Sterile tuberculous granuloma in a patient with XDR-TB treated with bedaquiline, pretomanid and linezolid

Drug-resistant tuberculosis (DR-TB) continues to pose a threat to the global eradication of TB. Regimens for extensively drug-resistant (XDR) TB are lengthy and poorly tolerated, often with unsuccessful outcomes. The TB Alliance Nix-TB trial investigated the safety and efficacy of a 26-week regimen of bedaquiline, pretomanid and linezolid (BPaL) in participants with XDR-TB, multidrug-resistant (MDR) TB treatment failure or intolerance. In this trial 9 out of 10 participants were cured. We describe a trial participant with XDR-TB who presented with new-onset seizures soon after BPaL treatment completion. Imaging showed a right temporal ring-enhancing lesion, and a sterile tuberculous granuloma was confirmed after a diagnostic, excisional biopsy. Learning points include management of a participant with a tuberculoma after BPaL completion, efficacy of new medications for central nervous system (CNS) TB and a review of their CNS penetration. This is the first case of pretomanid use in CNS TB.


BACKGROUND
In 2019, rifampicin-resistant tuberculosis (TB) accounted for an estimated half a million new cases worldwide. Of these, 6.2% had extensively drug-resistant (XDR) TB defined as resistance to a rifamycin, isoniazid, a fluoroquinolone and a second generation injectable. 1 Central nervous system (CNS) TB, a severe manifestation of extrapulmonary TB, develops in 1%-2% of TB cases. The incidence of CNS DR-TB is largely unknown due to diagnostic challenges, but when suspected the treatment is prolonged. Due to the relative impermeability of the CNS, regimens targeting pulmonary TB are inadequate for CNS TB. This leaves the brain vulnerable to injury with ensuing high morbidity and mortality. [2][3][4] We report a case of CNS DR-TB in a woman who presented with new-onset seizures after completing treatment for pulmonary XDR-TB on the Nix-TB clinical trial. Nix-TB was an open-label trial to assess the safety and efficacy of a novel 26-week regimen of bedaquiline (B), pretomanid (Pa) and high dose linezolid (L) (BPaL) in participants with pulmonary TB with either XDR-TB or treatment intolerant/non-responsive multidrug-resistant (MDR) TB. 5 This is the first patient with CNS TB treated with pretomanid. While it cannot be conclusively stated, this case makes a strong argument that this shortened, all oral regimen may have cured this participant's CNS DR-TB.

CASE PRESENTATION
A young woman was enrolled onto the Nix-TB trial with pulmonary XDR-TB (figure 1). She had a history of successfully treated drug susceptible pulmonary TB, HIV-1 infection, chronic hepatitis B and kanamycin-induced ototoxicity. She received 3 months of a kanamycin-based DR-TB regimen prior to receiving evidence of XDR-TB. She had no personal nor family history of seizures and an otherwise normal neurological examination. She had no history of childhood seizures, previous CNS infections, head injuries, stroke nor other vascular diseases.
She was diagnosed with HIV infection (CD4: 68 cells/µL) and initiated tenofovir/lamivudine/ efavirenz combination therapy 4 months prior to trial enrolment. Her regimen was changed to abacavir/lamivudine/nevirapine when initiating DR-TB treatment but did not achieve viral suppression, and nevirapine was substituted with lopinavir/ ritonavir. At enrolment (CD4: 101 cells/µL), she was not yet virally suppressed. Her sputum culture converted to negative for Mycobacterium tuberculosis at week 4. She completed 26 weeks of the BPaL regimen and had sustained culture negativity until end of trial, 30 months after enrolment. The treatment regimen was modestly tolerated. Linezolid was reduced from 600 mg two times per day to once daily at week 4 due to new-onset peripheral neuropathy. She complained of mild intermittent frontal headaches associated with photophobia from weeks 3 to 6 and was referred to an ophthalmologist who ruled out optic neuritis, a known adverse effect of linezolid, and diagnosed her with migraines based on history and normal ophthalmological examination. She did not present with neck stiffness nor any other focal neurological signs at this time, and no further investigations were done. The participant had a brief treatment interruption at week 9 for asymptomatic pancreatitis which resolved over 2 weeks. She caught up all missed doses, and on completing treatment, her CD4 count had recovered to 225 cells/µL.
The participant experienced a new-onset seizure 1 day after completing BPaL but did not report the event nor seek medical care. She was hospitalised at a regional hospital after a second seizure 3 months later. The participant did not report headaches, photophobia, nausea nor vomiting prior to either seizure. A family member reported she collapsed and became unresponsive; she was salivating but had no obvious abnormal limb movements. On admission, her general and neurological examinations were unremarkable. Blood tests were within normal limits except for a raised creatinine Case report (124 µmol/L; range 49-90 µmol/L) which normalised over a week, and a lumbar puncture (LP) was done. She was discharged on phenytoin and scheduled for an outpatient computed tomography (CT) brain 2 weeks later.
The contrast CT brain showed a rim-enhancing lesion in the right temporal lobe with otherwise normal anatomy. The participant defaulted phenytoin after the CT scan and experienced a third seizure, for which she delayed seeking care. The research team referred the participant for a neurologist consultation. Phenytoin was restarted and a contrast MRI of her brain was done a month after the initial CT brain. The phenytoin was later changed to sodium valproate by the research team after neurologist consultation, due to subtherapeutic levels.

