Customize your JAMA Network experience by selecting one or more topics from the list below.
Lather HD, Gornik HL, Olin JW, et al. Prevalence of Intracranial Aneurysm in Women With Fibromuscular Dysplasia: A Report From the US Registry for Fibromuscular Dysplasia. JAMA Neurol. 2017;74(9):1081–1087. doi:10.1001/jamaneurol.2017.1333
What is the prevalence of intracranial aneurysm in women with fibromuscular dysplasia?
In this cross-sectional registry study that included 669 women with a diagnosis of fibromuscular dysplasia and intracranial imaging, the prevalence of intracranial aneurysm was 12.9%.
The prevalence of intracranial aneurysm in women with fibromuscular dysplasia is high enough to warrant consideration of screening patients with fibromuscular dysplasia for intracranial aneurysm with noninvasive imaging.
The prevalence of intracranial aneurysm in patients with fibromuscular dysplasia (FMD) is uncertain.
To examine the prevalence of intracranial aneurysm in women diagnosed with FMD.
Design, Setting, and Participants
This cross-sectional study included 669 women with intracranial imaging registered in the US Registry for Fibromuscular Dysplasia, an observational disease-based registry of patients with FMD confirmed by vascular imaging and currently enrolling at 14 participating US academic centers. Registry enrollment began in 2008, and data were abstracted in September 2015. Patients younger than 18 years at the time of FMD diagnosis were excluded. Imaging reports of all patients with reported internal carotid, vertebral, or suspected intracranial artery aneurysms were reviewed. Only saccular or broad-based aneurysms 2 mm or larger in greatest dimension were included. Extradural aneurysms in the internal carotid artery were included; fusiform aneurysms, infundibulae, and vascular segments with uncertainty were excluded.
Main Outcomes and Measures
Percentage of women with FMD with intracranial imaging who had an intracranial aneurysm.
Of 1112 female patients in the registry, 669 (60.2%) had undergone intracranial imaging at the time of enrollment (mean [SD] age at enrollment, 55.6 [10.9] years). Of the 669 patients included in the analysis, 86 (12.9%; 95% CI, 10.3%-15.9%) had at least 1 intracranial aneurysm. Of these 86 patients, 25 (53.8%) had more than 1 intracranial aneurysm. Intracranial aneurysms 5 mm or larger occurred in 32 of 74 patients (43.2%), and 24 of 128 intracranial aneurysms (18.8%) were in the posterior communicating or posterior arteries. The presence of intracranial aneurysm did not vary with location of extracranial FMD involvement. A history of smoking was significantly associated with intracranial aneurysm: 42 of 78 patients with intracranial aneurysm (53.8%) had a smoking history vs 163 of 564 patients without intracranial aneurysm (28.9%; P < .001).
Conclusions and Relevance
The prevalence of intracranial aneurysm in women diagnosed with FMD is significantly higher than reported in the general population. Although the clinical benefit of screening for intracranial aneurysm in patients with FMD has yet to be proven, these data lend support to the recommendation that all patients with FMD undergo intracranial imaging if not already performed.
