Article Text


Survival of drug resistant tuberculosis patients in Lithuania: retrospective national cohort study
  1. Yanina Balabanova1,2,
  2. Birute Radiulyte3,
  3. Edita Davidaviciene3,
  4. Richard Hooper1,
  5. Olga Ignatyeva2,
  6. Vladyslav Nikolayevskyy1,
  7. Francis A Drobniewski1
  1. 1Blizard Institute, Queen Mary University of London, London, UK
  2. 2Samara Oblast Tuberculosis Dispensary, Samara, Russia
  3. 3National Tuberculosis and Infectious Diseases University Hospital, Vilnius, Lithuania
  1. Correspondence to Dr Yanina Balabanova; y.balabanova{at}


Objective To establish risk factors influencing survival of patients with multidrug-resistant and extensively drug-resistant tuberculosis (MDR/XDRTB).

Design All MDR/XDRTB cases (n=1809) reported from 2002 to 2008 in Lithuania with a known outcome were included in the survival analysis.

Results Median survival for MDRTB and XDRTB patients was 4.1 (95% CI 3.7 to 4.4) and 2.9 (95% CI 2.2 to 3.9) years. In a multivariable analysis adjusting for other patient characteristics, the difference in survival between MDRTB and XDRTB patients was not significant (HR=1.29 (0.91 to 1.81)). Older age (HR=4.80 (3.16 to 7.29)) for 60+ vs <30 years, rural living (HR=1.20 (1.02 to 1.40)), alcohol use (HR=1.49 (1.13 to 1.96)) for alcoholic versus moderate use, unemployment (HR=1.79 (1.31 to 2.46)), lower education levels (HR=1.50 (1.08 to 2.07)) for primary level versus tertiary level, cavitary disease (HR=1.54 (1.29 to 1.83)) and being smear positive at the time of MDR/XDRTB diagnosis (HR=1.47 (1.19 to 1.82)) were associated with poorer survival. HIV positivity significantly affected survival (HR=3.44 (1.92 to 6.19)) for HIV positive versus HIV negative; HR=1.60 (1.28 to 2.01) for HIV not tested versus HIV negative). There was no difference in survival of patients who acquired MDR/XDRTB during treatment compared with patients with primary MDR/XDRTB (HR=1.01 (0.85 to 1.19)). Treatment with a second-line drug improved survival (HR=0.40 (0.34 to 0.47)). In a subgroup with genotyped TB strains, a Beijing family of strains was associated with poorer survival (HR=1.71 (1.19 to 2.47)).

Conclusions Social factors, rural living, HIV infection and Beijing strain family impact on survival. Survival of MDR/XDRTB patients is short. Rapid drug resistance identification, early administration of appropriate treatment and achieving high cure rates, expansion of HIV testing and antiretroviral treatment are necessary for optimal management of MDR/XDRTB.

Statistics from

Article summary

Article focus

  • The Baltic States have consistently had one of the world's highest rates of drug resistant tuberculosis (TB) including multidrug- and extensively drug-resistant tuberculosis (MDR/XDRTB) despite an extensive tuberculosis control programmes.

  • Patients with TB in Lithuania show a relatively high mortality for unknown reasons.

  • The study analyses factors influencing survival of patients with MDR/XDRTB in Lithuania.

Key messages

  • Patients with MDR/XDRTB show a poor survival regardless of HIV status, but there was no significant difference between MDRTB and XDRTB.

  • Patients infected with Beijing TB family strains and co-infected with HIV with underlying social problems have worse survival. However, only few TB patients are tested for HIV.

  • Addressing accompanying social and health problems (eg, alcohol dependency), access to care in rural settings and expansion of HIV testing and antiretroviral treatment are necessary to improve survival.

Strength and limitations of the study

  • This is the first national study to analyse 7 years of national surveillance data covering a large cohort of MDR/XDRTB patients.

  • The study demonstrates the value of long-term survival cohorts, as well as pointing to the absence of comparable UK data.

  • Limitations: limited data on HIV status (due to the low HIV testing coverage in the early years of the TB programme) and limited genotyping data for the XDRTB isolates.


