Tratamiento de sustitución enzimática para la enfermedad de Anderson‐Fabry
Resumen
Antecedentes
La enfermedad de Anderson‐Fabry es un defecto del metabolismo de los glucoesfingolípidos vinculado al cromosoma X. La insuficiencia renal progresiva es una causa importante de morbilidad, las complicaciones adicionales resultan del compromiso cardio y cerebrovascular. La supervivencia es reducida entre los hombres afectados y las mujeres sintomáticas portadoras.
Ésta es una actualización de una revisión Cochrane publicada por primera vez en 2010 y actualizada previamente en 2013.
Objetivos
Evaluar la efectividad y la seguridad del tratamiento de sustitución enzimática en comparación con otras intervenciones, placebo o ninguna intervención, para tratar la enfermedad de Anderson‐Fabry.
Métodos de búsqueda
Se realizaron búsquedas en el registro de ensayos de errores innatos de metabolismo del Grupo de Fibrosis Quística y Trastornos Genéticos (fecha de la búsqueda más reciente: 08 de julio de 2016). También se buscó "Clinical Trials" en The Cochrane Library, MEDLINE, Embase y LILACS (fecha de la búsqueda más reciente: 24 de septiembre 2015).
Criterios de selección
Ensayos controlados aleatorizados de la agalsidasa alfa o beta en participantes con diagnóstico de enfermedad de Anderson‐Fabry.
Obtención y análisis de los datos
Dos autores seleccionaron los ensayos pertinentes, evaluaron la calidad metodológica y extrajeron los datos.
Resultados principales
Nueve ensayos que comparaban la agalsidasa alfa o beta en 351 participantes cumplieron los criterios de selección.
En ambos ensayos que comparaban la agalsidasa alfa con el placebo se informó sobre la concentración de globotriaosilceramida en el plasma y los tejidos; los resultados globales no fueron significativos. Un ensayo informó las puntuaciones de dolor medidas por la gravedad del Brief Pain Inventory, hubo una mejoría estadísticamente significativa para los participantes que recibieron tratamiento hasta los tres meses, diferencia de medias ‐2,10 (intervalo de confianza del 95%: ‐3,79 a ‐0,41; hasta los cinco meses, diferencia de medias ‐1,90 (intervalo de confianza del 95%: ‐3,65 a ‐0,15); y hasta los seis meses, diferencia de medias ‐2,00 (intervalo de confianza del 95%: ‐3,66 a ‐0,34). Hubo una diferencia significativa en la calidad de vida relacionada con el dolor del Brief Pain Inventory a más de cinco y hasta seis meses, diferencia media ‐2,10 (intervalo de confianza del 95%: ‐3,92 a ‐0,28), pero no en otros puntos temporales. La muerte no fue un resultado en ninguno de los ensayos.
Uno de los tres ensayos que comparaban la agalsidasa beta con el placebo informó sobre la concentración de globotriaosilceramida en el plasma y el tejido y mostró una mejoría significativa: riñón, diferencia de medias ‐1,70 (intervalo de confianza del 95%: ‐2,09 a ‐1,31); corazón, diferencia de medias ‐0,90 (intervalo de confianza del 95%: ‐1,18 a ‐0,62); y resultados compuestos (complicaciones renales, cardíacas y cerebrovasculares y muerte), diferencia de medias ‐4,80 (intervalo de confianza del 95%: ‐5,45 a ‐4,15). No hubo diferencias significativas entre los grupos en cuanto a la muerte; ningún ensayo informó sobre el dolor.
Sólo dos ensayos compararon la agalsidasa alfa con la agalsidasa beta. Uno de ellos no mostró diferencias significativas entre los grupos en cuanto a los eventos adversos, riesgo relativo 0,36 (intervalo de confianza del 95%: 0,08 a 1,59), o cualquier evento adverso grave; riesgo relativo 0,30; (intervalo de confianza del 95%: 0,03 a 2,57).
Dos ensayos compararon diferentes esquemas de dosificación de la agalsidasa alfa. En uno de ellos se utilizaron tres dosis diferentes (0,2 mg/kg cada dos semanas; 0,1 mg/kg semanalmente y; 0,2 mg/kg semanalmente), en el otro ensayo se evaluaron otras dos dosis de los esquemas de dosificación: 0.4 mg/kg cada semana y cada dos semanas. Ambos ensayos no mostraron diferencias significativas con los diversos programas de dosificación de los niveles de globotriaosilceramida. No se encontraron diferencias significativas entre los calendarios para el resultado de eficacia principal del estado de salud autoevaluado, o para las puntuaciones de dolor.
Un ensayo que comparaba la agalsidasa alfa con la agalsidasa beta no mostró ninguna diferencia significativa para ningún evento adverso como la disnea y la hipertensión.
En general, la calidad metodológica de los ensayos incluidos era incierta en lo que respecta a la generación de la secuencia aleatoria y la ocultación de la asignación.
Conclusiones de los autores
Los ensayos que comparan la terapia de sustitución de enzimas con el placebo muestran una mejora significativa con la terapia de sustitución de enzimas en lo que respecta a los depósitos endoteliales microvasculares de globotriaosilceramida y en la calidad de vida relacionada con el dolor. Sin embargo, no hay evidencia que identifique si la forma alfa o beta es superior o la dosis o frecuencia óptima de la terapia de reemplazo de enzimas. En lo que respecta a la seguridad, los eventos adversos (es decir, rigidez, fiebre) fueron más significativos en la agalsidasa beta en comparación con el placebo. Queda por establecer la influencia a largo plazo de la terapia de sustitución de enzimas en el riesgo de morbilidad y mortalidad relacionado con la enfermedad de Anderson‐Fabry. Esta revisión destaca la necesidad de continuar la investigación sobre el uso del tratamiento de sustitución enzimática para la enfermedad de Anderson‐Fabry.
PICO
Resumen en términos sencillos
Tratamiento de la enfermedad de Anderson‐Fabry
Antecedentes
La enfermedad de Anderson‐Fabry, un raro trastorno, es causada por una deficiencia de la enzima alfa‐galactosidasa A. Esto lleva a la acumulación de un material graso llamado globotriaosilceramida en varias células del cuerpo. La globotriaosilceramida está formada por tres azúcares y una sustancia grasa llamada ceramida, y se encuentra en la mayoría de las células del cuerpo. Los individuos no tratados pueden sufrir de dolor, problemas de piel, ojos y gastrointestinales. La enfermedad de Fabry puede causar complicaciones potencialmente mortales como daño renal, ataque cardíaco y accidente cerebrovascular. Un tipo de tratamiento disponible es la terapia de reemplazo de enzimas con agalsidasa alfa o beta, que reemplaza la enzima faltante o deficiente.
Fecha de la búsqueda
La evidencia está actualizada hasta: 08 de julio de 2016.
Características de los estudios
Nueve estudios incluyeron 351 participantes. Los estudios utilizaron diferentes formulaciones de la enzima, Agalsidase alfa o beta, y las compararon con un placebo (un tratamiento "falso") o entre sí. También se hizo una comparación con respecto a los diferentes esquemas de dosificación.
Resultados clave
Dos estudios que comparaban la agalsidasa alfa con el placebo informaron sobre la concentración de globotriaosilceramida en el plasma. Los efectos combinados no fueron significativos entre el grupo de tratamiento y el de placebo. El estudio que informó sobre el dolor y la calidad de vida relacionada con el dolor mostró una mejora en los participantes que recibieron tratamiento durante el período de observación de seis meses. La muerte no fue un resultado en ninguno de los dos estudios.
Uno de los tres estudios que comparaban la agalsidasa beta con el placebo informó sobre la globotriaosilceramida y mostró una mejora en los resultados de los riñones, el corazón y los resultados compuestos. No hubo diferencias significativas en cuanto a la muerte y ningún estudio informó sobre el dolor.
Sólo dos estudios compararon la agalsidasa alfa con la agalsidasa beta. Uno de ellos no mostró ninguna diferencia significativa para ningún evento adverso como la disnea, la hipertensión y los síntomas gastrointestinales; no se trata de eventos adversos, ya que los problemas gastrointestinales son en realidad un síntoma, como puede ser la hipertensión en el contexto de la enfermedad renal.
Dos estudios compararon diferentes esquemas de dosificación de la agalsidasa alfa. No se encontraron diferencias entre los esquemas de autoevaluación del estado de salud o de las puntuaciones de dolor.
En resumen, los estudios que comparan la terapia de sustitución de enzimas con el placebo muestran resultados significativos en lo que respecta a los depósitos endoteliales microvasculares de globotriaosilceramida y en la calidad de vida relacionada con el dolor. Sin embargo, no hay evidencia que identifique si la forma alfa o beta es superior, aunque los ensayos incluidos fueron pequeños en cuanto al tamaño de la muestra. En lo que respecta a la seguridad, los eventos adversos (es decir, rigidez, fiebre) fueron más significativos con la agalsidasa beta en comparación con el placebo.
Calidad de la evidencia
A partir de la información disponible en la mayoría de los informes de los estudios, no se pudo juzgar claramente si todos los voluntarios tenían las mismas posibilidades de estar en cualquiera de los grupos de tratamiento y si habrían sabido de antemano o durante el estudio qué tratamiento estaban recibiendo.