Fibromuscular dysplasia (FMD) is an uncommon, noninflammatory, nonatherosclerotic disease of the medium and large arteries.1,2 Its cause is not known; however, it is far more common in women.3 It can occur in any arterial bed, although it is most commonly found in the renal and cervical arteries.3 The clinical presentation of FMD varies, depending on the location of the arterial lesions. Often FMD is asymptomatic and therefore frequently found incidentally. In addition, FMD is associated with a substantial risk of aneurysm formation and rupture, as well as arterial dissection and occlusion.3,4
Previously published data suggest an increased risk of intracranial aneurysm (IA) in patients with carotid artery (CA) and vertebral artery (VA) FMD.2 Intracranial aneurysms may rupture, leading to subarachnoid hemorrhage (SAH) and significant morbidity and mortality. The mortality among patients who experience SAH is approximately 50%,5 with 12% dying immediately.6 The most commonly identified cause of SAH is IA rupture (85%).7 It is not known whether the natural history of IA in patients with FMD is similar to that in patients without FMD. Although outcomes in patients with SAH are poor,6 sizable IAs identified before rupture can be repaired or excluded with endovascular or surgical techniques.8 However, the risk of treatment must be balanced with the risk of rupture9; therefore, the clinical benefit of screening for IA, even in populations with known increased risk, is uncertain.10,11
Previous estimates of IA prevalence among patients with CA and/or VA FMD come from neurosurgery and radiology case series, which suffer from selection bias in patient inclusion.12 Previous studies12 estimated the prevalence of IA to be 21% to 51%. The most recent estimate was published in a 1998 meta-analysis and case series by Cloft and colleagues12 in which IA prevalence was calculated with and without IA symptoms (eg, SAH) among patients with CA and/or VA FMD who underwent cerebral angiography. In most patients with IA in that series, SAH was the indication for angiography; FMD was often incidentally found. Among asymptomatic patients with FMD, there was a mean (SD) prevalence of IA of 7.3% (2.2%; 38 of 517 patients).
In the 2014 American Heart Association (AHA) Scientific Statement on FMD, the writing committee identified the determination of the prevalence of IA among patients with FMD as a top research priority.2 Although there have been several estimates of IA prevalence among patients with FMD in the CAs and/or VAs, to our knowledge, there are no published reports of the prevalence of IA among patients with FMD in the renal arteries. Because of the life-threatening consequences of ruptured IAs and the possible association with FMD, the writing committee recommended IA screening for all patients with FMD found in any arterial bed. The current study uses the largest database of patients with FMD, the US Registry for Fibromuscular Dysplasia, to estimate the prevalence of IA and characterize IAs in FMD.
The US Registry for Fibromuscular Dysplasia is an observational disease-based registry of patients with diagnosis of FMD confirmed by vascular imaging and currently enrolling at 14 participating centers. Registry enrollment began in 2008, and data were abstracted in September 2015. Registry details and standardization of data collection have been described previously.3 Written informed consent was obtained from all participants, and all sites had institutional review board approval (Institutional Review Boards of the University of Michigan Medical School, Cleveland Clinic Foundation Institutional Review Board, Program for the Protection of Human Subjects at Icahn School of Medicine at Mount Sinai, University of Virginia Institutional Review Board for Health Sciences Research, Greenville Health System Institutional Review Board, Partners Institutional Review Boards, and University of California, Davis, Institutional Review Board). All data were deidentified.
Patients enrolled from all sites in the US Registry for Fibromuscular Dysplasia who had undergone intracranial imaging at or before enrollment, as documented on the initial registry data form, were included. For the purposes of this report, intracranial imaging was defined as catheter-based angiography, computed tomographic angiography, or magnetic resonance angiography (MRA). Patients were also included in this study if they had a known repaired IA. Patients younger than 18 years at the time of FMD diagnosis were excluded because of concern that FMD and IAs in children may be different from those in adults.8,13 Because male sex may affect the pathophysiologic mechanism of FMD14 and aneurysms,15 male patients were also excluded. In addition, male patients constituted only 5.7% of individuals in the registry. Although positive family history of IA or SAH increases the likelihood of IA by 3.4-fold,15 this information was not available in the database.
Intracranial aneurysm was defined as a saccular or broad-based aneurysm greater than or equal to 2 mm in greatest dimension occurring at or above the level of the skull base. Extradural (but intracranial) aneurysms in the petrous segment of the internal CA (C2 in the Bouthillier classification16) and above were included. To ensure that the definition of IA was met, all available imaging reports from each patient with a CA aneurysm, VA aneurysm, and/or IA were reviewed and assessed using these criteria. If patients had multiple imaging reports, the largest recorded dimension of each IA was used. Any registry site that was unable to provide deidentified imaging reports to the coordinating center was excluded from the analysis.