INVESTIGATIONS
At initial hospitalisation an LP was done with normal findings. Blood tests for toxoplasmosis IgG (0.2 IU/mL; range 28-110 IU/ mL), toxoplasmosis IgG and cysticercosis IgG which were negative.
The magnetic resonance imaging (MRI) of the brain (figure 2) confirmed a single rim-enhancing lesion in the right temporal lobe measuring 11×15×11 mm with mild perilesional vasogenic oedema. Restricted diffusion of the periphery of the lesion in the right temporal lobe was noted. The lesion was unchanged from the initial contrast CT brain.

DIFFERENTIAL DIAGNOSIS
At her initial admission she was investigated for meningitis, but the LP results were normal. Immune reconstitution inflammatory syndrome (IRIS) was possible. Undiagnosed subclinical CNS TB that had been inadequately treated was a concern.
Based on the initial CT brain and MRI, the differential diagnoses of rim-enhancing lesions in the brain of a patient in South Africa with HIV infection and recurrent TB were tuberculous granuloma (tuberculoma), toxoplasmosis, neurocysticercosis and lymphoma. Neurocysticercosis is considered to be the most common cause of new-onset epilepsy, especially in resourcelimited settings. 6

TREATMENT
The participant was referred to a neurosurgeon who determined the lesion was the likely cause of the seizures. The participant was admitted, and a second MRI brain was done with no change as compared with the MRI 6 weeks prior. An excision biopsy was done the following day for diagnostic and therapeutic purposes, 3 months after her initial admission for seizure workup and 6 months after completion of the BPaL regimen.

OUTCOME AND FOLLOW-UP
The surgery was uneventful. The tissue samples were sent for histology and a panel of TB tests. Histologically, the specimen was granulomatous with central necrosis, a tuberculous granuloma was confirmed. The GeneXpert was positive for M. tuberculosis complex with rifampicin resistance. The MTBDRplus line probe assay confirmed resistance to rifampicin and isoniazid. The MTBDRsl line probe assay was indeterminate for fluoroquinolone and aminoglycoside resistance.
Culture of the sample was negative for M. tuberculosis after 6 weeks. The granuloma was considered 'sterile'. Pharmacokinetic analysis of the tissue samples was not done.
The participant was discharged 4 days postoperatively. She complained of intermittent headaches in the right temporal region. A repeat MRI brain was conducted 6 months after the biopsy and showed no new lesions. Thereafter, the sodium valproate was tapered off over 3 months. Headaches resolved 10 months after the operation. She remained seizure free with

Case report
durable XDR-TB cure until her final trial follow-up visit 18 months after surgery.

