To the Editor:

Since the introduction of cranial ultrasound for the detection of cerebral abnormalities in preterm neonates a vast amount of knowledge has been gathered concerning white matter echolucencies (1). Longitudinal ultrasound studies as well as sequential MRI studies have identified echolucencies as disease entities with a different pathological background. Porencephalic cysts following a parenchymal hemorrhage, usually referred to as a venous infarction, subependymal pseudocysts, cystic periventricular leukomalacia (PVL), and cysts after a focal ischemic infarction, for instance of the middle cerebral artery, can all be identified using sequential ultrasonography (25). Ultrasound machines using 7.5 MHz transducers, available at the time of enrollment of the patients studied by Dammann et al. (6) in 1991–1993, enable identification of these diseases, provided that the children were examined often enough and for a sufficient number of weeks (7). Mixed lesions consisting of venous infarctions and cystic PVL were seen in only 11% of cases in a recent study (8). This is similar to our experience (9). All aforementioned white matter echolucencies are a “hole in the brain,” but have a totally different pathology, and a different long-term outcome (8, 9).

In this respect it is not surprising to see that Dammann et al. hardly found any association between low arterial Paco2 values on day 1 and all white matter echolucencies combined. Others have demonstrated a close relationship between hypocarbia and cystic PVL (1015). Associations of subependymal pseudocysts, venous infarction or neonatal stroke with hypocarbia, however, have never been reported. PVL is far less common at a gestational age ≤28 weeks than between 28 and 34 weeks, whereas major intracranial hemorrhages associated with a unilateral parenchymal hemorrhage are more common ≤28 weeks (personal observation). These observations are supported by the patterns of cerebral palsy that are observed in preterm neonates: hemiplegia seems to be more common in extremely preterm neonates, whereas spastic diplegia is seen more often in moderately preterm neonates (16). It is possible that some of the white matter echolucencies in the study by Dammann et al. were unilateral porencephalic cysts following a venous infarction. In our own clinical experience, substantial hypocarbia may also occur in preterm infants with a gestational age at birth ≤28 weeks when they are a few weeks old, e.g. during ventilation for septicemia or necrotizing enterocolitis, or following rapid pulmonary improvement after administration of corticosteroids. It is unclear whether these cases were identified with a maximum of three ultrasound scans, and whether they were eliminated from the control group. Of course, these white matter echolucencies related to so-called “late-onset PVL” can not be associated with hypocarbia during the 1st day of life. In addition to clinical studies, normoxic or hypoxic hypocarbia has been demonstrated to augment neonatal cerebral injury in different animal species underlining the potentially deleterious effects of hypocarbia (17, 18).

Which conclusions can be drawn from the study by Dammann et al.? Can hypocarbia on day 1 really be excluded as a risk factor for cystic PVL in preterm neonates with a gestational age ≤28 weeks? First, it is likely that the onset of periventricular leukomalacia is far less common than other white matter lesions at a gestational age ≤28 weeks, whatever happens to the brain. Second, it is possible that arterial blood vessels are less responsive to low Paco2 values ≤ 28 weeks than at more advanced gestational ages. Although both arguments support Dammann's hypothesis, these conclusions cannot be drawn from their study as presented. Is it important to identify different white matter lesions? In our view it is. If not interpreted correctly, the study by Dammann et al. could be used to deny the risks of hypocarbia in ventilated preterm neonates (1015).

Although the results of Dammann et al. are correct as presented, we feel that the term “white matter echolucencies” should no longer be used, as identification of different white matter lesions can easily be achieved with sequential ultrasound scans (9). In view of recent publications, hypocarbia remains a matter of concern for ventilated preterm neonates.