Abstract
BACKGROUND
Objective of this analysis was to identify predictors of death, failure, and default among MDR-TB patients treated with second-line drugs in DOTS-plus projects in Estonia, Latvia, Philippines, Russia, and Peru, 2000–2004.
METHODS
Risk ratios (RR) with 95% confidence intervals (CI) were calculated using multivariable regression.
RESULTS
Of 1,768 patients, treatment outcomes were: cure/completed – 1,156 (65%), died – 200 (11%), default - 241 (14%), failure - 118 (7%). Independent predictors of death included: age>45 years (RR=1.90 (95%CI 1.29–2.80), HIV infection (RR=4.22 (2.65–6.72)), extrapulmonary disease (RR=1.54 (1.04–2.26)), BMI<18.5 (RR=2.71 (1.91–3.85)), previous use of fluoroquinolones (RR=1.91 (1.31–2.78)), resistance to any thioamide (RR=1.59 (1.14–2.22)), baseline positive smear (RR=2.22 ()1.60–3.10), no culture conversion by 3rd month of treatment (RR=1.69 (1.19–2.41)); failure: cavitary disease (RR=1.73 (1.07–2.80)), resistance to any fluoroquinolone (RR=2.73 (1.71–4.37)) and any thioamide (RR=1.62 (1.12–2.34)), and no culture conversion by 3rd month (RR=5.84 (3.02–11.27)); default: unemployment (RR=1.50 (1.12–2.01)), homelessness (RR=1.52 (1.00–2.31)), imprisonment (RR=1.86 (1.42–2.45)), alcohol abuse (RR=1.60 (1.18–2.16)), and baseline positive smear (RR=1.35 (1.07–1.71)).
CONCLUSION
Patients with biomedical risk factors for treatment failure or death should receive heightened medical attention. To prevent treatment default, management of patients who are unemployed, homeless, alcoholic, or have a prison history requires extra measures to insure treatment completion.
Keywords: multidrug-resistant tuberculosis, treatment outcomes, risk factors
Introduction
Modern chemotherapy of drug-susceptible tuberculosis (TB) has >95% efficacy in randomized, controlled clinical trials, and the new World Health Organization (WHO) target for treatment success rates for TB Control programs is 90% by 2015.1 Treatment of multidrug-resistant (MDR) TB (i.e., TB caused by strains of Mycobacterium tuberculosis resistant to at least isoniazid and rifampicin) has far worse outcomes as summarized in two recent systematic reviews and meta-analyses.2, 3 Treatment of MDR-TB is more expensive, longer, and more toxic compared with standardized short-course chemotherapy for drug-susceptible TB.4
In 2000, the Green Light Committee (GLC) was formed to increase access to quality-assured second-line drugs (SLDs) at reduced prices while at the same time preventing the emergence of further drug resistance. From February 1999 to August 2001 the first five DOTS-plus projects approved by the GLC started treating patients, including Estonia; Latvia; Makati Medical Center, Manila, Philippines; the Harvard/Partners in Health Project in northern Lima, Peru; and Tomsk Oblast, Russia.5, 6 In an effort to improve MDR-TB treatment outcomes we analyzed data from these projects focusing on early predictors of poor treatment outcomes that should trigger timely clinical re-assessment and further action.
Methods
Study Design and Patient Population
This retrospective cohort included adult patients with MDR-TB starting treatment between 01/01/2000–12/31/2003 in five GLC-approved DOTS-Plus programs.6 The majority of patients in Russia, Latvia, and Estonia were hospitalized, while most patients in Peru and the Philippines received treatment as outpatients. Patients received individualized treatment regimens including 5–7 drugs (typically including a fluoroquinolone and injectable second-line drug) to ensure at least four drugs were effective, in accordance with the GLC-approved protocols.7 The study was approved by the U.S. Centers for Disease Control and Prevention Institutional Review Board (IRB) and IRBs at all participating sites.
Definitions
Treatment outcomes were programmatically assigned according to each site’s criteria. Consensus definitions were developed from 2000–2002 by an international panel of experts including representatives of these five projects and were used in GLC-approved pilot projects before their formal publication in 2005.8 Positive smear, positive culture, and culture conversion were defined according to WHO guidelines.4 Baseline drug resistance was determined by drug susceptibility tests (DST) in the local laboratory from sputum collected between 60 days prior to 30 days after MDR-TB treatment initiation. Extensively drug-resistant (XDR) TB and SLDs groups were defined according WHO guidelines.4
Data Collection and Statistical Analysis
The database was compiled from existing electronic medical records from five countries. Common variables were identified, common data definitions were operationalized, and disparate data formats were harmonized as previously described6. Statistical analyses were performed using SAS software, version 9.1 (SAS Institute Inc., Cary, NC). Risk ratios (RR) with 95% confidence intervals (CI) for each of the poor outcomes (death, failure, and default) were calculated using log-binomial regression. Factors significant at a value of P≤.05 in the univariate analysis as well as factors with plausible epidemiological or biological association with each outcome of interest were included in a multivariable log-binomial regression model. Confounding and interaction were assessed. Backward selection was used to choose prognostic variables for final models. A P value of ≤0.05 was considered statistically significant.
