Prevalence of Fabry Disease in Patients With Cryptogenic Strokes: A Systematic Review

Fabry disease (FD) is an X-linked disorder involving multiple organs. Stroke is a serious and frequent complication of FD. Cryptogenic stroke is a common presentation of FD, especially in the young population. The etiology of cryptogenic stroke is highly variable and difficult to assess, frequently leaving patients without a primary diagnosis. We conducted a systematic review to investigate the pooled prevalence of FD among patients with cryptogenic stroke, or patients with FD in whom a stroke was the presenting condition. English-language studies involving humans published in the last 20 years were included in this systematic review. FD was more common in male patients and tended to present at an earlier age. The frequency of hemorrhagic and ischemic strokes in this population was similar to that in the general population. There was a high rate of stroke recurrence in the study sample, even among patients undergoing enzyme replacement therapy. We conclude that screening for FD in patients with cryptogenic stroke is low yield and not cost-effective. However, it may be worthwhile to screen for FD among patients with recurrent strokes.


Introduction And Background
Fabry disease (FD) is an X-linked lysosomal storage disorder caused by an inborn error in the glycosphingolipid metabolic pathway, with deficient activity of the lysosomal enzyme alpha-galactosidase A that results in lysosomal accumulation of globotriaosylceramide (Gb3) in multiple organs [1]. Gb3 mainly accumulates in the endothelial, cardiac, renal, and dorsal root ganglion neuronal cells [2]. The prevalence of FD is approximately 1 in 117,000 individuals [1]. The majority of cases are reported in Caucasians, and most patients who present with symptoms are male [2]. There are different types of FD mutations on the alphagalactosidase (GAL) gene which causes the disease, including point mutations (83.4%), missense mutations (50%), nonsense mutations (29.2%), frameshift mutations (16.6%), and splicing (4.2%) [1].
Suspicion of FD can come through family history, clinical manifestations, and lab abnormalities, with the diagnosis confirmed by biochemical and molecular genetic testing [3]. Neuroradiologic evaluation is done using T2-weighted MRI, fluid-attenuated inversion recovery (FLAIR), or via evidence of pulvinar hyperintensity on T1-weighted images. These lesions are present in one-third of patients and are very suggestive of the disease when the diagnosis is uncertain [4].
There are two types of FD, namely, classical and late-onset FD. The classical form occurs in hemizygous males or heterozygous females [5]. Classical FD initiates at a young age with the main symptoms including acroparesthesias, gastrointestinal disturbances, angiokeratomas, impaired sweating, and heat and cold intolerance [5]. Complications include hypertrophic cardiomyopathy, renal failure, and stroke [5]. Female heterozygotes are also symptomatic, especially when they have skewed X chromosome inactivation [5]. Patients with late-onset FD usually present in the third to the seventh decade of life. Most patients do not display the classical symptoms as just one organ may be affected. The level of enzyme activity ranges from 2% to 30% of the normal, and an accumulation of Gb3 on blood vessels may not be evident [6][7][8].
There is a dearth of large-scale studies on FD. Although many observational studies have described the prevalence of stroke and FD, most included small cohorts. To reduce bias and determine a more accurate record, a pooled analysis of observational studies is needed. In this review, we aim to determine the prevalence of FD in patients with cryptogenic stroke or first-time stroke. We further aim to consolidate the knowledge of cryptogenic stroke and FD, explore the pathophysiology of ischemic stroke in FD, investigate the use of enzyme replacement therapy (ERT) as a preventive measure for stroke recurrence, and explore the influence of gender among these patients.

Protocol
This systematic review was conducted following the Meta-Analysis of Observational Studies in Epidemiology (MOOSE) reporting guidelines [9].

Eligibility Criteria and Study Selection
Only observational studies conducted among human subjects in the last 25 years and written in the English language with the full report available were included in this review. We excluded case reports, systematic reviews, literature reviews, and meta-analyses that did not fulfill the study criteria. We included studies reporting on patients aged 18 to 60. Additionally, any articles that were not deemed relevant to the study parameters were excluded.
After screening the results, only articles with at least one of the following characteristics were included: patients with cryptogenic ischemic stroke or stroke presenting as the first symptom in whom measurement of either plasma levels of Gb3 with α-GAL-A gene sequencing, or α-GAL-A enzyme levels, was done. Patients with a pathogenic mutation were compared to those with a mutation of unknown pathogenicity. Our primary aim was to calculate the prevalence of FD in cryptogenic stroke. The secondary aim was to describe the characteristics of patients with cryptogenic stroke who were diagnosed with FD. We investigated the mutation type, stroke type, and prognosis of each of these patients.