DISCUSSION
While it cannot be conclusively stated, this shortened, all oral regimen may have sterilised a drug-resistant tuberculoma.
The emergence of the tuberculoma is likely a paradoxical CNS TB-IRIS phenomenon. 2 3 The CD4 T-lymphocyte is one of the inflammatory cells integral to the formation of granulomas. 4 A granulomatous reaction to the participant's CNS TB is compatible with CD4 T-cell recovery from a low nadir count as a result of her HIV treatment, the sequelae of which only surfaced post-treatment.
The first novel TB drugs approved in decades, bedaquiline (diarylquinoline), delamanid and pretomanid (nitroimidazoles), have been groundbreaking for DR-TB treatment; however, little is known about their role in CNS TB. CNS DR-TB, especially in the face of HIV, has an almost 100% mortality rate. [7][8][9] In patients with tuberculoma, those with HIV are more likely to experience seizures, 10 which predict a poor prognosis. 11 Current treatment guidelines for CNS TB are derived from pulmonary TB regimens. Increased doses, treatment extension and, for CNS DR-TB, the use of CNS penetrating drugs are recommended. 12 Evaluation of a drug's potential use in CNS TB should consider its molecular, preclinical and clinical evidence. The CNS multiparameter optimisation (CNS MPO) score was designed and validated to identify drugs with molecular characteristics that are likely to cross the blood-brain or blood-cerebrospinal fluid (CSF) barriers and achieve good CNS penetration. A score of 4 or more out of 6 is considered desirable. 13 Linezolid is a small molecule, has relatively low protein binding (31%), high oral bioavailability and an excellent CNS MPO score of 5.8/6, making it an ideal agent for CNS infections where it has proven successful outcomes. When dosed intravenously, linezolid penetrates the CNS early with CSF concentrations equalling or exceeding those measured in plasma. 14 The WHO recommends that treatment for CNS DR-TB is guided by drug sensitivity testing where possible. 12 Linezolid penetrates well and should be considered for inclusion in a multidrug regimen.
Pretomanid is highly protein bound (86.4%) with an excellent calculated CNS MPO score of 5.8/6. It achieves excellent brain concentrations in a number of preclinical studies, attaining concentrations similar to those seen in the lungs. 14 15 In a study designed to analyse concentrations in the rat brain over 8 hours, pretomanid had early, widespread distribution throughout the brain. 15 Delamanid, another nitroimidazole, is very highly protein bound (>99%) with a low calculated CNS MPO score of 2.6/6. 16 Despite this, delamanid is measurable in CSF and has been reported to improve clinical outcomes as part of a strengthened multidrug regimen in patients failing CNS DR-TB regimens. 17 There are no published clinical studies on pretomanid in the CNS. Here, we present the first case report of a patient with CNS TB treated successfully with a regimen including pretomanid. Early clinical data is promising, and pretomanid may have a key role to play in the treatment of CNS TB.
The contribution of bedaquiline to the sterilisation of the tuberculoma is unclear. Bedaquiline has a low calculated CNS MPO score of 1.3/6. It achieved therapeutic brain concentrations in preclinical studies 18 19 but was undetectable in CSF at all timepoints in a patient with CNS DR-TB. 20 Bedaquiline is very highly protein bound (>99.9%) and, therefore, challenging to measure in a low protein environment such as the CSF. Only unbound drug is pharmacologically active and, therefore, low concentrations in CSF may in fact be sufficient to achieve antimicrobial activity. More research is needed for bedaquiline which forms the cornerstone of current DR-TB regimens.

Patient's perspective
I joined (the) study because I saw it as a chance for me to get shorter treatment with fewer side effects. I had seen how sick the other patients in the ward became after starting TB treatment. I had received injections for TB which had affected my hearing, and I was excited when I was told that the study will not have any injections. I was very happy when I completed the treatment. When I started having fits after completing treatment, I was scared. I thought that I was never going to have a normal life again. Blood and scans were performed to find out the cause of the fitting. I was counselled that I needed an operation to remove the mass. I was scared and my family, especially my children, were certain that I was not going to make it.
I felt overwhelmed. After more counselling, I finally accepted the operation but still felt that I will not survive it. I was shocked when I woke up after. and I was later discharged and continued to have follow-ups. The results of the mass was TB but was told that the TB was not active and that I will not need more TB treatment. I became hopeful that maybe I was going to make it and have a normal life.
It has been a few years now after the operation and I am healthy, found work and living a happy life with my family. I will always be grateful that I survived such a difficult time and lived to tell my story to other people who might be going through the same.

Learning points
► A new 6-month, all oral, three drug regimen led to the sterilisation of a drug-resistant tuberculoma. ► There is a need to investigate new, shortened antituberculous regimens targeting drugs with good central nervous system (CNS) penetration. ► Assessing the value of novel tuberculosis (TB) drugs early in development for CNS-TB must be considered to improve the outcomes of this devastating disease.
Contributors PH made substantial contributions to the conception of the work, the acquisition and interpretation of data for the work, drafted, revised, approved the work, and agrees to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. CU made substantial contributions to the conception and drafting of the work, revised and approved the work, and agrees to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. NM made substantial contributions to the acquisition and interpretation of data for the work, revised and approved the work, and agrees to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. MO made substantial contributions to the acquisition of data for the work, revised and approved the work, and agrees to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.
Funding This study was funded by Bill and Melinda Gates Foundation (OPP1129600).

Competing interests
We have read and understood BMJ policy on declaration of interests and declare the following interests: PH: trial sponsor paid for article