Abstract
BACKGROUND: Tuberculous meningitis (TBM) is one of the most serious types of extrapulmonary tuberculosis. However, low sensitivity of culture of cerebrospinal fluid (CSF) increases the difficulty in clinical diagnosis, leading to diagnostic delay, and misdiagnosis. Xpert MTB/RIF assay is a rapid and simple method to detect tuberculosis. However, the efficacy of this technique in diagnosing TBM remains unclear. Therefore, a meta-analysis was conducted to evaluate the diagnostic efficacy of Xpert MTB/RIF for TBM, which may enhance the development of early diagnosis of TBM. METHODS: Relevant studies in the PubMed, Embase, and Web of Science databases were retrieved using the keywords 'Xpert MTB/RIF', 'tuberculous meningitis (TBM)'. The pooled sensitivity, pooled specificity, positive likelihood ratio, negative likelihood ratio, diagnostic odds ratio, summary receiver operator characteristic curve, and area under the curve (AUC) of Xpert MTB/RIF were determined and analyzed. RESULTS: A total of 162 studies were enrolled and only 14 met the criteria for meta-analysis. The overall pooled sensitivity of Xpert MTB/RIF was 63% [95% confidence interval (CI), 59-66%], while the overall pooled specificity was 98.1% (95% CI, 97.5-98.5%). The pooled values of positive likelihood ratio, negative likelihood ratio, and diagnostic odds ratio were 20.91% (12.71-52.82%), 0.40% (0.32-0.50%), and 71.49% (32.64-156.56%), respectively. The AUC was 0.76. CONCLUSIONS: Xpert MTB/RIF exhibited high specificity in diagnosing TBM in CSF samples, but its sensitivity was relatively low. It is necessary to combine other high-sensitive detection methods for the early diagnosis of TBM. Moreover, the centrifugation of CSF samples was found to be beneficial in improving the sensitivity.