- Email: [email protected]
Predictors of Initial Presentation with Hemorrhage in Patients with Cavernous Malformations
Journal Pre-proof Predictors of Initial Presentation with Hemorrhage in Patients with Cavernous Malformations Kelly D. Flemming, Shivram Kumar, Robert D. Brown, Jr., Giuseppe Lanzino PII:
S1878-8750(19)32605-1
DOI:
https://doi.org/10.1016/j.wneu.2019.09.161
Reference:
WNEU 13461
To appear in:
World Neurosurgery
Received Date: 27 June 2019 Revised Date:
27 September 2019
Accepted Date: 28 September 2019
Please cite this article as: Flemming KD, Kumar S, Brown Jr. RD, Lanzino G, Predictors of Initial Presentation with Hemorrhage in Patients with Cavernous Malformations, World Neurosurgery (2019), doi: https://doi.org/10.1016/j.wneu.2019.09.161. This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. 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. © 2019 Elsevier Inc. All rights reserved.
Predictors of Initial Presentation with Hemorrhage in Patients with Cavernous Malformations Kelly D. Flemming (1), Shivram Kumar (1), Robert D. Brown, Jr. (1), Giuseppe Lanzino (2) Departments of (1) Neurology, (2) Neurosurgery
Corresponding Author Kelly D. Flemming, MD Mayo Clinic, W8B Mayo Building 200 First Street SW Rochester, MN 55905 Phone: 507-266-4143 Fax: 507-266-4419 Email: [email protected] This publication [or project] was supported by Grant Number UL1 TR002377 from the National Center for Advancing Translational Sciences (NCATS). Its contents are solely the responsibility of the authors and do not necessarily represent the official views of the NIH.
Key Words: Cavernous malformation, hemorrhage, cavernoma , cavernous, angioma, estrogen, aspirin, antithrombotic
Flemming, page 1 1
ASTRACT
2
Background. Cavernous malformations (CM) are low-flow vascular malformations of the central nervous
3
system, incidental or present with hemorrhage, seizures, or focal neurologic deficit (FND). Little is known
4
about the time course of symptoms and the predictors of initial hemorrhagic presentation. Methods.
5
Beginning in 2015, consecutive patients with radiologically confirmed CM were recruited and
6
completed a structured interview and written survey at baseline. Medical records and magnetic
7
resonance imaging studies were reviewed. Data collected included: comorbid conditions, select
8
medication use, location of CM. Logistic-regression models determined predictors of initial presentation
9
with hemorrhage. Results. Of 202 patients, 58.4% were female (average age at diagnosis: 43.7 +/-16.5
10
years). 37.1%, 6.5%, and 14.8% of patients presented with hemorrhage, FND, and seizures respectively.
11
CM was an incidental finding in 40.6%. The majority of patients presenting with hemorrhage (66.7%)
12
deteriorated over 2-30 days following symptom onset. Brainstem location correlated with a higher
13
likelihood of initial presentation with hemorrhage. Aspirin use and NSAID use were more common in CM
14
patients who did not present with hemorrhage. The effect of estrogen and aspirin was stronger in the
15
sporadic than the familial form of CM. In women, estrogen use increased the likelihood of presenting with
16
hemorrhage. Conclusion. This prospective cohort study confirms brainstem location increases the
17
likelihood to present with hemorrhage, but also supports a potential role of select medications on the
18
initial clinical presentation of CM. Further data from combined cohorts may aid in determining whether
19
the influence of medications is different in familial versus sporadic form CM.
20 21 22
Background: Cavernous malformations (CM) are angiographically occult, low flow vascular malformations
23
which occur throughout the central nervous system. Pathologically, CM consist of endothelial lined
24
caverns lacking appropriately formed tight junctions.
25
lesions, together with surrounding iron deposits that result from hemorrhage or vascular leak lead to the
26
clinical manifestations of the disease. Patients may come to clinical attention due to intracerebral
27
hemorrhage (ICH), seizure (with or without coexistent hemorrhage), focal neurologic deficit (FND) without
1
The formation, growth, and hemorrhage of these
Flemming, page 2 28
hemorrhage, or as an incidental finding. Hemorrhagic CM carry a worse prognosis than those presenting
29
with seizure or FND.
2
30
Cavernous malformations can occur as sporadic, single lesions or may be familial. Patients with
31
the familial form have multiple lesions, each with independent risk for growth or hemorrhage. 3 Sporadic
32
CM are usually single lesions that may result from an altered vascular milieu due to the frequent
33
associated finding of a developmental venous anomaly (DVA). Some have hypothesized that the
34
associated DVA not only plays a role in pathogenesis of the CM, but may influence the risk of bleeding .
35
That is, if a DVA radicle(s) thrombose, there may be increased outflow resistance from the CM leading to
36
hemorrhage. If the latter theory were true, there may be differences in the factors influencing hemorrhage
37
risk in the sporadic and familial forms.
4,5
38
Many retrospective and prospective cohorts have provided general information about the clinical
39
presentation of patients with CMs including age at presentation, proportion of familial cases, frequency of
40
symptomatic and asymptomatic lesions, and location of lesion(s).
41
prospective studies systematically assessing the time course (onset to maximal severity) in patients
42
presenting with hemorrhage or the influence of concurrent comorbid conditions, medication use or
43
radiologic factors on initial clinical presentation.
44
3,6-13
However, there are no large
We hypothesize that clinical factors influence CM hemorrhagic presentation and may be
45
important in management considerations of these patients. In this study the aim was to 1) assess the
46
onset and timing of clinical presentation in patients with CM and 2) identifypotential clinical and radiologic
47
factors influencing hemorrhagic presentation.
48
Methods:
49
Patient Selection. With IRB approval, in January, 2015 we established a prospective CM registry. Adult
50
patients (18 years and older) with CM of the brain or spinal cord are asked to participate. Patients were
51
recruited from neurology and neurosurgery clinical practices in addition to screening radiologic records for
52
the diagnosis of cavernous malformation. Those patients who consented were included in the study
53
(Figure 1). For this particular study, we considered cerebral CM only.
Flemming, page 3 54
Clinical Data Collection. Demographic data, comorbid conditions, medications at diagnosis, and ongoing
55
use of medications are all collected from in person interviews, medical record review, and an initial
56
questionnaire. Specific comorbid conditions of interest included: history of a major infectious illness
57
(requiring hospitalization), venous clotting disorder (e.g., deep venous thrombosis), chronic inflammatory
58
conditions
59
of the following at first CM symptoms:
60
heparin, clopidogrel, ticagrelor), non-steroidal anti-inflammatory drugs (NSAIDS), vitamin D
61
supplementation, statin, fish oil, vitamin E, and serotonin reuptake inhibitors (SSRI). Antithrombotic
62
agents were included given prior retrospective studies suggesting a lower rate of CM hemorrhage in
63
patients on anti-thrombotics.
64
bleeding previously reported in other conditions. Vitamin D supplementation, NSAIDS, and statins were
65
recorded because of preclinical data suggesting potential benefit against hemorrhage.
66
assessed estrogen use in women. Estrogen is known to increase risk of venous clotting in some patients.
67
Thus, if CM hemorrhage is due to thromboses of a DVA radicle with backflow into a cavernous
68
malformation, then estrogen may raise the risk of hemorrhage in sporadic form CM with DVA.
69
14
, neoplasm, and self-reported concussion. Specific medications of interest included the use
5,15-17
aspirin, any antithrombotic (warfarin, direct oral anticoagulants,
Fish oil, vitamin E and SSRI were considered given potential risks of
18-21
We further
We defined “clinical presentation” as the first time a patient presented to medical attention and
70
was diagnosed with a CM. The date of symptom onset was recorded when known. We recorded which
71
patients presented with symptoms between October to March (fall/winter) versus April to September
72
(Spring/Summer) due to a prior study suggesting a seasonal variation in hemorrhage onset.
73
symptoms and time course of symptoms leading to the diagnosis were collected in detail. An acute
74
presentation was considered one in which symptoms reach their maximal severity within 1 day. Subacute
75
refers to a presentation in which symptoms reach their maximal severity >1 day, up to 10 days.
76
Progressive refers to a presentation in which symptoms progress to a maximal severity more than 10
77
days after the initial symptoms. As per standard guidelines,
78
clinical event involving both symptoms (headache, seizure, impaired consciousness, new/worsened focal
79
neurologic deficit referable to the anatomic location of the CM) and radiological, pathological, surgical or
80
CSF evidence of hemorrhage. Non-hemorrhagic focal neurologic deficit (FND) was defined as a new or
81
worsened focal neurologic deficit referable to the anatomic location of the CM but without obvious
23
22
The type of
clinical hemorrhage was defined as a
Flemming, page 4 23
82
evidence of hemorrhage.
