Evaluation of the diagnostic accuracy of laboratory-based screening for hepatitis C in dried blood spot samples: A systematic review and meta-analysis

The dried blood spot (DBS) is increasingly used for the hepatitis C virus (HCV) screening. Our objective was to perform a meta-analysis of the methodology for HCV screening in DBS samples, particularly in the type of diagnostic assay used. We performed a meta-analysis of all eligible studies published to date (March 2018). The literature search revealed 26 studies: 21 for detection of anti-HCV antibodies and 10 for detection of HCV-RNA. Statistical analyses were performed using Meta-DiSc and STATA (MIDAS module). For detection of HCV antibodies, pooled diagnostic accuracy measures were as follows: sensitivity 96.1%, specificity 99.2%, positive likelihood ratio (PLR) 105, negative likelihood ratio (NLR) 0.04, diagnostic odds ratio (DOR) 2692.9, and summary receiver operating characteristic (SROC) 0.997 +/- 0.001. For detection of HCV-RNA, the pooled diagnostic accuracy measures were as follows: sensitivity 97.8%, specificity 99.2%, PLR 44.8, NLR 0.04, DOR 1966.9, and SROC 0.996 +/- 0.013. Similar values of pooled diagnostic accuracy measures were found according to the type of anti-HCV antibody detection assay (enzyme-linked immunosorbent assay, rapid diagnostic test, and chemiluminescence assays) and HCV-RNA detection assay (real-time polymerase chain reaction and transcription-mediated amplification). The analysis of external validity showed a high negative predicted value (NPV) for both approaches, but a low positive predicted value (PPV) when prevalence was < 10%, particularly in HCV-RNA tests. Finally, this meta-analysis is subject to limitations, especially publication bias and significant heterogeneity between studies. In conclusion, HCV screening in DBS samples has an outstanding diagnostic performance, with no relevant differences between the techniques used. However, external validity may be limited when the HCV prevalence is low.