Arteriovenous malformations (AVMs) of the brain are relatively rare congenital developmental vascular lesions. They may cause hemorrhagic stroke, epilepsy, chronic headache, or focal neurologic deficits, and the incidence of asymptomatic AVMs is increasing due to widespread availability of noninvasive imaging methods. Since the most severe complication of an AVM is hemorrhagic stroke, most epidemiologic studies have concentrated on the hemorrhage risk and its risk factors. In this article, the authors discuss the epidemiology, presenting symptoms, and hemorrhage risk associated with brain AVMs.
Brain arteriovenous malformations (AVMs) are complex vascular lesions in which arterial blood flows directly into draining veins without an intervening capillary bed. The lack of resistance normally created by small diameter arterioles and capillaries and the resulting direct transmission of arterial pressure to venous structures lead to markedly increased blood flow and, eventually, dilatation and tortuous growth of vessels. In addition to anatomic cerebrovascular changes, this process may bring about significant hemodynamic changes in the brain, such as reversal of venous flow, venous hypertension, and hypoperfusion of regions surrounding the AVM.
AVMs are probably congenital lesions that form during either the embryonic or fetal stages of development, but they are usually not hereditary lesions. Their etiology remains unknown. Although brain AVMs are highly heterogeneous and vary greatly in size and angioarchitecture, multiple AVMs in a single patient are rare. The most serious, and the most common, clinical symptom of an AVM is hemorrhagic stroke resulting from its rupture. AVMs may also cause symptoms due to irritation of surrounding cortex, leading to epilepsy, mass effect, and pathologic changes in hemodynamics as mentioned previously. This article reviews the literature concerning the epidemiology of brain AVMs, as well as presenting symptoms, the risk of hemorrhage, and angioarchitectural factors affecting these characteristics.
Epidemiology
AVMs are rather rare lesions. Although the pathogenesis of AVMs remains unknown, their angioarchitectural characteristics and presentation at any age indicate that they are probably either embryonic or fetal. Despite this, most AVMs seem not to be of hereditary origin, although rare cases of familial occurrence have been reported. Interestingly, AVMs have sometimes been observed to spontaneously disappear, or recur after angiographically complete obliteration.
The reported incidence rates of newly diagnosed AVMs have varied in different population-based studies from 0.89 to 1.34 cases per 100,000 person–years. The tendency for higher incidence rates in more recent studies reflects increased availability of noninvasive brain imaging, and is accompanied with increased proportion of unruptured and even asymptomatic AVMs in the patient population. Prevalence of AVMs has rarely been estimated; a community-based retrospective study from Scotland reported a prevalence of 18 cases in 100,000 people (ie, <0.02%). At the other end of the prevalence estimate continuum, a German study in over 2500 healthy young men (applicants for the military service in German Air Force) undergoing cranial magnetic resonance imaging (MRI) found a prevalence of 0.2% for incidental brain AVMs. In their neurosurgical unit in Helsinki, the authors are responsible for treating the cerebrovascular anomalies of a southern Finnish population of over 2 million people, and they have been treating annually approximately 25 AVM patients and 300 intracranial aneurysm patients. Considering that the prevalence of aneurysms is well established to be around 3%, the authors’ experience conforms very well with the prevalence of the German study, suggesting that prevalence of AVMs is 10- to 15-fold lower than aneurysms (ie, in the range of 0.2%). The total prevalence of AVMs, including all the asymptomatic lesions, is nevertheless extremely difficult to estimate reliably due to the rarity of the disease. The most common age of presentation is the third or fourth decade of life, but AVMs may become symptomatic at any age ( Fig. 1 ). There is no consistently observed gender predominance.
Although AVMs may remain asymptomatic for life in many carriers, they are by no means benign lesions. Overall annual mortality rates have varied from 0.7% to 2.9% in different study populations, but the mortality rates have usually not been compared with the background population. The authors have recently reported that AVM patients have significant excess mortality as compared with matched general population. The authors’ study population of 623 AVM patients from Finland also included a subset of 155 untreated patients. In patients with untreated AVMs, the overall annual mortality rate was 3.4% during a median follow-up period of 18.9 years. AVM-related (due to either acute case fatality or chronic sequelae of AVM-related morbidity) annual mortality rate in the same group was 1.6% (ie, almost 50% of mortality was explained by AVMs). AVM patients were compared with the general population using relative survival ratio (RSR) (ie, the survival of the patient population was compared with the survival of the whole population of Finland, matched for age, sex and historical era). Cumulative RSR in untreated AVM patients 30 years after presentation was 0.49 (ie, 51% excess in mortality compared with general population). In contrast, in AVM patients in whom the AVM was completely obliterated, the cumulative RSR had decreased to only 0.87 after 30 years. Untreated AVM patients thus seem to have a poor long-term prognosis.