83
hemorrhage status: hemorrhagic (with either FND or seizure) or non-hemorrhagic with seizures, FND or
84
unclear relationship, or incidental.
85
Radiologic Data Collected. The MRI performed to make the diagnosis of CM was reviewed by a staff
86
neuroradiologist in addition to the lead author (KDF). The initial MRI was often performed at other
87
institutions; thus no specific imaging protocol was used. We, however, felt it was important to use the
88
image at first diagnosis rather than one performed at a later time. The number of CMs was recorded
89
based on hemosiderin sensitive sequences [susceptibility weighted imaging (SWI), gradient recalled
90
echo (GRE)] when available or on T2 if SWI and GRE were not available. The locations of the cerebral
91
CM were divided into supratentorial-cortical, supratentorial-subcortical, infratentorial-brainstem,
92
infratentorial-cerebellum and other. Presence of an associated developmental venous anomaly (DVA),
93
and enhancement with gadoliniumwere recorded.
94
Data Analysis. Descriptive statistics including means, standard deviations, and frequencies were utilized
95
for patient characteristics and presenting symptoms. Univariate and multivariate logistic regression
96
models were used assess potential risk factors for hemorrhagic CM presentation versus other
97
presentation (related or incidental to the CM) and we report the odds ratio, 95% confidence intervals, and
98
likelihood ratio p values. Significant factors in the univariate analysis were included in the multivariate
99
analysis assessing the outcome of presentation with hemorrhage. JMP Pro software (version 14.1.0) was
100
used for analysis.
101
Results
102
Patients were classified into categories of presentation, subcategorized by
Between January 2015 and October 30, 2018, we had enrolled 202 patients with a cerebral CM 24
103
(Figure 1 and TABLE 1).
104
Half of patients were newly diagnosed within 1 year of their visit at our institution (median time from
105
diagnosis to consent: 1 year; range: 0-34.7 years). Fifty-seven (28.2%) of patients had multiple lesions
106
with 40 of these (19.8%) presumed to be familial CMs. The remainder had multiple CMs clustering around
107
a developmental venous anomaly. CMs were most commonly located in the supratentorial-cortical
The mean age at diagnosis was 43.7 +/-16.5 years and 58.4% were female.
Flemming, page 5 108
location (80; 39.6%), followed by brainstem (62; 30.7%), supratentorial-subcortical (46; 22.7%),
109
infratentorial cerebellum (12; 5.8%), and intraventricular (2; 1.0%).
110
At presentation, the CM was felt to be an incidental finding in 82 (40.6%) patients (Figure 1a).
111
Seventy-five (37.1%) presented with ICH, 13 (6.5%) with FND-non-hemorrhagic, 30 (14.8%) with seizure
112
(without hemorrhage), and 1 (0.5%) with a movement disorder and 1 (0.5%) with hydrocephalus. Of the
113
patients presenting with ICH, 57/75 (76.0%) suffered a focal neurologic deficit, 5 (6.7%) an associated
114
seizure, 10 (13.3%) presented with severe headache without focal neurologic deficit, and 3 (4%)
115
presented with a spell not clearly a seizure. Of those with a focal neurologic deficit due to hemorrhage
116
(n=57), 13 (22.8%) presented acutely, 27 (47.4%) sub-acutely and 11 (19.3%) in a progressive fashion
117
(Figure 1b). Six patients (10.5%) did not recall details of their presentation.
118
Brainstem location was more common in those presenting with ICH (58.7% vs. 15.1%; p<0.0001)
119
compared to those who presented without hemorrhage. There was no difference in the proportion of
120
patients with an associated developmental venous anomaly.
121
Patients who were taking aspirin, any antithrombotic, or NSAIDS were less likely to present with 24
122
hemorrhage (Table 1).
123
(49.7 +/-3.8 years vs. 42.8 +/-1.7; p=0.1). However, patients taking aspirin were more likely to be taking
124
a statin concomitantly (36.8% of patients on aspirin were also on a statin compared to 9.0% were on a
125
statin, but not taking aspirin; p=0.001). There was no other association of aspirin with other comorbidities
126
or interaction with other medications. However, there was a difference between sporadic and familial
127
CMs. Although numbers are small, when sub-grouped into those with familial disease and those with
128
sporadic disease, we found that the significance of these medications was only seen in the sporadic
129
group (Table 2).
130
Patients taking daily aspirin were not significantly older than those not on aspirin
When sub-setting only patients with known month of symptom onset (n=125), we did not find any
131
seasonal variation in ICH presentation. There were similar numbers of patients presenting with
132
hemorrhage during the fall/winter season as compared to the spring/summer season (54.4% vs. 54.5%;
133
p=0.96).
Flemming, page 6 134 135
In multivariate analysis, brainstem location (p<0.0001), aspirin (0.0018), and major infectious illness (p=0.024) remained significant as independent predictors of presentation with ICH (Table 3).
136
There were 118 women in the cohort, 42 of whom presented with hemorrhage and 76 without
137
hemorrhage. Twenty women were taking estrogen: 13 oral contraceptive, 3 estrogen topical patch, 3 oral
138
estrogen hormone replacement and 1 topical estrogen cream. 26.2% of women presenting with
139
hemorrhage were taking estrogen compared to 11.8% of women taking estrogen who did not present with
140
a hemorrhage (p=0.047). While numbers were small, we further assessed use of estrogen in women
141
with familial (n=23) versus sporadic CM (n=95). There were 5 familial CM patients taking oral estrogen.
142
Two of 5 patients taking estrogen presented with ICH, while the remaining 3 patients did not present with
143
hemorrhage (p =0.58). There were 15 sporadic CM patients taking estrogen, 9 (60%) of which presented
144
with ICH (p=0.05). In women, multivariate analysis confirms brainstem location (OR: 10.8; CI 3.68-32.34;
145
p<0.0001), estrogen (OR 5.51; CI 1.62-18.76; p=0.0047), aspirin (OR 0.098; CI 0.01-0.71; p=0.0075), and
146
major infectious illness (OR 0.0004; p=0.013) as independent predictors of presentation with hemorrhage.
147
Discussion
148
In a large, contemporary, prospective cohort of people with a CM, we characterized the onset and
149
clinical course and provided detailed data regarding the potential influence of comorbidities, medication
150
use, and radiologic factors on initial presentation with ICH.
151
The majority (66.7%) of those with a CM presenting with hemorrhage do so in a subacute to
152
progressive manner. Prior studies have suggested a step-wise decline, but none have quantified this
153
finding
154
Certain candidate medications proposed as an alternative to surgery to prevent future bleeding in CMs
155
may be used as a rescue-type method.
156
than a chronically, daily preventative. If a medication can prevent further leakage from the CM, patients
157
recognizing symptoms early could avoid the peak severity of symptoms which leads to increased
158
disability and adverse impact on employment and other life activities.
25,26
. This may become important when considering future potential pharmacological interventions.
27,28
That is, the treatment would start at symptom onset, rather
Flemming, page 7 159
Consistent with other studies,
6,8,11,29,30
patients with brainstem CM more commonly presented
160
with ICH. Why CM of the brainstem are more likely to bleed is a matter of debate with some suggesting it
161
to be a real phenomenon and others suggesting cohort bias to be more likely.
162
Taslimi and colleagues noted that the hemorrhage incidence rate correlated directly with the number of
163
brainstem lesions in each individual cohort study.
164
brainstem location more commonly presented initially with ICH as well in one of the only population based
165
studies . In that study, 32% of brainstem lesions presented with ICH.
29
31
In a meta-analysis,
Al-Shahi Salman and colleagues demonstrated that
6
166 167
We found that daily aspirin or any antithrombotic use was more common in those not presenting 5,15-17
168
with ICH. This finding is consistent with prior retrospective studies and one meta-analysis.
169
possible that antithrombotic drugs reduce venous stasis in an associated DVA in sporadic CMs thereby
170
reducing venous backflow into a CM and preventing ICH. This theory is supported by the findings that
171
the association was primarily seen in sporadic form CM in our study. It is also possible that this is a
172
spurious finding as patients requiring an antithrombotic are typically older while patients presenting with
173
CM ICH are younger. However, our data did not show an association with age. Moreover, a recent
174
meta-analysis has shown a potential protective role of antithrombotic medications against CM
175
hemorrhage.
176
use of statins. Statins may play a role in preventing leakiness of endothelium in CMs and may have had a
177
synergistic role in preventing hemorrhage form the CM.
