Author + information
- Received April 5, 2002
- Revision received June 8, 2002
- Accepted July 12, 2002
- Published online November 6, 2002.
- Christoph Kampmann, MD*,* (, )
- Frank Baehner, MD*,
- Catharina Whybra*,
- Claudia Martin, MD*,
- Christiane M Wiethoff*,
- Markus Ries, MD*,
- Andreas Gal, MD, PhD† and
- Michael Beck, MD, PhD*
- ↵*Reprint requests and correspondence:
Dr. Chistoph Kampmann, Universitätskinderklinik, Johannes Gutenberg Universität, Langenbeckstr. 1, D-55131 Mainz, Germany.
Objectives We sought to define the prevalence of cardiac involvement in female patients with Anderson–Fabry disease (AFD).
Background Anderson–Fabry disease is a rare inborn X-linked lysosomal storage disorder. Globotriaosylceramide (Gb3), the major substrate of the deficient α-galactosidase A enzyme, accumulates progressively in vulnerable cells, including the cardiovascular system. It has been believed that heterozygous females have less cardiac involvement than hemizygous males with AFD.
Methods We performed two-dimensional echocardiographic examinations of female patients heterozygous for AFD.
Results Since 1997, a total of 55 female patients (mean age, 39.6 years; range, 6.1 to 70.8 years) with proven AFD have been investigated prospectively at our hospital. Of these, 13 (23.6%) had normal left ventricular (LV) geometry and LV mass (LVM). Seven patients (12.7%) had concentric remodeling, 29 patients (52.7%) concentric LV hypertrophy (LVH), and 6 patients (10.9%) eccentric LVH (2 with subaortic pressure gradients). There was a strong correlation between age and the severity of LVH (r2= 0.905; p < 0.0001), and all patients older than 45 years had LVH. With increasing LVM, there was a significant age-independent decrease in systolic and diastolic LV function. Mild thickening of the aortic valve leaflets was present in 25.5% of patients, with the same percentage demonstrating mild thickening of the mitral valve leaflets. Mild mitral valve prolapse was documented in 10.9% of patients.
Conclusions Cardiac involvement, with LVH and structural valve abnormalities, is very common and worsens with age in females who are heterozygous for AFD, and they should therefore be considered candidates for enzyme replacement therapy.
Anderson–Fabry disease (AFD) is a rare inborn deficiency of the lysosomal enzyme α-galactosidase A, causing progressive intracellular accumulation of globotriaosylceramide (Gb3), the major glycosphingolipid substrate. The incidence has been estimated to be 1:117,000 male births (1). It affects different cell types, including myocardial cells and dorsal root ganglia neuronal cells, as well as endothelial cells and intimal and smooth muscle cells of the vascular system (2,3). The intracellular deposits of Gb3in the heart are similar to those found in other tissues (4). Cardiac involvement in hemizygotes is frequent, resulting in hypertrophic changes of the myocardium, in conduction delays, and in the thickening of the valves (5–8). Cardiomyopathy can be the sole manifestation of AFD, and it has been demonstrated that 3% to 4% of all idiopathic hypertrophic cardiomyopathies may be caused by AFD (9). The prevalence of these cardiac manifestations has been described in hemizygote male AFD patients.
Because the genetic abnormalities in AFD are X-linked, clinical manifestations of AFD in female heterozygotes, as obligate carriers, have been considered to be rare or mild. Severe or serious clinical manifestations have been estimated to affect only 1% of heterozygous females (2). Recently published data, however, indicate that females are affected much more commonly than previously believed (10). Questionnaire studies of patient pedigrees could confirm these findings (11). The purpose of the present study was to evaluate the prevalence and severity of cardiac manifestations of AFD in females genetically proven to be heterozygous for AFD.