Authors' conclusions
Summary of findings
| Enzyme replacement therapy for Anderson‐Fabry disease | ||||||
| Patient or population: Anderson‐Fabry disease Settings: outpatients Intervention: agalsidase alfa Comparison: placebo | ||||||
| Outcomes | Illustrative comparative risks (95% CI) | Relative effect | No of Participants | Quality of the evidence | Comments | |
| Assumed risk | Corresponding risk | |||||
| Placebo | Agalsidase alfa | |||||
| Changes in globotriaosylceramide (Gb3) concentration in plasma Follow up: 6 months (umol/l) | The mean in changes in Gb3 concentration in plasma score ranged between 0.63 and 10.19 umol/l | The mean in changes in Gb3 concentration in plasma score in the intervention group was on average 2.07 fewer | NA | 39 2 studies | ⊕⊝⊝⊝ | We found no statistically significant differences between studied groups in changes from baseline for Gb3for which data were available. |
| Death | Outcome not reported | Outcome not reported. | Not estimable | NA | NA | |
| Pain The Brief Pain Inventory severity Follow up: over 5 months and up to 6 months | The mean for change in pain by the BPI severity was 4.7 (SD 2.25) | The mean for changes in pain by the BPI severity in the intervention group was on average 2.00 fewer (3.66 fewer to 0.34 fewer) | NA | 26 1 study | ⊕⊕⊝⊝ | We found a statistically significant differences favouring agalsidase alfa compared to placebo in pain for which data were available. |
| Renal function by creatinine clearance Follow up: at up to 6 months (mg/dL) | The mean in changes in renal function was 84.5 (SD: 35.15) mg/dL | The mean for changes in renal function in the intervention group was on average 10.30 more (15.37 fewer to 35.97 more) mg/dL | NA | 24 1 study | ⊕⊝⊝⊝ | We found no statistically significant differences between studied groups in renal function for which data were available. |
| Symptoms and complications of disease:cardiac events | Outcome not reported. | Outcome not reported. | Not estimable | NA | NA | |
| Symptoms and complications of disease: cerebrovascular events | Outcome not reported. | Outcome not reported. | Not estimable | NA | NA | |
| *The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% CI) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). | ||||||
| GRADE Working Group grades of evidence | ||||||
| 1. We downgraded the quality of evidence by 1 for high risk due to conflict of interest in one trial. 2. We downgraded the quality of evidence by 1 due to imprecision. 3. We downgraded the quality of evidence by 1 due to inconsistency. 4. We downgraded the quality of evidence by 2 due to only one single study reporting this outcome (lack of statistical power). | ||||||
| Enzyme replacement therapy for Anderson‐Fabry disease | ||||||
| Patient or population: Anderson‐Fabry disease Settings: outpatients Intervention: agalsidase beta Comparison: placebo | ||||||
| Outcomes | Illustrative comparative risks (95% CI) | Relative effect | No of Participants | Quality of the evidence | Comments | |
| Assumed risk | Corresponding risk | |||||
| Placebo | Agalsidase beta | |||||
| Changes in globotriaosylceramide (Gb3) concentration in plasma Follow up: up to 18 months (umol/l) | The mean in changes in Gb3 concentration in plasma score was 5.5 (SD: 1.6) umol/l | The mean in changes in Gb3 concentration in plasma score in the intervention group was on average 4.80 fewer | NA | 58 1 study | ⊕⊝⊝⊝ | A statistically significant improvement on agalsidade beta compared to placebo was observed also observed in the domains of kidney (mean difference ‐1.70, (95% CI ‐2.09 to ‐1.31)) and heart (mean difference ‐0.90, 95% CI (‐1.18 to ‐0.62)) . |
| Death Follow up: 18 months | 0 per 1000 | 0 per 1000 (0 to 0) | 1.85 (0.08 to 43.96) | 82 1 study | ⊕⊝⊝⊝ | We found no statistically significant differences between studied groups in death which data were available. |
| Pain | Outcome not reported. | Outcome not reported. | Not estimable | NA | NA | |
| Renal function events (ITT population) Follow up: 18 months | 225 per 1000 | 195 (83 to 459) | 0.87 (0.37 to 2.04) | 82 1 study | ⊕⊝⊝⊝ | We found no statistically significant differences between studied groups in renal function events which data were available. |
| Symptoms and complications of disease: cardiac events Follow up: up to 18 months | 129 per 1000 | 59 (14 to 245) | 0.46 (0.11 to 1.90) | 82 1 study | ⊕⊝⊝⊝ | We found no statistically significant differences between studied groups in cardiac events which data were available. |
| Symptoms and complications of disease: cerebrovascular events (ITT population) Follow up: 18 months | 64 per 1000 | 7 (1 to 158) | 0.12 (0.01 to 2.48) | 82 1 study | ⊕⊝⊝⊝ | We found no statistically significant differences between studied groups in cerebrovascular events which data were available. |
| *The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% CI) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). | ||||||
| GRADE Working Group grades of evidence | ||||||
| 1. We downgraded the quality of evidence by 1 for high risk due to conflict of interest. 2. We downgraded the quality of evidence by 1 due to only one single study reporting this outcome (lack of statistical power). 3. We downgraded the quality of evidence by 1 due to imprecision. | ||||||
| Enzyme replacement therapy for Anderson‐Fabry disease | ||||||
| Patient or population: Anderson‐Fabry disease Settings: outpatients Intervention: agalsidase alfa Comparison: agalsidase beta | ||||||
| Outcomes | Illustrative comparative risks (95% CI) | Relative effect | No of Participants | Quality of the evidence | Comments | |
| Assumed risk | Corresponding risk | |||||
| Agalsidase alfa | Agalsidase beta | |||||
| Changes in globotriaosylceramide (Gb3) concentration in plasma and tissue Follow up: 12 months | Outcome not reported | Outcome not reported | Not estimable | NA | NA | |
| Death Follow up: not reported | 0 per 1000 | 0 (0 to 0) | 3.00 (0.13 to 69.09) | 36 1 study | ⊕⊝⊝⊝ | We found no statistically significant differences between studied groups in death which data were available. |
| Pain Acroparaesthesia and Fabry crises Follow up: not reported | Outcome not reported | Outcome not reported | Not estimable | NA | NA | |
| Renal function Follow up: up to 24 months. | Outcome not reported | Outcome not reported | Not estimable | NA | NA | |
| Symptoms and complications of disease: cardiac events Follow up: not reported | 142 per 1000 | 66 (7 to 653) | 0.47 (CI 0.05 to 4.60) | 29 1 study | ⊕⊝⊝⊝ | We found no statistically significant differences between studied groups in death which data were available. |
| Symptoms and complications of disease: cerebrovascular events Follow up: median 23.0 months | Outcome not reported. | Outcome not reported. | Not estimable. | NA | NA | |
| *The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% CI) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). | ||||||
| GRADE Working Group grades of evidence | ||||||
| 1. We downgraded the quality of evidence by 1 for high risk due to lack of blinding, incomplete outcome data and, selective reporting. 2. We downgraded the quality of evidence by 1 due to only one single study reporting this outcome (lack of statistical power). 3. We downgraded the quality of evidence by 1 due to imprecision. | ||||||
Background
Please see glossary in the appendices for definition of terms used throughout the review (Appendix 1).
This is a second update of a Cochrane review first published in The Cochrane Library in 2010 (El Dib 2010), which included five studies that did not provide evidence on patient‐important outcome measures. The first update of this review was then published in 2013 (El Dib 2013), which included only one new study and no information was changed. For this current version, a further three clinical trials not previously reviewed were included but they did not significantly alter the conclusions.
Description of the condition
Anderson‐Fabry disease (AFD) is an X‐linked, multi‐system disorder caused by a deficiency of the lysosomal enzyme alpha–galactosidase A (AGAL). The incidence of AFD is estimated at 1 in 117,000 live births for males (Meikle 1999); although recent newborn screening surveys suggest the incidence may be much higher, up to 1 in 3,100 (Spada 2006). One reasonable explanation for this difference in the incidence of AFD between these studies is that the first (Meikle 1999) refers to the incidence of classic AFD while the latter (Spada 2006) refers to the full phenotypic spectrum reflecting the expansion in phenotypic variation identified in the last decade.Variants include those with sole involvement of the heart, kidneys or brain. For instance, the renal variant was first described in 2003. Although the vast majority of reports have focused on symptomatic males, females with AFD can develop disease‐related problems (Wilcox 2008).
Clinically, AFD is characterized by major renal, cardiac and cerebrovascular complications consequent to the progressive deposition of incompletely metabolized glycosphingolipids, mainly globotrysilceramide (Gb3) in multiple cell types. The precise mechanism of tissue injury remains unclear though alterations in vascular reactivity and a propensity for thrombo‐embolic disease are believed to play a role in the increased risk for particular problems, such as stroke. Gb3 has been used as a biomarker in AFD being measured in plasma and urine by tandem mass spectrometry. The reference ranges for plasma and urinary Gb3 levels are 5.6 (3.6‐7.5) μg/ml and 0.016 (0.01 ‐ 0.03) mg/mmol of creatinine respectively. Gb3 levels are consistently elevated in most people with AFD and decrease after treatment initiation with ERT. Thus, it has been proposed to consider Gb3 levels as part of the diagnostic process as well as a biomarker of response to ERT. This said, the reliability of Gb3 levels as a surrogate marker of disease severity and of treatment response has been questioned by some investigators (Young 2005). Recently studies have shown that a deacylated form of Gb3 (globotriaosylsphingosine, lyso‐Gb3) is elevated in plasma obtained from individuals with AFD. Lyso‐Gb3 is a potent inhibitor of AGAL and alpha–galactosidase B (N‐acetylgalactosaminidase), and it has been found to promote smooth muscle cell proliferation in vitro (Aerts 2008). Furthermore, it has been suggested lyso‐Gb3 may identify the agalA mutations leading to classic Fabry disease (Niemann 2014).
Renal and cardiac failure represent major sources of morbidity, and account for the reduced survival among affected males and females compared to the normal population (median age of death is 50 to 57 years and 70 to 72 years, respectively). The pain crises, acroparaesthesia, hearing loss and gastrointestinal problems experienced lead to significant reduction in health‐related quality of life. While the majority of reports focus on males, females with AFD can develop disease‐related problems. Clinical expression among females tends to be more variable and onset of symptoms tends to occur at a later age. The expression of the disease in females appears to be influenced by the particular AGAL mutation and the pattern of X chromosome inactivation in each organ (Niemann 2014; Wang 2007).
Description of the intervention
Enzyme replacement therapy (ERT) for AFD consists of the regular intravenous infusion of a recombinant enzyme formulation. Two forms of recombinant AGAL exist; agalsidase alfa (Replagal™), Shire Human Genetic Therapies, Cambridge, MA) and agalsidase beta (Fabrazyme®, Genzyme Corporation, Cambridge, MA). Agalsidase alfa is generated by the activation of the AGAL gene in a continuous human cell line; whereas agalsidase beta is produced in a Chinese hamster ovary (CHO) mammalian cell expression system transduced with the human AGAL sequence. Both enzyme preparations are approved in Europe and many other countries, but in the United States the Federal Drug Administration approved only agalsidase beta (Eng 2001; Schiffmann 2001). Administration of ERT is usually once every two weeks, using a dose of 0.2 mg/kg body weight when using agalsidase alfa or 1 mg/kg for agalsidase beta.
Intravenous enzyme infusions appear to be reasonably well tolerated, with reported infusion reactions of about 10%; mostly consisting of fever and transient rigours of mild to moderate intensity. A proportion of people with AFD receiving ERT have sero‐converted (i.e. developed antibodies); the frequency of antibodies against agalsidase alfa and agalsidase beta has been reported at 55% and 83% of individuals treated respectively (Eng 2001; Schiffmann 2001). Antibody formation did not influence clinical efficacy or outcomes in either of the initial clinical studies undertaken, and antibody titres usually decreased over time. In a few cases, IgE antibodies have been reported after infusion of agalsidase beta (Eng 2001). Recent studies have shown the presence of antibodies may influence the Gb3 storage in skin capillaries and Gb3 excretion in urine, although no relation between antibody formation and plasma Gb3 levels or clinical outcome has been established thus far (Hollack 2009).
How the intervention might work
Treatment of Fabry disease consists of symptom management and replacement of the deficient or dysfunctional enzyme with enzyme replacement therapy. ERT supplies the organs with recombinant enzyme and therefore reduces the amount of Gb3 accumulation in tissues with consecutive multisystem damage. ERT is available since 2001, in the form of two recombinant GLA preparations: agalsidase alfa (Replagal, Shire Human Genetic Therapies, Cambridge, MA, 0.2 mg/kg per infusion), and agalsidase beta (Fabrazyme, Genzyme Corporation, Cambridge, MA, 1 mg/kg per infusion). ERT is administered intravenously either through a peripheral line or central access device, infusions typically occur once every two weeks.
Why it is important to do this review
A systematic review is needed to establish the evidence base for the effectiveness and safety of ERT for treating AFD. This review is of particular relevance since ERT has now been available for almost 15 years and a review of the most recent data may provide more guidance in its use.
This is an update of a Cochrane review first published in 2010, and previously updated in 2013 (El Dib 2010; El Dib 2013).
Objectives
To evaluate the effectiveness and safety of ERT compared to other interventions, placebo or no interventions for treating AFD.
Methods
Criteria for considering studies for this review
Types of studies
Randomized and quasi‐randomized controlled clinical trials.
Types of participants
Individuals with AFD of any age and any degree of disease severity. Diagnosis should be established either by accepted criteria based on concentration of enzyme activity or by mutation analysis.
Types of interventions
Enzyme replacement therapy (agalsidase beta or agalsidase alfa) in any amount given for a period of at least one month compared to: each other; another intervention (fat‐restricted diet, drugs, exercises, etc); no intervention; or to placebo.
Types of outcome measures
Primary outcomes
-
Changes in globotriaosylceramide (Gb3) concentration in plasma and tissue (i.e. endothelial cells)
-
All‐cause death
-
Pain (measured by the McGill Pain Questionnaire*)
-
Acroparaesthesia (measured by the intensity and duration of acroparaesthesia as reported by the individual)
-
Fabry crises
-
*Post hoc change: we also considered other validated pain questionnaires.
