Risk Factors for Subsequent Hemorrhage in Patients with Cerebellar Arteriovenous Malformations

Accepted Manuscript Risk factors for subsequent hemorrhage in patients with cerebellar arteriovenous malformations Xianzeng Tong, M.D., Jun Wu, M.D., Fuxin Lin, M.D., Yong Cao, M.D., Yuanli Zhao, M.D., Shuo Wang, M.D., Jizong Zhao, Chairman PII:

S1878-8750(16)30211-X

DOI:

10.1016/j.wneu.2016.04.082

Reference:

WNEU 4011

To appear in:

World Neurosurgery

Received Date: 20 March 2016 Revised Date:

20 April 2016

Accepted Date: 22 April 2016

Please cite this article as: Tong X, Wu J, Lin F, Cao Y, Zhao Y, Wang S, Zhao J, Risk factors for subsequent hemorrhage in patients with cerebellar arteriovenous malformations, World Neurosurgery (2016), doi: 10.1016/j.wneu.2016.04.082. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT Risk

factors

for subsequent

hemorrhage

in

patients

with

cerebellar

arteriovenous malformations Xianzeng Tong, M.D.,1-4 Jun Wu, M.D., Fuxin Lin, M.D., Yong Cao, M.D., Yuanli Zhao, M.D., Shuo Wang, M.D., Jizong Zhao, Chairman Department of Neurosurgery, Beijing Tiantan Hospital, Capital Medical University, Beijing, P. R.

China;

2

RI PT

1

China National Clinical Research Center for Neurological Diseases, Beijing, P. R.

China; 3Center of Stroke, Beijing Institute for Brain Disorders, Beijing, P. R. China; 4Beijing Key Laboratory of Translational Medicine for Cerebrovascular Diseases, Beijing, P. R. China

SC

Correspondence to: Prof. Shuo Wang, Department of Neurosurgery, Beijing Tiantan Hospital,

Email: [email protected])

M AN U

Capital Medical University, Beijing 100050, China (Tel: 86-10-65113440. Fax: 86-10-65113440.

AC C

EP

TE D

Key Words: cerebellar arteriovenous malformations (AVMs), natural history, hemorrhagic risk

ACCEPTED MANUSCRIPT Abstract OBJECTIVE: The aim of this study was to identify the risk factors for subsequent hemorrhage in patients with untreated cerebellar arteriovenous malformations (AVMs). METHODS: We searched our AVM database at Beijing Tiantan Hospital and identified 149 patients

RI PT

with cerebellar AVMs that were at least 1 month treatment-free after initial diagnosis between the year 2000 and 2015. The patients were followed up from initial diagnosis until subsequent hemorrhage, initiation of treatment or the end of 2015. The natural history of cerebellar AVMs was analyzed.

RESULTS: The overall annual rupture rate was 8.6% with a mean follow-up period of 4.2 years (range,

SC

1 month -15 years). The annual rupture rate for previously ruptured AVMs was 10.8% during the whole follow-up period, 12.4% in the first five years (18.8% in the first year and 9.0% in the subsequent 4

M AN U

years), and 6.7% in more than 5 years after initial diagnosis. The overall annual rupture rate for previously unruptured AVMs was 4.0%. Childhood at diagnosis, AVM size ≥3cm and exclusively deep venous drainage were independent risk factors for subsequent hemorrhage. Previous AVM rupture significantly increased the hemorrhagic risk during the first five years but did not significantly affect subsequent hemorrhage thereafter.

TE D

CONCLUSION: Childhood at diagnosis, large AVM size and AVMs with exclusively deep venous drainage are independent risk factors for subsequent hemorrhage in patients with cerebellar AVMs. Previous rupture may increase the hemorrhagic risk during the first five years after diagnosis but may

EP

not significantly increase the risk in the following years.

AC C

Key Words: cerebellar arteriovenous malformations (AVMs), natural history, hemorrhagic risk

ACCEPTED MANUSCRIPT Introduction Cerebellar arteriovenous malformations (AVMs) are rare lesions that comprise less than 15% of all brain AVMs.1 In our previous study of 3299 AVM patients, cerebellar AVMs accounted for only 9% (295/3299).2 Due to the rarity of cerebellar AVMs, previous studies often combine them with

RI PT

brainstem AVMs in reports on posterior fossa or infratentorial AVMs.3-11 However, cerebellar AVMs are different from brainstem AVMs in their natural history, clinical presentation and treatment outcomes. When describing the natural history of brain AVMs, previous studies found that infratentorial location is a risk factor of subsequent hemorrhage. However, the studies seldom point out

SC

whether brainstem AVMs or cerebellar AVMs contribute more to the hemorrhage risk. In the literature, the annual rupture rate for brain AVMs ranged from 0.78% to 34.3%.12-14 In a recent review of the

M AN U

natural history of brain AVMs, the overall annual rupture rate for patients with untreated brain AVMs ranged from 2.10% to 4.12%.15 Initial hemorrhagic presentation, exclusively deep venous drainage, and deep and infratentorial brain location are risk factors for subsequent hemorrhage.15 In 2009, Arnaout et al. reviewed the literature on posterior fossa AVMs and found their annual rupture rates to be as high as 11.6%.1 However, seldom studies have pointed out the annual rupture rate of cerebellar AVMs in their

TE D

natural history. To investigate the natural history of cerebellar AVMs, we retrospectively reviewed our AVM database at Beijing Tiantan Hospital to identify patients with cerebellar AVMs that were at least

AC C

EP

1 month treatment-free after initial diagnosis between 2000 and 2015.

ACCEPTED MANUSCRIPT Methods Patient population This study was approved by the Institutional Review Board of Beijing Tiantan Hospital Affiliated to Capital Medical University. Our prospectively maintained brain vascular malformation database was

RI PT

searched to identify patients with cerebellar arteriovenous malformations (AVMs) that were admitted at Beijing Tiantan Hospital between January 2000 and December 2015. All AVMs were confirmed by digital subtraction angiography (DSA). Ultimately, 225 patients were identified to harbor a cerebellar AVM. For patients with life-threatening posterior fossa hematoma due to AVM rupture, we performed

SC

emergency hematoma evacuation with AVM resection at the same time or with AVM resection at a deferred time (usually one or two months after hematoma evacuation). Early treatments were also

M AN U

initiated for patients presenting with associated aneurysms that caused posterior fossa hemorrhage or patients with life-threatening obstructive hydrocephalus. Although we recommended the patients without life-threatening hemorrhage to receive AVM resection one or two months after hemorrhagic presentation, the ultimate treatment modality and treatment timing were decided by patient preference. For those without life-threatening presentation, most patients were able to go back to work or school

