You are here

Article Text

Article menu
Heart failure and cardiomyopathies
Original article
Clinical and genetic predictors of major cardiac events in patients with Anderson–Fabry Disease
  1. Vimal Patel1,
  2. Constantinos O'Mahony1,
  3. Derralynn Hughes2,
  4. Mohammad Shafiqur Rahman3,
  5. Caroline Coats1,
  6. Elaine Murphy4,
  7. Robin Lachmann4,
  8. Atul Mehta2,
  9. Perry M Elliott1
  1. 1The Heart Hospital, University College London Hospitals NHS Foundation Trust, London, UK
  2. 2The Royal Free Hospital and University College Medical School, London, UK
  3. 3Institute of Statistical Research and Training, University of Dhaka, Dhaka, Bangladesh
  4. 4The National Hospital for Neurology and Neurosurgery, Charles Dent Metabolic Unit, London, UK
  1. Correspondence to Professor Perry M Elliott, The Heart Hospital, University College London Hospitals NHS Foundation Trust, 16-18 Westmoreland Street, London W1G 8PH, UK; perry.elliott{at}ucl.ac.uk

Abstract

Background Anderson–Fabry Disease (AFD) is an X linked lysosomal storage disorder caused by mutations in the α-galactosidase A gene. Some mutations are associated with prominent and, in many cases, exclusive cardiac involvement. The primary aims of this study were to determine the incidence of major cardiac events in AFD and to identify clinical and genetic predictors of adverse outcomes.

Methods and results We studied 207 patients with AFD (47% male, mean age 44 years, mean follow-up 7.1 years). Fifty-eight (28%) individuals carried mutations that have been previously associated with a cardiac predominant phenotype. Twenty-one (10%) developed severe heart failure (New York Heart Association functional class (NYHA) ≥3), 13 (6%) developed atrial fibrillation (AF), 13 (6%) received devices for the treatment of bradycardia; there were a total of 7 (3%) cardiac deaths. The incidence of the primary endpoint (a composite of new onset AF, NYHA ≥ 3 symptoms, device insertion for bradycardia and cardiac death) was 2.64 per 100 person-years (CI 1.78 to 3.77). Age (HR 1.04, CI 1.01 to 1.08, p=0.004), Mainz Severity Score Index score (HR 1.05, CI 1.01 to 1.09, p=0.012) and QRS duration (HR 1.03, CI 1.00 to 1.05, p=0.020) were significant independent predictors of the primary endpoint. The presence of a cardiac genetic variant did not predict the primary end point.

Conclusions AFD is associated with a high burden of cardiac morbidity and mortality. Adverse cardiac outcomes are associated with age, global disease severity and advanced cardiac disease but not the presence of cardiac genetic variants.

https://doi.org/10.1136/heartjnl-2014-306782

Statistics from Altmetric.com

    Request Permissions

    If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

    Introduction

    Anderson Fabry Disease (AFD) is an X-linked lysosomal storage disorder caused by mutations in the α-galactosidase A gene (GLA) that results in the accumulation of glycosphingolipid throughout the body. In its classic form, AFD is a multisystem disease with cutaneous, renal, cerebral and cardiac manifestations that causes progressive debilitating symptoms and premature death.1 ,2 Over 400 mutations have been identified in the GLA gene, most of which are associated with the classical phenotype. A group of mutations, the ‘cardiac variants’, are associated with a phenotype characterised by prominent and, in some cases, exclusive cardiac involvement3–11 but it is not known whether patients with these variants have a different prognosis from those with mutations that cause a classical multisystem presentation. The primary aims of this study were to determine the incidence of major cardiac events in a consecutively evaluated cohort of patients with AFD and to identify clinical and genetic predictors of adverse outcomes.

    Methods

    Study design

    An observational, longitudinal, prospective cohort design with retrospective analysis was used. The study conforms to the principles of the Helsinki declaration. Written informed consent for genetic investigation was obtained from all participants.

    Study population

    The cohort consisted of all patients with AFD aged ≥16 years who were evaluated at a dedicated cardiomyopathy clinic from 1 January 1988 until 3 December 2012. All patients were referred from two designated national centres for lysosomal storage disease at the Royal Free Hospital (London, UK) and the National Hospital for Neurology and Neurosurgery at Queen Square (London, UK).