To address the question of whether IA is more prevalent in patients with cervical FMD vs noncervical FMD, we analyzed the prevalence of IA in patients with renal and/or cervical FMD. Renal FMD was defined by imaging consistent with FMD in one or both renal arteries. Cervical FMD was defined as imaging consistent with FMD in extracranial CAs and/or VAs. Noncervical and nonrenal subgroups were defined by the absence of FMD in those vascular beds only if both renal and cervical imaging were performed. Patients with noncervical FMD were defined as having no FMD in the extracranial CAs and/or VAs. Most of these patients had renal FMD but could have had FMD in other vascular beds (eg, external iliac). Nonrenal FMD was defined similarly, and likewise, patients in this group had predominantly cervical FMD.
Comparisons of differences between groups were performed with 2-tailed, unpaired t tests or Wilcoxon rank sum tests for continuous variables and with χ2 or Fisher exact tests for categorical variables. A 2-sided P < .05 was considered to be statistically significant for all comparisons. All analyses were performed with SAS statistical software, version 9.3 (SAS Institute Inc).
At the time of data extraction (September 2015), there were 1145 female patients and 71 male patients across 11 sites. Of the female patients, 685 (59.8%) had undergone intracranial imaging before enrollment. Two of these female patients were younger than 18 years at the time of FMD diagnosis and thus were excluded. Fourteen patients (2.0%) with intracranial imaging were excluded because the 2 enrolling sites for those patients were unable to provide imaging reports. The final cohort for this study consisted of 669 women, all with intracranial imaging (mean [SD] age at enrollment, 55.6 [10.9] years). Demographic and clinical characteristics of these women are presented in Table 1.
The overall prevalence of saccular or broad-based IA of at least 2 mm in greatest dimension among women with FMD with intracranial imaging and enrolled in the registry was 12.9% (86 of 669 women; 95% CI, 10.3%-15.9%). The prevalence of IA was 11.9% (41 of 344; 95% CI, 8.6%-16.2%) among patients with renal FMD and 13.7% (77 of 563; 95% CI, 10.8%-17.1%) among patients with cervical FMD. A total of 242 patients were identified with renal and cervical FMD, and 32 (13.2%) of these patients had IA. The prevalence of IA among patients with intracranial FMD was not assessed because many centers consider IA to be a manifestation of intracranial FMD and thus record it as intracranial FMD even if traditional beading was not present in intracranial arteries.
The IA prevalence was not significantly different when stratified by arterial location of FMD. No significant difference was found between the prevalence of IA among patients with noncervical FMD (primarily renal FMD) (7.7% [8 of 104]; 95% CI, 3.3%-15.2%) and patients with nonrenal FMD (12.3% [31 of 252]; 95% CI, 8.4%-17.5%). Conversely, the prevalence of cervical FMD with IA (90.6% [77 of 85]) was not significantly different (P = .14) from that of cervical FMD without IA (83.9% [486 of 579]), and the percentage of patients with IA and renal FMD (56.9% [41 of 72]) was not significantly different from that of patients without IA with renal FMD (57.4% [303 of 528]; P = .94).
To account for the possible selection bias in imaging raised by Cloft and colleagues,12 unruptured IA (UIA) prevalence was also calculated by subtracting patients with a history of SAH (11.1% [66 of 593 women in the registry with no history of SAH had IA]; 95% CI, 8.6%-14.2%). Patient characteristics of those with and without intracranial imaging are presented in Table 1.
A total of 12 women with intracranial imaging had a potential IA that did not meet our criteria: 2 had saccular aneurysms smaller than 2 mm, 3 had fusiform aneurysms, and 7 had aneurysms that were not definitively saccular aneurysms. Many patients had multiple IAs; 128 saccular or broad-based IAs were found in 86 women with FMD. The number of IAs in each patient ranged from 1 to 8, with a median of 1 and a mean (SD) of 1.5 (1.0). A total of 26 of 86 women (30.2%) had more than 1 IA. Aneurysm locations are reported in Table 2. Of the 2 patients with unknown IA location, 1 patient had 2 aneurysms, and the IA location and repair status of that patient were unknown from the available medical records. The other patient had a remote history of aneurysm repair but no specific documentation of the site. Eleven IAs in the intracranial portion of the internal CA were extradural. The exact location of 5 additional IAs in the internal CA, although intracranial, could not be discerned from imaging reports. The IA sizes are reported in Table 3. Size was unavailable in some patients who had a record of IA repair but lacked original imaging reports. Of the 74 patients with all IAs of known size, the largest IA per patient was 5 mm or larger in 32 (43.2%).
