Trial of High-Dose Oral Rifampin in Adults with Tuberculous Meningitis.
Meya DB, Cresswell FV, Dai B, Engen N, Naidoo K, Ganiem AR, Imran D, Kabahubya M, Lessells RJ, Yunivita V, Estiasari R, Tugume L, Hlabisa B, Kurniawati MY, Sagita N, Kagimu E, Maharani K, Gakuru J, Gaharu MN, Mugabi T, Kimuda S, Namombwe S, Te Brake L, Aarnoutse R, Svensson EM, Bangdiwala AS, Namanda S, Bahr NC, Musubire AK, Moosa MYS, Hamers RL, Marais S, Boulware DR, van Crevel R, Ruslami R, HARVEST Trial Team
- DOI
- 10.1056/NEJMoa2502866
- Record issued
- 2026-08-15
- Engine
- 7.39.0
- Exported
- 2026-09-22
Prepared by Alpha1. This document is confidential: it is intended for the recipient it was shared with and must not be redistributed. The live record at alpha1science.com/verify/39302b4b-7321-41bd-b968-c5cd45130a0d is authoritative.
How this rating was calculated
- StatisticsImpossible or misreported statistic ×4−4★
- IntegrityIntegrity concern ×2−1★
- ReportingData & code availability partially met−0.25★
A demonstrable critical failure caps the rating at the minimum, regardless of the deductions above.
- No data or code availability links were detected to verify.
- 01Printed percentage does not match its own countdemonstrable
67.2% does not match the reported count 164/249
“Glasgow Coma Scale <15, no. (%) | 164 (67.2)”
Table 1Find in source - 02Printed percentage does not match its own countdemonstrable
64.8% does not match the reported count 160/250
“Glasgow Coma Scale <15, no. (%) | 160 (64.8)”
Table 1Find in source - 03Printed percentage does not match its own countdemonstrable
45.8% does not match the reported count 144/249
“Definite TBM | 144 (45.8)”
Table 1Find in source - 04Printed percentage does not match its own countdemonstrable
26% does not match the reported count 63/249
“White blood cells <5 cells/mm3 , no. (%) | 63 (26.0)”
Table 1Find in source
This Kaimen Rigor review uses Kaimen Rigor reviewers trained on a curated corpus of high-fidelity and retracted papers, with expert supervision and curation. It can still make mistakes; verify each finding against the source before relying on it.
This is a rigorously designed and reported phase III randomized controlled trial. The main methodological strengths are the robust randomization, blinding, power analysis, and transparent reporting of ethics, funding, and limitations. The primary weakness is the absence of a data availability statement, and there are minor reporting omissions (statistical software, CONSORT guideline) and a few copyedit issues including a potential percentage error in Table 1.
Both reviewers independently scored all eight dimensions and agreed on every status; no divergence required reconciliation. The statistics verification component recomputed only a subset of reported tests (11 total, 7 consistent, 4 inconsistent but unspecified), so the statistical analysis is not fully verified. The copyedit pass flagged three minor issues, including a possible percentage error in Table 1.
Numerical inconsistencies
2 findings · worst criticalValues that contradict each other or are impossible for the stated sample: recomputed p-values and test statistics, GRIM/GRIMMER checks on summary numbers, percentages against their own counts, totals against their parts, and estimates against their own confidence intervals.
- Summary statistic impossible for the stated N (GRIM/GRIMMER)Recomputed
- Internal contradictions in the reported numbersAssessed
Recomputed 7 tests: 7 consistent, 0 inconsistent; 7 via agent-written checks. 4 reported summary statistics mathematically impossible for the stated N (PERCENT).