178
but we are limited in determining the frequency and type of NSAID use in our patients because of the
179
design of our database and NSAIDs did not remain significant in the multivariate analysis. Many patients
180
with CM are told to avoid NSAIDS, but our data would suggest either a protective or neutral effect of
181
NSAID on hemorrhage at presentation. The significance of the antithrombotics and NSAIDs are intriguing
182
and deserve further investigation as it may have therapeutic implications in patients with sporadic CM.
183
17
It is
In patients taking aspirin in our prospective cohort, there was an increased concomitant
20,21,33-35
We noted similar findings with NSAIDS,
We found that women using estrogen hormones were more likely to present with ICH. We
184
hypothesize that the mechanism behind this interaction is similar to that of cerebral venous sinus
185
thrombosis. In patients taking estrogen-based medications, one or more radicles forming the DVA,
Flemming, page 8 186
invariably associated with sporadic CMs, may be prone to thrombosis, which in turn increases outflow
187
resistance resulting in higher risk of CM hemorrhage. Supporting this theory, we found the association to
188
be stronger in women with sporadic CMs and estrogen use but not in those with the familial form (which
189
lack associated DVA). Only case reports of estrogen associated with hemorrhage from CM exist until
190
now,
191
recommendations about estrogen use and its risk in patients with CM.
192
32
and confirmation of this findings in larger cohorts is warranted before making strong
Our study has limitations. It is not a population-based study. It is a prospective cohort from a
193
large academic Institution. In any academic center, there is tertiary referral bias. Specifically, our
194
institution may see more patients with symptomatic brainstem CMs due to the nature of our surgical
195
practice. Ascertainment of comorbidities and medications is subject to recall bias by patients. For
196
example, medications such as NSAIDs may be used intermittently and may not be included in a patient’s
197
medication list whereas daily medications may be more accurately recalled by patients, recorded in the
198
medical record, and collected. We tried to reduce this bias by thorough medical record review in addition
199
to patient surveys. Doses of medications and specific statin (i.e., atorvastatin vs. simvastatin) were not
200
collected which may have limited seeing a positive correlation of statin and presentation with ICH. Despite
201
these limitations, our study is, to our knowledge, the largest prospective CM registry with systematic
202
assessment of the impact of comorbidities and medications. As such it provides novel data on evolution
203
of clinical symptoms and factors associated with hemorrhagic presentation.
204
In conclusion, we have confirmed brainstem location as a high risk of presentation with
205
hemorrhage and identified several medications that may influence the initial clinical presentation of
206
patients with CM. We are assessing the same factors prospectively to assess whether they influence
207
further clinical activity and repeat hemorrhage over time. Further large combined prospective studies
208
could help verify these factors in the familial and sporadic forms which may have differing, important
209
mechanisms of bleeding and respond differently to therapeutics.
210
DECLARATIONS OF INTEREST: None
211
FUNDING
Flemming, page 9 212
This publication [or project] was supported by Grant Number UL1 TR002377 from the National Center for
213
Advancing Translational Sciences (NCATS). Its contents are solely the responsibility of the authors and
214
do not necessarily represent the official views of the NIH.
215
REFERENCES
216
1.
217 218
Med Genet. 2006; 43: 716-721 2.
219 220
Revencu N, Vikkula M. Cerebral cavernous malformation: new molecular and clinical insights. J
NIH State-of-the-Science Conference Statement on Multivitamin/Mineral Supplements and Chronic Disease Prevention. NIH consensus and state-of-the-science statements. 2006; 23: 1-30
3.
Labauge P, Brunereau L, Lévy C, Laberge S, Houtteville JP. The natural history of familial
221
cerebral cavernomas: a retrospective MRI study of 40 patients. Neuroradiology. 2000; 42: 327-
222
332
223
4.
224 225
Gross BA, Lin N, Du R, Day AL. The natural history of intracranial cavernous malformations. Neurosurg Focus. 2011; 30: E24.10.3171/2011.3.focus1165
5.
Schneble HM, Soumare A, Herve D, et al. Antithrombotic therapy and bleeding risk in a
226
prospective cohort study of patients with cerebral cavernous malformations. Stroke. 2012; 43:
227
3196-3199.10.1161/STROKEAHA.112.668533
228
6.
Al-Shahi Salman R, Hall JM, Horne MA, et al. Untreated clinical course of cerebral cavernous
229
malformations: a prospective, population-based cohort study. Lancet Neurol. 2012; 11: 217-
230
224.10.1016/S1474-4422(12)70004-2
231
7.
232 233
Flemming KD, Link MJ, Christianson TJ, Brown RD, Jr. . The Prospective Hemorrhage risk of intracerebral cavernous malformations. Neurology. 2012; 78: 632-636
8.
Horne MA FK, Su I-C, Stapf C, Jeon JP, Li D, Maxwell SS, White P, Christianson TJ, Agid R, Cho
234
W-S, Oh CW, Wu Z, Zhang J-T, Kim JE, ter Brugge K, Willinsky R, Brown Jr RD, Murray GD, Al-
235
Shahi Salman R, and the Cerebral Cavernous Malformations Individual Patient Data Meta-
236
analysis Collaborators. Clinical course of untreated cerebral cavernous malformations: an
237
individual patient data meta-analysis. Lancet Neurol. 2015; In Press:
238 239
9.
Kondziolka D, Lunsford LD, Kestle JR. The natural history of cerebral cavernous malformations. J Neurosurg. 1995; 83: 820-824
Flemming, page 10 240
10.
241 242
Kupersmith MJ, Kalish H, Epstein F, et al. Natural history of brainstem cavernous malformations. Neurosurgery. 2001; 48: 47-54
11.
Porter PJ, Willinsky RA, Harper W, Wallace MC. Cerebral cavernous malformations: natural
243
history and prognosis after clinical deterioration with or without hemorrhage. J Neurosurg. 1997;
244
87: 190-197
245
12.
246 247
familial occurrence. N Engl J Med. 1988; 319: 343-347 13.
248 249
Rigamonti D, Hadley MN, Drayer BP, et al. Cerebral cavernous malformations. Incidence and
Robinson JR, Awad IA, Little JR. Natural history of the cavernous angioma. J Neurosurg. 1991; 75: 709-714
14.
Kumar S, Lanzino G, Brinjikji W, Hocquard KW, Flemming KD. Infratentorial Developmental
250
Venous Abnormalities and Inflammation Increase Odds of Sporadic Cavernous Malformation. J
251
Stroke Cerebrovasc Dis. 2019.10.1016/j.jstrokecerebrovasdis.2019.02.025
252
15.
Bervini D, Jaeggi C, Mordasini P, Schucht P, Raabe A. Antithrombotic medication and bleeding
253
risk in patients with cerebral cavernous malformations: a cohort study. J Neurosurg. 2018: 1-
254
9.10.3171/2018.1.jns172547
255
16.
256 257
Flemming KD, Link MJ, Christianson TJ, Brown RD, Jr. Use of antithrombotic agents in patients with intracerebral cavernous malformations. J Neurosurg. 2013; 118: 43-46
17.
Zuurbier SM, Hickman CR, Tolias CS, et al. Long-term antithrombotic therapy and risk of
258
intracranial haemorrhage from cerebral cavernous malformations: a population-based cohort
259
study, systematic review, and meta-analysis. Lancet Neurol. 2019; 18: 935-
260
941.https://doi.org/10.1016/S1474-4422(19)30231-5
261
18.
Chohan MO, Marchio S, Morrison LA, et al. Emerging Pharmacologic Targets in Cerebral
262
Cavernous Malformation and Potential Strategies to Alter the Natural History of a Difficult
263
Disease: A Review. JAMA Neurol. 2019; 76: 492-500.10.1001/jamaneurol.2018.3634
264
19.
Gibson CC, Zhu W, Davis CT, et al. Strategy for identifying repurposed drugs for the treatment of
265
cerebral cavernous malformation. Circulation. 2015; 131: 289-
266
299.10.1161/circulationaha.114.010403
Flemming, page 11 267
20.
Polster SP, Stadnik A, Akers AL, et al. Atorvastatin Treatment of Cavernous Angiomas with
268
Symptomatic Hemorrhage Exploratory Proof of Concept (AT CASH EPOC) Trial. Neurosurgery.
269
2018.10.1093/neuros/nyy539
270
21.
Shenkar R, Peiper A, Pardo H, et al. Rho Kinase Inhibition Blunts Lesion Development and
271
Hemorrhage in Murine Models of Aggressive Pdcd10/Ccm3 Disease. Stroke. 2019; 50: 738-
272
744.10.1161/strokeaha.118.024058
273
22.
