This policy guideline anticipates the most common situations under which these recommendations will be applied, in order to provide the end-user with practical options when deciding upon clinical management, such as what to do while awaiting DST results, making best use of FDCs available to the programme, or principles to prevent the acquisition of additional resistance. It also addresses considerations for implementation, monitoring and adaptation in different subgroups (e.g. children; HIV-infected individuals) and other situations13.
The production of these guidelines has highlighted a number of gaps in knowledge about the treatment of Hr-TB, and further studies and operational research are strongly recommended.
Objectives
The objective of this guideline is to provide recommendations on the composition and duration of suitable treatment regimens for the treatment of Hr-TB, based on a review of the best available evidence. These recommendations complement existing WHO guidelines for treatment of DR-TB cases.
Target audience
These guidelines are primarily targeted at policy-makers in ministries of health or managers of national TB programmes (NTPs) who formulate country-specific TB treatment guidelines or who are involved in the planning of TB treatment programmes. In addition, health professionals, including doctors, nurses and educators working in governmental and non-governmental organizations, as well as technical agencies involved in treating patients and organizing treatment services are expected to use these guidelines.
Background and rationale
TB remains a threat to global public health and is the leading cause of death by a single infectious agent globally (1). In 2016, an estimated 10.4 million people developed TB and 1.7 million died from the disease. In the same year an estimated 600 000 TB patients developed rifampicin-resistant (RR-TB) or multidrug-resistant TB (MDR-TB, resistance to rifampicin and isoniazid); about 240 000 patients with MDR/RR-TB were estimated to have died. Patients with MDR/RR-TB roughly account for 4.1% of all new and 19% of retreatment TB cases globally, although wide regional and country differences occur. About 8% of TB cases worldwide are estimated to have Hr-TB, ranging from 5 to 11% between the WHO regions (1). In a recent systematic review, the comparison of treatment outcomes between Hr-TB cases and patients with drug-susceptible TB receiving the WHO standard regimen for new patients, suggested that patients with confirmed isoniazid resistance had worse outcomes – i.e., higher treatment failure (11% vs 1%); relapse (10% vs 5%); as well as higher rates of acquired drug resistance (8% vs 0.3%) (11).
A Guideline Development Group (GDG) convened by WHO in 2015 to revise policy guidelines on treatment of DR-TB also evaluated available evidence on the treatment of Hr-TB (4). The evidence reviewed could not locate cohort studies or randomized-controlled trials (RCTs) which included fluoroquinolones as part of standardized TB regimens designed primarily for Hr-TB. Three RCTs aimed at investigating whether the use of fluoroquinolones could shorten drug-susceptible TB regimens did not show an advantage when fluoroquinolones were included in the regimens of a limited number of patients who were ultimately diagnosed with Hr-TB (12-14). A study-level analysis based on these patients could not inform the composition of suitable regimens to treat Hr-TB. The GDG thus concluded that no policy recommendation on the treatment of Hr-TB could be formulated and suggested that an analysis of individual-patient data (IPD) from studies of subjects treated for Hr-TB using different regimens be done. An IPD analysis covering patients treated between January 2000 and April 2016 was subsequently commissioned by WHO in order to address major questions related to Hr-TB management.
Methods
These WHO guidelines were developed following the recommendations for standard guidelines as described in the WHO Handbook for Guideline Development, 2014 (2). The GRADE approach (Grading of Recommendations Assessment, Development and Evaluation) was used to rate the certainty in the estimate of effect (quality of evidence) as high, moderate, low or very low and to determine the strength of the recommendations (as strong or conditional) (2).
Preparation for evidence assessment and formulation of recommendations
In preparation for the in-person meeting of the GDG on 27 April 2017 (Annex 1), a WHO Guideline Steering Committee was formed to draft the initial scoping and planning documents (Annex 2). A proposal was submitted to the WHO GRC in February 2017 and was approved in March 2017. In preparation for the GDG meeting, two webinars (via WebEx) were held with GDG members to finalise the scoping, establish the PICO (Patients, Intervention, Comparator and Outcomes) questions, the scoring of the outcomes, and the results of the evidence reviews.
