RESUMO
BACKGROUND: Rapid and accurate diagnosis of herpes simplex virus (HSV)-1 and -2 (HSV1/2) in cerebrospinal fluid (CSF) is important for patient management. OBJECTIVES: Summarize the diagnostic accuracy of commercial rapid sample-to-answer PCR assays (results in <90 minutes, without a separate nucleic acid extraction step) for HSV1/2 detection in CSF. DATA SOURCES: Four databases (MEDLINE, EMBASE, Scopus, and CENTRAL) and five conference abstract datasets from January 2012 to March 2022. STUDY ELIGIBILITY CRITERIA: Eligible diagnostic accuracy studies provided sufficient data for the construction of a standard diagnostic accuracy two-by-two table. PARTICIPANTS: Patients with suspected meningitis and/or encephalitis. TESTS: FilmArray Meningitis-Encephalitis Panel and Simplexa HSV 1&2 Direct Kit PCR. REFERENCE STANDARD: Real-time PCR assay. ASSESSMENT OF RISK OF BIAS: Two investigators independently extracted data, rated risk of bias, and assessed quality using QUADAS-2. METHODS OF DATA SYNTHESIS: Accuracy estimates were pooled using Bayesian random effects models. RESULTS: Thirty-one studies were included (27 FilmArray; 4 Simplexa), comprising 9924 samples, with 95 HSV-1 and 247 HSV-2 infections. Pooled FilmArray sensitivities were 84.3% (95% credible interval, 72.3-93.0) and 92.9% (95% credible interval (CrI), 82.0-98.5) for HSV-1 and HSV-2, respectively; specificities were 99.8% (95% CrI, 99.6-99.9) and 99.9% (95% CrI, 99.9-100). Pooled Simplexa sensitivities were 97.1% (95% CrI, 88.1-99.6) and 97.9% (95% CrI, 89.6-99.9), respectively; specificities were 98.9% (95% CrI, 96.8-99.7) and 98.9% (95% CrI, 97.1-99.7). Pooled FilmArray sensitivities favoured industry-sponsored studies by 10.0 and 13.0 percentage points for HSV-1 and HSV-2, respectively. Incomplete reporting frequently led to unclear risk of bias. Several FilmArray studies did not fully report true negative data leading to their exclusion. CONCLUSIONS: Our results suggest Simplexa is accurate for HSV1/2 detection in CSF. Moderate FilmArray sensitivity for HSV-1 suggests additional testing and/or repeat CSF sampling is required for suspected HSV encephalitis when the HSV-1 result is negative. Low prevalence of HSV-1 infections limited summary estimates' precision. Underreporting of covariates limited exploration of heterogeneity.