Flemming KD, Brown RD, Link MJ. Seasonal variation in hemorrhage and focal neurologic deficit
274
due to intracerebral cavernous malformations. J Clin Neurosci. 2015; 22: 969-
275
971.10.1016/j.jocn.2015.01.007
276
23.
Al-Shahi Salman R, Berg MJ, Morrison L, Awad IA, Angioma Alliance Scientific Advisory B.
277
Hemorrhage from cavernous malformations of the brain: definition and reporting standards.
278
Angioma Alliance Scientific Advisory Board. Stroke. 2008; 39: 3222-
279
3230.10.1161/STROKEAHA.108.515544
280
24.
Kernan WN, Ovbiagele B, Black HR, et al. Guidelines for the prevention of stroke in patients with
281
stroke and transient ischemic attack: a guideline for healthcare professionals from the American
282
Heart Association/American Stroke Association. Stroke. 2014; 45: 2160-
283
2236.10.1161/str.0000000000000024
284
25.
Abe M, Kjellberg RN, Adams RD. Clinical presentations of vascular malformations of the brain
285
stem: comparison of angiographically positive and negative types. J Neurol Neurosurg
286
Psychiatry. 1989; 52: 167-175
287
26.
288 289
Aiba T, Tanaka R, Koike T, Kameyama S, Takeda N, Komata T. Natural history of intracranial cavernous malformations. J Neurosurg. 1995; 83: 56-59
27.
Chohan MO, Marchio S, Morrison LA, et al. Emerging Pharmacologic Targets in Cerebral
290
Cavernous Malformation and Potential Strategies to Alter the Natural History of a Difficult
291
Disease: A Review. JAMA Neurol. 2018.10.1001/jamaneurol.2018.3634
292 293 294
28.
Retta SF, Glading AJ. Oxidative stress and inflammation in cerebral cavernous malformation disease pathogenesis: Two sides of the same coin.
. Int J Biochem Cell Biol. 2016; 81: 254-270
Flemming, page 12 295
29.
Taslimi S, Modabbernia A, Amin-Hanjani S, Barker FG, Macdonald RL. Natural history of
296
cavernous malformation: Systematic review and meta-analysis of 25 studies. Neurology. 2016;
297
86: 1984-1991
298
30.
Dammann P, Jabbarli R, Wittek P, et al. Solitary Sporadic Cerebral Cavernous Malformations:
299
Risk Factors of First or Recurrent Symptomatic Hemorrhage and Associated Functional
300
Impairment. World Neurosurg. 2016; 91: 73-80.10.1016/j.wneu.2016.03.080
301
31.
302 303
review of neurotherapeutics. 2015; 15: 1109-1111.10.1586/14737175.2015.1071193 32.
304 305
Gazzaz M, Sichez J, Capelle L, Fohanno D. [Recurrent bleeding of thalamic cavernous angioma under hormonal treatment. A case report]. Neuro-Chirurgie. 1999; 45: 413-416
33.
306 307
Starke RM. Do brainstem cavernous malformations have a higher rate of hemorrhage? Expert
Whitehead KJ, Chan AC, Navankasattusas S, et al. The cerebral cavernous malformation signaling pathway promotes vascular integrity via Rho GTPases. Nat Med. 2009; 15: 177-184
34.
Girard R, Zeineddine HA, Koskimaki J, et al. Plasma Biomarkers of Inflammation and
308
Angiogenesis Predict Cerebral Cavernous Malformation Symptomatic Hemorrhage or Lesional
309
Growth. Circ Res. 2018; 122: 1716-1721.10.1161/CIRCRESAHA.118.312680
310
35.
311
Polster SP, Cao Y, Carroll T, et al. Trial Readiness in Cavernous Angiomas With Symptomatic Hemorrhage (CASH). Neurosurgery. 2019; 84: 954-964.10.1093/neuros/nyy108
312 313
Figures
314
Figure 1: Screening and Study Population
315
Figure 2: a) Modes of presentation in patients with CM b) Onset and time course to peak severity of ICH
316
symptoms in patients with CM
317 318
TABLE 1 - Patient Characteristics in Those With and Without Hemorrhage at Presentation All patients
N
202
Presentation
Presentation –
with ICH
no ICH
75
127
P Value
Odds ratio (95% CI)
Demographic Information Age at Diagnosis
43.7 +/-16.5
42.0 +/- 16.6
44.7+/-16.4
0.27
0.46 (0.12-1.8)
Sex (Female)
118 (58.4%)
42 (56.0%)
76 (59.8%)
0.59
0.8 (0.47-1.52)
Familial form
40 (19.8%)
13 (17.3%)
28 (22.0%)
0.42
0.74 (0.35-1.51)
Multiple lesions
57 (28.2%)
19 (25.3%)
38 (30.0%)
0.48
0.8 (0.42-1.51)
Symptom began
68/125 (54.4%)
38/70 (54.3%)
30/55 (54.5%)
0.96
0.98 (0.47-2.00)
Major infectious
15 (7.6%)
1 (1.3%)
14 (11.3%)
0.011
0.37 (0.014-0.87)
illness
7 missing data
4 missing data
3 missing data
Chronic inflammatory
56 (27.8%)
17 (22.7%)
39 (30.7%)
0.22
0.66 (0.33-1.26)
39 (19.4%)
12 (16.0%)
27 (21.4%)
0.35
0.69 (0.32-1.48)
(years)
between Octoberand March* (n=125) Medical history
disease Neoplastic disease
Medication use at presentation
Any antithrombotic
42 (20.8%)
6 (8.0%)
36 (28.3%)
0.0012
0.21 (0.08-0.55)
Aspirin
37 (18.3%)
5 (6.7%)
32 (25.2%)
0.0022
0.21 (0.07-0.57)
NSAIDs
74 (36.6%)
18 (24.0%)
56 (44.1%)
0.0047
0.40 (0.21-0.74)
Vitamin D
57 (28.2%)
17 (22.7%)
40 (31.5%)
0.18
0.64 (0.33-1.23)
Vitamin E
13 (6.4%)
5 (6.7%)
8 (6.3%)
0.92
1.06 (0.33-3.37)
Fish Oil
38 (18.8%)
12 (16.0%)
26 (20.5%)
0.42
0.74 (0.35-1.57)
Statin
34 (16.8%)
9 (12.0%)
25 (19.7%)
0.16
0.55 (0.24-1.27)
SSRI
21 (10.4%)
6 (6.7%)
16 (12.6%)
0.23
0.49 (0.15-1.32)
48 (23.8%)
20 (26.7%)
28 (22.0%)
0.46
0.46 (0.66-2.49)
18 (8.9%)
7 (9.3%)
11 (8.7%)
0.87
1.08 (0.44-2.93)
Tobacco use past or current Alcohol greater than recommended***
Radiologic Information Brainstem location
62 (30.7%)
44 (58.7%)
19 (15.1%)
<0.0001
8.07 (4.13-15.77)
Associated DVA
61 (38.1%)
23 (38.9%)
38 (37.6%)
0.86
0.86(0.55-2.05)
42 missing data
16 missing
26 missing
NSAID=nonsteroidal anti-inflammatory drug, SSR=serotonin reuptake inhibitor, DVA=developmental venous anomaly. * Denominator includes only those patients with known month of onset ** Only women included in this calculation *** Recommendations as per the American Heart Association for stroke prevention
24
Table 2 Comparison of the Significance of Antithrombotics and NSAIDS in Sporadic vs. Familial Cavernous Malformation Sporadic Form
Any antithrombotic Aspirin NSAIDS Familial Form
Present with ICH N=62 4 (6.4%) 4 (6.4%) 15 (24.2%)
Present with ICH N=13 Any antithrombotic 2 (15.4%) Aspirin 1 (7.7%) NSAIDS 3 (23.1%) NSAIDs=non-steroidal anti-inflammatory drugs
No ICH N=99 30 (30.3%) 26 (26.3%) 44 (44.4%)
P- value
No ICH N=28 6 (21.4%) 6 (21.4%) 12 (42.8%)
P-Value
0.0003 0.0016 0.0095
1.000 0.3986 0.3048
Table 3: Multivariate Analysis of Predictors of Presentation with Hemorrhage
Variable
Odds Ratio
95%
P value
Confidence Interval Brainstem Location
9.49
4.43-20.34
<0.0001
Aspirin
0.19
0.06-0.60
0.0018
NSAIDS
0.60
0.28-1.39
0.1950
Major Infectious Illness
0.07
0.02-0.008
0.0047
ABBREVIATIONS CM
cavernous malformation
CCMs
cerebral cavernous malformations
DVA
Developmental venous anomaly
FND
Focal neurologic deficit
GRE
Gradient recalled echo
ICH
Intracerebral hemorrhage
NSAIDS
Non-steroidal anti-inflammatory drugs
SWI
Susceptibility-weighted imaging
S1878-8750(19)32605-1
DOI:
https://doi.org/10.1016/j.wneu.2019.09.161
Reference:
WNEU 13461
To appear in:
World Neurosurgery
Received Date: 27 June 2019 Revised Date:
27 September 2019
Accepted Date: 28 September 2019
Please cite this article as: Flemming KD, Kumar S, Brown Jr. RD, Lanzino G, Predictors of Initial Presentation with Hemorrhage in Patients with Cavernous Malformations, World Neurosurgery (2019), doi: https://doi.org/10.1016/j.wneu.2019.09.161. This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. 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. © 2019 Elsevier Inc. All rights reserved.
