- Email: [email protected]
Hemorrhage, Seizures, and Dynamic Changes of Familial versus Nonfamilial Cavernous Malformation: Systematic Review and Meta-analysis
Literature Review
Hemorrhage, Seizures, and Dynamic Changes of Familial versus Nonfamilial Cavernous Malformation: Systematic Review and Meta-analysis Shervin Taslimi1, Jerry C. Ku1, Amirhossein Modabbernia3, R. Loch Macdonald1,2
Key words Cavernous malformation - Dynamic changes - Familial - Hemorrhage - Natural history - Seizure -
Abbreviations and Acronyms CCM: Cerebral cavernous malformation FEM: Fixed-effects model LY: Lesion-year PY: Person-year From the 1Division of Neurosurgery, Department of Surgery, University of Toronto, Toronto, Ontario, Canada; 2Division of Neurosurgery, Labatt Family Centre of Excellence in Brain Injury and Trauma Research, Keenan Research Centre for Biomedical Science, and Li Ka Shing Knowledge Institute, St. Michael’s Hospital, Toronto, Ontario, Canada; and 3 Department of Psychiatry, Icahn School of Medicine at Mount Sinai Hospital, New York, New York, USA To whom correspondence should be addressed: R. Loch Macdonald, M.D., Ph.D. [E-mail: [email protected]] Supplementary digital content available online. Citation: World Neurosurg. (2019) 126:241-246. https://doi.org/10.1016/j.wneu.2019.02.115 Journal homepage: www.journals.elsevier.com/worldneurosurgery Available online: www.sciencedirect.com 1878-8750/$ - see front matter ª 2019 Elsevier Inc. All rights reserved.
INTRODUCTION Cavernous malformations consist of abnormal blood vessels with a single layer of endothelium that do not exhibit intervening tight junctions and lack any intervening normal brain or spinal cord tissue.1-4 Cerebral cavernous malformations (CCMs) can be found incidentally or owing to various manifestations, such as nonspecific neurologic findings, focal neurologic deficits based on lesion location, seizures, and transient neurologic symptoms. They can be multiple or single.5 Most patients with multiple lesions have familial CCMs, with genetic mutation in germline cells.6 Familial CCMs are defined as the occurrence of CCMs in at least 2 family members, the presence of multiple CCMs,
- BACKGROUND:
Cerebral cavernous malformations (CCMs) may be familial or nonfamilial. This systematic review compared the natural history of CCMs in familial compared with nonfamilial cases.
- METHODS:
We searched MEDLINE, Web of Science, and EMBASE for natural history studies on CCMs up to September 2018. We included studies that followed at least 20 untreated patients. Primary outcomes were hemorrhage, seizures, and neuroimaging changes in familial and nonfamilial cases. Incidence rate per person-year (PY) or lesion-year (LY) of follow-up were used to pool the data using fixed-effects or random-effects models. We used the incidence rate ratio for comparison.
- RESULTS:
We could not compare hemorrhage rates between familial and nonfamilial cases mainly owing to mixtures of subgroups of patients. The seizure rate was similar in familial and nonfamilial cases with pooled incidence rate of 1.5%/PY (95% confidence interval 1.1%e2.2%). The reseizure rate was higher than the seizure rate (P < 0.001). New lesion development was higher in familial cases (32.1%/PY vs. 0.7%/PY, P < 0.001). Signal change on neuroimaging ranged from 0.2%/LY to 2.4%/LY in familial cases. In familial cases, incidence rate of size change was 8%/PY (95% confidence interval 5.2%e12.2%) and 1.1%/LY (95% confidence interval 0.6%e1.6%).
- CONCLUSIONS:
Familial CCMs show higher dynamic changes than nonfamilial cases. However, the presence of actual dynamic changes needs further assessment in nonfamilial cases. CCMs demonstrate a low incidence of seizure. First-time seizure increases the chance of recurrent seizure. Seizure rate based on the location and type of the lesion should be investigated further.
or the presence of a disease-causing mutation in 1 of the 3 genes in which mutations are known to cause familial CCM (CCM1, CCM2, and CCM3).6,7
cases and compare these characteristics with nonfamilial cases.
In familial cases, 4 types of lesions have been identified based on magnetic resonance imaging characteristics.8 Type 1 lesions are lesions with a subacute bleeding pattern. Type 2 lesions demonstrate mixed signal indicating hemorrhages and thrombosis of varying ages. Type 3 lesions are lesions with chronic hemorrhage. Type 4 lesions are detectable only on gradient echo sequences. The purpose of this systematic review and meta-analysis was to pool published data regarding the natural history of CCMs, including hemorrhage rate, seizure rate, and neuroimaging changes, in familial
MATERIALS AND METHODS
WORLD NEUROSURGERY 126: 241-246, JUNE 2019
This study was performed in accordance with Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines.9 Search strategy and method of the study have been previously published.10 In brief, MEDLINE, Web of Science, and EMBASE were searched with appropriate key words and methodology for natural history studies up to September 2018.11,12 Two reviewers (S.T. and J.C.K.) scanned the titles and abstracts of retrieved articles initially to identify potentially relevant articles. Any disagreements were resolved by discussion or involving the senior
www.journals.elsevier.com/world-neurosurgery
241
LITERATURE REVIEW SHERVIN TASLIMI ET AL.
FAMILIAL VERSUS NONFAMILIAL CAVERNOUS MALFORMATION
Table 1. Reported Hemorrhage Rates in Familial Cases
Hemorrhage Rate
Reference Labauge et al., 200014 13
Familial Series (Number of Patients/ Asymptomatic Patients at Beginning of Study)
Type of Study
Yes (40/5)
Prospective
Mixture of Hemorrhage and Rehemorrhage Rate
Symptomatic Multiple Supratentorial Symptomatic Bothy (Asymptomatic (Fraction of Lesions Lesions Bothy (Fraction of Brainstem) Fraction) Brainstem) 92%
75%
—
—
11% (14/21)
3.6%z (5/7)
Labauge et al., 2001
Yes (33/33)
Prospective
85%
85%
4.3%
1.43%z (1/1)
—
—
Zabramski et al., 19948
Yes (31/—)
Prospective
76%
94%
—
—
—
6.5% (0/3)
Moriarity et al., 199915
No*
Prospective
—
—
—
—
—
1.8% (—/—)
*Incidence rate of 13 familial cases among study population is also presented separately. yBoth symptomatic and asymptomatic mixed together. zThese numbers were manually calculated based on the information given in the article.
author (R.L.M.). Two authors (S.T. and J.C.K.) reviewed the full text of included articles to evaluate the inclusion and exclusion criteria and to extract predefined outcome measures. We included studies that followed, including presurgical follow-up, at least 20 untreated patients. Familial and nonfamilial cases should have been defined based on the history of a CCM in family members or genetic testing. Inception point and end of follow-up should have been explicitly mentioned in studies. For calculating dynamic changes, studies should have had prespecified imaging follow-up. Studies that did not report incidence rate and studies in which this could not be calculated based on the available data were excluded (we did not use odds ratio or relative risk). We contacted corresponding authors of the articles if any relevant information was missing. We also evaluated the risk of bias with our predefined variables in each article. Our predefined outcome measures were comparison of the following measures in familial and nonfamilial cases: 1) hemorrhage and rehemorrhage rate (per personyear [PY] or lesion-year [LY]), 2) seizure or reseizure rate per PY or LY of follow-up, 3) rate of new lesion development, 4) incidence of signal change in the lesions (PY or LY), and 5) incidence of size change in the lesions (PY or LY). We predefined some factors for investigating heterogeneity among included articles: 1) seizure vs. reseizure rate, 2) seizure based on different locations, and 3) new lesion development depending on the type of lesions.
