Brain Magnetic Resonance Imaging Findings Fail to Suspect Fabry Disease in Young Patients With an Acute Cerebrovascular Event
Background and Purpose—Fabry disease (FD) may cause stroke and is reportedly associated with typical brain findings on magnetic resonance imaging (MRI). In a large group of young patients with an acute cerebrovascular event, we wanted to test whether brain MRI findings can serve to suggest the presence of FD.
Methods—The Stroke in Young Fabry Patients (SIFAP 1) study prospectively collected clinical, laboratory, and radiological data of 5023 patients (18–55 years) with an acute cerebrovascular event. Their MRI was interpreted centrally and blinded to all other information. Biochemical findings and genetic testing served to diagnose FD in 45 (0.9%) patients. We compared the imaging findings between FD and non-FD patients in patients with at least a T2-weighted MRI of good quality.
Results—A total of 3203 (63.8%) patients had the required MRI data set. Among those were 34 patients with a diagnosis of FD (1.1%), which was definite in 21 and probable in 13 cases. The median age of patients with FD was slightly lower (45 versus 46 years) and women prevailed (70.6% versus 40.7%; P<0.001). Presence or extent of white matter hyperintensities, infarct localization, vertebrobasilar artery dilatation, T1-signal hyperintensity of the pulvinar thalami, or any other MRI finding did not distinguish patients with FD from non-FD cerebrovascular event patients. Pulvinar hyperintensity was not present in a single patient with FD but seen in 6 non-FD patients.
Conclusions—Brain MRI findings cannot serve to suspect FD in young patients presenting with an acute cerebrovascular event. This deserves consideration in the search for possible causes of young patients with stroke.
Fabry disease (FD) is an X-linked lysosomal storage disorder characterized by deficient activity of α-galactosidase A. This leads to pathological accumulation of predominantly globotriaosylceramide into vascular endothelial, neural, and renal cells, as well as cardiomyocytes interfering with their integrity and function.1,2 Early disease manifestations thus consist of acroparesthesias, angiokeratoma, and hypohidrosis, whereas kidney, cardiac, and central nervous system complications develop in later life.2 Because of this variability of presentation, affected individuals often have a long medical history sometimes with renal replacement therapy and even cardiac or renal transplantation before the correct diagnosis is established.1,3 Consequently, specific abnormalities that can raise the suspicion for FD are of great importance. For the brain, several findings have been reported to be associated with FD (Figure). These include extensive white matter changes,4,5 cerebral infarctions especially in regions supplied by the posterior circulation,6,7 and associated dilatation of the vertebrobasilar vessels6,7 up to extensive dolichoectasia.8 In addition, although such abnormalities can have several causes, the observation of high signal intensity of the pulvinar thalami on T1-weighted magnetic resonance imaging (MRI) in up to one quarter of individuals diagnosed with FD was suggested to be pathognomonic of this disorder.9,10
The Stroke in Young Fabry Patients (SIFAP 1) study prospectively investigated an unselected study population of young patients with an acute cerebrovascular event (CVE) and found a proportion of 0.9% to have definite or probable FD using genetic testing and determination of globotriaosylceramide in blood and urine and of lyso-globotriaosylceramide in human plasma.11 All study participants were asked to undergo MRI of the brain, which was collected and interpreted centrally and unaware of the patients’ clinical and laboratory findings.12 This provided a unique opportunity to investigate the extent to which brain abnormalities considered characteristic for FD exist in such patients and can serve to suspect this disorder and to prompt specific diagnostic workup.