TE D

with no deficit or minor deficit after a period of rehabilitation. Some patients sought for AVM treatment without a second hemorrhage over one or two months after hemorrhage. However, for some patients, considering the risk of treatment-related complications that may affect their quality of life, they

EP

deferred the treatment until a second hemorrhage occurred or until their neurological status declined. In this study, we included 149 patients with cerebellar AVMs that did not receive any treatment within 1

AC C

month after initial diagnosis. The 149 untreated patients were followed-up until subsequent AVM hemorrhage, initiation of treatment or the end of the year 2015. The baseline patient characteristics (sex, age at initial diagnosis and initial presentation) and the AVM features (size, location, angioarchitecture, associated aneurysms and Spetzler-Martin grade) were collected. According to previous studies, 16,17 we defined the cutoff age between childhood and adulthood as 20 years of age. Age ＜20 years was defined as childhood and age ≥20 was adulthood. An AVM hemorrhage was considered if there was any clinically symptomatic event (any focal neurological deficit, sudden-onset headache or seizure) that is associated with imaging findings of bleeding. The hemorrhage was detected by CT and/or MR brain imaging or the cerebrospinal fluid laboratory test. The primary end point was the first subsequent

ACCEPTED MANUSCRIPT hemorrhage after initial diagnosis. All available follow-up data were collected starting from initial diagnosis of a cerebellar AVM until subsequent AVM hemorrhage, initiation of treatment or the end of the year 2015. Statistical analysis

RI PT

Statistical analyses were performed by two authors (XT, JW) with the Statistical Package for the Social Sciences software (version 20.0; SPSS Inc., Chicago, Illinois). Patient demographics and AVM characteristics were summarized using descriptive statistics for continuous variables (mean ± standard deviation) and categorical variables (count and percentage). For patients that experienced subsequent

SC

hemorrhage and those that remained hemorrhage-free during the follow-up period, patient characteristics and AVM features were compared between the two groups using Pearson’s χ2 test for

M AN U

categorical variables and the Mann-Whitney U test for age and follow-up time. The person-years of follow-up were calculated from the date of initial AVM diagnosis to the first subsequent hemorrhage, the initiation of treatment or the end of the year 2015. The annual rupture rate was calculated as the number of patients experiencing subsequent hemorrhage after initial diagnosis divided by person-years of follow-up. Cumulative rupture rates were assessed using the Kaplan-Meier product-limit method,

TE D

and the significance of group differences were estimated using the log-rank test. Univariate Cox regression analysis was used to calculate the hazard ratios and 95% CIs for several variables. These variables were age group (childhood or adulthood); sex; AVM hemorrhage before admission; AVM size

EP

(categorized as small ＜3 cm in diameter and large ≥3 cm); cortical or deep location; pattern of venous drainage (categorized as exclusively deep and not exclusively deep); patterns of feeding artery (single

AC C

or not single) and associated aneurysm. Multivariate Cox proportional hazards model with a forward stepwise regression procedure was used to determine the risk factors for subsequent AVM rupture. Similar to the study of Hernesniemi et al.,18 we also found that the annual rupture rates were highest during the first five years after initial AVM diagnosis. We used similar statistical methods described by Hernesniemi et al.18 In this study, log-rank tests and Cox regression analyses were performed not only for the entire follow-up period but also for the first 5 years after initial diagnosis. A 2-tailed P value of less than 0.05 was considered statistically significant in each analysis. Results Baseline characteristics of the 149 patients with untreated cerebellar AVMs

ACCEPTED MANUSCRIPT Of the 225 patients, 76 patients received treatment within 1 month after intial AVM diagnosis, including microsurgical resection, hematoma evacuation, radiosurgical therapy, endovascular treatment or external ventricular drainage. The remaining 149 patients did not receive any of the above treatments within 1 month after initial diagnosis. Table 1 shows the characteristics of the 149 patients according to

RI PT

the occurrence of subsequent hemorrhage during follow-up. The patients included 86 males and 63 females (age at diagnosis range: 5-73 years, mean: 28.1 years). Children accounted for 37% (55/149) of the patients. Initial hemorrhage at presentation occurred in 81.2% (121/149) of the cases. The AVMs were Spetzler-Martin grade I in 37 patients, grade II in 47 patients, grade III in 35 patients, grade IV in

SC

22 patients and grade V in 8 patients. Due to the perceived risk of rupture, most of the AVM patients with associated aneurysms were treated within 1 month after admission. Only 13 patients with

M AN U

associated aneurysms were included in this follow-up series. Annual rupture rates during follow-up period

Table 2 shows the annual and cumulative rupture rates in relation to patient sex, age group, previous rupture, AVM size, cortical or deep location, pattern of venous drainage or feeding arteries. Annual rupture rates varied substantially, depending on stratifying factors and time point, from 1.9%

TE D

during the first 5 years after admission in previously unruptured AVMs to 19.6% during the first 5 years in AVMs with associated aneurysm (Table 2). Similarly, cumulative rupture rates were very variable, ranging from 23% in 10 years in previously unruptured AVMs to 77% in 10 years in AVMs

EP

with associated aneurysm (Table 2).

During the total follow-up period of 627.1 person-years, 54 patients experienced an AVM

AC C

hemorrhage, yielding an overall annual rupture rate of 8.6%. The annual rupture rate was 9.6% for the first five years after admission and 6.6% for the period over five years after admission. For all patients, the cumulative rupture rate was estimated as 26% (95%CI: 21%-31%) at five years after admission and 49% (95%CI: 43%-55%) at ten years after admission. For the 121 patients with previous hemorrhage at admission, the overall annual rupture rate was 10.8% during a follow-up period of 426.3 person-years. The annual rupture rate was 12.4% for the first five years. The annual rupture rate was the highest for the first year (18.8%), followed by 9.0% for the 2-5 years and 6.7% for the period over 5 years after diagnosis. The cumulative rupture rate was estimated as 41% (95%CI: 35%-47%) at five years after admission and 54% (95%CI: 47%-61%) at ten

ACCEPTED MANUSCRIPT years after admission. For the 28 patients without previous hemorrhage at admission, 8 patients experienced subsequent hemorrhage during a follow-up of 200.8 person-years, yielding an annual rupture rate of 4.0%. The cumulative rupture rate was estimated as 9% (95%CI: 3%-15%) at five years after admission and 23%

Univariate Analyses of risk factors for subsequent hemorrhage

RI PT

(95%CI: 12%-34%) at ten years after admission.

We used three models of univariate analyses to test the risk factors for subsequent hemorrhage (Table 1, Table 2, Table 3 and Table 4).