Predictors of Initial Presentation with Hemorrhage in Patients with Cavernous Malformations Kelly D. Flemming (1), Shivram Kumar (1), Robert D. Brown, Jr. (1), Giuseppe Lanzino (2) Departments of (1) Neurology, (2) Neurosurgery
Corresponding Author Kelly D. Flemming, MD Mayo Clinic, W8B Mayo Building 200 First Street SW Rochester, MN 55905 Phone: 507-266-4143 Fax: 507-266-4419 Email: [email protected] This publication [or project] was supported by Grant Number UL1 TR002377 from the National Center for Advancing Translational Sciences (NCATS). Its contents are solely the responsibility of the authors and do not necessarily represent the official views of the NIH.
Key Words: Cavernous malformation, hemorrhage, cavernoma , cavernous, angioma, estrogen, aspirin, antithrombotic
Flemming, page 1 1
ASTRACT
2
Background. Cavernous malformations (CM) are low-flow vascular malformations of the central nervous
3
system, incidental or present with hemorrhage, seizures, or focal neurologic deficit (FND). Little is known
4
about the time course of symptoms and the predictors of initial hemorrhagic presentation. Methods.
5
Beginning in 2015, consecutive patients with radiologically confirmed CM were recruited and
6
completed a structured interview and written survey at baseline. Medical records and magnetic
7
resonance imaging studies were reviewed. Data collected included: comorbid conditions, select
8
medication use, location of CM. Logistic-regression models determined predictors of initial presentation
9
with hemorrhage. Results. Of 202 patients, 58.4% were female (average age at diagnosis: 43.7 +/-16.5
10
years). 37.1%, 6.5%, and 14.8% of patients presented with hemorrhage, FND, and seizures respectively.
11
CM was an incidental finding in 40.6%. The majority of patients presenting with hemorrhage (66.7%)
12
deteriorated over 2-30 days following symptom onset. Brainstem location correlated with a higher
13
likelihood of initial presentation with hemorrhage. Aspirin use and NSAID use were more common in CM
14
patients who did not present with hemorrhage. The effect of estrogen and aspirin was stronger in the
15
sporadic than the familial form of CM. In women, estrogen use increased the likelihood of presenting with
16
hemorrhage. Conclusion. This prospective cohort study confirms brainstem location increases the
17
likelihood to present with hemorrhage, but also supports a potential role of select medications on the
18
initial clinical presentation of CM. Further data from combined cohorts may aid in determining whether
19
the influence of medications is different in familial versus sporadic form CM.
20 21 22
Background: Cavernous malformations (CM) are angiographically occult, low flow vascular malformations
23
which occur throughout the central nervous system. Pathologically, CM consist of endothelial lined
24
caverns lacking appropriately formed tight junctions.
25
lesions, together with surrounding iron deposits that result from hemorrhage or vascular leak lead to the
26
clinical manifestations of the disease. Patients may come to clinical attention due to intracerebral
27
hemorrhage (ICH), seizure (with or without coexistent hemorrhage), focal neurologic deficit (FND) without
1
The formation, growth, and hemorrhage of these
Flemming, page 2 28
hemorrhage, or as an incidental finding. Hemorrhagic CM carry a worse prognosis than those presenting
29
with seizure or FND.
2
30
Cavernous malformations can occur as sporadic, single lesions or may be familial. Patients with
31
the familial form have multiple lesions, each with independent risk for growth or hemorrhage. 3 Sporadic
32
CM are usually single lesions that may result from an altered vascular milieu due to the frequent
33
associated finding of a developmental venous anomaly (DVA). Some have hypothesized that the
34
associated DVA not only plays a role in pathogenesis of the CM, but may influence the risk of bleeding .
35
That is, if a DVA radicle(s) thrombose, there may be increased outflow resistance from the CM leading to
36
hemorrhage. If the latter theory were true, there may be differences in the factors influencing hemorrhage
37
risk in the sporadic and familial forms.
4,5
38
Many retrospective and prospective cohorts have provided general information about the clinical
39
presentation of patients with CMs including age at presentation, proportion of familial cases, frequency of
40
symptomatic and asymptomatic lesions, and location of lesion(s).
41
prospective studies systematically assessing the time course (onset to maximal severity) in patients
42
presenting with hemorrhage or the influence of concurrent comorbid conditions, medication use or
43
radiologic factors on initial clinical presentation.
44
3,6-13
However, there are no large
We hypothesize that clinical factors influence CM hemorrhagic presentation and may be
45
important in management considerations of these patients. In this study the aim was to 1) assess the
46
onset and timing of clinical presentation in patients with CM and 2) identifypotential clinical and radiologic
47
factors influencing hemorrhagic presentation.
48
Methods:
49
Patient Selection. With IRB approval, in January, 2015 we established a prospective CM registry. Adult
50
patients (18 years and older) with CM of the brain or spinal cord are asked to participate. Patients were
51
recruited from neurology and neurosurgery clinical practices in addition to screening radiologic records for
52
the diagnosis of cavernous malformation. Those patients who consented were included in the study
53
(Figure 1). For this particular study, we considered cerebral CM only.
Flemming, page 3 54
Clinical Data Collection. Demographic data, comorbid conditions, medications at diagnosis, and ongoing
55
use of medications are all collected from in person interviews, medical record review, and an initial
56
questionnaire. Specific comorbid conditions of interest included: history of a major infectious illness
57
(requiring hospitalization), venous clotting disorder (e.g., deep venous thrombosis), chronic inflammatory
58
conditions
59
of the following at first CM symptoms:
60
heparin, clopidogrel, ticagrelor), non-steroidal anti-inflammatory drugs (NSAIDS), vitamin D
61
supplementation, statin, fish oil, vitamin E, and serotonin reuptake inhibitors (SSRI). Antithrombotic
62
agents were included given prior retrospective studies suggesting a lower rate of CM hemorrhage in
63
patients on anti-thrombotics.
64
bleeding previously reported in other conditions. Vitamin D supplementation, NSAIDS, and statins were
65
recorded because of preclinical data suggesting potential benefit against hemorrhage.
66
assessed estrogen use in women. Estrogen is known to increase risk of venous clotting in some patients.
67
Thus, if CM hemorrhage is due to thromboses of a DVA radicle with backflow into a cavernous
68
malformation, then estrogen may raise the risk of hemorrhage in sporadic form CM with DVA.
69
14
, neoplasm, and self-reported concussion. Specific medications of interest included the use
5,15-17
aspirin, any antithrombotic (warfarin, direct oral anticoagulants,
Fish oil, vitamin E and SSRI were considered given potential risks of
18-21
We further
We defined “clinical presentation” as the first time a patient presented to medical attention and
70
was diagnosed with a CM. The date of symptom onset was recorded when known. We recorded which
71
patients presented with symptoms between October to March (fall/winter) versus April to September
72
(Spring/Summer) due to a prior study suggesting a seasonal variation in hemorrhage onset.
73
symptoms and time course of symptoms leading to the diagnosis were collected in detail. An acute
74
presentation was considered one in which symptoms reach their maximal severity within 1 day. Subacute
75
refers to a presentation in which symptoms reach their maximal severity >1 day, up to 10 days.
76
Progressive refers to a presentation in which symptoms progress to a maximal severity more than 10
77
days after the initial symptoms. As per standard guidelines,
78
clinical event involving both symptoms (headache, seizure, impaired consciousness, new/worsened focal
79
neurologic deficit referable to the anatomic location of the CM) and radiological, pathological, surgical or
80
CSF evidence of hemorrhage. Non-hemorrhagic focal neurologic deficit (FND) was defined as a new or
81
worsened focal neurologic deficit referable to the anatomic location of the CM but without obvious
23
22
The type of
clinical hemorrhage was defined as a
Flemming, page 4 23
82
evidence of hemorrhage.