Based on a large pedigree examination of families with hemizygous male patients with documented AFD, all possible female heterozygotes were genetically and enzymatically examined. The inclusion criterion for this prospective study was a mutation of the α-galactosidase A gene confirming the diagnosis of AFD. All female AFD patients received a clinical examination, electrocardiography, blood pressure (BP) measurements, and detailed echocardiography with a digital Toshiba 380 SSA Power-Vision ultrasound system using appropriate transducers (5.0 MHz, 3.5 MHz, 2.5 MHz). The following measurements were made from M-mode tracings, according to the recommendations of the American Society of Echocardiography: thickness of the interventricular septum at end-diastole (IVSd) and at end-systole (IVSs); left ventricular (LV) internal cavity diameter at end-diastole (LVED) and at end-systole (LVES); posterior wall thickness at end-diastole (PWd) and at end-systole (PWs); end-diastolic aortic root diameter (AoD); and left atrial diameter at atrial systole (LAD), as described elsewhere (12).
Additionally, the thickness of the left atrium at atrial diastole (LAT) was measured. The LV mass (LVM) was calculated by the cube formula, modified by Devereux (13), and indexed to body surface area (BSA) and body height2.7. Left ventricular hypertrophy (LVH) was considered to be present if LVM indexed to height was >47 g/m2.7in individuals with a body mass index (BMI) below 26 kg/m2or if LVM indexed to height was >60 g/m2.7in individuals with a BMI above 26 kg/m2(13). Relative wall thickness (RWT) was calculated as [(IVSd + PWd)/LVED]. Concentric remodeling was defined as being present if RWT was 0.45 and LVM was normal.
Overall LV geometry was classified into one of four categories: 1) normal (normal LVM and normal RWT); 2) concentric remodeling (normal LVM and increased RWT); 3) eccentric LVH (increased LVM and septal-to-posterior wall thickness ratio of ≥1.5); and 4) concentric LVH (increased LVM and septal-to-posterior wall thickness ratio of <1.5). Left ventricular dilation was present if the BSA-corrected LVED exceeded 32 mm/m2; left atrial dilation if LAD corrected for BSA exceeded 22 mm/m2, and aortic root dilation if aortic root exceeded 22 mm/m2after correction for BSA (14).
The following indices of overall systolic function were calculated: fractional shortening; fractional long axis shortening (LAX) (15); mean velocity of circumferential fiber shortening corrected for heart rate (mVcfc); pre-ejection period adjusted for heart rate and gender (PEPc); LV ejection time adjusted for heart rate (ETc); the ratio of the non-corrected pre-ejection period to ejection time (PEP/ET); ejection fraction; and cardiac output (CO) indexed to BSA (cardiac index [CI]). A CI below 2.0 l/min/m2was considered to be reduced. The following indices of overall diastolic function were also calculated: maximum amplitude of the E-wave (E-Vmax) and A-wave (A-Vmax); the ratio of E-max to A-max (E/A ratio); and the isovolumetric relaxation time (IVRT), the time from aortic valve closure to mitral valve opening, was obtained by Doppler echocardiography (16).
All data were calculated from three cardiac cycles, and mean values were used for further calculations.
Repeated BP measurements were performed with a Dynamap during echocardiographic examination. Systemic vascular resistance (SVRe: dyne·s·cm−5) was estimated as [(80 times mean arterial BP)/CO].
Patients were divided into two groups according to their LVM: those with an increased LVM and those with a normal LVM, based on normative height-corrected values.
Statistical analyses were performed using SPSS version 10.7 for Windows. Data are expressed as mean ± SD or as numbers or percentages of subjects. Differences between groups were analyzed using the Student unpaired ttest for continuous variables and the chi-square test for dichotomous variables. Least-square linear regression analysis was performed to assess bivariate correlations between variables. The given corrected r2are related to the partial correlation coefficients. Multivariate analysis was performed to assess co-factorial influences. Differences were considered to be statistically significant for p values below 0.05.
Since 1997, a total of 55 female patients with genetically proven AFD were included in this study. Forty-two of the 55 female subjects were relatives of affected males and were detected by pedigree examinations. The remaining 13 patients presented because of classical symptoms of AFD. The mean age of onset of symptoms was 9.3 years (mostly neuropathic pain), and the mean age at diagnosis was 40 years. Four patients (7.2%) had fewer than four classic AFD symptoms, 14 patients (25.5%) four to six symptoms, and 37 patients (67.3%) more than six different symptoms or organ manifestations dependent on the age of the patients. No female patients had clinical manifestations limited to the heart.
