A recent article by Roberts et al presented typical echocardiographic and gross pathology on 3 adults with noncompaction. The investigators raised the question whether isolated ventricular noncompaction was being overdiagnosed. I agree that morphology based only on echocardiography may be subjective.

Having studied 3 infants with this diagnosis in 2002 diagnosed by fetal echocardiography and confirmed by autopsies, I conclude that part of the problem is the extraordinary variation in the function and severity of noncompaction. For example, our cases were so severe that 2 died in utero, but the ventricles were not enlarged, and the striking pathophysiology was that of a restrictive cardiomyopathy with enlarged atria and normal-size ventricles with poor contractility. (Unfortunately, the report was published in an obstetrics journal and apparently was never seen in cardiac or anatomy circles.) Pulmonary, mitral, and aortic atresia were not present, conditions that are frequently associated with development of sinusoids connected to the coronary arteries and have no endocardial lining. Our patients had deep trabeculae that were lined with endocardium. Two of our 3 patients had involvement of the right ventricle as well as the left, and endocardial fibroelastosis was present. In contrast, the patients reported by Roberts et al did not require transplantation until 36, 45, and 49 years of age; we pointed out in 2002 that some cases have been reported who survived to the eighth decade.

Considering the overall poor prognosis of noncompaction, a degree of certainty is warranted before this diagnosis is applied.