Results
A total of 1,768 patients were included in analysis: 663 (37.5%) from Peru, 444 (25.1%) from Latvia, 280 (15.8%) from Estonia, 221 (12.5%) from Russia and 160 (9.1%) from the Philippines. The majority of patients were males (70%), median age was 36 years, and 1.6% were HIV-infected. Socio-demographic and clinical characteristics of this cohort were previously reported by Gammino et al.6 Pre-treatment isolates were resistant to a median and mean of 5 (interquartile range [IQR] 4–6; range 2–11) first- and second-line anti-TB drugs at the start of treatment. A total of 57 (4.7%) of 1,221 patients with available DST result for at least one injectable SLD and fluoroquinolone had extensively drug resistant tuberculosis at the start of treatment (ranged between sites from 2.9% to 7.9%).
Overall 92.9% (756 of 832 patients with available information) had at least one adverse event reported, with a median of 2 (mean 2.4±1.5; range 0–8) adverse event types per patient (Table 1). The most commonly recorded adverse events of treatment with second-line drugs included nausea (n=540, 65.6%), diarrhea (n=249, 30.3%), and ototoxicity/hearing disturbances (n=130, 15.8%).
Table 1.
Adverse events on treatment with second-line drugs (N=823)
| Adverse event | Number (percent*) |
|---|---|
| Nausea | 540 (65.6) |
| Diarrhea | 249 (30.3) |
| Ototoxicity (hearing disturbances) | 130 (15.8) |
| Headache | 95/602** (11.5) |
| Hypokalemia | 99/221** (14.9) |
| Peripheral neuropathy | 76 (9.2) |
| Hypothyroidism | 40 (4.9) |
| Seizure | 39 (4.7) |
| Psychosis | 35/379** (4.3) |
| Hepatitis | 19 (2.3) |
| Renal insufficiency (elevated creatinine) | 15/602** (1.8) |
Percentage from total 823 unless states otherwise.
Denominator is number of patients with available data on particular adverse event.
Of 1,768 patients, treatment outcomes were: cure/completed – 1,156 (65.4%), died – 200 (11.3%), failure - 118 (6.7%), default - 241 (13.6%), transferred/unknown - 53 (3.0%). The exact outcome dates were not available, but using the length of microbiological monitoring as a proxy for the length of treatment, patients in whom treatment was successful were on treatment for median of 21 months (IQR 17–25), patients who died - 8.5 months (IQR 2–16), patients who defaulted treatment – 8 months (IQR 3–15), and those in whom treatment failed – 13 months (IQR 7–23). Of 57 patients with XDR TB, treatment outcomes were: 31 (54.4%) patients had cure/completed treatment, 10 (17.5%) died, 13 (22.8%) failed, and 3 (5.3%) defaulted.
Results of univariate and multivariable analysis of predictors of poor treatment outcomes are shown in Tables 2–4. In multivariable analysis independent risk factors for death included age>45 years (RR=1.90 (95%CI 1.29–2.80), HIV infection (RR=4.22 (2.65–6.72)), extrapulmonary with pulmonary 6 disease (RR=1.54 (1.04–2.26)), BMI<18.5 (RR=2.71 (1.91–3.85)), previous use of fluoroquinolones (RR=1.91 (1.31–2.78)), resistance to any thioamide (RR=1.59 (1.14–2.22)), baseline positive smear (RR=2.22 (1.60–3.10)), and no culture conversion by 3rd month of treatment (RR=1.69 (1.19–2.41)). Patients with prior contact with any TB patient had decreased risk of death (RR=0.10 (0.01–0.68)). Treatment failure was associated in multivariable analysis with cavitary disease (RR=1.73 (1.07–2.80)), resistance to any fluoroquinolone (RR=2.73 (1.71–4.37)) and any thioamide (RR=1.62 (1.12–2.34)), and no culture conversion by 3rd month (RR=5.84 (3.02–11.27)). Independent predictors of default in multivariable analysis included unemployment (RR=1.50 (1.12–2.01)), homelessness (RR=1.52 (1.00–2.31)), imprisonment (RR=1.86 (1.42–2.45)), alcohol abuse (RR=1.60 (1.18–2.16)), and baseline positive smear (RR=1.35 (1.07–1.71)).
Table 2.
