Summary
OBJECTIVE
To evaluate the performance of a recently updated rapid molecular diagnostic test, MTBDRplus version 2 (v2), designed to detect drug resistance in both acid-fast bacilli (AFB) smear negative and positive specimens.
DESIGN
Sputum samples from 1,128 patients at risk for multidrug-resistant tuberculosis (MDR-TB) were tested by MTBDRplus v2 and compared to reference standard MGIT 960 drug susceptibility testing. The relationship of participant HIV status, diabetic status, previous treatment, and smear gradation to the likelihood of obtaining an interpretable result was assessed using logistic regression.
RESULTS
MTBDRplus v2 sensitivity and specificity for detecting MDR-TB, when compared to a reference standard, were 96.0% (95%CI 93.5–97.6) and 99.2% (95%CI 97.0–99.9) for AFB smear positive specimens and 82.8% (95%CI 63.5–93.5) and 98.3% (95%CI 89.9–99.9) for AFB smear negative specimens, respectively. A dose-response relationship was observed between the proportion of interpretable test results and AFB smear bacterial load after adjusting for age, sex, BMI, HIV status, previous treatment, and diabetic status.
CONCLUSION
While MTBDRplus v2 performs well among both AFB smear positive and negative specimens, smear gradation appears to influence both the probability of obtaining an interpretable result and test sensitivity, indicating a significant association between bacillary load and test performance.
Keywords: Diagnostic, Tuberculosis, Multidrug-resistance
BACKGROUND
Although global incidence rates of tuberculosis (TB) have been decreasing in recent years, multidrug-resistant TB (MDR-TB) or TB that is resistant to both isoniazid (INH) and rifampicin (RIF), the two most commonly used drugs for first-line treatment, threatens to undermine this recent progress. According to World Health Organization (WHO) estimates, in 2013, approximately 3.5% of new and 20.5% of previously-treated cases of TB were multidrug-resistant.1 These global estimates, however, do not provide a complete picture of the MDR-TB epidemic, as subnational rates of MDR-TB have risen to as high as 35% among new cases and 75% among previously treated cases in several eastern European and central Asian countries.1
Historically, phenotypic drug susceptibility testing (DST) of Mycobacterium tuberculosis (Mtb) on solid media took months to produce results.2 The advent of broth-based media reduced the time to results to weeks, and the recent introduction of rapid molecular-based diagnostics has further reduced this time to less than a day.3–5 Despite these gains, including WHO approval of the GenoType MTBDRplus assay (Hain, Nerhen, Germany) and Xpert MTB/RIF assay (Cephid, Sunnyvale, USA) in 2008 and 2010, gaps in MDR-TB diagnosis remain. In 2013, it is estimated that less than half (45%) of MDR-TB cases were detected globally.1 Differences in diagnostic test characteristics partially explain this gap in MDR-TB diagnosis. For example, Xpert MTB/RIF was developed to detect RIF resistance, but does not assess INH resistance. The MTBDRplus version 1 (v1) assay detects resistance to both RIF and INH, however given its performance, WHO recommended its usage among Acid-Fast Bacilli (AFB) smear positive samples only.6
In 2012, version 2 (v2) of the MTBDRplus assay was released to address performance issues associated with AFB smear status. Several small-scale studies have assessed the performance of this new assay version among homogenous populations, but a large-scale study in a diverse population has yet to be performed.7–9 Additionally, the impact of AFB smear status, HIV status, body mass index (BMI), and other clinical factors associated with disease progression on the performance of MTBDRplus v2 have yet to be assessed.10–13
Expanding the rapid testing and detection of MDR-TB cases is one of the five priority actions outlined by WHO in their goals to address the global MDR-TB crisis.1 Field trials of rapid point-of-care diagnostics provide valuable information for clinical decision making in resource limited settings.14 In this study we describe the performance of the MTBDRplus v2 assay in a large diverse field trial among both AFB smear positive and negative specimens.
METHODS
Study design
This prospective (diagnostic) cohort study was conducted as part of the Global Consortium for Drug-resistant TB Diagnostics (GCDD), a collaboration between UCSD and study sites in India, Moldova, and South Africa. Informed consent was obtained from all patients prior to study enrollment. This study was approved by institutional review boards at the University of California, San Diego as well as the participating institutions at the three study sites.
Details of patient recruitment and selection have been reported previously.15, 16 In brief, between 2012 and 2013, TB patients with risk factors for drug-resistant TB were screened at participating study sites. Patients were eligible if they were at least five years of age and had an AFB smear positive or a positive Xpert result in the previous fourteen days. It addition, patients had to meet at least one of the following MDR suspicion criteria: received >1 month of treatment for a prior TB episode, failing TB treatment with positive sputum smear or culture after ≥3 months of standard TB treatment, close contact with a known drug-resistant TB case, newly diagnosed with MDR-TB within the previous 30 days, or previously diagnosed with MDR-TB and failed TB treatment with positive sputum smear or culture after ≥3 months of a standard MDR-TB treatment regimen. Patients were excluded if they were unable to provide at least 7.5 mL combined sputum or had received second line drug susceptibility testing results within the previous three months.