83
hemorrhage status: hemorrhagic (with either FND or seizure) or non-hemorrhagic with seizures, FND or
84
unclear relationship, or incidental.
85
Radiologic Data Collected. The MRI performed to make the diagnosis of CM was reviewed by a staff
86
neuroradiologist in addition to the lead author (KDF). The initial MRI was often performed at other
87
institutions; thus no specific imaging protocol was used. We, however, felt it was important to use the
88
image at first diagnosis rather than one performed at a later time. The number of CMs was recorded
89
based on hemosiderin sensitive sequences [susceptibility weighted imaging (SWI), gradient recalled
90
echo (GRE)] when available or on T2 if SWI and GRE were not available. The locations of the cerebral
91
CM were divided into supratentorial-cortical, supratentorial-subcortical, infratentorial-brainstem,
92
infratentorial-cerebellum and other. Presence of an associated developmental venous anomaly (DVA),
93
and enhancement with gadoliniumwere recorded.
94
Data Analysis. Descriptive statistics including means, standard deviations, and frequencies were utilized
95
for patient characteristics and presenting symptoms. Univariate and multivariate logistic regression
96
models were used assess potential risk factors for hemorrhagic CM presentation versus other
97
presentation (related or incidental to the CM) and we report the odds ratio, 95% confidence intervals, and
98
likelihood ratio p values. Significant factors in the univariate analysis were included in the multivariate
99
analysis assessing the outcome of presentation with hemorrhage. JMP Pro software (version 14.1.0) was
100
used for analysis.
101
Results
102
Patients were classified into categories of presentation, subcategorized by
Between January 2015 and October 30, 2018, we had enrolled 202 patients with a cerebral CM 24
103
(Figure 1 and TABLE 1).
104
Half of patients were newly diagnosed within 1 year of their visit at our institution (median time from
105
diagnosis to consent: 1 year; range: 0-34.7 years). Fifty-seven (28.2%) of patients had multiple lesions
106
with 40 of these (19.8%) presumed to be familial CMs. The remainder had multiple CMs clustering around
107
a developmental venous anomaly. CMs were most commonly located in the supratentorial-cortical
The mean age at diagnosis was 43.7 +/-16.5 years and 58.4% were female.
Flemming, page 5 108
location (80; 39.6%), followed by brainstem (62; 30.7%), supratentorial-subcortical (46; 22.7%),
109
infratentorial cerebellum (12; 5.8%), and intraventricular (2; 1.0%).
110
At presentation, the CM was felt to be an incidental finding in 82 (40.6%) patients (Figure 1a).
111
Seventy-five (37.1%) presented with ICH, 13 (6.5%) with FND-non-hemorrhagic, 30 (14.8%) with seizure
112
(without hemorrhage), and 1 (0.5%) with a movement disorder and 1 (0.5%) with hydrocephalus. Of the
113
patients presenting with ICH, 57/75 (76.0%) suffered a focal neurologic deficit, 5 (6.7%) an associated
114
seizure, 10 (13.3%) presented with severe headache without focal neurologic deficit, and 3 (4%)
115
presented with a spell not clearly a seizure. Of those with a focal neurologic deficit due to hemorrhage
116
(n=57), 13 (22.8%) presented acutely, 27 (47.4%) sub-acutely and 11 (19.3%) in a progressive fashion
117
(Figure 1b). Six patients (10.5%) did not recall details of their presentation.
118
Brainstem location was more common in those presenting with ICH (58.7% vs. 15.1%; p<0.0001)
119
compared to those who presented without hemorrhage. There was no difference in the proportion of
120
patients with an associated developmental venous anomaly.
121
Patients who were taking aspirin, any antithrombotic, or NSAIDS were less likely to present with 24
122
hemorrhage (Table 1).
123
(49.7 +/-3.8 years vs. 42.8 +/-1.7; p=0.1). However, patients taking aspirin were more likely to be taking
124
a statin concomitantly (36.8% of patients on aspirin were also on a statin compared to 9.0% were on a
125
statin, but not taking aspirin; p=0.001). There was no other association of aspirin with other comorbidities
126
or interaction with other medications. However, there was a difference between sporadic and familial
127
CMs. Although numbers are small, when sub-grouped into those with familial disease and those with
128
sporadic disease, we found that the significance of these medications was only seen in the sporadic
129
group (Table 2).
130
Patients taking daily aspirin were not significantly older than those not on aspirin
When sub-setting only patients with known month of symptom onset (n=125), we did not find any
131
seasonal variation in ICH presentation. There were similar numbers of patients presenting with
132
hemorrhage during the fall/winter season as compared to the spring/summer season (54.4% vs. 54.5%;
133
p=0.96).
Flemming, page 6 134 135
In multivariate analysis, brainstem location (p<0.0001), aspirin (0.0018), and major infectious illness (p=0.024) remained significant as independent predictors of presentation with ICH (Table 3).
136
There were 118 women in the cohort, 42 of whom presented with hemorrhage and 76 without
137
hemorrhage. Twenty women were taking estrogen: 13 oral contraceptive, 3 estrogen topical patch, 3 oral
138
estrogen hormone replacement and 1 topical estrogen cream. 26.2% of women presenting with
139
hemorrhage were taking estrogen compared to 11.8% of women taking estrogen who did not present with
140
a hemorrhage (p=0.047). While numbers were small, we further assessed use of estrogen in women
141
with familial (n=23) versus sporadic CM (n=95). There were 5 familial CM patients taking oral estrogen.
142
Two of 5 patients taking estrogen presented with ICH, while the remaining 3 patients did not present with
143
hemorrhage (p =0.58). There were 15 sporadic CM patients taking estrogen, 9 (60%) of which presented
144
with ICH (p=0.05). In women, multivariate analysis confirms brainstem location (OR: 10.8; CI 3.68-32.34;
145
p<0.0001), estrogen (OR 5.51; CI 1.62-18.76; p=0.0047), aspirin (OR 0.098; CI 0.01-0.71; p=0.0075), and
146
major infectious illness (OR 0.0004; p=0.013) as independent predictors of presentation with hemorrhage.
147
Discussion
148
In a large, contemporary, prospective cohort of people with a CM, we characterized the onset and
149
clinical course and provided detailed data regarding the potential influence of comorbidities, medication
150
use, and radiologic factors on initial presentation with ICH.
151
The majority (66.7%) of those with a CM presenting with hemorrhage do so in a subacute to
152
progressive manner. Prior studies have suggested a step-wise decline, but none have quantified this
153
finding
154
Certain candidate medications proposed as an alternative to surgery to prevent future bleeding in CMs
155
may be used as a rescue-type method.
156
than a chronically, daily preventative. If a medication can prevent further leakage from the CM, patients
157
recognizing symptoms early could avoid the peak severity of symptoms which leads to increased
158
disability and adverse impact on employment and other life activities.
25,26
. This may become important when considering future potential pharmacological interventions.
27,28
That is, the treatment would start at symptom onset, rather
Flemming, page 7 159
Consistent with other studies,
6,8,11,29,30
patients with brainstem CM more commonly presented
160
with ICH. Why CM of the brainstem are more likely to bleed is a matter of debate with some suggesting it
161
to be a real phenomenon and others suggesting cohort bias to be more likely.
162
Taslimi and colleagues noted that the hemorrhage incidence rate correlated directly with the number of
163
brainstem lesions in each individual cohort study.
164
brainstem location more commonly presented initially with ICH as well in one of the only population based
165
studies . In that study, 32% of brainstem lesions presented with ICH.
29
31
In a meta-analysis,
Al-Shahi Salman and colleagues demonstrated that
6
166 167
We found that daily aspirin or any antithrombotic use was more common in those not presenting 5,15-17
168
with ICH. This finding is consistent with prior retrospective studies and one meta-analysis.
169
possible that antithrombotic drugs reduce venous stasis in an associated DVA in sporadic CMs thereby
170
reducing venous backflow into a CM and preventing ICH. This theory is supported by the findings that
171
the association was primarily seen in sporadic form CM in our study. It is also possible that this is a
172
spurious finding as patients requiring an antithrombotic are typically older while patients presenting with
173
CM ICH are younger. However, our data did not show an association with age. Moreover, a recent
174
meta-analysis has shown a potential protective role of antithrombotic medications against CM
175
hemorrhage.
176
use of statins. Statins may play a role in preventing leakiness of endothelium in CMs and may have had a
177
synergistic role in preventing hemorrhage form the CM.
178
but we are limited in determining the frequency and type of NSAID use in our patients because of the
179
design of our database and NSAIDs did not remain significant in the multivariate analysis. Many patients
180
with CM are told to avoid NSAIDS, but our data would suggest either a protective or neutral effect of
181
NSAID on hemorrhage at presentation. The significance of the antithrombotics and NSAIDs are intriguing
182
and deserve further investigation as it may have therapeutic implications in patients with sporadic CM.