The mean age at the time of cardiac examination was 39.6 ± 17.3 years, with a range from 6.1 to 70.8 years. Body weight ranged from 21 to 96 kg (mean, 63.3 ± 15.3 kg), and BSA from 0.8 to 2.1 m2(mean, 1.6 ± 0.2 m2). Mean BMI was 23.7 ± 4.8 kg/m2(range, 15.3 to 35.7 kg/m2). Mean heart rate was 69 ± 14 beats/min (range, 51 to 102 beats/min). Mean systolic BP was 123.6 ± 14.4 mm Hg (range, 95 to 158 mm Hg) and mean diastolic BP was 70.7 ± 8.5 mm Hg (range, 53 to 87 mm Hg). The SVRewas 1,935 ± 815 dyne·s·cm−5. Of the 55 patients, nine (16%) had arterial hypertension controlled by medication. Seven of the 55 patients were smokers, and no patients had diabetes or other diseases that could explain LVH.
General echocardiographic measurements of all 55 female patients with AFD are shown in Table 1, and they are grouped according to absence or presence of an increased LVM in Table 2. Patients with an increased LVM were older than patients with a normal LVM and tended to be more obese, although differences in BSA were not significant between the two groups. Patients in these two groups also differed significantly for AoD, LA cavity diameter (LAD), LA wall thickness (LAT), and by definition, for all LVM parameters. The upper 95% confidence limit for age of patients with a normal LVM was 36.8 years, and the lower 95% confidence limit for age of patients with an increased LVM was 39.7 years. Therefore, an age cut-point of 38.2 years was selected. Above 38.2 years, four of 28 patients were within the normal range of LVM, whereas 24 (85.7%) had an increased LVM. Below 38.2 years of age, 12 of 27 patients (44.4%) had a normal LVM (44.4%), and 15 (55.6%) had an increased LVM. All 25 patients older than 45 years had an increased LVM (Fig. 1). With increasing LVM there was an increase in LA thickness (F= 13.17; p < 0.0007 and r2= 0.465) and dimension (F= 41.8; p < 0.0001 and r2= 0.67).
Of the total population of 55 patients, only 13 (23.6%) had normal LV geometry; seven patients (12.7%) had concentric remodeling; 29 (52.7%) had concentric hypertrophy; and six (10.9%) had eccentric hypertrophy. Two of the six patients with eccentric hypertrophy had subaortic gradients of 15 and 28 mm Hg, respectively.
Multivariate analysis revealed a significant influence only of age on LVM (F= 22.27; p < 0.0001; r2= 0.905). Additionally, there was no correlation between LVM and the presence or absence of residual α-galactosidase A activity within leukocytes or with the gene mutation. Only two of the 55 (3.6%) female subjects had severely reduced leucocyte α-galactosidase A activity, comparable to findings in male patients. All others had normal enzyme activity.
Mean indices of systolic function in the groups of patients with normal versus increased LVM are presented in Table 3. Classic indices of systolic function, such as fractional shortening and ejection fraction, did not differ between the groups, although patients with an increased LVM tended to have a lower CI; but LAX was significantly reduced in patients with LVH. Additional indices of systolic function, PEPcand the ratio of PEP/ET, were prolonged in patients with increased LVM, whereas ETcdid not differ between the groups. Likewise, mVcfcwas decreased with increasing LVM (F= 13.51; p < 0.001; r2= −0.489). Even in young patients below the cut-off point of 38.2 years, LVH was associated with reductions in LAX (50.62 ± 9.5; 95% CI: 44.6 to 56.7 vs. 43.3 ± 6.7; 95% CI: 38.1 to 48.5; p < 0.01) and mVcfc(1.57 ± 0.26; 95% CI: 1.4 to 1.73 vs. 1.35 ± 0.37; 95% CI: 1.05 to 1.63; p < 0.05).