Univariate and multivariable analyses of socio-demographic risk factors for death, failure, and default versus successful treatment outcome in patients with MDR-TB
| Factor | Died N=200 n (%) |
cRR (95% CI) for death | aRR* (95% CI) for death | Failed N=118 n (%) |
cRR (95% CI) for failure | aRR* (95% CI) for failure | Defaulted N=241 n (%) |
cRR (95% CI) for default | aRR* (95% CI) for default | Success** N=1156 n |
|---|---|---|---|---|---|---|---|---|---|---|
| Age group¶ | ||||||||||
| Age <25 years | 29 (11.5) | 1.00 | 1.00 | 11 (4.7) | 1.00 | 29 (11.5) | 1.00 | 565 | ||
| Age 25–45 years | 97 (14.7) | 1.27 (0.86–1.88) | 1.35 (0.94–1.93) | 73 (11.4) | 2.43 (1.31–4.51) | - | 143 (20.2) | 1.76 (1.21–2.55) | - | 314 |
| Age >45 years | 64 (16.9) | 1.47 (0.98–2.21) | 1.90 (1.29–2.80) | 34 (9.8) | 2.08 (1.07–4.02) | 65 (17.2) | 1.49 (0.99–2.24) | 223 | ||
|
| ||||||||||
| Gender | ||||||||||
| Male | 133 (14.2) | 0.89 (0.68–1.17) | - | 82 (9.3) | 1.00 (0.69–1.46) | - | 82 (9.3) | 1.00 (0.69–1.46) | - | 802 |
| Female | 67 (15.9) | 1.00 | 36 (9.2) | 1.00 | 36 (9.2) | 1.00 | 354 | |||
|
| ||||||||||
| Homelessness | ||||||||||
| Yes | 2 (11.1) | 0.77 (0.21–2.89) | - | 4 (20.0) | 2.16 (0.88–5.29) | - | 16 (50.0) | 3.01 (2.08–4.36) | 1.52 (1.00–2.31) | 16 |
| No | 159 (14.3) | 1.00 | 97 (9.3) | 1.00 | 189 (16.6) | 1.00 | 1.00 | 950 | ||
| Unknown | 39 (17.0) | 1.19 (0.86–1.64) | 17 (8.2) | 0.89 (0.54–1.45) | 36 (15.9) | 0.96 (0.69–1.33) | 1.17 (0.62–2.21) | 190 | ||
|
| ||||||||||
| Imprisonment | ||||||||||
| Current or history | 26 (10.4) | 0.87 (0.57–1.33) | - | 44 (16.5) | 1.76 (1.21–2.56) | - | 83 (27.1) | 1.82 (1.40–2.38) | 1.86 (1.42–2.45) | 223 |
| Never | 69 (12.0) | 1.00 | 52 (9.4) | 1.00 | 88 (14.9) | 1.00 | 1.00 | 504 | ||
| Unknown | 105 (19.7) | 1.63 (1.23–2.16) | 22 (4.9) | 0.52 (0.32–0.85) | 70 (14.0) | 0.94 (0.71–1.26) | 1.24 (0.74–2.07) | 429 | ||
|
| ||||||||||
| Employment status | ||||||||||
| Unemployed | 101 (17.2) | 1.57 (1.13–2.17) | - | 55 (10.1) | 1.03 (0.70–1.51) | - | 139 (22.2) | 1.68 (1.27–2.23) | 1.50 (1.12–2.01) | 488 |
| Employed or other*** | 46 (10.9) | 1.00 | 41 (9.9) | 1.00 | 57 (13.2) | 1.00 | 1.00 | 375 | ||
| Unknown | 53 (15.3) | 1.40 (0.97–2.03) | 22 (7.0) | 0.71 (0.43–1.16) | 45 (13.3) | 1.01 (0.70–1.45) | 0.83 (0.51–1.36) | 293 | ||
|
| ||||||||||
| Alcohol abuse | ||||||||||
| Yes | 44 (13.5) | 1.69 (1.09–2.61) | - | 57 (16.8) | 2.04 (1.36–3.07) | - | 106 (27.3) | 1.95 (1.46–2.61) | 1.60 (1.18–2.16) | 282 |
| No | 31 (8.0) | 1.00 | 32 (8.2) | 1.00 | 58 (14.0) | 1.00 | 1.00 | 357 | ||
| Unknown | 125 (19.5) | 2.44 (1.68–3.54) | 29 (5.3) | 0.65 (0.40–1.05) | 77 (13.0) | 0.93 (0.68–1.27) | 0.93 (0.64–1.35) | 517 | ||
|
| ||||||||||
| Smoking | ||||||||||
| Yes | 22 (9.6) | 0.53 (0.34–0.82) | - | 11 (5.0) | 0.95 (0.47–1.93) | - | 31 (13.0) | 1.01 (0.67–1.52) | - | 208 |
| No | 87 (18.0) | 1.00 | 22 (5.3) | 1.00 | 58 (12.8) | 1.00 | 395 | |||
| Unknown | 91 (14.1) | 0.78 (0.60–1.02) | 85 (17.7) | 2.53 (1.61–3.97) | 152 (21.6) | 1.68 (1.27–2.22) | 553 | |||
|
| ||||||||||
| Injecting drug use | ||||||||||
| Yes | 2 (6.7) | 1.01 (0.25–4.04) | - | 0 (0) | undefined | - | 13 (31.7) | 2.06 (1.25–3.39) | - | 732 |
| No | 28 (6.6) | 1.00 | 58 (12.8) | 1.00 | 72 (15.4) | 1.00 | 28 | |||
| Unknown | 170 (30.0) | 2.85 (1.95–4.18) | 60 (13.2) | 0.59 (0.42–0.83) | 156 (28.3) | 1.14 (0.88–1.47) | 396 | |||
cRR = crude risk ratio (univariate analysis); aRR = adjusted risk ratio (multivariable analysis); CI = confidence interval.