242
www.SCIENCEDIRECT.com
Hemorrhage was defined as any radiologic evidence of new hemorrhage, seen as evidence of acute or subacute blood products on magnetic resonance imaging using T1-weighted or T2-weighted imaging or gradient echo sequences. We divided hemorrhage further into symptomatic or asymptomatic hemorrhage. For seizurerelated outcomes, we did not restrict our data extraction to any particular seizure type; all types of seizures were analyzed together. Reseizure was defined as the development of additional seizures after a first seizure, that is, the likelihood of progressing to develop epilepsy. None of the included studies mentioned the use of antiepileptic medications in included patients. Regarding changes in the neuroimaging characteristics of CCMs, a boardcertified radiologist must have reviewed the images to identify any new lesions, size changes, or signal changes in CCMs. Statistical Analysis We used the incidence rate as opposed to odds ratio or relative risk to pool the data to avoid unbalanced follow-up bias arising from the inclusion of odds ratio or relative risk. For comparisons, we used the incidence rate ratio. After logarithmic transformation of the incidence rate, we pooled the data either with fixed-effects model (FEM) or random-effects model, and then the point estimates were transferred back to the normal scale. We investigated heterogeneity using c2 test and calculating I2 statistic. We interpreted P value < 0.1 in c2 test or I2 >50% as heterogeneous. We used FEM
for homogeneous data and random-effects model for heterogeneous data. We performed all analyses using STATA 11 (StataCorp LLC, College Station, Texas, USA). RESULTS Literature review yielded 3960 deduplicated results, and full-text screening was performed for 166 articles. Eight articles were included in the final systematic review and meta-analysis. A flow chart of the study is available in the supplementary materials. Hemorrhage Rate In our previous meta-analysis, we showed that the symptomatic rehemorrhage rate was higher than hemorrhage rate and that symptomatic hemorrhage and rehemorrhage rates were higher in brainstem lesions.10 In this analysis, it was not possible to compare familial hemorrhage and rehemorrhage rates with nonfamilial cases for the following reasons: 1) familial series included asymptomatic hemorrhage or rehemorrhage in their calculations, whereas nonfamilial cases did not; 2) hemorrhage and rehemorrhage rates were mostly mixed in familial series; and 3) we could not find symptomatic hemorrhage or rehemorrhage rates separately for brainstem and nonbrainstem lesions in familial series. In familial cases, any hemorrhage rate was found to be 4.3%/PY (either asymptomatic or symptomatic), and symptomatic hemorrhage was reported to be 1.43%/PY.13 In studies of
WORLD NEUROSURGERY, https://doi.org/10.1016/j.wneu.2019.02.115
LITERATURE REVIEW SHERVIN TASLIMI ET AL.
FAMILIAL VERSUS NONFAMILIAL CAVERNOUS MALFORMATION
familial cases that reported mixed hemorrhage and rehemorrhage rates, both asymptomatic and symptomatic rate was 11%/PY, whereas symptomatic rate ranged from 1.8%/PY to 6.5%/PY (Table 1).8,14,15 Seizure and Epilepsy Rate Table 2 shows the reported seizure and reseizure incidence rates in familial and nonfamilial cases in included studies. The annual seizure rate was 1.5% in nonfamilial cases.15-18 There was no difference in first-time seizure rate between familial and nonfamilial cases (4 nonfamilial studies15-18 vs. 1 familial study,13 P ¼ 0.9). Combining all 5 studies with FEM, first-time annual seizure rate was 1.5%/PY (Table 3). Annual reseizure rate was higher than seizure rate (2 reseizure rates15,17 vs. 5 seizure rates,13,15-18 P < 0.001). Exclusion of 1 study with an uncharacteristically high reseizure rate17 did not change the result (P < 0.001). Data were insufficient to compare the reseizure rate between familial and nonfamilial cases. Development of New Lesions New lesion development in familial cases was 32.1%/PY based on 3 articles (Table 4).8,13,14 Comparing new lesion development between familial versus nonfamilial cases, we found only 1 study that clearly differentiated familial and nonfamilial cases.15,19 New lesion development was higher in familial cases (32.1%/PY [3 studies, 80 patients] vs. nonfamilial cases (0.7%/PY [1 study, 13 patients], P < 0.001). Two studies reported the development of new lesions
in a general population of patients without specifying whether these cases were familial or nonfamilial CCMs.20,21 We could not extract the follow-up period in these studies. Therefore, we did not enter them into meta-analysis.
could not find any reports meeting our study inclusion criteria that addressed the change in the size of lesions in nonfamilial cases.
DISCUSSION Neuroimaging Characteristics Signal Change. We excluded apparent hemorrhage as a cause of signal change in our meta-analysis. LY of follow-up was used, as PY of follow-up would not be accurate in the case of multiple lesions. In familial series, in the study in which all patients were asymptomatic at diagnosis,13 the rate of change in signal was much lower (0.2%/LY) than our calculated incidence rate in 2 other familial studies that included mixtures of symptomatic and asymptomatic patients (1.7%/LY14 and 2.4%/LY8). Owing to the heterogeneity of the data and low number of studies, we did not combine the data. We could not find any report regarding signal change in nonfamilial cases. Size Change. Three studies of familial cases reported the incidence rate of increase in size of lesions or change in size of lesions based on LY of follow-up.8,13,14 We pooled these data using FEM. Details of the point estimate are presented in Table 4. In 1 study, 4 out of 5 lesion changes were due to hemorrhage.8 In the rest of the articles, it could not be ascertained whether these changes were due to hemorrhage (increase in size owing to hemorrhage or regression after previous hemorrhage) or were true dynamic changes of the lesions. We
In this study, we pooled reports of the natural history of familial CCMs compared with nonfamilial CCMs and addressed some sources of heterogeneity among the results. Natural history studies in familial cases initially showed higher overall rates of hemorrhage compared with nonfamilial series.8,13,14 We attribute part of this difference to the inclusion of asymptomatic hemorrhage in familial studies, whereas other studies defined the hemorrhagic event as the symptomatic event with evidence of bleeding on the imaging or pathology result at the time of the surgery. A mixture of hemorrhage and rehemorrhage made the comparison between familial and nonfamilial cases challenging. As shown before, brainstem and previously bled cavernomas are more likely to have symptomatic hemorrhages10; simply pooling hemorrhage rates across studies that do not present these numbers separately would be inaccurate. Compared with the familial hemorrhage rates reported in Table 1, in our previous meta-analysis of CCMs, the symptomatic hemorrhage rate was 2.8%/PY in brainstem lesions.10 Similarly, Horne et al.22 in their individual patient data meta-analysis found that the 5-year hemorrhage risk was 8.0% in brainstem lesions without previous hemorrhage or focal neurologic deficit. These numbers are very close to
Table 2. Reported Seizure Rates in Familial and Nonfamilial Cases Familial Series
Type of Study
Multiple Lesions
Percentage of Supratentorial Lesions
Seizure Rate (Number of Patients, Events)
Del Curling et al., 199116
No
Retrospective
19%
86%
1.51% (32, 16)
—
—
Josephson et al., 201117
No
Prospective
24%
77%
1.47% (38, 2)
94%
—
Moriarity et al., 199915
No
Prospective
37%
81%
2.4% (35, 4)
5.5%
4.8%
Reference
18
Reseizure Rate
Mixture of Seizure and Reseizure
Kondziolka et al., 1995
No
Prospective
20%
60%*
1.5% (94, 4)
—
—
Labauge et al., 200113
Yes
Prospective
85%
86%
1.42% (33, 1)
—
—
Zabramski et al., 19948
Yes
Prospective
76%
95%
—
—
6.49%
*This is an approximation because cerebral and cerebellar lesions are combined.
