Management of patients with multidrug-resistant/extensively drug-resistant tuberculosis in Europe: a TBNET consensus statement

1 Introduction

Tuberculosis (TB) is a leading cause of morbidity and ranks among the 10 most common causes of death worldwide [1]. The emergence of drug-resistant strains of Mycobacterium tuberculosis is substantially challenging the goal of elimination of TB in the 21st century. The estimated global number of multidrug-resistant (MDR) and extensively drug-resistant (XDR)-TB cases among newly diagnosed patients with pulmonary TB in the year 2012 was 450 000 (range 300 000–600 000) [2]. Only 28% of these estimated MDR/XDR-TB cases are notified, and the majority of these cases are from eastern European and central Asian countries [2]. Out of 83 714 notified cases of pulmonary MDR/XDR-TB in the year 2012 worldwide, 36 700 (44%) cases were from the World Health Organization (WHO) European Region [2]. At least 92 countries have now reported patients with XDR-TB and the proportion of MDR-TB cases with XDR-TB is highest in the WHO European Region [2]. A recent study showed that 35.3% of patients with a first episode of TB and 76.5% of patients with a new TB episode who had been previously treated for TB from Minsk, Belarus, were diagnosed with MDR/XDR-TB [3]. This observation was also made countrywide for Belarus and represents the highest rate of MDR/XDR-TB for a country ever recorded [4].

Treatment success rates of MDR/XDR-TB vary between 36% and 79% [5, 6]. The 2013 joint WHO and European Centre for Disease Prevention and Control (ECDC) surveillance report recorded only 31.6% successful treatment outcome for MDR/XDR-TB cases in the European Union (EU)/European Economic Area (EEA) region [7], comparable to the situation in the pre-antibiotic era [8]. It is estimated that, globally, <20% of patients with pulmonary MDR/XDR-TB are currently receiving adequate treatment [2].

Infectious diseases caused by mycobacteria differ from almost all other bacterial infections by the requirement for a very long duration of treatment in order to achieve relapse-free cure with the currently available therapies. This is especially the case for TB when M. tuberculosis is MDR (i.e. resistant to the most potent anti-TB drugs rifampicin and isoniazid) [9]. The current recommendation from WHO for the duration of treatment of patients with MDR/XDR-TB is for a period of 20 months with a combination of four or more anti-TB drugs to which the organism is sensitive [10]. Suboptimal treatment adherence, adverse drug events and high treatment costs during the long course of therapy in MDR/XDR-TB severely compromise treatment success [11, 12].

Major advances have been achieved in recent years for the rapid identification of patients with MDR/XDR-TB [13, 14], but these technologies are not yet universally available and developments in the area of diagnostics do not match the developments in therapeutics or prevention. However, for the first time in more than four decades, a new drug for the treatment of TB was licensed by the US Food and Drug Administration for the treatment against TB (specifically for the treatment of MDR/XDR-TB) at the end of 2012 [15], and following a positive recommendation by the European Medicines Agency (EMA) at the end of 2013, authorisation for the European market for two new antituberculosis drugs is expected soon.

In response to the alarming problem of MDR/XDR-TB, all 53 member states of the WHO European Region have fully endorsed the Consolidated Action Plan to Prevent and Combat MDR/XDR-TB in the WHO European Region 2011–2015, and its accompanying resolution at the 61st session of the WHO Regional Committee for Europe. This Plan has six strategic directions and seven areas of intervention, and aims to contain the spread of drug-resistant TB by achieving universal access to prevention, diagnosis and treatment of MDR/XDR-TB in all Member States of the Region [16].

Up to now, many issues in the management of patients with MDR/XDR-TB rely on expert opinions rather than on evidence. Examples are the evaluation of the duration of infectiveness of the patients, the choice of preventive chemotherapy for contacts of patients with MDR/XDR-TB [17], the optimal duration of therapy in individual patients and the best combination of anti-TB drugs given the drug resistance pattern of M. tuberculosis, to name but a few.

2 Methodology

This document has been developed predominantly by physicians treating patients with MDR/XDR-TB to reach a consensus about difficult management issues relating to the care of these patients. We did not intended to perform a systematic review of the literature with evidence grading or to substitute for recommendations by the WHO or ECDC. Following planning of the document structure by the coordinating authors (C. Lange and D.M. Cirillo) and agreement by the TBNET steering committee, co-authors were invited from among international experts in the field of MDR/XDR-TB (chapter leaders and authors are identified in the Acknowledgements section). Review of the available literature was accomplished by the members of the writing committee and the search for evidence included hand-searching journals, reviewing previous guidelines, and searching electronic databases including MEDLINE and PubMed. Final decisions for formulating recommendations were based upon the result of literature review and practical experience by committee members.

The coordinating authors met on a regular basis while communication between co-authors and exchange of manuscript drafts was performed by e-mail correspondence. Each chapter was reviewed in a three-step process by the chapter co-authors, coordinating authors and all manuscript authors. Consensus statements (chapter 20) were developed in a stepwise approach.

3 Definitions and reference documents

4 Epidemiology of MDR/XDR-TB in Europe

15 of the 27 high MDR-TB burden countries worldwide are in the European Region [31]. In 2011, 49 of the 53 countries of the WHO European Region reported on first-line anti-TB DST; 89 877 new TB cases and 31 929 previously treated TB cases were subjected to DST to first-line anti-TB drugs. In 2011, the prevalence of MDR-TB was 13.2% among new cases and 46.4% among previously treated cases, respectively [7]. In 2011, nine countries had a MDR-TB rate of 10–30% among all TB cases and in five countries this rate was >30% (fig. 1). These countries are Belarus (52.6%), Kazakhstan (39.6%), the Republic of Moldova (42.0%), Tajikistan (66.0%) and Uzbekistan (62.5%) [7]. The data from two of these countries (Kazakhstan and Tajikistan) are based on partial national data only and should be interpreted with caution. The prevalence of MDR-TB among new cases was ≤4% in all EU/EEA countries, except for the Baltic states, where this prevalence ranged from 11.1% in Lithuania to 22.9% in Estonia. Owing to limited DST to second-line drugs (SLDs), data on the extent of XDR-TB are limited. WHO estimated that ∼8100 XDR-TB cases emerged world-wide in 2010 [31]. The vast majority of reported XDR-TB cases reported occurred in the Eastern part of the Region. Of the 3448 MDR-TB cases tested for SLD resistance in 2011, 381 (11.0%) were XDR [7]. The MDR-TB epidemic in the Eastern part of the region coincides with an increasing HIV prevalence; HIV co-infection among TB cases increased from 3.6% in 2008 to 6.2% in 2011 [7].

• Download figure
• Open in new tab
• Download powerpoint

Figure 1–

Percentage of notified tuberculosis (TB) cases with multidrug resistance among all TB cases with drug susceptibility results in the World Health Organization European Region, 2011. Note that the data from Azerbaijan, Italy, Kyrgyzstan, Serbia, Spain, Tajikistan, Turkey, Turkmenistan and Uzbekistan cover only part of the country, and may therefore not be representative and should be interpreted with care. In addition, note that data from the United Nations Administrated Province of Kosovo (in accordance with Security Council Resolution 1244 (1999)) is not included in the figures reported for Serbia. Reproduced from [7] with permission from the publisher.

