We read with great interest the paper by Scala et al who assessed the value of reduced fetal movements (RFMs), showing that repeated RFMs at term were associated strongly with small-for-gestational-age (SGA) fetuses and infants. Moreover, they found that RFMs were more likely to occur in women with high second-trimester uterine artery (UtA) Doppler resistance indices. This conclusion was based on a logistic regression analysis.

These are very interesting findings with important clinical implication, because SGA fetuses or placental dysfunction should be suspected in cases of RFMs and be managed as such. This article addresses a relevant question about the predictive value of one vs repeated RFMs, which is a frequent indication for emergency consultation.

However, from our point of view, some methodologic concerns should be answered because they may lead to erroneous results. First, the quantitative variables (maternal age, body mass index, PAPP-A, mean UtA-Pulsatility Index [PI]) were entered in the model as quantitative predictors, which relies on the existence of a linear relation between the predictor and the outcome that is not assessed in the Methods section. For example, it is not obvious that the relation between the fetal weight and the maternal age is linear. Moreover, the relevance of the use of the linear mean UtA-PI is questionable, because most studies use the 90e percentile as a predictor of adverse fetal outcome. Caution should be taken regarding this point because it could lead to a misinterpretation of the results.

Second, ethnicity seems to have the “other” class as reference (odds ratio, 1) in the logistic regression. In Table 2, “other” includes only 0.7% of the whole population. The “caucasian” class would have been more relevant as a reference.

Finally, a short sentence above the table explains that “values are adjusted for other confounding variable (maternal age, body mass index, ethnicity). We wonder how and why were these 3 factors chosen? Were they known as confounding factors for RFM and SGA?

Although these results seem promising, more details regarding the statistical method would improve the clinical relevance of the study.