Bedaquiline as Treatment for Disseminated Nontuberculous Mycobacteria Infection in 2 Patients Co-Infected With HIV

Eliza Gil; Nicola Sweeney; Veronica Barrett; Stephen Morris-Jones; Robert F. Miller; Victoria J. Johnston; Michael Brown


Emerging Infectious Diseases. 2021;27(3):944-948. 

In This Article

The Study

Case-patient 1 was a 54-year-old HIV-infected man who had colonic perforation secondary to rectal trauma. He underwent an emergency Hartmann's procedure and showed an uncomplicated immediate recovery. Two months later, he showed development of fevers, breathlessness, and a purulent exudate at the abdominal wound site, which did not improve after receiving antimicrobial drug therapy. Imaging showed pleural effusions and perihepatic collections; mycobacterial liquid culture of effusions, collections, and wound exudate contained M. abscessus, presumed secondary to fecal abdominal cavity contamination. Mycobacterial blood cultures were negative.

At diagnosis of his disseminated NTM infection, HIV viral load was undetectable (CD4 count >900 cells/μL). Empirical treatment was begun and then refined after speciation as M. abscessus (Figure 1, panel A). Susceptibility testing subsequently demonstrated extensive drug resistance (Table 1). MIC estimations for bedaquiline showed in vitro susceptibility (MIC ≤0.0625 mg/L). There was no information on Clinical and Laboratory Standards Institute/European Committee on Antimicrobial Susceptibility Testing for M. abscessus. The MIC breakpoint for this drug with M. tuberculosis was 0.25 mg/L (T. McHugh, University College London, pers. comm., 2020 Jun 29). Compassionate access to bedaquiline was obtained from Janssen-Cilag Treatment was initiated (400 mg/d for 2 wks, followed by 200 mg 3×/wk), as per treatment for tuberculosis. Bedaquiline was well tolerated and treatment continued for a year; there was a brief interruption because of a delay in reapproval.

Figure 1.

Summary of treatment and monitoring of 2 HIV-positive persons who had disseminated Mycobacterium abscessus infections, London, UK. A) Case-patient 1. B) Case-patient 2. The infection in case-patient 1 was secondary to fecal abdominal cavity contamination after rectal perforation. Bars in top section show timing of treatments; red indicates bedaquiline. Drug regimens and treatment responses were measured by using Hb (g/L), Neut (× 109 cells/L), and CRP (mg/L). Values (+ and –) on the bottom of panel B are results for mycobacterial blood cultures. CRP, C-reactive protein; Hb, hemoglobin; Neut, neutrophils.

During his treatment for NTM, the patient showed adverse effects caused by intravenous amikacin (1 g/d) and mild renal impairment. Therefore, this drug was withheld until his renal function recovered. On its reinitiation at the same dose, tinnitus developed prompting permanent withdrawal of amikacin. Because of excellent progress, tigecycline was replaced with linezolid at that time.

18Fluorodeoxyglucose-positron emission tomography/computed tomography imaging was used to monitor disease response (Figure 2). Although pulmonary and hepatic lesions emerged intermittently, they were consistently culture negative and are believed to represent immune-mediated lesions. His most recent scan 36 months after treatment began showed ongoing, but greatly reduced, 18fluorodeoxyglucose avidity in all areas. Therefore, he continues maintenance therapy with clofazimine and azithromycin, despite evidence of clarithromycin resistance in vitro. This combination (clofazimine and azithromycin) has been well tolerated, and the condition of the patient continues to be favorable.

Figure 2.

Serial 18fluorodeoxyglucose-positron emission tomography/computed tomography imaging quantification of disease burden for an HIV-positive person (case-patient 1) given treatment for disseminated Mycobacterium abscessus infection, London, UK. Images demonstrate marked reduction in fluorodeoxyglucose avidity over time.

Case-patient 2 was a 30-year-old man who had pyrexia, pancytopenia, and lymphadenopathy. Advanced HIV infection had been diagnosed 1 month earlier (CD4 count 10 cells/μL, viral load >1 million copies/mL). He started antiretroviral therapy 10 days before he came to the hospital. Culture of lymph node, peripheral blood, and sputum all yielded M. avium. Treatment with azithromycin, rifabutin, and ciprofloxacin was initiated. Ethambutol was excluded because of color blindness.

The patient initially transferred his care to another hospital but returned 6 months later because of abdominal pain. Mycobacterial blood cultures had remained persistently positive throughout the intervening period, and at his return cultures of blood, bone marrow, and lymph node were positive for M. avium. We performed sensitivity testing (Table 2). Few drugs have MIC values for M. avium, and the correlation between MIC and clinical outcomes for drugs other than clarithromycin is unclear.[5] The isolate demonstrated new clarithromycin resistance, postulated to have emerged because of persistence through treatment with a macrolide-containing regimen without ethambutol.[6] His treatment was intensified by addition of amikacin and, after discussion with ophthalmologists, ethambutol. Because he continued to have fevers, meropenem, which covered the possibility of bacterial sepsis, and clofazimine were initiated, leading to symptomatic improvement. Because his symptoms worsened again on brief cessation of meropenem, it was continued.

Treatment for this patient required multiple modifications because of adverse drug effects (Figure 1, panel B). Amikacin was stopped because of renal toxicity. Given the extensive in vitro resistance, best practice recommendation to add 2 drugs synchronously, lack of access to bedaquiline at this time, and evidence for a possible benefit of tigecycline in combination with a macrolide,[7] tigecycline and linezolid were initiated in its place, causing nausea and anemia, respectively. He also had arthralgia secondary to moxifloxacin and QTc prolongation caused by azithromycin, which required their cessation.

Shortly after linezolid and azithromycin were discontinued his fevers and neutropenia returned. Mycobacterial blood cultures again showed M. avium despite prolonged treatment and immune reconstitution. Although bedaquiline sensitivity of this isolate was not determined in vitro, compassionate access to bedaquiline was obtained from Janssen-Cilag, and treatment was initiated at month 13 (dosing as reported for case-patient 1), along with tedizolid. Rifabutin was discontinued because of concerns over its effect on bedaquiline pharmacokinetics.

After bedaquiline and tedizolid were initiated, his fevers resolved, and he made a steady recovery and had no side effects. Given his high risk for relapse, he received bedaquiline for 18 months on the advice of the British Thoracic Society panel, ensuring a year of effective therapy since his last positive blood culture. He has successfully immune reconstituted, continues to receive only antiretroviral therapy, and remains healthy.