Demographic characteristics and medical history were compared between patients with and without IA (Table 4). Only history of smoking, increased age at smoking cessation, number of tobacco pack-years, and SAH were significantly associated with IA. The presence of a nonintracranial aneurysm was not associated with IA: 9 of 86 patients with IA (10.5%) had a nonintracranial aneurysm, whereas 60 of 583 patients without IA (10.3%) had a nonintracranial aneurysm.
We found a 12.9% prevalence of IAs among women with FMD with intracranial imaging at the time of enrollment in the US Registry for Fibromuscular Dysplasia. The prevalence of IA was not significantly different when stratified by location of FMD.
Unlike previous reports of IA prevalence among patients with FMD, the women currently described in this study are not from a neurosurgical case series. Furthermore, a prior report12 performed before noninvasive imaging was widely used to evaluate less serious presentations, such as headache and/or pulsatile tinnitus. Other investigators have reported estimates of IA prevalence as high as 50% in cervical FMD.12 Cloft and colleagues12 tried to correct for the bias inherent in selecting patients with FMD who were undergoing catheter-based intracranial angiography by excluding patients with signs and symptoms of IA (namely, SAH and cranial nerve defects). Adjusting for symptomatic IAs in that series decreased reported IA prevalence in cervical FMD from 23.9% to 6.3%.12 In contrast, the IA prevalence found in the current report from the registry was minimally different, excluding patients with SAH (decreasing from 12.9% to 11.1%), suggesting that intracranial imaging in the registry is not nearly as biased toward those with symptomatic aneurysms as in previous reports. Aside from IA, few clinical characteristics are specific for SAH; therefore, imaging performed to investigate symptoms, such as strokes and headaches without SAH, is less likely to bias toward IA prevalence. Estimates of the prevalence of UIAs in the general population have improved greatly since Cloft and colleagues12 estimated the IA prevalence among patients with FMD in 1998. Vlak and colleagues15 performed a systematic review and meta-analysis of UIAs in the general population in 2011 that identified 1450 UIAs in 94 912 patients, for an overall prevalence of 3.2% (95% CI, 1.9%-5.2%) in a population without comorbidity, with a mean age of 50 years, and consisting of 50% men. Smoking and hypertension were not evaluated as comorbidities. The prevalence ratios in populations of women were 1.61 (95% CI, 1.02-2.54) for a mean age of 50 years and 2.2 (95% CI, 1.3-3.6) for a mean age older than 50 years. The overall prevalence of UIA among women of all ages was 6.0% (95% CI, 4.5%-8.0%). This meta-analysis included a notable study17 in which 7345 healthy Japanese volunteers (mean age, 55.5 years) were screened with MRA for intracranial abnormalities, including aneurysm. The UIA prevalence among female patients was 2.7%, and overall (male and female) mean (SD) diameter was 3.9 (1.6) mm. In a subsequent cross-sectional study,18 4813 Chinese adults aged 35 to 75 years were screened with MRA for UIA. The prevalence was 7.0% (95% CI, 6.3%-7.7%) overall and 8.4% (7.3%-9.5%) among women. Among women aged 55 to 64 years, the prevalence was 11.0% (95% CI, 8.4%-13.6%), and among women aged 65 to 75 years, the prevalence was 9.9% (95% CI, 6.9%-13.0%). The prevalence of UIA (ie, no SAH) in the US Registry of Fibromuscular Dysplasia (11.1%) is significantly greater than the 6.0%15 (P < .001) and 8.4%18 (P = .03) reported in the other aforementioned studies.
