Accuracy of the interferon-gamma release assay for the diagnosis of tuberculous pleurisy: an updated meta-analysis.

Cai-Shuang Pang,Chun Wan,Jing Zhu,Mei Feng,Fu-Qiang Wen,Pan-Wen Tian,Yong-Chun Shen

doi:10.7717/peerj.951

Abstract

Background and Objectives. The best method for diagnosing tuberculous pleurisy (TP) remains controversial. Since a growing number of publications focus on the interferon-gamma release assay (IGRA), we meta-analyzed the available evidence on the overall diagnostic performance of IGRA applied to pleural fluid and peripheral blood.Materials and Methods. PubMed and Embase were searched for relevant English papers up to October 31, 2014. Statistical analyses were performed using Stata and Meta-DiSc. Pooled sensitivity, specificity, positive likelihood ratio (PLR), negative likelihood ratio (NLR), positive predictive value (PPV), negative predictive value (NPV) and diagnostic odds ratio (DOR) were count. Summary receiver operating characteristic curves and area under the curve (AUC) were used to summarize the overall diagnostic performance.Results. Fifteen publications met our inclusion criteria and were included in the meta analysis. The following pooled estimates for diagnostic parameters of pleural IGRA were obtained: sensitivity, 0.82 (95% CI [0.79–0.85]); specificity, 0.87 (95% CI [0.84–0.90]); PLR, 4.94 (95% CI [2.60–9.39]); NLR, 0.22 (95% CI [0.13–0.38]); PPV, 0.91 (95% CI [0.85–0.96]); NPV, 0.79 (95% CI [0.71–0.85]); DOR, 28.37 (95% CI [10.53–76.40]); and AUC, 0.91. The corresponding estimates for blood IGRA were as follows: sensitivity, 0.80 (95% CI [0.76–0.83]); specificity, 0.70 (95% CI [0.65–0.75]); PLR, 2.48 (95% CI [1.95–3.17]); NLR, 0.30 (95% CI [0.24–0.37]); PPV, 0.79 (95% CI [0.60–0.87]); NPV, 0.75 (95% CI [0.62–0.83]); DOR, 9.96 (95% CI [6.02–16.48]); and AUC, 0.89.Conclusions. This meta analysis suggested that pleural IGRA has potential for serving as a complementary method for diagnosing TP; however, its cost, high turn around time, and sub-optimal performance make it unsuitable as a stand-alone diagnostic tool. Better tests for the diagnosis of TP are required.

Highlights

Tuberculous pleurisy (TP) is the most common form of extrapulmonary tuberculosis, accounting for 23% of all tuberculosis cases and 30% of cases of disease-causing pleural effusion (PE) (Vidal et al, 1986; Corbett et al, 2003; Valdes et al, 2003), which involves exudate containing primarily lymphocytes
This meta analysis suggested that pleural interferon-gamma release assay (IGRA) has potential for serving as a complementary method for diagnosing TP; its cost, high turn around time, and sub-optimal performance make it unsuitable as a stand-alone diagnostic tool
In the studies by Dheda et al (2009) and Kang et al (2012), IGRA was performed in two formats: as an enzyme-linked immunosorbent spot (ELISPOT) assay, and as an enzyme-linked immunosorbent assay (ELISA)

Summary

Introduction

Tuberculous pleurisy (TP) is the most common form of extrapulmonary tuberculosis, accounting for 23% of all tuberculosis cases and 30% of cases of disease-causing pleural effusion (PE) (Vidal et al, 1986; Corbett et al, 2003; Valdes et al, 2003), which involves exudate containing primarily lymphocytes. The tuberculin skin test is cross-reactive for Bacille Calmette Guerin (BCG) and many non-tuberculous mycobacteria, increasing the risk of misdiagnosis (Lawrence, 2000; Stead & To, 1987; Liebeschuetz et al, 2004) The limitations of these conventional approaches to diagnosing TP highlight the need to identify new diagnostic tools. The following pooled estimates for diagnostic parameters of pleural IGRA were obtained: sensitivity, 0.82 (95% CI [0.79–0.85]); specificity, 0.87 (95% CI [0.84–0.90]); PLR, 4.94 (95% CI [2.60–9.39]); NLR, 0.22 (95% CI [0.13–0.38]); PPV, 0.91 (95% CI [0.85–0.96]); NPV, 0.79 (95% CI [0.71–0.85]); DOR, 28.37 (95% CI [10.53–76.40]); and AUC, 0.91. This meta analysis suggested that pleural IGRA has potential for serving as a complementary method for diagnosing TP; its cost, high turn around time, and sub-optimal performance make it unsuitable as a stand-alone diagnostic tool.

Methods

Results

Conclusion