Using Symptom Scores, Lifestyle Measures and Biochemical Markers to Create a Test for Endometriosis

• Published in: Journal of endometriosis and pelvic pain disorders 2014-07, Vol.6 (3), p.135-143
• Main Authors: Paiva, Premila, Lappas, Martha, Barker, Gillian, Healey, Martin
• Format: Article
• Language: English
• ISSN: 2284-0265; 2284-0273
• DOI: 10.5301/je.5000194

Purpose The purpose of this study was to develop a test to discriminate between women suffering from pelvic pain associated with presence or absence of endometriosis, using symptom visual analogue scale (VAS) scores, demographic and lifestyle factors and known and novel plasma biomarkers. Methods This was a prospective cohort study. We recruited women presenting with pelvic pain symptoms and undergoing laparoscopy. They provided a plasma sample and completed a questionnaire assessing symptom VAS scores, lifestyle factors and demographic details. Plasma was assayed for 18 potential biomarkers. Logistic regression was used to develop a test to discriminate between women with pelvic pain associated with presence or absence of endometriosis. Results A series of 101 women were analyzed. On univariate analysis, women with endometriosis were less likely to be parous or have previously used an intrauterine contraceptive device (IUD) but have a higher weekly alcohol intake. These women had higher VAS scores for nausea and dyspareunia, and higher plasma levels of cancer antigen 125 (CA-125). The optimum logistic regression model included parity, previous IUD use, history of endometriosis, weekly alcohol intake, dyspareunia, VAS score and CA-125. This model had a sensitivity of 93%, specificity of 62% and accuracy of 83%. Conclusions Combining symptom scores, historical measures and CA-125 provides a reasonable means to discriminate between women with pelvic pain associated with presence or absence of endometriosis, but greater specificity is needed before such a model could replace laparoscopy.