Fabry disease (OMIM 301500) is an X-linked lysosomal storage disorder caused by mutations in the GLA gene, which is located on Xq22.1 and codes for the enzyme alpha-galactosidase A (α-Gal A). Deficiency in the enzyme activity of α-Gal A leads to the lysosomal accumulation of globotriaosylceramide (Gb3) and related glycosphingolipids, causing multiorgan damage.1

Fabry disease is a rare condition, estimated to occur in 1 in 3100 hemizygous males.2 The diagnosis in males is based on biochemical determination of α-Gal A activity. In heterozygous females, α-Gal A activity may be normal, so diagnosis requires molecular analysis of the GLA gene.1 Clinical manifestations may arise early, in childhood or adolescence, and include acroparesthesias, cornea verticillata, angiokeratomas and microalbuminuria or proteinuria. In adulthood, patients may develop sensorineural deafness, renal failure, cerebrovascular events and cardiac manifestations.1

Fabry disease affects the heart, causing left ventricular hypertrophy (LVH) and fibrosis, heart failure, angina, valve thickening or regurgitation, dysrhythmias, cardiac conduction abnormalities and sudden death.1,3 The main cardiac manifestation is hypertrophic cardiomyopathy, which occurs in 43% of male patients (mean age of onset 39±10 years) and 26% of female patients (mean age of onset 50±11 years).4 LVH in Fabry disease is usually non-obstructive and symmetrical3 with prominent papillary muscles5 and, in nearly half of cases, myocardial late gadolinium enhancement is typically observed in the mid-wall of the inferolateral basal segments of the left ventricle.6

Left ventricular noncompaction (LVNC) is morphologically characterized by prominent left ventricular (LV) trabeculations and deep intertrabecular recesses communicating with the LV cavity, located in the apical and lateral regions of the left ventricle, distal to the papillary muscles. The myocardium has a two-layer appearance, composed of a thin compacted epicardial layer and a thick spongy noncompacted endocardial layer.7

LVNC has been found in 0.045% of adult individuals referred for echocardiography8 and 0.029% of children admitted for heart failure due to cardiomyopathy.9 It predominantly affects males8 and individuals of African origin.10 First described on autopsy in 1969 by Feldt et al.11 and on echocardiography by Engberding et al. in 1984,12 it is most commonly diagnosed by transthoracic echocardiography or cardiac magnetic resonance imaging (MRI). Several diagnostic criteria have been proposed using echocardiography or cardiac MRI13–19 (Table 1).

Patients with LVNC may be asymptomatic or develop complications such as heart failure, thromboembolism or ventricular dysrhythmias and sudden death.20

The etiology and pathogenesis of LVNC are unknown. It has been hypothesized to result from a disturbance of the compaction process of the LV myocardium during early development, but growing evidence suggests that, although the papillary muscles are formed by compaction of pre-existing trabeculations, these trabeculations contribute little, if anything, to the compacted layer of the ventricular walls.21 Moreover, not all cases are congenital and LVNC may be acquired at any stage of life after birth.7

LVNC is characterized by marked genetic heterogeneity. Mutations have been described in genes that code for proteins in the mitochondria (tafazzins), Z line (ZASP/Cypher), cytoskeleton (alpha-dystrobrevin), nuclear lamina (lamin A/C) and sarcomere (beta-myosin heavy chain, alpha-cardiac actin and cardiac troponin T).7 However, the causal relationship between these mutations and LVNC remains to be established.