SC

Model 1. Table 1 shows the difference in characteristics between patients who experienced a subsequent AVM rupture after admission and those who remained rupture-free during the whole

M AN U

follow-up period (Pearson’s χ2 test for categorical variables and the Mann-Whitney U test for age and follow-up time). A subsequent hemorrhage was more likely to occur in patients who were younger at the time of admission and those with AVMs that were deep-seated, larger-sized or had an exclusively deep venous drainage (all p＜0.05, Table 1). In this univariate analyses, patients who experienced a subsequent hemorrhage during the entire follow-up period seemed to be more likely to have a

TE D

previously ruptured AVM. However the difference did not reach statistical significance (p=0.238). For the first five years after admission, 38 of the 121 patients with previously ruptured AVMs only 2 of the 28 patients with previously unruptured AVMs experienced a subsequent hemorrhage (p=0.009).

EP

Previous AVM rupture may be significantly associated with the occurrence of subsequent hemorrhage in the first five years after admission and may be insignificant in the following time over five years.

AC C

AVMs with more than one feeding arteries may increase the occurrence of subsequent hemorrhage compared with AVMs that had single feeding artery, however, the difference also did not reach statistical significance (p=0.080). Model 2. As demonstrated in Table 2, Figure 1 (B-D) and Figure 2 (B-D), previous rupture, young

age at admission (childhood), AVMs with exclusively deep venous drainage significantly increased the annual rupture rates and cumulative rupture rates both in the first five years and during the entire follow-up period. AVMs with larger size (≥3cm) and deep locations may increase the annual rupture rate and cumulative rupture rate compared with small-sized and cortical located AVMs respectively (Table 2; Figure 1, E and F; Figure 2, E and F). However, the differences did not reach statistical

ACCEPTED MANUSCRIPT significance in log-rank tests either in the first five years or during the entire follow-up period (all p>0.05, Table 2, Figures 1 and 2). Sex and patterns of arterial supply did not significantly affect the rupture rate (all p>0.05, Table 2). Model 3. Univariate Cox regression analysis (Table 3 and Table 4) shows that previous rupture,

RI PT

AVM diagnosis at childhood and exclusively deep venous drainage were significantly associated with the risk for subsequent AVM hemorrhage. Previously rupture increased the relative risk almost six-fold for the first five years and approximately 2.5-fold for the entire follow-up period. AVM diagnosis at childhood and exclusively deep venous drainage increased the relative risk approximately 2- to 3-fold

SC

both for the first five years and for the entire follow-up period. Although larger AVM size (≥3cm) and deep AVM location may seem to increase the relative risk of subsequent hemorrhage compared with

M AN U

small sized AVMs and deep located AVMs respectively, the difference did not reach statistical significance (all p>0.05). Sex and patterns of arterial supply did not significantly affect the rupture rate in this univariate analyses model (all p>0.05, Table 3 and 4). Multivariate analyses of risk factors for subsequent rupture

Multivariate Cox regression analyses with a forward stepwise regression procedure were used to

TE D

determine the independent risk factors for subsequent AVM rupture during the first five years after admission and during the entire follow-up period (Tables 3 and 4). All variables in Table 3 and Table 4 were tested. AVM diagnosis at childhood, larger AVM size and exclusively deep venous drainage were

EP

independent risk factors for subsequent AVM rupture both in the first five years and during the entire follow-up period. Previous rupture was an independent factor for subsequent AVM hemorrhage during

AC C

the first five years, increasing the relative risk almost five-fold (Table 3). However, according to this regression model, previous rupture did not significantly affect AVM rupture during the entire follow-up period (Table 4).

Discussion

Cerebellar AVMs are a small group of neurovascular lesions that comprise only 5-10% of all brain AVMs.19 Due to the rarity of cerebellar AVMs, most publications have combined them with brainstem AVMs in reports on posterior fossa or infratentorial AVMs, which account for 7–15% of brain AVMs.3,5,13,20-22 Cerebellar AVMs are different from their supratentorial counterparts in their natural history, clinical presentation and treatment outcomes.4,6,11,23,24 Cerebellar AVMs pose an

ACCEPTED MANUSCRIPT increased risk of hemorrhagic presentation, which may lead to high rate of morbidity and mortality.1,8,25-27 In our previous report on 3299 patients with brain AVMs in World Neurosurgery, cerebellar AVMs accounted for 9% and infratentorial AVMs accounted for 11.8%.2 Hemorrhagic presentation was reported in 78% of the cerebellar AVMs.2 In this series of 149 patients with cerebellar

RI PT

AVMs that did not receive any treatment within 1 month after admission, 81.2% presented with hemorrhage.

Despite advances in microsurgical techniques, radiosurgery and endovascular embolization during the last decades, treatment of cerebellar AVMs remains challenging, due to the concentration of

SC

adjacent eloquent structures in the narrow confines of the posterior fossa. However, the natural history of cerebellar AVMs is still unclear. In this study, cerebellar AVM patients with at least 1 month of

M AN U

follow-up before any treatment were included. Most of the patients received treatment or experienced subsequent hemorrhage in the first five years after initial diagnosis. Only 39 patients were followed over five years. The perceived risk of hemorrhagic presentation has led the patients and doctors to initiate treatment at an early stage. Annual rupture rate of cerebellar AVMs

TE D

In the literature, numerous studies have demonstrated that an infratentorial location is an independent risk factor for subsequent AVM rupture. Recently in 2008, Kelly et al. reviewed their surgical series of 76 patients with posterior fossa AVMs and reported an annual rupture rate of

EP

8.4-9.4%.6 Hernesniemi et al. retrospectively reviewed 238 patients with untreated brain AVMs between 1942 and 2005 (mean follow-up 13.5 years).18 Infratentorial AVM location was found to be an

AC C

independent risk factor for subsequent hemorrhages. Infratentorial AVMs had the highest annual rupture rate: 11.6% for the first five years after presentation, 3.6% for the period over five years and 6.7% for the entire follow-up period.18 For patients with infratentorial AVMs, the cumulative rupture rates were 45% and 76% respectively for five years and 20 years after initial presentation.18 In our series, the overall annual rupture rate was 8.6%-9.6% for the first five years and 6.6% for the time over five years after initial presentation. The cumulative rupture rates were 26% and 49% respectively for five years and ten years after initial presentation. Our findings were consistent with previous studies. Infratentorial AVM location may increase the proclivity of hemorrhage by some anatomical and hemodynamic factors that are unique to this location.