    The diagnosis of AFD was based on plasma and leucocyte α-galactosidase A enzyme activity and sequencing of the GLA gene. A priori, GLA mutations previously associated with a cardiac predominant phenotype were labelled cardiac variants while all other mutations were considered classical (table 1).

    View this table:
    Table 1

    α Galactosidase mutations associated with a cardiac phenotype

    Baseline assessment

    The Mainz Severity Score Index (MSSI) was used to assess the overall severity of AFD.12 All patients were evaluated using resting 12-lead ECG and echocardiography in accordance with previously published methods.13 Recorded ECG parameters included rhythm, PR interval and QRS duration. Echocardiographic variables included: left atrial (LA) diameter, LV end-systolic dimensions (LVesd), LV end-diastolic dimension (LVedd), maximal wall thickness (MWT), posterior wall thickness in diastole (PWTd) and interventricular septal thickness in diastole (SWTd). MWT was defined as the greatest thickness in any single LV segment measured in the parasternal short-axis plane at the level of the mitral valve, mid-ventricle and apex at end-diastole. LVEF was calculated using LV volumes and Teichholz's method.14 LV mass (LVM) was calculated from the formula [0.8×(1.04×[(LVedd+PWTd+SWTd)3 2 (LVedd)3])+0.6 g] and indexed to body surface area to obtain the LVM index.13

    Follow-up and study end points

    A clinical review was performed every 6–12 months or earlier if there was a clinical event.

    The primary end point was a composite of:

    1. Severe heart failure symptoms: The development of New York Heart Association (NYHA) functional class III/IV symptoms.15

    2. Atrial fibrillation (AF): new onset paroxysmal, persistent or permanent AF.

    3. Bradycardia requiring device implantation: sinus node disease (inappropriate sinus bradycardia, tachycardia–bradycardia syndrome, sinoatrial arrest and sinoatrial exit block) and atrioventricular nodal disease (second-degree and third-degree atrioventricular block).

    4. Cardiac death: Cardiac death was classified as either sudden or heart failure related death; sudden cardiac death was defined as witnessed sudden cardiac death with or without documented ventricular fibrillation or death within 1 h of new symptoms or an aborted cardiac death due to successful cardiac resuscitation or appropriate internal cardioverter defibrillator (ICD) therapy. Nocturnal deaths with no antecedent history of worsening symptoms and aborted sudden death were included in this category. Death due to progressive heart failure included deaths in individuals with symptoms of progressive heart failure including cardiogenic shock.16 Non-cardiac death was not considered as an end point in this study.

    Development of severe heart failure, AF, bradycardia requiring device insertion and cardiac death were also examined individually as secondary end points.

    Statistical analysis

    SPSS (V.21.0), MedCalc (V.12.7.2.0) and coxphf available in R statistics were used for all statistical analyses. Normally distributed continuous data are expressed as mean±SD and as median and IQR for non-normally distributed data. Differences between means were compared using the Student t test and the Mann–Whitney U test for normally distributed and non- normally distributed continuous data, respectively. The χ2 and Fisher's exact tests were used for comparison of categorical data. The incidence rate for each end point was calculated by dividing the number of patients reaching the end point by the total follow-up period.

    Time-to-event (survival) analysis was used to examine the primary and secondary end-points. The follow-up period was calculated from the date of first evaluation at the Heart Hospital to the time of the first endpoint or censoring event. In individuals that did not reach an end point or censoring event, the follow-up period extended to the most recent clinical evaluation available. Patients that had reached an endpoint prior to their initial assessment were excluded from the time to event analysis for that end-point (eg, individuals with AF at baseline were excluded from the analysis for time to AF and the primary end point). Non-cardiac death was a censoring event. Implantation of cardiac devices exclusively for indications other than bradycardia (eg, ICD for the prevention of SCD) was used as a censoring event for the device implantation and primary composite end-point. The relation of the end point to prespecified baseline clinical characteristics (mutation, age at first evaluation, sex, MSSI score, LA diameter, EF and indexed LV mass) was assessed using the Cox regression with Firth's penalised likelihood estimation.17 ,18 Following univariate analysis, a multivariable model was fitted in accordance to the principles proposed by Harrell19 to identify independent predictors with a significance of 10%. The performance of QRS duration on the primary endpoint was examined using receiver operating characteristic curves and calculating the area under the curve (c-statistic). The positive predictive value (PPV) of a predictor was calculated by dividing sensitivity×prevalence by sensitivity×prevalence+ (1−specificity)×(1−prevalence) and expressed as a percentage. The negative predictive value (NPV) of a predictor was calculated by dividing specificity×(1−prevalence) by (1−sensitivity)×prevalence+specificity×(1−prevalence) and expressed as a percentage.

    Results

    Two hundred and seven patients were evaluated during the study period; the median follow-up period was 7.1 (4.0–9.1) years. The study cohort and outcomes are shown in figure 1. The characteristics of the population are shown in table 2. Fifty-eight individuals carried a cardiac genetic variant (50 with N215S, seven with R301Q and one with I91T). Sex distribution and duration of follow-up were similar in patients with cardiac and classical AGal mutations. At first evaluation, patients with cardiac variant mutations were older (42.1±13.8 vs 48.9±16.7 years, p=0.005), had a longer PR interval (142.5±24.4 vs 161.4±26.3 ms, p≤0.001) and a lower MSSI score (22.5±11.5 vs 17.1±10.4, p=0.002). A smaller proportion of patients with cardiac variants were receiving enzyme replacement therapy (ERT) at baseline (27.5 vs 10.3%, p=0.009) and a larger proportion of individuals with cardiac variants had devices for bradycardia at the first visit (0 vs 5, p=0.001). The incidence and multivariate predictors of the composite and individual end points are displayed in tables 3 and 4, respectively. Univariate predictors are provided in online supplementary table S1.

    View this table:
    Table 2

    Baseline characteristics of the study population. Normally distributed continuous data are expressed as mean+(SD) and as median and (IQR) for non-normally distributed data

    View this table:
    Table 3

    Annual incidence (per 100 person-years) and CI of study end-points

    View this table:
    Table 4

    The relation of baseline clinical characteristics to cardiac outcomes (Cox regression with Firth's penalised likelihood estimation. Multivariate analysis)

    Figure 1
    Figure 1

    Study cohort and outcomes. AF, atrial fibrillation; NYHA, New York Heart Association.

    Primary end-point

    Thirty patients reached the primary end-point (24 with classical mutations and six with cardiac variants). The annual incidence of the primary endpoint was 2.64 per 100 person-years (CI 1.78 to 3.77). In a univariable analysis, age, MSSI score, LA diameter, indexed LV mass and QRS duration were predictors for the composite endpoint. In multivariable analysis, age, MSSI score and QRS duration remained significant independent predictors. The QRS duration c-statistic (area under the curve) obtained from receiver operator characteristic curve was 0.675. The sensitivity, specificity, PPV and NPV for QRS duration on the composite endpoint are displayed in table 5. The presence of a cardiac genetic variant was not a significant predictor of the primary end point, as shown in figure 2. At the time of composite end point, 25 of the 30 individuals were receiving ERT. The mean duration of therapy prior to achieving the composite end point was 3.27±2.34 years.

    View this table:
    Table 5

    Receiver operator characteristics for QRS duration

    Figure 2
    Figure 2

    Cumulative incidence of the composite end point for both the cardiac and classical variants. Number at risk at each time point displayed on the x-axis.

    Heart failure symptoms

    During the follow-up period, 21 individuals (six with cardiac variants) developed severe heart failure symptoms with an annual incidence of 1.62 per 100 person-years (CI 1.00 to 2.48). There was no statistical difference in the incidence between the groups with and without cardiac variant mutations. Age at first evaluation, MSSI score, LA diameter and indexed LV mass were univariable predictors for the development of severe heart failure symptoms. In a multivariable analysis, age and MSSI score were the only significant predictor. The presence of a cardiac variant was not a significant predictor of progression to severe heart failure symptoms. In total, 22 of the 207 (10.6%) the patients in the study developed severe heart failure symptoms either prior to or during the study follow-up.

    Atrial fibrillation

    During the follow-up period, a total of 13 individuals (two with cardiac variants) developed AF with an annual incidence of 1.0 per 100 person-years (0.53–1.71). There were no differences in incidence of AF between those with cardiac and classical mutations. Age at first evaluation, MSSI score, LA diameter, indexed LV mass and QRS duration were univariable predictors for the development of AF. In multivariable analysis, age and indexed LV mass were significant predictors for the development of AF, while LA diameter showed a trend towards significance (p=0.053). In total, 17 of the 207 (8.2%) of the patients in the study developed AF either prior to or during the study follow-up.

    Device implantation

    During the follow-up period 13 patients (four with cardiac variants) received devices for the treatment of bradycardia. Of these 13, 11 received pacemakers and two received ICDs for concurrent indications warranting primary arrhythmic prophylaxis. During the follow-up period, three individuals (one with cardiac variant) received ICD for primary arrhythmic prophylaxis (unexplained syncope/presyncope and non-sustained ventricular tachycardia (NSVT)) and two individuals with classical variants received devices for cardiac resynchronisation with defibrillator function (NSVT and severe heart failure symptoms). Two individuals with cardiac variants received a pacemaker for other indications unrelated to disease progression (one for the treatment of LV outflow tract obstruction and another for the treatment of complete heart block following alcohol septal ablation). One individual with a cardiac variant had a pacemaker upgraded to a dual chamber ICD for primary prevention and two patients with cardiac variants received an upgrade of their pacing system to a device capable of cardiac resynchronisation and two individuals (one with a classical variant and one with a cardiac variant) received an upgrade of their pacing system to a device capable of cardiac resynchronisation with defibrillator function. In total, one individual with a cardiac variant was fitted with an ICD prior to baseline and 10 additional individuals (four with cardiac variants and six with classical variants) were fitted or upgraded to devices capable of treating ventricular arrhythmias during the follow-up period. One individual (N215S) received an appropriate shock prior to their baseline assessment and another (R301G) was treated by antitachycardia pacing for ventricular arrhythmia.

    The annual incidence of device insertion for bradycardia was 1.07 per 100 person years (CI 0.57 to 1.84). There were no differences in the incidence of device insertion for bradycardia between the classical and cardiac variants. Age at first evaluation, male sex, MSSI score, indexed LV mass, PR interval and QRS duration were independent univariable predictors for requiring a device for bradycardia. In a multivariable analysis age and QRS duration remained the only significant predictors. The presence of a cardiac variant was not a predictor of device implantation. In total 18 of the 207 (8.7%) of the patients in the study received a device for bradycardia either prior to or during the study follow-up.

    Cardiac death

    One individual was censored for an appropriate ICD discharge prior to the baseline study. There were a total of seven cardiac deaths with annual incidence of 0.52 per 100 person years (CI 0.21 to 1.06) during the follow-up period: five sudden cardiac deaths and two heart failure related deaths. Characteristics of those experiencing cardiac related mortality are displayed in table 6. Two of these deaths were in patients with cardiac variants. All the deaths occurred in men aged 40 years or more (range 41–85 years). The incidence of cardiac death was similar in patients with cardiac and classical variants. In the time to event analysis, age at evaluation, male sex, MSSI score, LA diameter, indexed LV mass and QRS duration were univariable predictors of death. Indexed LV mass was the only predictor following multivariable analysis. The presence of a cardiac variant was not a predictor of cardiac death. Five individuals including one with a cardiac variant experienced non-cardiac death.

    View this table:
    Table 6

    Characteristic of patients experiencing cardiac related mortality

    Discussion

    In this study, we found that individuals with AFD have considerable cardiac morbidity and a high cardiac mortality. Adverse outcomes were associated with advancing age and global severity of AFD. There was no relation between genotype (classical vs cardiac) and the primary end-point.

    Cardiac disease in AFD

    Numerous studies have shown that the heart and major vessels are commonly involved in patients with AFD and that the majority of affected individuals experience cardiovascular symptoms.21 Accumulation of the principal substrate of α galactosidase A, Gb3, is seen in cardiomyocytes, the specialised conduction tissue, vascular endothelium and in valvular fibroblasts. Together, these abnormalities contribute to a complex pathophysiology that includes progressive LV diastolic and systolic dysfunction, bradyarrhythmia and tachyarrhythmia, microvascular ischaemia and valve dysfunction.20 ,22 With advancing age, these abnormalities are associated with secondary changes such as myocardial fibrosis23 ,24 that promote further deterioration in cardiac function and arrhythmia.

    Clinical predictors of cardiac events

    Comparison of cardiac outcomes with other investigations is made difficult by the use of different end points and variable inclusion criteria of patients on ERT. A summary of other studies evaluating cardiac outcomes is provided in the online supplementary table S2. The importance of disease duration and overall severity in determining adverse events in patients with AFD is reflected in the association of age and MSSI score with the primary end-point. Age was also an independent predictor of device insertion for bradycardia reflecting the progressive nature of conduction system dysfunction in AFD. Indexed LV mass was an independent predictor AF and cardiac death, confirming the importance of cardiac hypertrophy in risk prediction.

    We have shown previously that QRS duration is a predictor for pacemaker insertion,25 but in this study QRS duration was also a univariable predictor of cardiac death and AF as well as being an independent predictor for the composite end-point suggesting that it is a useful surrogate for the severity of cardiac involvement more generally.

    Of the seven cardiac deaths that occurred during the follow-up period, all were in male patients. In a time to event analysis, male sex was univariate predictor for cardiac death, suggesting that men were more vulnerable from early mortality from cardiac disease; however, this association did not reach statistical significance following multivariable analysis.

    Influence of genotype on outcomes

    While the existence of organ specific subtypes of AFD remains controversial, it is clear that cardiac involvement can be the predominant clinical manifestation of disease in some individuals.3–11 At baseline the cardiac variant cohort had lower MSSI scores despite being older. Individuals with cardiac variants also had longer PR interval and more pacemakers for bradyarrhythmia at baseline, suggesting that individuals with cardiac variants may be more prone to develop clinically significant conduction disease. However, during follow-up the primary and secondary outcome was similar in individuals with cardiac and classical genetic variants.

    Limitations

    The implantation of ICDs, biventricular pacemakers and devices for other indications, and censoring of these individuals may underestimate the incidence of device implantation for antibradycardia pacing. The small number of endpoints for individual outcomes makes statistical modelling prone to overfitting and did not allow for the effect of ERT to be examined. The classification of variants into cardiac and non-cardiac may be simplistic in some cases as some individual mutations are described in patients with both phenotypes.7 ,26 ,27 Nevertheless, most data suggest that a cardiac phenotype predominates in patients with the mutations selected in this study.

    Conclusions

    AFD is associated with a high burden of cardiac morbidity and mortality. Adverse cardiac outcomes are associated with age, global disease severity and advanced cardiac disease but not the presence of cardiac genetic variants.

    Key messages

    What is already known on this subject?

    • Anderson–Fabry Disease (AFD) is an X linked lysosomal storage disorder caused by mutations in the α-galactosidase A gene (GLA). It is a multisystem disorder and cardiac manifestations include LV hypertrophy, conduction system disease, valve dysfunction, and arrhythmias. Cardiac involvement contributes to substantial morbidity and mortality and some GLA mutations, ‘cardiac variants’, are associated with prominent and, in many cases, exclusive cardiac involvement. It is not known whether patients with these variants have a different prognosis from those with mutations that cause a classical multisystem presentation.

    What might this study add?

    • In this study we show that the incidence of the primary endpoint (a composite of new onset atrial fibrillation, New York Heart Association functional class ≥3 heart failure symptoms, device insertion for bradycardia and cardiac death) was 2.64 per 100 person-years (CI 1.78 to 3.77). Age (HR 1.04, CI 1.01 to 1.08, p=0.004), Mainz Severity Score Index score (HR 1.05, CI 1.01 to 1.09, p=0.012) and QRS duration (HR 1.03, CI 1.00 to 1.05, p=0.020) were independent predictors of the primary endpoint. The presence of a cardiac genetic variant was not a significant predictor of the primary end point.

    How might this impact on clinical practice?

    • This study provides further demonstration that individuals with AFD have considerable cardiac morbidity and mortality irrespective of the underlying mutation. The results of this study facilitate the identification of ‘high-risk’ individuals who warrant closer clinical surveillance.