Specimen collection and processing
The initial sputum sample was collected by study staff at the clinic site during enrollment and the second sample by the participant themselves upon waking the following morning. These two samples were pooled and used for multiple diagnostic tests including AFB smear, MTBDRplus v2 line probe assay, and MGIT 960 culture (the inoculum for MGIT 960 DST). All tests were performed by local study site laboratories according to manufacturer instructions or per standard procedures.17, 18 Mtb culture positive specimens were tested for drug susceptibility to RIF and INH resistance at concentrations of 1.0 and 0.1 (μg/ml) respectively.19
MTBDRplus v2
The updated MTBDRplus v2 assay strip contains 27 hybridization probes. Two probes, the conjugate and amplification probes, ensure test function. Four control probes confirm the presence of Mtb and identify wild type loci sequences for the rpoB, katG, and inhA promoter gene regions. Eight rpoB wild type probes overlap gene regions between codons 505 and 533. Four rpoB mutant probes identify the specific mutations D516V, H526Y, H526D, and S531L. One wild type probe is specific for the katG315 codon and two mutant probes identify S315T mutations at this codon. Two inhA wild type probes cover the inhA promoter from position −15 to −8. And, four mutant probes identify specific nucleic acid changes in the inhA promotor region (C-15T, A-16G, T-8C, and T-8A). The absence of one or more wild type probes for a given resistance-associated gene region, or the absence of one or more wild type probes in addition to the presence of one or more mutant probes for a given gene region, is considered indicative of resistance. Assay results are classified as indeterminate if either of the test function probes are absent; if the Mtb or gene loci probes are absent; or if all wild type probes, in addition to mutation probes, are present.18
Data Analysis
Performance of MTBDRplus v2 was described by computing sensitivity, specificity, likelihood ratio positive (LR+), and likelihood ratio negative (LR−) for RIF resistance, INH resistance, and MDR-TB; confidence intervals for all proportions were calculated using the Wilson Score method.20 Logistic regression was used to determine if any demographic or clinical data were associated with obtaining valid MTBDRplus v2 results. Analyses were performed using Stata 13 (College Station, TX).
RESULTS
A total of 1,128 participants were enrolled in the study. Nine-hundred and fourteen (81%) were Mtb culture positive, and of those culture positive specimens 540 (59%) were phenotypically RIF resistant and 592 (65%) were phenotypically INH resistant. Among the 540 phenotypically RIF resistant specimens, 462 (86%) were classified as resistant, 16 (3%) were classified as susceptible, and 62 (11%) were indeterminate by MTBDRplus v2. Among the 592 phenotypically INH resistant specimens, 478 (82%) were classified as resistant, 31 (5%) were classified as susceptible, and 83 (14%) were indeterminate by MTBDRplus v2 (Table 1). Sub-analysis revealed that among six phenotypically RIF mono-resistant specimens, two were classified as resistant, one as susceptible, and three as indeterminate by MTBDRplus v2; and among 58 phenotypically INH mono-resistant specimens, 32 (55%) were classified as resistant, 15 (26%) were classified as susceptible, and 11 (19%) were indeterminate by MTBDRplus v2. Culture negative specimens were also assessed using the MTBDRplus v2 assay and 42 (20%) specimens were classified as either INH and/or RIF resistant or susceptible.
Table 1.
Detection of rifampicin and isoniazid resistance by MTBDRplus v2 versus MGIT 960
| MTBDRplus v2 | MGIT Results
|
||||
|---|---|---|---|---|---|
| Resistant | Susceptible | Culture Negative | Null/Not Done | Total | |
| Rifampicin | |||||
| Resistant | 462 | 7 | 9 | 0 | 478 |
| Susceptible | 16 | 324 | 32 | 3 | 375 |
| Indeterminate | 62 | 36 | 173 | 4 | 275 |
| Total | 540 | 367 | 214 | 7 | 1128 |
| Isoniazid | |||||
| Resistant | 478 | 1 | 5 | 0 | 484 |
| Susceptible | 31 | 280 | 33 | 3 | 347 |
| Indeterminate | 83 | 34 | 176 | 4 | 298 |
| Total | 592 | 315 | 214 | 7 | 1128 |
Overall, sensitivity and specificity of MTBDRplus v2 when compared to culture for the detection of Mtb were 91.8% (95%CI 89.9–93.5) and 76.2% (95%CI 69.8–81.6). Among specimens with both valid phenotypic drug susceptibility and valid MTBDRplus v2 results, sensitivity and specificity of MTBDRplus v2 were 96.7% (95%CI 94.5–98.0) and 97.9% (95%CI 95.5–99.1) for detection of RIF resistance; 93.9% (95%CI 91.4–95.8) and 99.6% (95%CI 97.7–100.0) for INH resistance; and 95.1% (95%CI 92.6–96.9) and 99.1% (95%CI 97.1–99.8) for MDR-TB detection. Additionally, performance characteristics were stratified by AFB status (Table 2) and study site (Table 3). Point estimates for both sensitivity and specificity were consistently lower for AFB smear negative specimens than for AFB smear positive specimens. MTBDRplus v2 sensitivity and specificity were consistent across study sites with the exception of INH and RIF sensitivity estimates in South Africa, which were significantly lower than estimates in India and Moldova.
