Tuberculous pleurisy is the most common extrapulmonary form of tuberculosis and remains diagnostically challenging because the pleural space is typically paucibacillary. Low bacillary burden reduces the sensitivity of conventional tests: culture yields are slow and limited, smear microscopy performs poorly, and nucleic acid amplification tests in pleural fluid have shown only modest sensitivity. As a result, clinicians frequently rely on invasive pleural biopsy for definitive diagnosis.
Analysis of cell-free DNA (cfDNA) — fragmented genetic material released into biofluids from host cells and pathogens — has emerged as a minimally invasive diagnostic approach. In tuberculous pleurisy, cfDNA may include M. tuberculosis–derived fragments that could be more readily detected than intact organisms in paucibacillary effusions. Molecular platforms such as quantitative PCR, droplet digital PCR (ddPCR), and metagenomic next-generation sequencing (mNGS) can target pathogen-specific sequences and may offer improved rapid detection compared with traditional methods. However, reported diagnostic accuracy across studies varies, likely reflecting differences in sample type (pleural fluid versus plasma/serum), assay technology, and pre-analytical and analytical procedures. A comprehensive synthesis of the evidence is lacking.
This protocol describes a systematic review and meta-analysis designed to:
The protocol follows PRISMA-P guidance and has been prospectively registered on PROSPERO (registration number CRD420261424501). The final systematic review will be reported according to PRISMA-DTA. Because the project will analyze published, publicly available data only, institutional ethics approval is not required.
Studies will be selected using a PIRT framework:
Eligible study designs include cross-sectional diagnostic accuracy studies, prospective or retrospective cohorts, and case–control studies, provided sufficient data exist to reconstruct a 2×2 contingency table (true positives, false positives, false negatives, true negatives). Studies that report only sensitivity or specificity without extractable 2×2 data will be excluded.
Excluded publications will include case reports, narrative reviews, editorials, commentaries, conference abstracts without full data, protocols, existing meta-analyses, and any study for which the full text is unavailable. Studies lacking sufficient data to populate a 2×2 table, even after attempts to contact authors, will also be excluded.
A comprehensive literature search will be performed in PubMed, Embase, Web of Science, Scopus, and The Cochrane Library from database inception to June 2027. The search strategy combines controlled vocabulary (MeSH/Emtree) and free-text terms addressing two core concepts: tuberculous pleurisy/pleural effusion, and cell-free DNA/circulating DNA. No language or date restrictions will be applied initially. Reference lists of included studies and relevant reviews will be manually screened to identify additional eligible reports.
The planned timeline reported in the protocol is: literature search between 1–30 June 2027; data extraction completed by 31 December 2027; and final results expected to be reported by 31 July 2028.
After deduplication in reference management software, two reviewers will independently screen titles and abstracts and perform full-text assessments against eligibility criteria. Discrepancies will be resolved by discussion and, if necessary, adjudication by a third senior reviewer.
Data extraction will use a piloted standardized form completed independently by two reviewers. Extracted items will include study identifiers (first author, year, country), study design and enrollment method, participant characteristics (sample size, age groups, HIV status, comorbidities), index test and sample details (platform, target gene(s) such as IS6110 or 16S rRNA, sample type, pre-analytical processing, extraction kits), components of the composite reference standard, and diagnostic performance data required to construct 2×2 contingency tables.
Risk of bias and applicability concerns will be assessed with the QUADAS-2 tool across four domains: patient selection, index test, reference standard, and flow and timing. Two reviewers will perform assessments independently.
The certainty of evidence for pooled sensitivity and specificity will be evaluated using the GRADE approach for diagnostic tests, considering risk of bias, inconsistency, indirectness, imprecision, and publication bias. Evidence will be graded as high, moderate, low, or very low.
When at least four eligible studies are identified, a bivariate random-effects meta-analysis will be conducted (planned using Stata v18.0) to jointly pool sensitivity and specificity with 95% confidence intervals, accounting for within-study variance and the correlation between sensitivity and specificity. Summary PLR, NLR, and DOR with 95% CIs will be calculated, and a hierarchical SROC curve will be produced.
Subgroup analyses and meta-regression are planned to explore heterogeneity by sample type (pleural fluid versus plasma/serum), detection platform (e.g., ddPCR, mNGS), target genes, patient characteristics (age groups, HIV status), and pre-analytical variables. If quantitative pooling is not appropriate, a narrative synthesis of study-level findings will be provided.
Publication bias and small-study effects will be evaluated where appropriate, and sensitivity analyses will examine the impact of study quality and reference standard definitions on pooled estimates.
The review aims to provide pooled estimates of diagnostic sensitivity and specificity for cfDNA assays in tuberculous pleurisy, evaluate clinical utility via likelihood ratios and DOR, and identify methodological and pre-analytical factors associated with optimal performance. The authors note that all supporting files are available on the OSF repository linked in the protocol. The protocol specifies that no original datasets were generated at the time of publication and that all relevant data will be shared upon study completion.
The work is supported by a Hangzhou Biomedicine and Health Industry Development Supporting Science and Technology Special Project (Yanqin Shen, 2023WJC162). The funder is stated not to have played a role in study design, data collection and analysis, decision to publish, or manuscript preparation. The authors declared no competing interests.
This prospectively registered protocol sets out a structured approach to synthesize diagnostic accuracy data for cfDNA in tuberculous pleurisy, using established methodological frameworks (PRISMA-P, PRISMA-DTA, QUADAS-2, GRADE). By pooling performance metrics and exploring sources of heterogeneity, the planned review intends to inform clinical application and standardization of cfDNA assays for this challenging extrapulmonary manifestation of tuberculosis.