Secondary outcomes
-
Effect of therapy on renal function (e.g. reduction in serum creatinine and proteinuria; creatinine and inulin clearance)
-
Symptoms and complications of disease such as occurrence of renal failure, skin, cerebrovascular and cardiac complications based on the following observations: serum creatinine level, proteinuria (ratio of urinary protein to urinary creatinine in mg/dL), 12‐lead electrocardiography, echocardiography parameters (such as thickness of cardiac structures, left ventricular volume, measures of systolic and diastolic function, heart rate), neurologic examination, head magnetic resonance imaging, exercise tolerance, and AFD symptom assessment
-
Histologic analysis of microvascular capillary endothelial deposits of Gb3 in biopsy specimens
-
Adverse effects of treatment (including severe adverse effects such as dyspnoea, malaise, hypertension and gastrointestinal symptoms, and antibody formation), based on the type and frequency of adverse events in treated participants and those on placebo.
-
Quality of life (QoL) (as determined by the Short Form 36 (SF‐36) and the impact of clinical variables on domain scores within the SF‐36* (Smith 2000).
*Post hoc change: we also considered other validated QoL questionnaires.
Search methods for identification of studies
There was no language restriction and the trials were identified from the sources listed below.
Electronic searches
Relevant trials were identified from the Cystic Fibrosis and Genetic Disorders Group's Inborn Errors of Metabolism Trials Register using the term: fabry disease.
The Register is compiled from electronic searches of the Cochrane Central Register of Controlled Trials (CENTRAL) (updated with each new issue of The Cochrane Library), weekly searches of MEDLINE and the prospective handsearching of one journal ‐ Journal of Inherited Metabolic Disease. Unpublished work were identified by searching through the abstract books of the Society for the Study of Inborn Errors of Metabolism conference and the SHS Inborn Error Review Series. For full details of all searching activities for the register, please see the relevant section of the Cystic Fibrosis and Genetic Disorders Group Module.
Date of the most recent search of the Cystic Fibrosis and Genetic Disorders Group's Inborn Errors of Metabolism Trials Register: 08 July 2016.
We also searched Cochrane Central Register of Controlled Trials (CENTRAL) (Issue 9, 2015, www.cochranelibrary.com) Ovid Embase (1980 to 24 September 2015), PubMed (1980 to 24 September 2015; www.ncbi.nlm.nih.gov/pubmed) and the Literature Latino‐Americana e do Caribe em Ciências da Saúde ‐ LILACS (1982 to 24 September 2015; http://lilacs.bvsalud.org). These search strategies are listed in the appendices (Appendix 2; Appendix 3; Appendix 4; Appendix 5).
Searching other resources
Reference lists of the identified relevant trials were scrutinized for additional citations.
Specialists in the field and authors of the included trials were contacted for any possible unpublished data.
We also searched the abstract books of WORLD LDN.
Data collection and analysis
Selection of studies
Two authors (RED and HG) independently screened the trials identified. Discrepancies were resolved by discussion.
Data extraction and management
Two authors (RED and HG) independently extracted data and discrepancies were resolved by discussion. We used a standard form to extract the following information: characteristics of the trial (design, methods of randomisation), participants, interventions and outcomes (types of outcome measures, timing of outcomes, adverse events).
We presented different enzyme therapies (alfa and beta) as well as different control groups (placebo or active controls) as separate interventions as we did not judge these sufficiently comparable to combine.
When possible, outcome data were grouped into those measured at up to one month, over one month and up to three months, over three months and up to six months and over six months and up to 12 months and annually thereafter. In a post hoc change, we also considered other time points such as 'over five months and up to six months' where multiple data sets from an individual trial were available for a single planned time point.
Assessment of risk of bias in included studies
We assessed every trial using a simple form and followed the domain‐based evaluation as described in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011).
We assessed the following domains as having either low, unclear or high risk of bias:
-
randomisation
-
concealment of allocation
-
blinding (of participants, caregivers and outcome assessors)
-
incomplete outcome data
-
selective reporting
-
other potential sources of bias
1. Randomisation
Low risk: adequate generation of allocation
Unclear risk: not described in the paper or by contacting authors
High risk: inadequate generation of allocation
2. Concealment of allocation
Low risk: adequate allocation concealment
Unclear risk: not described in the paper or by contacting authors
High risk: inadequate allocation concealment
3. Blinding (of participants, caregivers and outcome assessors)
Low risk: adequately described; adequate method
Unclear risk: described as blinded, but no information provided
High risk: not blinded; inadequate method
If a trial measured an objective outcome then lack of blinding was rated as low risk of bias.
4. Incomplete outcome data
We recorded the rates of withdrawal for each outcome as follows.
Low risk: less than 20% and equal for both groups;
Unclear risk: not reported in paper or by authors;
High risk: greater than 20% or not equal for any comparison groups or both.
5. Selective reporting
We considered the possibility of selective reporting of outcomes when data were not available in the 'Results' section of the published paper, but according to the 'Methods' section these outcome data were collected.
6. Other potential sources of bias
We recorded any other factors we felt might cause bias.
Measures of treatment effect
For dichotomous data (such as death, presence or absence of Fabry crises and adverse effects), we used the risk ratio (RR), with 95% confidence intervals (95% CIs) as the effect measure.
For continuous data (such as changes in Gb3 concentration, echocardiographic parameters, pain and quality of life scores) we used the mean difference (MD) (in which the effect estimates of individual trials are weighted by dispersion measures), with 95% CIs. When standard errors (SE) were reported within the trial reports, we converted these to standard deviations (SD) (SD = SE x square root of n). For future updates of this review we plan to use standardized mean difference (SDM).
For time‐to‐event data (such as effects on renal function) we planned to consider a joint model for longitudinal and discrete time‐to‐event data in which the discrete event time distribution is modelled as a linear function of the slope of an individual's longitudinal process on the probit scale (Jones 2009). Effect of therapy on renal function (e.g. increase in serum creatinine and proteinuria; creatinine and inulin clearance) could be measured as time‐to‐event‐data or number of events. However, the only trial reporting on this outcome only reported on one time point, which has been presented as continuous data (Schiffmann 2001).
Unit of analysis issues
We have included data from an eligible cross‐over trial and analysed this according to a method recommended by Elbourne, whereby we inserted the mean and SD of the participant‐specific differences between the intervention and control measurements taking all these measurements from intervention periods and all measurements from control periods and, we analysed these as if the trial was a parallel group trial of intervention versus control group (Elbourne 2002). We appreciate that this is a conservative method and will overestimate the SE of the MD by treating (paired) within differences as independent data.
Dealing with missing data
For any new trial which may be included in a future update of this review, we will contact authors, as necessary, to clarify methodological issues (such as generation and allocation concealment, blind method and withdrawals) as well as for any missing outcome data.
Assessment of heterogeneity
We planned to qualify inconsistency among the pooled estimates using the I2 statistic. This illustrates the percentage of the variability in effect estimates resulting from heterogeneity rather than sampling error (Higgins 2003; Higgins 2011). We considered there to be a low degree of heterogeneity if I2 was below 25%, a moderate degree if I2 between 25% and 50%, and a high degree if I2 was over 50%.
Assessment of reporting biases
We planned to assess publication bias by drawing a funnel plot (trial effect versus trial size), but there were not sufficient trials included in the review. In future analysis, funnel plots will be used in an exploratory data analysis to assess for the potential existence of small trial bias. There are a number of explanations for the asymmetry of a funnel plot, including true heterogeneity of effect with respect to trial size, poor methodological design of small trials (Sterne 2001) and publication bias and selective reporting of outcomes. Thus, this exploratory data tool may be misleading (Tang 2000; Thornton 2000) and we will not place undue emphasis on this tool.
Data synthesis
We used the fixed‐effect model to analyse data, if in future updates, significant heterogeneity (e.g. I2 higher than 50%) is identified, we will compute pooled estimates of the treatment effect for each outcome under a random‐effects model (with two or more trials).
Subgroup analysis and investigation of heterogeneity
For future reviews (including 10 or more trials) we plan to investigate heterogeneity by conducting meta‐analyses by subgroups as described below.
-
Dose. We plan to sub‐divide dose in to lower dose (less than 0.2 mg/kg) and higher dose (more than 0.2 mg/kg). These cut off doses were chosen arbitrarily.
-
Duration of intervention. We plan to sub‐divide duration of intervention up to one year and more than one year. These time divisions were chosen arbitrarily.
-
Age at commencement of treatment, for example, up to the age of 18 years and over age 18 years.
-
Severity of disease as stratified into less severe and more severe (as reported within the trials).
Sensitivity analysis
If we had included an adequate number of trials, we planned to perform a sensitivity analysis to test the robustness of analyses to certain assumptions in the results. In future analysis, we will include the following factors in the sensitivity analysis, separating trials according to:
-
allocation concealment quality (low risk, high risk or unclear risk);
-
blinding of participants, caregiver and outcome assessment (low risk, high risk or unclear risk);
-
rates of withdrawal for each outcome.
Summary of findings tables
In our review, we used the principles of the GRADE (Grades of Recommendation, Assessment, Development and Evaluation) system (Guyatt 2008) to assess the quality of the body of evidence associated with specific outcomes (changes in globotriaosylceramide (Gb3) concentration in plasma and tissue; death; pain; renal function; and symptoms and complications of disease (i.e., cardiac and cerebrovascular events)) and constructed a summary of findings (SoF) table using GRADE software. The GRADE approach appraises the quality of a body of evidence according to the extent to which one can be confident that an estimate of effect or association reflects the item being assessed. Assessment of the quality of a body of evidence considers within‐trial risk of bias (methodological quality), directness of the evidence, heterogeneity of the data, precision of effect estimates and risk of publication bias. The quality of the evidence for a specific outcome will be altered by a level according to the performance of trials against these five factors.
High‐quality evidence: findings are consistent among at least 75% of RCTs with low risk of bias; data are consistent, direct and precise, and no publication biases are known or suspected. Further research is unlikely to change the estimate or our confidence in the results.
Moderate‐quality evidence: one of the domains is not met. Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate.
Low‐quality evidence: two of the domains are not met. Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate.
Very low‐quality evidence: three of the domains are not met. We are very uncertain about the results.
No evidence: no RCTs that addressed this outcome were identified.
Results
Description of studies
See: Characteristics of included studies; Characteristics of excluded studies.
Results of the search
For the original review, we identified 273 references from the electronic searches (El Dib 2010). After initial assessment of these references, 56 were looked at in more detail. A total of 31 were excluded from the review (Excluded studies). Thus, five trials (Banikazemi 2007; Bierer 2006; Hughes 2008; Schiffmann 2001; Eng 2001) (represented by 25 individual references), which enrolled 187 participants were included (Included studies).
For the first update of the review in 2012, 368 references were identified by the searches. After an initial assessment seven references were considered. Of these we included one new trial (Vedder 2007) (represented by five individual references) and, excluded two further trials (Fernhoff 2011; West 2011).
For the second update of the review in 2016, 19 references were identified. We selected all references for careful consideration and obtained them in full text, where available. Following assessment of the full articles, we included three new clinical trials (Clarke 2007; Hughes 2013; Sirrs 2014) (represented by four individual references) and, we excluded 12 further studies (Benjamin 2014; Fellgiebel 2014; Germain 2013; Hughes 2014; Kim 2014; Rombach 2012; Schiffmann 2013; Schiffmann 2014; Sirrs 2011; Tøndel 2013; Tsuboi 2014; Weidemann 2014). One further trial is currently listed in 'Studies awaiting classification' (represented by three individual references) and will be fully assessed in the next update of this review (Wijburg 2015).
Therefore, for this second update we have a total of nine included trials (Banikazemi 2007; Bierer 2006; Clarke 2007; Eng 2001; Hughes 2008; Hughes 2013; Schiffmann 2001; Sirrs 2014; Vedder 2007) (represented by 34 individual references) (Included studies) involving a total of 351 participants (Figure 1), 43 excluded trials (Excluded studies) and three trials awaiting assessment (Benjamin 2014; Hughes 2014; Wijburg 2015).
Study flow diagram for the second update.
Included studies
Agalsidase alfa versus placebo
Two trials (n = 41) were included in this comparison and are described in detail below (Hughes 2008; Schiffmann 2001).
