Outcome prediction in pregnancies of unknown location using endometrial thickness measurement: is this of real clinical value?

OBJECTIVE

To re-evaluate the role of measuring endometrial thickness (ET) in prediction of intrauterine pregnancy (IUP) among women with pregnancy of unknown location (PUL).

STUDY

DESIGN

987 women with PUL were included in a prospective observational multicenter study. Transvaginal ultrasonography was performed to measure ET and a blood sample was taken to measure serum β-hCG and progesterone levels. All patients were then managed expectantly till the final PUL outcome was diagnosed.

RESULTS

78 patients (8.9%) were finally diagnosed as having IUP. The best cutoff point of ET as a possible predictor for IUP was 10mm, with an area under receiver-operating characteristic (ROC) curve of 69.0%. At this cutoff point, ET was able to predict IUP with positive likelihood ratio (PLR) and negative likelihood ratio (NLR) of 1.43 and 0.19, respectively. Serum progesterone at a cutoff point of 50 nmol/L was able to predict IUP with PLR and NLR of 9.0 and 0.06, respectively. Variables showing statistically significant differences among those with IUP and those with the other PUL outcomes using univariate analysis (ET, gestational age, β-hCG, parity, serum progesterone and maternal age) were entered into logistic regression analysis. Logistic regression models were constructed. The performance of these models was better than using ET alone to predict the outcome of PUL.

CONCLUSION

Measurement of ET is not recommended as a single clinical test for intrauterine pregnancy prediction in women with pregnancy of unknown location.