68 patients had missing information on age and were not included in multivariable analysis (including 10 patients with death, 4 patients with default and 54 with success outcome).
Controlling for GLC site.
Successful treatment outcome is referent group for each poor outcome (death, treatment failure and default).
Other employment status includes retired, student, disabled. Variables for which aRRs are not shown were not included in the final multivariable model. Risk ratios for predictors significant at P=0.05 are highlighted in bold.
Table 4.
Univariate and multivariable analyses of laboratory predictors for death, failure, and default versus successful treatment outcome in patients with MDR-TB
| Factor | Died N=200 n (%) |
cRR (95% CI) for death | aRR* (95% CI) for death | Failed N=118 n (%) |
cRR (95% CI) for failure | aRR* (95% CI) for failure | Defaulted N=241 n (%) |
cRR (95% CI) for default | aRR* (95% CI) for default | Success** N=1156 n |
|---|---|---|---|---|---|---|---|---|---|---|
| Baseline AFB smear status | ||||||||||
| Positive | 160 (19.3) | 2.67 (1.90–3.75) | 2.22 (1.60–3.10) | 89 (11.7) | 2.11 (1.40–3.17) | - | 156 (18.9) | 1.27 (0.99–1.62) | 1.35 (1.07–1.71) | 669 |
| Negative | 37 (7.2) | 1.00 | 28 (5.6) | 1.00 | 83 (14.9) | 1.00 | 1.00 | 474 | ||
| Unknown | 3 (18.8) | 2.59 (0.89–7.52) | 1.67 (0.52–5.38) | 1 (7.1) | 1.28 (0.19–8.76) | 2 (13.3) | 0.89 (0.24–3.30) | 1.47 (0.39–5.59) | 13 | |
|
| ||||||||||
| Baseline resistance to any fluoroquinolone | ||||||||||
| Yes | 29 (24.8) | 2.01 (1.40–2.88) | - | 22 (20.0) | 2.75 (1.76–4.30) | 2.73 (1.71–4.37) | 12 (12.0) | 0.79 (0.45–1.37) | - | 88 |
| No | 106 (12.4) | 1.00 | 59 (7.3) | 1.00 | 1.00 | 135 (15.2) | 1.00 | 752 | ||
| Unknown | 65 (17.1) | 1.38 (1.04–1.83) | 37 (10.5) | 1.44 (0.97–2.13) | 0.80 (0.43–1.47) | 94 (22.9) | 1.51 (1.19–1.91) | 316 | ||
|
| ||||||||||
| Baseline resistance to any injectable SLD | ||||||||||
| Yes | 50 (15.5) | 1.27 (0.92–1.75) | - | 47 (14.7) | 2.22 (1.51–3.25) | - | 41 (13.1) | 0.88 (0.63–1.22) | - | 272 |
| No | 92 (12.2) | 1.00 | 47 (6.6) | 1.00 | 116 (14.9) | 1.00 | 661 | |||
| Unknown | 58 (20.6) | 1.69 (1.25–2.28) | 24 (9.7) | 1.46 (0.91–2.34) | 84 (27.4) | 1.83 (1.43–2.35) | 223 | |||
|
| ||||||||||
| Baseline resistance to any thioamide | ||||||||||
| Yes | 58 (19.9) | 1.90 (1.37–2.62) | 1.59 (1.14–2.22) | 33 (12.4) | 1.93 (1.24–3.01) | 1.62 (1.12–2.34) | 38 (14.0) | 1.01 (0.71–1.44) | - | 233 |
| No | 65 (10.5) | 1.00 | 1.00 | 38 (6.4) | 1.00 | 1.00 | 89 (13.9) | 1.00 | 553 | |
| Unknown | 77 (17.2) | 1.64 (1.21–2.23) | 1.36 (0.89–2.07) | 47 (11.3) | 1.75 (1.16–2.64) | 2.07 (1.07–3.98) | 114 (23.6) | 1.70 (1.32–2.18) | 370 | |
|
| ||||||||||
| Baseline XDR | ||||||||||
| Yes | 10 (24.4) | 1.82 (1.03–3.19) | - | 13 (29.5) | 3.80 (2.28–6.33) | - | 3 (8.8) | 0.59 (0.20–1.74) | - | 31 |
| No | 125 (13.4) | 1.00 | - | 68 (7.8) | 1.00 | - | 143 (15.1) | 1.00 | - | 806 |
| Unknown | 65 (16.9) | 1.26 (0.96–1.66) | - | 37 (10.4) | 1.34 (0.91–1.96) | - | 95 (22.9) | 1.52 (1.21–1.92) | - | 319 |
|
| ||||||||||
| Initial sputum culture conversion within 3 months of treatment*** | ||||||||||
| No | 77 (26.1) | 2.07 (1.44–2.96) | 1.69 (1.19–2.41) | 84 (27.8) | 7.20 (3.82–13.58) | 5.84 (3.02–11.27) | 70 (24.3) | 1.43 (1.04–1.97) | - | 218 |
| Yes | 87 (11.2) | 0.89 (0.62–1.28) | 0.94 (0.66–1.33) | 24 (3.4) | 0.87 (0.42–1.80) | 0.96 (0.46–1.99) | 120 (14.8) | 0.87 (0.65–1.18) | 689 | |
| Baseline culture negative | 36 (12.6) | 1.00 | 1.00 | 10 (3.9) | 1.00 | 1.00 | 51 (17.0) | 1.00 | 249 | |
cRR = crude risk ratio (univariate analysis); aRR = adjusted risk ratio (multivariable analysis); CI = confidence interval.
