Abstract
Linezolid is an effective treatment option for MDR/XDR-TB, including among HIV-co-infected patients on ART http://ow.ly/JoPzO
To the Editor:
Two recent systematic reviews found that treatment outcomes with regimens containing linezolid for complicated cases of multidrug-resistant (MDR) tuberculosis (TB) are equal to or better than those reported for uncomplicated MDR-TB [1, 2] and better than those reported among patients treated for extensively drug-resistant (XDR)-TB [3–6]. Existing data on treatment outcomes with linezolid for MDR-TB are predominantly among HIV-uninfected patients; <10% of patients were HIV co-infected in the reviews [1, 2]. There is concern over the safety of using linezolid within MDR-TB regimens for HIV-infected patients due to underlying HIV-related neuropathy and bone marrow dysfunction as well as a potentially higher incidence of adverse events in patients on antiretroviral therapy (ART), notably anaemia with zidovudine and peripheral neuropathy with stavudine [7]. Given the high rate of HIV/TB co-infection in many settings, we report here our clinical experience with the use of linezolid for treatment of XDR-TB and complicated MDR-TB among HIV-infected and -uninfected patients in community-based programmes supported by Médecins Sans Frontières (MSF) in two high-burden settings in Khayelitsha, South Africa [8], and Mumbai, India [7]. We carried out a retrospective analysis of routinely collected data. Programmes at both sites were approved by either MSF or local university ethics review boards. Due to restricted access to linezolid in South Africa as well as limited capacity for operational research in both programmes, the numbers are small, and data were collected and analysed over a limited time period only.
At time of analysis in August 2013, 34 patients (17 from each site); one with MDR-TB, 16 with pre-XDR-TB (defined as MDR plus resistance to either a fluoroquinolone or an injectable agent) and 17 with XDR-TB) had received individualised regimens with a median of eight drugs (range 7–10), including linezolid, for ≥1 month. 18 (53%) patients started treatment on initial detection of second-line drug resistance; the remainder only after second-line TB treatment appeared to be failing due to continued culture positivity or reconversion from negative to positive. The median second-line treatment duration prior to linezolid start in the latter group was 12 months (range 6–49 months). 18 (53%) patients in the cohort were female; median age was 29 years (range 12–46 years) and median weight 51 kg (range 26–94 kg). While a starting oral dose of 600 mg linezolid once daily was preferred, two patients with low body mass index (≤14 kg·m−2) were started on a lower dose of 300 mg once daily. 19 (56%) patients were HIV infected (median CD4 count 290 cells·mm−3, range 67–490 cells mm−3) and all were on ART (two with stavudine and six with zidovudine-containing regimens) at the time of linezolid start.
The median duration of linezolid treatment among all 34 patients was 8 months (range 1–24 months) at the time of censoring, and at this point, 29 (85%) patients were still alive and on treatment. No patients were lost from treatment (default) but two (6%) died during treatment; both had been started on linezolid due to failure of second-line therapy. One (HIV uninfected) died within 2 months of starting linezolid and without culture conversion, while the other (HIV infected) died of an unknown cause after 17 months of treatment with linezolid, 14 months after culture conversion. Treatment failed and was withdrawn for a further three (9%) patients, after a median of 8 months (range 6–10 months), all of whom subsequently died due to clinical progression of TB disease (only one was HIV infected).
Only 28 patients had positive Mycobacterium tuberculosis cultures at linezolid start and, among these, culture conversion was 50% by 3 months of treatment (fig. 1a). There was no difference in culture conversion over time related to HIV infection (hazard ratio (HR) 1.12, 95% CI 0.39–3.22; p=0.83), resistance pattern (MDR/pre-XDR versus XDR) (HR 1.27, 95% CI 0.40–4.00; p=0.69) or weight (<50 versus >50 kg) (HR 1.64, 95% CI 1.10–4.76; p=0.36). Those who started linezolid on initial detection of second-line drug resistance had more rapid culture conversion than those treated after failure of previous treatment (HR 3.33, 95% CI 1.10–10.20; p=0.03), suggesting that individualised regimens containing linezolid should be offered early to improve chance of treatment success rather than reserved as a salvage option later in treatment.