Multidrug-resistant (MDR) and extensively drug-resistant (XDR) tuberculosis (TB) threaten effective treatment and undermine global efforts towards elimination of TB.1 2 The Baltic States have consistently had one of the world's highest rates of drug resistance including MDRTB; MDRTB rates for 2009 were 62.1, 43.2 and 30.7/100 000 for Lithuania, Latvia and Estonia, respectively.3 Our knowledge of drug resistance epidemiology and its impact on patients' population has come principally from studies from Estonia and Latvia4–7 but relatively little is known about Lithuania.8–11 However, Lithuania is one of the 18 high priority TB countries in the European Region, has seen an increase in rates of both primary and acquired MDRTB (corresponding rates were 9% and 50% in 2010, respectively) and appearance of XDRTB cases that in 2010 constituted 4.3% of all MDRTB cases (figure 1).12 Drug resistance is accompanied by high rates of default (around 30%) and low treatment success rates (40% in newly diagnosed and 19% in retreatment cases in 2009) among MDRTB patients despite a well-established TB control programme with relatively good indicators of treatment success and low default rates (7%) among patients with sensitive TB.12 Mortality rates among newly diagnosed culture-confirmed cases is high (10.3%) and is one of the highest in retreatment TB cases in the WHO European Region (22.3%); the explanation for the high mortality remains unknown.12

Figure 1

Trends in drug-resistant cases of tuberculosis in Lithuania, 2002–2010.

To answer this question, we analysed 7 years (2002–2008) of Lithuanian national tuberculosis surveillance data. Our aims were to describe the epidemiological, clinical and socioeconomic features and survival of a large national cohort of MDR/XDRTB cases and to establish risk factors influencing their survival.

Material and methods

Study population and data sources

The analysis was based on all MDRTB and XDRTB cases including new and retreatment cases, confirmed by conventional microbiological drug susceptibility testing (DST) methods and registered for treatment from 2002 to 2008. The National TB Register (of the National TB Surveillance system) was created in 2002 and used as the source of data.

Standard case reporting includes demographic and clinical information along with initial and follow-up DST. The initial DST is performed on the first positive mycobacterial culture for all cases; a follow-up test is repeated if treatment failure or developed drug resistance is suspected on a specimen collected at least 30 days after the initial specimen. DST is performed using solid or automated liquid culture media system (BACTEC MGIT 960; Becton Dickinson, Sparks, Maryland, USA) according to standard procedures.13

A proportion of strains (all available MDR/XDRTB isolates during 2004–2006 or approximately 20% of the total with MDR/XDRTB) were genotyped (by IS6110 restriction fragment length polymorphism typing and spoligotyping)14 15 by the Lithuanian Institute of Biotechnology.

Case definitions

An MDRTB case was defined as a person infected with an isolate resistant to at least isoniazid (INH) and rifampicin (RIF), excluding cases confirmed to have XDRTB. MDRTB cases that had resistance to a fluoroquinolone (FQ) and a second-line injectable (INJ, not streptomycin) drug were defined as XDRTB cases. If an MDR or XDRTB patient was registered for treatment and had never received TB treatment in the past for longer than 4 weeks, he/she was considered to have primary drug resistance. MDR/XDRTB patients who were retreatment cases with a first episode of TB before 2002 were assumed to have acquired TB.