Trial design
Both trials were described as randomized, double‐blind, placebo‐controlled, clinical trials (Hughes 2008; Schiffmann 2001). Both trials were for six months (Hughes 2008; Schiffmann 2001), but one had an additional 24 months open‐label follow up (Hughes 2008).
Both trials recruited only males and included 15 and 26 participants respectively (Hughes 2008; Schiffmann 2001). The ratio of participants in the treatment and placebo groups varied; the Schiffmann trial achieved an almost equal split, randomising 14 participants to treatment and 12 to placebo (Schiffmann 2001); and in the Hughes trial seven and eight participants were randomised to the treatment and to the placebo groups respectively (Hughes 2008).
Types of interventions
Both trials compared ERT to placebo and used the same dose regime of Agalsidase alfa (Replagal™) administered intravenously at a dose of 0.2 mg/kg over a period of 20 or 40 minutes for six months. The treatment regimen was the same for the additional 24‐month open‐label follow up (Hughes 2008).
Types of outcomes measured
Moore (in a report of the Schiffmann trial) evaluated transversal doppler measurements, left ventricular volume, heart rate and stroke volume. The following Doppler parameters were obtained: peak flow velocity, mean flow velocity, end‐diastolic velocity, pulsatility index, and resistance index (Schiffmann 2001).
At baseline and six months, left ventricular mass assessed by magnetic resonance imaging (MRI), QRS duration, levels of Gb3 in cardiac tissue, urine sediment and plasma, and adverse effects were evaluated in the Hughes trial (Hughes 2008).
Agalsidase beta versus placebo
Three trials (n = 146) were included in this comparison and are described in detail below (Banikazemi 2007; Bierer 2006; Eng 2001).
Trial design
All three trials were described as randomized, placebo‐controlled, clinical trials (Banikazemi 2007; Bierer 2006; Eng 2001); and two of these were described as double blind (Banikazemi 2007; Eng 2001). Trial duration ranged from five (Eng 2001) to 35 months (Banikazemi 2007).
The number of participants included in each trial ranged from six (Bierer 2006) to 82 (Banikazemi 2007) and the three trials randomized mostly males (Banikazemi 2007; Bierer 2006; Eng 2001). The ratio of participants in the treatment and placebo groups varied across trials; two had a ratio of 2:1 treatment‐to‐placebo (Banikazemi 2007; Bierer 2006); and one had an even split of 29 participants in each group (Eng 2001).
In one trial, the investigators were part of the Genzyme Corporation in Cambridge, Massachussetts (Eng 2001). Only one trial was described as multicenter; with there being 26 referral centers from six countries from North America and Europe (Banikazemi 2007).
Types of interventions
All three trials compared ERT ‐ agalsidase beta (recombinant human alfa‐galactosidase A) ‐ to placebo (Banikazemi 2007; Bierer 2006; Eng 2001). The treatment regimen was the same in these trials with Agalsidase beta administered at a dose of 1mg/kg intravenously every two weeks. Duration of the treatment varied from 20 weeks (Eng 2001), to 35 months (median 18.5 months) (Banikazemi 2007), and 18 months (Bierer 2006).
Types of outcomes measured
Thurberg (in a report of the Eng trial) evaluated urinary creatinine, urinary protein excretion, renal function and dermatologic characteristics of Gb3 accumulation in the dermis (Eng 2001). Measurements were taken at baseline and after infusion 11 at five months.
Banikazemi measured the time to first clinical event (renal, cardiac, cerebrovascular event or death) (Banikazemi 2007). The following measures were taken: serum creatinine level, proteinuria (ratio of urinary albumin to urinary creatinine), 12‐lead electrocardiography, echocardiography, neurologic examination, head MRI, Brief Pain Inventory, exercise tolerance, plasma Gb3 level, Fabry symptom assessment, physical examination, blood chemistry, urinalysis, IgG antibody titers to agalsidase beta and optional skin biopsy. All measurements were taken at baseline. Serum creatinine levels were measured every four weeks aechocardiography, head MRI, and exercise tolerance,were repeated every 24 weeks and all other baseline measurements were repeated every 12 weeks. The estimated glomerular filtration rate (GFR) was determined by using the 4‐variable Modification of Diet in Renal Disease formula.
Bierer measured cardiopulmonary exercise performance, forced expiratory volume and forced vital capacity every three months over an 18‐month period (Bierer 2006).
Agalsidase alfa versus agalsidase beta
Two trials (n = 184) were included in this comparison and are described in detail below (Sirrs 2014; Vedder 2007).
Trial design
The trial conducted by Vedder was a randomized clinical trial with participants being treated for a period of at least 12 months. A total of 36 participants (18 males and 18 females) were included. A total of 34 participants had at least 12 months of follow up and a subgroup of participants (25 out of 34) was followed up for more than 24 months of treatment. Two female participants withdrew after six months of agalsidase alfa or agalsidase beta treatment (Vedder 2007).
The Sirrs trial was a nationwide study of all individuals in Canada with Fabry disease between ages five and 85 years old. This trial looked at the two Canadian Fabry Disease Initiative (CFDI) cohorts of enzyme replacement (ERT) treated individuals. Cohort 1a: participants on ERT when the CFDI began who maintained their baseline treatment assignment (these participants were not randomised and therefore not included in the review). Cohort 1b: participants newly meeting criteria for ERT and randomized 1:1 to agalsidase beta (1.0 mg/kg every two weeks) or agalsidase alfa (0.2 mg/kg every two weeks). A total of 67 participants were randomized; mean follow up was at 23.00 months (Sirrs 2014).
Types of interventions
In one trial, participants were treated with either agalsidase alfa or agalsidase beta at an equal dose of 0.2 mg/kg bi‐weekly (Vedder 2007). In the other trial participants were randomized 1:1 to agalsidase beta 1.0 mg/kg every two weeks or agalsidase alfa 0.2 mg/kg every two weeks (Sirrs 2014).
Types of outcomes measured
In the Vedder trial, reduction in left ventricular mass after 12 and 24 months of treatment was considered the primary outcome. The authors also evaluated the occurrence of treatment failure (defined as progression of cardiac, renal or cerebral disease), GFR, pain, anti‐agalsidase antibodies and GL3 levels in plasma and urine (Vedder 2007).
In the Sirrs trial comparison was made between the effects of agalsidase alfa and agalsidase beta on a composite clinical outcome consisting of renal (dialysis, transplant, or reduction in GFR by 50%), cardiac (admission for cardiac event), neurologic (stroke or sudden unilateral hearing loss) or death (Sirrs 2014).
Multiple agalsidase alfa dosing schedules
Two trials (n = 36) were included in this comparison and are described in detail below (Clarke 2007; Hughes 2013).
Trial design
Both trials were described as randomized, double‐blind, placebo controlled, clinical trials (Clarke 2007; Hughes 2013). One trial was 10 weeks in length (Clarke 2007), while the other was four weeks (Hughes 2013). Each trial recruited 18 participants all of whom were male in one trial (Clarke 2007), while in the other participants of both genders were included (12 males and 6 females). One trial was a cross‐over design where all participants received the same interventions sequentially (Hughes 2013). The other achieved an almost equal split between five arms (Clarke 2007).
Types of interventions
Both trials compared ERT at different dosage regimens. The Hughes trial used three treatment schedules: treatment A 0.2 mg/kg every other week; treatment B, 0.1 mg/kg weekly and treatment C, 0.2 mg/kg weekly (Hughes 2013). The Clarke trial compared five different dosage regimens of agalsidase alfa: 0.1, 0.2, or 0.4 mg/kg weekly; 0.2 mg/kg every other week (the approved dose) or 0.4 mg/kg every other week (Clarke 2007).
Types of outcomes measured
For the Hughes trial, the primary outcome variable was self‐assessed health state measured by the visual analogue scale (‘thermometer’) item of the European quality of life questionnaire (the EQ‐5D: Euroqol Group 1990).
Secondary outcome variables were:
-
pain as assessed as the average composite pain severity dimension of the brief pain inventory (BPI) short form, a standard pain assessment tool, each individual item of the BPI short form, health state as calculated from items 1–5 of the EQ‐5D questionnaire (mobility, self‐care, usual activities, pain/discomfort, and anxiety/depression), each individual item of the EQ‐5D questionnaire;
-
results of quantitative sudomotor axon reflex test (QSART) testing which tests for abnormalities of resting and evoked sweat production;
-
the Mainz severity score index (MSSI), a scoring system to measure the severity of Fabry disease;
-
plasma and urine Gb3; and
-
analgesic use.
All questionnaire‐based data were completed by the participant on a daily diary sheet (health state, pain and analgesic use) and using weekly investigator administered questionnaires (BPI and EQ‐5D). Safety assessments included physical examination, 12‐lead ECG, clinical laboratory testing, and vital sign measurements. Adverse events (AEs) and concomitant medications were recorded throughout the trial. Anti‐agalsidase alfa antibodies were assayed at baseline using an enzyme‐linked immunosorbent assay (ELISA) (Hughes 2013).
The primary efficacy endpoint variable for the Clarke trial was change in plasma Gb3 level from baseline to the day of the final infusion (Clarke 2007).
Excluded studies
A total of 43 studies were excluded from the review (Alamartine 2005; Banikasemi 2005; Beck 2004; Beer 2006; Breunig 2006; Cartwright 2004; Elliott 2006; Eto 2005; Fellgiebel 2014; Fernhoff 2011; Germain 2007; Guffon 2002; Guffon 2004; Hajioff 2006; Hilz 2004; Jardim 2006; Jardim 2006b; Kalliokoshi 2006; Kampmann 2002; Kobayashi 2005; Kosch 2004; Linthorst 2004; Linthorst 2006; Mignani 2004; Mills 2004; Pisani 2005; Ramaswami 2007; Ries 2006; Schiffmann 2003; Schiffmann 2006; Schiffmann 2014; Sirrs 2011; Spinelli 2004; Tsuboi 2014; Utsumi 2005; Weidemann 2003; Weidemann 2014; West 2011). The main reasons for exclusion being that these were case series, cohort studies, retrospective studies or non‐randomized trials.
Risk of bias in included studies
See: Figure 2; Figure 3 and Characteristics of included studies.
Risk of bias graph: review authors' judgements about each risk of bias item presented as percentages across all included studies.
Risk of bias summary: review authors' judgements about each risk of bias item for each included study.
Agalsidase alfa versus placebo
Generation of randomization sequence
The Hughes trial did not describe the generation of allocation, thus, this trial was classified as having an unclear risk of bias (Hughes 2008). The Schiffmann trial reported the generation of allocation was made by random tables and, therefore we rated this trial as low risk of bias for this domain (Schiffmann 2001).
Allocation concealment
The Hughes trial did not describe the concealment of allocation so this was judged to have an unclear risk of bias (Hughes 2008). The Schiffmann trial reported that randomization was provided by an unblinded pharmacist and, therefore we rated this trial as low risk of bias for this domain (Schiffmann 2001).
Blinding
The Hughes trial reported data remained blinded until the database was unlocked and the statistical analyses performed (Hughes 2008). We therefore judge this to be at a low risk of bias regarding this domain. However, the Schiffmann trial did not describe any blinding process and, therefore it was classified as having an unclear risk of bias (Schiffmann 2001).
Incomplete outcome data
The Hughes trial did not describe withdrawals, dropouts or ITT, therefore we judged this to have an unclear risk of bias (Hughes 2008). The Schiffmann trial describes only four withdrawals and, therefore we rated this trial as low risk of bias for this domain (Schiffmann 2001).
Selective reporting
We judged both included trials to be free of selective reporting and having a low risk of bias (Hughes 2008; Schiffmann 2001).
Other potential sources of bias
We did not identify any other potential sources of bias (Hughes 2008; Schiffmann 2001).