In our cohort of 34 patients, there were 17 episodes of any adverse event attributed to linezolid, reported among 14 (41%) patients: six patients with peripheral neuropathy only, five with anaemia only, two with anaemia and peripheral neuropathy, and one with anaemia and optic neuritis. This incidence was not as high as that previously reported from a cohort of 18 MDR-TB patients treated with linezolid utilising individualised regimens in India [9]. Adverse events were reported pragmatically and it is likely that mild events were not always reported, particularly peripheral neuropathy, which is common among HIV-infected people in these settings [10]. As data were collected retrospectively, this may be an underestimation of the true incidence in our cohort. Encouragingly, six patients showed improvement in adverse event symptoms following dose reduction from 600 mg once daily to 300 mg once daily, whereas seven patients were able to continue linezolid at the same dose after the adverse event was reported. Linezolid was withdrawn in one other patient with severe peripheral neuropathy, who also had prolonged stavudine exposure. Optic neuropathy is reported to be relatively rare but can be irreversible [11]; this event occurred in only one patient after 6 months on linezolid. The patient insisted on continuing linezolid despite persistent visual disturbance and extensive counselling and advice to stop, and eventually died of an unknown cause at 17 months. Neither of the two patients who started linezolid at 300 mg once daily experienced any adverse events. It should be noted that while anaemia and neuropathies were occasionally reported to be severe, their occurrence should be viewed in the context of the serious, debilitating and irreversible adverse events associated with other second-line TB medications, such as renal toxicity, deafness and psychosis [12, 13]. Therapeutic drug monitoring has a potential role in the management of complicated MDR-TB cases with limited treatment options. Measurement of serum concentrations of second-line TB drugs may allow individual dosages to be modified in patients at risk of multiple drug–drug interactions, and may minimise adverse events while maintaining optimal exposure to effective treatment [14, 15].
We found that while the cumulative incidence of adverse events was high in the whole cohort, HIV-infected patients appeared to have a higher risk of any adverse event associated with linezolid, particularly beyond 3 months of exposure, although this did not reach significance (HR 2.92, 95% CI 0.9–9.4; p=0.073) (fig. 1b). This pattern was similar between HIV-infected and uninfected patients for peripheral neuropathy (HR 2.99, 95% CI 0.60–14.87; p=0.18) and anaemia (HR 3.39, 95% CI 0.68–16.93; p=0.14). There was no increased risk of peripheral neuropathy among patients weighing <50 kg (HR 1.09, 95% CI 1.27–4.39; p=0.90) but there appeared to be a higher risk of anaemia with weight <50 kg (HR 3.56, 95% CI 0.81–15.64; p=0.092), particularly among those who were HIV-infected (HR 5.43, 95% CI 1.25–23.55; p=0.024).
In summary, our data indicate that the rate of culture conversion among pre-XDR and XDR-TB patients treated with linezolid is better than previously reported among XDR cohorts [3–6] and that HIV-infected patients on ART were able to tolerate prolonged linezolid exposure, with dose reduction in some cases. HIV infection was not associated with poorer treatment response, despite low CD4 counts; however, all HIV-infected patients in this cohort were on either first- or second-line ART regimens and most had suppressed viral loads, potentially contributing to encouraging interim outcomes. Routine monitoring for specific adverse events is recommended for all patients receiving linezolid and, although not uncommon, these events may be managed appropriately among both HIV-infected and -uninfected patients, even on an ambulatory basis.
Footnotes
Conflict of interest: None declared.
- Received July 21, 2014.
- Accepted February 3, 2015.
- Copyright ©ERS 2015