Presenting symptoms
Traditionally, most AVMs come into clinical attention because of hemorrhage, with epilepsy coming far behind as the second most common type of presentation. This pattern is gradually changing, however. Increasing availability of noninvasive imaging methods, mainly MRI, have led to more and more frequent detection of unruptured and even incidental AVMs. A few decades ago, over 70% of AVMs presented with hemorrhage, whereas more recent patient series report a hemorrhagic presentation rate of less than 50%. Despite this trend, hemorrhagic stroke still remains the most common type of symptom leading to diagnosis, and 45% to 72% (median 52%) of patients have presented with hemorrhage in 10 large AVM series.
Although AVMs cause only 2% of all and 4% of hemorrhagic strokes, the victims are often younger than most other stroke patients, and AVMs explain one-third of hemorrhagic strokes in young adults. Hemorrhage from AVM is generally considered less hazardous than the rupture of an intracranial aneurysm or spontaneous hypertensive intracranial hemorrhage, but AVM rupture with associated intraparenchymatous hemorrhage often results in significant neurologic disability. Observed case fatality rates and risk of permanent disability caused by AVM hemorrhage vary widely in different reports, but are usually in the range of 5% to 25% and 10% to 40%, respectively.
Various anatomic factors associated with hemorrhagic presentation have been identified. Independent risk factors for hemorrhagic presentation according to multivariate analyses include small size, deep venous drainage, deep, nonborder zone (watershed) and infratentorial locations, associated aneurysms, hypertension, small number of draining veins, venous ectasias, and high feeding artery pressure. Although some characteristics of AVMs, such as small size, are undoubtedly more common in AVMs presenting with rupture, it should be noted that factors associated with hemorrhagic presentation are not necessarily the same as independent risk factors predicting future AVM rupture. Small AVM size is probably the most obvious example; it is found consistently to associate with hemorrhagic presentation, whereas large AVM size seems to increase the risk for hemorrhage during the follow-up after the initial diagnosis. This is most likely due to the fact that small AVMs are not easily diagnosed unless they bleed, whereas large AVMs may cause a variety of symptoms leading to diagnosis before they rupture. In other words, this apparent discrepancy also suggests that significant proportion of small, unruptured AVMs remain unnoticed and undiagnosed for the whole lifetimes of their carriers.
The second most common form of AVM presentation is symptomatic epilepsy, with 18% to 35% (median 26%) of patients diagnosed because of seizures. However, less than 1% of all first episodes of unprovoked seizures are caused by AVMs according to a Swedish prospective, population-based study. Reported anatomic AVM characteristics associated with epileptic presentation have included large size, cortical location of the nidus or the feeders, and location of the AVM in the middle cerebral artery territory. Less common presenting symptoms are chronic headache (unrelated to bleeding) in 6% to 14% of patients, and focal neurologic deficit (temporary, fixed or progressive) due to mass effect or hemodynamic disturbances in 3% to 10% of patients. The proportion of patients with incidentally found AVMs has increased from less than 2% in early studies to 10% in contemporary series.
Presenting symptoms
Traditionally, most AVMs come into clinical attention because of hemorrhage, with epilepsy coming far behind as the second most common type of presentation. This pattern is gradually changing, however. Increasing availability of noninvasive imaging methods, mainly MRI, have led to more and more frequent detection of unruptured and even incidental AVMs. A few decades ago, over 70% of AVMs presented with hemorrhage, whereas more recent patient series report a hemorrhagic presentation rate of less than 50%. Despite this trend, hemorrhagic stroke still remains the most common type of symptom leading to diagnosis, and 45% to 72% (median 52%) of patients have presented with hemorrhage in 10 large AVM series.
Although AVMs cause only 2% of all and 4% of hemorrhagic strokes, the victims are often younger than most other stroke patients, and AVMs explain one-third of hemorrhagic strokes in young adults. Hemorrhage from AVM is generally considered less hazardous than the rupture of an intracranial aneurysm or spontaneous hypertensive intracranial hemorrhage, but AVM rupture with associated intraparenchymatous hemorrhage often results in significant neurologic disability. Observed case fatality rates and risk of permanent disability caused by AVM hemorrhage vary widely in different reports, but are usually in the range of 5% to 25% and 10% to 40%, respectively.
