Evaluation of interferon-gamma release assays for the diagnosis of tuberculosis: an updated meta-analysis

Abstract

The objective of this investigation was to systematically evaluate the diagnostic accuracy of interferon-gamma release assays (IGRAs) for tuberculosis disease. Both English and Chinese databases were searched for relevant articles through January 2012. We included studies that were restricted to diagnostic applications of IGRAs in patients with active tuberculosis and excluded studies performed in the immune-compromised population. We used Meta-DiSc software to handle the data. We calculated the sensitivity, specificity, positive likelihood ratio (PLR), negative likelihood ratio (NLR), diagnostic odds ratio (DOR), and 95 % confidence interval (CI) for each study. We also calculated the pooled sensitivity, specificity, PLR, NLR, DOR, and produced forest plots and summary receiver operating characteristic (SROC) curves. A total of 61 papers (73 studies) were eligible for meta-analysis, including 36 published in English and 25 published in the Chinese language. The overall sensitivity, specificity, PLR, NLR, DOR, and 95 % CI of IGRAs were 0.85 (95 % CI: 0.84-0.86), 0.84 (95 % CI: 0.83-0.85), 7.82 (95 % CI: 6.01-10.19), 0.17 (95 % CI: 0.14-0.21), and 59.27 (95 % CI: 40.19-87.42), respectively. For ten studies evaluating T-SPOT.TB in China, the combined sensitivity, specificity, PLR, NLR, DOR, and 95 % CI were 0.88 (95 % CI: 0.86-0.91), 0.89 (95 % CI: 0.86-0.92), 8.86 (95 % CI: 5.42-14.46), 0.13 (95 % CI: 0.10-0.17), and 88.15 (95 % CI: 41.76-186.07), respectively. The SROC area under the curve (AUC) was 0.9548 (95 % CI: 0.9323-0.9773). Though IGRAs showed good sensitivity and specificity for the detection of tuberculosis in this meta-analysis, the decision to use an IGRA should be based on the local prevalence of the disease and the country guidelines, as well as resources and logistical considerations.