- PERCENT67.2% does not match the reported count 164/249
“Glasgow Coma Scale <15, no. (%) | 164 (67.2)”
Table 1Find in source - PERCENT64.8% does not match the reported count 160/250
“Glasgow Coma Scale <15, no. (%) | 160 (64.8)”
Table 1Find in source - PERCENT45.8% does not match the reported count 144/249
“Definite TBM | 144 (45.8)”
Table 1Find in source - PERCENT26% does not match the reported count 63/249
“White blood cells <5 cells/mm3 , no. (%) | 63 (26.0)”
Table 1Find in source
- CONSISTENTreported p = .250 · recomputed p = .262Reviewers 1, 2Primary outcome hazard ratio p-value
“Hazard Ratio, 1.17; 95% confidence interval [CI], 0.89 to 1.54; P=0.25”
Taken as given: The hazard ratio is 1.17 with 95% CI 0.89-1.54.; The CI is two-sided at 95%.; The p-value is two-tailed.Method: Recomputed p-value from the hazard ratio and its 95% confidence interval using the normal approximation for the log hazard ratio.How we recomputed it: pCI(1.17, 0.89, 1.54, 1) - CONSISTENTreported p = .030 · recomputed p = .030Reviewer 1Subgroup analysis: antiretroviral therapy hazard ratio p-value
“Hazard Ratio, 2.01; 95% CI, 1.07 to 3.78”
Taken as given: The hazard ratio is 2.01 with 95% CI 1.07-3.78.; The CI is two-sided at 95%.; The p-value is two-tailed.Method: Recomputed p-value from the hazard ratio and its 95% confidence interval using the normal approximation for the log hazard ratio.How we recomputed it: pCI(2.01, 1.07, 3.78, 1) - CONSISTENTreported p = .016 · recomputed p = .016Reviewer 1Subgroup analysis: CSF white cells <5 hazard ratio p-value
“Hazard Ratio, 2.01; 95% CI, 1.14 to 3.54”
Taken as given: The hazard ratio is 2.01 with 95% CI 1.14-3.54.; The CI is two-sided at 95%.; The p-value is two-tailed.Method: Recomputed p-value from the hazard ratio and its 95% confidence interval using the normal approximation for the log hazard ratio.How we recomputed it: pCI(2.01, 1.14, 3.54, 1) - CONSISTENTreported p = .030 · recomputed p = .030Reviewer 2Subgroup hazard ratio p-value for ART
“Hazard Ratio, 2.01; 95% CI, 1.07 to 3.78”
Taken as given: The hazard ratio is 2.01 with 95% CI 1.07-3.78.; The CI is two-sided at 95%.; The p-value is two-tailed.Method: Recomputed p from HR and CI using normal approximation on log scale.How we recomputed it: pCI(2.01, 1.07, 3.78, 1) - CONSISTENTreported p = .020 · recomputed p = .016Reviewer 2Subgroup hazard ratio p-value for CSF WBC <5
“Hazard Ratio, 2.01; 95% CI, 1.14 to 3.54”
Taken as given: The hazard ratio is 2.01 with 95% CI 1.14-3.54.; The CI is two-sided at 95%.; The p-value is two-tailed.Method: Recomputed p from HR and CI using normal approximation on log scale.How we recomputed it: pCI(2.01, 1.14, 3.54, 1) - CONSISTENTreported p = .310 · recomputed p = .329Reviewer 212-month mortality hazard ratio p-value
“Hazard Ratio, 1.14; 95% CI 0.88 to 1.49”
Taken as given: The hazard ratio is 1.14 with 95% CI 0.88-1.49.; The CI is two-sided at 95%.; The p-value is two-tailed.Method: Recomputed p from HR and CI using normal approximation on log scale.How we recomputed it: pCI(1.14, 0.88, 1.49, 1) - CONSISTENTreported p = .180 · recomputed p = .178Reviewer 2Modified Rankin score odds ratio p-value
“Odds Ratio, 0.80; 95% CI, 0.58 to 1.11”
Taken as given: The odds ratio is 0.80 with 95% CI 0.58-1.11.; The CI is two-sided at 95%.; The p-value is two-tailed.Method: Recomputed p from OR and CI using normal approximation on log scale.How we recomputed it: pCI(0.80, 0.58, 1.11, 1)
- lowinternal contradictionAbstract states 428 (86%) had confirmed or probable TBM, but Table 1 shows definite+probable = 144+89=233 (high-dose) and 104+91=195 (standard-dose) = 428 total, which is 85.8% of 499. This is consistent, but the abstract says 'confirmed or probable' while Table 1 includes 'possible' and 'not TBM' categories.