Database and Search Strategy
For this systematic review, we used PubMed as a database. The search was done between July 25, 2020, and July 18, 2021. We used the following combination of search terms: "Fabry disease"[Title/Abstract] AND "stroke"[Title/Abstract]) OR ("cryptogenic stroke"[Title/Abstract] AND "Fabry disease"[Title/Abstract]. In addition to PubMed, Embase and Scopus were searched, but it did not result in any additional studies that fit the criteria.

Data Extraction
The following information was collected: author, year of publication, country of publication, study type, methodology, outcomes, and prevalence of FD in stroke patients. For all patients who presented with stroke and had a diagnosis of FD, the following additional information was extracted: type of mutation, Gb3 levels or plasma α-GAL-A levels, type of stroke, and prognosis.

Bias Analysis
To assess and minimize bias the risk of bias in non-randomized studies (ROBINS-I) tool for observational studies was used [10]. Figure 1 shows a flowchart with the results of this systematic review.

Study Characteristics and Outcomes
A total of 11 studies from 10 different countries were included. Probable and definitive diagnosis was assessed in the final prevalence among the observational studies. Table 1 shows the characteristics and outcomes of the cohorts investigated in this systematic review [1,2,[11][12][13][14][15][16][17][18][19].  We found 36 patients with cryptogenic stroke, or stroke as a presenting condition, in the 11 studies involving 1,986 patients in various cohort studies. The total prevalence of FD among these studies was 1.81%.  Among the stroke patients, there were 32 ischemic infarcts (86.48%) and five hemorrhagic infarcts (13.51%). The number of patients with re-infarctions was 16 (43.24%). The mean age of males was 37.24 and for females was 41.11 years. In general, Gb3 levels were normal or mildly elevated in patients for whom the value was available. Measurement of α-GAL-A activity was a more sensitive marker than Gb3 level. Among the cohorts, there was heterogeneity in the way the different studies described mutations. We included patients with pathogenic mutations, mutations of unknown pathogenicity, and neutral mutations with related pathogenicity. Table 3 shows the bias of the studies included in this systematic review [1,2,[11][12][13][14][15][16]19].

ROBINS-I: risk of bias in non-randomized studies of interventions
In five studies, the authors found that the sample was too small to reach statistical power [1,11,[14][15][16]. In the study by Kinoshita et al., genetic analysis was not conducted in each patient, which could lead to a measurement bias [14]. Lanthier et al. stated that they could have missed some FD cases as all patients did not undergo genetic analysis [11]. In the study by Alhazzaa et al., the inclusion criteria were somewhat restrictive and more than half of the cohort suffered regular vascular risk factors such as diabetes, hypertension, and/or dyslipidemia, which could have contributed to the stroke. Further, in the study by Brouns et al., genetic testing was not performed among women which could introduce bias as the enzyme levels in female FD patients are often normal (false negatives) [16]. Finally, this study did not include cases of lacunar infarcts which further limited their numbers [15].

Pathophysiology of Stroke in Patients With Fabry Disease
Most studies agree that the accumulation of glycosphingolipids in vascular endothelium causes disturbances in vasoreactivity and autoregulation, increasing the likelihood of stroke. Other theories have also been suggested in the literature. According to Sawada et al., vasospasm may play an important role in the pathophysiological mechanism of stroke in FD patients. This is supported by the high incidence of spastic coronary angina seen in FD [20]. Hilz et al. extensively studied stroke mechanisms and concluded that reduced cerebral blood flow velocities and impaired autoregulation are key factors [21]. A review by Rombach et al. suggested that smooth muscle cells are primarily involved in the vasculopathy of FD and that, in the early stages of Fabry vasculopathy, angiotensin II production becomes upregulated [22]. Further, other mechanisms such as increased reactive oxygen species production, enhanced nitric oxide levels, and the presence of a prothrombotic state have also been identified to contribute to the development of stroke [23]. Lastly, cardiac arrythmias, hypertrophic cardiomyopathy and myocardial infarction may lead to stroke in patients with FD [14].

Stroke in Patients with Fabry Disease and Prevalence of Fabry Disease in all Types of Strokes
FD affects both the peripheral nervous system and the central nervous system (CNS). The CNS symptoms in FD include psychiatric manifestations, headaches, and strokes [20]. In this systematic review, we only included information on cryptogenic strokes or patients with stroke as the first sign of presentation; therefore, all patients in our sample pool were, by definition, late-onset FD patients. Stroke is approximately 12 times more common in patients with FD than in the general population [1]. Approximately 25% of FD patients develop transient ischemic attack and stroke at a mean age of 40 (PM1) [24]. From the FD database (PM2), 138 out of 2,446 patients developed stroke [24]. The median age of first-time stroke was 39 years in males and 45.7 in females. Ischemic stroke was seen in 86.8% of the patients and hemorrhagic stroke in 13.9% of the patients, which is very similar to the general population [24]. In the same study, 70.4% of the strokes were lacunar [24]. Five of 52 females (8.6%) and 43 of 86 males (50%) who experienced stroke did not have the diagnosis of FD [25]. In comparison to our review, the rates of ischemic and hemorrhagic strokes were similar, and the age of presentation and the ratio of males to females were also similar. Table 4 lists the studies that included all types of strokes [25][26][27].
Author, year, country  Prevalence in studies that included all types of strokes ranged from 0% to 0.5% compared to our pooled analysis (0-3.88%), suggesting that the prevalence of FD is higher with cryptogenic strokes than other stroke types. The study by Rolfs et al. reported the highest prevalence among all the studies [26]. Most studies included only patients who developed a stroke for the first time. However, the study by Rolfs et al. also included patients with recurrent strokes, suggesting that the prevalence of FD among patients with recurrent stroke is higher compared to the general population.

Screening Fabry Disease in Patients With Ischemic Stroke
Early recognition of FD in patients with cryptogenic stroke is critical to prevent recurrent strokes [11]. However, most studies agree that generalized screening for FD among these patients is expensive and has low yield due to the low disease prevalence [2]. However, patients with recurrent strokes in some cohorts had higher rates of FD, suggesting that screening could be beneficial in this population [11,26].
Females tend to have normal levels of α-GAL-A enzyme activity due to the random inactivation of chromosome X, making gene sequencing more essential among females [12]. Measuring the levels of Gb3 has shown to be a more effective screening tool than α-GAL-A enzyme activity among females. However, measuring the enzyme activity is a more sensitive test in males [14]. Gündoğdu et al. suggested that screening of FD in countries with a high prevalence of genetic disease could be beneficial [1]. Moreover, recognizing neutral mutations is vital to ensure that patients do not go through unnecessary investigations and treatment [11]. Alhazzaa et al. also argued that screening for FD in patients with cryptogenic stroke is not cost-effective [15].

Management and Role of Enzyme Replacement Therapy in the Treatment of Fabry Disease
ERT has been shown to improve kidney function, reduce cardiomyopathy, and improve circulation by increasing NO levels. However, the prevention of stroke with ERT has not been demonstrated [20]. There is no consensus regarding when to initiate ERT in FD. Some experts suggest that it should be started among men as soon as the diagnosis is made and among women if clinical manifestations are present. The initial consensus suggests that ERT does not prevent stroke. However, a pooled analysis including seven cohort studies and two randomized controlled trials indicated that ERT has a beneficial effect in preventing stroke and stroke recurrence in patients with genetically, enzymatically, and/or biopsy-proven FD [28]. In one report, stroke recurrence rates were 0.082 in the ERT group versus 0.160 in the control group, with the difference being statistically significant [28]. The effects of ERT occurred gradually, but prevented strokes in the long term [29]. Sheng et al. suggested that women should consider initiating ERT at the time of FD diagnosis because stroke often occurs before the diagnosis is made [28].
Rolfs et al. argued that patients should be on a more effective antiplatelet medication such as clopidogrel or a combination of dipyridamole and aspirin [13]. Only two studies in our systematic review described patients who received ERT after stroke. In these studies, three patients received ERT but two developed stroke recurrence. We recommend further research to help determine whether ERT prevents recurrent strokes in patients with cryptogenic stroke due to FD.

Conclusions
Although the pathogenesis of stroke in FD remains unclear, various mechanisms such as vasospasm, reduced cerebral blood flow velocity, abnormal autoregulation and upregulation of angiotensin II, and accumulation of glycosphingolipids in vascular endothelium have been proposed.
The prevalence of cryptogenic stroke in FD is highly variable among studies. However, compared to single strokes, FD was seen more commonly in patients with recurrent strokes. Hence, it may be useful to perform FD screening in such cases. Screening should be done by determining α-GAL-A enzyme activity in males and plasma Gb3 levels in females, followed by α-GAL-A gene sequencing for confirmation.
Finally, we conclude that ERT may decrease the recurrence of stroke in FD patients in the long term. Despite the initial consensus that ERT does not prevent stroke, the difference was statistically significant in recent systematic reviews.

Conflicts of interest:
In compliance with the ICMJE uniform disclosure form, all authors declare the following: Payment/services info: All authors have declared that no financial support was received from any organization for the submitted work. Financial relationships: All authors have declared that they have no financial relationships at present or within the previous three years with any organizations that might have an interest in the submitted work. Other relationships: All authors have declared that there are no other relationships or activities that could appear to have influenced the submitted work.