Grading intraventricular hemorrhage with no grades.

Abstract

To the Editor.—The first classification strategy for intraventricular hemorrhage (IVH grade I–IV) was developed by Papile et al1 for computed tomography (CT) scan studies in 1978. It is still widely used by most cranial ultrasonographers, radiologists, and neonatologists. Volpe,2 based on neuropathologic findings, has suggested that the presence of parenchymal lesions (IVH grade IV) should be noted separately. Still, the grading systems imply, to a different extent, that IVH (presence of blood within the ventricular system) is the expansion of a germinal matrix hemorrhage (GMH), and also, in grade III IVH, the ventricular dilation is posthemorrhagic. However, this is not always the case.The choroid plexus can be a source of bleeding and may be the exclusive site of origin of IVH.3,4 IVH may lead to acute or progressive ventricular dilation as a neuropathologic consequence, but other, more global or diffused brain injury such as ischemic white matter injury and periventricular leukomalacia may accompany IVH.5,6 Based on routine ultrasonography, it is difficult to distinguish between IVH grade III, subacute-chronic hydrocephalus, and ventriculomegaly.Furthermore, grading IVH is often not well standardized. The determination of the ventricular area occupied with blood is not always applicable. In many cases, ventricular dilation is observed only with moderate presence of blood in the lateral ventricles. Parenchymal hemorrhage can be caused by periventricular hemorrhage infarction (PVH), but the etiology and association with IVH is not always certain.2,7Most importantly, many developmental follow-up studies use the variables of low-grade (I and II) and high-grade (III and IV) IVH.7–10 Based on the pathophysiology or the outcome, this grouping is not justified. In IVH grade III (with no persistent ventriculomegaly), the known pathology is the same with grade I and II, and it is the GMH and destruction. From the outcome point of view, studies have shown that, apart from high-grade IVH, low-grade IVH is also associated with a relatively increased risk of impaired outcome.9,10These issues, plus the fact that the timing of scanning is crucial in staging IVH, make the grading systems complex and in some cases may be misleading. In that direction, a suggestion would be to use GMH, IVH, ventricular dilation (VD, the progress would provide information on whether it is posthemorrhagic or ventriculomegaly due to different pathology), and PVH. The hemorrhage could be described as a possible combination (eg, GMH and IVH and VD or GMH and VD). This system would reflect most of the known variations of IVH. The diagnosis could follow the evolution of IVH and also allow better consideration of IVH attendant but different brain pathology. This suggestion might be a more descriptive system with no grades, more informative, and more flexible and precise.In Reply.—We were delighted to read the comments of Dr Vasileiadis, who underscores many of the limitations, for both clinicians and researchers, that are inherent in using a cranial ultrasound grading system as a shorthand means of communicating results. Since 1984, we and our colleagues have advocated describing the findings rather than using a grading system.1–3For the clinician, use of a 4-grade system has a number of limitations, both practically and conceptually. First, this classification does not lend itself to classifying infants who have significant echoabnormalities without intraventricular hemorrhage (IVH). Cerebral white matter echolucency, the most powerful ultrasonographic predictor of adverse outcomes,4,5 can occur with minimal or no IVH and no ventriculomegaly, although both of these latter findings increase the risk of white matter abnormalities.6Similarly, the 4-grade system also does not have a place for infants who have isolated ventriculomegaly. Persistently enlarged ventricles, regardless of the severity or degree of hemorrhage even in the absence of recognizable sonographic white matter echoabnormality, also strongly predict later neurologic abnormalities.7,8Second, this classification implies that grade IV hemorrhage is generally an extension of subependymal hemorrhage. This concept has virtually no pathologic support.9 The suggestion that parenchymal lesions be considered separately10 is commendable, but it is built on acceptance of a 4-grade system. Once again, we suggest that the 4-grade system has long passed its time of usefulness.For the researcher, creating a single nomenclature oversimplifies the complex nature of several disorders that have heterogeneous risk factors and different pathophysiologies, most of which have not been fully elucidated yet. Separating distinct disorders from each other rather than grouping heterogeneous imaging outcomes together enables investigators to make better sense of factors that may contribute to each of the underlying disorders as well as to the interrelationships of the several disorders to each other.