Various anatomic factors associated with hemorrhagic presentation have been identified. Independent risk factors for hemorrhagic presentation according to multivariate analyses include small size, deep venous drainage, deep, nonborder zone (watershed) and infratentorial locations, associated aneurysms, hypertension, small number of draining veins, venous ectasias, and high feeding artery pressure. Although some characteristics of AVMs, such as small size, are undoubtedly more common in AVMs presenting with rupture, it should be noted that factors associated with hemorrhagic presentation are not necessarily the same as independent risk factors predicting future AVM rupture. Small AVM size is probably the most obvious example; it is found consistently to associate with hemorrhagic presentation, whereas large AVM size seems to increase the risk for hemorrhage during the follow-up after the initial diagnosis. This is most likely due to the fact that small AVMs are not easily diagnosed unless they bleed, whereas large AVMs may cause a variety of symptoms leading to diagnosis before they rupture. In other words, this apparent discrepancy also suggests that significant proportion of small, unruptured AVMs remain unnoticed and undiagnosed for the whole lifetimes of their carriers.
The second most common form of AVM presentation is symptomatic epilepsy, with 18% to 35% (median 26%) of patients diagnosed because of seizures. However, less than 1% of all first episodes of unprovoked seizures are caused by AVMs according to a Swedish prospective, population-based study. Reported anatomic AVM characteristics associated with epileptic presentation have included large size, cortical location of the nidus or the feeders, and location of the AVM in the middle cerebral artery territory. Less common presenting symptoms are chronic headache (unrelated to bleeding) in 6% to 14% of patients, and focal neurologic deficit (temporary, fixed or progressive) due to mass effect or hemodynamic disturbances in 3% to 10% of patients. The proportion of patients with incidentally found AVMs has increased from less than 2% in early studies to 10% in contemporary series.
Risk of hemorrhage from untreated AVM s
Since hemorrhagic stroke is the most severe consequence of harboring an AVM, it is understandable that major efforts have been undertaken to identify risk factors predicting AVM rupture. This is all the more important considering that the treatment of complex AVMs is associated with significant risks of morbidity and mortality, and these should not, naturally, exceed the estimated lifetime risk of harboring an untreated lesion. Establishing the probability of rupture and possible anatomic and demographic factors affecting this risk requires long-term follow-up studies in large cohorts of patients with untreated AVMs, desirably with as little selection bias favoring the lack of treatment as possible. Table 1 reviews published long-term follow-up studies using Kaplan-Meier life table analyses and univariate or multivariate models to identify risk factors predicting AVM rupture during follow-up. The average annual AVM rupture rate is approximately 2% to 4% in most of these cohorts, but the rate varies highly depending on various risk factors. When estimating the lifetime risk of hemorrhage from an untreated AVM based on these figures, it is important to remember that the annual risk of rupture should not be multiplied with the (estimated) remaining years at risk; the proper formula for the cumulative probability of hemorrhage is 1-(1- p ) t , where p = the annual probability of hemorrhage and t = time at risk in years, given that the risk remains constant over time. Unfortunately, to make things more complicated, the risk of rupture in untreated AVMs seems not to be stable over the years. Many studies with sufficiently long average follow-up times suggest a change in the hemorrhage rate over time, the rate being highest during the first years after the diagnosis. This observation also explains why studies with short follow-up durations generally report higher average annual rupture rates than those with longer follow-up periods (see Table 1 ). This phenomenon is not likely explained by drop-out bias during long follow-ups, since it was also observed in the authors’ cohort with almost no patients lost to follow-up during the average observation period of 13.5 years (up to 53 years; the follow-up data were 98.7% complete in a cohort of 238 patients). One hypothetical explanation is some kind of hemodynamic destabilization of the AVM around the time it becomes symptomatic.
Related posts:
Dural Arteriovenous Fistulas: Epidemiology and Clinical Presentation
Historical Perspective of Treatments of Cranial Arteriovenous Malformations and Dural Arteriovenous Fistulas
Selection of Treatment Modalities or Observation of Arteriovenous Malformations
Acute Management of Ruptured Arteriovenous Malformations and Dural Arteriovenous Fistulas
Occlusive Hyperemia Versus Normal Perfusion Pressure Breakthrough after Treatment of Cranial Arteriovenous Malformations
Anesthesia Considerations and Intraoperative Monitoring During Surgery for Arteriovenous Malformations and Dural Arteriovenous Fistulas
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
