| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Letters to the Editor |
and
Hans Jörg Meisel, MD
Institute of Medical Biometry and Informatics, University of Heidelberg , In Neuenheimer Feld 305, 69120 Heidelberg, Germany* e-mail: mansmann@imbi.uni-heidelberg.de
Department of Interventional Neuroradiology, CHU Bicêtre, Kremlin-Bicêtre, France
Department of Neurosurgery, BG-Kliniken Bergmannstrost, Halle, Germany
Editor:
Dr Piotin and colleagues (1) describe an endovascular treatment strategy for patients with arterial aneurysms associated with arteriovenous malformations (AVMs) as follows: "If neither the aneurysm nor the AVM has bled, we then consider treatment of the aneurysm first." Their recommendation is based on the observation that the coexistence of AVMs and aneurysms correlates significantly with intracranial hemorrhage at presentation. Arterial aneurysms were present in 30 (11%) of 270 patients.
This recommendation is opposite to the strategy described by Meisel et al (2), who observed shrinkage of arterial aneurysms after endovascular treatment of the AVM and concluded that aneurysms should not be the primary targets compared with AVMs; arterial aneurysms were present in 138 (31%) of the 450 endovascularly treated patients. In accordance with this strategy, no treatment of associated aneurysms does not mean abstinence from treatment. However, on the basis of this experience, the primary treatment of associated aneurysms is not necessary.
It is important to study the possible reasons for developing recommendations for opposite treatments:
1. Case Mix: Possible differences in the structure of the patient population were demonstrated by Hofmeister et al (3). These differences may influence the correlation between the aneurysm rate and the rate of hemorrhage at presentation. Further differences may be found in the anatomic structure of the diseases under treatment in both populations. However, Dr Piotin and colleagues (1) do not give any description of the AVMs diagnosed in their patients.
2. Confounder: Dr Piotin and colleagues (1) interpret the correlation between the aneurysm rate and the rate of hemorrhage at presentation as causal for the future course of disease, while Meisel et al (2) present an analysis of prospective follow-up data. A cross-sectional observation, such as that of Dr Piotin and colleagues, cannot give insight into a causal relationship. Meisel et al show that the size of the AVM is correlated with a higher rate of intracranial hemorrhage and also with a higher susceptibility of associated aneurysms.
Dr Piotin and colleagues (1) do not allow invalidation of points 1 and 2. Steps to overcome this problem could be either an analysis of large series (so far not available) of observational data to study risk factors and to define relevant subgroups in which the efficacy and safety of treatment strategies can be assessed or comparative experimental studies can be performed to settle the argument. The simplest way would be to combine both data sets and perform a case-control study, in which cases from one group are matched (on the basis of risk factors) with the controls of the other group. This is a cooperative task, which would be successful if both groups use the same data model.
Dr Piotin and colleagues (1) report neurologic deficits in five of 30 patients treated with the proposed strategy, which implies a morbidity rate of 16% (95% CI: 5%, 34%). The fact that no deaths occurred among the 30 patients allows the conclusion that the mortality rate is significantly below 12%. Are these procedural rates or are these rates related to a long-term follow-up of the patients? In the case of long-term follow-up, with use of information of 150 patient years of follow-up, the raw yearly hemorrhage rate is 0.03; therefore, it is possible to calculate a 95% CI of 0.01, 0.09. Meisel et al (2) report a 95% CI of 0.00, 0.03 for the raw yearly hemorrhage rate. No conclusion about a possible superiority of one strategy over the other can be drawn from these numbers.
The problem of a comparative study follows from a simple calculation of the sample size: To show a difference in the hemorrhage rate between the two treatment strategies with a significance level of 5% and a power of 80%, it is necessary to recruit regularly 100 patients per group over a period of 10 years with a total follow-up of 15 years, if the yearly hemorrhage rate for treatment A is 0.01 and for treatment B, 0.03. The study of Meisel et al (2) included 83 treated patients with associated aneurysms, that of Piotin et al (1) study, 30 patients. The patients were recruited for more than 10 years in both series. But unfortunately, combining the information of both studies does not meet the information requirement obtained with the sample size calculation. This makes clear which cooperative effort is needed to answer crucial questions about the treatment of patients with cerebral AVMs.
REFERENCES
Department of Interventional Neuroradiology, Hôpital de la Fondation Ophtalmologique Adolphe de Rothschild, 25-29 rue Manin, 75940 Paris Cedex 19, France e-mail: mpiotin@fo-rothschild.fr
We read with interest the comments from Dr Mansmann and colleagues about our article (1), in which the endovascular treatment strategy for patients with arterial aneurysms associated with AVMs was described. The recommendations extracted from our series of 270 consecutive patients highlight the importance of recognizing true arterial aneurysms to focus on the endovascular treatment. In our study, it is true that AVM size was not specified; therefore, the possibility that patients with larger AVMs could harbor aneurysm(s) more frequently than could those with smaller AVMs cannot be ruled out.
On the basis of our experience, our assumption has been that proximal AVM-associated aneurysms should be treated first, especially when the aneurysm has been identified as a source of hemorrhage. It appeared to us that proximal aneurysms associated with AVMs have a greater propensity to rupture when compared with aneurysms in patients without AVMs. In many cases, we advocate treatment of these proximal aneurysms specifically when they are the source of the hemorrhage or when endovascular treatment of the AVM will require repeated crossings (navigation of microcatheters into the aneurysm’s patent artery) during staged embolizations. In many instances, a subarachnoid hemorrhage (more likely due to aneurysm rupture than to the AVM) is more deleterious to the patient than is an intraparenchymal bleeding, which is more likely to be attributed to an AVM.
It may be argued that this series included selection and referral bias, which are always difficult to avoid when dealing with retrospective studies. Although our center is a tertiary referral center to which a large patient population is referred at the acute and subacute stage of intracranial bleeding, this center-specific referral pattern may have been the causative bias that showed the higher prevalence of bleeding in our patients with both AVMs and aneurysms.
REFERENCES
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
