Periventricular echodensities in preterm infants

Abstract

Sir, Periventricular echodensities (PVE) in preterm infants are not necessarily abnormal. It has been found that periventricular “blush” is very common [1, 2]. The findings of the study by Resch and associates [3] would be significantly strengthened were there no serious limitations to it. Readings should have been made by more than one observer, because in a retrospective study the quality of the images often varies. The assertion that a neuroradiologist’s interpretation was “confirmed” by a pediatric radiologist suggests an unusual 100% agreement. Grades of severity of PVE were based on comparison of the areas of PVE with the density of the choroids plexus. This is subjective, so that the use of more than one reader is critical. In fact, as I look at Fig. 2 of the article I wonder whether there are early low density areas within the PVE in the reproduction of the parasagittal image. The method of selection of preterm infants for the study is not clear. I assume that PVEs were not universal, otherwise the finding would lead me to the conclusion that at least in some cases it represented the well-known “blush”. It is surprising, therefore, that those infants without PVE were not chosen as controls. Inclusion of controls would give the article the needed reliability. The text and Tables 3 and 4 show that persistence of echodensities rather than grade was the main factor of adverse neurodevelopmental outcome. Numbers are small and evidently a Fisher exact test was used instead of the more reliable chi-square test. Nevertheless, the small sample makes the power of the test low (or the likelihood of type II or beta error high). This means that there might indeed be a significant difference missed between grades 1 and 2 because there were so few patients studied. Moreover, I found no consideration of the combined effect of both the grade and the persistence of PVE. The mixed information regarding the significance of PVE in premature infants, as quoted in the authors’ review of the literature on the subject, is the result of small numbers, subjective criteria, and absence of controls. Unfortunately, these problems, many admittedly difficult to control, prevent conclusions being drawn from this report along with others regarding the significance of demonstrating PVE in neonatal cranial sonograms.