CRD summary

This generally well-conducted review concluded that microscopic-observation drug susceptibility and thin layer agar assays were inexpensive and rapid alternatives to conventional methods for drug susceptibility testing of Mycobacterium tuberculosis. The authors acknowledged some weaknesses in the available evidence but their conclusions are likely to be reliable.

Authors' objectives

To investigate the accuracy of microscopic-observation drug susceptibility (MODS) and thin-layer agar assays for rapid screening of patients at risk of drug-resistant Mycobacterium tuberculosis (TB).

Searching

PubMed, EMBASE and BIOSIS Previews were searched for studies published in English, French, or Spanish from 1990 to February 2009; search terms were reported. Reference lists from selected studies were hand-searched. Experts and test developers were contacted to identify further studies.

Study selection

Studies of microscopic-observation drug susceptibility or thin-layer agar assays used to assess drug susceptibility of M. tuberculosis compared with an accepted reference standard were eligible for inclusion. Accepted reference standards were indirect proportion methods, absolute concentration, resistance ratio, commercial liquid systems, or microdilution methods. Studies using direct (direct patient specimen) or indirect (previously isolated strain of M. tuberculosis) inoculation were included.

The most commonly tested drugs were rifampicin and isoniazid. Most studies used the proportion method as the reference standard, and direct application of a specimen from a patient to drug-containing and drug-free media using sputum specimens. Most studies were conducted in South American countries.

Study selection was conducted by two independent reviewers; disagreements were resolved by consensus.

Assessment of study quality

Study quality was assessed using criteria on blinded interpretation of test results, verification of all tests with the reference standard, consecutive or random recruitment of patients/collection of specimens, and study design.

The number of reviewers performing the quality assessment was not reported.

Data extraction

Data were extracted to produce 2x2 tables of test performance, from which sensitivity and specificity, with 95% confidence intervals (CI) were calculated. Other outcomes extracted included turnaround time for processing of specimens, specimen contamination rate, and costs.

Two reviewers independently extracted data; disagreements were resolved by consensus.

Methods of synthesis

Pooled estimates of sensitivity and specificity, with 95% confidence intervals, were calculated separately for MODS assay and thin-layer agar assay using a bivariate random-effects model. Heterogeneity was assessed using I².

Subgroup analyses investigated the drug concentration used to define the resistance cut-off; these were combined if the confidence interval for individual cut-off levels overlapped. Hierarchical summary receiver operating characteristic (HSROC) curves were used to investigate the effect of the thresholds, and produce an overall summary of test accuracy. Subgroups with four or fewer studies were combined using univariate random-effects models due to lack of convergence with bivariate models. Rates of contamination were compared with rates seen in solid and liquid culture reference standards. Combination rates were combined using weighted means and turnaround times using simple means.

Results of the review

Twelve studies met the inclusion criteria (n=1,913 patients, range 53 to 338). Six studies were prospective, three retrospective and three were of unclear direction of data collection. All studies avoided partial verification bias, 50% of studies recruited participants consecutively or randomly, and 50% used blinded interpreters of the index test.

Microscopic-observation drug susceptibility (MODS) assay: (nine studies; n=1,474 patients): Testing for M. tuberculosis rifampicin resistance (eight studies), sensitivity was 98% (95% CI 94.5 to 99.3) and specificity was 99% (95% CI 95.7 to 99.9). Testing for isoniazid resistance with a 0.1μg/mL cut-off (six studies), sensitivity was 98% (95% CI 94.4 to 99.1) and specificity was 96% (95% CI 88.1 to 98.6); with a 0.4μg/mL cut-off (seven studies), sensitivity was 90% (95% CI 84.5 to 93.7) and specificity was 99% (95% CI 96.9 to 99.4).

Thin-layer agar assay: (three studies; n=439 patients) Diagnostic sensitivity and specificity was 100% for all thin-layer agar tests for rifampicin and isoniazid resistance.

Turnaround times were slightly shorter and contamination rates higher for MODS compared with thin-layer agar assay (where reported); further results are reported. Results were provided for other drugs; these were tested in fewer studies. Indirect inoculation resulted in higher sensitivity and specificity than direct inoculation, but this was only done in four studies. Quality criteria showed no impact on the results. Contamination rates were lower for the thin-layer agar than the MODS assay, but data were sparsely reported.

Cost information

Across the four studies that reported on costs, the cost of MODS ranged from 1.38 US dollars ($) to $2 .17 per sample; thin-layer agar assay costs ranged from $1.60 to $2.92. Some of the variability in cost seemed to depend on the drug being tested. Estimates for the cost for an incubator were $8,000 and $4,000 for an inverted microscope.

Authors' conclusions

Microscopic-observation drug susceptibility and thin-layer agar assays were inexpensive and rapid alternatives to conventional methods for drug susceptibility testing of Mycobacterium tuberculosis.

CRD commentary

The review addressed a clear research question supported by appropriate inclusion criteria. Several relevant sources were searched, but the restriction to published studies in three languages meant publication and language bias could not be ruled out. Study selection and data extraction were conducted in duplicate, but it was unclear whether similar methods were employed to reduce error and bias during the quality assessment.

Appropriate criteria were used to assess study quality, although only summary results were presented, so it was not possible to identify which studies in the analyses were prone to which methodological weaknesses. However, the impact of study quality on the results was investigated. Appropriate methods of synthesis were employed; heterogeneity was investigated.

This was a generally well-conducted review, the authors acknowledged the weaknesses of the available evidence, and their conclusions are likely to be reliable.

Implications of the review for practice and research

Practice: The authors stated that this research informed the World Health Organization (WHO) recommendation for use of selected non-commercial drug susceptibility tests, including MODS, as an interim solution until capacity for genotypic or automated liquid culture drug susceptibility testing is developed.

Research: The authors stated that direct comparisons of direct versus indirect inoculation are required. They also highlighted the need for the continued development and evaluation of novel tuberculosis diagnostics due to the time to detection and the inability to detect drug resistant tuberculosis.

Funding

World Health Organization, Stop TB Department; individual authors received funding from a range of organisations.

Bibliographic details

Minion J, Leung E, Menzies D, Pai M. Microscopic-observation drug susceptibility and thin layer agar assays for the detection of drug resistant tuberculosis: a systematic review and meta-analysis. Lancet Infectious Diseases 2010; 10(10): 688-698. [PubMed: 20813587]

Original Paper URL

http://www.thelancet.com/journals/laninf/article/PIIS1473-3099(10)70165-1/abstract

Indexing Status

Subject indexing assigned by NLM

MeSH

Antitubercular Agents /pharmacology; Drug Resistance, Bacterial; Humans; Microbial Sensitivity Tests /economics /methods; Microscopy /methods; Mycobacterium tuberculosis /drug effects /isolation & purification; ROC Curve; Sensitivity and Specificity; Time Factors; Tuberculosis, Multidrug-Resistant /diagnosis /microbiology

AccessionNumber

12010007158

Database entry date

24/08/2011

Record Status

This is a critical abstract of a systematic review that meets the criteria for inclusion on DARE. Each critical abstract contains a brief summary of the review methods, results and conclusions followed by a detailed critical assessment on the reliability of the review and the conclusions drawn.