ACCEPTED MANUSCRIPT Risk factors for subsequent hemorrhage Previous rupture as a risk factor for subsequent hemorrhage There is a growing body of evidence demonstrating a correlation between previous rupture and an increased risk of subsequent hemorrhage. In the study of Mast et al., the annual rupture rate was

RI PT

reported as high as 17.8% in patients with previously ruptured AVMs.28 The authors concluded that hemorrhagic presentation was the most important feature influencing subsequent hemorrhage rates. In their prospective study of 622 patients with brain AVMs (including 74 infratentorial AVMs), Stapf et al., reported annual rupture rates ranging from 4.5 to 34.3% for previously ruptured AVMs, depending

SC

on lesion location and deep venous drainage.14 In their prospective analysis of 678 patients, da Costa et al. reported an annual rupture rate of 7.48% for previously ruptured AVMs and found that hemorrhagic

M AN U

presentation was a significant independent predictor of future hemorrhage.29 According to our univariate and multivariate analyses, previous rupture is an independent predictor for subsequent hemorrhage only during the first five years after initial diagnosis. Previous rupture did not significantly affect subsequent hemorrhage in view of the entire follow-up period. In this point of view, we can conclude that previous AVM rupture may significantly influence future rupture only in the first few

TE D

years after initial diagnosis and may not significantly affect the occurrence of subsequent hemorrhage thereafter. In other words, the effect of previous rupture on subsequent hemorrhage may diminish over time. Previous studies showed that the increased risk of hemorrhage caused by previous rupture was

EP

highest (6-33%) in the first year after the initial hemorrhage, and thereafter gradually decreased.17,18,28-33 In our series, for the patients with previous AVM rupture, the annual rupture rate

AC C

was as high as 18.8% for the first year, 9.0% in the subsequent 4 years and 6.7% in more than 5 years. Our findings are consistent with previous studies. Yamada et al. reported 15.42% in the first year, 5.32% in the subsequent 4 years, and 1.72% in more than 5 years. Compared with that, in the nonhemorrhagic group in their series, the annual rupture rate was 3.12%.17 Our findings are in accordant with the assumption of Yamada et al. who stated that if patients with initially hemorrhagic AVMs did not have a recurrent hemorrhage in the first five years, they subsequently had a fairly low risk of subsequent hemorrhage.17 This phenomenon also supports our conclusion that previous AVM rupture may significantly affect subsequent hemorrhage only in the first few years after initial diagnosis. However, even after the first five years, cerebellar AVMs still pose a high risk of

ACCEPTED MANUSCRIPT hemorrhage compared with their supratentorial counterparts in previous reports, implicating aggressive treatment for these lesions with aggressive natural history. Childhood or young age at diagnosis as a risk factor for subsequent hemorrhage In this study, we found that childhood at diagnosis had a significant (two to three fold) higher risk

RI PT

of subsequent hemorrhage both for the first five years after diagnosis and during the entire follow-up period. An AVM patient with a younger age at diagnosis was also associated with a higher rate of subsequent hemorrhage. In the study of Yamada et al., children with previously ruptured AVMs had a significant (threefold) higher risk of subsequent hemorrhage and the annual rupture rate was as high as

SC

20.28%.17 Conflicting data exist regarding the association between age and subsequent AVM hemorrhage after initial diagnosis. In their studies, Stapf et al. found that increasing age correlated

M AN U

positively with initial hemorrhagic presentation and subsequent hemorrhage.14,34,35 Other studies also reported that advancing age increased the risk of subsequent hemorrhage.30,36 In their large cohort study, Fullerton and colleagues reported that although children with brain AVMs were more likely to present with hemorrhage, they were not at increased risk for a subsequent hemorrhage and may even be relatively protected.16 A recent study on the natural history of untreated brain AVMs showed that

TE D

hemorrhagic presentation and increasing age are independent predictors of subsequent hemorrhage.37 Since previous reports are mostly studies on all brain AVMs, not exclusively on cerebellar AVMs, our results need to be warranted. Our findings suggest that aggressive treatment should be initiated sooner

EP

rather than later for children with hemorrhagic cerebellar AVM presentation. Even some previous studies suggest that brain AVMs in children do not necessarily need to be treated more aggressively

AC C

than those in adults due to similar annual rupture risk, they still state that the cumulative risk in children with brain AVMs is greater given their greater number of years left to live.16 Exclusively deep venous drainage as a risk factor for subsequent hemorrhage Several multivariate analyses have found that deep vein drainage plays a role in subsequent

hemorrhage.14,28,37-39 In our study, exclusively deep venous drainage was a significant risk factor of subsequent hemorrhages both for the first five years and for the entire follow-up period. However, the unexpectedly high annual rupture rate of 15.6% in AVMs with exclusively deep venous drainage is somewhat hard to explain. In the literature, Stapf et al. reported that the combination of a deep or infratentorial AVM location with deep vein drainage predicted an annual rupture rate of 8.0%,

ACCEPTED MANUSCRIPT hemorrhagic presentation alone predicted an annual rupture rate of 14.8% and the combine of the above (infratentorial location with deep venous drainage and initial hemorrhage) predicted an annual rupture rate of 34.3%.14 From this point of view, we can elucidate the high annual rupture rate of cerebellar AVMs in the subgroup of exclusively deep venous drainage. In our series, 18 hemorrhages

RI PT

occurred during a follow-up period of 103.8 person-years in the subgroup of previously ruptured cerebellar AVMs with exclusively deep venous drainage, yielding an annual rupture rate of 17.3%, which is significantly lower than the reported rate of 34.3%.14 Cerebellar AVMs with exclusively deep venous drainage may contribute to the high rupture rate by increasing the pressure gradient in the AVM

SC

nidus.1 Unique local hemodynamic factors produced by the convergence of the draining veins into the vein of Galen and straight sinus may contribute to the increased rupture rates.40

M AN U

As Table 2 showed in our series, the hemorrhagic risk for AVMs with previous rupture decreased dramatically over 5 years after initial diagnosis. However, for cerebellar AVMs with exclusively deep venous drainage, the risk of hemorrhage decreased slightly and remained at a high rupture rate over time. Just as previous studies concluded, whatever the mechanism is behind the diminishing rupture risk over time for previously ruptured AVMs, it seems to not affect the rupture risk conferred by deep

TE D

venous drainage.15,29 However, current data are limited for deep venous drainage as a risk factor for hemorrhagic risk, not as much as hemorrhage at initial presentation.15 More data are needed to elucidate the true annual rupture rate for AVMs with deep vein drainage alone and deep vein drainage

EP

in combination with other putative risk factors.15

AVM size as a risk factor for subsequent hemorrhage

AC C

In our study, AVM size larger than 3 cm increased the relative risk almost two fold of that in small AVMs. Several studies have demonstrated that larger AVM size is associated with subsequent hemorrhage. In 1996, Brown et al. reported no statistically significant association between AVM size and subsequent hemorrhage, but the rupture rates were higher for medium- and large-sized AVMs.41 In their series of 24 brain AVMs, Hirai et al. reported annual rupture rates of 5.4% for large AVMs and 2.1% for small AVMs.42 In 2000, Mine et al. reported the natural history of 55 brain AVMs with the annual rupture rate of 6.4% for large AVMs and 2.3% for small and medium AVMs for the first five years after initial diagnosis.43 After 20 years, the annual rupture rates decreased to 3.9% for large AVMs and 1.6% for small and medium AVMs.43 In a study of 390 patients with brain AVMs, Stefani