Table 2.
Performance parameters of MTBDRplus v2 in detecting rifampicin, isoniazid, and multidrug-resistant TB (MDR-TB) compared to conventional DST, stratified by smear status*
| Drug Resistance Smear Status | Isolates† | Sensitivity % | Specificity % | LR+ | LR− |
|---|---|---|---|---|---|
| Mtb | 1128 | 91.8 (89.8, 93.5) | 76.2 (69.8, 81.6) | 4 (3, 5) | 0.11 (0.09, 0.14) |
| AFB + | 826 | 95.6 (93.8, 96.9) | 74.1 (60.7, 84.4) | 4 (2, 6) | 0.06 (0.04, 0.09) |
| AFB − | 302 | 71.9 (63.8, 78.9) | 76.9 (69.4, 83.1) | 3 (2,4) | 0.37 (0.28, 0.48) |
| RIFR | 809 | 96.7 (94.5, 98.0) | 97.9 (95.5, 99.1) | 46 (22, 95) | 0.03 (0.02, 0.06) |
| AFB + | 712 | 97.1 (94.9, 98.4) | 98.5 (96.0, 99.5) | 65 (25, 173) | 0.03 (0.02, 0.05) |
| AFB − | 97 | 91.4 (75.8, 97.8) | 95.2 (85.6, 98.7) | 19 (6, 57) | 0.09 (0.03, 0.27) |
| INHR | 790 | 93.9 (91.4, 95.8) | 99.6 (97.7, 100.0) | 264 (37, 1867) | 0.06 (0.04, 0.09) |
| AFB + | 699 | 94.9 (92.4, 96.6) | 100.0 (97.9, 100.0) | ∞ | 0.05 (0.03, 0.08) |
| AFB − | 91 | 81.6 (65.1, 91.7) | 98.1 (88.6, 99.9) | 43 (6, 303) | 0.19 (0.10, 0.37) |
| MDR-TB | 775 | 95.1 (92.6, 96.9) | 99.1 (97.1, 99.8) | 102 (33, 316) | 0.05 (0.03, 0.07) |
| AFB + | 686 | 96.0 (93.5, 97.6) | 99.2 (97.0, 99.9) | 126 (32, 502) | 0.04 (0.03, 0.06) |
| AFB − | 89 | 82.8 (63.5, 93.5) | 98.3 (89.9, 99.9) | 50 (7,349) | 0.18 (0.08, 0.39) |
Values are percentages with 95% confidence interval in parentheses.
Samples that did not have both interpretable phenotypic and genotypic results were excluded from analysis.
Table 3.
