Highlight
– The HARVEST randomized, double‑blind trial (n=499) compared high‑dose oral rifampin (cumulative 35 mg/kg/day for 8 weeks) versus standard rifampin dosing (10 mg/kg/day) in adults with tuberculous meningitis (TBM).
– Six‑month mortality was 44.6% in the high‑dose group versus 40.7% in the standard‑dose group (HR 1.17; 95% CI 0.89–1.54; P=0.25), with no evidence of benefit from higher rifampin dosing.
– Drug‑induced liver injury occurred more often with high‑dose rifampin (8.0% vs 4.4%), and median time to death was shorter in the high‑dose group (13 vs 24 days among those who died).
Background and Unmet Need
Tuberculous meningitis (TBM) is the most lethal form of extrapulmonary tuberculosis, with high early mortality and frequent neurologic disability among survivors. Global management of TBM commonly combines standard antitubercular drugs with adjunctive corticosteroids; despite this approach, outcomes remain poor, particularly in people living with HIV. Rifampin is a cornerstone drug for Mycobacterium tuberculosis, but its penetration into cerebrospinal fluid (CSF) is limited when used at standard doses, raising the biologic argument that higher rifampin exposures could increase mycobacterial killing in the central nervous system and improve outcomes.
Study Design
The HARVEST trial (Meya et al., N Engl J Med 2025) was a double‑blind, randomized, placebo‑controlled multicenter study conducted in Indonesia, South Africa, and Uganda. Adults with definite, probable, or suspected TBM were enrolled, including persons with and without HIV coinfection.
All participants received a standard four‑drug antitubercular regimen daily (isoniazid, rifampin at 10 mg/kg, ethambutol, and pyrazinamide) plus either additional rifampin or matched placebo for the first 8 weeks. The experimental arm received supplementary rifampin so that the cumulative rifampin dose during the intensive phase equaled 35 mg/kg/day (termed “high‑dose” in the trial); the control arm received placebo to maintain blinding (standard 10 mg/kg rifampin). After 8 weeks, both groups continued standard TB therapy to complete a 9–12 month treatment course. The primary endpoint was 6‑month all‑cause mortality. Safety endpoints included adverse events of special interest such as drug‑induced liver injury (DILI).
Key Results
Population and baseline characteristics: 499 participants were randomized and included in the intention‑to‑treat (ITT) population (249 in the high‑dose group; 250 in the standard‑dose group). Of the cohort, 304 (60.9%) were persons living with HIV; 428 (85.8%) met criteria for definite or probable TBM.
Primary outcome: During 6 months of follow‑up, 109 participants in the high‑dose group died (Kaplan‑Meier estimate 44.6%) versus 100 in the standard‑dose group (Kaplan‑Meier estimate 40.7%). The hazard ratio (HR) for death with high‑dose rifampin compared with standard dosing was 1.17 (95% confidence interval [CI], 0.89–1.54; P=0.25). The trial therefore did not demonstrate a survival benefit with high‑dose oral rifampin.
Timing of deaths: Among participants who died within 6 months, the median time to death was earlier in the high‑dose group (13 days; interquartile range [IQR], 4–39) than in the standard‑dose group (24 days; IQR, 6–56). This finding raises concern that higher early dosing may be associated with earlier deterioration in some patients, although causality was not established.
Safety and adverse events: Drug‑induced liver injury was more frequent in the high‑dose arm (8.0%) than in the standard‑dose arm (4.4%), although no DILI‑related deaths were reported. Other safety signals were not described as leading to clear differences in mortality, but the overall adverse‑event profile favored caution with the higher rifampin exposure.
Subgroup findings: The trial report concludes there was no evidence of benefit across the enrolled population, and the possibility of harm could not be excluded. The primary publication did not support routine adoption of the high‑dose oral rifampin regimen for adults with TBM.
Interpretation and Biological Plausibility
The negative result contrasts with the pharmacologic rationale that higher rifampin concentrations might enhance bactericidal activity in the central nervous system. Several mechanistic and practical explanations may explain the lack of clinical benefit:
- Pharmacokinetics and CNS penetration: Rifampin is highly protein‑bound and its CSF penetration is limited, especially once the blood‑brain barrier starts to recover during therapy. Oral dose escalation may not result in proportionally higher free CSF concentrations because of saturation of absorption, first‑pass metabolism, and active efflux transporters at the blood‑brain barrier.
- Variable absorption in critically ill patients: Gastrointestinal dysfunction, malabsorption, or altered oral bioavailability early in severe TBM may blunt expected increases in systemic and CSF exposure from high oral dosing.