183
17
It is
In patients taking aspirin in our prospective cohort, there was an increased concomitant
20,21,33-35
We noted similar findings with NSAIDS,
We found that women using estrogen hormones were more likely to present with ICH. We
184
hypothesize that the mechanism behind this interaction is similar to that of cerebral venous sinus
185
thrombosis. In patients taking estrogen-based medications, one or more radicles forming the DVA,
Flemming, page 8 186
invariably associated with sporadic CMs, may be prone to thrombosis, which in turn increases outflow
187
resistance resulting in higher risk of CM hemorrhage. Supporting this theory, we found the association to
188
be stronger in women with sporadic CMs and estrogen use but not in those with the familial form (which
189
lack associated DVA). Only case reports of estrogen associated with hemorrhage from CM exist until
190
now,
191
recommendations about estrogen use and its risk in patients with CM.
192
32
and confirmation of this findings in larger cohorts is warranted before making strong
Our study has limitations. It is not a population-based study. It is a prospective cohort from a
193
large academic Institution. In any academic center, there is tertiary referral bias. Specifically, our
194
institution may see more patients with symptomatic brainstem CMs due to the nature of our surgical
195
practice. Ascertainment of comorbidities and medications is subject to recall bias by patients. For
196
example, medications such as NSAIDs may be used intermittently and may not be included in a patient’s
197
medication list whereas daily medications may be more accurately recalled by patients, recorded in the
198
medical record, and collected. We tried to reduce this bias by thorough medical record review in addition
199
to patient surveys. Doses of medications and specific statin (i.e., atorvastatin vs. simvastatin) were not
200
collected which may have limited seeing a positive correlation of statin and presentation with ICH. Despite
201
these limitations, our study is, to our knowledge, the largest prospective CM registry with systematic
202
assessment of the impact of comorbidities and medications. As such it provides novel data on evolution
203
of clinical symptoms and factors associated with hemorrhagic presentation.
204
In conclusion, we have confirmed brainstem location as a high risk of presentation with
205
hemorrhage and identified several medications that may influence the initial clinical presentation of
206
patients with CM. We are assessing the same factors prospectively to assess whether they influence
207
further clinical activity and repeat hemorrhage over time. Further large combined prospective studies
208
could help verify these factors in the familial and sporadic forms which may have differing, important
209
mechanisms of bleeding and respond differently to therapeutics.
210
DECLARATIONS OF INTEREST: None
211
FUNDING
Flemming, page 9 212
This publication [or project] was supported by Grant Number UL1 TR002377 from the National Center for
213
Advancing Translational Sciences (NCATS). Its contents are solely the responsibility of the authors and
214
do not necessarily represent the official views of the NIH.
215
REFERENCES
216
1.
217 218
Med Genet. 2006; 43: 716-721 2.
219 220
Revencu N, Vikkula M. Cerebral cavernous malformation: new molecular and clinical insights. J
NIH State-of-the-Science Conference Statement on Multivitamin/Mineral Supplements and Chronic Disease Prevention. NIH consensus and state-of-the-science statements. 2006; 23: 1-30
3.
Labauge P, Brunereau L, Lévy C, Laberge S, Houtteville JP. The natural history of familial
221
cerebral cavernomas: a retrospective MRI study of 40 patients. Neuroradiology. 2000; 42: 327-
222
332
223
4.
224 225
Gross BA, Lin N, Du R, Day AL. The natural history of intracranial cavernous malformations. Neurosurg Focus. 2011; 30: E24.10.3171/2011.3.focus1165
5.
Schneble HM, Soumare A, Herve D, et al. Antithrombotic therapy and bleeding risk in a
226
prospective cohort study of patients with cerebral cavernous malformations. Stroke. 2012; 43:
227
3196-3199.10.1161/STROKEAHA.112.668533
228
6.
Al-Shahi Salman R, Hall JM, Horne MA, et al. Untreated clinical course of cerebral cavernous
229
malformations: a prospective, population-based cohort study. Lancet Neurol. 2012; 11: 217-
230
224.10.1016/S1474-4422(12)70004-2
231
7.
232 233
Flemming KD, Link MJ, Christianson TJ, Brown RD, Jr. . The Prospective Hemorrhage risk of intracerebral cavernous malformations. Neurology. 2012; 78: 632-636
8.
Horne MA FK, Su I-C, Stapf C, Jeon JP, Li D, Maxwell SS, White P, Christianson TJ, Agid R, Cho
234
W-S, Oh CW, Wu Z, Zhang J-T, Kim JE, ter Brugge K, Willinsky R, Brown Jr RD, Murray GD, Al-
235
Shahi Salman R, and the Cerebral Cavernous Malformations Individual Patient Data Meta-
236
analysis Collaborators. Clinical course of untreated cerebral cavernous malformations: an
237
individual patient data meta-analysis. Lancet Neurol. 2015; In Press:
238 239
9.
Kondziolka D, Lunsford LD, Kestle JR. The natural history of cerebral cavernous malformations. J Neurosurg. 1995; 83: 820-824
Flemming, page 10 240
10.
241 242
Kupersmith MJ, Kalish H, Epstein F, et al. Natural history of brainstem cavernous malformations. Neurosurgery. 2001; 48: 47-54
11.
Porter PJ, Willinsky RA, Harper W, Wallace MC. Cerebral cavernous malformations: natural
243
history and prognosis after clinical deterioration with or without hemorrhage. J Neurosurg. 1997;
244
87: 190-197
245
12.
246 247
familial occurrence. N Engl J Med. 1988; 319: 343-347 13.
248 249
Rigamonti D, Hadley MN, Drayer BP, et al. Cerebral cavernous malformations. Incidence and
Robinson JR, Awad IA, Little JR. Natural history of the cavernous angioma. J Neurosurg. 1991; 75: 709-714
14.
Kumar S, Lanzino G, Brinjikji W, Hocquard KW, Flemming KD. Infratentorial Developmental
250
Venous Abnormalities and Inflammation Increase Odds of Sporadic Cavernous Malformation. J
251
Stroke Cerebrovasc Dis. 2019.10.1016/j.jstrokecerebrovasdis.2019.02.025
252
15.
Bervini D, Jaeggi C, Mordasini P, Schucht P, Raabe A. Antithrombotic medication and bleeding
253
risk in patients with cerebral cavernous malformations: a cohort study. J Neurosurg. 2018: 1-
254
9.10.3171/2018.1.jns172547
255
16.
256 257
Flemming KD, Link MJ, Christianson TJ, Brown RD, Jr. Use of antithrombotic agents in patients with intracerebral cavernous malformations. J Neurosurg. 2013; 118: 43-46
17.
Zuurbier SM, Hickman CR, Tolias CS, et al. Long-term antithrombotic therapy and risk of
258
intracranial haemorrhage from cerebral cavernous malformations: a population-based cohort
259
study, systematic review, and meta-analysis. Lancet Neurol. 2019; 18: 935-
260
941.https://doi.org/10.1016/S1474-4422(19)30231-5
261
18.
Chohan MO, Marchio S, Morrison LA, et al. Emerging Pharmacologic Targets in Cerebral
262
Cavernous Malformation and Potential Strategies to Alter the Natural History of a Difficult
263
Disease: A Review. JAMA Neurol. 2019; 76: 492-500.10.1001/jamaneurol.2018.3634
264
19.
Gibson CC, Zhu W, Davis CT, et al. Strategy for identifying repurposed drugs for the treatment of
265
cerebral cavernous malformation. Circulation. 2015; 131: 289-
266
299.10.1161/circulationaha.114.010403
Flemming, page 11 267
20.
Polster SP, Stadnik A, Akers AL, et al. Atorvastatin Treatment of Cavernous Angiomas with
268
Symptomatic Hemorrhage Exploratory Proof of Concept (AT CASH EPOC) Trial. Neurosurgery.
269
2018.10.1093/neuros/nyy539
270
21.
Shenkar R, Peiper A, Pardo H, et al. Rho Kinase Inhibition Blunts Lesion Development and
271
Hemorrhage in Murine Models of Aggressive Pdcd10/Ccm3 Disease. Stroke. 2019; 50: 738-
272
744.10.1161/strokeaha.118.024058
273
22.
Flemming KD, Brown RD, Link MJ. Seasonal variation in hemorrhage and focal neurologic deficit
274
due to intracerebral cavernous malformations. J Clin Neurosci. 2015; 22: 969-
275
971.10.1016/j.jocn.2015.01.007
276
23.
Al-Shahi Salman R, Berg MJ, Morrison L, Awad IA, Angioma Alliance Scientific Advisory B.
277
Hemorrhage from cavernous malformations of the brain: definition and reporting standards.
278
Angioma Alliance Scientific Advisory Board. Stroke. 2008; 39: 3222-
279
3230.10.1161/STROKEAHA.108.515544
280
24.
Kernan WN, Ovbiagele B, Black HR, et al. Guidelines for the prevention of stroke in patients with
281
stroke and transient ischemic attack: a guideline for healthcare professionals from the American
282
Heart Association/American Stroke Association. Stroke. 2014; 45: 2160-
283
2236.10.1161/str.0000000000000024
284
25.
Abe M, Kjellberg RN, Adams RD. Clinical presentations of vascular malformations of the brain
285
stem: comparison of angiographically positive and negative types. J Neurol Neurosurg
286
Psychiatry. 1989; 52: 167-175
287
26.
288 289
Aiba T, Tanaka R, Koike T, Kameyama S, Takeda N, Komata T. Natural history of intracranial cavernous malformations. J Neurosurg. 1995; 83: 56-59
27.
Chohan MO, Marchio S, Morrison LA, et al. Emerging Pharmacologic Targets in Cerebral
290
Cavernous Malformation and Potential Strategies to Alter the Natural History of a Difficult
291
Disease: A Review. JAMA Neurol. 2018.10.1001/jamaneurol.2018.3634
292 293 294
28.
Retta SF, Glading AJ. Oxidative stress and inflammation in cerebral cavernous malformation disease pathogenesis: Two sides of the same coin.
. Int J Biochem Cell Biol. 2016; 81: 254-270
Flemming, page 12 295
29.
Taslimi S, Modabbernia A, Amin-Hanjani S, Barker FG, Macdonald RL. Natural history of
296
cavernous malformation: Systematic review and meta-analysis of 25 studies. Neurology. 2016;
297
86: 1984-1991
298
30.
Dammann P, Jabbarli R, Wittek P, et al. Solitary Sporadic Cerebral Cavernous Malformations:
299
Risk Factors of First or Recurrent Symptomatic Hemorrhage and Associated Functional
300
Impairment. World Neurosurg. 2016; 91: 73-80.10.1016/j.wneu.2016.03.080
301
31.
302 303
review of neurotherapeutics. 2015; 15: 1109-1111.10.1586/14737175.2015.1071193 32.
304 305
Gazzaz M, Sichez J, Capelle L, Fohanno D. [Recurrent bleeding of thalamic cavernous angioma under hormonal treatment. A case report]. Neuro-Chirurgie. 1999; 45: 413-416
33.
306 307
Starke RM. Do brainstem cavernous malformations have a higher rate of hemorrhage? Expert
Whitehead KJ, Chan AC, Navankasattusas S, et al. The cerebral cavernous malformation signaling pathway promotes vascular integrity via Rho GTPases. Nat Med. 2009; 15: 177-184
34.
Girard R, Zeineddine HA, Koskimaki J, et al. Plasma Biomarkers of Inflammation and
308
Angiogenesis Predict Cerebral Cavernous Malformation Symptomatic Hemorrhage or Lesional
309
Growth. Circ Res. 2018; 122: 1716-1721.10.1161/CIRCRESAHA.118.312680
310
35.
311
Polster SP, Cao Y, Carroll T, et al. Trial Readiness in Cavernous Angiomas With Symptomatic Hemorrhage (CASH). Neurosurgery. 2019; 84: 954-964.10.1093/neuros/nyy108
312 313
Figures
314
Figure 1: Screening and Study Population
315
Figure 2: a) Modes of presentation in patients with CM b) Onset and time course to peak severity of ICH
316
symptoms in patients with CM
317 318
TABLE 1 - Patient Characteristics in Those With and Without Hemorrhage at Presentation All patients
N
202
Presentation
Presentation –
with ICH
no ICH
75
127
P Value
Odds ratio (95% CI)
Demographic Information Age at Diagnosis
43.7 +/-16.5
42.0 +/- 16.6
44.7+/-16.4
0.27
0.46 (0.12-1.8)
Sex (Female)
118 (58.4%)
42 (56.0%)
76 (59.8%)
0.59
0.8 (0.47-1.52)
Familial form
40 (19.8%)
13 (17.3%)
28 (22.0%)
0.42
0.74 (0.35-1.51)
Multiple lesions
57 (28.2%)
19 (25.3%)
38 (30.0%)
0.48
0.8 (0.42-1.51)
Symptom began
68/125 (54.4%)
38/70 (54.3%)
30/55 (54.5%)
0.96
0.98 (0.47-2.00)
Major infectious
15 (7.6%)
1 (1.3%)
14 (11.3%)
0.011
0.37 (0.014-0.87)
illness
7 missing data
4 missing data
3 missing data
Chronic inflammatory
56 (27.8%)
17 (22.7%)
39 (30.7%)
0.22
0.66 (0.33-1.26)
39 (19.4%)
12 (16.0%)
27 (21.4%)
0.35
0.69 (0.32-1.48)
(years)
between Octoberand March* (n=125) Medical history
disease Neoplastic disease
Medication use at presentation
Any antithrombotic
42 (20.8%)
6 (8.0%)
36 (28.3%)
0.0012
0.21 (0.08-0.55)
Aspirin
37 (18.3%)
5 (6.7%)
32 (25.2%)
0.0022
0.21 (0.07-0.57)
NSAIDs
74 (36.6%)
18 (24.0%)
56 (44.1%)
0.0047
0.40 (0.21-0.74)
Vitamin D
57 (28.2%)
17 (22.7%)
40 (31.5%)
0.18
0.64 (0.33-1.23)
Vitamin E
13 (6.4%)
5 (6.7%)
8 (6.3%)
0.92
1.06 (0.33-3.37)
Fish Oil
38 (18.8%)
12 (16.0%)
26 (20.5%)
0.42
0.74 (0.35-1.57)
Statin
34 (16.8%)
9 (12.0%)
25 (19.7%)
0.16
0.55 (0.24-1.27)
SSRI
21 (10.4%)
6 (6.7%)
16 (12.6%)
0.23
0.49 (0.15-1.32)
48 (23.8%)
20 (26.7%)
28 (22.0%)
0.46
0.46 (0.66-2.49)
18 (8.9%)
7 (9.3%)
11 (8.7%)
0.87
1.08 (0.44-2.93)
Tobacco use past or current Alcohol greater than recommended***
Radiologic Information Brainstem location
62 (30.7%)
44 (58.7%)
19 (15.1%)
<0.0001
8.07 (4.13-15.77)
Associated DVA
61 (38.1%)
23 (38.9%)
38 (37.6%)
0.86
0.86(0.55-2.05)
42 missing data
16 missing
26 missing
NSAID=nonsteroidal anti-inflammatory drug, SSR=serotonin reuptake inhibitor, DVA=developmental venous anomaly. * Denominator includes only those patients with known month of onset ** Only women included in this calculation *** Recommendations as per the American Heart Association for stroke prevention
24
Table 2 Comparison of the Significance of Antithrombotics and NSAIDS in Sporadic vs. Familial Cavernous Malformation Sporadic Form
Any antithrombotic Aspirin NSAIDS Familial Form
Present with ICH N=62 4 (6.4%) 4 (6.4%) 15 (24.2%)
Present with ICH N=13 Any antithrombotic 2 (15.4%) Aspirin 1 (7.7%) NSAIDS 3 (23.1%) NSAIDs=non-steroidal anti-inflammatory drugs
No ICH N=99 30 (30.3%) 26 (26.3%) 44 (44.4%)
P- value
No ICH N=28 6 (21.4%) 6 (21.4%) 12 (42.8%)
P-Value
0.0003 0.0016 0.0095
1.000 0.3986 0.3048
Table 3: Multivariate Analysis of Predictors of Presentation with Hemorrhage
Variable
Odds Ratio
95%
P value
Confidence Interval Brainstem Location
9.49
4.43-20.34
<0.0001
Aspirin
0.19
0.06-0.60
0.0018
NSAIDS
0.60
0.28-1.39
0.1950
Major Infectious Illness
0.07
0.02-0.008
0.0047
ABBREVIATIONS CM
cavernous malformation
CCMs
cerebral cavernous malformations
DVA
Developmental venous anomaly
FND
Focal neurologic deficit
GRE
Gradient recalled echo
ICH
Intracerebral hemorrhage
NSAIDS
Non-steroidal anti-inflammatory drugs
SWI
Susceptibility-weighted imaging