Tuberculous meningitis (TBM) is the main form of tuberculosis that affects the central nervous system and is associated with high rates of death and disability. Most international guidelines recommend longer antituberculous treatment (ATT) regimens for TBM than for pulmonary tuberculosis disease to prevent relapse. However, longer regimens are associated with poor adherence, which could contribute to increased relapse, development of drug resistance, and increased costs to patients and healthcare systems.
To compare the effects of short-course (six months) regimens versus prolonged-course regimens for people with tuberculous meningitis (TBM).
We searched the following databases up to 31 March 2016: the Cochrane Infectious Diseases Group Specialized Register; the Cochrane Central Register of Controlled Trials (CENTRAL), published in the Cochrane Library; MEDLINE; EMBASE; LILACS; INDMED; and the South Asian Database of Controlled Clinical Trials. We searched the World Health Organization International Clinical Trials Registry Platform (WHO ICTRP) and ClinicalTrials.gov for ongoing trials. We also checked article reference lists and contacted researchers in the field.
We included randomized controlled trials (RCTs) and prospective cohort studies of adults and children with TBM treated with antituberculous regimens that included rifampicin for six months or longer than six months. The primary outcome was relapse, and included studies required a minimum of six months follow-up after completion of treatment.
Data collection and analysis
Two review authors (SJ and HR) independently assessed the literature search results for eligibility, and performed data extraction and ‘Risk of bias’ assessments of the included studies. We contacted study authors for additional information when necessary. Most data came from single arm cohort studies without a direct comparison so we pooled the findings for each group of cohorts and presented them separately using a complete-case analysis. We assessed the quality of the evidence narratively, as using the Grading of Recommendations Assessment, Development and Evaluation (GRADE) approach was inappropriate with no direct comparisons between short- and prolonged-course regimens.
Four RCTs and 12 prospective cohort studies met our inclusion criteria, and included a total of 1881 participants with TBM. None of the included RCTs directly compared six months versus longer regimens, so we analysed all data as individual cohorts to obtain relapse rates in each set of cohorts.
We included seven cohorts of participants treated for six months, with a total of 458 participants. Three studies were conducted in Thailand, two in South Africa, and one each in Ecuador and Papua New Guinea between the 1980s and 2009. We included 12 cohorts of participants treated for longer than six months (ranging from eight to 16 months), with a total of 1423 participants. Four studies were conducted in India, three in Thailand and one each in China, South Africa, Romania, Turkey and Vietnam, between the late 1970s and 2011.
The proportion of participants classified as having stage III disease (severe) was higher in the cohorts treated for six months (33.2% versus 16.9%), but the proportion with known concurrent HIV was higher in the cohorts treated for longer (0/458 versus 122/1423). Although there were variations in the treatment regimens, most cohorts received isoniazid, rifampicin, and pyrazinamide during the intensive phase.
Investigators achieved follow-up beyond 18 months after completing treatment in three out of the seven cohorts treated for six months, and five out of the 12 cohorts treated for eight to 16 months. All studies had potential sources of bias in their estimation of the relapse rate, and comparisons between the cohorts could be confounded.
Relapse was an uncommon event across both groups of cohorts (3/369 (0.8%) with six months treatment versus 7/915 (0.8%) with longer), with only one death attributed to relapse in each group.
Overall, the proportion of participants who died was higher in the cohorts treated for longer than six months (447/1423 (31.4%) versus 58/458 (12.7%)). However, most deaths occurred during the first six months in both treatment cohorts, which suggested that the difference in death rate was not directly related to duration of ATT but was due to confounding. Clinical cure was higher in the group of cohorts treated for six months (408/458 (89.1%) versus longer than six months (984/1336 (73.7%)), consistent with the observations for deaths.
Few participants defaulted from treatment with six months treatment (4/370 (1.1%)) versus longer treatment (8/355 (2.3%)), and adherence was not well reported.
In all cohorts most deaths occurred in the first six months; and relapse was uncommon in all participants irrespective of the regimen. Further inferences are probably inappropriate given this is observational data and confounding is likely. These data are almost all from participants who are HIV-negative, and thus the inferences will not apply to the efficacy and safety of the six months regimens in HIV-positive people. Well-designed RCTs, or large prospective cohort studies, comparing six months with longer treatment regimens with long follow-up periods established at initiation of ATT are needed to resolve the uncertainty regarding the safety and efficacy of six months regimens for TBM.
Plain language summary
Six months therapy for patients with tuberculous meningitis
What is tuberculous meningitis and why is the duration of treatment important?
Tuberculous meningitis (TBM) is a severe form of tuberculosis, which affects the membranes that cover the brain and spine. It is associated with high rates of death and disability. While there are standardized international recommendations for treating people with pulmonary tuberculosis (tuberculosis of the lungs) for six months with antituberculous therapy, there is a wide range of differing recommendations and practices for treating people with TBM worldwide. Some specialists recommend nine months, 12 months, or even longer treatment for TBM in order to prevent relapse of the disease. Longer regimens have potential disadvantages: they are associated with poor adherence to treatment, which could contribute to increased relapse and development of drug resistance; and increased costs to patients and healthcare systems.
What the evidence shows
This Cochrane review assessed the effects of six months regimens for treating people with TBM, compared with longer regimens. Cochrane researchers examined the available evidence up to 31 March 2016. They did not find any trial that directly compared people with TBM treated for six months with people with TBM treated for longer. They included seven studies with 458 participants that evaluated six months of treatment, and 12 studies with 1423 participants that evaluated longer treatment. Although the treatment regimens in the included studies varied, most participants received standard first-line antituberculous drugs, and were followed up for more than a year after the end of treatment. The studies included adults and children with TBM, but few participants were HIV-positive.
Relapse was an uncommon event across both groups of studies, with only one death attributed to relapse in each group. Most deaths occurred during the first six months of treatment in both groups of studies, which showed that treatment duration did not have a direct impact on the risk of death in these studies. There was a higher death rate in participants treated for longer than six months, and this probably reflects the differences between the participants in the two groups of studies. Few participants defaulted from treatment, and adherence was not clearly documented.
They found no evidence of high relapse rates in people treated for six months, and relapse was uncommon in all patients irrespective of regimen. There may be differences between the participants treated for six months and longer than six months that could have led to bias (confounding factors), so further research would help determine if shorter regimens are safe. Most of the data were in patients without HIV, and so these inferences do not apply to patients who are HIV-positive.