WORLD NEUROSURGERY 126: 241-246, JUNE 2019
www.journals.elsevier.com/world-neurosurgery
243
LITERATURE REVIEW SHERVIN TASLIMI ET AL.
FAMILIAL VERSUS NONFAMILIAL CAVERNOUS MALFORMATION
Table 3. Synthesis of Pooled Seizure and Reseizure Rates in Familial and Nonfamilial Cases Number of Studies (Number of Subjects)
Seizure Rate Reseizure versus seizure (incidence rate difference)
Model P Value for Overall Pooled Used Heterogeneity (I2) Estimate (95% CI) P Value References
5 (seizure, 289, reseizure, 33)
—
—
3.7% (2.9%e4.54%) <0.001
13,15e18
Seizure rate in nonfamilial series
4 (218)
FEM
0.73
1.5% (1.1%e2.2%)
<0.001
13,15,17,18
Seizure rate in familial and nonfamilial cases combined
5 (251)
FEM
0.86
1.5% (1.1%e2.2%)
<0.001
13,15e18
CI, confidence interval; FEM, fixed-effects model.
annual the 1.43% hemorrhage rate in familial brainstem cavernomas. Another important potential factor for hemorrhage may be the type of lesion, which can change over time.8,20 For instance, type 1 CCM lesions can become type 2 lesions if hemorrhage and thrombosis occur multiple times. Type 3 lesions, which show remote bleeding, can change into type 1 or 2 lesions if they rebleed. Type 1 or 2 lesions can change into type 3 lesions if they do not hemorrhage for some time. As previously shown, the rehemorrhage rate is higher during the first 2 years after bleeding.10 Taken together, type 3 lesions seem to be less active compared with type 1 and 2 lesions and to be associated with less hemorrhage. A sequential decrease in incidence of hemorrhage from type 1 to type 3 has been shown by Labauge et al.14 Nikoubashman et al.23 found the annual hemorrhage rate to be 29.8% for Zabramski type 1, 20.1% for type 2, 3.4% for type 3, and 1.3% for type 4 CCMs. The hemorrhage rate was 23.1% for their newly proposed type 5 lesions, which were CCMs that showed gross extralesional hemorrhage.23 In their
article, Nikoubashman et al.23 did not report hemorrhage and rehemorrhage rates separately; therefore, these were not added into our meta-analysis. The relationship of the type of lesion and the hemorrhage rate should be further assessed in future studies. Based on the present data, it is hard to draw a rigorous conclusion regarding the difference in hemorrhage rate between familial and nonfamilial cases stratified by the type of lesion. In future studies reporting natural history, we suggest stratifying for hemorrhage status (hemorrhage vs. rehemorrhage), timing after the hemorrhage, and type of lesions. This would help to reduce possible effect modification in comparison of familial and nonfamilial cases. Overall, hemorrhage rate may be higher in familial cases because of the presence of multiple lesions, as higher risk lesions bleed sooner and increase the overall incidence rate. Based on the available data, no distinction has been made between different locations and the incidence rate of seizure. This can potentially add to the heterogeneity of the data. Because lesions in different locations may confer different
incidence rates of seizure, seizure and reseizure incidence rates should be reported based on the location of the lesions. Josephson et al.17 did not find an association between the anatomic location of CCMs and the occurrence of seizures but did note that cavernomas were more likely to be multiple in adults with seizures. Moreover, the type of lesion may be an important factor for seizure development, such as an acute hemorrhage in the lesion causing a seizure. Given the different possible propensities for hemorrhage in the different types of lesions, lesion type may also be an important factor in predicting the seizure rate. This should be differentiated with other processes that can also lead to seizures, such as hemosiderin ring or gliosis around the CCM. CCMs appear to show dynamic changes in familial cases. We found new lesion development in familial cases to be 32.1%/ PY.8,13,14 Gross et al.24 found new lesion development in patients with multiple cavernomas on presentation to be 3.0%/ PY compared with 0.4% in patients with solitary nonfamilial cavernomas. Notably,
Table 4. Dynamic Changes in Familial Cases Number of Studies (Number of Subjects)
Model Used
P Value for Heterogeneity (I2)
Overall Pooled Estimate (95% CI)
P Value
References
Development of new lesions (PY) in familial cases
3 (94)
REM
0.049 (69%)
32.1% (20%e51.2%)
<0.001
13,14,19
Change in size (LY)
3 (94)
FEM
0.6
1.1% (0.6%e1.6%)
<0.001
13,14,19
Change in size (PY)
3 (94)
FEM
0.5
8% (5.2%e12.2%)
<0.001
13,14,19
Increase in size (LY)
3 (94)
FEM
0.640
0.8% (0.4%e1.3%)
<0.001
13,14,19
Increase in size (PY)
3 (94)
FEM
0.63
5.9% (5.9%e9.9%)
<0.001
13,14,19
Incidence Rate
CI, confidence interval; PY, person-year; REM, random-effects model; LY, lesion-year; FEM, fixed-effects model.
244
www.SCIENCEDIRECT.com
WORLD NEUROSURGERY, https://doi.org/10.1016/j.wneu.2019.02.115
LITERATURE REVIEW SHERVIN TASLIMI ET AL.