Agalsidase beta versus placebo
Generation of randomization sequence
The Banikazemi trial described the generation of randomization as computer‐generated and, therefore we rated this trial as low risk of bias (Banikazemi 2007).The Bierer and Eng trials did not describe details on how the sequences were generated; therefore these trials presented an unclear risk of bias (Bierer 2006; Eng 2001).
Allocation concealment
The Banikazemi trial was classified as having a low risk of bias since the randomization codes were maintained centrally at a secure location (Banikazemi 2007). The Bierer and Eng trials did not describe details regarding the concealment of allocation; thus, these trials were classified as having an unclear risk of bias (Bierer 2006; Eng 2001)
Blinding
The Banikazemi trial described that the sponsor staff, investigators, and participants were blinded to treatment allocation leading to a low risk of bias (Banikazemi 2007). The Bierer and Eng trials were described as double blind, but provided no further details, we therefore judge these two trials to have an unclear risk of bias (Bierer 2006; Eng 2001).
Incomplete outcome data
The Banikazemi trial reported less than 20% of withdrawals and dropouts, we therefore assessed this as having a low risk of bias (Banikazemi 2007). The Bierer and Eng trials did not describe withdrawals, dropouts or ITT, therefore we judged this to have an unclear risk of bias (Bierer 2006; Eng 2001).
Selective reporting
We judged all three included trials to be free of selective reporting (Banikazemi 2007; Bierer 2006; Eng 2001).
Other potential sources of bias
We judged all three included trials to be at high risk of bias as the investigators were part of the Genzyme Corporation (Banikazemi 2007; Bierer 2006; Eng 2001).
Agalsidase alfa versus agalsidase beta
Generation of randomization sequence
Both trials described the generation of randomization as manual randomization and, therefore we rated this domain as low risk of bias (Sirrs 2014; Vedder 2007).
Allocation concealment
The Vedder trial was classified as having a low risk of bias since the investigators used envelopes that were checked by two people not involved in obtaining informed consent, thus we classified the trial as having a low risk of bias (Vedder 2007). The Sirrs trial reported that a third party was involved in the generation of allocation; thus, this trial was classified as having a low risk of bias (Sirrs 2014).
Blinding
The Vedder trial was classified as having a high risk of bias since there was no blinding (Vedder 2007). The Sirrs trial was not blinded but the composite outcome consisting of renal, cardiac, neurologic and, death is unlikely to be affected by lack of blinding, hence it was judged to have low risk of bias (Sirrs 2014).
Incomplete outcome data
The Vedder trial reported more than 20% of withdrawals and dropouts, we therefore assessed this as having a high risk of bias (Vedder 2007). The Sirrs trial did not describe withdrawals, dropouts or ITT, therefore we judged this to have an unclear risk of bias (Sirrs 2014).
Selective reporting
he Vedder trial reported quality of life as an outcome in the protocol, but this was not stated in the paper, we therefore assessed this as having a high risk of bias (Vedder 2007). The Sirrs trial was free of selective reporting and, therefore we classified this domain as having a low risk of bias (Sirrs 2014).
Other potential sources of bias
We did not identify any other potential sources of bias (Vedder 2007; Sirrs 2014).
Agalsidase alfa multiple dose schedules comparison
Generation of randomisation sequence
The Clarke and Hughes trials did not provide a description regarding the generation of allocation; thus, they were classified as having an unclear risk of bias (Clarke 2007; Hughes 2013).
Allocation concealment
The Clarke and Hughes trials did not provide a description regarding the concealment of allocation; thus, they were judged to have an unclear risk of bias (Clarke 2007; Hughes 2013).
Blinding
The Hughes trial was classified as having a low risk of bias since there was blinding (Hughes 2013). The Clarke trial was also classified as having a low risk of bias, although this was an open‐label trial, the outcome was objective (Clarke 2007).
Incomplete outcome data
The Clarke and Hughes trials reported less than 20% of withdrawals and dropouts, we therefore assessed this as having a low risk of bias (Clarke 2007; Hughes 2013).
Selective reporting
We judged both included trials to be free of selective reporting (Clarke 2007; Hughes 2013).
Other potential sources of bias
We did not identify any other potential sources of bias (Clarke 2007; Hughes 2013).
Effects of interventions
See: Summary of findings for the main comparison ; Summary of findings 2 ; Summary of findings 3
We presented the data from agalsidase alfa and agalsidase beta separately because of differences in trial design and treatment doses.
Agalsidase alfa versus placebo
Two trials were included in this comparison with a total of 41 participants. 15 and 26 participants were enrolled in the trial conducted by Hughes and Schiffmann respectively (Hughes 2008; Schiffmann 2001).
Primary outcomes
1. Changes in globotriaosylceramide (Gb3) concentration in plasma and tissue
Two trials reported on this outcome at six months (end of treatment) (Hughes 2008; Schiffmann 2001). There was no statistically significant difference between treatment groups in the plasma Gb3 concentration, pooled MD ‐2.07 (95% CI ‐6.64 to 2.50) (with data entered from 39 of 41 participants) (Hughes 2008; Schiffmann 2001) (Analysis 1.1). There were no statistically significant differences between treatment groups regarding the subcategories: urine sediment Gb3, MD ‐812.00 (95% CI ‐1897.83 to 273.83); and kidney, MD ‐2.50 (95% CI ‐9.47 to ‐4.47) (Schiffmann 2001) (Analysis 1.2; Analysis 1.3). Hughes reported a statistically significant difference between the participants receiving agalsidase alfa and those receiving placebo for myocardial Gb3 levels at over three months and up to six months, MD 0.07 (95% CI ‐0.35 to 0.49) (Hughes 2008) (Analysis 1.4).
2. Death
Neither of the included trials reported on this outcome (Hughes 2008; Schiffmann 2001).
3. Pain (measured by the McGill Pain Questionnaire)
a. Acroparaesthesia
We intended to measure this outcome by the intensity and duration of acroparaesthesia as reported by the individual, however none of the included trials reported on this outcome (Hughes 2008; Schiffmann 2001).
b. Fabry crises
We planned to present the presence or absence of crises as a binary outcome, but none of the included trials reported on this outcome (Hughes 2008; Schiffmann 2001).
c. The Brief Pain Inventory severity
One trial reported on this outcome (Schiffmann 2001). There was a statistically significant difference favouring the participants receiving agalsidase alfa versus placebo in all the subcategories studied at over one month and up to three months, MD ‐2.10 (95% CI ‐3.79 to ‐0.41), at over three months and up to five months, MD ‐1.90 (95% CI ‐3.65 to ‐0.15), and at over five months and up to six months, MD ‐2.00 (95% CI ‐3.66 to ‐0.34) (Schiffmann 2001) (Analysis 1.5).
d. The Brief Pain Inventory pain‐related quality of life
One trial reported on this outcome (Schiffmann 2001). There was no statistically significant difference between groups in the subcategories studied at over one month and up to three months, MD ‐0.90 (95% CI ‐2.73 to 0.93) and at over three months and up to five months, MD ‐1.80 (95% CI ‐3.77 to 0.17). However, at over five months and up to six months a significant difference favouring agalsidase alfa was noted, MD ‐2.10 (95% CI ‐3.92 to ‐0.28) (Schiffmann 2001) (Analysis 1.6).
Secondary outcomes
1. Effect of therapy on renal function
One trial reported on this outcome (Schiffmann 2001). Creatinine clearance and inulin clearance were used to estimate glomerular filtration rate. There were no statistically significant difference between both groups in the subcategories creatinine clearance at end of treatment (up to six months), and inulin clearance (up to six months) respectively, MD 10.30 (95% CI ‐15.37 to 35.97) and MD ‐0.50 (95% CI ‐21.36 to 20.36) (data entered for 24 of 26 participants) (Schiffmann 2001) (Analysis 1.7; Analysis 1.8).
There was no statistically significant difference between groups in the subcategories: glomeruli with mesangial widening at up to six months, MD ‐14.70 (95% CI ‐36.72 to 7.32); glomeruli with segmental sclerosis at up to six months and obsolescent glomeruli at up to six months; MD 3.80 (95% CI ‐2.35 to 9.95) and MD 6.50 (95% CI ‐8.93 to 21.93) (data entered for 21 out of 26 participants) (Schiffmann 2001) (Analysis 1.9; Analysis 1.10; Analysis 1.11).
2. Symptoms and complications of disease (such as renal failure, skin, cerebrovascular and cardiac complications)
Neither of the included trials reported on this outcome (Hughes 2008; Schiffmann 2001).
3. Parameters: echocardiographic
There was no statistically significant difference between groups for mean left ventricular wall thickness, MD ‐0.79 (95% CI ‐3.62 to 2.04); left ventricular internal diameter (diastolic), MD ‐3.70 (95% CI ‐11.73 to 4.33); left ventricular internal diameter (systolic), MD ‐2.70 (95%CI ‐9.91 to 4.51); and left ventricular ejection fraction, MD 1.88 (95%CI ‐4.68 to 8.44) (data entered for 14 out of 15 participants) (Hughes 2008) (Analysis 1.12).
4. Histologic analysis of microvascular capillary endothelial deposits of Gb3 in biopsy specimens
Neither of the included trials reported on this outcome (Hughes 2008; Schiffmann 2001).
5. Adverse effects of treatment
Neither of the included trials reported on this outcome (Hughes 2008; Schiffmann 2001).
6. Quality of life (as determined by the Short Form 36 (SF‐36) and the impact of clinical variables on domain scores within the SF‐36 (Smith 2000))
Neither of the included trials reported on this outcome (Hughes 2008; Schiffmann 2001).
Agalsidase beta versus placebo
Three trials were included in this comparison with a total of 146 participants. Banikazemi enrolled 82 participants, Bierer enrolled six participants and Eng enrolled 58 participants (Banikazemi 2007; Bierer 2006; Eng 2001).
Primary outcomes
1. Changes in globotriaosylceramide (Gb3) concentration in plasma and tissue
One of the included trials reported on this outcome in tissue (Eng 2001). There was a statistically significant difference favouring the participants receiving agalsidase beta versus placebo in three subcategories: kidney, MD ‐1.70 (95% CI ‐2.09 to ‐1.31); heart, MD ‐0.90 (95% CI ‐1.18 to ‐0.62); and composite, MD ‐4.80 (95% CI ‐5.45 to ‐4.15) (Eng 2001) (Analysis 2.1).
None of the included trials reported the effect on plasma concentrations (Banikazemi 2007; Bierer 2006; Eng 2001).
2. Death
There was no statistically significant difference in death as reported in one trial, RR 1.85 (95% CI 0.08 to 43.96) (Analysis 2.2). There was one death from a total of 51 participants in agalsidase beta and no deaths in the placebo group (n = 31) (Banikazemi 2007).
3. Pain (measured by the McGill Pain Questionnaire)
a. Acroparaesthesia
We intended to measure this outcome by the intensity and duration of acroparaesthesia as reported by the individual however none of the included trials reported this outcome (Banikazemi 2007; Bierer 2006; Eng 2001).
b. Fabry crises
We planned to present whether the participants had reported the presence or absence of crises as a binary outcome, but none of the included trials reported this outcome (Banikazemi 2007; Bierer 2006; Eng 2001).
Only one trial reported chest pain and fatigue; there was no statistically significant difference for any of these at 24 months; chest pain,RR 6.77 (95% CI 0.39 to 118.36) and fatigue, RR 6.77 (95% CI 0.39 to 118.36) (Banikazemi 2007) (Analysis 2.10; Analysis 2.12). Eng reported on pain related to Fabry disease at over three months and up to six months; there was no statistically significant difference with pain related to Fabry disease, RR 3.00 (95% CI 0.33 to 27.18) (Eng 2001) (Analysis 2.11).
Secondary outcomes
1. Effect of therapy on renal function
The number of renal events, defined by the authors as 33% increase in serum creatinine level; end‐stage renal disease, were reported by Banikazemi. There was no statistically significant difference in the number of events between groups, RR 0.87 (95% CI 0.37 to 2.04) (Analysis 2.3) (Banikazemi 2007).