In the last decade, the proportion of MDR-TB cases increased steadily in the European Region; it tripled among previously treated cases and increased five-fold among new cases. These data suggest active transmission of MDR-TB. According to the latest ECDC report, out of the 78 000 (range 31 000–85 000) estimated MDR-TB cases in the region, only ∼30 000 MDR-TB cases were notified in 2011, mainly because of limited laboratory capacity [7]. These estimated 78 000 MDR-TB cases represent 25% of the 310 000 (range 220 000–400 000) MDR-TB cases estimated to occur globally in 2011, whereas the number of new and relapse TB cases in the region (380 000) only represents just 4.4% of the global burden of TB, underscoring the specific problems with MDR-TB in the WHO European Region [7, 32]. The estimated number of patients with MDR/XDR-TB has increased again in 2012 to 450 000 (range 300 000–600 000) globally [2].

5 Risk factors for M. tuberculosis drug resistance and MDR/XDR-TB

While a formal systematic review would have been beyond the scope of this document, a selection of risk factors for MDR-TB, based on in-depth knowledge of the literature for this topic, is reported in table 3. Among them, prior TB treatment for >1 month is the most important. This could particularly be the case if the treatment was inappropriate for the drug susceptibility of the strain [42], although the evidence is scarce. Other prominent risk factors include close contact to a patient with MDR-TB, migration, HIV infection and young age. TB contact and previous imprisonment showed an increased risk for XDR-TB in patients with MDR-TB, although this trend is not statistically significant due to the low number of patients, even in the larger studies [36, 40].

Early discontinuation of adequate therapy is a risk factor for relapse but does not seem to induce drug resistance [43]. On the contrary, failure of treatment or relapse after therapy may be the consequence of the use of an inappropriate treatment against MDR/XDR-TB or with the causative M. tuberculosis strain [44]. It can be postulated that inadequate quality of anti-TB drugs and low drug exposure contribute to the development of drug resistance in M. tuberculosis [45, 46].

It is alarming that MDR/XDR-TB is now observed in a substantial proportion of patients without risk factors for MDR/XDR-TB, other than living in an area of high MDR/XDR-TB prevalence in Europe [4, 39], or in new patients who have never been treated for TB [47]. This indicates active transmission of drug-resistant strains in these countries, with an important role for imprisonment [48] and hospitalisation [41], possibly because of insufficient infection control measures in these facilities.

Important for clinical management is the possibility that strains of M. tuberculosis may have acquired additional drug resistance beyond isoniazid and rifampicin. A prospective cohort study in eight countries demonstrated that bacterial strains of 43.7% of patients with MDR-TB already showed resistance against at least one SLD, 20.0% against one injectable drug and 12.9% against fluoroquinolones [36]. Previous treatment with SLDs was the strongest risk factor for the development of resistance against these drugs, thus increasing the risk of XDR-TB. A study in Russia demonstrated that 6% of patients with MDR-TB developed XDR-TB during TB treatment [49]. The risk factors were extensive lesions, the prior use of injectable SLDs and long duration of inappropriate treatment with insufficient intake of the medication.

In most Eastern European countries, MDR/XDR-TB must be highly suspected in all patients who were treated for TB in the past. However, as transmission of MDR/XDR M. tuberculosis strains in the community is high, all patients with TB in these countries must be evaluated for MDR/XDR-TB by rapid diagnostics. In Western European countries, MDR/XDR-TB must be considered especially in migrants with TB coming from high MDR/XDR-TB prevalence countries. According to unpublished data from Switzerland, 90% of MDR-TB cases were observed among foreign-born patients, who represent some 20% of the local population. Among MDR-TB patients, the absolute number of new cases was higher than previously treated patients. Although never investigated in detail, the risk of MDR/XDR-TB is probably very high in those patients who migrate to other countries for the purpose of seeking TB treatment [50].

6 Contact investigations and preventive chemotherapy in MDR/XDR-TB contacts

Molecular epidemiology studies suggest that, in some settings, the majority of MDR/XDR-TB cases are related to ongoing transmission of MDR strains of M. tuberculosis [51]. This emphasises the need for active case finding, rapid diagnosis, comprehensive contact tracing and infection control in order to reduce disease transmission within the community [52, 53]. The consequences of acquiring an infection with MDR/XDR-TB are more serious than infection with a drug-susceptible strain. This warrants high priority for the screening of close contacts of MDR/XDR-TB patients [54, 55].

The need to screen contacts should be assessed on the basis of 1) the infectiousness of the index case, 2) intensity of exposure and 3) susceptibility of contacts to TB [28]. Patients with acid-fast bacilli (AFB) sputum smear-positive TB are most infectious. Intensity of exposure is determined by proximity to the index patient, duration of exposure, and the physical conditions of the location where transmission may have occurred in terms of shared breathing space, ventilation and ambient light. Consequently, household contacts and contacts sharing breathing space for prolonged periods are most at risk. Young children and immunocompromised persons have a higher risk of developing TB [52]. Contact investigation should include all close contacts and may be expanded to lower risk contacts [28, 56] as recommended in national guidelines.

Following an incubation period of about 6–8 weeks, the risk of progression to TB is highest shortly after infection, followed by an exponential decline in subsequent years [57, 58]. Therefore, contacts should be evaluated for infection with M. tuberculosis with the tuberculin skin test (TST) and/or an interferon-γ release assay (IGRA) ≥8 weeks after the last infectious exposure. Immunocompromised contacts, children and contacts with symptoms suggestive of TB should be evaluated as soon as possible after diagnosis of the index case. If a contact is symptomatic, or considered to be infected, TB disease should be excluded. In children <5 years of age and immunocompromised persons, TB disease should be excluded, irrespective of the TST or IGRA response. All contacts or caretakers should receive information regarding the likelihood of latent infection with M. tuberculosis (LTBI) developing into TB disease and symptoms of TB disease.

The evidence for the effectiveness of chemoprophylaxis or preventive therapy of MDR/XDR-TB contacts when infected is scarce [59, 60], although there are indications of a positive effectiveness in children [61] and in adults [62, 63]. Given the lack of evidence from controlled clinical trials, both preventive treatment and careful clinical follow-up are valid options for the management of contacts, provided they are informed of the benefits and harm [17, 59, 64]. Special attention with regard to individual risk assessment should be paid to children <5 years of age and immunocompromised persons, particularly those with evidence of LTBI. If chemoprophylaxis or preventive treatment is considered, the drug regimen is dictated by the drug-resistance pattern of the source case, whether a drug with bactericidal activity is available, the likelihood of adherence and the risk of adverse events. There are no data defining the minimum duration of such a preventive therapy or whether it is safe to give the drug as monotherapy. Therefore, until further data are available, monotherapy may be inadvisable. It is essential to monitor adverse events closely and to provide support to ensure adherence during preventive treatment. Chemoprophylaxis may be discontinued if, based on further examination, infection is found to be unlikely.

When preventive therapy is not an option, or risks outweigh the benefits, contacts and their general practitioners should be educated about the risks of progressing to TB disease. Temporary periodical follow-up can be considered, but easy access to a specialised TB clinic in case symptoms appear is of higher importance [65, 66].