PICO question
The PICO questions, inclusive of sub-populations, treatment regimen composition and duration and outcomes, were agreed upon by the GDG members (Annex 3). The questions were framed to capture the effect of different treatment regimen compositions and duration, when compared with six or more months of treatment with rifampicin-pyrazinamide-ethambutol combination therapy (Annex 3).
GDG members were invited to score the outcomes and the mean scores for the 14 responses received were all in the “critical” or “important” range ().
Scoring of outcomes considered relevant by the GDG for the evidence review related to the WHO treatment guidelines for Hr-TB.
Evidence gathering and analysis
McGill University coordinated the consolidation of an IPD database for Hr-TB during 2016. By November 2016, data on 5418 Hr-TB patients from 33 global datasets were identified and retained for the analysis (Annex 4). All studies identified were observational; no cohort studies or RCTs which included fluoroquinolones as part of standardized TB regimens designed for Hr-TB were identified. Estimates of effect for each outcome were adjusted for age, sex, HIV coinfection, sputum microscopy positivity, cavitation identified in chest radiography, history of TB treatment and resistance to first-line medicines other than isoniazid. Propensity score matching (caliper method with difference of 0.02 allowed, with replacement) was used to estimate the adjusted odds ratios of outcome and their 95% confidence intervals (15).
Decision-making during the Guideline Development Group meeting
Decision-making was based on unanimous agreement among all GDG members or by reaching consensus. No recourse to voting was required during the GDG process.
Certainty of evidence and strength of recommendations
In assessing the quality of evidence, a number of factors can increase or decrease the quality of evidence(16, 17). The highest quality rating is usually assigned to evidence gathered from RCTs while evidence from observational studies is usually assigned a low or very low-quality value. The higher the quality of evidence, the more likely a strong recommendation can be made. The criteria used by the GDG to determine the quality of available evidence are summarised in the GRADE tables annexed to these guidelines (see Online Annexes). The certainty in the estimates of effect (quality of evidence) was assessed and either rated down or up based on: risk of bias; inconsistency or heterogeneity; indirectness; imprecision; and other considerations () (17).
Classification of the certainty in the evidence.
Through the GRADE system, the strength of a recommendation is classified as “strong” or “conditional”. The strength of a recommendation is determined by the balance between desirable and undesirable effects, values and preferences, resource use, equity considerations, acceptability and feasibility to implement the intervention (17). For strong recommendations, the GDG is confident that the desirable effects of adherence to the recommendation outweigh the undesirable effects. For conditional recommendations, the GDG considers that desirable effects probably outweigh the undesirable effects. The strength of a recommendation has different implications for the individuals affected by these guidelines ().
Implications of the strength of a recommendation for different users.
Assessment of the quality of the evidence
One of the advantages of IPD analyses is that they allow the examination of patient-level characteristics, outcome harmonization, and exploration of variability in effectiveness (18). IPD analyses also allow the investigation of whether an intervention is more or less effective for different sub-populations (19). Additionally, between-study heterogeneity can be reduced by IPD analysis, given that results for specific subgroups of participants can be obtained across studies and the differential (treatment) effects can be assessed across individuals.
The Hr-TB IPD was composed of observational studies, and despite the adjustment done for potential confounding using propensity score matching, bias in exposure-effect estimates could still occur due to residual or unmeasured confounding. Residual confounding could also have arisen from unknown factors, associated both with the exposure and the outcome, for which data were not collected. Specific analyses could only be done using variable and limited subsets of the IPD due to limitations in comparability and incompleteness of the data (see Online Annexes). This led to serious imprecision for most of the estimates of effect. The GDG concluded that, overall, the studies included posed serious risk of bias attributed to residual confounding. In view of these factors, the certainty in the estimates of effect was judged to be “low” or “very low”. This influenced the GDG’s decision in favour of conditional rather than strong recommendations for the proposed treatment options (see Annexes 5 & 6 online).
External review
A draft of the guidelines document complete with the recommendations, accompanying remarks and GRADE tables was circulated to the External Review Group for their comments. Feedback provided was incorporated into the subsequent version of the guidelines.
- 13
Concurrent with the release of the new guidelines, and ahead of the update of the Companion handbook, practical implementation issues are being provided in the document “Frequently asked questions on the WHO treatment guidelines for isoniazid-resistant tuberculosis”, available on the same website as the new guidelines.