Controlling for GLC site.
Successful treatment outcome is referent group for each poor outcome (death, treatment failure and default).
Cut point for timing of initial sputum culture conversion was chosen based on median time to conversion at 3 months of treatment (interquartile range [IQR]: 2.0, 5.0). Risk ratios for predictors significant at P=0.05 are highlighted in bold.
Variables for which aRRs are not shown were not included in the final multivariable model.
Discussion
In summary, despite the use of DOTS-plus to treat this large retrospective cohort of 1,768 patients with drug-resistant tuberculosis, treatment succeeded only in 65% of patients, while 32% had poor treatment outcomes, including death (11%), failure (7%), and default (14%). Treatment outcomes were strikingly similar to the results of two recent meta-analyses and recent surveillance data from 71 countries.9 Johnston et al2 reported from 36 studies that 62% of patients achieved successful outcomes, while 11% died, 8% failed, and 13% defaulted. Orenstein et al3 reported a mean proportion of patients with treatment success in 29 studies of individualized treatment regimens for MDR-TB was 64%, while 11% died, 6% failed, and 12% defaulted treatment. Surveillance data from 71 countries or territories on treatment outcomes for MDR-TB cases showed similar overall treatment success rate (60% after adjustment for clustering by countries) and deaths (8%-13% in new and retreatment cases).9
When patients are heading for a poor outcome, prompt action by the physician/program is required. This may include repeating or broadening DST, changing treatment, social/behavioral support, surgery, and further evaluation. Death and treatment failure were associated mainly with biomedical factors, while default from treatment was associated mostly with social factors.
Independent predictors of death and failure, including HIV infection, low body mass index, extensive disease, previous treatment with specific drugs, baseline positive smear, additional resistance to specific drugs, and persistently positive cultures at the 3rd month of treatment should trigger additional steps and may be amenable to specific interventions. Older age, HIV infection, and extensive TB disease are well-known risk factors for mortality of patients with tuberculosis.10 Association of prior contact with any TB patient with reduced mortality may be related to earlier diagnosis of MDR TB through contact investigation. Malnutrition, positive baseline smear, and delayed culture conversion were previously demonstrated as risk factors for poor outcomes of MDR-TB treatment.2, 11 Our findings suggest that fluoroquinolones play an important part in TB treatment, which concurs with previous studies demonstrating that fluoroquinolone resistance was associated with poor MDR-TB treatment outcomes.2, 12, 13 Regimens for treatment of TB that include fluoroquinolones need to be strong enough not to amplify resistance to fluoroquinolones. Also patients who receive fluoroquinolones for other reasons (e.g. pneumonia) should be screened for TB first. The impact of thioamide resistance on poor outcomes should be further studied.
Socioeconomic factors that increase the risk of default such as unemployment, homelessness, imprisonment, and alcoholism should trigger different types of measures to improve treatment adherence, such as disability stipends, different types of incentives and/or enablers, consultation with a social worker, home visiting, reimbursement of transportation to/from the place of treatment, psychological counseling, legal consulting, etc.14 Collaboration between TB services and substance abuse services may improve TB treatment adherence, but this needs further research.