The SIFAP 1 study enrolled 5023 young (18–55 years) patients with an acute CVE in 15 European countries and 47 study centers experienced in stroke care and the main result on the prevalence of FD, as well as the study’s inclusion and exclusion criteria have already been reported.11,12 In short, CVE was defined as an acute cerebral neurological deficit of any vascular cause. Besides comprehensive work-up, the diagnosis of CVE had to be verified by brain MRI (82%), or by a qualified stroke neurologist with at least 5 years of experience in treating stroke (18%) in case of a negative or missing MRI. Patients not meeting these criteria were excluded from the study.11 The diagnosis of FD was performed centrally blinded to clinical and imaging data and was based on both biochemical markers indicating a reduced activity of the enzyme α galactosidase A by increased levels of globotriaosylceramide and genetic testing for mutations in the α-galactosidase gene. Definite FD was diagnosed if a given mutation significantly reduced the enzyme activity and was a known causative mutation, and a significant increase in at least 2 independent biochemical markers was detected (globotriaosylceramide) in blood (>4.0 mg/L), lyso-globotriaosylceramide (>0.5 ng/mL), and globotriaosylceramide-C24 in urine (>35 mg/L). Probable FD was diagnosed in patients carrying either the mutation D313Y or a complex intronic haplotype15 in conjunction with an increase of at least 2 of the above-mentioned biochemical markers. A detailed description of used methodologies is provided elsewhere.11 MRI of the brain had to be performed where possible with the center-specific set of sequences and the study protocol recommended T2/PD-weighted or fluid-attenuated inversion recovery images and a diffusion-weighted imaging sequence as a minimum. Investigators were also asked to obtain T1- and T2*-weighted sequences if possible, however these sequences were performed only in a subset of patients. The MRI data were collected and interpreted centrally in a predefined and standardized manner.12,13 For current analysis, we included only those individuals in whom MRI examinations had been obtained with a quality sufficient for interpretation and which contained at least T2/PD-weighted or fluid-attenuated inversion recovery images. Demographic information (age and sex) was taken from the mandatory core data set collected by all centers.12 We also collected additional clinical data on those patients who showed hyperintensity of the pulvinar thalami on T1-weighted MRI but did not have FD.
Standard Protocol Approvals, Patient Consents, and Registrations
The study was conducted according to the Declaration of Helsinki and approved by the local ethics committee of the lead study center at the University of Rostock and by the regional review boards of each individual study center. All patients or their guardians gave written informed consent. The SIFAP 1 study (http://www.sifap.eu) has been registered at http://clinicaltrials.gov/ct2/show/NCT00414583, clinical trial identifier number: NCT00414583.
MRI scans were interpreted in a standardized fashion by 3 experienced readers (C.E., F.F., and R.S.) blinded to clinical and demographic data. They recorded the presence of acute and old cerebral hemorrhage or ischemic infarcts including infarct type, infarct location, and vascular territories involved among others.13 Old infarcts were defined as areas of T2 hyperintensity with associated focal atrophy or containing fluid-filled spaces. White matter hyperintensities (WMH) were defined as lesions with high signal intensity on T2-weighted images in the absence of evidence for complete tissue destruction14 and were rated according to the Fazekas scale as deep WMH (0=absent; 1=punctate; 2=early confluent; and 3=confluent) and periventricular WMH (0=absent; 1=pencil-thin lining; 2=halo of ≥5 mm thickness; 3=irregular WMH extending into deep white matter).15 The presence of old microbleeds was identified on gradient-echo T2*-weighted images as previously reported.16 Hyperintensity of the pulvinar was specifically searched for on all T1-weighted images available.10 The tortuosity of the basilar artery (BA) was rated as none, mild (some tortuosity of BA with a deviation from midline >5 to ≤10 mm), moderate (deviation of BA from midline >1 cm or diameter >5 mm), and severe (tortuosity with impression of brain stem or diameter >10 mm). The BA artery diameter was also directly measured at its maximum on axial scans. The inter-rater reliability of MRI interpretations was assessed and indicated substantial to excellent agreement.13 We also rereviewed all MRI scans of patients identified as definite or probable FD on potential hyperintensity of the pulvinar because subtle increases in signal intensity on T1-weighted images might have been overlooked, as well as those scans with such a finding but no diagnosis of FD.
Descriptive analyses were performed, and we compared demographic characteristics and MRI findings of patients identified by genetic and laboratory investigations as having definite or possible FD with those of the remaining patients. Comparisons were also made according to the diagnostic certainty of FD. P values were calculated using mixed-effects models to account for the center heterogeneity. We used linear mixed models for continuous, normally distributed measures, logistic models with random effects for dichotomous measures and ordinal models with random effects in case of ordinal variables. All statistical analyses were conducted using IBM SPSS Statistics 22.0 (SPSS, Inc, IBM Company, Chicago, IL, 2010) and STATA IC 13.1 (Stata Statistical Software: Release 13; StataCorp, College Station, TX, 2013: StataCorp LP). A 2-sided P<0.05 was considered statistically significant. Analyses were not corrected for multiple testing.