multiplicity of cavernomas does not equate with familial cavernoma, and the definition by Gross et al.24 of a new lesion to be at least 4 mm is different from previous studies; therefore, direct comparison and addition of these data into our metaanalysis were not possible. In addition, several nonfamilial studies have shown some dynamic changes as well. Kupersmith et al.25 evaluated nonfamilial brainstem cases (familial cases assessed only with obtaining a family history of CCM) and observed some size changes in the lesions. These data were not included in our study because they did not meet our criteria for data extraction. Some other series also reported dynamic changes in patients with CCMs, but we could not find any criteria in these studies to differentiate familial from nonfamilial cases.16,20,21,26-28 However, it cannot be ascertained whether the cases that showed dynamic changes were nonfamilial or if these changes were related to bleeding episodes. It may be important to evaluate changes in lesions that are not due to hemorrhage by examining size and signal changes as well as the development of new lesions. So far, no studies have done this. Taken together, even though the presence of dynamic changes is obvious in familial cases, it is hard to draw any conclusion regarding dynamic changes in nonfamilial cases because most series pooled familial and nonfamilial cases. The current study has several limitations. We were not able to complete some subgroup analyses owing to the limited number of studies. For some of our outcome measures, direct statistical comparisons between familial and nonfamilial cases were not possible mainly because of the mixture of the patients in different subgroups or inadequate reporting of data. Based on the recent definition of familial CCM, individuals with multiple lesions are considered to have familial CCMs. In our included articles, none of the nonfamilial series considered this criterion. This may have affected our results. CONCLUSIONS Familial CCMs exhibit dynamic changes, including change in the size of lesions, development of new lesions, change in the signal of lesions, and change in the type of
FAMILIAL VERSUS NONFAMILIAL CAVERNOUS MALFORMATION
lesions. Hemorrhage can explain only part of these changes; therefore, other processes apparently are involved in these dynamic changes. The risk of hemorrhage in familial cases may be higher than in nonfamilial cases because of the presence of multiple lesions, but further study is needed to evaluate this. Potential sources of heterogeneity in estimating the incidence rate of hemorrhage, rehemorrhage, and seizures in these lesions are the type of presentation, location, and type of lesions. Improved reporting and additional studies regarding the clinical behavior and natural history of familial and nonfamilial cavernomas are required to determine best management. REFERENCES 1. Tomlinson FH, Houser OW, Scheithauer BW, Sundt TM Jr, Okazaki H, Parisi JE. Angiographically occult vascular malformations: a correlative study of features on magnetic resonance imaging and histological examination. Neurosurgery. 1994; 34:792-799 [discussion 799-800]. 2. Raychaudhuri R, Batjer HH, Awad IA. Intracranial cavernous angioma: a practical review of clinical and biological aspects. Surg Neurol. 2005;63: 319-328 [discussion: 328]. 3. Maraire JN, Awad IA. Intracranial cavernous malformations: lesion behavior and management strategies. Neurosurgery. 1995;37:591-605. 4. Frischer JM, Pipp I, Stavrou I, Trattnig S, Hainfellner JA, Knosp E. Cerebral cavernous malformations: congruency of histopathological features with the current clinical definition. J Neurol Neurosurg Psychiatry. 2008;79:783-788. 5. Washington CW, McCoy KE, Zipfel GJ. Update on the natural history of cavernous malformations and factors predicting aggressive clinical presentation. Neurosurg Focus. 2010;29:E7. 6. Labauge P, Denier C, Bergametti F, TournierLasserve E. Genetics of cavernous angiomas. Lancet Neurol. 2007;6:237-244. 7. Riant F, Bergametti F, Ayrignac X, Boulday G, Tournier-Lasserve E. Recent insights into cerebral cavernous malformations: the molecular genetics of CCM. FEBS J. 2010;277:1070-1075.
meta-analysis of 25 studies. Neurology. 2016;86: 1984-1991. 11. Wilczynski NL, Haynes RB. Developing optimal search strategies for detecting clinically sound prognostic studies in MEDLINE: an analytic survey. BMC Med. 2004;2:23. 12. Wilczynski NL, Haynes RB. Optimal search strategies for detecting clinically sound prognostic studies in EMBASE: an analytic survey. J Am Med Inform Assoc. 2005;12:481-485. 13. Labauge P, Brunereau L, Laberge S, Houtteville JP. Prospective follow-up of 33 asymptomatic patients with familial cerebral cavernous malformations. Neurology. 2001;57:1825-1828. 14. Labauge P, Brunereau L, Levy C, Laberge S, Houtteville JP. The natural history of familial cerebral cavernomas: a retrospective MRI study of 40 patients. Neuroradiology. 2000;42:327-332. 15. Moriarity JL, Wetzel M, Clatterbuck RE, et al. The natural history of cavernous malformations: a prospective study of 68 patients. Neurosurgery. 1999;44:1166-1173. 16. Del Curling O Jr, Kelly DL Jr, Elster AD, Craven TE. An analysis of the natural history of cavernous angiomas. J Neurosurg. 1991;75:702-708. 17. Josephson CB, Leach JP, Duncan R, et al. Seizure risk from cavernous or arteriovenous malformations: prospective population-based study. Neurology. 2011;76:1548-1554. 18. Kondziolka D, Lunsford LD, Kestle JR. The natural history of cerebral cavernous malformations. J Neurosurg. 1995;83:820-824. 19. Clatterbuck RE, Moriarity JL, Elmaci I, Lee RR, Breiter SN, Rigamonti D. Dynamic nature of cavernous malformations: a prospective magnetic resonance imaging study with volumetric analysis. J Neurosurg. 2000;93:981-986. 20. Willinsky R, Harper W, Wallace MC, et al. Followup MR of intracranial cavernomas: the relationship between haemorrhagic events and morphology. Interv Neuroradiol. 1996;2:127-135. 21. Kim DS, Park YG, Choi JU, Chung SS, Lee KC. An analysis of the natural history of cavernous malformations. Surg Neurol. 1997;48:9-18. 22. Horne MA, Flemming KD, Su IC, et al. Clinical course of untreated cerebral cavernous malformations: a meta-analysis of individual patient data. Lancet Neurol. 2016;15:166-173.
8. Zabramski JM, Wascher TM, Spetzler RF, et al. The natural history of familial cavernous malformations: results of an ongoing study. J Neurosurg. 1994;80:422-432.
23. Nikoubashman O, Di Rocco F, Davagnanam I, Mankad K, Zerah M, Wiesmann M. Prospective hemorrhage rates of cerebral cavernous malformations in children and adolescents based on MRI appearance. AJNR Am J Neuroradiol. 2015;36: 2177-2183.
9. Moher D, Liberati A, Tetzlaff J, Altman DG, PRISMA Group. Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. PLoS Med. 2009;6:e1000097.
24. Gross BA, Du R, Orbach DB, Scott RM, Smith ER. The natural history of cerebral cavernous malformations in children. J Neurosurg Pediatr. 2016;17: 123-128.
10. Taslimi S, Modabbernia A, Amin-Hanjani S, Barker FG II, Macdonald RL. Natural history of cavernous malformation: systematic review and
25. Kupersmith MJ, Kalish H, Epstein F, et al. Natural history of brainstem cavernous malformations. Neurosurgery. 2001;48:47-53 [discussion: 53-54].
WORLD NEUROSURGERY 126: 241-246, JUNE 2019
www.journals.elsevier.com/world-neurosurgery
245
LITERATURE REVIEW SHERVIN TASLIMI ET AL.
26. Robinson JR, Awad IA, Little JR. Natural-history of the cavernous angioma. J Neurosurg. 1991;75:709-714. 27. Li D, Yang Y, Hao SY, et al. Hemorrhage risk, surgical management, and functional outcome of brainstem cavernous malformations. J Neurosurg. 2013;119:996-1008. 28. Pozzati E, Giuliani G, Nuzzo G, Poppi M, Piepgras DG, Hunt Batjer H. The growth of
246
www.SCIENCEDIRECT.com
FAMILIAL VERSUS NONFAMILIAL CAVERNOUS MALFORMATION
cerebral cavernous angiomas. Neurosurgery. 1989; 25:92-97.
Conflict of interest statement: R. L. Macdonald receives grants from the Brain Aneurysm Foundation and Genome Canada. The remaining authors have no conflicts to report. Received 17 August 2016; accepted 10 February 2019
Citation: World Neurosurg. (2019) 126:241-246. https://doi.org/10.1016/j.wneu.2019.02.115 Journal homepage: www.journals.elsevier.com/worldneurosurgery Available online: www.sciencedirect.com 1878-8750/$ - see front matter ª 2019 Elsevier Inc. All rights reserved.