7 Infection control measures for MDR/XDR-TB

Infection control measures, especially in congregate settings, are necessary to prevent transmission of M. tuberculosis. CDC and WHO have developed policies and guidelines on TB infection control [67, 68]. A toolbox with documents that assist in implementing the WHO policy is available from the TB CARE website [69]. When it comes to preventing transmission, there is no essential difference between MDR/XDR-TB strains and other strains of M. tuberculosis [70, 71].

The persons to be protected are household contacts (e.g. family members), and, in a hospital setting, other patients, healthcare workers, cleaners, administrative staff, etc., and visitors. The best way to prevent transmission of M. tuberculosis is prompt diagnosis and start of effective treatment [72–74]. Contagiousness of the index case can be measured by investigating the sputum for AFB, but ∼15% of transmission in the community comes from sputum AFB smear-negative index cases [75]. There is a wide variability in contagiousness of AFB sputum smear-positive index cases [76, 77]. It must be considered that, in general, the duration of contagiousness of patients with MDR/XDR-TB is longer than that of patients with non-MDR/XDR-TB [78].

Individuals suspected of having TB should be separated from other patients and evaluated for TB without waiting in general areas. Hospitalisation should include airborne isolation precautions and be limited primarily to contagious AFB sputum smear-positive TB patients. Infectiousness is substantially reduced once a patient is on an adequate regimen and it is probably not necessary to keep a patient in hospital until their cultures become negative [72, 73].

In some settings, patients with MDR/XDR-TB are discharged from the hospital after 2 weeks of an adequate treatment [74, 79–81], although many centres require a negative result from three sputum cultures collected over a 14-day period in patients with MDR/XDR-TB before they are considered for discharge. At this time, the optimal time for safe discharge of patients with MDR/XDR-TB is not known.

In daily practice, strict isolation of a patient is often not possible, such as in a patient with ongoing nicotine addiction. Smoking is generally forbidden in patient rooms and patients with contagious TB, especially MDR/XDR-TB, should not use shared smoking shelters. If patients are unwilling or unable to stop smoking, they should smoke outside. There must be awareness that in-patients will socialise in the evenings, at nights and on weekends, when infection control measures by hospital staff cannot be strictly enforced.

Any M. tuberculosis transmission risk reduction programme should include the issues of administrative and environmental measures and personal respiratory protection [68]. In hospitals specialising in the care of patients with MDR/XDR-TB, M. tuberculosis transmission risk assessment should be conducted regularly within a TB infection control plan. In high-risk areas of M. tuberculosis transmission, procedures and staff should be identified for personal respiratory protection (e.g. education and training on donning respirators, including their seal checks, use, care and disposal); respirator fit tests (qualitative or quantitative) should be organised on an annual basis [81–83].

Hospital staff sharing air with contagious patients with MDR/XDR-TB should be wearing FFP2-certified respirators [84]. Use of surgical masks by contagious MDR/XDR-TB patients indoors reduces transmission risk by half [83]. Because of the very high air change rate outdoors, any recommendation to use masks for TB patients in the open air is not justified.

Natural ventilation should be maximised by appropriate architectural design and rational administrative support [85]. Sustainable environmental controls should be implemented for high-risk areas for TB transmission where administrative controls cannot further substantially reduce the risk. Mechanical ventilation should provide directional airflow, and negative pressure is required in isolation rooms and other high-risk areas. 12 air changes per hour (ACH) are recommended for new facilities, but 6 ACH may be acceptable for older facilities [86]. Contaminated air exhaust outlets should by located ≥8 m from windows, air intakes and occupied areas, or effectively decontaminated by in-duct ultraviolet germicidal irrigation (UVGI) or high-efficiency particulate air filtration. Upper room UVGI fixtures, professionally designed, installed and maintained, are an effective and sustainable option for limited-resource settings [87–89]. Air recirculation and room air cleaners are not recommended for high M. tuberculosis transmission risk areas. To ensure sustainability of engineering controls, their application should be rationally limited to high-risk areas; systems should be professionally designed, installed, commissioned and maintained.

As bronchoscopy and endotracheal intubation of patients with pulmonary TB pose a very high risk of transmission for all occupants in the room, a clear protocol of TB hygiene precautions should be available for these interventions in patients with pulmonary TB or those with an unclear diagnosis where pulmonary TB is part of the differential diagnosis. Indications for such procedures should be strictly limited; they should be performed in isolated, ideally negative pressure mechanically ventilated rooms with 12 ACH. All occupants should wear FFP2 or FFP3 respirators. If effective mechanical ventilation is not available, the whole room should receive UVGI after completion of the procedure with sufficient exposure. To help reduce the risk of contaminating a ventilator or discharging M. tuberculosis into the ambient air when mechanically ventilating (i.e. with a ventilator or manual resuscitator) a patient with suspected or confirmed TB, a bacterial filter should be placed in the patient’s endotracheal tube (or at the expiratory side of the breathing circuit of a ventilator). In selecting a bacterial filter, preference should be given to models specified by the manufacturer to filter particles ≥0.2 μm in size in both the unloaded and loaded states with a filter efficiency of >95% (i.e. filter penetration of <5%) at the maximum design flow rates of the ventilator for the service life of the filter, as specified by the manufacturer [68].

Transport of contagious patients inside the hospital should be limited (avoid use of lifts, sitting in waiting areas, etc.). Transport of contagious patients outside the hospital should preferably be carried out in vehicles with a separate compartment for the patient and for accompanying persons. If weather permits, open windows allow very high air change rates inside the vehicle. At least a surgical mask or a respirator without an exhalation valve should be worn by the patient, and a FFP2 respirator worn by the driver and accompanying person during the journey.

Some of these measures recommended by the international airborne infection control expert community have no solid evidence base and there is still a lot of research to be done in this field.

8 Principles and interpretation of M. tuberculosis DST

Quality-assured bacteriological examination is an essential element for diagnosis and management of patients infected with susceptible or resistant TB bacteria. Phenotypic DST remains the mainstay for the detection of drug resistance in M. tuberculosis and is based on detection of bacterial growth in the presence of antibiotics. Traditional solid-media culture techniques are reliable and cheap but slow, with an overall turnaround time of 6–12 weeks when a patient specimen is taken. Automated liquid-media systems are faster (average turnaround time of 14–21 days) [90, 91], and are recommended for use even in low- and middle-income countries. Whatever the test used, results from the Supranational Reference Laboratory Network have clearly shown that, while DST for rifampicin, isoniazid, fluoroquinolones and second-line injectables is reliable in solid- and liquid-media systems, there is more controversy around the standardisation of DST for other drugs (ethambutol, pyrazinamide and other SLDs) [91–99]. The WHO policy of testing for SLD susceptibilities suggests testing MDR-TB isolates for sensitivity to amikacin, kanamycin, capreomycin and newer fluoroquinolones (moxifloxacin at two concentrations or ofloxacin/levofloxacin plus high concentration of moxifloxacin). For fluoroquinolones, testing of the drug in use in the country is recommended.