LVNC has been reported in patients with congenital cardiac abnormalities, such as LV outflow tract abnormalities, Ebstein anomaly, tetralogy of Fallot and double outlet right ventricle; chromosomal defects, such as trisomy 13, Cornelia de Lange, DiGeorge, Turner and 5q minus syndromes; and neuromuscular disorders, such as mitochondrial myopathies, Barth syndrome, zaspopathy, myotonic dystrophy type 1, dystrobrevinopathy, and Emery-Dreifuss muscular dystrophy.7 LVNC has also been reported in patients with storage disorders such as glycogen storage disease types Ib and IV,22,23 Danon disease24 and amyloidosis.25

In 2011, we reported the first case of LVNC in a Fabry disease patient.26 This was a young asymptomatic female who fulfilled the Jenni echocardiographic criteria and the Petersen cardiac MRI criteria for LVNC. She had been diagnosed with Fabry disease following family screening, and the pathogenic mutation of the GLA gene detected was a missense mutation (p.F113L) that is known to be associated with a predominantly cardiac later-onset phenotype. However, thorough diagnostic study revealed none of the known clinical manifestations of Fabry disease, including LVH, late gadolinium enhancement or other cardiac or extracardiac manifestations, nor did she present LV dilation or dysfunction or any clinical complication of LVNC, such as heart failure, dysrhythmias or thromboembolism. Moreover, all of her relatives with Fabry disease underwent echocardiography and cardiac MRI and none presented LVNC. In this report, we pointed out the limitations of the current diagnostic criteria of LVNC and raised the question whether this was an overdiagnosis of LVNC, a coincidental finding of LVNC and Fabry disease, or if indeed the two conditions were clinically related.26,27

More recently, two additional cases of LVNC in Fabry disease patients have been reported,28–31 also in young females. In the second case, the diagnosis of LVNC was based only on cardiac MRI (Petersen criteria), without clear echocardiographic criteria of LVNC due to insufficient quality of the acoustic window. She had classical Fabry disease due to a nonsense GLA gene mutation (p.R220X). As in our case, she also did not present LVH, dilation, dysfunction, late gadolinium enhancement or other cardiac abnormalities, and there was no history of heart failure, dysrhythmias or thromboembolism. There was also no family history of LVNC. She had no clinical manifestations of Fabry disease besides cornea verticillata and angiokeratomas. An endomyocardial biopsy showed deposition of glycosphingolipids in cardiomyocytes and based on this finding the patient was referred for enzyme replacement therapy.28,29

In the third case, the diagnosis of LVNC was based on cardiac MRI (Petersen criteria) and echocardiography (although it was not reported which specific criteria were fulfilled). She was diagnosed with Fabry disease based on low levels of α-Gal A, but the pathogenic GLA gene mutation was never found. The only clinical manifestations of Fabry disease were angiokeratomas and acroparesthesias. As in the two previous cases, there was no LVH, dilation or dysfunction. Cardiac MRI showed subendocardial late gadolinium enhancement, but its location and extension was not reported. Interestingly, she had Wolff-Parkinson-White syndrome, interpreted by the authors as an unrelated finding, and underwent ablation of a para-Hisian accessory pathway, which was followed by disappearance of the short PR interval and delta waves. However, as in the two previous cases, she had no history of heart failure or thromboembolism. Moreover, there were no available data consistent with a family history of LVNC or Fabry disease.30,31

Taken together, it remains essential to clarify whether LVNC could be a cardiac manifestation of Fabry disease, given the implications for therapeutic decisions on whether to start enzyme replacement or chaperone therapy, particularly in young asymptomatic females with no other known clinical manifestations of Fabry disease.

The only published study to date on Fabry disease screening in patients with LVNC included 26 male patients who fulfilled the Stöllberger echocardiographic criteria for LVNC. Fabry disease screening was based on the enzymatic assay of α-Gal A and found no patients with the disease.32

Herein, we report the results of Fabry disease screening in a larger cohort of patients with LVNC.

Patients with LVNC were predominantly males (62.8%). Only four patients (5.1%) were identified as being of African origin. Mean age at diagnosis of LVNC was 64.4±14.1 years. A family history of LVNC was recorded in five patients (6.4%). The patients did not present concomitant neuromuscular disorders or any other medical conditions reported to be associated with LVNC.