ACCEPTED MANUSCRIPT et al. found that AVM size larger than 3cm increased the hemorrhagic risk 2.5-fold compared with that for small AVMs.44 In 2008, Hernesniemi et al. found that large AVM size is an independent risk factor for subsequent hemorrhage.18 The annual rupture rates for large AVMs were 5.5% and 2.7% respectively in the first five years and in more than 5 years after initial diagnosis, while for the

RI PT

small-sized AVMs that rates were 4.7% and 1.6% respectively.18 In our series, larger AVM size predicted a significantly higher annual rupture rates than small AVM size (overall, 10.5% vs 6.8%; 1-5 years, 11.4% vs 8.0%; more than 5 years, 8.7% vs 4.6%). Based on our findings and previous reports,

after initial diagnosis. Other factors that may affect subsequent hemorrhage

SC

we can conclude that large cerebellar AVMs may be more prone to hemorrhage in the subsequent years

M AN U

In our study, we did not find a statistically significant association between other factors and future rupture. In the literature, sex, deep AVM location and associated aneurysm may be hemorrhagic risk factors. Some studies reported that female sex may be an independent risk for rebleeding of cerebral AVMs,17,37 while other found that male patients had a higher risk of subsequent hemorrhage than female patients.28 Still others, like our study, reported no association between sex and future AVM

TE D

rupture.14 Previous studies tend to combine deep location with infratentorial location as one factor (deep and infratentorial location) when analyzing the risk factors for subsequent hemorrhage. Stapf et al. reported that deep and infratentorial location was one of the three major risk factors for subsequent

EP

hemorrhage.14 Stefani et al. found that deep and infratentorial location alone significantly increased the risk (almost six-fold) for subsequent hemorrhage.44 Mine et al. found a significantly higher rupture rate

AC C

in deep and infratentorial AVMs (7.5% vs 2.3% overall) from univariate analysis.43 In the series of Crawford et al., deep and infratentorial location was not found to be a predictor of future hemorrhage, while a temporal AVM location predicted a higher rupture rate.36 In our study, although cerebellar location belongs to previously called “deep and infratentorial location”, we still put them into two categories: cortical location and deep location. Deep location was borderline significant in univariate analyses (Chi-square test, Log-rank test and Cox proportional hazards model), but multivariate analyses did not reveal a statistically significant association between deep location and future hemorrhage. Previous studies reported that AVM-associated aneurysms were significantly associated with a risk of an initial hemorrhagic presentation.7,24,34,39,45-48 However, Stapf et al. reported that

ACCEPTED MANUSCRIPT AVM-associated aneurysms were not significantly associated with subsequent hemorrhage, although they were significantly associated with an initial hemorrhagic presentation.14 In their series of 305 AVM patients, Yamada et al. also did not find an increased risk of subsequent hemorrhage in AVMs with an associated aneurysm.17 In the recent multi-center study of untreated brain AVMs, Kim et al.

RI PT

found that the presence of AVM associated aneurysm increased hemorrhage risk but this effect seemed to diminish over time.37 This can also be seen from our series. In our series, associated aneurysm was borderline significant according to log-rank test and univariate Cox regression model. Only 13 patients with associated aneurysms were included in this study. Most of the patients with AVM associated

SC

aneurysms (prenidal or intranidal) were treated within 1 month after initial diagnosis, due to the perceived risk of hemorrhage caused by these aneurysms. The selection bias may have affected the

M AN U

results. Limitations

In this study, we retrospectively reviewed 149 patients with cerebellar AVMs that were treatment-free at least 1 month after initial diagnosis. Since our study was based on patients admitted to our tertiary neurosurgical center and the study was not population based, selection bias may exist in

TE D

this one single center, retrospective study. Meanwhile, many patients dropped out and sought for treatment in the first five years after initial diagnosis due to the perceived risk of hemorrhage for infratentorial AVMs. Selection bias, retrospective analysis and short follow-up period may weaken the

EP

strengths of our findings in the natural history of cerebellar AVMs. A prospective, multi-center study may help to improve the precision of risk estimates. However, even in a prospective, multi-center, large

AC C

cohort study, the limitation of short follow-up period may not be overcome due to censoring for treatment for these cerebellar AVMs with aggressive natural history. Conclusions

Our data suggest that young age at diagnosis, large AVM size and AVMs with exclusively deep

venous drainage are independent risk factors for subsequent hemorrhage in patients with cerebellar AVMs. Previous AVM rupture may increase the hemorrhagic risk during the first five years after diagnosis but may not significantly increase the risk in the following years. The annual rupture rate is highest in the first year after initial hemorrhage and gradually decreases thereafter. The high rupture rate in the natural history of cerebellar AVMs may implicate aggressive treatment sooner rather than

ACCEPTED MANUSCRIPT later for these lesions with special location. Disclosure and Acknowledgement This study was supported by Key Projects in the National Science & Technology Pillar Program during the Twelfth Five-year Plan Period (Grant No. 2011BAI08B08). The authors report no conflict of

RI PT

interest concerning the materials or methods used in this study or the findings specified in this paper. Author contributions to the study and manuscript preparation include the following. Conception and design: all authors. Acquisition of data: Tong, Wu, Lin. Analysis and interpretation of data: Tong, Wu, Lin. Drafting the article: Tong. Critically revising the article: all authors. Reviewed submitted version

SC

of manuscript: all authors. Approving the final version of the manuscript on behalf of all authors: Shuo Wang. Statistical analysis: Tong, Wu. Administrative/technical/material support: Shuo Wang. Study

M AN U

supervision: Yong Cao, Yuanli Zhao, Shuo Wang and Jizong Zhao. References

1. Arnaout OM, Gross BA, Eddleman CS, Bendok BR, Getch CC, Batjer HH. Posterior fossa arteriovenous malformations. Neurosurg Focus. 2009; 26(5):E12.