Recently, with the insights from mycobacterial genomics, molecular techniques to detect antibiotic resistance have been established. They offer advantages like turnaround times of hours and the possibility of omitting microbiological culture. A prerequisite is the knowledge of specific genetic mutations that are undoubtedly associated with resistance. The development of drug resistance in M. tuberculosis isolates results from random genetic mutations in genes involved with the drug mechanism of action. Several molecular assays have been established, allowing the prediction of drug resistance in clinical M. tuberculosis isolates within one working day or less.

Rifampicin is the most important drug for the treatment of TB and is considered a surrogate marker of MDR for its frequent association (80–95% of cases) with resistance to isoniazid [100, 101].

An 81 base pair region of the rpoB gene (encoding the β-subunit of RNA polymerase) harbours mutations responsible for rifampicin resistance in ∼97% of resistant strains [102]. This makes rifampicin resistance the ideal target for molecular testing.

Isoniazid requires activation in vivo by catalase–peroxidase (KatG); it is therefore not surprising that mutations in the katG gene cause the majority of isoniazid resistance (50–90% of isoniazid-resistant strains) [103]. Mutations in other regions may also be responsible for resistance, primarily by mutations in the inhA promoter and coding region, which encodes the target of activated isoniazid. On average, 80–90% of isoniazid-resistant strains present mutations in either katG or inhA [104]. Mutations in inhA confer cross-resistance to ethionamide while strains mutated in katG may still retain ethionamide sensitivity [105]. The different mutations confer different levels of resistance to isoniazid. In some cases, a low level of resistance can be overcome by the use of higher doses of isoniazid [106, 107] and it is therefore worth clarifying the resistance level with the microbiological laboratory.

For other first-line drugs (FLDs), suboptimal correlation or disagreement exists in the interpretation of the results: mutations in codon 306 of embB are present in 47–62% of ethambutol-resistant strains [108, 109], mutations in rpsL (encoding ribosomal protein S12) are present in 54% of the streptomycin-resistant strains and mutations in pncA (encoding pyrazinamidase) characterise 72–97% of pyrazinamide-resistant strains [110]. In some cases, the disagreement between phenotypic and genotypic results is due to the suboptimal performance of in vitro DST. The main mutations responsible for resistance identified for SLDs are in the gyrA gene (encoding a DNA gyrase) for fluoroquinolones (60–70%) [111, 112], and in the rrs (encoding the 16S ribosomal RNA) and eis (conferring enhanced intracellular survival) genes for injectable drugs [113, 114].

Gene sequencing is the gold standard for mutation detection. The current commercial PCR-based methods focus mainly on the detection of rifampicin resistance. They target the hot-spot region for mutations of the rpoB gene and include two line probe assays (LPAs) (INNO-LiPA Rif.TB (Fujirebio, Ghent, Belgium) and GenoType MTBDRplus (Hain Lifescience GmbH, Nehren, Germany)) and the Xpert MTB/RIF system (Cepheid, Sunnyvale, CA, USA). Very recent data have shown that testing for rifampicin in liquid media may produce false-negative results due to the low growth rate of strains harbouring specific mutations in rpoB [115]. These uncommon mutations are well identified by the Xpert MTB/RIF system but their presence needs to be confirmed by rpoB gene sequencing, not by liquid culture.

Among the genetic tests, the LPAs are best suited for the rapid diagnosis of MDR-TB in reference or well-equipped laboratories from sputum smear-positive samples or strains isolated in culture. Molecular assays are highly accurate for identifying rifampicin resistance, as the mechanism of resistance is known. The detection rate of isoniazid resistance is lower due to the heterogeneity of the molecular biology background of this resistance.

Although the LPAs are endorsed by WHO for the rapid detection of MDR in smear-positive samples only, the newer generation tests show a high sensitivity in smear negative respiratory samples [110].

The commercially available LPA GenoTypeMTBDRsl aims to detect resistance to ethambutol, the second-line injectable drugs (SLIDs) and the fluoroquinolones. Because of the low negative predictive value (NPV) (probes included only detect a fraction of mutations known to cause resistance [109]), the test can only “rule in” XDR-TB and has not been endorsed by WHO [91]. Newer generations of tests targeting a larger number of mutated genes may overcome this problem in the future.

The only completely automated test able to detect simultaneously the presence of M. tuberculosis complex DNA and rifampicin resistance in clinical samples is the Xpert MTB/RIF assay [13, 117]. This assay simplifies molecular testing by fully integrating and automating the whole process required for sample preparation, real-time PCR and result interpretation. The Xpert MTB/RIF assay is currently the only technology suitable for near-patient use and it is recommended by WHO as the first test when MDR-TB is suspected and where TB is suspected in HIV-infected individuals [118]. Sensitivity and specificity for the detection of TB and MDR-TB in both pulmonary and extrapulmonary samples are high [119]. If the test is used in a low-prevalence MDR-TB population to test subjects not considered at risk for MDR-TB, rifampicin-resistant cases should be reconfirmed by a different rapid test or liquid culture. A second Xpert MTB/RIF test on a different sample, although not optimal, could be considered as a fast re-testing strategy.

9 Treatment of patients with MDR/XDR-TB

Treating drug-resistant TB, especially rifampicin-resistant TB, has become a major challenge for TB control worldwide. With high bactericidal and sterilising activities, and low toxicity, rifampicin is by far the most active TB drug. As rifampicin monoresistance is relatively rare, except in HIV-infected patients, it serves as a surrogate marker of MDR-TB [120]. Among SLDs, newer-generation fluoroquinolones have demonstrated prominent bactericidal [121] and sterilising activities [122]. Bacillary resistance to fluoroquinolones clearly influences the prognosis in XDR-TB [123, 124].

Regardless of the drug-resistance pattern, all TB forms should be treated with a combination of drugs for a sufficient duration to avoid relapse. The number of drugs in a multidrug regimen and the total treatment duration depend on the bactericidal and sterilising activities of the TB drugs included [123, 125]. Availability of drugs that are likely to be effective decreases as drug resistance becomes more extensive.

9.1 Number of drugs for treating MDR/XDR-TB

If M. tuberculosis is fully susceptible to all the TB drugs used in a regimen, two active core drugs, such as isoniazid and rifampicin, are probably adequate for cure [123]. In the treatment of MDR-TB, WHO advises giving at least four drugs that are known or likely to be effective against the drug-resistant M. tuberculosis strain harboured by the patient, plus pyrazinamide (fig. 2) [126]. If possible, WHO group 2 (amikacin, capreomycin or kanamycin) and WHO group 3 (fluoroquinolones) should be core drugs with the rest (ethionamide/prothionamide and cycloserine) being accompanying drugs that may help protect the core drugs in the initial phase when bacillary load is high [124, 125].

• Download figure
• Open in new tab
• Download powerpoint

Figure 2–

Choosing an initial drug regimen for the treatment of patients with suspected/confirmed multidrug-resistant (MDR)/extensively drug-resistant (XDR) tuberculosis (TB). DST: drug susceptibility testing; M. tuberculosis: Mycobacterium tuberculosis.