“428 (86%) had confirmed or probable tuberculous meningitis”
AbstractFind in source - lowinternal contradictionTable 1 reports 'Definite TBM' as 144 (45.8%) for the high-dose group, but 144/249 = 57.8%. The percentage appears inconsistent with the numerator and denominator.
“Definite TBM | 144 (45.8) | 104 (41.6)”
Table 1Find in source
Overstated conclusions
1 finding · worst lowConclusions that reach past what the paper's own results support — including a significance claim that no longer holds when the statistic is recomputed, and efficacy resting on an unvalidated surrogate endpoint.
- Conclusions only partially backed by the presented evidenceAssessed
5 major claims checked against the paper's own evidence: 2 only partially supported (evidence backs part of the claim; gaps or caveats remain); the rest adequately supported.
- partialReviewer 1High-dose rifampicin may be harmful, as suggested by higher early mortality and subgroup findings.The paper reports higher early mortality and two subgroups with increased hazard, but the overall result is not statistically significant, so harm is not definitively established.Evidence: Higher early mortality (27.6% vs 19.4% at day 21) and subgroup HRs of 2.01 for those on ART and with CSF WBC <5.
“the potential for a harmful effect cannot be ruled out”
Discussion ¶1Find in source - partialReviewer 2High-dose rifampicin may have a harmful effect, particularly in certain subgroups.The overall result is not significant, but subgroup analyses show increased mortality in patients with CSF WBC <5 and those on ART, though these are exploratory.Evidence: Subgroup analyses show HR 2.01 for both subgroups.
“We observed higher risk for mortality in the high-dose group compared to the standard-dose group among those receiving antiretroviral therapy at the time of presentation (Hazard Ratio, 2.01; 95% CI, 1.07 to 3.78).”
ResultsFind in source - supportedReviewers 1, 2High-dose rifampicin does not improve 6-month survival in adults with tuberculous meningitis.The primary outcome analysis shows no significant difference in 6-month mortality between groups, supporting the claim.Evidence: Primary outcome: HR 1.17, 95% CI 0.89-1.54, p=0.25.
“There is no evidence of beneficial effect from high-dose rifampicin for tuberculous meningitis”
AbstractFind in source - supportedReviewer 1High-dose rifampicin is associated with increased drug-induced liver injury.The paper reports a higher incidence of drug-induced liver injury in the high-dose group, though not statistically significant.Evidence: Drug-induced liver injury occurred in 8.0% vs 4.4% (p=0.09).
“Drug-induced liver injury occurred in 20 of 249 (8.0%) participants in the high-dose group and 11 of 250 (4.4%) in the standard-dose group.”
ResultsFind in source - supportedReviewer 2High-dose rifampicin is not associated with increased overall toxicity.Safety outcomes show no significant differences in serious adverse events, though there were more cases of drug-induced liver injury and hyperbilirubinemia in the high-dose group.Evidence: Safety results: DILI 8.0% vs 4.4%, total bilirubin elevations 9.6% vs 3.6%.
“We did not observe differences in safety.”
ResultsFind in source
Data authenticity concerns
None foundAn adversarial read for patterns associated with data that may not be genuine: results that look too clean, implausibly large effects, duplicated data or images, and methods that do not match the results reported.
Checked — nothing surfaced.
Reporting gaps
1 finding · worst mediumRequired detail the manuscript never states — study design, biological variables, ethics approval and consent, key resources, statistical reporting, data and code availability, and overall transparency.
- Data/code availability incompleteAssessed
The introduction cites multiple prior trials and a meta-analysis, acknowledges mixed evidence, and justifies the need for a well-powered trial. The hypothesis follows logically from the cited evidence. Limitations of prior research (e.g., uncertainty in meta-analysis) are explicitly addressed.
“Four phase II trials have shown a dose-exposure relationship for rifampicin, with higher doses enhancing plasma and CSF drug concentrations.”
“A Vietnamese trial of 817 persons with tuberculous meningitis testing 15 mg per kilogram oral rifampicin showed no benefit.”
“Given mixed evidence, a well powered randomized clinical trial evaluating high-dose rifampicin was deemed necessary.”
“Four phase II trials have shown a dose-exposure relationship for rifampicin, with higher doses enhancing plasma and CSF drug concentrations.”
“Given mixed evidence, a well powered randomized clinical trial evaluating high-dose rifampicin was deemed necessary.”
“albeit the uncertainty in the estimate of the effect was large (relative standard error 86%).”
Randomization used permuted blocks with stratification, blinding was double-blind with identical placebo, and a power analysis was provided. Inclusion/exclusion criteria were pre-specified, and the analysis population (ITT) was defined. Outlier handling is addressed through pre-specified analysis populations and missing-data approaches.
“We randomized participants in a 1:1 ratio to either high-dose or standard-dose oral rifampicin, using permutated blocks with sizes of two or four.”
“Both participants and investigators were blinded to treatment allocation, with identical-appearing intervention and placebo tablets.”
“Assuming 50% survival in the control group, an overall sample size of 500 participants (250 per group) provided 80% power with a two-sided 0.05 significance level to detect a hazard ratio for death of 0.68”
“We randomized participants in a 1:1 ratio to either high-dose or standard-dose oral rifampicin, using permutated blocks with sizes of two or four.”
“Both participants and investigators were blinded to treatment allocation, with identical-appearing intervention and placebo tablets.”
“Assuming 50% survival in the control group, an overall sample size of 500 participants (250 per group) provided 80% power with a two-sided 0.05 significance level to detect a hazard ratio for death of 0.68”
Sex, age, weight, and HIV status are reported in baseline characteristics. Demographics are reported. Sex-justification is not applicable as both sexes were enrolled.
“The median age was 37 years (interquartile range, 28 to 45), and 222 of 499 (44%) were female.”
“Median weight (IQR), kg | 54 (49-60) | 55 (45-62)”
“Overall, 304 of 499 (61%) were persons living with HIV”
“The median age was 37 years (interquartile range, 28 to 45), and 222 of 499 (44%) were female.”
“Overall, 304 of 499 (61%) were persons living with HIV; 125 of 304 (41%) were receiving antiretroviral therapy, and 115 of 226 (51%) had CD4 counts <100 cells/mm 3 .”
The paper states that research ethics committees and regulatory approvals occurred at all sites, and written informed consent was obtained from all participants or surrogates. Regulatory compliance is implied through the mention of approvals.
“Research ethics committees and relevant regulatory approvals occurred at all sites.”
“Written informed consent was obtained from all participants, or surrogates for people lacking capacity to consent.”
“Research ethics committees and relevant regulatory approvals occurred at all sites.”
“Written informed consent was obtained from all participants, or surrogates for people lacking capacity to consent.”
The investigational drug rifampicin is identified with dosing and regimen. The Xpert MTB/Rif Ultra assay is identified with manufacturer. Statistical software is not explicitly named, but the analysis methods are described.
“Participants in the intervention arm received high-dose oral rifampicin of ~35 mg per kilogram daily, administered for 8 weeks.”
“Participants in the intervention arm received high-dose oral rifampicin of ~35 mg per kilogram daily, administered for 8 weeks.”
Tests are named (Cox regression, cumulative logistic regression, Fine-Gray model). Exact p-values are reported (e.g., p=0.25). Effect sizes with confidence intervals are provided. Software is not explicitly identified, but this is a minor omission. Data presentation includes Kaplan-Meier curves and per-group n.
“The treatment effect of high-dose rifampicin vs standard-dose rifampicin was estimated via hazard ratio from an unadjusted Cox regression.”
“Hazard Ratio, 1.17; 95% confidence interval [CI], 0.89 to 1.54; P=0.25”
“Hazard Ratio, 1.17; 95% confidence interval [CI], 0.89 to 1.54”
“The treatment effect of high-dose rifampicin vs standard-dose rifampicin was estimated via hazard ratio from an unadjusted Cox regression.”
“Hazard Ratio, 1.17; 95% confidence interval [CI], 0.89 to 1.54; P=0.25”
“Hazard Ratio, 1.17; 95% confidence interval [CI], 0.89 to 1.54”
The paper does not include a data availability statement. It mentions the protocol is available at NEJM.org, but not the data. No repository deposit or accession numbers are provided. Code sharing is not applicable as no custom code is described.
The trial is registered (ISRCTN15668391). Limitations are discussed, including potential misdiagnosis and limited CSF sampling. Conclusions are proportional, noting no benefit and potential harm. Funding and COI are disclosed.
“Clinical Trial numberISRCTN15668391”
“The major limitation common to all tuberculous meningitis studies is the potential for misdiagnosis.”
“Funded by the Department of Health and Social Care, the Foreign, Commonwealth & Development Office, the United Kingdom Medical Research Council, and Wellcome Trust Medical Research Council Joint Global Health Trials (MR/S004963/1).”
“Clinical Trial numberISRCTN15668391”
“The major limitation common to all tuberculous meningitis studies is the potential for misdiagnosis.”
“Funded by the Department of Health and Social Care, the Foreign, Commonwealth & Development Office, the United Kingdom Medical Research Council, and Wellcome Trust Medical Research Council Joint Global Health Trials (MR/S004963/1).”
Registered (1 ID: ISRCTN). No reporting guideline cited.
Broken references and links
None found · partly checkedReferences checked against Crossref, OpenAlex and Retraction Watch for retractions and resolvability, plus declared data and code links probed for whether they resolve to content matching the paper.
Nothing surfaced — but not everything feeding this category ran (missing: data/code link verification), so read this as a partial clean bill.
Checked 33 references by DOI: 32 verified — 1 no DOI (shown, not verified).
- NO DOIWHO consolidated guidelines on tuberculosis: Module 4: Treatment - Drug-susceptible tuberculosis treatmentNo DOI in the reference — shown for manual review; not independently verifiable (not a fabrication signal).
Copyediting
3 minorWording, consistency and formatting errors that need correcting before submission.
No major wording or formatting errors. 3 minor suggestions below.
3 copyedit issues flagged: mostly typo, consistency, clarity.
- MINORtypoAbstract, Results“interquartile rage”→ interquartile rangeTypo in 'interquartile rage'.
- MINORconsistencyResults, Participants“529 adult patients”→ 529 participantsInconsistent use of 'patients' vs 'participants'.
- MINORclarityTable 1, Diagnostic category“Definite TBM | 144 (45.8)”→ 144 (57.8)Percentage appears incorrect; 144/249 = 57.8%, not 45.8%.
The published work is methodologically robust and the conclusions are appropriately cautious. An informed reader should weigh the missing data availability statement and the minor reporting omissions (statistical software, CONSORT) as transparency gaps, and the Table 1 percentage discrepancy warrants verification. The 4 inconsistent statistics recomputations (unspecified) suggest a cautious interpretation of the numerical results until clarified.
- 1.CRITICALstatisticsCorrect or explain the statistically impossible value: PERCENT: 67.2% does not match the reported count 164/249Demonstrable critical failure — blocks the verdict from passing.
