Liquid-based cervical cytologic smear study and conventional Papanicolaou smears: a metaanalysis of prospective studies comparing cytologic diagnosis and sample adequacy.

OBJECTIVE We sought to evaluate the cytologic diagnosis and sample adequacy of the liquid-based cervical cytologic smear (ThinPrep) compared with that of the conventional Papanicolaou smear. STUDY DESIGN Prospective studies of ThinPrep and conventional Papanicolaou smears were analyzed for cytologic diagnosis and sample adequacy. Computerized databases, references in published studies, and index reviews published in English were used to identify direct-to-vial and split-sample clinical trials of cervical smears performed by conventional and liquid-based techniques. Only published studies that used the Bethesda system nomenclature with clearly documented outcome data were included. Each trial was assessed for the quality of its method, inclusion and exclusion criteria, adequacy of randomization, sampling protocols, definition of outcome, and statistical analyses. RESULTS Twenty-five studies met inclusion criteria for this review. Odds ratios with 95% confidence intervals were calculated for each outcome. Estimates of odds ratios and risk differences for dichotomous outcomes were calculated by use of random and fixed-effects models. Homogeneity was tested across the studies. Results indicate that the ThinPrep test is as good as or superior to the conventional Papanicolaou smear in diagnosing uterine cervical premalignant abnormalities. Also the ThinPrep test provides improved sample adequacy when compared with the conventional Papanicolaou test. CONCLUSION The ThinPrep test improved sample adequacy and led to improved diagnosis of low-grade and high-grade squamous intraepithelial lesions. However, there is no difference in the rate of atypical cells of undetermined significance diagnosis between ThinPrep and conventional smear groups. The added cost of ThinPrep cytologic screening and, hence, its cost-effectiveness are not evaluated in this study.