Main outcome measures and statistical analysis

We used Kaplan–Meier survival curves and multivariable Cox regression to analyse time until death from any cause during patient's treatment or follow-up, from the time of the first-recorded diagnosis of MDR or XDRTB in the database. Patients who died or defaulted before a diagnosis of TB was established were excluded from the analysis. Cases from recent years who had not completed their therapy or were still being followed up were censored in April 2010 when the analysis was conducted. The date of the last visit to a TB clinic was recorded as the last day when a patient was documented to be alive. Factors analysed in relation to survival were MDR versus XDRTB and primary versus acquired TB, as well as other characteristics assessed at the first TB diagnosis: sex, age, rural/urban residence, contact with TB, smoking, alcohol use, drug abuse, homelessness, unemployment, education level, HIV status, co-morbidity, TB type, smear positivity and cavitary disease. Only patients with complete data were included in the main analysis. (Note that this includes patients in three categories of HIV status: positive, negative or not tested. In a subsidiary analysis, we looked at the effect on survival of TB strain family (Beijing/non-Beijing) in the subsample of patients with genotyped strains, adjusting for the variables in the main analysis.) The effect of treatment with second-line drugs (SLDs; in the treatment cycle where MDR/XDRTB was first diagnosed) was modelled in a number of ways: we first looked for a trend according to the number of SLDs used and compared this with a model which compared any second-line treatment with none. We also looked at effects of individual drugs and of particular drug combinations, adjusting for all variables in the main analysis. In addition to comparing the survival of XDR/MDRTB cases, we assessed the effect of different resistance patterns at the first TB diagnosis on survival, adjusting for the same variables as in the analysis of MDR versus XDRTB. For this analysis, patients were divided into eight resistance patterns: (1) INH+RIF only, (2) INH+RIF+Ofloxacin (‘Oflox’), (3) INH+RIF+INJ, (4) INH+RIF+Ethambutol/Prothionamide (‘ETH/PT’), (5) INH+RIF+Oflox+INJ (‘XDR’), (6) INH+RIF+Oflox+INJ+ETH/PT (‘XDR’), (7) INH+RIF+Oflox+ETH/PT and (8) INH+RIF+INJ+ETH/PT.

All analysis was performed using Stata V.11 (Stata Corporation).

Ethics review

The project was reviewed by the Vilnius Regional Committee for Biomedical Research Ethics, Vilnius University and Queen Mary College Research Ethics Committee and received a waver of informed consent as anonymised data were used.


Study population

There were 1841 patients in the database with a diagnosis of MDR or XDRTB. Twenty-five were diagnosed after the patient had died or defaulted and were excluded. The last date of follow-up was unknown for a further seven patients, and two patients had missing data on smear positivity at the time of MDR/XDRTB diagnosis. The analysis was done on the 1807 (98%) patients with complete data.

Socio-demographic and clinical characteristics

The majority of MDR/XDRTB patients were young Lithuanian-born males aged between 30 and 49 years living in urban settings, unemployed, with primary or secondary education and frequently consuming alcohol (table 1). Drug abuse was relatively uncommon. Most patients had pulmonary disease and were smear positive at diagnosis. Over 60% of patients had extensive lung damage with cavities identified on radiographs. Relatively few patients (17.2%) were tested for HIV infection as testing is not imposed by national policies (although usually more MDR/XDRTB patients are offered HIV testing than non-MDRTB patients).

Table 1

Socio-demographic and clinical characteristic of MDR/XDRTB cases

A proportion of MDRTB patients (13%) were resistant to other SLDs in addition to resistance to isoniazid and rifampicin but not meeting the definition of XDR.

Analysis of treatment regimes showed that 62.0% of patients with MDRTB and 40.8% of patients with XDRTB received fluoroquinolones and 34.9% and 35.2%, respectively, received injectable SLDs. Ethionamide/prothionamide was administered to over half of MDR/XDRTB patients and cycloserine (CS) to 42.3% and 52.1% of MDRTB and XDRTB patients, respectively. Negligible proportions of patients received terizidone (TRD), amoxicillin/clavulanate (AMC) or thiacetazone (THZ). P-aminosalicyclic acid (PAS) was administered to 1.3% of MDRTB and 28.2% of XDRTB patients (table 1). The most commonly administered drug combinations were FQ combined with group 4 drugs (ETH/PT, PAS, terizidone (TRD)), or cycloserine and FQ, injectables and group 4 drugs. For treatment of primary XDRTB in addition to these regimens, a combination of injectables and group 4 drugs was used in 15.5% of patients.

Survival analysis

The 1807 patients were followed for a total of 4089.3 person-years. Figure 2 shows Kaplan–Meier plots of the probability of survival depending on resistance and HIV status. Median survival for MDR and XDRTB patients was 4.0 (95% CI 3.7 to 4.4) and 2.9 (95% CI 2.2 to 4.3) years, respectively, and for HIV positive versus HIV negative was 1.9 (95% CI 0.4 to 3.5) and 4.9 (95% CI 4.3 to 6.8) years, respectively. Median survival of patients with primary and acquired MDRTB was 4.2 (95% CI 3.7 to 5.1) and 3.7 (95% CI 3.4 to 4.3) years, respectively; it was 2.7 (95% CI 1.8 to no upper limit) and 2.9 (95% CI 1.4 to 4.9) years for primary and acquired XDRTB patients, respectively.