The diagnosis of LVNC was established by echocardiography in 91.0% of the cases, based on the Jenni criteria in 84.6%, the Stöllberger criteria in 46.2%, the Chin criteria in 21.8% and the three echocardiographic criteria in 20.5% of the cases. Nearly half of the patients presented LV systolic dysfunction (48.7%) and mean LV ejection fraction was 47.2±15.2%. Cardiac MRI was performed in 68 patients (87.2%) and the diagnosis of LVNC was established in 58 patients (85.3%), mainly based on the Petersen criteria (83.8%). The trabeculated LV mass was calculated in only 11 patients, all of them fulfilling the Jacquier criteria of LVNC. The presence of the Stacey or Captur criteria was not assessed in any cardiac MRI. The diagnosis of LVNC was established by both echocardiography and cardiac MRI in 53 of the 68 patients who underwent cardiac MRI (77.9%).

LVNC was an incidental finding in 44.9% of the patients. Heart failure was the most common clinical manifestation at the time of diagnosis (34.6%). In the follow-up (53±33 months), heart failure remained the most frequent clinical complication (60.3%), with New York Heart Association functional class ranging from I to IV (I: 8.5%; II: 59.6%; III: 23.4%; IV: 8.5%). Dysrhythmias were also common (35.9%). Twenty-four-hour Holter monitoring was performed in 74 patients (94.9%) and revealed atrial fibrillation in 10 patients (13.5%) and ventricular tachycardia (VT) in 16 patients (21.6%), although only one case exhibited sustained VT. Thromboembolism occurred in nine patients (11.5%), as stroke in eight and as myocardial infarction in one (Table 2).

Twenty-eight patients (35.9%) had undergone genetic study that included a wide range of genes (MYH7: 32.1%; MYBPC3: 30.8%; TNNT2: 28.2%; TNNI3: 19.2%; ACTC1: 28.2%; TPM1: 28.2%; MYL2: 11.5%; MYL3: 11.5%; DTNA: 7.7%; DMD: 10.3%; LDB3: 28.2%; TAZ: 26.9%; LMNA: 24.4%; CSRP3: 24.4%; TCAP: 20.5%; SGCD: 21.8%; PLN: 23.1%; TTN: 2.6%; SCN5A: 1.3%). A pathogenic mutation was found in the MYH7 gene in one patient (c.1003G>C) (unpublished data). Eight genetic variants of uncertain significance were found in six cases: four in MYBPC3 (c.1505G>A; c.2762A>T; c.1624G>C; c.3676C>T), the last two in the same patient; two in TTN (c.16425T>G; c.60259C>G); one in SCN5A (c.393-5C>A); and one in SGCD (c.772-774delAAG) in the same patient carrying two genetic variants of the MYBPC3 gene.

The screening of 78 patients with LVNC revealed no additional cases of Fabry disease besides the one already identified.26 This Fabry patient was never under enzyme replacement or chaperone therapy. During the eight-year follow-up, this patient has remained asymptomatic and without known clinical manifestation of Fabry disease or complication of LVNC. Moreover, there were no significant changes regarding the location or extension of LVNC or LV chamber size, mass or function during follow-up (Figure 1).

Figure 1.

Left ventricular noncompaction in a Fabry disease patient: cardiac magnetic resonance imaging showing hypertrabeculation mostly affecting the left ventricular apical region and lateral wall in apical 4-chamber view.

(0.03MB).

This multicenter study of Fabry disease screening in 78 consecutive patients with LVNC revealed no additional Fabry patients besides the single case that had been previously reported by our center, which argues against the hypothesis that LVNC is a clinical manifestation of Fabry disease.

The most likely explanation for the finding of LVNC in this Fabry patient is an overdiagnosis of LVNC. Studies have increasingly exposed the lack of specificity of the current diagnostic criteria for LVNC, potentially leading to overdiagnosis.10,33,34 Moreover, there is growing evidence that LVNC has no independent prognostic impact compared to conventional risk factors, such as LV dilation, dysfunction or fibrosis, and that patients with LVNC without LV dilation, dysfunction or fibrosis have an excellent prognosis.34–36

In a study of 199 patients referred to a heart failure clinic and 60 healthy controls, Kohli et al. found that 23.6% of the patients and 8.3% of the controls fulfilled one or more echocardiographic criteria for LVNC. Among the heart failure patients with LVNC criteria, 29.8% fulfilled all three echocardiographic criteria and 36.2% fulfilled only a single criterion, with the frequency of each criterion ranging from 12.1% to 18.6%.10 In a large population-based cohort of 1000 subjects from the Multi-Ethnic Study of Atherosclerosis (MESA) registry who underwent cardiac MRI, the Petersen criteria for LVNC were fulfilled in 43% of the subjects without cardiac disease.33 Another study based on the MESA registry including 2742 subjects undergoing cardiac MRI showed that a greater extent of LV trabeculations in asymptomatic individuals appears to be benign and is not associated with deterioration in LV volumes or function over an almost 10-year period.36 In a study including 700 patients referred for cardiac MRI, Ivanov et al. found an LVNC prevalence of 39%, 23%, 25% and 3%, according to the Petersen, Stacey, Jacquier and Captur criteria, respectively. In this study, during a median follow-up of seven years, there were no statistically significant differences in the assessed outcomes (death, ischemic stroke, ventricular tachycardia/fibrillation, and heart failure hospitalization) between patients with and without LVNC, irrespective of the criteria applied.34 Finally, in a study including 113 patients with LVNC according to the Jenni and Petersen criteria, Andreini et al. found that the degree of hypertrabeculation had no prognostic impact over and above LV dilation, dysfunction or late gadolinium enhancement. Indeed, patients without LV dilation, dysfunction or fibrosis had an excellent prognosis at four-year follow-up.35

In light of this evidence, the possibility of an overdiagnosis of LVNC, with no associated prognostic impact, must be considered in these three cases of Fabry disease published in the literature. In fact, all three case reports of LVNC occurred in young females with no family history of LVNC in Fabry or non-Fabry relatives, no LV hypertrophy, dilation or dysfunction, no late gadolinium enhancement suggestive of Fabry disease involvement, no or few clinical manifestations of Fabry disease, and no clinical complications of LVNC.26–31

Appropriate follow-up would also be required to correctly interpret the significance of LVNC in these Fabry patients. In our case, the patient was never under enzyme replacement or chaperone therapy, because at baseline she did not present any known manifestation of Fabry disease that would constitute a clinical indication to start specific therapy according to the guidelines. During the eight-year follow-up of this patient, no significant change has been documented in the location or extension of LVNC or in LV chamber size, mass or function. Moreover, this patient has remained asymptomatic and with no known clinical manifestation of Fabry disease or complication of LVNC, which supports the hypothesis of an overdiagnosis of LVNC in this Fabry patient.

Another possibility is that LVNC and Fabry disease in these patients are coincidental and unrelated conditions. Although the simultaneous finding of two unrelated rare conditions may seem unlikely, it is now known that LVNC has a much higher prevalence10,33,34 than the previous estimate of 0.045%8 and the currently estimated Fabry disease prevalence of 1 in 31002 is much higher than the previous estimate of 1 in 117 000.37

The finding of LVNC in a wide variety of congenital cardiac abnormalities, chromosomal defects, neuromuscular disorders7 and storage disorders,22–24 as well as in elite athletes38 and pregnant women,39 has raised the question of whether LVNC is indeed a distinct cardiomyopathy or simply a common epiphenomenon in cardiac overload conditions. In our case, this young woman was not an athlete, but she had been pregnant once, although we have no record of echocardiograms before, during or just after pregnancy. A history of pregnancy or vigorous exercise was not reported in the other two cases.