Of interest, the UIAs that we report from the registry have higher-risk features for rupture (ie, they are larger and more often in the posterior circulation) than do those in screening studies of healthy volunteers. In an MRI screening study19 in the Netherlands with 2000 participants (mean age, 63.3 years; 52% women), UIA was found in 35 individuals (1.8%), but only 3 UIAs (8.6%) were 7 mm or larger, and the largest was only 12 mm. In the Chinese study,18 90.2% of UIAs had a maximum diameter less than 5 mm, with a mean diameter of 3.7 mm in women. Only 0.9% of UIAs were 10 mm or larger. The authors noted that the increased prevalence and preponderance of small aneurysms might be in part attributable to the high-quality 3.0-T 3-dimensional time-of-flight MRA used in the study. In contrast, the mean size of UIAs in the current study of women with FMD was at least 5.0 mm, and the largest UIA in 22.0% of patients was 7 mm or larger (although 10.6% of patients had at least 1 UIA of unknown size owing to repair and thus were excluded from this calculation). Rupture risk also depends on location; UIAs in the posterior communicating (relative risk, 2.4) and posterior circulation (relative risk, 2.5) have the greatest risk of rupture, compared with middle cerebral arteries.9 In the current report, 18.3% of UIAs were in these higher-risk arteries, compared with 5.7%19 and 2.2%18 in other studies. In addition, only 8.0% of patients in the study by Li et al18 had multiple aneurysms compared with 24.2% of patients with FMD. However, even after correcting for sex and age, significant differences were found in the patient populations; therefore, direct comparison is difficult.
Finally, smoking was strongly associated with IA in patients in the registry, consistent with related registry findings.20 Smoking is considered a risk factor for the development of IA in patients without FMD,11 although it was not associated with IA prevalence in the Chinese study.18 Smoking has also been associated with onset and severity of renal FMD.21 Perhaps smoking similarly worsens the underlying FMD disease state that leads to IA formation. In addition, smoking is thought to be a risk factor for rupture of IA in patients without FMD.22 Although it has long been recommended that patients with FMD do not smoke,23 the association with IA and particularly age of cessation and number of pack-years provides more compelling evidence in favor of smoking cessation.
The benefits of screening for IA must be balanced with the risks of intervention, among other considerations. Although one mathematical model found that the clinical benefit of screening depended not on the prevalence of IA but the risk of rupture,10 the AHA/American Stroke Association guideline on the management of patients with UIA still recommends screening in selected populations based on increased prevalence.11
This study has a number of limitations. Not all patients enrolled in the registry had undergone intracranial imaging at the time of initial enrollment in the registry, and there were no standard protocols to recommend it; thus, potential for bias in patient selection for imaging exists. Differences were found in demographic and clinical characteristics between those patients with FMD with and without intracranial imaging (Table 1), although patients without intracranial imaging were older and had a higher prevalence of hypertension and smoking history, established risk factors for IA. Furthermore, imaging protocols for IA were not standardized across centers; thus, choice of modality and quality of images may have varied significantly. This likely underestimates IA prevalence because 1.5-T MRA and lower-quality computed tomographic angiography likely miss rather than falsely report IAs. All registry research introduces selection bias and has other inherent limitations, including potential issues with data quality and completeness, recruitment and enrollment strategies, and confounding variables, both known and unknown. Specifically, the US Registry for Fibromuscular Dysplasia has unique limitations because of the nature of FMD, including its underdiagnosis, heterogeneity of presentation, and poorly understood mechanism and natural history. Data derived from the registry are also likely to be enriched with patients with symptomatic and complicated FMD who seek care at the referral centers that comprise registry centers.
Reports of imaging studies were reviewed according to standardized criteria by 2 of us (H.D.L. and J.B.F.), but the existing resources of the US Registry of Fibromuscular Dysplasia did not allow for centralized review of imaging studies in a core laboratory. Some imaging reports were not available for review, such as those performed at nonparticipating hospitals. These factors make comparison of these findings with those of several large, high-quality studies17-19 in patients without FMD less definitive.