WORLD NEUROSURGERY, https://doi.org/10.1016/j.wneu.2019.02.115
Hemorrhage, Seizures, and Dynamic Changes of Familial versus Nonfamilial Cavernous Malformation: Systematic Review and Meta-analysis Shervin Taslimi1, Jerry C. Ku1, Amirhossein Modabbernia3, R. Loch Macdonald1,2
Key words Cavernous malformation - Dynamic changes - Familial - Hemorrhage - Natural history - Seizure -
Abbreviations and Acronyms CCM: Cerebral cavernous malformation FEM: Fixed-effects model LY: Lesion-year PY: Person-year From the 1Division of Neurosurgery, Department of Surgery, University of Toronto, Toronto, Ontario, Canada; 2Division of Neurosurgery, Labatt Family Centre of Excellence in Brain Injury and Trauma Research, Keenan Research Centre for Biomedical Science, and Li Ka Shing Knowledge Institute, St. Michael’s Hospital, Toronto, Ontario, Canada; and 3 Department of Psychiatry, Icahn School of Medicine at Mount Sinai Hospital, New York, New York, USA To whom correspondence should be addressed: R. Loch Macdonald, M.D., Ph.D. [E-mail: [email protected]] Supplementary digital content available online. Citation: World Neurosurg. (2019) 126:241-246. https://doi.org/10.1016/j.wneu.2019.02.115 Journal homepage: www.journals.elsevier.com/worldneurosurgery Available online: www.sciencedirect.com 1878-8750/$ - see front matter ª 2019 Elsevier Inc. All rights reserved.
INTRODUCTION Cavernous malformations consist of abnormal blood vessels with a single layer of endothelium that do not exhibit intervening tight junctions and lack any intervening normal brain or spinal cord tissue.1-4 Cerebral cavernous malformations (CCMs) can be found incidentally or owing to various manifestations, such as nonspecific neurologic findings, focal neurologic deficits based on lesion location, seizures, and transient neurologic symptoms. They can be multiple or single.5 Most patients with multiple lesions have familial CCMs, with genetic mutation in germline cells.6 Familial CCMs are defined as the occurrence of CCMs in at least 2 family members, the presence of multiple CCMs,
- BACKGROUND:
Cerebral cavernous malformations (CCMs) may be familial or nonfamilial. This systematic review compared the natural history of CCMs in familial compared with nonfamilial cases.
- METHODS:
We searched MEDLINE, Web of Science, and EMBASE for natural history studies on CCMs up to September 2018. We included studies that followed at least 20 untreated patients. Primary outcomes were hemorrhage, seizures, and neuroimaging changes in familial and nonfamilial cases. Incidence rate per person-year (PY) or lesion-year (LY) of follow-up were used to pool the data using fixed-effects or random-effects models. We used the incidence rate ratio for comparison.
- RESULTS:
We could not compare hemorrhage rates between familial and nonfamilial cases mainly owing to mixtures of subgroups of patients. The seizure rate was similar in familial and nonfamilial cases with pooled incidence rate of 1.5%/PY (95% confidence interval 1.1%e2.2%). The reseizure rate was higher than the seizure rate (P < 0.001). New lesion development was higher in familial cases (32.1%/PY vs. 0.7%/PY, P < 0.001). Signal change on neuroimaging ranged from 0.2%/LY to 2.4%/LY in familial cases. In familial cases, incidence rate of size change was 8%/PY (95% confidence interval 5.2%e12.2%) and 1.1%/LY (95% confidence interval 0.6%e1.6%).
- CONCLUSIONS:
Familial CCMs show higher dynamic changes than nonfamilial cases. However, the presence of actual dynamic changes needs further assessment in nonfamilial cases. CCMs demonstrate a low incidence of seizure. First-time seizure increases the chance of recurrent seizure. Seizure rate based on the location and type of the lesion should be investigated further.
or the presence of a disease-causing mutation in 1 of the 3 genes in which mutations are known to cause familial CCM (CCM1, CCM2, and CCM3).6,7
cases and compare these characteristics with nonfamilial cases.
In familial cases, 4 types of lesions have been identified based on magnetic resonance imaging characteristics.8 Type 1 lesions are lesions with a subacute bleeding pattern. Type 2 lesions demonstrate mixed signal indicating hemorrhages and thrombosis of varying ages. Type 3 lesions are lesions with chronic hemorrhage. Type 4 lesions are detectable only on gradient echo sequences. The purpose of this systematic review and meta-analysis was to pool published data regarding the natural history of CCMs, including hemorrhage rate, seizure rate, and neuroimaging changes, in familial
MATERIALS AND METHODS
WORLD NEUROSURGERY 126: 241-246, JUNE 2019
This study was performed in accordance with Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines.9 Search strategy and method of the study have been previously published.10 In brief, MEDLINE, Web of Science, and EMBASE were searched with appropriate key words and methodology for natural history studies up to September 2018.11,12 Two reviewers (S.T. and J.C.K.) scanned the titles and abstracts of retrieved articles initially to identify potentially relevant articles. Any disagreements were resolved by discussion or involving the senior
www.journals.elsevier.com/world-neurosurgery
241
LITERATURE REVIEW SHERVIN TASLIMI ET AL.
FAMILIAL VERSUS NONFAMILIAL CAVERNOUS MALFORMATION
Table 1. Reported Hemorrhage Rates in Familial Cases
Hemorrhage Rate
Reference Labauge et al., 200014 13
Familial Series (Number of Patients/ Asymptomatic Patients at Beginning of Study)
Type of Study
Yes (40/5)
Prospective
Mixture of Hemorrhage and Rehemorrhage Rate
Symptomatic Multiple Supratentorial Symptomatic Bothy (Asymptomatic (Fraction of Lesions Lesions Bothy (Fraction of Brainstem) Fraction) Brainstem) 92%
75%
—
—
11% (14/21)
3.6%z (5/7)
Labauge et al., 2001
Yes (33/33)
Prospective
85%
85%
4.3%
1.43%z (1/1)
—
—
Zabramski et al., 19948
Yes (31/—)
Prospective
76%
94%
—
—
—
6.5% (0/3)
Moriarity et al., 199915
No*
Prospective
—
—
—
—
—
1.8% (—/—)
*Incidence rate of 13 familial cases among study population is also presented separately. yBoth symptomatic and asymptomatic mixed together. zThese numbers were manually calculated based on the information given in the article.
author (R.L.M.). Two authors (S.T. and J.C.K.) reviewed the full text of included articles to evaluate the inclusion and exclusion criteria and to extract predefined outcome measures. We included studies that followed, including presurgical follow-up, at least 20 untreated patients. Familial and nonfamilial cases should have been defined based on the history of a CCM in family members or genetic testing. Inception point and end of follow-up should have been explicitly mentioned in studies. For calculating dynamic changes, studies should have had prespecified imaging follow-up. Studies that did not report incidence rate and studies in which this could not be calculated based on the available data were excluded (we did not use odds ratio or relative risk). We contacted corresponding authors of the articles if any relevant information was missing. We also evaluated the risk of bias with our predefined variables in each article. Our predefined outcome measures were comparison of the following measures in familial and nonfamilial cases: 1) hemorrhage and rehemorrhage rate (per personyear [PY] or lesion-year [LY]), 2) seizure or reseizure rate per PY or LY of follow-up, 3) rate of new lesion development, 4) incidence of signal change in the lesions (PY or LY), and 5) incidence of size change in the lesions (PY or LY). We predefined some factors for investigating heterogeneity among included articles: 1) seizure vs. reseizure rate, 2) seizure based on different locations, and 3) new lesion development depending on the type of lesions.