10 Duration of MDR/XDR-TB treatment

One of the most difficult current controversies in the treatment of MDR/XDR-TB is the optimal duration of therapy [124]. Neither the optimal duration of the initial intensive phase, when an injectable agent such as amikacin, capreomycin or kanamycin is usually given, nor the total duration of therapy have been tested in a randomised trial. Prior to 1970, standard therapy for drug-sensitive TB was ≥18 months with three or more drugs [174, 175]. Subsequent to the introduction of rifampicin into clinical practice in 1970, multiple randomised trials were conducted, which demonstrated that, with use of rifampicin, the total duration of therapy could be reduced to 9 months [176, 177]. Subsequent trials demonstrated that, with the addition of pyrazinamide, total therapy duration could be further reduced to 6 months [176, 177]. Given the historic experience with regimens without rifampicin, it has been conventional wisdom that 18 months should be the minimum total duration of therapy for patients with TB caused by rifampicin-resistant bacteria.

Recently, individual patient records from 32 centres that had reported the results from cohorts of patients were assembled into a single pooled data set. Most of these received individualised (and nonrandomised) treatment. These data comprised a total of 9898 patients with MDR-TB of whom 9153 had pulmonary MDR-TB [178]. The availability of detailed individual patient data enabled analysis of treatment covariates with stratification, restriction or adjustment for a number of confounding variables such as age, previous treatment, drug-resistance pattern or HIV infection. Analysis of this large data set revealed that the optimal duration of the initial intensive phase was 7–8.4 months and the optimal total duration of therapy was 19–21 months. These findings are summarised in figure 3. A number of additional sensitivity analyses were performed, such as restricting to adult patients who had never been treated with SLDs before, excluding patients who were treated only with FLDs or excluding HIV-infected patients [178]. These sensitivity analyses confirmed the findings, which form the basis of recent recommendations by WHO on the management of MDR-TB, which recommend that the duration of treatment is modified according to the response to therapy [126]. Evidence-based recommendations for the duration of treatment for patients with XDR-TB are not available. In the absence of evidence, recommendations for the duration of treatment in MDR-TB currently also apply to patients with XDR-TB.

• Download figure
• Open in new tab
• Download powerpoint

Figure 3–

Association of the duration of initial intensive phase and total treatment with adjusted odds of treatment success in patients with multidrug-resistant tuberculosis. aOR: adjusted odds ratio. Data from [178].

Of course, this long duration creates a number of problems, including a greater risk of toxicity, potentially greater default rates, greater health system costs and greater economic losses for patients who often cannot work. The high costs associated with the treatment of MDR/XDR-TB may represent an intolerable burden to some National TB Programmes [179].

Two drugs that might allow shortening of current MDR/XDR-TB therapy are pyrazinamide and the fluoroquinolones. However, resistance to pyrazinamide is common among patients with MDR/XDR-TB (in the individual patient data meta-analysis, slightly over half of those tested had resistance to pyrazinamide), reflecting prior use in most of these patients [178]. Fortunately, resistance to fluoroquinolones is much less common and the fluoroquinolones are highly active against M. tuberculosis, particularly the later generation fluoroquinolones, such as levofloxacin, moxifloxacin or gatifloxacin [180]. A nonrandomised study from Bangladesh reported excellent results with a much shorter regimen, which used gatifloxacin throughout [106]. Of the 9153 MDR-TB patients referred to earlier, only 14% had access to later generation fluoroquinolones [178]. It is possible that the results would have been different if more centres had these drugs available.

Many experts will recommend shorter regimens for patients with smear-negative or less radiologically extensive disease and patients with little prior exposure to TB treatment, particularly to SLDs, as well as those with smear or culture conversion in <3 months. It should be noted that these suggestions are based on expert opinion and personal experience [25, 181]. WHO has provided guidance to countries listing the criteria under which these shorter regimens for MDR-TB treatment can be used [24]. In order to individualise the duration of therapy for patients with MDR/XDR-TB, identification and validation of biomarkers that indicate a point in time during therapy corresponding to relapse-free treatment success are urgently needed.

A randomised controlled trial (STREAM) is ongoing to test a shorter treatment duration regimen. This trial is expected to end in October 2016 [182]. Results from the STREAM trial will be crucial in determining whether therapy for MDR-TB can be successfully shortened and, if so, in which groups of patients. Treatment with new drugs, like bedaquiline and delamanid, may allow shortening of MDR/XDR-TB therapy, although it is likely that several trials will be required to determine the optimal duration of therapy with these new agents.

11 Management of adverse drug events in MDR/XDR-TB

Drugs used to treat MDR/XDR-TB are often associated with adverse effects [66, 183, 184]. These can occur early or late in treatment. It is very important to anticipate the development of adverse events and to inform patients about this. Giving treatment and, in some cases, discontinuing or changing medication can all help improve adherence and enable the patient to complete their course of treatment. Ensuring that the patient understands the importance of each medication, the possible side-effects of medication they are taking and what to do when they occur can reduce the risk of serious adverse events. Herein, we present a brief description of adverse events by system. table 5 outlines the main adverse effects of each drug class and provides management guidance.

12 Special considerations for MDR/XDR-TB in-patient case management

It is agreed that the initial management of patients with MDR/XDR-TB in Europe should generally take place in a hospital setting. All patients suspected of having pulmonary MDR/XDR-TB should be individually isolated after admission to a hospital to avoid nosocomial transmission of M. tuberculosis (see chapter 7 for infection control measures). Patients need to be informed about TB, preventive measures, diagnostic procedures and treatment modalities in a language that they understand. Brochures and videos (www.explaintb.org) can be very helpful for patient education and to promote adherence. When the diagnosis of MDR/XDR-TB is confirmed, the long-anticipated treatment duration and related issues of isolation and infection control and adherence need to be discussed with the patient. In the great majority of low TB incidence countries in Europe, patients with MDR/XDR-TB often have a background of recent migration [192, 193]. In fact, MDR/XDR-TB is sometimes the reason for healthcare-seeking migration. Psychosocial support is particularly important in patients who lack local social networks including recent visits from family and friends.

Rapid molecular resistance testing should be the standard of care for all patients evaluated for MDR/XDR-TB [26, 66]. Treatment should ideally be initiated when molecular test results are available. The cost of molecular testing, available at national and supranational reference laboratories throughout Europe, is less than the cost to the patient and to a health system of an inadequate regimen.

Unless a patient is very ill, treatment initiation is not an emergency. However, the risk of nosocomial transmission decreases substantially as soon as a patient receives adequate treatment and treatment should not be unnecessarily deferred. Comorbidities, such as HIV infection (see chapter 16), viral hepatitis, diabetes mellitus and potential drug–drug interactions, need to be thoroughly evaluated before anti-TB therapy is started. The decision to initiate anti-TB therapy and the choice of the combination drug regimen should be made by a physician with experience in the care of patients with MDR/XDR-TB, preferably in agreement with a consilium. When patients are initially admitted to nonspecialised departments, transfer to a specialised referral centre is to be preferred. This is especially important in low TB incidence countries in western Europe, where experience with complicated cases of TB is usually restricted to very few centres in a country.

Prior to the initiation of therapy, a set of baseline investigations is required (table 6). When the choice of a drug regimen is made, all drugs should be administered at once without stepwise introduction. Directly observed treatment should be ensured in the in-patient setting. Drug concentration measurement can be used to optimise drug exposure.

