Background
Fabry disease (FD) is characterized by the accumulation of sphingolipids in multiple organs, including the left atrium. It is uncertain if the left atrial (LA) reservoir, conduit, and contractile functions evaluated by speckle-tracking echocardiography are affected in Fabry cardiomyopathy and whether enzyme replacement therapy can improve LA function.
Methods
In this retrospective cohort study, LA strain, strain rates, and phasic LA volumes were studied in 50 patients with FD and compared with values in 50 healthy control subjects.
Results
All three LA phasic functions were altered. Peak positive strain (reservoir function) was 38.9 ± 14.9% versus 46.5 ± 10.9% ( P = .004), and late diastolic strain (contractile function) was 12.6 ± 5.9% versus 15.6 ± 5.3% ( P = .010). In 15 patients who started enzyme replacement therapy during the study, most of the LA parameters improved at 1-year follow-up (peak positive strain from 32.0 ± 13.5% to 38.0 ± 13.5%, P = .006), whereas there was a trend toward deterioration in 15 patients who never received treatment (peak positive strain from 47.3 ± 10.8% to 41.3 ± 9.3%, P = .058). Nine patients with FD (21%) experienced new-onset atrial fibrillation or stroke during 4-year follow-up. By univariate analysis, peak positive strain and early diastolic strain demonstrated significant associations with clinical events, surpassing conventional echocardiographic parameters and clinical characteristics.
Conclusions
LA reservoir, conduit, and contractile functions by speckle-tracking echocardiography were all affected in FD. Enzyme replacement therapy improved LA function. LA strain parameters were associated with atrial fibrillation and stroke.
Highlights
- •
LA function by STE is altered in FD.
- •
ERT improves LA function after 1 year of treatment.
- •
Strain parameters are associated with AF and stroke.
Fabry disease (FD) is a genetic lysosomal storage disease characterized by a deficient activity of the enzyme α-galactosidase A. This defect causes accumulation of sphingolipids in multiple organs, including the heart. Cardiovascular death is the leading cause of mortality in patients with FD. Enzyme replacement therapy (ERT) can reverse left ventricular (LV) hypertrophy (LVH) and improve exercise capacity, but its efficacy is uncertain when started after the development of myocardial fibrosis. It is therefore crucial to diagnose cardiac involvement early and implement effective and timely treatment to prevent irreversible cardiomyopathy. Because of the high cost of ERT, most international guidelines currently suggest strict eligibility criteria, such as significant LVH, diastolic dysfunction, and increased left atrial (LA) volume.
Speckle-tracking echocardiography (STE) is a sensitive tool allowing quantification of the three LA phasic functions: reservoir, conduit, and contractile. Cross-sectional studies have demonstrated impaired LA reservoir function in patients with FD. However, LA dysfunction remains incompletely characterized, and the potential beneficial role of ERT is unknown. Therefore, we sought to comprehensively study LA phasic function by STE in patients with FD in comparison with healthy subjects and to assess the effects of ERT. We also aimed to determine the association between these parameters and new-onset atrial fibrillation (AF) and stroke in patients with FD.
Methods
Study Design
Study Population
In this retrospective cohort study, we included all patients followed at the FD clinic of Hôpital du Sacré-Coeur de Montréal from 2008 to 2015. Patients with inadequate endocardial definition of the LA walls resulting in poor tracking with STE were excluded. The control group consisted of 50 healthy subjects matched for age (40 ± 14 years) and gender (42% men). To study the effect of ERT, patients with FD were divided into two groups: patients who started treatment during the study period (the ERT cohort), in whom transthoracic echocardiography (TTE) was performed before and after ERT, and patients who never received ERT (the natural evolution cohort). The study protocol was approved by the institutional ethics committee.
FD
The diagnosis of FD was confirmed in all patients by mutation identification. Patients were followed twice a year with multiple tests, including electrocardiography. Holter monitoring was performed every 2 years and when clinically indicated. ERT was prescribed to patients fulfilling the Canadian FD treatment guidelines criteria ( Supplemental Table 1 ). The residual enzyme activity of α-galactosidase A was collected.
Clinical Data
All adult patients with FD in Canada are offered to enroll in a national database, the Canadian Fabry Disease Initiative ( ClinicalTrials.gov identifier NCT00455104 ). Baseline characteristics, comorbidities, medications, test results, and clinical events, including AF and stroke, were prospectively collected in this database. AF was defined as any clinically significant episode or episode lasting >30 sec in case of clinically silent AF found during Holter monitoring.
