A multiple regression model for predicting a high cytomegalovirus immunoglobulin G avidity level in pregnant women with IgM positivity

Abstract Objective To establish a model to predict a high cytomegalovirus (CMV) immunoglobulin (Ig)G avidity index (AI) using clinical information to contribute to the mental health of CMV-IgM positive pregnant women. Methods We studied 371 women with IgM positivity at ≤14 weeks of gestation. Information on the age, parity, occupation, clinical signs, IgM and G values, and IgG AI was collected. The IgG AI cut-off value for diagnosing congenital infection was calculated based on a receiver operating characteristic curve analysis. Between-group differences were assessed using the Mann–Whitney U-test or χ2 analysis. The factors predicting a high IgG AI were determined using multiple logistic regression. Results The women were divided into high or low IgG AI groups based on the IgG AI cut-off value of 31.75. There were significant differences in the IgG and IgM levels, age, clinical signs, and number of women with one parity between the two groups. In multiple logistic regression analysis, IgM and the number of women with one parity were independent predictors. This result helped us to establish a mathematical model which correctly classified the IgG AI level for 84.6% women. Conclusion We established a highly effective model for predicting a high IgG AI immediately after demonstrating IgM positivity.