Performance parameters of MTBDRplus v2 by Study Site
| Resistance Smear Status | Isolates | Sensitivity % | Specificity % | |
|---|---|---|---|---|
| India | RIF | 438 | 98.0 (95.8, 99.1) | 98.8 (92.5, 99.9) |
|
| ||||
| AFB+ | 422 | 98.0 (95.7, 99.1) | 98.7 (91.9, 99.9) | |
| AFB− | 16 | 100.0 (62.9, 100.0) | 100.0 (56.1, 100.0) | |
|
| ||||
| INH | 433 | 97.3 (94.9, 98.6) | 100.0 (93.3, 100.0) | |
|
| ||||
| AFB+ | 418 | 97.5 (95.1, 98.9) | 100.0 (92.9, 100.0) | |
| AFB− | 15 | 90.9 (57.1, 99.5) | 100.0 (39.6, 100.0) | |
|
| ||||
| MDR-TB | 426 | 96.2 (93.4, 97.9) | 98.8 (92.5, 99.9) | |
|
| ||||
| AFB+ | 411 | 96.4 (93.6, 98.0) | 98.7 (92.0, 99.9) | |
| AFB− | 15 | 88.9 (50.7, 99.4) | 100.0 (51.7, 100.0) | |
|
| ||||
| Moldova | RIF | 204 | 97.0 (90.9, 99.2) | 98.1 (92.5, 99.7) |
|
| ||||
| AFB+ | 166 | 98.7 (92.0, 99.9) | 97.8 (91.4, 99.6) | |
| AFB− | 38 | 91.7 (71.5, 98.5) | 100.0 (73.2, 100.0) | |
|
| ||||
| INH | 193 | 92.7 (85.7, 96.6) | 100.0 (94.5, 100.0) | |
|
| ||||
| AFB+ | 161 | 94.4 (86.8, 97.9) | 100.0 (93.7, 100.0) | |
| AFB− | 32 | 85.7 (62.6, 96.2) | 100.0 (67.9, 100.0) | |
|
| ||||
| MDR-TB | 190 | 94.5 (87.1, 98.0) | 99.0 (93.7, 99.9) | |
|
| ||||
| AFB+ | 158 | 98.6 (91.5, 99.9) | 98.8 (92.8, 99.9) | |
| AFB− | 32 | 78.9 (53.9, 93.0) | 100.0 (71.7, 100.0) | |
|
| ||||
| South Africa | RIF | 167 | 72.2 (49.6, 88.4) | 97.2 (92.6, 99.1) |
|
| ||||
| AFB+ | 124 | 75.0 (50.6, 90.4) | 99.0 (94.0, 99.9) | |
| AFB− | 43 | 50.0 (2.7, 97.3) | 92.7 (79.0, 98.1) | |
|
| ||||
| INH | 164 | 61.8 (43.6, 77.3) | 99.2 (95.2, 100.0) | |
|
| ||||
| AFB+ | 120 | 64.3 (44.1, 80.7) | 100.0 (95.0, 100.0) | |
| AFB− | 44 | 50.0 (13.9, 86.1) | 97.4 (84.6, 99.9) | |
|
| ||||
| MDR-TB | 159 | 77.8 (51.9, 92.6) | 99.3 (95.5, 100.0) | |
|
| ||||
| AFB+ | 117 | 76.5 (49.8, 92.2) | 100.0 (95.4, 100.0) | |
| AFB− | 42 | 100.0 (5.5 100.0) | 97.6 (85.6, 99.9) | |
Among the 914 Mtb culture positive specimens, 102 (11%) were classified as indeterminate for RIF resistance and 121 (13%) were classified as indeterminate for INH resistance by MTBDRplus v2 assay. The absence of the Mtb control probe was the primary (~70%) cause of indeterminate classification. A majority of the remaining of specimens were classified as indeterminate due to the presence of uninterpretable banding patterns. The proportion of indeterminates did not vary significantly by study site.
Correlates of Test Performance
The distribution of demographic and clinical factors among valid (resistant or susceptible) versus invalid (indeterminate) MTBDRplus v2 assay results are presented in Table 4. Univariate analysis was used to assess any significant difference in variable distribution. AFB smear grade of +1 or greater was significantly associated with valid results for both RIF and INH. Body mass index (BMI) appeared borderline significantly inversely associated with obtaining a valid result for INH, and significantly inversely associated with obtaining a valid result for RIF, indicating that those with a higher BMI were more likely to have an invalid test result. The likelihood of obtaining a valid test was not significantly associated with age, sex, HIV status, previous treatment, diabetic status, or study site.
Table 4.
Distribution of patient demographic and clinical characteristics by validity of MTBDRplus v2 assay results for rifampicin (RIF) and isoniazid (INH)*
| RIF results
|
INH results
|
|||||
|---|---|---|---|---|---|---|
| Valid (n=812) | Invalid (n=102) | p-value | Valid (n=793) | Invalid (n=121) | p-value | |
| Smear gradation | ||||||
| Negative (ref) | 98 (12.1) | 48 (47.1) | 92 (11.6) | 54 (44.6) | ||
| Rare or Scanty | 58 (7.1) | 17 (16.7) | 0.12 | 55 (6.9) | 20 (16.5) | 0.13 |
| Few (+1) | 171 (21.1) | 14 (13.7) | <0.001 | 166 (20.9) | 19 (15.7) | <0.001 |
| Many (+2) | 168 (20.7) | 7 (6.9) | <0.001 | 165 (20.8) | 10 (8.3) | <0.001 |
| TNTC (+3) | 317 (39.0) | 16 (15.7) | <0.001 | 315 (39.7) | 18 (14.9) | <0.001 |
| Age (continuous) | 33 (24,45) | 33 (24, 47) | 0.39 | 33 (24,45) | 33 (24, 46) | 0.63 |
| Male | 525 (64.7) | 61 (59.8) | 0.34 | 514 (64.8) | 72 (59.5) | 0.26 |
| HIV Positive (vs. Negative/Unknown) | 99 (12.2) | 14 (13.7) | 0.66 | 97 (12.2) | 16 (13.2) | 0.75 |
| Previous treatment | 629 (77.5) | 78 (76.5) | 0.82 | 619 (78.1) | 88 (72.7) | 0.20 |
| BMI† (continuous) | 18.0 (16.0, 20.4) | 19.6 (17.3, 21.8) | 0.005 | 18.0 (16.0, 20.4) | 18.8 (16.9, 21.3) | 0.06 |