Echocardiography
Annual TTE was performed in accordance with the American Society of Echocardiography guidelines. For each patient, we used the earliest retrievable digitally recorded study available for offline analysis, from 2008 onward. Standard echocardiographic parameters, such as LV mass and parameters of diastolic function, were measured offline by the researchers. For longitudinal follow-up, we selected the study closest to 1 year after the baseline study.
LA Myocardial Mechanics and Volumetric Phasic Functions
All transthoracic echocardiograms were analyzed offline by STE by a single operator who was blinded to clinical data. Using apical four- and two-chamber views, the LA endocardial borders were tracked, and strain and strain rate values were derived from the time-strain and strain rate curves, averaging six atrial segments (Velocity Vector Imaging version 3; Siemens Medical Solutions USA, Mountain View, CA). Peak positive, early diastolic, and late diastolic strain and strain rates represent respectively the reservoir, conduit, and contractile functions, taking the electrocardiographic R wave as the reference point ( Figure 1 ). The mean frame rate for the STE analysis was 48 ± 16 frames/sec.
The maximal, minimal, and pre–atrial contraction LA volumes were measured (Simpson method of disks) and the three phasic emptying fractions quantified ( Figure 1 ). The LA stiffness index was calculated by dividing the E/e′ ratio by peak positive strain. For each parameter, LA dysfunction was defined as a lower value than the mean of the control group minus 2 SDs.
LV Myocardial Mechanics
In patients with FD, LV subendocardial mechanics (global longitudinal systolic strain [GLS], global longitudinal systolic strain rate [GLSR], and early diastolic strain rate) was measured at baseline and at follow-up using the same STE software previously described. Apical four-, three-, and two-chamber views were used to obtain longitudinal strain and strain rate, averaging the 16 myocardial segments.
Intraobserver and Interobserver Variability
Ten randomly selected studies were reanalyzed by the same operator several months after the initial analysis. A second experienced observer, also blinded to previously obtained data, analyzed the same loops for the assessment of interobserver variability.
Statistical Analyses
Categorical variables are expressed as frequencies and percentages. Continuous variables are summarized as mean ± SD or median (interquartile range). Comparisons between groups were done using independent t tests or Wilcoxon Mann-Whitney U tests for continuous variables and Fisher exact tests for categorical variables. Pearson coefficients were used to assess the correlation between LA phasic function parameters and echocardiographic and clinical variables. LA parameters at baseline and at follow-up were compared using paired t tests. Comparisons of the temporal changes in LA echocardiographic parameters between the ERT cohort and the natural evolution cohort were performed by mixed-design repeated-measures analysis of variance using treatment group as the intersubject factor and the echocardiographic parameters at baseline and at follow-up as the intrasubject factor. Univariate logistic regression models were used to identify factors associated with clinical events. Receiver operating characteristic curve analysis was performed to assess the discriminatory power of LA peak positive and early diastolic strain and LV GLS for the occurrence of the combined end point of AF or stroke. Intraobserver and interobserver variability was assessed by the intraclass correlation coefficient and the coefficient of variation. A level of significance of .05 was set for all analyses. All statistical analyses were conducted using SPSS version 20 (SPSS, Chicago, IL).
Results
Study Population
Fifty-one patients with FD followed at our clinic were eligible for this study ( Figure 2 ). One patient with poor image quality was excluded (FD group, n = 50). The mean age was 40 ± 15 years, and 42% were men ( Table 1 ). Twelve patients were receiving ERT at baseline and were not studied longitudinally. Thirty-eight patients were not receiving ERT at baseline. In 30 of these patients, follow-up TTE was performed, and they were studied longitudinally at a median time of 435 days (interquartile range, 369–758 days). Fifteen patients started ERT after baseline TTE (ERT cohort), and 15 had no indication (natural evolution cohort). Three patients (20%) started ERT on the basis of cardiac criteria, nine (60%) after developing Fabry nephropathy (glomerular filtration rate < 90 mL/m 2 ), two (13%) after transient ischemic attacks, and one (7%) for compassionate use without fulfilling major criteria for ERT. All patients, except one bearing a cardiac variant mutation, had classical FD ( Supplemental Table 2 ).
| All patients ( n = 50) | ERT cohort ( n = 15) | Natural evolution cohort ( n = 15) | P ∗ | |
|---|---|---|---|---|
| Age (y) | 40 ± 15 | 39 ± 14 | 38 ± 15 | .863 |
| Men | 21 (42) | 7 (47) | 4 (27) | .450 |
| Hypertension | 7 (15) | 2 (17) | 0 (0) | .450 |
| Atrial fibrillation | 4 (9) | 1 (8) | 0 (0) | .444 |
| Prior stroke | 6 (13) | 1 (8) | 1 (7) | 1.000 |
| Glomerular filtration rate (mL/min/1.73 m 2 ) | 83 ± 33 | 88 ± 23 | 96 ± 22 | .403 |
| Chronic renal insufficiency † | 8 (17) | 1 (8) | 0 (0) | .444 |
| Medication at baseline | ||||
| ERT | 12 (24) | 0 (0) | 0 (0) | 1.000 |
| Aspirin | 13 (28) | 3 (25) | 1 (7) | .294 |
| Statin | 5 (11) | 1 (8) | 0 (0) | .444 |
| β-blocker | 7 (15) | 1 (8) | 0 (0) | .444 |
| ACE inhibitor or ARB | 12 (26) | 2 (17) | 0 (0) | .188 |
| α-galactosidase A activity (nmol/h/mg) | 19 (2–37) | 24 (1–65) | 18 (4–37) | .212 |
| Women | 32 (23–46) | 49 (26–67) | 26 (16–42) | |
| Men | 2 (1–3) | 1 (1–3) | 2 (0–4) |
∗ ERT cohort versus natural evolution cohort.
Echocardiography
Significant differences were observed in the baseline echocardiographic parameters between the FD group and the age-balanced control group ( Table 2 ). Patients with FD had significantly higher LV mass and larger anteroposterior LA diameter and maximal volume (41.2 ± 27.5 vs 30.1 ± 8.6 mL/m 2 , P = .009). LVH was present in 14 patients with FD (28%). The majority of patients with FD had normal diastolic function (72%). LV strain parameters were measured in 49 patients with FD (98%). Measurements were not possible in one patient (2%), because of poor endocardial definition. The mean GLS, GLSR, and early diastolic strain rate were −18.8 ± 3.0%, −0.8 ± 0.2 sec −1 , and 1.0 ± 0.3 sec −1 , respectively.
| Patients with FD ( n = 50) | Healthy subjects ( n = 50) | P | |
|---|---|---|---|
| LV size and function | |||
| LV end-diastolic diameter (mm) | 48 ± 6 | 46 ± 5 | .097 |
| LV end-systolic diameter (mm) | 30 ± 7 | 30 ± 4 | .867 |
| Relative wall thickness | 0.40 ± 0.12 | 0.31 ± 0.05 | <.001 |
| LV mass (g/m 2 ) | 95 ± 49 | 58 ± 14 | <.001 |
| LV ejection fraction (%) | 65 ± 5 | 63 ± 4 | .088 |
| Doppler parameters | |||
| E/A ratio | 1.7 ± 0.7 | 1.5 ± 0.8 | .242 |
| Deceleration time (msec) | 215 ± 59 | 195 ± 48 | .067 |
| e′ lateral | 12 ± 5 | 13 ± 4 | .072 |
| E/e′ ratio lateral | 8 ± 4 | 6 ± 2 | .001 |
| LA diameter (mm) | 36 ± 8 | 32 ± 5 | .019 |
LA Myocardial Mechanics and Volumetric Phasic Functions
Patients in the FD group had higher phasic LA volumes and lower total and active emptying fractions compared with those in the control group ( Table 3 ). The FD group also had significantly lower absolute values of LA strain and strain rates compared with the control group. These changes were present for all three LA functions. The analyses were repeated after excluding the four patients in the FD group who were in AF during echocardiography or who previously had AF. The results of the comparison of LA deformation parameters with those in control subjects were very similar after this exclusion. All the differences remained significant except peak positive strain rate (1.5 ± 0.4 vs 1.6 ± 0.4 sec −1 , P = .084). Using cutoff values of 24.7% for peak positive strain and 56.5% for total emptying fraction (2 SDs below mean value of the control subjects), nine (18%) and 10 (20%) patients with FD had reduced LA reservoir function by strain and volumetric analysis, respectively. The LA stiffness index was higher in patients with FD. The baseline characteristics of patients with FD with versus without LA reservoir dysfunction by strain analysis are shown in Supplemental Table 3 . Patients with LA reservoir dysfunction were older and had greater LV mass and filling pressures and lower LV systolic function.
| Patients with FD ( n = 50) | Healthy subjects ( n = 50) | P | |
|---|---|---|---|
| LA phasic functions | |||
| LA reservoir function | |||
| Peak positive strain (%) | 38.9 ± 14.9 | 46.5 ± 10.9 | .004 |
| Peak positive strain rate (sec −1 ) | 1.4 ± 0.5 | 1.6 ± 0.5 | .013 |
| LA conduit function | |||
| Early diastolic strain (%) | 28.2 ± 10.1 | 31.0 ± 8.7 | .160 |
| Early diastolic strain rate (sec −1 ) | −1.4 ± 0.7 | −1.7 ± 0.5 | .007 |
| LA contractile function | |||
| Late diastolic strain (%) | 12.6 ± 5.9 | 15.6 ± 5.3 | .010 |
| Late diastolic strain rate (sec −1 ) | −1.2 ± 0.5 | −1.5 ± 0.4 | .012 |
| LA volumes (mL/m 2 ) | |||
| Maximal | 41.2 ± 27.5 | 30.1 ± 8.6 | .009 |
| Minimal | 17.4 ± 24.3 | 8.4 ± 4.1 | .013 |
| Pre–atrial contraction | 19.8 ± 11.0 | 15.3 ± 5.5 | .014 |
| LA reservoir function | |||
| Total emptying fraction (%) | 65.1 ± 14.8 | 72.7 ± 8.1 | .002 |
| LA conduit function | |||
| Passive emptying fraction (%) | 46.5 ± 11.5 | 49.4 ± 8.9 | .168 |
| LA contractile function | |||
| Active emptying fraction (%) | 39.8 ± 12.8 | 45.8 ± 14.8 | .037 |
| LA stiffness index | 0.34 ± 0.49 | 0.14 ± 0.07 | .007 |
Correlation between Parameters of LA Phasic Function and Other Echocardiographic and Clinical Data
Strain, strain rate, and emptying fractions for the three LA phasic functions in patients with FD were correlated with other echocardiographic and clinical parameters in univariate linear regression analyses ( Table 4 ). Reservoir and conduit functions were associated more closely with age, LV mass, E/e′ ratio, LV diastolic strain rate, and systolic strain (less so for conduit function). In addition, conduit function was linked to diastolic filling variables. Inconstant weak associations were found between LA contraction and E/e′ ratio, LV systolic strain, and diastolic strain rate.
| Pearson correlation coefficient | |||||||||
|---|---|---|---|---|---|---|---|---|---|
| Age | Enzyme activity | LVEF | LV mass | E/A ratio | E/e′ ratio lateral | DT | LV GLS | LV early diastolic SR | |
| LA phasic functions | |||||||||
| LA reservoir | |||||||||
| Peak positive strain | −0.603 ∗ | 0.008 | 0.202 | −0.632 ∗ | 0.378 ∗ | −0.626 ∗ | −0.176 | −0.576 ∗ | 0.512 ∗ |
| Peak positive strain rate | −0.521 ∗ | 0.052 | 0.248 | −0.596 ∗ | 0.276 | −0.538 ∗ | −0.228 | −0.525 ∗ | 0.493 ∗ |
| Total emptying fraction | −0.494 ∗ | 0.156 | 0.341 ∗ | −0.675 ∗ | 0.122 | −0.517 ∗ | −0.082 | −0.580 ∗ | 0.437 ∗ |
| LA conduit | |||||||||
| Early diastolic strain | −0.687 ∗ | −0.232 | −0.064 | −0.604 ∗ | 0.586 ∗ | −0.581 ∗ | −0.311 ∗ | −0.407 ∗ | 0.424 ∗ |
| Early diastolic strain rate | 0.818 ∗ | 0.068 | −0.040 | 0.545 ∗ | −0.511 ∗ | 0.626 ∗ | 0.426 ∗ | 0.421 ∗ | −0.483 ∗ |
| Passive emptying fraction | −0.632 ∗ | −0.177 | −0.035 | −0.510 ∗ | 0.514 ∗ | −0.411 ∗ | −0.326 ∗ | −0.330 ∗ | 0.278 |
| LA contractile | |||||||||
| Late diastolic strain | 0.008 | −0.048 | 0.017 | −0.128 | −0.163 | −0.297 ∗ | −0.011 | −0.282 | 0.290 |
| Late diastolic strain rate | 0.071 | 0.052 | 0.058 | 0.142 | 0.194 | 0.355 ∗ | 0.135 | 0.269 | −0.401 ∗ |
| Active emptying fraction | 0.009 | 0.040 | 0.052 | −0.148 | −0.278 | −0.254 | −0.032 | −0.331 ∗ | 0.244 |
Effect of Enzyme Replacement Therapy versus Natural Evolution of FD
LA function was reassessed at a median time of 355 days (interquartile range, 316–382 days) after the treatment was started in the 15 patients in the ERT cohort. Significant improvements in all STE- and volume-derived parameters of reservoir function and some parameters of conduit and contractile functions were observed. LA phasic volumes all significantly regressed after institution of treatment ( Table 5 ). Table 6 demonstrates the proportion of patients in the ERT cohort with LA dysfunction at baseline and at follow-up and the proportion of patients with ≥10% improvement at follow-up. For example, LA reservoir dysfunction, defined by a total emptying fraction <56.5%, was present in seven patients (47%) in the ERT cohort at baseline. In six of these patients (86%), LA reservoir function normalized at follow-up. Ten patients (67%) had ≥10% improvement in peak positive reservoir strain.
| Patients with therapy ( n = 15) | Patients without therapy ( n = 15) | P ∗ | |||||
|---|---|---|---|---|---|---|---|
| Baseline | Follow-up | P † | Baseline | Follow-up | P † | ||
| LA phasic functions | |||||||
| LA reservoir function | |||||||
| Peak positive strain (%) | 32.0 ± 13.5 | 38.0 ± 13.5 | .006 | 47.3 ± 10.8 ‡ | 41.3 ± 9.3 | .058 | .002 |
| Peak positive strain rate (sec −1 ) | 1.3 ± 0.4 | 1.6 ± 0.5 | .015 | 1.6 ± 0.4 | 1.5 ± 0.4 | .411 | .017 |
| LA conduit function | |||||||
| Early diastolic strain (%) | 23.9 ± 8.9 | 27.7 ± 9.4 | .030 | 33.5 ± 9.0 ‡ | 28.9 ± 7.7 | .066 | .006 |
| Early diastolic strain rate (s −1 ) | −1.3 ± 0.7 | −1.4 ± 0.6 | .412 | −1.6 ± 0.7 | −1.6 ± 0.6 | .656 | .381 |
| LA contractile function | |||||||
| Late diastolic strain (%) | 9.5 ± 5.8 | 12.2 ± 8.0 | .069 | 13.8 ± 5.2 ‡ | 12.5 ± 6.7 | .539 | .120 |
| Late diastolic strain rate (s −1 ) | −1.0 ± 0.5 | −1.2 ± 0.7 | .125 | −1.4 ± 0.5 | −1.2 ± 0.6 | .559 | .166 |
| LA volumes (mL/m 2 ) | |||||||
| Maximal | 44.1 ± 17.3 | 35.2 ± 11.8 | .007 | 31.9 ± 8.0 ‡ | 32.1 ± 11.6 | .939 | .050 |
| Minimal | 19.4 ± 13.4 | 13.1 ± 10.1 | .017 | 9.1 ± 3.6 ‡ | 9.3 ± 4.6 | .864 | .019 |
| Pre–atrial contraction | 26.2 ± 13.6 | 18.0 ± 6.3 | .006 | 14.9 ± 3.8 ‡ | 16.2 ± 7.8 | .543 | .006 |
| LA reservoir function | |||||||
| Total emptying fraction (%) | 59.1 ± 13.3 | 66.2 ± 12.9 | .015 | 71.3 ± 9.1 ‡ | 72.1 ± 8.3 | .744 | .082 |
| LA conduit function | |||||||
| Passive emptying fraction (%) | 39.7 ± 9.9 | 45.7 ± 12.1 | .055 | 52.7 ± 8.7 ‡ | 51.4 ± 10.7 | .632 | .075 |
| LA contractile function | |||||||
| Active emptying fraction (%) | 34.7 ± 13.7 | 41.9 ± 12.4 | .009 | 39.9 ± 12.9 | 42.5 ± 12.4 | .505 | .305 |
| LA stiffness index | 0.42 ± 0.43 | 0.27 ± 0.21 | .038 | 0.17 ± 0.08 | 0.18 ± 0.07 | .321 | .024 |
∗ Comparison between the changes in patients with therapy and those in patients without therapy.
‡ Comparison of baseline values in patients with therapy versus patients without therapy ( P < .05).