A total of 3203 (63.8% of the study sample) patients fulfilled the criteria for inclusion into this study, ie, this MRI had a quality sufficient for interpretation with T2/PD-weighted or fluid-attenuated inversion recovery images available. Among those were 34 patients with a diagnosis of FD that was definite in 21 and probable in 13 cases according to previously reported criteria.11 The median age was slightly higher in non-FD patients (46 versus 45 years) and women were more frequent in the FD group than among non-FD patients (70.6% versus 40.7%; P<0.001; Table 1).
When comparing the results of the blinded MRI analysis between patients with and without FD, no significant differences emerged (Table 1). Specifically, a similar rate of infarcts in the posterior circulation was noted in patients with and without FD for both acute and old lesions. There was a somewhat higher proportion of early confluent to confluent deep WMH (23.6% versus 12.5%) and of irregular periventricular WMH extending into the deep white matter (11.8% versus 3.8%) in FD than in non-FD patients, but the overall differences in WMH severity between groups were not statistically significant (Table 1). The proportion of patients with a hematoma or microbleeds, rated severity of tortuosity of the BA, and the maximum diameter of the BA were comparable between both groups. Finally, hyperintensity of the pulvinar thalami was not seen in any of 23 patients with FD in whom a T1-weighted MRI was available but was noted in 6 of 1398 (0.3%) non-FD patients. These numbers did not change after rereview of the respective T1-weighted images of both patients with FD and patients with a T1 pulvinar hyperintensity but no diagnosis of FD.
To explore whether diagnostic certainty might have an impact on the presence of certain MRI abnormalities, we also performed a comparison between patients with definite and probable FD (Table 2). Here, we also looked at the direct location of acute infarcts. However, no significant differences emerged between both groups.
Subsequent analysis of patients with a T1 pulvinar hyperintensity but no diagnosis of FD showed additional involvement of the basal ganglia in 3 of the 6 patients. Possible reasons for the observed signal changes in the pulvinar were systemic lupus erythematosus (n=1) and suspected mitochondriopathy (n=1), whereas no obvious explanation emerged in the remaining 4.
FD is a well-recognized but infrequent cause for stroke especially in the young,17 which may for long escape detection because of its variability in clinical presentation. It would, therefore, be desirable to have some clues that should alert for this disorder. Along these lines, certain MRI features of the brain like extensive WMH, infarcts in the posterior circulation, dilatation of the vertebrobasilar vessels, and high signal intensity of the pulvinar thalami on T1-weighted MRI have been suggested as characteristic for FD.5 Our findings in a large study population of young patients with an acute CVE do not support this expectation. This is an important information because it indicates that the diagnostic consideration of FD must not rely on a certain type or pattern of MRI abnormalities.