The prevalence of IA among women with FMD at enrollment in the US Registry for Fibromuscular Dysplasia was 12.9%, significantly greater than that estimated for the general population of women with a mean age older than 50 years. There was no difference in IA prevalence among patients in the registry with FMD in the renal or cervical arteries, suggesting that location of FMD does not affect the prevalence of IA. Many registry patients with FMD had multiple IAs, IAs of significant size, and IAs in rupture-prone locations. These data support a strong association of IAs with FMD. In light of the AHA/American Stroke Association guidelines on treating patients with unruptured IA,11 the current data support the recommendation for screening intracranial imaging in patients with FMD.2 These data also suggest that a prospective intracranial imaging study of consecutive patients with FMD should be performed.
Corresponding Author: James B. Froehlich, MD, MPH, Cardiovascular Center, University of Michigan Medical School, 1500 E Medical Center Dr, SPC 5853, Ann Arbor, MI 48109 (email@example.com).
Accepted for Publication: April 19, 2017.
Correction: This article was corrected on January 2, 2017, to change “dystrophy” to “dysplasia” throughout the article.
Published Online: July 17, 2017. doi:10.1001/jamaneurol.2017.1333
Author Contributions: Ms Gu and Dr Froehlich had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Froehlich, Lather, Gornik, Olin, Heidt, Sharma, Kline-Rogers.
Acquisition, analysis, or interpretation of data: Froehlich, Lather, Gornik, Olin, Gu, Heidt, Kim, Kadian-Dodov, Gray, Jaff, Chi, Mace, Kline-Rogers.
Drafting of the manuscript: Lather, Gornik, Chi, Mace.
Critical revision of the manuscript for important intellectual content: Froehlich, Lather, Gornik, Olin, Gu, Heidt, Kim, Kadian-Dodov, Sharma, Gray, Jaff, Kline-Rogers.
Statistical analysis: Lather, Gu.
Obtained funding: Froehlich, Olin.
Administrative, technical, or material support: Froehlich, Olin, Heidt, Sharma, Gray, Chi, Kline-Rogers.
Study supervision: Froehlich, Gornik, Olin, Heidt, Kim, Sharma, Kline-Rogers.
Conflict of Interest Disclosures: Dr Gornik reports being a noncompensated board member and medical advisory committee member of the Fibromuscular Dysplasia Society of America (FMDSA). Dr Olin reports being a noncompensated board member and medical advisory committee chair of the FMDSA, a steering committee member for the AstraZeneca EUCLID (Effects of Ticagrelor and Clopidogrel in Patients with Peripheral Artery Disease) Trial, and a member of the AstraZeneca medical advisory board. Dr Kim is a consultant for Philips Ultrasound. Dr Gray is a consultant for Medtronic and Gore. Dr Jaff reports being a compensated VIVA Physicians board member. Ms Mace reports being the Executive Director of the FMDSA. Dr Froehlich reports being a consultant for Janssen Pharmaceuticals, Merck, and Novartis and an advisory committee member of Boehringer-Ingelheim and Pfizer and receiving grant/research support from Blue Cross Blue Shield of Michigan and the FMDSA. No other disclosures were reported.
Funding/Support: This work was supported by the FMDSA, a nonprofit organization, and the University of Michigan Medical School through the Summer Biomedical Research Program (Mr Lather).
Role of the Funder/Sponsor: The funding source had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and the decision to submit the manuscript for publication.
Additional Contributions: Ellen Brinza, BS, Cleveland Clinic, Cleveland, Ohio, assisted in abstracting and preparing imaging reports and Rachel Krallman, BS, Michigan Medicine, University of Michigan, Ann Arbor, assisted in preparing the manuscript; they were not compensated for their work on this article. Cleveland Clinic, Mount Sinai Health System, University of Virginia Health System, University of California Davis Health System, Greenville Health System, Massachusetts General Hospital, and Michigan Medicine provided data and responded to imaging queries.
Create a personal account or sign in to: