Michael Jacques Rossouw 1, Robert Wallis 2, Hardy Kornfeld 3, Amit Singhal 4, Gary Maartens 1, Paolo Denti 1, Roeland Wasmann 1
1 Division of Clinical Pharmacology, Department of Medicine, University of Cape Town (Cape Town, South Africa), 2 The Aurum Institute (Johannesburg, South Africa), 3 Department of Medicine, UMass Chan Medical School (Worcester, United States of America), 4 A*STAR Infectious Diseases Labs, Agency for Science Technology and Research (A*STAR) (, Republic of Singapore)
Objective
Tuberculosis (TB) remains the leading infectious cause of death, particularly among people with HIV (1,2). Metformin is being investigated as an immunomodulating, host-directed adjunct to tuberculosis therapy (3,4) and is also widely used for type-2 diabetes, a common comorbidity in TB. As metformin use expands, drug–drug interactions with first-line antitubercular drugs are likely to be encountered more frequently in clinical practice (5). We aimed to quantify the effect of metformin coadministration on the pharmacokinetics of rifampicin, isoniazid, and pyrazinamide in adults with HIV-associated pulmonary-TB receiving standard first-line TB treatment.
Methods
This pharmacokinetic substudy was nested within METHOD, a Phase IIa randomised, open-label trial conducted in South Africa (2021–2025; NCT04930744). Adults with HIV-1 and rifampicin-susceptible pulmonary TB, but without diabetes, were randomised (1:1) to standard TB therapy with or without adjunctive metformin (daily 500mg for 1 week, then twice daily until week 12). ART-naïve participants initiated ART within 8 weeks and received dolutegravir-based ART with an additional 50mg evening dolutegravir dose to account for rifampicin drug-drug interactions (6).
Intensive (predose and 1, 2, 4, 6, 8, 10h) or semi-intensive (predose, 2, 6h) sampling was performed at week 5 after an observed dose, following an overnight fast with a standardised meal at 1h post-dose. Concentrations were quantified using validated LC–MS/MS assays.
Data were analysed using nonlinear mixed-effects modelling (NONMEM 7.5.1, FOCE-I). One- or two-compartment models with first-order absorption and elimination were evaluated for all drugs, with allometric scaling of clearance and volume parameters. BLQ (below limit of quantification) data were handled using the M7+ method (7). NAT2 genotype data were unavailable; therefore, a mixture model was implemented for isoniazid alongside hepatic extraction. Rifampicin models additionally evaluated saturable hepatic elimination, first-pass metabolism, and transit-compartment absorption. Steady-state AUC0-24h was derived from population model predictions and used for simulation of dose-adjustment scenarios and DDI evaluation.
Results
Seventy-eight participants (310 samples; 43 receiving metformin) were included; median weight was 60.8 kg (IQR 53.9–67.8), 37% were female, and 81% were receiving antiretroviral therapy at the PK visit. BLQ plasma concentrations were observed for isoniazid (21.5%), rifampicin (23.2%), and pyrazinamide (1.6%), predominantly at pre-dose. Rifampicin and pyrazinamide were adequately described by one-compartment models, with fat-free mass outperforming total body weight for allometric scaling on all three drugs. Isoniazid pharmacokinetics was described by a two-compartment model with hepatic extraction and a mixture model. Estimated clearance was 10.5 L/h (95% CI: 9.44–11.9) in slow acetylators and 28.8 L/h (95% CI: 25.6–31.0) in intermediate/rapid acetylators, with approximately 30% of participants classified as slow acetylators. Further subdivision of NAT2 phenotypes was not supported by the available data. Metformin coadministration lowered the oral bioavailability of rifampicin by 24.1% (95% CI 7.83−36.3; p<0.007) and isoniazid by 15.6% (95% CI 4.45−26.4−; p<0.009), with no significant effect on pyrazinamide. However, pyrazinamide showed a statistically significant effect (p<0.05) on absorption rate in univariate analysis, this was not retained in the final model, as it did not meet the predefined backward elimination criterion (p < 0.01). Visual predictive checks stratified indicated adequate model performance. Model-based simulations indicated that metformin shifted rifampicin exposure toward the lower end of historical reference ranges (8,9); a simulated rifampicin dose increase of +150 mg mitigated this effect.
Conclusions
In adults with HIV-associated pulmonary TB receiving standard first-line therapy, metformin coadministration was associated with reduced oral bioavailability of rifampicin (~24%) and isoniazid (~16%), while pyrazinamide exposure was not significantly affected. Given the fixed-dose combination formulation, these findings may suggest an effect of metformin on drug absorption kinetics rather than isolated drug-specific interactions. Given rifampicin’s exposure–response relationship (10), these findings support consideration of pragmatic rifampicin dose adaptation during metformin coadministration; simulations suggest that an additional 150 mg rifampicin dose may offset the reduced exposure. Importantly, because metformin is widely used in people with diabetes, these drug–drug interactions are likely to be clinically relevant in TB patients with diabetes, who are already at increased risk of poor outcomes. This is supported by a recent study showing that diabetes independently predicted lower rifampicin and isoniazid exposure, underscoring the need for further investigation in patients with diabetes, as the potential contribution of metformin was not assessed (11).
References:
1. World Health Organization. Global Tuberculosis Report 2024. Geneva: WHO; 2024.
2. National Department of Health (South Africa). National TB Recovery Plan 4. Pretoria; 2025.
3. Padmapriyadarsini C, Mamulwar M, Mohan A, et al. Randomized trial of metformin with anti-tuberculosis drugs for early sputum conversion in pulmonary TB. Clin Infect Dis. 2022;75(3):425–434.
4. Smith SG, Bowness R, Cliff JM. Host-directed therapy in diabetes and tuberculosis comorbidity toward global TB elimination. Int J Infect Dis. 2025;155.
5. Kumar AKH, Kadam A, Karunaianantham R, et al. Effect of metformin on plasma exposure of rifampicin, isoniazid, and pyrazinamide in pulmonary TB. Ther Drug Monit. 2022.
6. Dooley KE, Sayre P, Borland J, et al. Safety, tolerability, and pharmacokinetics of dolutegravir with rifampin or rifabutin in healthy subjects. 2012.
7. Wijk M, Wasmann RE, Jacobson KR, et al. Handling censored data below the LLOQ in pharmacokinetic modeling. CPT Pharmacometrics Syst Pharmacol. 2025;14(6):1042–1049.
8. Daskapan A, Idrus LR, Postma MJ, et al. Effect of HIV infection on first-line TB drug pharmacokinetics: systematic review. Clin Pharmacokinet. 2019;58:747–766.
9. Stott KE, Pertinez H, Sturkenboom MGG, et al. Pharmacokinetics of rifampicin in adult TB patients and healthy volunteers: systematic review and meta-analysis. J Antimicrob Chemother. 2018;73(9):2305–2313.
10. Wijk M, Denti P, Gausi K, et al. Rifampicin exposure and baseline covariates influencing pulmonary TB treatment response. Clin Infect Dis. 2025.
11. Mtabho CM, Semvua HH, Van den Boogaard J, et al. Effect of diabetes mellitus on TB drug concentrations in Tanzanian patients. J Antimicrob Chemother. 2019;74(12):3537–3545.
Reference: PAGE 34 (2026) Abstr 11985 [www.page-meeting.org/?abstract=11985]
Poster: Drug/Disease Modelling - Infection