Strengths and Limitations
The main strength of this study is that we have a considerably large sample of Brazilian women from 17 centers in three regions, thus covering possible internal population differences. In Brazil, previous TVU performance analyses to predict PTB were from single-center studies18,23 with smaller samples. All cervical measurements were performed by expert medical sonographers in tertiary reference centers, along with checking of the ultrasound images to correct and reinforce the pattern technique. We analyzed TVU using different accuracy tests, different cutoff points and specific PTB subgroups for gestational age.
The vaginal progesterone used for women with CL ≤30mm is a limitation in our study because progesterone reduces the occurrence of PTB. Nevertheless, in our prenatal clinical assistance, women with CL ≤25mm are encouraged to use progesterone, so maintaining this intervention in our sample allows the possibility to pragmatically infer the results to medical practice. Unfortunately, we cannot identify if progesterone has caused any reduction in PTB between women with CL 25-≤30mm, which could have underestimated PTB incidence in this subgroup. Another limitation is that some participating centers did not perform universal TVU screening, which could introduce some selection bias in our sample and the tendency to have a shorter CL. However, the mean CL identified was very similar to other previous Brazilian studies16,24,25.