| Diabetes | 42 (5.2) | 7 (6.9) | 0.49 | 43 (5.4) | 6 (5.0) | 0.83 |
| Site | ||||||
| India (ref) | 439 (53.8) | 53 (52.0) | 433 (54.6) | 59 (48.8) | ||
| Moldova | 204 (25.1) | 22 (21.6) | 0.67 | 193 (24.3) | 33 (27.3) | 0.33 |
| South Africa | 169 (21.4) | 27 (26.5) | 0.27 | 167 (21.1) | 29 (24.0) | 0.32 |
Categorical data are reported as percentages; continuous data as medians with interquartile ranges (Q1, Q3)
Weight (kg)/height (m)2
Three models were generated to determine the relationship between clinical and demographic factors and the likelihood of a valid or successful test (Table 5). The first model included only AFB smear gradation, the only significant factor associated with valid results for both RIF and INH in the univariate analysis. The second model included BMI to account for any residual association of bacilli load and test performance based on the assumption that individuals with low BMI would likely produce sputum with larger amounts of bacilli due to disease progression. The third model included three more variables, HIV status, a known confounder of test performance, and previous TB treatment and diabetic status, both associated with disease progression. The models were compared using the likelihood ratio test; models two and three were not statistically superior to model one. After controlling for smear status, BMI was no longer significantly associated with valid results in any of the models. HIV status, previous TB treatment status, BMI and diabetic status were not significantly associated with test performance (valid results). However, AFB smear grade was significantly associated with both RIF and INH test performance. Higher amounts of bacteria in the sputum as defined by AFB smear grade were associated with significantly more valid test results. A dose response relationship was evident; interpretable test results increased with each progressive AFB smear grade, culminating in an AFB smear +3 being approximately 10 times more likely to result in an interpretable MTBDRplus v2result compared to smear negative specimens.
Table 5.
Logistic regression of demographic and clinical characteristics and the odds of obtaining a valid (resistant or susceptible) versus invalid (indeterminate) MTBDRplus v2 result
| Characteristics | Model 1 OR (95%CI) |
Model 2 OR (95%CI) |
Model 3 OR (95%CI) |
|---|---|---|---|
| Rifampicin Resistance | |||
|
| |||
| Smear gradation | |||
| Negative (ref) | |||
| Rare | 1.7 (0.9, 3.2) | 1.6 (0.9, 3.1) | 1.8 (0.9, 3.4) |
| Few | 6.0 (3.1, 11.4)* | 6.6 (3.0, 11.1)* | 6.6 (3.4, 12.7)* |
| Many | 11.8 (5.1, 27.0* | 11.6 (5.1, 26.7)* | 12.4 (5.4, 28.8)* |
| TNTC | 9.7 (5.3, 17.8)* | 8.9 (4.8, 16.6)* | 10.6 (5.6, 20.1)* |
| BMI | 0.9 (0.9, 1.0) | 0.9 (0.9, 1.0) | |
| Previously Treated | 0.6 (0.4, 1.1) | ||
| HIV status | 1.3 (0.7, 2.7) | ||
| Diabetic status | 0.7 (0.3, 1.7) | ||
|
| |||
| Isoniazid Resistance | |||
|
| |||
| Smear gradation | |||
| Negative (ref) | |||
| Rare | 1.6 (0.9, 3.0) | 1.6 (0.9, 3.0) | 1.7 (0.9, 3.2) |
| Few | 5.1 (2.9, 9.2)* | 5.1 (2.8, 9.1)* | 5.3 (2.9, 9.6)* |
| Many | 9.7(4.7, 20.0)* | 9.6 (4.7, 19.8)* | 10.0 (4.8, 20.6)* |
| TNTC | 10.3 (5.7, 18.4)* | 10.0 (5.5, 17.9)* | 10.6 (5.8, 19.5)* |
| BMI | 1.0 (0.9, 1.0) | 0.9 (0.9, 1.0) | |
| Previously Treated | 0.9 (0.5, 1.4) | ||
| HIV status | 1.3 (0.7, 2.5) | ||
| Diabetic status | 1.0 (0.4, 2.5) | ||
P values meeting significance criterion with an alpha < 0.05
OR= odds ratio; CI = confidence interval
DISCUSSION
Overall, MTBDRplus v2 performed well among both AFB smear positive and negative specimens. However, performance of the test, particularly sensitivity, was affected by AFB smear status. Sensitivity for the detection of MDR-TB among AFB smear positive specimens was 96.0% (95%CI 93.5–97.6) compared to 82.8% (95%CI 63.5–93.5) among AFB smear negative specimens. This difference in sensitivity between AFB smear negative and positive specimens was more pronounced for INH resistance, 81.6 (95%CI 65.1–91.7) versus 94.9% (95%CI 92.4–96.6) than for RIF resistance, 91.4 (95%CI 75.8–97.8) versus 97.1% (95%CI 94.9–98.4). Additionally, the proportion of results that were valid was strongly correlated with AFB smear gradation. After adjusting for BMI, HIV status, diabetic status, and history of previous TB treatment, AFB smear +3 specimens were approximately 10 times more likely to result in a valid result when compared to AFB smear negative specimens.