2. Symptoms and complications of disease (such as renal failure, skin, cerebrovascular and cardiac complications)
a. Cardiac events
Bierer reported cardiopulmonary exercise testing (Bierer 2006). There were no significant differences between groups in any of the sub‐categories evaluated (average heart rate reserve, MD 21.30 (95% CI ‐1.28 to 43.88); average maximum oxygen uptake measured at peak exercise (in mL/(kg·min)), MD 0.22 (95% CI ‐0.94 to 1.38); maximum oxygen uptake measured at peak exercise, MD 2.60 (95% CI ‐13.16 to 18.36); and oxygen pulse average at peak exercise, MD 2.10 (95% CI ‐3.67 to 7.87)) (Analysis 2.4).
There were no significant differences in decrease in cardiopulmonary exercise test, RR 1.50 (95% CI 0.34 to 6.70) (Analysis 2.5) (Bierer 2006).
Banikazemi reported cardiac events and there was no significant differences between studied groups, RR 0.46 (95% CI 0.11 to 1.90) (Analysis 2.6) (Banikazemi 2007).
b. Cerebrovascular events
There was no significant difference in the number of cerebrovascular events between groups in the Banikazemi trial, RR 0.12 (95% CI 0.01 to 2.48) (Analysis 2.7). (Banikazemi 2007).
c. Percentage of participants achieving zero scores in microvascular endothelial deposits of globotriaosylceramide in the skin
For participants achieving zero scores in skin at five months, there was a statistically significant difference between groups in the Eng trial favouring the participants receiving agalsidase beta versus placebo in two subcategories: superficial endothelial cells, RR 19.67 (95% CI 4.13 to 93.63) and deep endothelial cells, RR 11.00 (95% CI 2.88 to 42.08) (Eng 2001). There was no significant difference between the analysed participants in each group in two further sub‐categories: smooth muscle cells, RR 1.50 (95% CI 0.10 to 22.62), three and one participants contributed to this analysis in the agalsidase beta and placebo groups, respectively; or in perineurium cells, RR 1.04 (95% CI 0.07 to 15.72) (note not all the trial participants contribute to the analyses, please refer to the forest plot) (Analysis 2.8) (Eng 2001).
Eng reported data for participants achieving a zero score or a reduction in microvascular endothelial deposits of globotriaosylceramide in the skin (Analysis 2.9). There was a statistically significant difference favouring the participants receiving agalsidase beta versus placebo in two sub‐categories: superficial endothelial cells, RR 2.81 (95% CI 1.72 to 4.59); and deep endothelial cells, RR 2.79 (95% CI 1.67 to 4.67) (Eng 2001). There was no significant difference between groups in two further sub‐categories: smooth muscle cells, RR 1.50 (95% CI 0.10 to 22.62) and in perineurium cells, RR 1.49 (95% CI 0.68 to 3.25). Three and one participants contributed to this analysis in the agalsidase beta and placebo groups, respectively.
3. Parameters echocardiographic
None of the included trials reported this outcome (Banikazemi 2007; Bierer 2006; Eng 2001).
4. Histologic analysis of microvascular capillary endothelial deposits of Gb3 in biopsy specimens
None of the included trials reported this outcome (Banikazemi 2007; Bierer 2006; Eng 2001).
5. Adverse effects of treatment
a. Any adverse event
Two trials reported adverse events; one at over three month and up to six months (Eng 2001) and one at 24 months (Banikazemi 2007).
There were statistically significant differences in favour of placebo regarding rigours at the following: over three and up to six months, RR 29.00 (95% CI 1.81 to 464.38) and, at 24 months, RR 10.94 (95% CI 1.54 to 77.95) (Analysis 2.13). Over three and up to six months there was no statistically significant difference comparing agalsidase beta to placebo, RR 7.00 (95% CI 0.92 to 53.36); however, at 24 months there was a statistically significant difference in favour of agalsidase beta compared to placebo in the reduction of fever, RR 8.51 (95% CI 1.18 to 61.58) (Analysis 2.14). One trial reported chills and there was no statistically significant difference at over three months and up to six months, RR 9.00 (95% CI 0.51 to 159.94) (Eng 2001) (Analysis 2.15). With regards to hypertension, there were no statistically significant differences between groups at both studied time points: over three and up to six months, RR 7.00 (95% CI 0.38 to 129.74); and at 24 months, RR 2.13 (95% CI 0.47 to 9.60) (Analysis 2.16).
Only one trial reported vomiting; there was no statistically significant difference at 24 months with vomiting, RR 8.00 (95% CI 0.47 to 137.27) (Banikazemi 2007) (Analysis 2.17). Eng reported on headache at over three months and up to six months; there was no statistically significant difference with headache, RR 2.50 (95% CI 0.53 to 11.86) (Eng 2001) (Analysis 2.18).
b. Any serious adverse events
None of the included trials reported this outcome (Banikazemi 2007; Bierer 2006; Eng 2001).
6. Quality of life (as determined by the Short Form 36 (SF‐36) and the impact of clinical variables on domain scores within the SF‐36 (Smith 2000))
None of the included trials reported this outcome (Banikazemi 2007; Bierer 2006; Eng 2001).
Agalsidase alfa versus agalsidase beta
Two trials were included in this comparison with a total of 103 participants (Sirrs 2014; Vedder 2007).
Primary outcomes
1. Changes in globotriaosylceramide (Gb3) concentration in plasma and tissue
The Vedder trial reported a median of 3.3 umol/l (range 1.45 to 6.42) and 3 umol/l (range 1.89 to 7.85) for 12 and 24 months of treatment, respectively in agalsidase alfa arm, while there was a median of 2.48 umol/l (range 1.40 to 3.93) and 2.23 umol/l (range 0.92 to 4.3) for 12 and 24 months of treatment, respectively in agalsidase beta arm (Vedder 2007).
Sirrs did not report on this outcome (Sirrs 2014).
2. Death
There was a death from multiple cerebral infarctions after 20 months in the agalsidase alfa treatment group, RR 3.00 (95% CI 0.13 to 69.09) (Vedder 2007) (Analysis 3.1). There was one death from a total of 18 participants in agalsidase alfa and no deaths in agalsidase beta (n = 18) (Vedder 2007).
3. Pain (measured by BPI‐3)
a. Acroparaesthesia and Fabry crises
One trial described no significant reduction of pain score (BPI‐3) after 12 months of treatment in either of the studied groups: alfa 0 (range 25 to 1); and beta 21.5 (range 24 to 3). The values did not change for analysis at the 24 month time point (Vedder 2007). The Sirrs trial did not evaluate pain (Sirrs 2014).
b. Fabry crises
We planned to present whether the participants had reported the presence or absence of crises as a binary outcome, but neither of the included trials reported this outcome (Sirrs 2014; Vedder 2007).
Secondary outcomes
1. Effect of therapy on renal function
The trial reported that renal insufficiency, defined as an increase of serum creatinine greater than 33%, progressed in two participants treated with agalsidase alfa, who had baseline GFR measurements of 22 and 30 ml/min (within 24 months). In the agalsidase alfa group, progression of renal insufficiency was seen in one other participant after 30 months of treatment (Vedder 2007).
2. Symptoms and complications of disease (such as renal failure, skin, cerebrovascular and cardiac complications)
a. Cardiac events
Two participants suffered from atrial fibrillation after 42 and 36 months of agalsidase beta treatment (Vedder 2007). Atrial fibrillation occurred in a male participant treated with agalsidase alfa for 30 months. There was no statistically significant difference for cardiac events, RR 0.47 (95% CI 0.05 to 4.60) (26 out of 29 participants analysed) (Vedder 2007) (Analysis 3.2).
b. Cerebrovascular events
One of the included trials did not report this outcome (Vedder 2007).
c. Percentage of participants achieving zero scores in skin
Neither of the included trials reported on this outcome (Sirrs 2014; Vedder 2007).
3. Echocradiographic Parameters
The Vedder trial reported LV Mass (g). There was a median of 244 g (range 157 to 424) and 294 g (range 196 to 502) for 12 and 24 months of treatment, respectively in the agalsidase alfa arm, while there was a median of 296 g (range 169 to 401) and 308 g (range 196 to 471) for 12 and 24 months of treatment, respectively in the agalsidase beta arm (Vedder 2007).
The Sirrs trial did not report on this outcome (Sirrs 2014).
4. Histologic analysis of microvascular capillary endothelial deposits of Gb3 in biopsy specimens
Neither of the included trials reported on this outcome (Vedder 2007; Sirrs 2014).
5. Adverse effects of treatment
a. Any adverse event
There was no significant difference in the number of any adverse events between groups in the Vedder trial, RR 0.36 (95% CI 0.08 to 1.59) (Vedder 2007) (Analysis 3.3).
The Sirrs trial did not report on this outcome (Sirrs 2014).
b. Any serious adverse events
There was no significant difference in the number of any serious adverse events between groups in the Vedder trial, RR 0.30 (95% CI 0.03 to 2.57) (Vedder 2007) (Analysis 3.4).
6. Quality of life (as determined by the Short Form 36 (SF‐36) and the impact of clinical variables on domain scores within the SF‐36 (Smith 2000))
One of the included trials did not report on this outcome, but it is included as an outcome in the protocol (Vedder 2007). We contacted the main author for further clarification and hope these data can be included in a future update of this review.
Sirrs did not report on this outcome (Sirrs 2014).
Multiple dose schedules of agalsidase alfa
Two trials were included in this comparison with a total of 36 participants (Clarke 2007; Hughes 2013).
Primary outcomes
1. Changes in globotriaosylceramide (Gb3) concentration in plasma and tissue
No significant difference was found among the treatment schedules for fasting plasma GB3 levels in the Hughes trial (Hughes 2013).
No statistically significant association between the magnitude of the reduction in plasma Gb3 and dose or dose frequency was described in the Clarke trial (Clarke 2007).
2. Death
Neither of the included trials reported on this outcome (Clarke 2007; Hughes 2013).
3. Pain (measured by BPI‐3)
For the cross‐over trial by Hughes (18 participants), there were no significant differences comparing 0.2 mg/ kg/ every two weeks versus 0.1 mg/ kg/ every week for mean pain score (Hughes 2013), MD ‐0.06 (95% CI ‐1.26 to 1.38), and for self assessed health state MD ‐1.20 (95% CI ‐14.10 to 11.70) (Hughes 2013) (Analysis 4.1; Analysis 4.2). There were no significant differences comparing 0.1 mg/kg/every week versus 0.2 mg/kg/every week for mean pain score MD 0.06 (95% CI ‐1.25 to 1.37) and for self assessed health state, MD 0.30 (95% CI ‐12.70 to 13.30) (Hughes 2013) (Analysis 5.1; Analysis 5.2).
The Clarke trial did not report on this outcome (Clarke 2007).
Secondary outcomes
1. Effect of therapy on renal function
Neither of the included trials reported on this outcome (Clarke 2007; Hughes 2013).
2. Symptoms and complications of disease (such as renal failure, skin, cerebrovascular and cardiac complications)
Neither of the included trials reported on this outcome (Clarke 2007; Hughes 2013).
3. Echocardiographic Parameters
Neither of the included trials reported on this outcome (Clarke 2007; Hughes 2013).
4. Histologic analysis of microvascular capillary endothelial deposits of Gb3 in biopsy specimens
Neither of the included trials reported on this outcome (Clarke 2007; Hughes 2013).
5. Adverse effects of treatment
In the trial conducted by Clarke, the incidence of adverse events was not related to dose or dosing frequency (Clarke 2007).
The majority of moderate and severe AEs were reported by participants after treatment schedule A (0.2 mg/kg every other week) (Hughes 2013).
6. Quality of life (as determined by the Short Form 36 (SF‐36) and the impact of clinical variables on domain scores within the SF‐36 (Smith 2000))
Neither of the included trials reported on this outcome (Clarke 2007; Hughes 2013).
