A Logistic Regression Model Including Dna Status and Morphology of Spermatozoa for Prediction of Fertilization in Vitro

Abstract

To determine predictive values of routine semen analysis, sperm morphology evaluation using strict criteria and DNA status for in-vitro fertilization (IVF), 66 consecutive couples undergoing IVF in a university hospital IVF programme were prospectively investigated. Semen samples from 66 men were evaluated by routine semen analysis, morphology evaluation using strict criteria and acridine orange staining for determination of DNA status. A new technique is described for acridine orange scoring which consisted of evaluation of two smears per case, with and without heat treatment. Resistance to heat-provoked denaturation was determined by the difference between two evaluations. A logistic regression model was built and receiver operating characteristic curves were constructed to determine the threshold values and to compare diagnostic properties. Morphology evaluation using strict criteria and concentration of progressively motile spermatozoa were found to be the principal parameters determining the sperm fertilizing capacity in vitro. The logistic regression model composed of morphology evaluation using strict criteria and acridine orange score had a powerful diagnostic capability for prediction of fertilization in vitro.