2. Tong X, Wu J, Lin F, Cao Y, Zhao Y, Ning B, Zhao B, Wang L, Zhang S, Wang S, Zhao J. The effect

TE D

of age, sex, and lesion location on initial presentation in patients with brain arteriovenous malformations. World Neurosurg. 2016; 87:598-606. 3. Batjer H, Samson D. Arteriovenous malformations of the posterior fossa: clinical presentation,

EP

diagnostic evaluation and surgical treatment. Neurosurg Rev. 1986; 9:287-296. 4. Da Costa L, Thines L, Dehdashti AR, Wallace MC, Willinsky RA, Tymianski M, Schwartz ML, ter

AC C

Brugge KG. Management and clinical outcome of posterior fossa arteriovenous malformations: report on a single-centre 15-year experience. J Neurol Neurosurg Psychiatry. 2009; 8:376-379. 5. Drake CG, Friedman AH, Peerless SJ. Posterior fossa arteriovenous malformations. J Neurosurg. 1986; 64:1-10.

6. Kelly ME, Guzman R, Sinclair J, Bell-Stephens TE, Bower R, Hamilton S, Marks MP, Do HM, Chang SD, Adler JR, Levy RP, Steinberg GK. Multimodality treatment of posterior fossa arteriovenous malformations. J Neurosurg. 2008; 108:1152-1161. 7. Khaw AV, Mohr JP, Sciacca RR, Schumacher HC, Hartmann A, Pile-Spellman J, et al. Association of infratentorial brain arteriovenous malformations with hemorrhage at initial presentation. Stroke.

ACCEPTED MANUSCRIPT 2004; 35:660-663. 8. Neacsu A, Ciurea AV. General considerations on posterior fossa arteriovenous malformations (clinics, imaging and therapy). Actual concepts and literature review. J Med Life. 2010; 3:26-35. 9. O'Shaughnessy BA, Getch CC, Bendok BR, Batjer HH. Microsurgical resection of infratentorial

RI PT

arteriovenous malformations. Neurosurg Focus. 2005; 19(2):E5. 10. Sinclair J, Kelly ME, Steinberg GK. Surgical management of posterior fossa arteriovenous malformations. Neurosurgery. 2006; 58(4 Suppl 2):189–201.

11. Symon L, Tacconi L, Mendoza N, Nakaji P. Arteriovenous malformations of the posterior fossa: a

SC

report on 28 cases and review of the literature. Br J Neurosurg. 1995; 9:721–732.

12. ApSimon HT, Reef H, Phadke RV, Popovic EA: A population based study of brain arteriovenous

M AN U

malformation: long-term treatment outcomes. Stroke. 2002; 33:2794-2800.

13. Ondra SL, Troupp H, George ED, Schwab K: The natural history of symptomatic arteriovenous malformations of the brain: a 24-year follow-up assessment. J Neurosurg. 1990; 73:387-391. 14. Stapf C, Mast H, Sciacca RR, Choi JH, Khaw AV, Connolly ES, et al: Predictors of hemorrhage in patients with untreated brain arteriovenous malformation. Neurology. 2006; 66:1350-1355.

TE D

15. Abecassis IJ, Xu DS, Batjer HH, Bendok BR. Natural history of brain arteriovenous malformations: a systematic review. Neurosurg Focus. 2014;37: E7. 16. Fullerton HJ, Achrol AS, Johnston SC, McCulloch CE, Higashida RT, Lawton MT, et al:

EP

Long-term hemorrhage risk in children versus adults with brain arteriovenous malformations. Stroke. 2005; 36: 2099-2104.

AC C

17. Yamada S, Takagi Y, Nozaki K, Kikuta K, Hashimoto N: Risk factors for subsequent hemorrhage in patients with cerebral arteriovenous malformations. J Neurosurg. 2007; 107:965-972. 18. Hernesniemi JA, Dashti R, Juvela S, Väärt K, Niemelä M, Laakso A: Natural history of brain arteriovenous malformations: a long-term follow-up study of risk of hemorrhage in 238 patients. Neurosurgery. 2008; 63:823-831. 19. Robert T, Blanc R, Ciccio G, Redjem H, Fahed R, Smajda S, Piotin M. Anatomic and angiographic findings of cerebellar arteriovenous malformations: Report of a single centerexperience. J Neurol Sci. 2015; 358:357-361. 20. Al-Shahi R, Warlow C: A systematic review of the frequency and prognosis of arteriovenous

ACCEPTED MANUSCRIPT malformations of the brain in adults. Brain. 2001; 124:1900-1926. 21. Fleetwood IG, Steinberg GK: Arteriovenous malformations. Lancet. 2002; 359:863–873. 22. Wilkins RH: Natural history of intracranial vascular malformations: a review. Neurosurgery. 1985; 16:421-430.

RI PT

23. Mpotsaris A, Loehr C, Harati A, Lohmann F, Puchner M, Weber W. Interdisciplinary clinical management of high grade arteriovenous malformations and ruptured flow-related aneurysms in the posterior fossa, Interv. Neuroradiol. 2010; 16: 400-408.

24. Stefani MA, Porter PJ, terBrugge KG, Montanera W, Willinsky RA, Wallace MC.

SC

Angioarchitectural factors present in brain arteriovenous malformations associated with hemorrhagic presentation. Stroke. 2002; 33: 920-924

M AN U

25. Redekop G, TerBrugge K, Montanera W, Willinsky R. Arterial aneurysms associated with cerebral arteriovenous malformations: classification, incidence, and risk of hemorrhage, J Neurosurg. 1998; 89:539-546.

26. Rodriguez-Hernandez a, Zador Z, Rodriguez-Mena R, Lawton MT, Distal aneurysms of intracranial arteries: application of numerical nomenclature, predilection for cerebellar arteries, and

TE D

results of surgical management. World Neurosurg. 2013; 80:103-112. 27. Yilmaz A, Musluman AM, Kanat A, Cavusoglu H, Terzi Y, Aydin Y, The correlation between hematoma volume and outcome in ruptured posterior fossa arteriovenous malformations indicates the

EP

importance of surgical evacuation of hematomas, Turk Neurosurg. 2011; 21:152-159. 28. Mast H, Young WL, Koennecke HC, Sciacca RR, Osipov A, Pile-Spellman J, et al: Risk of

AC C

spontaneous haemorrhage after diagnosis of cerebral arteriovenous malformation. Lancet. 1997; 350:1065-1068.

29. Da Costa L, Wallace MC, Ter Brugge KG, O’Kelly C, Willinsky RA, Tymianski M: The natural history and predictive features of hemorrhage from brain arteriovenous malformations. Stroke. 2009; 40:100-105. 30. Graf CJ, Perret GE, Torner JC: Bleeding from cerebral arteriovenous malformations as part of their natural history. J Neurosurg. 1983; 58:331-337. 31. Fults D, Kelly DL Jr: Natural history of arteriovenous malformations of the brain: a clinical study. Neurosurgery.1984; 15:658-662.

ACCEPTED MANUSCRIPT 32. Itoyama Y, Uemura S, Ushio Y, Kuratsu J, Nonaka N, Wada H, et al: Natural course of unoperated intracranial arteriovenous malformations: study of 50 cases. J Neurosurg. 1989; 71:805-809. 33. Halim AX, Johnston SC, Singh V, McCulloch CE, Bennett JP, Achrol AS, et al: Longitudinal risk of intracranial hemorrhage in patients with arteriovenous malformation of the brain within a defined

RI PT

population. Stroke. 2004; 35:1697-1702. 34. Stapf C, Mohr JP, Sciacca RR, Hartmann A, Aagaard BD, Pile-Spellman J, et al: Incident hemorrhage risk of brain arteriovenous malformations located in the arterial borderzones. Stroke. 2000; 31: 2365-2368.

SC

35. Stapf C, Khaw AV, Sciacca RR, Hofmeister C, Schumacher HC, Pile-Spellman J, et al: Effect of

Stroke. 2003; 34:2664-2669.

M AN U

age on clinical and morphological characteristics in patients with brain arteriovenous malformation.

36. Crawford PM, West CR, Chadwick DW, Shaw MD: Arteriovenous malformations of the brain: natural history in unoperated patients. J Neurol Neurosurg Psychiatry. 1986; 49:1-10. 37. Kim H, Al-Shahi Salman R, McCulloch CE, Stapf C, Young WL; MARS Coinvestigators. Untreated brain arteriovenous malformation: patient-level meta-analysis of hemorrhage predictors.

TE D

Neurology. 2014; 83:590-597.

38. Brown RD Jr, Wiebers DO, Forbes G, O’Fallon WM, Piepgras DG, Marsh WR, et al: The natural history of unruptured intracranial arteriovenous malformations. J Neurosurg. 1988; 68:352-357.

EP

39. Marks MP, Lane B, Steinberg GK, Chang PJ: Hemorrhage in intracerebral arteriovenous malformations: angiographic determinants. Radiology. 1990; 176:807-813.

AC C

40. Vinuela F, Nombela L, Roach MR, Fox AJ, Pelz DM: Stenotic and occlusive disease of the venous drainage system of deep brain AVM’s. J Neurosurg. 1985; 63:180-184. 41. Brown RD Jr, Wiebers DO, Torner JC, O’Fallon WM: Frequency of intracranial hemorrhage as a presenting symptom and subtype analysis: a population-based study of intracranial vascular malformations in Olmsted Country, Minnesota. J Neurosurg. 1996; 85:29-32. 42. Hirai S, Mine S, Yamakami I, Ono J, Yamaura A: Angioarchitecture related to hemorrhage in cerebral arteriovenous malformations. Neurol Med Chir (Tokyo). 1998; 38 Suppl:165–170. 43. Mine S, Hirai S, Ono J, Yamaura A: Risk factors for poor outcome of untreated arteriovenous malformation. J Clin Neurosci. 2000; 7:503-506.

ACCEPTED MANUSCRIPT 44. Stefani MA, Porter PJ, terBrugge KG, Montanera W, Willinsky RA, Wallace MC: Large and deep brain arteriovenous malformations are associated with risk of future hemorrhage. Stroke. 2002; 33:1220-1224. 45. Brown RD Jr, Wiebers DO, Forbes GS: Unruptured intracranial aneurysms and arteriovenous

RI PT

malformations: frequency of intracranial hemorrhage and relationship of lesions. J Neurosurg. 1990; 73: 859-863.

46. Stapf C, Mohr JP, Pile-Spellman J, Sciacca RR, Hartmann A, Schumacher HC, et al: Concurrent arterial aneurysms in brain arteriovenous malformations with haemorrhagic presentation. J Neurol

SC

Neurosurg Psychiatry. 2002; 73:294-298.

47. Turjman F, Massoud TF, Vinuela F, Sayre JW, Guglielmi G, Duckwiler G: Correlation of the

M AN U

angioarchitectural features of cerebral arteriovenous malformations with clinical presentation of hemorrhage. Neurosurgery. 1995; 37:856-862.

48.Thompson RC, Steinberg GK, Levy RP, Marks MP: The management of patients with arteriovenous malformations and associated intracranial aneurysms. Neurosurgery. 1998; 43:202-211. Figure legends

TE D

Figure 1. Graphs of Kaplan–Meier curves of hemorrhage-free ratios according to the year of follow-up during the first five years after initial diagnosis: all patients together (A) and patients stratified by previous rupture (B), patient age group (C), pattern of venous drainage (D), AVM size (E)

EP

and cortical or deep location (F).

Figure 2. Graphs of Kaplan–Meier curves of hemorrhage-free ratios according to the year of

AC C

follow-up during the entire follow-up period: all patients together (A) and patients stratified by previous rupture (B), patient age group (C), pattern of venous drainage (D), AVM size (E) and cortical or deep location (F)

ACCEPTED MANUSCRIPT Table 1. Summary of the characteristics of the 149 patients with cerebellar AVMs according to subsequent hemorrhage during follow-up Characteristic Age

at

admission

All patients

No rupture

Rupture

(n=149)

(n=95)

(n=54)

28.1±14.8

31.3±15.1

22.4±12.5

P value

(years),

mean±SD Age group

0.001

Childhood

55 (37)

28 (30)

Adult

94 (63)

67 (70)

Sex, no. (%)

RI PT

0.014 27 (50) 27 (50)

0.732

63 (42)

39 (41)

Male

86 (58)

56 (59)

AVM rupture before admission, no. (%)c

24 (44) 30 (56)

SC

Female

121 (81)

75 (79)

46 (85)

No

28 (19)

20 (21)

8 (15)

AVM size, no. (%) Small (＜3cm) Large (≥3cm) AVM location, no. (%)

M AN U

Yes

0.016

68 (46)

36 (38)

32 (59)

81 (54)

59 (62)

22 (41) 0.679

Hemispheric

117 (79)

76 (80)

41 (76)

Vermis

32 (21)

19 (20)

13 (24)

Cortical Deep Single artery supply Yes

0.047

101 (68)

70 (74)

31 (57)

48 (32)

25 (26)

23 (43)

TE D

Cortical or deep location, no. (%)

0.080

94 (63)

65 (68)

29 (54)

55 (37)

30 (32)

25 (46)

Exclusively deep

42 (28)

20 (21)

22 (41)

Not exclusively deep

107 (72)

75 (79)

32 (59)

No

EP

Venous drainage, no. (%)

0.014

Associated aneurysm, no. (%)

AC C

0.238

0.548

Yes

13 (9)

7 (7)

6 (11)

No

136 (91)

88 (93)

48 (89)

627.1

204

423.1

-

4.2 (0.1-15.0)

3.8 (0.1-13.8)

4.5 (0.2-15.0)

0.060

Follow-up

Total person-years Median (range)

ACCEPTED MANUSCRIPT Table 2. Annual and cumulative rupture rates in relation to patient sex, age group, previous rupture, AVM size, cortical or deep location, pattern of venous drainage or feeding arteries Annual rupture rate, %

Cumulative rupture

Log-rank P values

rates, No. of patients

All patients

% (95% CI) 0-5 years

＞5 years

Whole

5 years

10 years

0-5 years

Whole

after

after

follow-up

after

after

after

follow-up

diagnosis

diagnosis

period

diagnosis

diagnosis

9.6

6.6

8.6

26 (21-31)

49 (43-55)

149

Sex

RI PT

Characteristic

86

10.7

5.2

8.8

35 (29-41)

44 (36-52)

Female

63

8.4

8.4

8.4

35 (28-42)

53 (44-62)

Age group Childhood

55

18.0

10.7

16.0

Adult

94

6.1

5.5

5.9

51 (43-59)

64 (54-74)

25 (20-30)

37 (30-44)

M AN U

Previous rupture

SC

Male

Ruptured

121

12.4

6.7

10.8

41 (35-47)

54 (47-61)

Unruptured

28

1.9

6.5

4.0

9 (3-15)

23 (12-34)

AVM size Small (＜3cm)

81

8.0

4.6

6.8

29 (23-35)

34 (26-42)

large (≥3cm)

68

11.4

8.7

10.5

39 (32-46)

56 (48-64)

AVM location 101

7.8

6.1

7.2

27 (22-32)

38 (31-45)

Deep

48

13.5

7.7

11.6

48 (39-57)

62 (52-72)

Artery supply Single

94

More than one

55

Venous drainage

TE D

Cortical

9.0

5.5

7.8

30 (24-36)

41 (33-49)

10.4

8.3

9.8

39 (31-47)

52 (43-61)

42

16.1

14.0

15.6

45 (37-53)

62 (50-74)

Not exclusively deep

107

7.4

5.1

6.6

29 (23-35)

40 (33-47)

13

19.6

7.7

15.6

55 (37-73)

77 (59-95)

136

9.0

6.5

8.2

32 (27-37)

43 (37-49)

Yes No

AC C

Associated aneurysm

EP

Exclusively deep

diagnosis

period

0.543

0.923

0.001

0.0004

0.005

0.022

0.247

0.115

0.086

0.111

0.533

0.314

0.017

0.002

0.116

0.159

ACCEPTED MANUSCRIPT Table 3. Risk factors for subsequent hemorrhage in the first five years using Cox proportional hazards model Variables

Univariate analysis

Multivariate analysis

HR (95%CI)

P value

HR (95%CI)

Male sex

0.547

1.215 (0.645-2.289)

-

-

Childhood

0.001

2.762 (1.478-5.160)

0.015

2.218 (1.171-4.202)

Previous rupture

0.014

5.924 (1.423-24.658)

0.032

4.949 (1.150-21.293)

AVM size ≥3cm

0.254

1.441 (0.769-2.698)

0.049

1.920 (1.003-3.677)

Exclusively deep venous drainage

0.021

2.093 (1.117-3.920)

0.009

2.370 (1.240-4.531)

Deep location

0.092

1.709 (0.916-3.187)

-

-

Single artery supply

0.537

0.822 (0.440-1.534)

-

-

Associated aneurysm

0.127

-

-

M AN U

SC

RI PT

P value

AC C

EP

TE D

2.077 (0.812-5.313)

ACCEPTED MANUSCRIPT Table 4. Risk factors for subsequent hemorrhage during the entire follow-up period using Cox proportional hazards model Variables

Univariate analysis

Multivariate analysis

RR (95%CI)

P value

HR (95%CI)

Male sex

0.924

0.974 (0.568-1.671)

-

-

Childhood

0.001

2.517 (1.470-4.311)

0.0003

2.804 (1.613-4.874)

Previous rupture

0.027

2.364 (1.101-5.076)

-

-

AVM size ≥3cm

0.121

1.537 (0.893-2.647)

0.017

1.988 (1.128-3.502)

Exclusively deep venous drainage

0.003

2.314 (1.332-4.019)

0.001

2.572 (1.466-4.514)

Deep location

0.117

1.541 (0.898-2.646)

-

-

Single artery supply

0.320

0.761 (0.445-1.302)

-

-

Associated aneurysm

0.169

-

-

M AN U

SC

RI PT

P value

AC C

EP

TE D

1.819 (0.775-4.265)

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

ACCEPTED MANUSCRIPT Highlights The natural history of cerebellar AVMs is more aggressive than their supratentorial counterparts. Childhood at diagnosis, large AVM size and AVMs with exclusively deep venous drainage are independent risk factors for subsequent hemorrhage in patients with cerebellar AVMs.

AC C

EP

TE D

M AN U

SC

may not significantly increase the risk in the following years.

RI PT

Previous rupture may increase the hemorrhagic risk during the first five years after diagnosis but

ACCEPTED MANUSCRIPT AUTHOR DECLARATION We wish to confirm that there are no known conflicts of interest associated with this publication and there has been no significant financial support for this work that could have influenced its outcome.

RI PT

We confirm that the manuscript has been read and approved by all named authors and that there are no other persons who satisfied the criteria for authorship but are not listed. We further confirm that the order of authors listed in the manuscript has been approved by all of us.

SC

We confirm that we have given due consideration to the protection of intellectual property associated with this work and that there are no impediments to publication, including the timing of publication, with respect to intellectual property. In so doing we confirm that we have followed the regulations of our institutions concerning intellectual property.

M AN U

We further confirm that any aspect of the work covered in this manuscript that has involved human patients has been conducted with the ethical approval of all relevant bodies and that such approvals are acknowledged within the manuscript.

TE D

We understand that the Corresponding Author is the sole contact for the Editorial process (including Editorial Manager and direct communications with the office). He is responsible for communicating with the other authors about progress, submissions of revisions and final approval of proofs. We confirm that we have provided a current, correct email address which is accessible by the Corresponding Author and which has been configured to accept email from ([email protected]) Signed by all authors as follows:

EP

Xianzeng Tong, 20-March-2016 Jun Wu, 20-March-2016

AC C

Fuxin Lin, 20-March-2016

Yong Cao, 20-March-2016 Yuanli Zhao, 20-March-2016 Shuo Wang, 20-March-2016 Jizong Zhao, 20-March-2016

ACCEPTED MANUSCRIPT Abbreviations AVM: arteriovenous malformation AVMs: arteriovenous malformations CT: computed tomography

RI PT

DSA: digital subtraction angiography

AC C

EP

TE D

M AN U

SC

MR: magnetic resonance imaging