Figure 2

Survival among (A) multidrug-resistant (MDR) versus extensively drug-resistant tuberculosis (XDRTB) patients; (B) MDR/XDRTB HIV-positive versus MDR/XDRTB HIV-negative cases.

Table 2 shows results of the multivariable analysis of survival. Number of SLDs was associated with survival after adjusting for other patient characteristics (HR per drug 0.77, 95% CI 0.73 to 0.81, p<0.001). However, there was no trend over and above the simple effect of receiving any SLD treatment (mutually adjusted HR: any treatment vs no treatment 0.42, 95% CI 0.29 to 0.59, p<0.001; per drug 0.98, 95% CI 0.88 to 1.10, p=0.78), so the results in table 2 adjust only for the effect of any treatment. Older age, rural living, alcohol use, unemployment, lower levels of education, positive or unknown HIV status, cavity disease and being smear positive at the time of MDR/XDRTB diagnosis were all independently associated with poorer survival (table 2). Once other patient characteristics were adjusted for, there was no association of survival either with acquired versus primary or with XDRTB versus MDRTB. In the subsample of patients with genotyped TB strains (n=306), there was evidence that a Beijing strain was associated with poorer survival after adjusting for other factors (HR 1.70, 95% CI 1.18 to 2.45, p<0.004).

Table 2

Factors influencing survival of MDR/XDRTB patients*

We compared survival in eight different patient subgroups defined by their patterns of resistance to SLDs (see Methods). There was no evidence of differences between these patient subgroups (p=0.20).

We also looked at effects of individual drugs and of individual drug combinations (table 3). Treatment with fluoroquinolones, injectables, ethionamide or prothionamide and cycloserine as well as with combinations of these drug groups was associated with better survival.

Table 3

Effect of different SLDs and their combination on survival of MDR/XDRTB patients*


This is the first national study in Lithuania to examine long-term survival in a large cohort of patients with MDR/XDRTB. It complements smaller studies in other Baltic States but with a longer follow-up period.5–7 16

Most MDR/XDRTB patients in Lithuania were young men with accompanying social problems. These findings describe a common distinct profile of TB patients in other Eastern European regions5 7 16–21 drawing attention to the high social marginalisation of this group.

Median survival of MDRTB and XDRTB patients was around 4 and 3 years, respectively (similar to findings from other settings).22 Although other studies demonstrated a significant impact of XDR on survival,5 22–27 in our study, median survival of XDRTB patients was not significantly shorter than survival of MDRTB patients. This finding might reflect the overwhelming effect of MDRTB on survival and highlights the fact that many patients although not XDRTB were ‘MDRTB plus’ with resistance to many other SLDs. In these setting, the difference between MDR and XDRTB might be less prominent than elsewhere. Individuals with lower survival rates were more likely to be older, alcohol consuming, unemployed rural-based individuals who had cavities in their lungs and remained sputum smear positive. The association between smear positivity and lethality is in line with poorer treatment outcome in patients reported from Estonia.5

Extremely high rates of the Beijing TB strain family among MDRTB cases support findings from other studies in Eastern Europe including Russia,27–32 Estonia33 34 and Latvia.35 36 However, this study also presented evidence that these strains are independently associated with worse survival even after adjusting for the effect of other characteristics.

HIV infection was associated with lower survival; only half survived for up to 1.9 years from MDR/XDRTB diagnosis. Unlike other Baltic States, this largely Catholic country does not routinely offer HIV testing to TB patients; initiation of testing would be of value particularly for patients shown to have MDR/XDRTB where highly active antiretroviral therapy would be of as much importance for survival as anti-TB drugs.

There was no difference in survival of patients with primary MDR or XDRTB compared with those who developed drug resistance during treatment highlighting the overwhelming effect of drug resistance on life expectancy.

