Transvaginal ultrasonography examination of the uterine cervix in hospitalised women undergoing preterm labour

Objective: Our purpose was to assess the accuracy of cervical mensuration ( length, wedging) by transvaginal ultrasonography in the prediction of prete rm birth among hospitalised women undergoing preterm labour. Method: Eviden ce was extracted from articles through a MEDLINE search from 1987 to 1997, followed up by examining the references at the end of pertinent articles. T he key words were: pregnancy, cervix, ultrasonography, preterm labour, pret erm delivery, cervical length, cervical wedging or funnelling. Final criter ia for the selection of the studies were cervical length and/or wedging at internal os in hospitalised women undergoing preterm labour without restric tion on the study design. Results: Nine articles from 12 publications were analysed and incorporated in a discussion. Study design and selection crite ria vary among publications. Preterm delivery was the principal outcome, bu t its definition is not the same according to the authors. The best cut-off for cervical length varies from 18 to 30 mm. The size of the samples is li mited. The sensitivity for predicting preterm birth ranges from 68 to 100%, and the specificity ranges from 30 to 78%. Cervical wedging has a sensitiv ity varying from 23 to 100%, and a specificity ranging from 54 to 90%. Conc lusion: Ultrasonography of the cervix is an interesting technology as it is cheap and easy to use. However this technique remains to be assessed befor e implementing it in daily clinical practice. Finally, studies incorporatin g transvaginal ultrasonography into interventional trials are needed to det ermine the efficacy of measuring cervical length in the prevention of prete rm birth. (C) 2000 Elsevier Science Ireland Ltd. All rights reserved.