Among the 914 culture positive specimens, less than 1% of MTBDRplus v2 test results were indeterminate due to test failure. Our observed rates of indeterminate assay results (11 to 13%) among culture positive specimens, were similar to those reported by WHO of 10 to 14%.6 If the study protocol had included repeat testing for indeterminate results, most likely indeterminate rates would have been comparable to those reported by Luetkemeyer et al. of 3 to 5% among culture positive specimens.21
Mtb was detected, as evidenced by hybridization of the Mtb probe, in 92% of all culture positive specimens (both AFB smear negative and positive). This result is slightly higher than two smaller studies evaluating MTBDRplus v2 among a mix of AFB smear positive and negative specimens, which reported sensitivities to Mtb of 73% and 88%.7, 8 When comparing performance among only AFB smear negative specimens, the Mtb sensitivity for the present study, 74%, was lower than the 79–80% reported by Crudu et al. and higher than the 57% reported by Bernard et al.7, 8 As a comparison, MTBDRplus v1, although not recommended for use on AFB smear negative samples, had previously reported Mtb detection rates among exclusively AFB smear negative specimens of 46%, 48% and 65%.21–23 When compared to Xpert MTB/RIF as reported by the Cochrane Review, the sensitivity of MTBDRplus v2 to Mtb among AFB smear positive specimens of 92% was lower than the pooled sensitivity of Xpert MTB/RIF of 98%; however the sensitivity of MTBDRplus v2 to Mtb among AFB smear negative specimens of 74%, was higher than the Xpert MTB/RIF sensitivity of 67%.24
The MTBDRplus v2 assay also detected Mtb among 24% of culture negative specimens in the current study. Probe hybridization in these culture negative cases may have been caused by the presence of non-viable bacteria in the specimen or may indicate that bacteria were present in amounts too small to sustain culture growth.
Direct comparison of assay results by smear status to previously published MTBDRplus v2 results was difficult as all previous studies assessed only small numbers of drug resistant specimens. However, the sensitivity and specificity of the test for RIF resistance in this study, stratified by smear status, was similar (within 95% confidence intervals) to those reported by Crudu et al.7 The only evident difference in INH resistance performance measures between the current study and the Crudu et al. study was the sensitivity and specificity of the test for AFB smear negative specimens. INH sensitivity and specificity among AFB smear negative specimens in this study was 81.6% (95%CI 65.1–91.7) and 98.1% (95%CI 88.6–99.9) respectively, compared to 93.5% and 82.3% reported by Crudu et al. N’guessan et al. reported RIF sensitivity and specificity only among AFB smear positive specimens, and the only evident difference in assay performance was the sensitivity of RIF resistance (73.2%), which was significantly lower than reported in the current study (97.1%, 95%CI 94.9–98.4).9 The sensitivity and specificity of MTBDRplus v2, as reported by Barnard et al., were 100% for both RIF and INH resistance, higher than both estimates from the current study.8
Given an absence of large diverse studies assessing the overall performance of MTBDRplus v2, sensitivity and specificity for RIF and INH could only be compared to the performance of MTBDRplus v1, as reported by the WHO Expert Group Report and a meta-analysis of the MTBDRplus v1 by Ling et al.6, 25 Overall, the sensitivity for RIF resistance found in this study (96.7%, 95%CI 94.5–98.0) was slightly lower than reported by the WHO (98.4%) and by Ling et al. (98.4%).6, 25 In contrast, overall sensitivity for INH resistance in this study (93.9% 95%C: 91.4–95.8) was slightly higher than reported by the WHO (91.4%) and by Ling et al. (88.7%). Specificities of both RIF and INH resistance from the current study were similar to those reported by the WHO and Ling et al.
