Background:

No preeclampsia screening test has been validated in our country.

Objective:

To assess the fit and performance of the FMF 4.0 Bayesian algorithm in a Mexican population.

Method:

Cohort study in singleton pregnancies, according to the competing risks model for preeclampsia. A priori risk was calculated based on medical history. Mean blood pressure (MBP), uterine artery mean pulsatility index (UtAmPI) and pregnancy-associated plasma protein A (PAPP-A) were measured at 11-14 weeks of gestation using standardized methodology. The value of each marker was transformed into multiples of the median (MoM) by the originala algorithm and by the population correctionb. Multivariate normal distribution and Bayes theorem were applied to obtain the post-test probabilities.

Results:

There was 5.0% (40/807) preeclampsia. Distributions in women without preeclampsia were MBP = 0.927a vs. 0.991b, UtAmPI = 0.895a vs. 1.030b, and PAPP-A = 0.815a vs 0.963b (p < 0.001). The AUC was 0.821a and 0.823b. For a detection rate of 69%, cutoff points were 1 in 240a and 1 in 120b, with false positive rates of 31%a and 29%b.

Conclusions:

The model has to be adjusted to the target population.