    References

    1. ↵
      1. Zarate YA,
      2. Hopkin RJ
      . Fabry's disease. Lancet 2008;372:1427–35. doi:10.1016/S0140-6736(08)61589-5
      OpenUrlCrossRefPubMedWeb of Science
    2. ↵
      1. Mehta A,
      2. Ricci R,
      3. Widmer U, et al
      . Fabry disease defined: baseline clinical manifestations of 366 patients in the Fabry Outcome Survey. Eur J Clin Invest 2004;34:236–42. doi:10.1111/j.1365-2362.2004.01309.x
      OpenUrlCrossRefPubMedWeb of Science
    3. ↵
      1. Nakao S,
      2. Takenaka T,
      3. Maeda M, et al
      . An atypical variant of Fabry's disease in men with left ventricular hypertrophy. N Engl J Med 1995;333:288–93. doi:10.1056/NEJM199508033330504
      OpenUrlCrossRefPubMedWeb of Science
    4. ↵
      1. Eng CM,
      2. Resnick-Silverman LA,
      3. Niehaus DJ, et al
      . Nature and frequency of mutations in the alpha-galactosidase A gene that cause Fabry disease. Am J Hum Genet 1993;53:1186–97.
      OpenUrlPubMedWeb of Science
    5. ↵
      1. Elliott PM
      . Anderson-Fabry Disease: an important differential diagnosis in patients with unexplained left ventricular hypertrophy. Heart Metab 2006;33:25–9.
      OpenUrl
    6. ↵
      1. von Scheidt W,
      2. Eng CM,
      3. Fitzmaurice TF, et al
      . An atypical variant of Fabry's disease with manifestations confined to the myocardium. N Engl J Med 1991;324:395–9. doi:10.1056/NEJM199102073240607
      OpenUrlCrossRefPubMedWeb of Science
    7. ↵
      1. Ishii S,
      2. Sakuraba H,
      3. Suzuki Y
      . Point mutations in the upstream region of the alpha-galactosidase A gene exon 6 in an atypical variant of Fabry disease. Hum Genet 1992;89:29–32. doi:10.1007/BF00207037
      OpenUrlCrossRefPubMed
    8. ↵
      1. Eng CM,
      2. Ashley GA,
      3. Burgert TS, et al
      . Fabry disease: thirty-five mutations in the alpha-galactosidase A gene in patients with classic and variant phenotypes. Mol Med 1997;3:174–82.
      OpenUrlPubMedWeb of Science
    9. ↵
      1. Eng CM,
      2. Niehaus DJ,
      3. Enriquez AL, et al
      . Fabry disease: twenty-three mutations including sense and antisense CpG alterations and identification of a deletional hot-spot in the alpha-galactosidase A gene. Hum Mol Genet 1994;3:1795–9. doi:10.1093/hmg/3.10.1795
      OpenUrlAbstract/FREE Full Text
    10. ↵
      1. Frustaci A,
      2. Chimenti C,
      3. Ricci R, et al
      . Improvement in cardiac function in the cardiac variant of Fabry's disease with galactose-infusion therapy. N Engl J Med 2001;345:25–32. doi:10.1056/NEJM200107053450104
      OpenUrlCrossRefPubMedWeb of Science
    11. ↵
      1. Ishii S,
      2. Nakao S,
      3. Minamikawa-Tachino R, et al
      . Alternative splicing in the alpha-galactosidase A gene: increased exon inclusion results in the Fabry cardiac phenotype. Am J Hum Genet 2002;70:994–1002. doi:10.1086/339431
      OpenUrlCrossRefPubMedWeb of Science
    12. ↵
      1. Beck M
      . The Mainz Severity Score Index (MSSI): development and validation of a system for scoring the signs and symptoms of Fabry disease. Acta Paediatr Suppl 2006;95:43–6. doi:10.1080/08035320600618825
      OpenUrlCrossRefPubMed
    13. ↵
      1. Lang RM,
      2. Bierig M,
      3. Devereux RB, et al
      . Recommendations for chamber quantification. Eur J Echocardiogr 2006;7:79–108. doi:10.1016/j.euje.2005.12.014
      OpenUrlCrossRefPubMedWeb of Science
    14. ↵
      1. Teicholz LE,
      2. Cohen MV,
      3. Sonnenblick EH, et al
      . Study of left ventricular geometry and function by B-scan ultrasonography in patients with and without asynergy. N Engl J Med 1974;291:1220–6. doi:10.1056/NEJM197412052912304
      OpenUrlCrossRefPubMedWeb of Science
    15. ↵
      The Criteria Committee of the New York Heart Association. Nomenclature and Criteria for Diagnosis of Diseases of the Heart and Great Vessels. 9th edn. Boston: Mass: Little, Brown & Co, 1994.
    16. ↵
      1. Elliott PM,
      2. Poloniecki J,
      3. Dickie S, et al
      . Sudden death in hypertrophic cardiomyopathy: identification of high risk patients. J Am Coll Cardiol 2000;36:2212–18. doi:10.1016/S0735-1097(00)01003-2
      OpenUrlCrossRefPubMedWeb of Science
    17. ↵
      1. Firth D
      . Bias Reduction of Maximum Likelihood Estimates. Biometrika 1993;80:27–38. doi:10.1093/biomet/80.1.27
      OpenUrlAbstract/FREE Full Text
    18. ↵
      1. Heinze G,
      2. Schemper M
      . A Solution to the Problem of Monotone Likelihood in Cox Regression. Biometrics 2001;57:114–19. doi:10.1111/j.0006-341X.2001.00114.x
      OpenUrlCrossRefPubMedWeb of Science
    19. ↵
      1. Harrell FE Jr.
      . Regression Modeling Strategies: With Applications to Linear Models, Logistic Regression, and Survival Analysis, 2010.
    20. ↵
      1. Linhart A,
      2. Elliott PM
      . The heart in Anderson-Fabry disease and other lysosomal storage disorders. Heart 2007;93:528–35. doi:10.1136/hrt.2005.063818
      OpenUrlFREE Full Text
    21. ↵
      1. Linhart A,
      2. Kampmann C,
      3. Zamorano JL, et al
      . Cardiac manifestations of Anderson-Fabry disease: results from the international Fabry outcome survey. Eur Heart J 2007;28:1228–35. doi:10.1093/eurheartj/ehm153
      OpenUrlAbstract/FREE Full Text
    22. ↵
      1. Elliott PM,
      2. Kindler H,
      3. Shah JS, et al
      . Coronary microvascular dysfunction in male patients with Anderson-Fabry disease and the effect of treatment with alpha galactosidase A. Heart 2006;92:357–60. doi:10.1136/hrt.2004.054015
      OpenUrlAbstract/FREE Full Text
    23. ↵
      1. Moon JC,
      2. Sheppard M,
      3. Reed E, et al
      . The histological basis of late gadolinium enhancement cardiovascular magnetic resonance in a patient with Anderson-Fabry disease. J Cardiovasc Magn Reson 2006;8:479–82. doi:10.1080/10976640600605002
      OpenUrlCrossRefPubMedWeb of Science
    24. ↵
      1. Moon JC,
      2. Sachdev B,
      3. Elkington AG, et al
      . Gadolinium enhanced cardiovascular magnetic resonance in Anderson-Fabry disease. Evidence for a disease specific abnormality of the myocardial interstitium. Eur Heart J 2003;24:2151–5. doi:10.1016/j.ehj.2003.09.017
      OpenUrlAbstract/FREE Full Text
    25. ↵
      1. O'Mahony C,
      2. Coats C,
      3. Cardona M, et al
      . Incidence and predictors of anti-bradycardia pacing in patients with Anderson-Fabry disease. Europace 2011;13:1781–8.
    26. ↵
      1. Germain DP,
      2. Poenaru L
      . Fabry disease: identification of novel alpha-galactosidase A mutations and molecular carrier detection by use of fluorescent chemical cleavage of mismatches. Biochem Biophys Res Commun 1999;257:708–13. doi:10.1006/bbrc.1999.0310
      OpenUrlCrossRefPubMedWeb of Science
    27. ↵
      1. Rodriguez-Mari A,
      2. Coll MJ,
      3. Chabas A
      . Molecular analysis in Fabry disease in Spain: fifteen novel GLA mutations and identification of a homozygous female. Hum Mutat 2003;22:258. doi:10.1002/humu.9172
      OpenUrlPubMed