Indices of LV diastolic function in patients with normal versus increased LVM are presented in Table 4. Nearly all measured non-corrected diastolic function indices differed significantly between patients with a normal LVM compared with those with LVH. The LVH was associated with an increase in A-Vmax, and IVRT. The increase in A-Vmax led to a reduction in the E/A ratio. Additionally, LVH was associated with a significant increase in ΔT E/A. There was a significant correlation between increasing LVM and increasing IVRT (r2= 0.598; p < 0.0001) and between increasing LVM and decreasing E/A ratio (r2= 0.499; p < 0.0001). Even after adjustment for age, there was still a reduction in the E/A ratio in patients with an increased LVM compared with patients with a normal LVM (F= 18.13; p < 0.0001; r2= −0.509). Furthermore, an increase in LAD and LAT was accompanied by an increase in IVRT (F= 12.4; p < 0.001; r2= 0.671 and F= 7.17; p < 0.01; r2= 0.44) and a decrease in E/A ratio (F= 24.35; p < 0.0001; r2= −0.572 and F= 12.17; p < 0.001; r2= −0.455).
Of the 55 patients investigated, 27 (49.1%) had an E/A ratio above 1.5, while 23 patients (41.9%) had a ratio between 1.5 and 1.0, and five patients (9%) had a ratio below 1.0. Increasing mass was correlated with a decrease in the E/A ratio (F= 6.546; p = 0.014). This decrease in E/A ratio could be noted as dependent on age and LVM (F= 4.980; p = 0.012), and on age, LVM, and IVRT corrected for heart rate (IVRTc) (F= 3.316; p = 0.02). There was a significant correlation between IVRTcand the E/A ratio (F= 6.621; p = 0.016; r2= −0.444). Patients with a severely reduced E/A ratio below 1.0 had a significantly higher LVM than those with a normal E/A ratio (83.06 ± 13.4 g; 95% CI: 66.4 to 99.7 g vs. 64.2 ± 36.5 g; 95% CI: 53.8 to 74.6 g; p = 0.035). Eight of 27 patients with reduced E/A ratio and reduced CI and LVH were in New York Heart Association (NYHA) functional class III to IV, although having normal or minimally reduced indices of systolic function.
Additional echocardiographic findings
There was thickening of the aortic valve leaflets in 14 patients (25.5%) and mild mitral valve thickening combined with very mild insufficiency in 14 others (25.5%). Six patients (10.9%) had mild mitral valve prolapse. No patient had a gradient across the aortic valve of more than 20 mm Hg. Patients with aortic valve thickening were significantly older than those without (54.9 ± 13.4 years vs. 35.49 ± 16 years; p < 0.0001), as were patients with mitral valve thickening and regurgitation (56.5 ± 12.9 years vs. 34.9 ± 15.4 years; p < 0.0001). Both the aortic and mitral valves were affected in seven patients (12.7%); these seven patients were significantly older than those who were not affected (64.2 ± 6.6 years vs. 37.0 ± 15.8 years; p < 0.0001). There was a strong correlation between LVM increase and bi-valvular manifestations of AFD (F= 58.23; p < 0.0001; r2= 0.724).
In light of new therapeutic strategies for treatment of lysosomal storage diseases with enzyme replacement therapy, it is of major interest to know who might be affected. No previous studies have focused solely on the presence of cardiac involvement in female heterozygotes; prevalence data have been presented combined with data from hemizygote male patients (5,7,17). One reason for this might be that female patients, except those who presented with severe clinical symptoms of heart failure, have not been examined systematically for cardiac involvement. This lack of attention to female subjects was probably driven by the common belief that heterozygous “carriers” of an X-linked recessive disease are not likely to be affected as seriously as hemizygote males (2). The present study shows that the prevalence of cardiac involvement in heterozygote females is high and might be as high as in males. This is in contrast to previous published data (11). It should therefore be considered that AFD follows X-linked dominant rather than recessive transmission.
