Improving the Diagnosis and Treatment of HIV-associated Tuberculous Meningitis

Abstract

HIV-1 has caused a marked shift in the epidemiology of meningitis in sub-Saharan Africa and Mycobacterium tuberculosis is now a leading cause of meningitis in adults. Diagnosis of tuberculous meningitis (TBM) is notoriously difficult, compounded by an atypical cerebrospinal (CSF) picture in HIV co-infection. The WHO-recommended standard quadruple TB therapy may not be the optimal regimen in TBM as rifampicin does not readily penetrate into the CSF at the current dose. Better diagnostics and treatments are needed to tackle the disconcertingly high TBM mortality of at least 50% in people living with HIV (PLHIV). The goals of this PhD thesis were to first identify improved approaches to the diagnosis of TBM; and second, to investigate the safety and pharmacokinetics of a novel approach to treatment with high-dose rifampicin through a phase II open-label randomised controlled trial in predominantly HIV-positive adults. To begin, using retrospective data from two Ugandan referral hospitals, I showed that the introduction of the Cepheid Xpert MTB/Rif assay had markedly increased the microbiological confirmation of TBM. However, around half of TBM cases were treated on the basis of clinical suspicion alone, despite negative CSF TB tests. So, better diagnostics were still needed. I went on to examine the novel Xpert Ultra assay in a prospective diagnostic accuracy study, and found it to be the best CSF test for TBM, with a sensitivity of 77% and a negative predictive value of 93% for probable/ definite TBM, significantly better than Xpert MTB/Rif. A high CSF lactate and low CSF glucose were good discriminators of definite TBM. Given that CSF collection can be delayed or contraindicated, a non-CSF dependent test would be valuable, so i explored the potential of urine TB testing. Urine TBlipoarabinomannan (Alere LAM) and/or urine Xpert Ultra were positive in 50% of probable/definite TBM cases, suggesting extra-meningeal disseminated TB disease is highly prevalent in adults with HIV/TBM. In this exploratory study urine Xpert Ultra had prognostic value in definite TBM. The phase II randomised clinical trial demonstrated that with WHO-recommended TB treatment (containing rifampicin 10 mg/kg/day), CSF rifampicin was undetectable (<0.25 mg/L) in around half of participants, and only 11% achieved a CSF rifampicin concentration above the minimal inhibitory concentration (MIC) of the predominant M. tuberculosis strain in Uganda. Intravenous rifampicin 20 mg/kg/day yielded a ~6-fold increase in serum and CSF exposures; and oral rifampicin 35 mg/kg/day yielded an ~8-fold increase in serum and CSF exposures compared to standard of care. All participants in the investigational arms had detectable CSF rifampicin, and almost all had CSF levels above the rifampicin MIC. Importantly, there was no excess toxicity in the high-dose rifampicin arms. Overall, the most common adverse event was elevated alanine transaminase, attributed to ii drug-induced liver injury, which occurred in 7 (12%) of participants overall, 4 of which were in the control arm. Together, this research has shown that, in Uganda, TBM is a usually a manifestation of disseminated TB disease resultant upon advanced HIV-related immune suppression, and carries a high case-fatality. Xpert Ultra is a step forward but cannot be used as a ‘rule-out’ test. A diagnostic test that combines pathogen-detection with a host biomarker may be the most sensitive approach for TBM. The WHO recommended TB therapy for TBM is sub-optimal and high-dose rifampicin is a promising and safe intervention that needs to be further investigated in a phase III trial.