Polymerase Chain Reaction for the Diagnosis of HIV Infection in Adults

摘要: Purpose : To do a meta-analysis of studies that have evaluated the sensitivity and specificity polymerase chain reaction (PCR) assay for diagnosis human immunodeficiency virus (HIV) infection in adults. Evaluating performance PCR is difficult because certain clinical situations, or may exceed those current reference standard tests (enzyme immunoassay followed by confirmatory Western blot analysis). Therefore, an additional goal was to develop recommendations 1) design future evaluative 2) use persons with suspected HIV infection. Data Sources Studies published between 1988 1994 were identified search 17 computer databases, including MEDLINE, abstracts from conference proceedings. Study Selection included if DNA amplification done on peripheral blood mononuclear cells Ninety-six met inclusion criteria. Extraction extracted independently two reviewers. assessed investigators blinded study results. Results Reported sensitivities range 10% 100%, specificities 40% 100%. A summary receiver-operating characteristic curve based all 96 has maximum joint (upper left point curve, where equals specificity) 97.0% 98.1%. If threshold value defines positive result chosen so higher than 98.1%, will decrease less Conversely, greater are be equal, corresponding false-positive rate 1.9% 3.0%. At specificity, predictive ranges 34% 85% as prevalence increases 1.0% 10%. We seven areas which could modified reduce susceptibility bias estimates increase generalizability These modifications also help overcome methodologic problems created lack test. Conclusions The not sufficiently accurate used without confirmation. Use adults should limited situations antibody known insufficient. Future large adequate standardized methods PCR. Specimens status results other diagnostic