242
www.SCIENCEDIRECT.com
Hemorrhage was defined as any radiologic evidence of new hemorrhage, seen as evidence of acute or subacute blood products on magnetic resonance imaging using T1-weighted or T2-weighted imaging or gradient echo sequences. We divided hemorrhage further into symptomatic or asymptomatic hemorrhage. For seizurerelated outcomes, we did not restrict our data extraction to any particular seizure type; all types of seizures were analyzed together. Reseizure was defined as the development of additional seizures after a first seizure, that is, the likelihood of progressing to develop epilepsy. None of the included studies mentioned the use of antiepileptic medications in included patients. Regarding changes in the neuroimaging characteristics of CCMs, a boardcertified radiologist must have reviewed the images to identify any new lesions, size changes, or signal changes in CCMs. Statistical Analysis We used the incidence rate as opposed to odds ratio or relative risk to pool the data to avoid unbalanced follow-up bias arising from the inclusion of odds ratio or relative risk. For comparisons, we used the incidence rate ratio. After logarithmic transformation of the incidence rate, we pooled the data either with fixed-effects model (FEM) or random-effects model, and then the point estimates were transferred back to the normal scale. We investigated heterogeneity using c2 test and calculating I2 statistic. We interpreted P value < 0.1 in c2 test or I2 >50% as heterogeneous. We used FEM
for homogeneous data and random-effects model for heterogeneous data. We performed all analyses using STATA 11 (StataCorp LLC, College Station, Texas, USA). RESULTS Literature review yielded 3960 deduplicated results, and full-text screening was performed for 166 articles. Eight articles were included in the final systematic review and meta-analysis. A flow chart of the study is available in the supplementary materials. Hemorrhage Rate In our previous meta-analysis, we showed that the symptomatic rehemorrhage rate was higher than hemorrhage rate and that symptomatic hemorrhage and rehemorrhage rates were higher in brainstem lesions.10 In this analysis, it was not possible to compare familial hemorrhage and rehemorrhage rates with nonfamilial cases for the following reasons: 1) familial series included asymptomatic hemorrhage or rehemorrhage in their calculations, whereas nonfamilial cases did not; 2) hemorrhage and rehemorrhage rates were mostly mixed in familial series; and 3) we could not find symptomatic hemorrhage or rehemorrhage rates separately for brainstem and nonbrainstem lesions in familial series. In familial cases, any hemorrhage rate was found to be 4.3%/PY (either asymptomatic or symptomatic), and symptomatic hemorrhage was reported to be 1.43%/PY.13 In studies of
WORLD NEUROSURGERY, https://doi.org/10.1016/j.wneu.2019.02.115
LITERATURE REVIEW SHERVIN TASLIMI ET AL.
FAMILIAL VERSUS NONFAMILIAL CAVERNOUS MALFORMATION
familial cases that reported mixed hemorrhage and rehemorrhage rates, both asymptomatic and symptomatic rate was 11%/PY, whereas symptomatic rate ranged from 1.8%/PY to 6.5%/PY (Table 1).8,14,15 Seizure and Epilepsy Rate Table 2 shows the reported seizure and reseizure incidence rates in familial and nonfamilial cases in included studies. The annual seizure rate was 1.5% in nonfamilial cases.15-18 There was no difference in first-time seizure rate between familial and nonfamilial cases (4 nonfamilial studies15-18 vs. 1 familial study,13 P ¼ 0.9). Combining all 5 studies with FEM, first-time annual seizure rate was 1.5%/PY (Table 3). Annual reseizure rate was higher than seizure rate (2 reseizure rates15,17 vs. 5 seizure rates,13,15-18 P < 0.001). Exclusion of 1 study with an uncharacteristically high reseizure rate17 did not change the result (P < 0.001). Data were insufficient to compare the reseizure rate between familial and nonfamilial cases. Development of New Lesions New lesion development in familial cases was 32.1%/PY based on 3 articles (Table 4).8,13,14 Comparing new lesion development between familial versus nonfamilial cases, we found only 1 study that clearly differentiated familial and nonfamilial cases.15,19 New lesion development was higher in familial cases (32.1%/PY [3 studies, 80 patients] vs. nonfamilial cases (0.7%/PY [1 study, 13 patients], P < 0.001). Two studies reported the development of new lesions
in a general population of patients without specifying whether these cases were familial or nonfamilial CCMs.20,21 We could not extract the follow-up period in these studies. Therefore, we did not enter them into meta-analysis.
could not find any reports meeting our study inclusion criteria that addressed the change in the size of lesions in nonfamilial cases.
DISCUSSION Neuroimaging Characteristics Signal Change. We excluded apparent hemorrhage as a cause of signal change in our meta-analysis. LY of follow-up was used, as PY of follow-up would not be accurate in the case of multiple lesions. In familial series, in the study in which all patients were asymptomatic at diagnosis,13 the rate of change in signal was much lower (0.2%/LY) than our calculated incidence rate in 2 other familial studies that included mixtures of symptomatic and asymptomatic patients (1.7%/LY14 and 2.4%/LY8). Owing to the heterogeneity of the data and low number of studies, we did not combine the data. We could not find any report regarding signal change in nonfamilial cases. Size Change. Three studies of familial cases reported the incidence rate of increase in size of lesions or change in size of lesions based on LY of follow-up.8,13,14 We pooled these data using FEM. Details of the point estimate are presented in Table 4. In 1 study, 4 out of 5 lesion changes were due to hemorrhage.8 In the rest of the articles, it could not be ascertained whether these changes were due to hemorrhage (increase in size owing to hemorrhage or regression after previous hemorrhage) or were true dynamic changes of the lesions. We
In this study, we pooled reports of the natural history of familial CCMs compared with nonfamilial CCMs and addressed some sources of heterogeneity among the results. Natural history studies in familial cases initially showed higher overall rates of hemorrhage compared with nonfamilial series.8,13,14 We attribute part of this difference to the inclusion of asymptomatic hemorrhage in familial studies, whereas other studies defined the hemorrhagic event as the symptomatic event with evidence of bleeding on the imaging or pathology result at the time of the surgery. A mixture of hemorrhage and rehemorrhage made the comparison between familial and nonfamilial cases challenging. As shown before, brainstem and previously bled cavernomas are more likely to have symptomatic hemorrhages10; simply pooling hemorrhage rates across studies that do not present these numbers separately would be inaccurate. Compared with the familial hemorrhage rates reported in Table 1, in our previous meta-analysis of CCMs, the symptomatic hemorrhage rate was 2.8%/PY in brainstem lesions.10 Similarly, Horne et al.22 in their individual patient data meta-analysis found that the 5-year hemorrhage risk was 8.0% in brainstem lesions without previous hemorrhage or focal neurologic deficit. These numbers are very close to
Table 2. Reported Seizure Rates in Familial and Nonfamilial Cases Familial Series
Type of Study
Multiple Lesions
Percentage of Supratentorial Lesions
Seizure Rate (Number of Patients, Events)
Del Curling et al., 199116
No
Retrospective
19%
86%
1.51% (32, 16)
—
—
Josephson et al., 201117
No
Prospective
24%
77%
1.47% (38, 2)
94%
—
Moriarity et al., 199915
No
Prospective
37%
81%
2.4% (35, 4)
5.5%
4.8%
Reference
18
Reseizure Rate
Mixture of Seizure and Reseizure
Kondziolka et al., 1995
No
Prospective
20%
60%*
1.5% (94, 4)
—
—
Labauge et al., 200113
Yes
Prospective
85%
86%
1.42% (33, 1)
—
—
Zabramski et al., 19948
Yes
Prospective
76%
95%
—
—
6.49%
*This is an approximation because cerebral and cerebellar lesions are combined.