- Timing and disease biology: TBM mortality typically occurs early; if higher rifampin exposure is not reliably achieved in the first days of therapy (or if irreversible neurologic injury has already occurred), dose escalation may not alter the clinical course.
- Host toxicity and competing risks: Increased hepatotoxicity and drug interactions (particularly in persons receiving antiretroviral therapy) can offset any microbiologic benefit and contribute to earlier clinical decline.
Clinical and Research Implications
For practicing clinicians:
- Do not adopt routine high‑dose oral rifampin (35 mg/kg for 8 weeks achieved by supplemental dosing) for adults with TBM based on current evidence; standard‑dose regimens remain appropriate.
- Maintain vigilance for early clinical deterioration and hepatotoxicity in TBM patients, particularly when drug exposure is modified or when concomitant hepatotoxic drugs are used.
- Manage coexisting HIV infection, raised intracranial pressure, and supportive care aggressively—these remain critical determinants of outcome.
For researchers and policy makers:
- Pharmacokinetic approaches: Future trials should incorporate intensive PK sampling (plasma and CSF), therapeutic‑drug‑monitoring guided dosing, and possibly alternative routes (intravenous rifampin) to determine whether reliably higher CSF exposures can be achieved safely.
- Alternative rifamycins and combinations: Investigation of other rifamycins with different PK properties (e.g., rifapentine) or regimens including adjunctive host‑directed therapies may be warranted.
- Stratified or early‑intervention trials: Enrolling patients earlier in their disease course and stratifying by severity or HIV status could clarify whether particular subgroups benefit from altered dosing or adjuncts.
- Biomarker development and rapid diagnostics: Improving early diagnosis and stratification of TBM severity remains essential to identify patients most likely to gain from intensified antimicrobial strategies.
Study Strengths and Limitations
Strengths of HARVEST include its randomized, double‑blind, placebo‑controlled design; multinational conduct across settings with high TBM burden; inclusion of persons with HIV; and clinically meaningful primary endpoint (6‑month mortality). These features enhance the trial’s relevance to global practice.
Limitations include absence (in the primary report) of detailed PK‑pharmacodynamic data showing whether target CSF exposures were achieved in individuals; potential heterogeneity across sites in ancillary care (including neurosurgical management and HIV treatment timing); and the possibility that oral dose escalation does not reliably translate into increased CNS exposure in the earliest, most critical phase of disease. The median earlier time to death among the high‑dose deceased subgroup is a signal that merits further mechanistic investigation but does not prove causation.
Conclusion
The HARVEST trial provides high‑quality evidence that adding high‑dose oral rifampin to standard TBM therapy (to reach a cumulative 35 mg/kg during the intensive phase) did not reduce 6‑month mortality and was associated with higher rates of drug‑induced liver injury and a pattern of earlier deaths among those who died. These data do not support routine use of this high‑dose oral rifampin strategy in adult TBM and emphasize the need for mechanistic PK studies and alternative strategies to improve outcomes in this devastating disease.
Funding and Registration
The trial was funded by the U.K. Medical Research Council and other partners. The study is registered at ISRCTN (ISRCTN15668391).
Selected References
1) Meya DB, Cresswell FV, Dai B, Engen N, Naidoo K, Ganiem AR, et al.; HARVEST Trial Team. Trial of High‑Dose Oral Rifampin in Adults with Tuberculous Meningitis. N Engl J Med. 2025 Dec 18;393(24):2434–2446. doi:10.1056/NEJMoa2502866. PMID: 41406445.
2) Thwaites GE, Nguyen DB, Nguyen HD, Hoang TT, Do TT, Pham PM, et al. Dexamethasone for the treatment of tuberculous meningitis. N Engl J Med. 2004;351:1741–1751. doi:10.1056/NEJMoa040573.
3) World Health Organization. Global Tuberculosis Report 2023. Geneva: WHO; 2023. Available at: https://www.who.int/teams/global-tuberculosis-programme/tb-reports (accessed 2025).
Author Note
This article summarizes and interprets the primary report of the HARVEST randomized trial (Meya et al., NEJM 2025) and places its findings in clinical and research context. Clinicians should consult the full trial report for detailed methods, subgroup analyses, and safety data before applying these findings to practice.
AI image prompt for article thumbnail
A worried clinician in a modern hospital reviewing a brain MRI and CSF laboratory printouts at a desk, pill bottles labeled ‘Rifampin’ visible in the foreground, cool clinical lighting, neutral muted colors, photographic realism, medium shot.
