An assessment of the subjectivity of sperm scoring

Current histological investigation of vaginal swabs after alleged sexual assault includes the scoring of spermatozoa (0, + to ++++) and the recording of visible tails. It is a method that is universally employed. Despite this method being used for 40 years, there has never been a study investigating its suitability for forensic science. Here, we investigate the reproducibility and subjectivity of sperm scoring among different investigators. Dilutions of seminal fluid were randomly distributed onto 20 slides, stained with haematoxylin/eosin and assessed by 37 investigators, over 2 years. Slides were assessed for levels of spermatozoa and the presence of tails. Each slide was scored by a minimum of 25 investigators. On no slide was there a consensus between all scores. Standard deviation remained below 1, but relative standard deviation (RSD) ranged from 6 to 105% in a positive correlation as the average score decreased. Spermatozoa were not observed 56 times (9.6%) and 27 investigators (73%) did not observe spermatozoa on at least one slide. Spermatozoa with tails were observed on every slide by at least 10 examiners, but as the average score of the slide decreased, so did the observation of tails. The current sperm scoring method is highly subjective with a particularly high %RSD in slides with low overall sperm counts. Moreover, the recording of tails does not add value to the current technique of sperm scoring. Further research might improve the objectivity of sperm scoring and the reliability of recording of tails