One major difference between male and female patients with AFD is the degree of residual activity of α-galactosidase A. Female patients usually present with low to even normal residual enzyme activity. Nevertheless, even female patients with normal plasma enzyme levels can show a number of clinical symptoms of AFD, and they may have severe cardiac involvement.
All the women investigated who were over 45 years of age had LVH, and their hypertrophy was not related to arterial hypertension or other diseases. Those nine females with controlled arterial hypertension have been treated mainly with diuretics and angiotensin-converting inhibitors between three and 18 years. Three of them started treatment before the final diagnosis of AFD had been established.
Cardiac involvement below 38 years of age was more variable. The youngest female patient with LVH was 14 years old. The severity of hypertrophy appears to progress with age, in agreement with previous data (5,7), and LVH apparently occurs 10 to 15 years later in heterozygote females than in hemizygote males. There were four female patients with severe cardiomyopathy (LVM >150 g/m2.7) who were older than 66 years (Fig. 1). They were all in NYHA functional class IV, and one had reduced leukocyte α-galactosidase activity.
The major cardiac abnormality was LVH. Only 23.6% of the female patients had normal LV geometry and mass. Concentric hypertrophy, with or without very mild cavity dilation, appears to be the most common abnormality (52.7%), followed by concentric remodeling (12.7%) and eccentric hypertrophy (10.9%). In addition to increased wall thickness of the LV, there was an increase in LA thickness and cavity diameter. The increased LAT is consistent with the increased LV thickness and likely indicates atrial deposition of Gb3. Furthermore, it is known that increased LVM in patients with hypertension correlates with increased LAD, indicating an impaired LV diastolic filling (18).
Classic indices of overall systolic function were normal, as has previously been reported in males (5,7,17). However, with increasing mass we observed a decrease in long axis shortening and mean velocity of circumferential fiber shortening. Decreased long axis shortening with preserved fractional shortening is also known to occur in severe LVH due to pressure overload (15).
Additionally, a disturbed long axis function reflects early diastolic dysfunction when systolic function remains normal (19,20). Likewise, an increase in LVM was also associated with a prolonged PEPcand an increase in the PEP/ET ratio, which may additionally indicate a reduced rate of increase of LV filling pressure during isovolumetric systole (LV dP/dt). These findings were age independent, as demonstrated by comparisons of patients with normal mass or LVH below the age of 38 years. Although all compared measurements of systolic function were within the normal ranges, there were significant differences between both groups. Nevertheless, patients with AFD cardiomyopathy without clearly abnormal hemodynamic functions (i.e., without pressure or volume overload) can be distinguished from AFD patients with normal hearts by the presence of depressed ejection phase indices and lower mVcf values (21). Of the investigated patients with LVH, 22.8% (8 of 35) were in NYHA functional class III or IV, and their dyspnea was not related to other AFD-related organ involvement.
Although most parameters of diastolic function also remained within the normal range, LA cavity size and wall thickness were increased. In fact, there was a good correlation between increased LA wall thickness and cavity size and reduced diastolic function parameters. Nearly all non-corrected parameters of diastolic function, however, showed significant differences between patients with normal LVM and those with increased LVM. Even after correction for age and heart rate, differences in IVRT, A-Vmax, and the E/A ratio remained. This result contrasts with the results of a previously published study (5), in which only three of the 17 patients had, by our definition, severely increased LVM. Likewise, the mean age of the females studied by Linhart et al. (5)was 35 ± 19 years, while the mean age of the hemizygotes in that study was 39 ± 10 years. In our study population, five of the patients (9%) with a significantly increased LVM had evidence of diastolic dysfunction, as indicated by an E/A ratio below 1.0 and a prolonged IVRT.