Although the total number of MDR/XDRTB patients receiving currently recommended treatment is relatively small,37 38 it reflects the fact that as at the beginning of the study treatment, guidelines were not developed and SLDs not widely available. At the same time, treatment regimens that include ethionamide/prothionamide and cycloserine were commonly administered for several months up to a year. However, the prescription of any SLDs (even singly for ofloxacin, injectables, ethionamide/prothionamide, cycloserine) was associated with better survival. Taken together, the results supported the importance of ofloxacin (and so presumably FQ treatment in general) and injectable agents in improving survival in line with the findings of studies including a meta-analysis25 showing an improved survival for XDRTB patients who received late-generation FQs. Particularly, interesting was the importance of ethionamide/prothionamide therapy used in combinations with ofloxacin and/or injectable agents. A combination of ethionamide/prothionamide with ofloxacin appeared to be at least as effective as an FQ with an injectable. Although methods for DST of ethionamide/prothionamide exist,39 40 the ability to reliably demonstrate resistance remains difficult and it is reasonable to argue that prothionamide should always be added to an MDR/XDRTB treatment regimen regardless of DST data. This provides support for the current WHO combination drug class approach.38

Even though survival time of MDR/XDRTB patients is relatively short, it is long enough to establish a large pool of individuals potentially infectious for others and facilitate further transmission of drug-resistant strains in the community and in hospital settings where patients spend up to 3 months. Early detection of MDR/XDRTB with better infection control is therefore vital to interrupt further transmission. Identification of MDRTB patients should lead to their isolation with a regime consisting of ethambutol, pyrazinamide, FQ, injectable and ethionamide/prothionamide.

The strong association of drug resistance and worse survival with social factors including alcohol abuse and high rates of default among MDR/XDRTB cases are in line with the findings of others41 and emphasise an urgent need for non-medical interventions to improve treatment outcomes. When this has been introduced (eg, treatment of accompanying alcohol dependency), it has improved treatment adherence and outcomes significantly.42–44 The independent impact of rural living on survival may indicate possible obstacles in accessing TB treatment facilities in the country despite a well-established system of TB care; the issue warrants further investigation by local agencies.

The study has some limitations. HIV status was not known for the majority of TB cases and therefore we were unable to investigate further the association between HIV, drug resistance and survival. Genotyping was done for a relatively limited number of strains; however, it provided sufficient data to identify statistically significant association at least for a group of MDRTB patients. Larger genetic studies are needed to answer the question on influence of the strain type on survival among XDRTB patients. Nevertheless, despite these limitations, the study results can be generalised at least for the Eastern European countries with similarly high levels of TB, drug resistance and similar profile of patients.

In conclusion, rapid identification of drug resistance, early administration of appropriate treatment, achievement of high cure rates, adequate infection control measures, expansion of HIV testing and antiretroviral treatment are necessary to improve patients' survival and prevent further spread of MDR and XDRTB in Lithuania.


View Abstract
  • Supplementary Data

    This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

    Files in this Data Supplement:


  • To cite: Balabanova Y, Radiulyte B, Davidaviciene E, et al. Survival of drug resistant tuberculosis patients in Lithuania: retrospective national cohort study. BMJ Open 2011;1:e000351. doi:10.1136/bmjopen-2011-000351

  • Funding The research leading to these results has received funding from the European Community's Seventh Framework Programme (FP7/2007-2013) under grant agreement FP7-223681. The funders had no role in the design or analysis of the study.

  • Competing interests None.

  • Ethics approval Ethics approval was provided by Vilnius Regional Committee for Biomedical Research Ethics, Vilnius University and Queen Mary College Research Ethics Committee.

  • Contributors YB, FAD, OI, VN and RH designed and implemented the study; BR and ED assisted with data collection. RH performed the statistical analysis; YB and FAD drafted the paper. All authors contributed to drafts and approved the final draft of the manuscript. All authors, external and internal, had full access to all of the data (including statistical reports and tables) in the study and can take responsibility for the integrity of the data and the accuracy of the data analysis.

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Data sharing statement Technical appendix, statistical code and partial dataset (anonymised, without any patients-related identifiers) available from the corresponding author at y.balabanova{at} Informed consent was not obtained from the study participants as the data present routinely collected surveillance data and the risk of identification is low.

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.