Smear gradation was the strongest predictor of obtaining valid results (either susceptible or resistant) for either RIF and INH, even after adjusting for age, gender, BMI, previous history of TB treatment, HIV status, and diabetic status. As smear positivity increased, so did the odds of obtaining a valid assay result. This correlates with both the WHO findings for version 1 of the MTBDRplus assay and those reported by Singhal et al. for both version 1 and 2 of the MTBDRplus assay, where a significant association was found between smear gradation and the proportion of interpretable or valid results.6, 26
One notable limitation of this study was the lack of repeat testing of indeterminate MTBDRplus v2 assay results. This limitation likely contributed to a higher rate of indeterminate results than may have otherwise been reported if indeterminate MTBDRplus v2 assay results had been repeated.
CONCLUSIONS
The MTBDRplus v2 assay performed well among both AFB smear positive and negative specimens in the current study, sensitivity and specificity for the detection of RIF resistance, INH resistance, and MDR-TB by MTBDRplus v2 were 96.7% (95%CI 94.5–98.0) and 97.9% (95%CI 95.5–99.1), 93.9% (95%CI 91.4–95.8) and 99.6% (95%CI 97.7–100.0), and 95.1%(95%CI 92.6–96.9) and 99.1% (95%CI 97.1–99.8), respectively. Smear gradation appeared to influence both the probability of obtaining an interpretable result and test sensitivity and specificity, indicating a significant association between bacillary load and test performance. MTBDRplus v2 does however appear to consistently provide accurate results for both RIF and INH resistance.
Acknowledgments
The authors would like to acknowledge RL Jackson, N Hillery, and K Eisenach for their contribution to the study of laboratory data quality assurance and specimen validation. Specimens, data and funding for this project were provided by the Global Consortium for Drug-resistant TB Diagnostics ([GCDD] website: http://gcdd.ucsd.edu). This project was funded by the National Institute of Allergy and Infectious Diseases (NIH U01-AI082229), PI Antonino Catanzaro. MS receives funding through National Heart, Lung, and Blood Institute (NIH T32 HL098062-05). TCR receives funding through National Institute of Allergy and Infectious Diseases (P30 AI036214-20) and receives salary support from the Foundation for Innovative New Diagnostics (FIND), a nonprofit organization. The terms of this arrangement have been reviewed and approved by the University of California, San Diego in accordance with its conflict of interest policies. TCV acknowledges support from the SA MRC Centre for TB Research, DST/NRF Centre of Excellence for Biomedical Tuberculosis Research, Division of Molecular Biology and Human Genetics Stellenbosch University for infrastructure support.
References
- 1.Global tuberculosis report 2014. Geneva, Switzerland: WHO Press; 2014. [Google Scholar]
- 2.Bwanga F, Hoffner S, Haile M, Joloba ML. Direct susceptibility testing for multi drug resistant tuberculosis: a meta-analysis. BMC infectious diseases. 2009;9:67. doi: 10.1186/1471-2334-9-67. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Drobniewski F, Nikolayevskyy V, Maxeiner H, et al. Rapid diagnostics of tuberculosis and drug resistance in the industrialized world: clinical and public health benefits and barriers to implementation. BMC medicine. 2013;11:190. doi: 10.1186/1741-7015-11-190. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Ahmad S, Mokaddas E. Recent advances in the diagnosis and treatment of multidrug-resistant tuberculosis. Respiratory medicine. 2009;103:1777–90. doi: 10.1016/j.rmed.2009.07.010. [DOI] [PubMed] [Google Scholar]
- 5.Alcaide F, Coll P. Advances in rapid diagnosis of tuberculosis disease and anti-tuberculous drug resistance. Enfermedades infecciosas y microbiologia clinica. 2011;29(Suppl 1):34–40. doi: 10.1016/S0213-005X(11)70016-7. [DOI] [PubMed] [Google Scholar]
- 6.Organization WH, editor. Molecular Line Probe Assays for Rapid Screening of Patients at Risk of Multi-Drug Resistant Tuberculosis. 2008. [Google Scholar]
- 7.Crudu V, Stratan E, Romancenco E, Allerheiligen V, Hillemann A, Moraru N. First evaluation of an improved assay for molecular genetic detection of tuberculosis as well as rifampin and isoniazid resistances. Journal of clinical microbiology. 2012;50:1264–9. doi: 10.1128/JCM.05903-11. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Barnard M, Gey van Pittius NC, van Helden PD, Bosman M, Coetzee G, Warren RM. The diagnostic performance of the GenoType MTBDRplus version 2 line probe assay is equivalent to that of the Xpert MTB/RIF assay. Journal of clinical microbiology. 2012;50:3712–6. doi: 10.1128/JCM.01958-12. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.N’Guessan K, Assi JS, Ouassa T, et al. Assessment of the genotype MTBDRplus assay for rifampin and isoniazid resistance detection on sputum samples in Cote d’Ivoire. European journal of microbiology & immunology. 2014;4:166–73. doi: 10.1556/EUJMI-D-14-00014. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Li XX, Jiang SW, Zhang H, et al. Clinical and radiographic predictors in diagnosing sputum smear-negative pulmonary tuberculosis in HIV-negative patients: a cross-sectional study in China. Chinese medical journal. 2013;126:3662–7. [PubMed] [Google Scholar]