Concordantly with previous reports,15–17 we found that adverse events during MDR TB treatment were very common – 93% of patients had at least one adverse event reported. Gastrointestinal disturbances (nausea and diarrhea) were most often reported side effects. It is important to emphasize that adverse events may have a severe impact on adherence to treatment, so it is crucial to clinically monitor patients for adverse events and timely manage side effects if they occur.4, 7
This study has several limitations. The data were retrospective and not originally collected for the purpose of this study; thus there was site-to-site variation in data definitions and recording format. We operationalized the variables for this analysis so they would be consistent across sites. However, not all sites collected data on the same variables; consequently data were missing for individual variables (for up to 30% of observations for certain variables) for some sites; however, we included a category for “missing” results as a dummy variable in the multivariable regression analysis for variables missing more than 7% of data. Data were not available regarding details of treatment, but all sites followed GLC approved treatment protocols consisting of at least four effective drugs, generally treating with aggressive regimens of five to seven drugs. Treatment outcomes at each of the DOTS-plus projects were defined by national norms, but these were generally consistent with international consensus recommendations developed 2000–2002.8 Generalizability of the study findings may be limited to sites with comparable patient populations, i.e. with low HIV rates, similar drug-resistance epidemiology, and on individualized MDR-TB treatment.
The overall findings of our study corroborate with findings of recent individual patient data (IPD) meta-analysis of 9,153 patients18 which was used to inform the revision of the WHO Guidelines for the Programmatic Management of Drug-resistant Tuberculosis (PMDT)19. The median duration of treatment of 21 months among those with success outcome in our study was similar to total treatment duration of 18.6–21.5 months associated with increased odds of success in IDP meta-analysis. Similarly to Menzies at al18, we found an independent association of poor outcomes with fluoroquinolone resistance or use of fluoroquinolone in the past for treatment of TB and thioamide resistance, and no independent association of resistance to injectable agents with poor treatment outcomes. Thus, the findings of our study support current WHO recommendations on including a later generation quinolone and thioamide in treatment regimen and total duration of treatment of at least 21 months.19 Given gross lack of prospective evaluation of MDR regimens (with and without new drugs), exemplified by the finding that fluoroquinolones resistance was the strongest risk factor for failure, better understanding on how to best combine old drugs with novel drugs in the pipeline is also needed. Randomized clinical trials are urgently needed to determine the optimal treatment regimens for patients with MDR-TB.
In conclusion, patients with identified biomedical risk factors for death or treatment failure should receive targeted medical attention from the beginning of MDR-TB treatment including aggressive treatment with maximum number of effective drugs, adjuvant interventions, including nutritional supplementation and consideration of surgery. To prevent treatment default, management of patients who are unemployed, homeless, alcoholic, or have a history of imprisonment requires extra measures to insure treatment completion. Operations research should be carried out to identify effective interventions for both adherence and treatment outcome improvement in specific settings.
Table 3.
Univariate and multivariable analyses of clinical risk factors for death, failure, and default versus successful treatment outcome in patients with MDR-TB
| Factor | Died N=200 n (%) |
cRR (95% CI) for death | aRR* (95% CI) for death | Failed N=118 n (%) |
cRR (95% CI) for failure | aRR* (95% CI) for failure | Defaulted N=241 n (%) |
cRR (95% CI) for default | aRR* (95% CI) for default | Success** N=1156 n |
|---|---|---|---|---|---|---|---|---|---|---|
| HIV status | ||||||||||
| Positive | 9 (50.0) | 3.82 (2.35–6.22) | 4.22 (2.65–6.72) | 1 (10.0) | 1.20 (0.18–7.77) | - | 8 (47.1) | 2.88 (1.71–4.86) | - | 9 |
| Unknown | 49 (19.4) | 1.49 (1.11–2.00) | 1.57 (1.13–2.18) | 31 (13.2) | 1.59 (1.08–2.33) | 49 (19.4) | 1.19 (0.90–1.58) | 201 | ||