WORLD NEUROSURGERY 126: 241-246, JUNE 2019
www.journals.elsevier.com/world-neurosurgery
243
LITERATURE REVIEW SHERVIN TASLIMI ET AL.
FAMILIAL VERSUS NONFAMILIAL CAVERNOUS MALFORMATION
Table 3. Synthesis of Pooled Seizure and Reseizure Rates in Familial and Nonfamilial Cases Number of Studies (Number of Subjects)
Seizure Rate Reseizure versus seizure (incidence rate difference)
Model P Value for Overall Pooled Used Heterogeneity (I2) Estimate (95% CI) P Value References
5 (seizure, 289, reseizure, 33)
—
—
3.7% (2.9%e4.54%) <0.001
13,15e18
Seizure rate in nonfamilial series
4 (218)
FEM
0.73
1.5% (1.1%e2.2%)
<0.001
13,15,17,18
Seizure rate in familial and nonfamilial cases combined
5 (251)
FEM
0.86
1.5% (1.1%e2.2%)
<0.001
13,15e18
CI, confidence interval; FEM, fixed-effects model.
annual the 1.43% hemorrhage rate in familial brainstem cavernomas. Another important potential factor for hemorrhage may be the type of lesion, which can change over time.8,20 For instance, type 1 CCM lesions can become type 2 lesions if hemorrhage and thrombosis occur multiple times. Type 3 lesions, which show remote bleeding, can change into type 1 or 2 lesions if they rebleed. Type 1 or 2 lesions can change into type 3 lesions if they do not hemorrhage for some time. As previously shown, the rehemorrhage rate is higher during the first 2 years after bleeding.10 Taken together, type 3 lesions seem to be less active compared with type 1 and 2 lesions and to be associated with less hemorrhage. A sequential decrease in incidence of hemorrhage from type 1 to type 3 has been shown by Labauge et al.14 Nikoubashman et al.23 found the annual hemorrhage rate to be 29.8% for Zabramski type 1, 20.1% for type 2, 3.4% for type 3, and 1.3% for type 4 CCMs. The hemorrhage rate was 23.1% for their newly proposed type 5 lesions, which were CCMs that showed gross extralesional hemorrhage.23 In their
article, Nikoubashman et al.23 did not report hemorrhage and rehemorrhage rates separately; therefore, these were not added into our meta-analysis. The relationship of the type of lesion and the hemorrhage rate should be further assessed in future studies. Based on the present data, it is hard to draw a rigorous conclusion regarding the difference in hemorrhage rate between familial and nonfamilial cases stratified by the type of lesion. In future studies reporting natural history, we suggest stratifying for hemorrhage status (hemorrhage vs. rehemorrhage), timing after the hemorrhage, and type of lesions. This would help to reduce possible effect modification in comparison of familial and nonfamilial cases. Overall, hemorrhage rate may be higher in familial cases because of the presence of multiple lesions, as higher risk lesions bleed sooner and increase the overall incidence rate. Based on the available data, no distinction has been made between different locations and the incidence rate of seizure. This can potentially add to the heterogeneity of the data. Because lesions in different locations may confer different
incidence rates of seizure, seizure and reseizure incidence rates should be reported based on the location of the lesions. Josephson et al.17 did not find an association between the anatomic location of CCMs and the occurrence of seizures but did note that cavernomas were more likely to be multiple in adults with seizures. Moreover, the type of lesion may be an important factor for seizure development, such as an acute hemorrhage in the lesion causing a seizure. Given the different possible propensities for hemorrhage in the different types of lesions, lesion type may also be an important factor in predicting the seizure rate. This should be differentiated with other processes that can also lead to seizures, such as hemosiderin ring or gliosis around the CCM. CCMs appear to show dynamic changes in familial cases. We found new lesion development in familial cases to be 32.1%/ PY.8,13,14 Gross et al.24 found new lesion development in patients with multiple cavernomas on presentation to be 3.0%/ PY compared with 0.4% in patients with solitary nonfamilial cavernomas. Notably,
Table 4. Dynamic Changes in Familial Cases Number of Studies (Number of Subjects)
Model Used
P Value for Heterogeneity (I2)
Overall Pooled Estimate (95% CI)
P Value
References
Development of new lesions (PY) in familial cases
3 (94)
REM
0.049 (69%)
32.1% (20%e51.2%)
<0.001
13,14,19
Change in size (LY)
3 (94)
FEM
0.6
1.1% (0.6%e1.6%)
<0.001
13,14,19
Change in size (PY)
3 (94)
FEM
0.5
8% (5.2%e12.2%)
<0.001
13,14,19
Increase in size (LY)
3 (94)
FEM
0.640
0.8% (0.4%e1.3%)
<0.001
13,14,19
Increase in size (PY)
3 (94)
FEM
0.63
5.9% (5.9%e9.9%)
<0.001
13,14,19
Incidence Rate
CI, confidence interval; PY, person-year; REM, random-effects model; LY, lesion-year; FEM, fixed-effects model.
244
www.SCIENCEDIRECT.com
WORLD NEUROSURGERY, https://doi.org/10.1016/j.wneu.2019.02.115
LITERATURE REVIEW SHERVIN TASLIMI ET AL.
