First‐trimester ultrasound features associated with subsequent miscarriage: A prospective study. (6th February 2019)

• Record Type: Journal Article
• Authors: Taylor, Tracey J.; Quinton, Ann E.; de Vries, Bradley S.; Hyett, Jonathon A.
• Is Part Of: Australian and New Zealand journal of obstetrics and gynaecology. Volume 59:Number 5(2019)
• Page Start: 641
• Page End: 648
• Publication Date: 2019-02-06
• Subjects: first trimester -- logistic regression -- miscarriage -- trophoblast -- ultrasound -- uterine artery; Obstetrics -- Periodicals; Gynecology -- Periodicals; 618.05
• DOI: 10.1111/ajo.12944
• Languages: English
• ISSNs: 0004-8666

Abstract

Abstract : Background: First‐trimester miscarriage is common, with women increasingly offered an ultrasound scan early in the first trimester to assess the status of their pregnancy. Ultrasound is uniquely situated to significantly impact the clinical management of these women. Aims: This study aims to determine whether there were any differences in the early ultrasound appearances of pregnancies that continued to be viable or resulted in miscarriage before 12 weeks gestation. Materials and Methods: This was a prospective cohort study including ultrasound measurements: mean sac diameter (MSD), yolk sac diameter (YSD), crown‐rump length (CRL), fetal heart rate (FHR), trophoblast thickness, trophoblast volume (TTV) and mean uterine artery pulsatility index (meanUAPI). Regression models were fitted for each parameter and Z‐scores compared between cohorts that progressed or miscarried after the scan but before 12 weeks gestation. Logistic regression analysis was used to create a prediction model for miscarriage prior to 12 weeks gestation based on the standardised ultrasound measurements recorded during the early first‐trimester scan. Results: Comparison of Z‐Scores for meanUAPI, TTV, FHR and MSD demonstrated significant variation between the two groups. The proposed logistic regression model resulted in an area under the receiver operator curve of 0.81. At a false‐positive rate of 30%, the model resulted in a sensitivity of 76% (95% CI 64–89%). Conclusion: The combination of FHR, meanUAPI, TTV in a prediction model for miscarriage may prove to be of value for ongoing pregnancy management in the first trimester.