| Negative | 142 (13.1) | 1.00 | 1.00 | 86 (8.3) | 1.00 | 184 (16.3) | 1.00 | 944 | ||
|
| ||||||||||
| Diabetes mellitus | ||||||||||
| Yes | 9 (16.7) | 1.24 (0.67–2.30) | - | 6 (11.8) | 1.39 (0.64–3.04) | - | 8 (15.1) | 1.02 (0.53–1.96) | 45 | |
| No | 145 (13.4) | 1.00 | 87 (8.5) | 1.00 | 163 (14.8) | 1.00 | 938 | |||
| Unknown | 46 (21.0) | 1.57 (1.16–2.11) | 25 (12.6) | 1.49 (0.98–2.26) | 70 (28.8) | 1.95 (1.53–2.48) | 173 | |||
|
| ||||||||||
| Body mass index (BMI) | ||||||||||
| BMI<18.5 | 53 (24.3) | 3.36 (2.35–4.79) | 2.71 (1.91–3.85) | 28 (14.5) | 1.80 (1.18–2.76) | - | 24 (12.0) | 0.83 (0.55–1.26) | - | 165 |
| BMI>=18.5 | 50 (7.2) | 1.00 | 1.00 | 56 (8.0) | 1.00 | 115 (15.2) | 1.00 | 640 | ||
| Unknown | 97 (21.7) | 2.99 (2.17–4.11) | 2.01 (1.40–2.88) | 34 (8.8) | 1.10 (0.73–1.65) | 102 (22.5) | 1.48 (1.16–1.88) | 351 | ||
|
| ||||||||||
| Site of TB | ||||||||||
| Pulmonary with EPTB | 17 (28.8) | 2.02 (1.32–3.08) | 1.54 (1.04–2.26) | 4 (8.7) | 1.13 (0.44–2.93) | 7 (14.30 | 1.26 (0.63–2.53) | - | 42 | |
| Pulmonary only | 170 (14.3) | 1.00 | 1.00 | 111 (9.8) | 1.00 | 224 (18.0) | 1.00 | 1020 | ||
| Unknown | 13 (12.2) | 0.85 (0.50–1.44) | 0.95 (0.29–3.09) | 3 (3.1) | 0.36 (0.08–1.52) | 10 (9.6) | 0.67 (0.27–1.66) | 94 | ||
|
| ||||||||||
| Cavitary TB disease | ||||||||||
| Yes | 131 (15.9) | 1.55 (1.08–2.23) | - | 93 (11.8) | 2.33 (1.37–3.95) | 1.73 (1.07–2.80) | 168 (19.5) | 1.31 (0.98–1.75) | - | 693 |
| No | 32 (10.3) | 1.00 | 15 (5.1) | 1.00 | 1.00 | 49 (14.9) | 1.00 | 280 | ||
| Unknown | 37 (18.5) | 1.64 (1.06–2.55) | 10 (5.2) | 1.02 (0.47–2.22) | 1.49 (0.62–3.58) | 24 (11.6) | 0.78 (0.49–1.23) | 183 | ||
|
| ||||||||||
| Outcome of previous treatment of TB under DOTS | ||||||||||
| Cure | 14 (9.3) | 1.77 (0.74–4.27) | - | 13 (8.7) | 1.46 (0.63–3.42) | 1.25 (0.57–2.75) | 24 (14.9) | 0.88 (0.53–1.46) | - | 137 |
| Failure or default | 105 (16.1) | 3.07 (1.46–6.46) | 35 (6.0) | 1.01 (0.48–2.13) | 2.04 (0.98–4.24) | 78 (12.4) | 0.73 (0.49–1.10) | 549 | ||
| Unknown | 74 (17.7) | 3.40 (1.60–7.19) | 62 (15.3) | 2.58 (1.27–5.25) | 2.86 (1.43–5.74) | 113 (24.8) | 1.46 (0.99–2.14) | 343 | ||
| New MDR-TB case | 7 (5.2) | 1.00 | 8 (5.9) | 1.00 | 1.00 | 26 (17.0) | 1.00 | 127 | ||
|
| ||||||||||
| History of TB treatment with fluoroquinolones | ||||||||||
| Yes | 96 (26.1) | 2.48 (1.93–3.18) | 1.91 (1.31–2.78) | 30 (9.9) | 1.10 (0.74–1.63) | - | 47 (14.7) | 0.82 (0.61–1.10) | - | 272 |
| No | 104 (10.5) | 1.00 | 1.00 | 88 (9.1) | 1.00 | 194 (18.0) | 1.00 | 884 | ||
|
| ||||||||||
| History of TB treatment with any injectable SLD | ||||||||||
| Yes | 99 (20.9) | 1.82 (1.42–2.35) | - | 43 (10.3) | 1.17 (0.82–1.68) | - | 65 (14.8) | 0.80 (0.62–1.04) | - | 375 |
| No | 101 (11.5) | 1.00 | 75 (8.8) | 1.00 | 176 (18.4) | 1.00 | 781 | |||
Note. cRR = crude risk ratio (univariate analysis); aRR = adjusted risk ratio (multivariable analysis); CI = confidence interval; EPTB=Extrapulmonary TB.
Controlling for GLC site.
Successful treatment outcome is referent group for each poor outcome (death, treatment failure and default). Variables for which aRRs are not shown were not included in the final multivariable model. Risk ratios for predictors significant at P=0.05 are highlighted in bold.
Acknowledgments
Funding. The United States Agency for International Development (USAID). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Footnotes
Conflict of Interest. All authors declare no conflict of interest.
IRB approval. The original study was approved by the U.S. Centers for Disease Control and Prevention (CDC) Institutional Review Board (IRB). Secondary analysis of these data reported in current manuscript received non-research determination by the U.S. Centers for Disease Control and Prevention (CDC) because this activity was considered to be a research but does NOT involve identifiable human subjects (i.e. personal identifiers were stripped from the combined database, and the data were anonymized; the data cannot be linked or re-linked with identifiable human subjects; there was no additional data collection or further interaction with patient medical records).
Disclaimer. The conclusions and interpretations of data presented in this report are solely those of the authors and do not necessarily represent an official position of CDC, USAID, WHO or the governments of the five countries.
D.F and M.Z. are staff members of the World Health Organization (WHO). The authors alone are responsible for the views expressed in this publication and they do not necessarily represent the decisions or policies of WHO.
