, with BCG recovered in only 6/54 (11%) of BAL samples. (5) No serious adverse events were reported in any BCG challenge study; all intradermal injection sites healed by 90 days. (6) Two serious adverse events occurred in vaccine trials: one participant receiving MTBVAC in South Africa developed aseptic meningitis (unrelated to vaccine), and two participants receiving high-dose AERAS-422 developed varicella zoster virus reactivation, leading to programme discontinuation. (7) The MVA85A vaccine showed no effect in a BCG CHIM, consistent with its negative phase 2b clinical trial.

**Clinical Implications:** A TB-CHIM using BCG as a surrogate is feasible and safe, with the intradermal route being the most practical for initial development. The intradermal model has limitations—cutaneous immune responses may not fully represent pulmonary responses to M.tb—but molecular signatures from skin biopsy sites closely reflect those in lungs. The authors recommend a stepwise approach: first establish cutaneous BCG CHIM in well-resourced settings (UK), then transfer to Malawi after de-risking. In Malawi, initial studies should recruit IGRA-negative healthy volunteers to minimise risk. Modified BCG strains with fluorophore insertions could enable continuous optical readouts, and conditionally replicating M.tb strains with suicide switches are in development for future use. A TB-CHIM could test not only vaccines but also new drug regimens, with the potential to identify optimal drug combinations from the >25 drugs currently in the clinical pipeline. The model also offers scientific opportunities to study BCG's non-specific effects on trained immunity and protection against respiratory infections.