Morbidity of Intracranial Hemorrhage in Patients With Cerebral Arteriovenous Malformation
To the Editor:
The Columbia-Presbyterian Medical Center Arteriovenous Malformation Study Project has made, and continues to make, a significant contribution to our understanding of arteriovenous malformations of the brain. In the recent contribution of Hartmann et al,1 a number of interesting observations were made with regard to hemorrhage. The first is the high incidence among those that bleed of subarachnoid and intraventricular hemorrhage. Only 54% of initial hemorrhages and 49% of follow-up hemorrhage were intraparenchymal. This is at considerable variance with our experience at the Northern and Western Medical School, The University of Sydney, where we have followed all arteriovenous malformations (AVMs) seen since 1991, and of 114 patients presenting with hemorrhage, 82% have a significant intraparenchymal component.
One is left with the feeling from this article that hemorrhage from AVMs is relatively benign. However, it must be borne in mind that this is a specially selected subset of patients referred to a tertiary service. The authors of this article acknowledge that they were not able to ascertain the number of patients that were ineligible for referral because of death. One also wonders whether patients with a poor quality of life as a result of hemorrhage were also excluded from referral. A much more significant study in this regard is the excellent report of 31 patients followed up at the Mayo Clinic until initial hemorrhage, with 29% mortality and 16% permanent neurological deficit.2 Unfortunately, there were only 4 such patients reported by the Columbia-Presbyterian Medical Center Arteriovenous Malformation Study Project.1
A most interesting aspect of the article is the relatively benign second hemorrhage experienced by 27 patients. As with the initial hemorrhage, many of these cases appear to have been nonparenchymal hemorrhage. Although not too much should be made of such small numbers, it does raise the prospect that in those patients for whom the first hemorrhage was benign, a second hemorrhage experienced within 4 years of the first is also relatively benign. This should not be a surprise given that the pathophysiology of an AVM may not have been much changed by an initial benign hemorrhage; therefore, the force of a subsequent hemorrhage may well be similar to that of the first hemorrhage. Although we have seen patients with a subsequent hemorrhage die, we have also a small group of patients in whom repeat hemorrhages were a clinical copy of their initial hemorrhage. However, we have been concerned that the hemodynamics may change with time, and more temporally distant subsequent hemorrhages may be less likely to mimic the initial hemorrhage. This may explain the devastating consequences of harboring an AVM in the study reported by Ondra et al,3 in which 85% of 64 patients sustaining a hemorrhage after study enrollment died or suffered major permanent morbidity. The mean follow-up of this study was 24 years.
Therefore, we suggest that if these findings are confirmed with larger numbers, one should conclude that for those who experience a second hemorrhage shortly after an initial benign hemorrhage, the likelihood is that this second hemorrhage is also benign. No statement can be made regarding the initial hemorrhage or remote subsequent hemorrhage.
- Copyright © 1998 by American Heart Association
We thank Drs Morgan et al for their thoughtful comments. They represent a group that has contributed substantially to the study of brain AVM.
Regarding the issue of possible referral biases to tertiary centers such as ours, we recently presented data from 1266 patients from several AVM databases (Berlin, New York, Paris, Toronto).R1 The results suggest that our sample may be representative in many major clinical and morphological aspects. Also, we offer the opinion that the important contributions of other authors, cited by Morgan et al and discussed in our article, may be influenced by the retrospective nature of the observations and the selection of patients deemed untreatable. Our study is the first to date providing prospective data for patients that survived their diagnostic event. Given the current therapeutic options and the higher risk of further hemorrhage for patients presenting with a bleed,R2 long-term follow-up studies in unselected untreated samples as proposed by Morgan et al will be difficult to conduct because of the strong desire for interventional treatment of a potentially life-threatening disease.
We agree, and so stated in the article, that the paucity of natural history data remains a major concern, especially in that serious hemorrhages may be underrepresented in case series from referral centers. To address these concerns, we have proposed a prospective population-based study (http://cpmcnet.columbia.edu/dept/avm).
Hofmeister C, Hartmann A, Meisel J, Mansmann U, Lasjaunias P, terBruegge K, Pile-Spellman J, Young WL, Mohr JP, Mast H. Epidemiological, clinical, and morphological characteristics of 1266 patients with cerebral arteriovenous malformation. 7th European Stroke Conference, Edinburgh, UK. Cerebrovasc Dis. 1998;8(suppl 4):65. Abstract.