Discussion
Summary of main results
This review included nine trials that were split between low and uncertain risk of bias and in addition statistical analysis was difficult due to the manner in which results were presented. Overall though, compared with participants receiving placebo, those treated with either agalsidase alfa or agalsidase beta experienced a reduction in Gb3 concentration in plasma and tissues. Although this finding is indicative of ‘proof of concept’, its clinical significance is uncertain as the trials examined did not provide specific information on correlation with clinical events or survival. In the agalsidase alfa comparison, even though we used a random‐effects model (to better deal with heterogeneity), we found a high level of heterogeneity (I2 = 91%) in the outcome plasma Gb3 at up to six months. This method assumes that the effects being estimated in the different trials are not identical, but follow some distribution. In this case the effect may have varied due to different population or intervention characteristics (such the dose, duration and regimen). Trial results indicate a significant effect of enzyme replacement therapy (ERT) on neuropathic pain, cardiac morphology and renal function, and a positive influence on pain‐related quality of life. Treatment with ERT appears to be relatively well‐tolerated. Infusion‐related adverse events may be controlled by the use of pre‐medication and the administration of the enzymes at a slower rate. Antibody formation has been reported in a significant proportion of treated individuals with trials indicating that the presence of neutralizing antibodies may potentially influence therapeutic outcome.
The effectiveness of therapy in delaying the onset or reducing the incidence and severity of Anderson‐Fabry disease (AFD)‐related complications, and its impact on long‐term survival remains unclear though clearly it does have a positive effect on some aspects of the disease. Heterogeneity in clinical phenotype means long‐term, large cohort studies are required. This is being addressed primarily through patient registries. It will also be important to look in detail at effectiveness and safety of ERT according to patient sub‐populations, such as gender, later‐onset versus classic phenotypes, type of mutation, age and presence of chronic injury at initiation of treatment.
The low number of eligible trials for this Cochrane analysis together with the low number of participants included in the trials, due to the rare nature of AFD, prevented us performing these sub‐analyses. It is likely that advanced disease prior to the start of treatment influences outcome and that the earlier the initiation of treatment the greater the effect on preventing complications. Trials in asymptomatic or minimally affected individuals may provide information on the effectiveness of therapy in preventing or delaying the onset of symptoms or disease‐related complications (Ramaswami 2007; Ries 2006; Ries 2007; Wraith 2008).
Overall completeness and applicability of evidence
A limitation is the fact that pain often does not feature in the trials. It is known that most cases of pain resolve spontaneously, even in the absence of any treatment, and pain is a subjective outcome and thus depends on the self‐reports provided by the participants, so it is difficult to provide objective evidence. Only Schiffmann reported the Brief Pain Inventory comparing agalsidase alfa versus placebo (Schiffmann 2001). No included trial (Banikazemi 2007; Bierer 2006; Eng 2001 ) comparing agalsidase beta versus placebo reported on pain and, only Hughes 2013 reported on it comparing 0.2 mg/kg/every 2 weeks versus 0.1 mg/kg/every week and, 0.1 mg/kg/every week versus 0.2 mg/kg/every week.
Quality of the evidence
The methodological quality of the included trials was generally unclear for the random sequence generation and allocation concealment (Bierer 2006; Clarke 2007; Eng 2001; Hughes 2008; Hughes 2013; Sirrs 2014). Some methodological aspects of one trial (with inadequate blinding of outcome assessment, participants and caregivers, as well as incomplete outcome data and selective reporting outcome) resulted in this trial having a high risk of bias (Vedder 2007).
Potential biases in the review process
Despite our thorough search in various databases we might have overlooked trials, especially with regard to the grey literature. However, we contacted the authors of the included trials to ask whether they had undertaken any other relevant trials comparing ERT to other interventions, placebo or no interventions, for treating AFD.
Agreements and disagreements with other studies or reviews
The results of this review reflect those of previous versions of the review (El Dib 2010; El Dib 2013). A recent systematic review undertaken to analyse the currently available data concerning quality of life (QoL) measurement, concluded no definite conclusions could be drawn from the trials on the effect of ERT on QoL (Arends 2015). A review of the literature comparing two different products the Replagal™) (agalsidase alfa) and Fabrazyme® (agalsidase beta) formed the same conclusion as this review in that there is little evidence for choosing one over the other (Riccio 2013).
There are some clinical single entity studies demonstrating beneficial effect of ERT. The effect of agalsidase beta 1 mg/kg bi‐weekly was assessed in a phase 4 placebo‐controlled study of both untreated and treated individuals with Fabry disease with mild‐to‐moderate renal involvement (i.e., serum creatinine measurements of ≥1.2 mg/dl and < 3.0 mg/dl) on progressive white matter lesions and stroke. The authors concluded that "ERT may reduce the progression of vascular disease, even in advanced FD patients, suggesting that early treatment may stabilize WML progression and stroke risk" (Fellgiebel 2014). A cohort study evaluated the progression of left ventricular hypertrophy in untreated men with Fabry disease and assessed the effects of agalsidase beta on left ventricular hypertrophy (Germain 2013). The authors analysed 115 men treated with agalsidase beta (1 mg/kg/2 weeks) and 48 untreated men. The authors found for men in whom treatment was initiated at the age of 18 years to under 30 years, mean left ventricular mass slope was ‐3.6 g/year (n = 31) compared with + 9.5g/year in untreated men of that age (n = 15). The authors also found that untreated men had a higher risk of having faster increases in left ventricular mass compared with treated men (Germain 2013).
Schiffmann conducted an open‐label study with 11 children who completed 26‐weeks of intravenous 0.2 mg/kg agalsidase alfa (Schiffmann 2014). The authors found all participants experienced at least one adverse event, but no death occurred and cardiac outcomes remained stable. Tøndel evaluated five years of treatment with agalsidase alfa or agalsidase beta in 12 people aged seven to 33 years (Tøndel 2013). Biopsy findings from all participants showed total clearance of glomerular endothelial and mesangial cell inclusions after a median of follow up of 65 months. Another study assessed 105 adults with Fabry disease who had received agalsidase beta (1.0 mg/kg body weight) for more than a year and, the authors found that "patients receiving regular agalsidase beta dose had a stable disease course, but dose reduction led to worsening of renal function and symptoms" (Weidemann 2014).
A randomized controlled trial that evaluated pharmacokinetic parameters of a recombinant agalsidase beta (Agal) ERT in 18 healthy adult volunteers found no immunogenicity or any significant infusion‐related reactions; although overall there were seven adverse events that were resolved without any complications (Kim 2014).
Rombach study evaluated 59 people with Fabry disease treated with either agalsidase alfa or beta for at least one year (Rombach 2012). They found 17 out of 59 individuals developed anti‐agalsidase antibodies during the first year of treatment and, the authors concluded that the presence of these antibodies is associated with a less robust decrease in plasma lysoGb3 as well as a negative urinary Gb3 reduction.
Study flow diagram for the second update.
Risk of bias graph: review authors' judgements about each risk of bias item presented as percentages across all included studies.
Risk of bias summary: review authors' judgements about each risk of bias item for each included study.
Comparison 1 Agalsidase alfa vs placebo, Outcome 1 Plasma Gb3.
Comparison 1 Agalsidase alfa vs placebo, Outcome 2 Urine sediment Gb3.
Comparison 1 Agalsidase alfa vs placebo, Outcome 3 Kidney Gb3.
Comparison 1 Agalsidase alfa vs placebo, Outcome 4 Myocardial globotriaosylceramide levels.
Comparison 1 Agalsidase alfa vs placebo, Outcome 5 The Brief Pain Inventory Severity.
Comparison 1 Agalsidase alfa vs placebo, Outcome 6 The Brief Pain Inventory pain‐related QoL.
Comparison 1 Agalsidase alfa vs placebo, Outcome 7 Creatinine clearance.
Comparison 1 Agalsidase alfa vs placebo, Outcome 8 Inulin clearance.
Comparison 1 Agalsidase alfa vs placebo, Outcome 9 Glomeruli with mesangial widening.
Comparison 1 Agalsidase alfa vs placebo, Outcome 10 Glomeruli with segmental sclerosis.
Comparison 1 Agalsidase alfa vs placebo, Outcome 11 Obsolescent glomeruli.
Comparison 1 Agalsidase alfa vs placebo, Outcome 12 Echocardiographic data on left ventricular structure and function over 3 months and up to 6 months.
Comparison 2 Agalsidase beta vs placebo, Outcome 1 Microvascular endothelial deposits of globotriaosylceramide.
Comparison 2 Agalsidase beta vs placebo, Outcome 2 Death (ITT population).
Comparison 2 Agalsidase beta vs placebo, Outcome 3 Renal events (ITT population).
Comparison 2 Agalsidase beta vs placebo, Outcome 4 Cardiopulmonary exercise test.
Comparison 2 Agalsidase beta vs placebo, Outcome 5 Cardiopulmonary exercise test.
Comparison 2 Agalsidase beta vs placebo, Outcome 6 Cardiac events.
Comparison 2 Agalsidase beta vs placebo, Outcome 7 Cerebrovascular events (ITT population).
Comparison 2 Agalsidase beta vs placebo, Outcome 8 Participants achieving zero scores in skin over 3 months and up to 6 months.
Comparison 2 Agalsidase beta vs placebo, Outcome 9 Participants achieving zero score or reduction in skin over 3 months and up to 6 months.
Comparison 2 Agalsidase beta vs placebo, Outcome 10 Chest pain.
Comparison 2 Agalsidase beta vs placebo, Outcome 11 Pain related to AFD.
Comparison 2 Agalsidase beta vs placebo, Outcome 12 Fatigue.
Comparison 2 Agalsidase beta vs placebo, Outcome 13 Adverse event: rigors.
Comparison 2 Agalsidase beta vs placebo, Outcome 14 Adverse event: fever.
Comparison 2 Agalsidase beta vs placebo, Outcome 15 Adverse event: chills.
Comparison 2 Agalsidase beta vs placebo, Outcome 16 Adverse event: hypertension.
Comparison 2 Agalsidase beta vs placebo, Outcome 17 Adverse event: vomiting.
Comparison 2 Agalsidase beta vs placebo, Outcome 18 Adverse event: headache.
Comparison 3 Agalsidase alfa vs agalsidase beta, Outcome 1 Death.
Comparison 3 Agalsidase alfa vs agalsidase beta, Outcome 2 Cardiac events.
Comparison 3 Agalsidase alfa vs agalsidase beta, Outcome 3 Any adverse event.
Comparison 3 Agalsidase alfa vs agalsidase beta, Outcome 4 Any serious adverse events.
Comparison 4 0.2 mg/kg/every 2 weeks vs 0.1 mg/kg/every week, Outcome 1 Pain.
Comparison 4 0.2 mg/kg/every 2 weeks vs 0.1 mg/kg/every week, Outcome 2 Self‐assessed health state.
Comparison 5 0.1 mg/kg/every week vs 0.2 mg/kg/every week, Outcome 1 Pain.