An increase in LVM may lead to diastolic dysfunction of the LV independent of the age of the patient and may, thereby, result in restrictive cardiomyopathy (22–24). Although classic infiltrative cardiomyopathies are frequently restrictive, the infiltrative cardiomyopathy associated with AFD seems to be different. By contrast with other cardiomyopathies (e.g., cardiac amyloidosis), where there is interstitial infiltration, in AFD the lipid accumulation is intra-lysosomal; furthermore, the amount of Gb3deposited in the myocytes is <1% of the LVM (11 mg/g heart wet weight) (24, 25). Therefore, there appears to be a true increase in LV myocyte mass in AFD cardiomyopathy. It is known that an increase in LVM can result in impaired ventricular compliance and increased filling pressures, leading to the restriction of diastolic filling and symptoms of congestive heart failure. Interpretation of the given data indicates an impaired compliance, a reduced rate of increase of filling pressure during isovolumetric systole (LV dP/dt), and an increased filling pressure in all those patients with the most severe cardiac involvement.
The present study also showed frequent valvular abnormalities, mostly consisting of thickening of the aortic and mitral valves. The prevalence of mitral and aortic valve abnormalities increased with age. Thickening of the aortic valve was not observed below the age of 23 years, and mitral valve thickening was not observed below 33 years of age. Combined mitral and aortic valve thickening was not observed below the age of 54 years. Furthermore, this study did not confirm the high incidence of mitral valve prolapse previously reported by others (7,26). Those patients with mitral valve prolapse had no or only very mild regurgitation. A high incidence of aortic root dilation has been described previously in severely affected male patients (7,17). In our female study population, the aortic root diameter was significantly larger in patients with an increased LVM than in patients with a normal LVM; but there was only one patient in whom the BSA-indexed AoD exceeded 22 mm/m2, and this patient had eccentric LVH. Aortic root dilation appears, therefore, to be less frequent in female heterozygotes than in male hemizygotes.
The present study has several limitations. Even though its single-observer design has the advantage of limiting interobserver variability, it has the disadvantages inherent in a single-observer study. Parameters presented in this study were obtained non-invasively, and though they were therefore available from more patients, it is recognized that there are more sophisticated parameters available for describing left ventricular diastolic function by using more invasive monitoring techniques. Additionally, the use of these frequently reported parameters offers the possibility of comparing our data in AFD patients with data from other hypertrophic cardiomyopathies.
This is the first study to demonstrate the high incidence of cardiac involvement in females heterozygous for AFD. The most frequent cardiac structural anomaly was concentric LVH, followed by cardiac remodeling and eccentric hypertrophy. The structural changes were highly correlated with the age of the patients, and the increasing LVM caused impairment of systolic and diastolic function. Valvular abnormalities, such as aortic or mitral leaflet thickening, while frequent, were mild. Mitral valve prolapse was found not to be as frequent as previously reported. Female patients with AFD have equally severe cardiac involvement and should therefore be considered for enzyme replacement therapies.
- Anderson–Fabry disease
- body mass index
- blood pressure
- left atrial diameter
- thickness of the left atrial wall
- fractional long axis shortening
- left ventricular/ventricle
- left ventricular hypertrophy
- left ventricular mass
- left ventricular mass, indexed to body surface area
- left ventricular mass, indexed to height
- estimated systemic vascular resistance
- Received April 5, 2002.
- Revision received June 8, 2002.
- Accepted July 12, 2002.
- American College of Cardiology Foundation
- Desnick R.J.,
- Ioannou Y.A.,
- Eng C.M.
- Desnick R.J.,
- Blieden L.C.,
- Sharp H.L.,
- et al.
- Goldman M.E.,
- Cantor R.,
- Schwartz M.F.,
- et al.
- MacDermot K.D.,
- Holmes A.,
- Miners A.H.
- Kampmann C.,
- Wiethoff C.M.,
- Wenzel A.,
- et al.
- Devereux R.B.,
- Koren M.J.,
- de Simone G.,
- et al.
- Feigenbaum H.
- Dumesnil J.G.,
- Shourci M.,
- Laurenceau J.L.,
- et al.
- Schmailzl K.J.G.
- Gerdts E.,
- Palmieri V.,
- Otterstad J.E.,
- et al.
- Henein M.Y.,
- Gibson D.G.
- Henein M.Y.,
- Gibson D.G.
- Peterson K.L.,
- Sklovan D.,
- Ludbrook P.,
- et al.
- Cantor W.J.,
- Butany J.,
- Iwanochko M.,
- et al.