There are several possible explanations for our negative findings. Descriptions of high signal intensity in the pulvinar thalami on T1-weighted images as a pathognomonic MRI sign of FD come from patients who mostly had a longstanding diagnosis of this disorder9,10 Therefore, many of these patients are likely to have had multiorgan involvement including secondary manifestations like hypertension and renal failure. This certainly contrasts with the investigated population who consisted of nonselected young individuals with an acute CVE and in whom FD was diagnosed based on sophisticated genetic and laboratory examinations rather than a typical clinical presentation.11 As a consequence, FD was probably detected at a stage that might not yet have allowed for diffuse changes in the pulvinar to develop. This would fit with the observation of an increasing rate of the pulvinar sign with age and probably the duration of the disease.10 Furthermore, this precluded significant organ manifestations that per se may cause calcification of the deep brain nuclei.18
About the other radiological findings that have been previously associated with FD such as extensive WMH, infarctions in the posterior circulation, and dilatation of the vertebrobasilar arteries long-standing disease together with the impact of coexisting disorders such as hypertension are also likely to have played a role. Furthermore, reported findings were compared with the normal population rather than to a large group of patients examined for the same clinical presentation, ie, an acute CVE. However, this is necessary for defining differences and their significance within diseased populations. Thus, the SIFAP study has shown that the rate of posterior circulation infarction overall is unexpectedly high in the young, which makes this finding in FD less specific.13
Pulvinar hyperintensity on T1-weighted MRI was observed in the absence of FD in 6 (0.3%) individuals and thus may be considered a false-positive finding. In this context, it is important to note that several disorders are known to cause or to be associated with such signal abnormality including idiopathic strio-pallido-dentate calcinosis or Fahr disease.18 A possible cause for this abnormality was present in only 2 of these patients, but a complete work-up for possible causes of T1 signal abnormality of the deep gray nuclei was not a part of SIFAP 1.18
Limitations of our study come from the fact that our study population contains only a relatively small number of patients with FD and that the MRI protocol was not uniform among centers, which reflects clinical practice. Thus, some sequences specific for certain cerebral abnormalities like the pulvinar hyperintensity or for hemosiderin deposits were not available on all patients. Furthermore, it may be argued that only somewhat more than half of patients with FD had a definite diagnosis although we observed no difference in the type and pattern of MRI abnormalities according to diagnostic certainty. Again small numbers preclude a firm conclusion on this. In addition, we cannot tell to what extent minor clinical abnormalities possibly suggestive of FD existed in these patients because we did not screen for those in a prospective manner. Certainly, none of included patients had a multiorgan involvement suggestive of FD. We also did not perform refined measurements on the diameter of the BA, which have shown group differences between FD, healthy controls, and patients with stroke.5,19 However, these results were derived from relatively small surveys and the fact that the mean maximal BA in our series is comparable with that reported for patients with FD in the literature suggests that BA damage in other young patients with stroke may have been underestimated before. Finally, all sorts of bias need to be considered because this analysis could not be based on the entire data set, which in itself may already be biased. For the latter aspect, we have collected evidence that the group of patients included in SIFAP 1 was representative of young patients admitted to participating centers12 and of young patients with stroke overall when compared with a regional stroke registry.20 About selection for present study, which was driven primarily by availability and quality of MRI, it is likely that we have excluded more patients from centers with limited access to MRI and who were severely affected by their CVE. However, it is unlikely that this would have influenced our findings in a systematic manner. Strengths of our investigation come from the large and unselected collection of individuals with a common clinical presentation and from a careful prospective and blinded MRI interpretation. Stored MRI data also allowed to reinterpret all scans of patients diagnosed with FD after the study had been ended. Thus, we can exclude that low sensitivity reading may have caused our overall negative findings, but it has to be noted that a comprehensive MR protocol including also T1-weighted MRI was obtained in only half of all individuals available for this analysis.
In conclusion, our findings do not contradict previous observations of specific MRI features including T1-weighted hyperintensity of the pulvinar thalami on MRI in patients with a clinical spectrum of FD. However, they indicate that these features fail to identify patients with FD where the classical spectrum of FD has not developed. Thus, genetic and laboratory testing and consideration of other organ abnormalities are imperative to search for an association of an acute CVE with FD, whereas MRI findings are largely noncontributory.
We thank all centers and all patients who contributed to the study (see online-only Data Supplement). All authors contributed to study concept, data collection, and critical revision of the article. Dr Fazekas drafted the article. Drs Fazekas, Enzinger, and Schmidt reviewed the MRI scans. Drs Grittner and Martus performed the statistical analysis.
Sources of Funding
The Stroke in Young Fabry Patients 1 (SIFAP 1) study (http://www.sifap.eu; NCT00414583) has been supported by an unrestricted scientific grant from Shire Human Genetic Therapies to the University of Rostock.