    Supplementary materials

    • Supplementary Data

      This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

      Files in this Data Supplement:

    Footnotes

    • Contributors All authors have been involved in one or more of the following: Conception and design, acquisition of data or analysis and interpretation of data, drafting the article or revising it critically for important intellectual content, final approval of the enclosed manuscript.

    • Funding This research was supported by grants from the British Heart Foundation (to Caroline Coats). This work was undertaken at UCLH/UCL who received a proportion of funding from the Department of Health's NIHR Biomedical Research Centres funding scheme.

    • Competing interests VP: Received travel grants from Shire HGT, outside the submitted work. CO: Received travel grants from Shire HGT, outside the submitted work. DH: Reports grants and personal fees from Shire HGT, Genzyme Sanofi, Protaliz and Amicus, outside the submitted work. EM: Reports unrestricted educational grants from both Shire HGT and Genzyme Sanofi, outside the submitted work. RL: Reports personal fees and non-financial support from Genzyme Sanofi and grants and personal fees from Shire HGT, outside the submitted work. AM: Reports grants and personal fees from Genzyme Sanofi and grants and personal fees from Shire HGT, outside the submitted work. PME: Reports speaker fees from Shire HGT and consultancy and speaker fees from Genzyme Sanofi, Pfizer and Gilead Sciences.

    • Provenance and peer review Not commissioned; externally peer reviewed.

    Linked Articles