multiplicity of cavernomas does not equate with familial cavernoma, and the definition by Gross et al.24 of a new lesion to be at least 4 mm is different from previous studies; therefore, direct comparison and addition of these data into our metaanalysis were not possible. In addition, several nonfamilial studies have shown some dynamic changes as well. Kupersmith et al.25 evaluated nonfamilial brainstem cases (familial cases assessed only with obtaining a family history of CCM) and observed some size changes in the lesions. These data were not included in our study because they did not meet our criteria for data extraction. Some other series also reported dynamic changes in patients with CCMs, but we could not find any criteria in these studies to differentiate familial from nonfamilial cases.16,20,21,26-28 However, it cannot be ascertained whether the cases that showed dynamic changes were nonfamilial or if these changes were related to bleeding episodes. It may be important to evaluate changes in lesions that are not due to hemorrhage by examining size and signal changes as well as the development of new lesions. So far, no studies have done this. Taken together, even though the presence of dynamic changes is obvious in familial cases, it is hard to draw any conclusion regarding dynamic changes in nonfamilial cases because most series pooled familial and nonfamilial cases. The current study has several limitations. We were not able to complete some subgroup analyses owing to the limited number of studies. For some of our outcome measures, direct statistical comparisons between familial and nonfamilial cases were not possible mainly because of the mixture of the patients in different subgroups or inadequate reporting of data. Based on the recent definition of familial CCM, individuals with multiple lesions are considered to have familial CCMs. In our included articles, none of the nonfamilial series considered this criterion. This may have affected our results. CONCLUSIONS Familial CCMs exhibit dynamic changes, including change in the size of lesions, development of new lesions, change in the signal of lesions, and change in the type of
FAMILIAL VERSUS NONFAMILIAL CAVERNOUS MALFORMATION
lesions. Hemorrhage can explain only part of these changes; therefore, other processes apparently are involved in these dynamic changes. The risk of hemorrhage in familial cases may be higher than in nonfamilial cases because of the presence of multiple lesions, but further study is needed to evaluate this. Potential sources of heterogeneity in estimating the incidence rate of hemorrhage, rehemorrhage, and seizures in these lesions are the type of presentation, location, and type of lesions. Improved reporting and additional studies regarding the clinical behavior and natural history of familial and nonfamilial cavernomas are required to determine best management. REFERENCES 1. Tomlinson FH, Houser OW, Scheithauer BW, Sundt TM Jr, Okazaki H, Parisi JE. Angiographically occult vascular malformations: a correlative study of features on magnetic resonance imaging and histological examination. Neurosurgery. 1994; 34:792-799 [discussion 799-800]. 2. Raychaudhuri R, Batjer HH, Awad IA. Intracranial cavernous angioma: a practical review of clinical and biological aspects. Surg Neurol. 2005;63: 319-328 [discussion: 328]. 3. Maraire JN, Awad IA. Intracranial cavernous malformations: lesion behavior and management strategies. Neurosurgery. 1995;37:591-605. 4. Frischer JM, Pipp I, Stavrou I, Trattnig S, Hainfellner JA, Knosp E. Cerebral cavernous malformations: congruency of histopathological features with the current clinical definition. J Neurol Neurosurg Psychiatry. 2008;79:783-788. 5. Washington CW, McCoy KE, Zipfel GJ. Update on the natural history of cavernous malformations and factors predicting aggressive clinical presentation. Neurosurg Focus. 2010;29:E7. 6. Labauge P, Denier C, Bergametti F, TournierLasserve E. Genetics of cavernous angiomas. Lancet Neurol. 2007;6:237-244. 7. Riant F, Bergametti F, Ayrignac X, Boulday G, Tournier-Lasserve E. Recent insights into cerebral cavernous malformations: the molecular genetics of CCM. FEBS J. 2010;277:1070-1075.
meta-analysis of 25 studies. Neurology. 2016;86: 1984-1991. 11. Wilczynski NL, Haynes RB. Developing optimal search strategies for detecting clinically sound prognostic studies in MEDLINE: an analytic survey. BMC Med. 2004;2:23. 12. Wilczynski NL, Haynes RB. Optimal search strategies for detecting clinically sound prognostic studies in EMBASE: an analytic survey. J Am Med Inform Assoc. 2005;12:481-485. 13. Labauge P, Brunereau L, Laberge S, Houtteville JP. Prospective follow-up of 33 asymptomatic patients with familial cerebral cavernous malformations. Neurology. 2001;57:1825-1828. 14. Labauge P, Brunereau L, Levy C, Laberge S, Houtteville JP. The natural history of familial cerebral cavernomas: a retrospective MRI study of 40 patients. Neuroradiology. 2000;42:327-332. 15. Moriarity JL, Wetzel M, Clatterbuck RE, et al. The natural history of cavernous malformations: a prospective study of 68 patients. Neurosurgery. 1999;44:1166-1173. 16. Del Curling O Jr, Kelly DL Jr, Elster AD, Craven TE. An analysis of the natural history of cavernous angiomas. J Neurosurg. 1991;75:702-708. 17. Josephson CB, Leach JP, Duncan R, et al. Seizure risk from cavernous or arteriovenous malformations: prospective population-based study. Neurology. 2011;76:1548-1554. 18. Kondziolka D, Lunsford LD, Kestle JR. The natural history of cerebral cavernous malformations. J Neurosurg. 1995;83:820-824. 19. Clatterbuck RE, Moriarity JL, Elmaci I, Lee RR, Breiter SN, Rigamonti D. Dynamic nature of cavernous malformations: a prospective magnetic resonance imaging study with volumetric analysis. J Neurosurg. 2000;93:981-986. 20. Willinsky R, Harper W, Wallace MC, et al. Followup MR of intracranial cavernomas: the relationship between haemorrhagic events and morphology. Interv Neuroradiol. 1996;2:127-135. 21. Kim DS, Park YG, Choi JU, Chung SS, Lee KC. An analysis of the natural history of cavernous malformations. Surg Neurol. 1997;48:9-18. 22. Horne MA, Flemming KD, Su IC, et al. Clinical course of untreated cerebral cavernous malformations: a meta-analysis of individual patient data. Lancet Neurol. 2016;15:166-173.
8. Zabramski JM, Wascher TM, Spetzler RF, et al. The natural history of familial cavernous malformations: results of an ongoing study. J Neurosurg. 1994;80:422-432.
23. Nikoubashman O, Di Rocco F, Davagnanam I, Mankad K, Zerah M, Wiesmann M. Prospective hemorrhage rates of cerebral cavernous malformations in children and adolescents based on MRI appearance. AJNR Am J Neuroradiol. 2015;36: 2177-2183.
9. Moher D, Liberati A, Tetzlaff J, Altman DG, PRISMA Group. Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. PLoS Med. 2009;6:e1000097.
24. Gross BA, Du R, Orbach DB, Scott RM, Smith ER. The natural history of cerebral cavernous malformations in children. J Neurosurg Pediatr. 2016;17: 123-128.
10. Taslimi S, Modabbernia A, Amin-Hanjani S, Barker FG II, Macdonald RL. Natural history of cavernous malformation: systematic review and
25. Kupersmith MJ, Kalish H, Epstein F, et al. Natural history of brainstem cavernous malformations. Neurosurgery. 2001;48:47-53 [discussion: 53-54].
WORLD NEUROSURGERY 126: 241-246, JUNE 2019
www.journals.elsevier.com/world-neurosurgery
245
LITERATURE REVIEW SHERVIN TASLIMI ET AL.
26. Robinson JR, Awad IA, Little JR. Natural-history of the cavernous angioma. J Neurosurg. 1991;75:709-714. 27. Li D, Yang Y, Hao SY, et al. Hemorrhage risk, surgical management, and functional outcome of brainstem cavernous malformations. J Neurosurg. 2013;119:996-1008. 28. Pozzati E, Giuliani G, Nuzzo G, Poppi M, Piepgras DG, Hunt Batjer H. The growth of
246
www.SCIENCEDIRECT.com
FAMILIAL VERSUS NONFAMILIAL CAVERNOUS MALFORMATION
cerebral cavernous angiomas. Neurosurgery. 1989; 25:92-97.
Conflict of interest statement: R. L. Macdonald receives grants from the Brain Aneurysm Foundation and Genome Canada. The remaining authors have no conflicts to report. Received 17 August 2016; accepted 10 February 2019
Citation: World Neurosurg. (2019) 126:241-246. https://doi.org/10.1016/j.wneu.2019.02.115 Journal homepage: www.journals.elsevier.com/worldneurosurgery Available online: www.sciencedirect.com 1878-8750/$ - see front matter ª 2019 Elsevier Inc. All rights reserved.
WORLD NEUROSURGERY, https://doi.org/10.1016/j.wneu.2019.02.115