- 11.Gupta RK, Lawn SD, Bekker LG, Caldwell J, Kaplan R, Wood R. Impact of human immunodeficiency virus and CD4 count on tuberculosis diagnosis: analysis of city-wide data from Cape Town, South Africa. The international journal of tuberculosis and lung disease: the official journal of the International Union against Tuberculosis and Lung Disease. 2013;17:1014–22. doi: 10.5588/ijtld.13.0032. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Norbis L, Miotto P, Alagna R, Cirillo DM. Tuberculosis: lights and shadows in the current diagnostic landscape. The new microbiologica. 2013;36:111–20. [PubMed] [Google Scholar]
- 13.Putri FA, Burhan E, Nawas A, et al. Body mass index predictive of sputum culture conversion among MDR-TB patients in Indonesia. The international journal of tuberculosis and lung disease: the official journal of the International Union against Tuberculosis and Lung Disease. 2014;18:564–70. doi: 10.5588/ijtld.13.0602. [DOI] [PubMed] [Google Scholar]
- 14.Drain PK, Hyle EP, Noubary F, et al. Diagnostic point-of-care tests in resource-limited settings. The Lancet Infectious diseases. 2014;14:239–49. doi: 10.1016/S1473-3099(13)70250-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Hillery N, Groessl EJ, Trollip A, et al. The Global Consortium for Drug-resistant Tuberculosis Diagnostics (GCDD): design of a multi-site, head-to-head study of three rapid tests to detect extensively drug-resistant tuberculosis. Trials. 2014;15:434. doi: 10.1186/1745-6215-15-434. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Catanzaro A, Rodwell TC, Catanzaro DG, et al. Performance Comparison of Three Rapid Tests for the Diagnosis of Drug-Resistant Tuberculosis. PloS one. 2015;10:e0136861. doi: 10.1371/journal.pone.0136861. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Laboratory services in tuberculosis control: Microscopy Part II. WHO Press; 1998. [Google Scholar]
- 18.GmbH HL. GenoType MTBDRplus version 2.0: Instruction Manual. Nehren, Germany: 2012. [Google Scholar]
- 19.Updated Interim Critical Concentrations for First-ine and Second-line DST. World Health Organization; 2012. http://wwwstoptborg/wg/gli/assets/documents/Updated%20critical%20concentration%20table_1st%20and%202nd%20line%20drugspdf. [Google Scholar]
- 20.Newcombe RG. Improved confidence intervals for the difference between binomial proportions based on paired data. Statistics in medicine. 1998;17:2635–50. [PubMed] [Google Scholar]
- 21.Luetkemeyer AF, Kendall MA, Wu X, et al. Evaluation of two line probe assays for rapid detection of Mycobacterium tuberculosis, tuberculosis (TB) drug resistance, and non-TB Mycobacteria in HIV-infected individuals with suspected TB. Journal of clinical microbiology. 2014;52:1052–9. doi: 10.1128/JCM.02639-13. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Tukvadze N, Kempker RR, Kalandadze I, et al. Use of a molecular diagnostic test in AFB smear positive tuberculosis suspects greatly reduces time to detection of multidrug resistant tuberculosis. PloS one. 2012;7:e31563. doi: 10.1371/journal.pone.0031563. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Dorman SE, Chihota VN, Lewis JJ, et al. Genotype MTBDRplus for direct detection of Mycobacterium tuberculosis and drug resistance in strains from gold miners in South Africa. Journal of clinical microbiology. 2012;50:1189–94. doi: 10.1128/JCM.05723-11. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Steingart KR, Schiller I, Horne DJ, Pai M, Boehme CC, Dendukuri N. Xpert(R) MTB/RIF assay for pulmonary tuberculosis and rifampicin resistance in adults. The Cochrane database of systematic reviews. 2014;1:CD009593. doi: 10.1002/14651858.CD009593.pub3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Ling DI, Zwerling AA, Pai M. GenoType MTBDR assays for the diagnosis of multidrug-resistant tuberculosis: a meta-analysis. The European respiratory journal. 2008;32:1165–74. doi: 10.1183/09031936.00061808. [DOI] [PubMed] [Google Scholar]
- 26.Singhal R, Myneedu VP, Arora J, et al. Early detection of multi-drug resistance and common mutations in Mycobacterium tuberculosis isolates from Delhi using GenoType MTBDRplus assay. Indian journal of medical microbiology. 2015;33(Suppl):46–52. doi: 10.4103/0255-0857.150879. [DOI] [PubMed] [Google Scholar]