Dr Fazekas serves on scientific advisory boards for Bayer-Schering, Biogen Idec, Genzyme, Merck-Serono, Pfizer, Novartis, and Teva Pharmaceutical Industries Ltd; serves on the editorial boards of Cerebrovascular Diseases, Multiple Sclerosis, the Polish Journal of Neurology and Neurosurgery, Stroke, and the Swiss Archives of Neurology and Psychiatry; and has received speaker honoraria and support from Biogen Idec, Bayer Schering, Merck-Serono, Novartis, Pfizer, Sanofi-Aventis, Shire and Teva Pharmaceutical Industries Ltd. Dr Enzinger is board member of the Austrain Science Fund; serves on scientific advisory boards for Bayer-Schering, Biogen Idec, Genzyme, Merck-Serono, Novartis, and Pharmaceutical Industries Ltd; has received travel grants and speaker honoraria from Bayer-Schering, Biogen-Idec, Teva-Aventis, Merck-Serono; and has received unrestricted research grants from Teva-Aventis, Biogen-Idec and Merck-Serono. Dr Schmidt received honoraria for consultancy from Pfizer and Axon Neuroscience and for lectures from Pfizer, Novartis, Merz Austria, Takeda, and Genericon. Dr Hennerici served on scientific advisory boards for Bayer, Pfizer, and Boehringer and is editor/consulting editor of Cerebrovascular Diseases, Interventional Neurology and International Stroke Journal. Dr Huber has received speaker honoraria from Genzyme. Dr Jungehulsing has received funding from the German Ministry for Education and Research for the fund Center for Stroke Research Berlin; serves on critical event committees of Edvard Life Sciences; has received speaker honoraria from Bayer, Pfizer, and Genzyme and consulting honoraria from Cipio Partners and Elron Electronic Industries. Dr Kaps serves on advisory board for Daiichi-Sankyo. Dr Kessler has received speaker honoraria from Bristol Meyer Squibb, Pfizer and Boehringer-Ingelheim. Dr Putaala has received speaker honoraria from Genzyme and research grants for related research by the Finnish Medical Foundation and Helsinki University Hospital. Dr Tatlisumak has served on scientific advisory boards for Boehringer-Ingelheim, Mitsubishi Pharma, Bayer, and Pfizer; serves/has served as a consultant to Boehringer-Ingelheim, PhotoThera, BrainsGate, Schering Plough, H. Lundbeck A/S, Sanofi-Aventis, Orion Pharma, Bayer, Pfizer, and Concentric Medical; has/has had research contracts with Boehringer-Ingelheim, PhotoThera, BrainsGate, Schering Plough, H. Lundbeck A/S, Sanofi-Aventis, Concentric Medical, Mitsubishi Pharma, Portola Pharmaceuticals, Bayer, and Pfizer. Dr Norrving serves on the editorial boards of International Journal of Stroke, Neuroepidemiology and Stroke, and has received speaker honoraria from Bayer and Daiichi-Sankyo. Dr Rolfs has received speaker honoraria and support from Shire, Biomarin, Genzyme, and Actelion and serves on Advisory board of Biomarin and Shire.
Guest Editor for this article was Tatjana Rundek, MD, PhD.
The online-only Data Supplement is available with this article at http://stroke.ahajournals.org/lookup/suppl/doi:10.1161/STROKEAHA.114.008548/-/DC1.
- Received December 22, 2014.
- Revision received March 6, 2015.
- Accepted March 16, 2015.
- © 2015 American Heart Association, Inc.
- Garzuly F,
- Maródi L,
- Erdös M,
- Grubits J,
- Varga Z,
- Gelpi E,
- et al
- Takanashi J,
- Barkovich AJ,
- Dillon WP,
- Sherr EH,
- Hart KA,
- Packman S
- Moore DF,
- Ye F,
- Schiffmann R,
- Butman JA
- Rolfs A,
- Fazekas F,
- Grittner U,
- Dichgans M,
- Martus P,
- Holzhausen M,
- et al
- Rolfs A,
- Martus P,
- Heuschmann PU,
- Grittner U,
- Holzhausen M,
- Tatlisumak T,
- et al
- Roob G,
- Lechner A,
- Schmidt R,
- Flooh E,
- Hartung HP,
- Fazekas F
- Tanislav C,
- Grittner U,
- Misselwitz B,
- Jungehuelsing GJ,
- Enzinger C,
- von Sarnowski B,
- et al