When patients receive injectable drugs, implantation of a peripherally inserted central catheter (PICC line) or, preferably, a central venous catheter with a subcutaneous silicon reservoir for injections (e.g. PORT-A-CATH; Smiths Medical, St Paul, MN, USA) should be encouraged to avoid daily intramuscular or intravenous injections through peripheral veins over a period of several months.

Depending on the treatment regimen, patients should be informed about signs and symptoms of adverse drug events. Patients need to know that the medication must be discontinued immediately under certain circumstances, such as when hearing or balance impairment develops.

Patients who smoke cigarettes should be strongly encouraged to stop smoking as it improves the chance of a positive treatment outcome [194]; nicotine replacement therapy should be offered.

Continued alcohol intake may lead to hepatotoxicity, which can be difficult to distinguish from drug toxicity. Abstinence from alcohol is recommended and treatment with benzodiazepines and thiamine should be given to cover withdrawal. Opioid substance abuse should be treated with methadone in a substitution programme. The problem of substance abuse should be re-addressed repeatedly and support must be offered to maximise treatment success.

Malnutrition is common. In addition to a good diet, vitamin D may improve sputum smear conversion in patients, depending on a vitamin D receptor polymorphism [195]. A diet rich in cholesterol has been associated with accelerated sputum conversion [196]. In severely malnourished individuals, enteric feeding should be considered if providing high-caloric drinks does not result in weight gain. Feeding via a percutaneous endoscopic gastrostomy may be required for those who cannot take regular meals (both methods can deliver drugs whose taste affects adherence). Vitamin B₆ (50 mg·day⁻¹ for adults, optimal dosage unknown) should be given with high-dose isoniazid and cycloserine, and has also been suggested for linezolid and the thionamides [197].

Physiotherapy and occupational therapy are often withheld from in-patients with TB, especially patients with MDR/XDR-TB, because of the risk of infection for the hospital staff. Due to the long duration of in-patient treatment for those with MDR/XDR-TB [198], regular exercises for body and mind are especially important for these patients. Hospitals must provide possibilities for regular physical activities (e.g. exercise bikes in isolation rooms and open-air sports) to ensure muscle mass is regained with treatment. Occupational therapy and the use of the new social media can reduce the sense of isolation and stimulate the mind. When patients are unable to communicate in the language of the hospital staff, regular conversations with interpreters should be arranged. Face-to-face interpretation is preferred. Sometimes telephone conferences with an interpreter may be helpful to communicate with a patient.

Response to treatment and adverse drug events should be assessed and monitored at regular intervals (table 6). Sputum smear and culture examination should be performed at least once a month [199]. Time to culture positivity is a valuable measure of response to treatment [200]. In contrast, clinical and radiological scores are poorly validated.

Although patients with MDR/XDR-TB are sometimes discharged from the hospital as early as 2 weeks after initiation of an adequate treatment regimen [74, 79–81], in many centres, patients with MDR/XDR-TB are considered for discharge when three sputum cultures collected over a 14-day period are culture-negative. In practice, some patients may stay in isolation for >6 months.

Prior to the discharge of a patient with MDR/XDR-TB from the hospital, continuation of care and monitoring during the outpatient phase of treatment must be ensured [66]. table 7 provides a practical checklist of issues that need to be addressed and solved before a hospitalised patient with MDR/XDR-TB should be discharged. This checklist is also available as a form for clinical use in the online supplementary material (table S1).

13 Special considerations for MDR/XDR-TB outpatient case management

Outpatient treatment of patients with MDR/XDR-TB has been considered problematic because of the need for daily directly observed therapy (DOT), particularly if treatment includes multiple daily doses of medications (including i.v. or i.m. administered medications). Nonetheless, several reports of outpatient treatment of patients with MDR/XDR-TB have demonstrated encouraging results [201–204]. WHO now recommends that patients with MDR-TB should be treated using mainly ambulatory care [66]. In some countries where in-patient capacity is low at the beginning of the treatment, isolation may be required in the patient’s own home with the “hospital at home” approach that was employed in the early treatment of drug-sensitive TB [205].

Each patient should have a specifically assigned healthcare worker, so that there is a single point of contact (e.g. for adverse effects, loss of tablets and change in social circumstances). Training of assigned healthcare workers is essential in order to coordinate care across community, hospital and public health services. Outpatient care and the provision of DOT are managed by healthcare providers in a different manner in, and even within, each European country. Coordination, the timely transfer of information by the assigned healthcare worker and a record of each dose of each drug that has been administered are important to ensure that treatment is successfully completed. The response to treatment should be assessed by monthly sputum analysis (including recording of the times of sputum smear and culture conversion), ideally until the end of treatment. However, most patients are unable to provide sputum samples shortly after sputum culture conversion. The inability to produce sputum for microbiological examinations should then be recorded. Repeated DST is required if a patient who had a negative sputum smear develops positive sputum smears again or if negative cultures become positive again.

Recommendations for monitoring treatment response (smear, culture and imaging studies), screening for adverse events (audiogram and laboratory tests) and infection control precautions should follow a similar pattern proposed for the in-patient care (table 6).

Great care should be taken to ensure that patients continue to take all their treatment for the duration prescribed. The full participation of the patient is helped by a good relationship with a key worker and by an understanding of the illness [206]. Patient and family education is often needed in order to reduce stigma. Patients must be continuously educated about infection control measures, the importance of ventilation, the need for separate sleeping rooms, and the importance of avoiding exposure to young children and immunocompromised individuals [207, 208]. Incentives, such as transportation subsidy or food, have been shown to increase adherence to treatment, particularly in high-risk groups such as drug users, the homeless and prison inmates [209].

A flexible, organised outpatient centre on the model of the One-Stop TB Shop [210] should be available to manage the cases, provide expert medical advice and deal with adverse events. Regular (e.g. 3-monthly) review of patients with MDR/XDR-TB by a multidisciplinary team (consilium) is helpful to ensure good treatment outcomes (see chapter 18).

Kept hospital appointments are a marker for adherence and follow-up by clinical assessment, and chest radiography is valuable as an early pointer to treatment failure [211]. Identification of potential causes of nonadherence should be sought and resolved.

Facilities for MDR/XDR-TB outpatients should also be available for contact investigations. In order to prevent loss of information between different health services, including the laboratory, and to allow comparability of data, there is a need to define early the information required, harmonising support, responsibilities and flow of information. At the last outpatient visit, treatment outcome, based on standardised definitions, must be recorded, and previous healthcare providers (hospital, consilium, etc.) and surveillance systems must be informed. End-of-treatment imaging studies are very helpful as patients may be reassessed later for suspected relapse or reinfection. Patients must have clear information about signs and symptoms of recurrent TB. They should already have the date of their first follow-up appointment and should be actively followed for ≥2 years. The end-of-treatment checklist is presented in table 8 and is also available as a form for clinical use in the online supplementary material (table S2).

14 Special considerations for MDR/XDR-TB in children

The well-described limitations surrounding correct and timely diagnosis of drug-sensitive childhood TB apply equally to MDR/XDR-TB: the paucibacillary nature of the disease implies that in the best case scenario, only 30–40% of children diagnosed with TB have testable isolates in culture systems, depending on age and site of disease [212–215]. In the absence of bacteriological confirmation, children are often not included in global surveys of MDR/XDR-TB incidence, and data on the global burden [32, 216, 217] and cohort outcome [218–223] are sparse. In addition, the treatment regimens are deduced from the adult recommendations and information on drug pharmacokinetics; safety and dosing data of paediatric MDR/XDR-TB are currently lacking.

Judging by adult MDR/XDR-TB prevalence figures, it is likely that MDR/XDR-TB in children is seriously under-reported. TB in children is usually transmitted from a sputum smear-positive adult index case and the same applies to MDR/XDR strains [224, 225]. However, rather than developing primary resistance through incomplete or inadequate regimens, as seen in adults, MDR/XDR-TB in children is usually caused by the acquisition of an already-resistant strain [226]. In line with drug-susceptible TB, therefore, a case of MDR/XDR-TB in a child equally represents a sentinel event of transmission in the community and can serve as an indicator for TB control programmes.

The main challenges are not only the management of active cases (confirmed or suspected) but also the uncertainty of what to do with children who have been exposed but do not yet have signs of active disease. We do not know for certain if they require prophylactic treatment to prevent infection or treatment of LTBI with a presumed drug-resistant strain, or if the risks of potential side-effects and use of complicated drug regimens outweigh the benefits.

It is well established that the likelihood of progression to active disease varies with age in drug-susceptible disease and there is no reason to assume that this should be any different for MDR/XDR cases of TB. Hence, management of LTBI and MDR/XDR exposure needs to take age and other risk factors into account. The complexity of these issues has resulted in a dearth of evidence-based guidelines for paediatric MDR/XDR-TB. Here, we summarise the available evidence to date and outline the further research priorities.

15 Special considerations for MDR/XDR-TB in pregnancy

There is consensus that neither LTBI following contact of a patient with MDR/XDR-TB nor active MDR/XDR-TB during pregnancy requires termination of pregnancy [230]. While treatment of females who develop MDR/XDR-TB during pregnancy or become pregnant during treatment can be successful without adverse events for the newborn [230–233], the safety of many drugs for the unborn child is unknown. FLDs are not considered teratogenic but there is little knowledge about human use of SLDs during pregnancy. A case series of treatment of pregnant females with amikacin, ofloxacin, prothionamide, cycloserine and pyrazinamide did not show any harm to the children or their mothers [233]. In contrast, streptomycin and kanamycin caused ototoxicity in the children [234]. Hence, these drugs and other aminoglycosides/polypeptides are not recommended for treatment during pregnancy. Fluoroquinolones have caused arthropathy in animals but have been safely used in pregnant women [235]. Ethionamide has been implicated in congenital defects in animals [236], although human data are not available [232].

Breastfeeding should be recommended only in females who are not infectious. In general, although drugs are found in the breast milk, the benefits of breastfeeding of children with mothers on MDR/XDR-TB treatment outweigh the potential harm related to ingestion of TB drugs through breastfeeding [237].

16 Special considerations for MDR/XDR-TB and HIV co-infection

The interaction between TB and HIV infection is characterised by a high probability of rapid progression to primary TB shortly after M. tuberculosis infection and from LTBI to overt disease in persons with pre-existing HIV infection [238, 239].

Updated global drug-resistant TB surveillance data show that HIV-infected patients are more likely than HIV-uninfected cases to have MDR/XDR-TB, but this association is not statistically significant (pooled OR 1.4, 95% CI 0.7–3.0; p=0.19) and not consistent throughout the different regions surveyed [240]. In the future, eastern Europe may witness a large increase in the number of co-infected patients as it includes 15 out of 27 high MDR-TB burden countries [241] and, in contrast to other regions, both HIV incidence and AIDS-related deaths continue to increase [242]. In this region, it is of paramount importance that integrated TB, HIV, and drug substitution and needle exchange programmes are strengthened for civil society and the penitentiary system, in concert with harm- and stigma-reduction policies. Expansion of HIV testing is also a high priority in the region and worldwide.

There is firm evidence that HIV-infected persons with MDR/XDR-TB have higher mortality rates compared with HIV-seronegative cases. High mortality with extremely short survival was reported from South Africa [243, 244] among MDR/XDR-TB and HIV co-infected cases with very low CD4 cell counts and limited access to antiretroviral therapy (ART). HIV infection was an independent predictor of death (relative risk 4.22, 95% CI 2.65–6.72) in patients treated with SLDs in Estonia, Latvia, the Philippines, Russia and Peru [245]. In Europe, MDR-TB cases had a poorer treatment outcome compared with non-MDR-TB among HIV-infected persons (27% versus 44%) [246]. However, recent reports from Latvia [247] and Russia [248] did not identify HIV as a predictor of death among MDR-TB cases, suggesting that other factors may be more important as determinants of TB treatment outcome in specific settings.

Timely diagnosis of MDR/XDR-TB among HIV-infected persons is crucial to reduce mortality: conventional DST methods have a turnaround time of a few weeks, which is unsuitable for immediate action. In HIV-infected patients, molecular tests allow a rate of case detection that is increased by 45%, as compared with smear microscopy [249]. Xpert MTB/RIF is a rapid point-of-care molecular assay that is sensitive and specific for detection of TB when it is used as an initial diagnostic test in patients suspected of having HIV-associated TB [250]. WHO recommends Xpert MTB/RIF as a primary diagnostic test for TB in persons living with AIDS: the test should be made widely available in HIV-prevalent settings to improve patient care [251]. Rapid point-of-care diagnostic tests such as molecular assays should primarily be made available to HIV-infected communities and individuals [252].

The role of ART in reducing TB incidence, recurrence and mortality is well established. Limited data are available regarding ART initiation in MDR/XDR-TB patients. Evidence from 10 studies (including two European cohorts [11, 247]) of HIV-infected patients using SLDs suggests a lower risk of death and a higher likelihood of TB cure in patients who were treated with ART versus those who were not [10]. Based on these data, WHO recommends that, among patients with MDR/XDR-TB, ART should be initiated as soon as possible (within the first 8 weeks) following anti-TB treatment initiation, irrespective of CD4 cell-count [241]. It may be speculated that, as demonstrated for TB/HIV co-infected persons with drug-susceptible M. tuberculosis, persons with CD4⁺ T-cell counts of <50 per cubic millimetre should start ART within 2 weeks of starting TB treatment in order to prevent new AIDS-defining illnesses and death [253, 254].

Systematic assessment of serious adverse events, treatment adherence, drug interactions and the incidence of immune reconstitution inflammation syndrome associated with SLDs in HIV-infected patients with MDR/XDR-TB is rather scarce and needs to be improved. Drug absorption in HIV-infected patients can be significantly reduced. Drug concentration measurement can be performed to assure drug exposure (table 4).

While the evidence of the importance of acquired resistance among HIV-infected persons is elusive, the risk of primary disease after infection with drug-resistant strains is substantial, as demonstrated by the outbreaks of MDR-TB among hospitalised HIV-infected patients and healthcare workers in Europe and the USA [238]. Strategies for infection control based on administrative measures, environmental control and respiratory protection should become a priority in countries where both MDR/XDR-TB and HIV are prevalent. HIV-infected and −uninfected TB cases should be managed to similar standards, based on the principle that adequate adherence to standard regimens will prevent drug resistance in both situations [26].