- 2.CRITICALstatisticsCorrect or explain the statistically impossible value: PERCENT: 64.8% does not match the reported count 160/250Demonstrable critical failure — blocks the verdict from passing.
- 3.CRITICALstatisticsCorrect or explain the statistically impossible value: PERCENT: 45.8% does not match the reported count 144/249Demonstrable critical failure — blocks the verdict from passing.
- 4.CRITICALstatisticsCorrect or explain the statistically impossible value: PERCENT: 26% does not match the reported count 63/249Demonstrable critical failure — blocks the verdict from passing.
- 5.HIGHdata codeAdd a data availability statement to the manuscript (e.g., in Methods or a dedicated section) specifying how de-identified participant data can be accessed, including any conditions or a data access committee.The paper currently has no data availability statement, which is a transparency gap for a data-driven clinical trial.
- 6.HIGHstatisticsVerify and correct the percentage reported for 'Definite TBM' in Table 1: 144/249 = 57.8%, not 45.8%.The copyedit and integrity checks both flagged this internal inconsistency, which could mislead readers.
- 7.HIGHreportingExplicitly name the statistical software (e.g., R version, SAS) used for analyses in the Statistical Analysis section.Both reviewers noted the software is not identified, which is a reproducibility gap.
- 8.HIGHreportingAdd a statement about adherence to CONSORT guidelines or provide a CONSORT checklist as supplementary material.The reporting guideline is not referenced, which is expected for a randomized trial.
- 9.MEDIUMcopyeditFix the typo 'interquartile rage' to 'interquartile range' in the Abstract, Results.This is a clear typographical error that should be corrected.
- 10.MEDIUMcopyeditStandardize terminology: use 'participants' consistently instead of alternating with 'patients' in the Results, Participants section.Inconsistent terminology can confuse readers.
- 11.MEDIUMdata codeIf applicable, deposit aggregate or non-identifiable data in a public repository with a DOI or accession number.This would enhance reproducibility and is a common expectation for clinical trials.
- 12.MEDIUMdata codeIf custom analysis code was used, provide it in a public repository (e.g., GitHub) with a version and link.Sharing code improves transparency and reproducibility.
- 13.LOWstatisticsClarify the handling of missing data for secondary outcomes (e.g., imputation methods) in the Statistical Analysis section.This would improve the completeness of the statistical methods description.
The star rating is the report’s one-glance summary. Every paper starts at 5★ and loses stars for the concrete problems the review finds — so a rating is never a vague average, it’s a running total you can read line by line under “How this rating was calculated.”
- Reporting — 8 dimensionseach dimension that fully fails−½★
- each dimension partially met−¼★
- Statistics · Integrity · Claimseach serious problem−1★
- each medium problem−½★
- Citationseach retracted or unverifiable reference−¼★
- Copyeditonly when the manuscript needs a full edit−½★
The rating never drops below 1★, and a demonstrable critical failure (an impossible statistic, a proven ethics violation) caps it at 1★ on its own — so the stars can never look healthy when the verdict is CRITICAL.
The rating draws on a panel of agents. Three independent Kaimen Rigor reviewers grade the eight dimensions below across several independent passes (the shown verdict is their majority vote — steadier than any single run), isolate the paper’s major claims and check its own evidence backs them, and flag integrity concerns. Alongside them, a citation agent resolves every reference against Crossref, OpenAlex, and Retraction Watch; a statistics agent recomputes reported tests; and rule-based checks verify that declared data/code links actually resolve. Full text is required — an abstract-only submission is not analyzed.
Graded against NIH, MDAR, ARRIVE 2.0, CONSORT, EQUATOR, and RRID guidelines. A dimension that doesn’t apply to the study type is skipped, never penalized.
This Kaimen Rigor review is model-assisted and is not a substitute for formal expert review. It complements human evaluation by surfacing potential methodological concerns — verify each finding against the source.