In Reply: Silent Arteriovenous Malformation Hemorrhage and the Recognition of “Unruptured” Arteriovenous Malformation Patients Who Benefit From Surgical Intervention

We appreciate the comments of Drs Defillo and Kennedy and enjoyed their paper on microbleeding from intracranial aneurysms. We hoped that our publication would draw attention to this similarity between sentinel aneurysm hemorrhage and silent arteriovenous malformation (AVM) hemorrhage, with both processes portending a dangerous natural history and indicating aggressive treatment. In fact, we intended in our article to describe intralesional microhemorrhage from AVMs as “sentinel AVM hemorrhage” to emphasize this similarity with aneurysms, but during peer review, it was argued that because these AVM events are clinically silent without symptoms or signs, they should not be called sentinel events. Nonetheless, the idea is clear: hemosiderin is an important finding in a patient with an “unruptured” vascular lesion, be it aneurysm or AVM, and the lesion should be managed more aggressively than a truly unruptured lesion without hemosiderin or evidence of old hemorrhage.

Regarding the groupings, it is logical to think of the 4 patient groups as defined in our paper: unruptured AVM (group 1), silent AVM hemorrhage with hemosiderin on MR imaging (group 2), ruptured AVM without evidence of old hemorrhage (group 3), and reruptured AVM with evidence of old hemorrhage (group 4). It is a matter of semantics whether to label reruptured AVM patients as a group 3 or 4. Our intention was to study these 4 groups independently, but we found that AVM specimens develop hemosiderin positivity quickly (within 5–7 days), making it difficult to differentiate silent previous hemorrhage in group 4 patients (true group 4) from acute hemorrhage in patients without previous microhemorrhage that were “cooled down” before surgery (false group 4). We therefore combined groups 3 and 4 because of this issue with pathological analysis. However, from a clinical perspective, the 4 groups are separate and distinct. Both groups 3 and 4 are typically managed aggressively and, in our practice, most often surgically. The group 4 patients are important ones because these patients sometimes come to clinical attention with seizures or other symptoms before rerupture, as group 2 patients, and intervention at that point would spare them the morbidity of rerupture. Indeed, our study showed that patients treated in group 2 fared better than those in group 4 by nearly 1 modified Rankin Scale point.

Therefore, we agree with Drs Defillo and Kennedy that group 2 patients with hemosiderin associated with their AVMs are unstable. We also agree that group 4 patients are “the result of chronic undetected hemorrhages” or, in other words, of silently hemorrhaging AVMs that eluded diagnosis. This letter and our paper stress that it is incumbent upon us to sharpen our diagnostic acumen to identify warning leaks in AVM patients, which is far more difficult than identifying the textbook warning leaks in aneurysm patients. The diagnosis of silent AVM hemorrhage requires a high index of suspicion when ordering magnetic resonance imaging, as special sequences (eg, iron-sensitive imaging, susceptibility-weighted imaging) must be specified and/or 3-T magnets must be used. In addition, the diagnosis of silent AVM hemorrhage requires a keen eye for hemosiderin when reviewing the studies.