Finally, there is the possibility that LVNC could be a cardiac manifestation of Fabry disease. However, there is no evidence that Fabry disease causes LVNC by either compensatory or genetic mechanisms. Indeed, although they had Fabry disease, none of these patients had concomitant LVH, dilation or dysfunction that could have led to cardiac overload and hence to LVNC. Furthermore, there is no evidence that the GLA gene mutation, the mutated α-Gal A, or the intramyocardial deposition of glycosphingolipids due to Fabry disease are the cause of or are in any way related to the development of LVNC, and any assumption on this point remains speculative. Lysosomal accumulation of Gb3 begins in utero and has been demonstrated as early as the 17th week of gestation in various fetal cell types, such as cardiomyocytes,40 kidney podocytes,40–42 hepatocytes,41,43 corneal epithelial cells44 and myenteric plexus cells.42 Although Gb3 storage begins in utero, clinical manifestations only develop several years later. Therefore, Gb3 deposits are expected findings in biopsies, including endomyocardial biopsies, of all Fabry patients independently of the presence of clinical manifestations, and such deposition should not be interpreted as evidence that LVNC is due to Fabry disease.

The only study screening for Fabry disease in patients with LVNC did not reveal any positive case. However, this was a small cohort of 26 male patients and the only criteria for LVNC diagnosis were the Stöllberger echocardiographic criteria.32 Our study, on the other hand, included a larger cohort of patients and the inclusion criteria allowed (LVNC) diagnosis based on one or more criteria, by echocardiography or cardiac MRI. Our cohort of patients appears to be representative of LVNC and its wide clinical spectrum. Although a significant proportion of cases were incidental findings, there was good agreement between echocardiography and cardiac MRI regarding the diagnosis of LVNC, nearly half of the patients presented LV systolic dysfunction, and in the follow-up most patients developed complications. Even so, no Fabry disease patient was diagnosed besides the single case already reported in the literature. Nevertheless, our study is also limited by its small sample size, which may have prevented the detection of Fabry disease if LVNC was a rare manifestation of this disease. However, the common finding of LVNC in other patient populations, such as heart failure, also argues against the hypothesis of LVNC being a rare clinical manifestation of Fabry disease. The fact that our patient remained stable and did not develop any complications during an eight-year follow-up without specific therapy for Fabry disease also argues against the hypothesis that LVNC is a clinical manifestation or a risk marker for the development of other cardiac manifestations of Fabry disease.

Based on recent evidence exposing the low specificity of the current diagnostic criteria and the increasing rate of overdiagnosis of LVNC,10,33,34 it is plausible that more cases of LVNC will be diagnosed among Fabry patients, raising concerns about its therapeutic implications. This is particularly important in Fabry females, who, in part due to the inactivation pattern of the X chromosome, present a highly variable prognosis, some of them exhibiting a benign disease course and a lifelong prognosis without significant clinical manifestations in the absence of specific therapy.45

In summary, our study found no evidence to support the hypothesis that LVNC is a clinical manifestation of Fabry disease. Taken together with the available evidence, our results do not support the use of the current diagnostic criteria of LVNC for the decision to start specific therapy for Fabry disease, either enzyme replacement or chaperone therapy, in young asymptomatic females presenting LVNC criteria in the absence of other cardiac abnormalities or known manifestations of Fabry disease that would constitute a clinical indication to start therapy according to the current guidelines.

Olga Azevedo, MD, and Gabriel Miltenberger-Miltenyi, PhD, have received educational and/or research grants from Shire Human Genetic Therapies, Inc., Genzyme Corporation and Amicus Nuno Marques, MD, Hugo Antunes, MD, and Liliana Reis, MD, have received educational grants from Shire Human Genetic Therapies, Inc. and Genzyme Corporation. Nuno Craveiro, MD, Rui Azevedo Guerreiro, MD, and Rui Pontes dos Santos, MD, have received educational grants from Shire Human Genetic Therapies, Inc. The other authors have no disclosures.

This research was supported by a research grant from Shire Human Genetic Therapies, Inc.

Shire Human Genetic Therapies, Inc. had no influence on the design and conduct of the study, in the collection, analysis, and interpretation of the data, or in the preparation, review, or approval of the manuscript.

The authors have no conflicts of interest to declare.