17 Special considerations for patient management in areas of high MDR/XDR-TB prevalence

MDR/XDR-TB increasingly occurs in economically disadvantaged regions of Europe [7]. Countries with a high burden of MDR/XDR-TB share common features of health systems that are not capable of coping with the challenges related to the emergence of drug-resistant TB, such as poverty and insufficient resources, inadequate technical equipment, shortage of medical education and training, lack of community support, and failing political commitment [255].

The high frequency of MDR-TB among new laboratory-confirmed cases in some countries of the WHO European region indicates active ongoing transmission of drug-resistant strains of M. tuberculosis [7]. Infection control measures to prevent the spread of MDR/XDR M. tuberculosis, especially in congregate settings like prisons and hospitals, are likely to be insufficient. The extent of possible nosocomial transmission is unclear, as molecular surveillance of M. tuberculosis strain transmission is not established in the high-burden countries. Exposure to patients who are infectious can be minimised by adequate infection control measures including reducing the time of hospitalisation, reducing the number of outpatient visits, avoiding overcrowding in wards and waiting areas, and prioritising community-care approaches for TB management, where possible [66, 256–261].

Delayed case detection and lack of access to rapid and accurate diagnosis of TB and M. tuberculosis DST are major obstacles for TB control in countries with high MDR/XDR-TB prevalence. Only 37% of TB cases in the 18 high-priority countries in the European region are confirmed by culture and can be further evaluated by conventional DST [7]. This highlights the enormous gap in the identification of patients with MDR/XDR-TB in the European Region. Knowledge of M. tuberculosis drug susceptibility/resistance pattern, particularly to SLDs, is especially important for the management of patients with MDR/XDR-TB in Europe, due to the increasing frequency of M. tuberculosis strains resistance to SLDs [36]. Scale-up for universal availability of diagnostic methods for the rapid TB identification and DST by molecular methods is an utmost priority to combat MDR/XDR-TB in Europe [262]. Moreover, universal access to standard (conventional/culture-based) first- and second-line DST is still essential as the current generation of molecular tests does not identify all relevant mutations that lead to phenotypic drug resistance [262].

Access to treatment, availability of quality-assured SLDs throughout therapy and continuous treatment monitoring are often lacking in countries with a high prevalence of MDR/XDR-TB in Europe. At present, only ∼20% of all estimated cases with MDR/XDR-TB in the European region receive adequate treatment with quality-assured drugs [16]. Although the Green Light Committee (GLC) and the Global Drug Facility are providing very important support for the sustained provision of essential anti-TB drugs in priority countries, political willingness of the countries to provide their own resources is essential and often insufficient. In addition, access to new drugs like linezolid, carbapenems, bedaquiline and clofazimine is not yet ensured by the GLC mechanism, although it is intended [263]. Availability of SLDs from quality-assured suppliers, regulated access and local governments must ensure distribution of drugs.

Default from treatment is a major problem in countries of high MDR/XDR-TB prevalence. Economic needs (e.g. migration for work) often overrule adherence and jeopardise treatment success.

Necessary interventions to address the enormous challenges for the prevention and management of MDR/XDR-TB, especially in the high-burden countries of the WHO European Region are outlined in the Roadmap to Prevent and Combat Drug-Resistant Tuberculosis [16].

18 MDR/XDR-TB consilia

Due to limited experience in the care of patients with MDR/XDR-TB even by infectious diseases or respiratory diseases specialists in many countries of the European region, there is a danger that these patients are managed in a less than optimum way. Sharing information and discussing clinical management among experts together with the caregivers in charge ensure better care for these patients.

In some European countries, there is a single consilium or committee to which all MDR/XDR-TB cases are reported and regularly discussed among experienced physicians, microbiologists and nurses. Data on drug-sensitivity patterns are used to plan which drugs will be used in the regimen. As treatment progresses and new drug resistance patterns emerge or patients appear to be failing treatment, alternative regimens can be devised and implemented.

In this way MDR/XDR-TB patients are managed only by a small group of experts within the country, who can share their knowledge and build on it with further experience.

Regular meetings in a face-to-face consilium are not feasible for larger European countries. In this case, use of electronic information transfer for a centrally coordinated advisory service enables a small number of experts to give advice on patients about whom the service had been informed and to discuss the progress of each patient as required by clinical need. By central coordination, the aim of the service is to ensure that the best advice is given for every MDR/XDR-TB case in the country. A successful model for an electronic consilium has been established by the British Thoracic Society in cooperation with other relevant professional bodies [153], which is now being adapted in other countries. In 2012, WHO and ERS launched the ERS/WHO electronic consilium [160].

19 Palliative care of patients with MDR/XDR-TB

MDR/XDR-TB is a life-threatening condition in which most patients experience major physical and emotional suffering resulting from the disease and the treatment itself. This suffering does not always end with the end of successful treatment, as the disease quite often leaves long-term sequelae. Yet major suffering goes unnoticed in MDR/XDR-TB patients in whom all treatment options are exhausted. Although there are almost no systematic representative data describing the fate of these patients, clinical experience indicates in XDR-TB patients, there is high mortality within months, especially when they are co-infected with HIV, not on ART or within a few years after treatment failure is declared [264, 265].

As part of a comprehensive response to the epidemic of MDR/XDR-TB, there is an ethical obligation to alleviate suffering in patients through palliative care services. This has been increasingly recognised by TB and palliative care experts [266, 267].

Palliative care is defined as “an approach that improves the quality of life of patients and their families facing the problem associated with life-threatening illness, through the prevention and relief of suffering by means of early identification and impeccable assessment and treatment of pain and other problems, physical, psychosocial and spiritual” [268].

Data from the WHO Global TB Report 2011 suggest that the TB epidemic in several countries of eastern Europe is slowly being replaced by one of MDR-TB [266]. In 2011, the treatment success rate in the WHO European Region was only 48.5%. This confirms the massive, largely unaddressed need for delivery of palliative care and end-of-life care to MDR/XDR-TB patients.

Existing expertise in palliative care should be sought and applied to MDR/XDR-TB cases. New expertise needs to be developed for carers for MDR/XDR-TB patients. TB infection control is a true challenge for patients and carers. General principles have to be applied. There is no consensus about continued drug treatment after declaration of treatment failure with no further options of cure.

Practical guidelines for palliative care in MDR/XDR- TB are available in South Africa [269].

In the Consolidated Action Plan to Prevent and Combat Multidrug- and Extensively Drug-Resistant Tuberculosis in the WHO European Region 2011–2015 [16], countries committed to establish, and strengthen as needed, palliative care mechanisms for MDR/XDR-TB patients who fail treatment, and to conduct operational research on service models (needs of patients, cost and resources) for provision of palliative care. Technical assistance on designing and running palliative care services in TB control programmes have to be urgently intensified.

20 Consensus statements

21 Priorities for improvements of the management of patients with MDR/XDR-TB

As is evident from the previous chapters, improvements in several areas, including basic research, capacity building and education, prevention, diagnosis, and treatment, are necessary to achieve the goal of TB eradication in times of emerging M. tuberculosis drug resistance.