References
- 1.World Health Organization. The Global Plan to Stop TB 2011–2015. Geneva, Switzerland: 2011. [Google Scholar]
- 2.Johnston JC, Shahidi NC, Sadatsafavi M, Fitzgerald JM. Treatment outcomes of multidrug-resistant tuberculosis: a systematic review and meta-analysis. PLoS One. 2009;4(9):e6914. doi: 10.1371/journal.pone.0006914. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Orenstein EW, Basu S, Shah NS, et al. Treatment outcomes among patients with multidrug-resistant tuberculosis: systematic review and meta-analysis. Lancet Infect Dis. 2009;9(3):153–61. doi: 10.1016/S1473-3099(09)70041-6. [DOI] [PubMed] [Google Scholar]
- 4.World Health Organization. Emergency Update. Geneva, Switzerland: 2008. Guidelines for the programmatic management of drug-resistant tuberculosis. [Google Scholar]
- 5.Nathanson E, Lambregts-van Weezenbeek C, Rich ML, et al. Multidrug-resistant tuberculosis management in resource-limited settings. Emerg Infect Dis. 2006;12(9):1389–97. doi: 10.3201/eid1209.051618. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Gammino V, Taylor A, Rich M, et al. Bacteriologic monitoring of multidrug-resistant tuberculosis patients in five DOTS-Plus pilot projects. Int J Tuberc Lung Dis. 2011;15(10):1315–22. doi: 10.5588/ijtld.10.0221. [DOI] [PubMed] [Google Scholar]
- 7.World Health Organization. Guidelines for establishing DOTS-PLUS pilot projects for the management of multidrug-resistant tuberculosis (MDR-TB) Geneva, Switzerland: 2000. [Google Scholar]
- 8.Laserson KF, Thorpe LE, Leimane V, et al. Speaking the same language: treatment outcome definitions for multidrug-resistant tuberculosis. Int J Tuberc Lung Dis. 2005;9(6):640–5. [PubMed] [Google Scholar]
- 9.World Health Organization. 2010 Global Report on Surveillance and response. Geneva, Switzerland: 2010. Multidrug and extensively drug-resistant TB (M/XDR-TB) [Google Scholar]
- 10.Waitt CJ, Squire SB. A systematic review of risk factors for death in adults during and after tuberculosis treatment. Int J Tuberc Lung Dis. 2011 doi: 10.5588/ijtld.10.0352. [DOI] [PubMed] [Google Scholar]
- 11.Holtz TH, Sternberg M, Kammerer S, et al. Time to sputum culture conversion in multidrug-resistant tuberculosis: predictors and relationship to treatment outcome. Ann Intern Med. 2006;144(9):650–9. doi: 10.7326/0003-4819-144-9-200605020-00008. [DOI] [PubMed] [Google Scholar]
- 12.Leimane V, Riekstina V, Holtz TH, et al. Clinical outcome of individualised treatment of multidrug-resistant tuberculosis in Latvia: a retrospective cohort study. Lancet. 2005;365(9456):318–26. doi: 10.1016/S0140-6736(05)17786-1. [DOI] [PubMed] [Google Scholar]
- 13.Migliori GB, Lange C, Girardi E, et al. Fluoroquinolones: are they essential to treat multidrug-resistant tuberculosis? Eur Respir J. 2008;31(4):904–5. doi: 10.1183/09031936.00159807. [DOI] [PubMed] [Google Scholar]
- 14.Jakubowiak WM, Bogorodskaya EM, Borisov SE, Danilova ID, Lomakina OB, Kourbatova EV. Social support and incentives programme for patients with tuberculosis: experience from the Russian Federation. Int J Tuberc Lung Dis. 2007;11(11):1210–5. [PubMed] [Google Scholar]
- 15.Bloss E, Kuksa L, Holtz TH, et al. Adverse events related to multidrug-resistant tuberculosis treatment, Latvia, 2000–2004. Int J Tuberc Lung Dis. 2010;14(3):275–81. [PubMed] [Google Scholar]
- 16.Nathanson E, Gupta R, Huamani P, et al. Adverse events in the treatment of multidrug-resistant tuberculosis: results from the DOTS-Plus initiative. Int J Tuberc Lung Dis. 2004;8(11):1382–4. [PubMed] [Google Scholar]
- 17.Shin SS, Pasechnikov AD, Gelmanova IY, et al. Adverse reactions among patients being treated for MDR-TB in Tomsk, Russia. Int J Tuberc Lung Dis. 2007;11(12):1314–20. [PubMed] [Google Scholar]
- 18.The Collaborative Group for Meta-Analysis of Individual Patient Data in MDR-TB. Specific treatment parameters and treatment outcomes of multidrug-resistant tuberculosis: an individual patient data (IPD) meta-analysis of 9153 patients. In preparation. [Google Scholar]
- 19.World Health Organization. Guidelines for the programmatic management of drug-resistant tuberculosis. Geneva, Switzerland: 2011. 2011 update. [PubMed] [Google Scholar]