Comparison 5 0.1 mg/kg/every week vs 0.2 mg/kg/every week, Outcome 2 Self‐assessed health state.
| Enzyme replacement therapy for Anderson‐Fabry disease | ||||||
| Patient or population: Anderson‐Fabry disease Settings: outpatients Intervention: agalsidase alfa Comparison: placebo | ||||||
| Outcomes | Illustrative comparative risks (95% CI) | Relative effect | No of Participants | Quality of the evidence | Comments | |
| Assumed risk | Corresponding risk | |||||
| Placebo | Agalsidase alfa | |||||
| Changes in globotriaosylceramide (Gb3) concentration in plasma Follow up: 6 months (umol/l) | The mean in changes in Gb3 concentration in plasma score ranged between 0.63 and 10.19 umol/l | The mean in changes in Gb3 concentration in plasma score in the intervention group was on average 2.07 fewer | NA | 39 2 studies | ⊕⊝⊝⊝ | We found no statistically significant differences between studied groups in changes from baseline for Gb3for which data were available. |
| Death | Outcome not reported | Outcome not reported. | Not estimable | NA | NA | |
| Pain The Brief Pain Inventory severity Follow up: over 5 months and up to 6 months | The mean for change in pain by the BPI severity was 4.7 (SD 2.25) | The mean for changes in pain by the BPI severity in the intervention group was on average 2.00 fewer (3.66 fewer to 0.34 fewer) | NA | 26 1 study | ⊕⊕⊝⊝ | We found a statistically significant differences favouring agalsidase alfa compared to placebo in pain for which data were available. |
| Renal function by creatinine clearance Follow up: at up to 6 months (mg/dL) | The mean in changes in renal function was 84.5 (SD: 35.15) mg/dL | The mean for changes in renal function in the intervention group was on average 10.30 more (15.37 fewer to 35.97 more) mg/dL | NA | 24 1 study | ⊕⊝⊝⊝ | We found no statistically significant differences between studied groups in renal function for which data were available. |
| Symptoms and complications of disease:cardiac events | Outcome not reported. | Outcome not reported. | Not estimable | NA | NA | |
| Symptoms and complications of disease: cerebrovascular events | Outcome not reported. | Outcome not reported. | Not estimable | NA | NA | |
| *The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% CI) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). | ||||||
| GRADE Working Group grades of evidence | ||||||
| 1. We downgraded the quality of evidence by 1 for high risk due to conflict of interest in one trial. 2. We downgraded the quality of evidence by 1 due to imprecision. 3. We downgraded the quality of evidence by 1 due to inconsistency. 4. We downgraded the quality of evidence by 2 due to only one single study reporting this outcome (lack of statistical power). | ||||||
| Enzyme replacement therapy for Anderson‐Fabry disease | ||||||
| Patient or population: Anderson‐Fabry disease Settings: outpatients Intervention: agalsidase beta Comparison: placebo | ||||||
| Outcomes | Illustrative comparative risks (95% CI) | Relative effect | No of Participants | Quality of the evidence | Comments | |
| Assumed risk | Corresponding risk | |||||
| Placebo | Agalsidase beta | |||||
| Changes in globotriaosylceramide (Gb3) concentration in plasma Follow up: up to 18 months (umol/l) | The mean in changes in Gb3 concentration in plasma score was 5.5 (SD: 1.6) umol/l | The mean in changes in Gb3 concentration in plasma score in the intervention group was on average 4.80 fewer | NA | 58 1 study | ⊕⊝⊝⊝ | A statistically significant improvement on agalsidade beta compared to placebo was observed also observed in the domains of kidney (mean difference ‐1.70, (95% CI ‐2.09 to ‐1.31)) and heart (mean difference ‐0.90, 95% CI (‐1.18 to ‐0.62)) . |
| Death Follow up: 18 months | 0 per 1000 | 0 per 1000 (0 to 0) | 1.85 (0.08 to 43.96) | 82 1 study | ⊕⊝⊝⊝ | We found no statistically significant differences between studied groups in death which data were available. |
| Pain | Outcome not reported. | Outcome not reported. | Not estimable | NA | NA | |
| Renal function events (ITT population) Follow up: 18 months | 225 per 1000 | 195 (83 to 459) | 0.87 (0.37 to 2.04) | 82 1 study | ⊕⊝⊝⊝ | We found no statistically significant differences between studied groups in renal function events which data were available. |
| Symptoms and complications of disease: cardiac events Follow up: up to 18 months | 129 per 1000 | 59 (14 to 245) | 0.46 (0.11 to 1.90) | 82 1 study | ⊕⊝⊝⊝ | We found no statistically significant differences between studied groups in cardiac events which data were available. |
| Symptoms and complications of disease: cerebrovascular events (ITT population) Follow up: 18 months | 64 per 1000 | 7 (1 to 158) | 0.12 (0.01 to 2.48) | 82 1 study | ⊕⊝⊝⊝ | We found no statistically significant differences between studied groups in cerebrovascular events which data were available. |
| *The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% CI) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). | ||||||
| GRADE Working Group grades of evidence | ||||||
| 1. We downgraded the quality of evidence by 1 for high risk due to conflict of interest. 2. We downgraded the quality of evidence by 1 due to only one single study reporting this outcome (lack of statistical power). 3. We downgraded the quality of evidence by 1 due to imprecision. | ||||||
| Enzyme replacement therapy for Anderson‐Fabry disease | ||||||
| Patient or population: Anderson‐Fabry disease Settings: outpatients Intervention: agalsidase alfa Comparison: agalsidase beta | ||||||
| Outcomes | Illustrative comparative risks (95% CI) | Relative effect | No of Participants | Quality of the evidence | Comments | |
| Assumed risk | Corresponding risk | |||||
| Agalsidase alfa | Agalsidase beta | |||||
| Changes in globotriaosylceramide (Gb3) concentration in plasma and tissue Follow up: 12 months | Outcome not reported | Outcome not reported | Not estimable | NA | NA | |
| Death Follow up: not reported | 0 per 1000 | 0 (0 to 0) | 3.00 (0.13 to 69.09) | 36 1 study | ⊕⊝⊝⊝ | We found no statistically significant differences between studied groups in death which data were available. |
| Pain Acroparaesthesia and Fabry crises Follow up: not reported | Outcome not reported | Outcome not reported | Not estimable | NA | NA | |
| Renal function Follow up: up to 24 months. | Outcome not reported | Outcome not reported | Not estimable | NA | NA | |
| Symptoms and complications of disease: cardiac events Follow up: not reported | 142 per 1000 | 66 (7 to 653) | 0.47 (CI 0.05 to 4.60) | 29 1 study | ⊕⊝⊝⊝ | We found no statistically significant differences between studied groups in death which data were available. |
| Symptoms and complications of disease: cerebrovascular events Follow up: median 23.0 months | Outcome not reported. | Outcome not reported. | Not estimable. | NA | NA | |
| *The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% CI) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). | ||||||
| GRADE Working Group grades of evidence | ||||||
| 1. We downgraded the quality of evidence by 1 for high risk due to lack of blinding, incomplete outcome data and, selective reporting. 2. We downgraded the quality of evidence by 1 due to only one single study reporting this outcome (lack of statistical power). 3. We downgraded the quality of evidence by 1 due to imprecision. | ||||||
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 1 Plasma Gb3 Show forest plot | 2 | Mean Difference (IV, Random, 95% CI) | Subtotals only | |
| 1.1 At up to 6 months | 2 | 39 | Mean Difference (IV, Random, 95% CI) | ‐2.07 [‐6.64, 2.50] |
| 2 Urine sediment Gb3 Show forest plot | 1 | Mean Difference (IV, Fixed, 95% CI) | Totals not selected | |
| 2.1 At up to 6 months | 1 | Mean Difference (IV, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 3 Kidney Gb3 Show forest plot | 1 | Mean Difference (IV, Fixed, 95% CI) | Totals not selected | |
| 3.1 At up to 6 months | 1 | Mean Difference (IV, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 4 Myocardial globotriaosylceramide levels Show forest plot | 1 | Mean Difference (IV, Fixed, 95% CI) | Totals not selected | |
| 4.1 Over 3 months and up to 6 months | 1 | Mean Difference (IV, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 5 The Brief Pain Inventory Severity Show forest plot | 1 | Mean Difference (IV, Fixed, 95% CI) | Totals not selected | |
| 5.1 Over 1 month and up to 3 months | 1 | Mean Difference (IV, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 5.2 Over 3 months and up to 5 months | 1 | Mean Difference (IV, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 5.3 Over 5 months and up to 6 months | 1 | Mean Difference (IV, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 6 The Brief Pain Inventory pain‐related QoL Show forest plot | 1 | Mean Difference (IV, Fixed, 95% CI) | Totals not selected | |
| 6.1 Over 1 month and up to 3 months | 1 | Mean Difference (IV, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 6.2 Over 3 months and up to 5 months | 1 | Mean Difference (IV, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 6.3 Over 5 months and up to 6 months | 1 | Mean Difference (IV, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 7 Creatinine clearance Show forest plot | 1 | Mean Difference (IV, Fixed, 95% CI) | Totals not selected | |
| 7.1 At up to 6 months | 1 | Mean Difference (IV, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 8 Inulin clearance Show forest plot | 1 | Mean Difference (IV, Fixed, 95% CI) | Totals not selected | |
| 8.1 At up to 6 months | 1 | Mean Difference (IV, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 9 Glomeruli with mesangial widening Show forest plot | 1 | Mean Difference (IV, Fixed, 95% CI) | Totals not selected | |
| 9.1 At up to 6 months | 1 | Mean Difference (IV, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 10 Glomeruli with segmental sclerosis Show forest plot | 1 | Mean Difference (IV, Fixed, 95% CI) | Totals not selected | |
| 10.1 At up to 6 months | 1 | Mean Difference (IV, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 11 Obsolescent glomeruli Show forest plot | 1 | Mean Difference (IV, Fixed, 95% CI) | Totals not selected | |
| 11.1 At up to 6 months | 1 | Mean Difference (IV, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 12 Echocardiographic data on left ventricular structure and function over 3 months and up to 6 months Show forest plot | 1 | Mean Difference (IV, Fixed, 95% CI) | Totals not selected | |
| 12.1 Mean left ventricular wall thickness (mm) | 1 | Mean Difference (IV, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 12.2 Left ventricular internal diameter (mm) (diastolic) | 1 | Mean Difference (IV, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 12.3 Left ventricular internal diameter (mm) (systolic) | 1 | Mean Difference (IV, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 12.4 Left ventricular ejection fraction | 1 | Mean Difference (IV, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 1 Microvascular endothelial deposits of globotriaosylceramide Show forest plot | 1 | Mean Difference (IV, Fixed, 95% CI) | Totals not selected | |
| 1.1 Kidney ‐ over 3 months and up to 6 months | 1 | Mean Difference (IV, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 1.2 Heart ‐ over 3 months and up to 6 months | 1 | Mean Difference (IV, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 1.3 Composite ‐ over 3 months and up to 6 months | 1 | Mean Difference (IV, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 2 Death (ITT population) Show forest plot | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | Totals not selected | |
| 3 Renal events (ITT population) Show forest plot | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | Totals not selected | |
| 4 Cardiopulmonary exercise test Show forest plot | 1 | Mean Difference (IV, Fixed, 95% CI) | Totals not selected | |
| 4.1 Average heart rate reserve | 1 | Mean Difference (IV, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 4.2 Average maximum oxygen uptake measured at peak exercise (L/min) | 1 | Mean Difference (IV, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 4.3 Maximum oxygen uptake measured at peak exercise (ml/kg/min) | 1 | Mean Difference (IV, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 4.4 Oxygen pulse average at peak exercise | 1 | Mean Difference (IV, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 5 Cardiopulmonary exercise test Show forest plot | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | Totals not selected | |
| 5.1 Decrease in diastolic blood pressure | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 6 Cardiac events Show forest plot | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | Totals not selected | |
| 7 Cerebrovascular events (ITT population) Show forest plot | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | Totals not selected | |
| 8 Participants achieving zero scores in skin over 3 months and up to 6 months Show forest plot | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | Totals not selected | |
| 8.1 Superficial endothelial cells | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 8.2 Deep endothelial cells | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 8.3 Smooth muscle cells | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 8.4 Perineurium | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 9 Participants achieving zero score or reduction in skin over 3 months and up to 6 months Show forest plot | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | Totals not selected | |
| 9.1 Superficial endothelial cells | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 9.2 Deep endothelial cells | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 9.3 Smooth muscle cells | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 9.4 Perineurium | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 10 Chest pain Show forest plot | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | Totals not selected | |
| 10.1 At 24 months | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 11 Pain related to AFD Show forest plot | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | Totals not selected | |
| 11.1 Over 3 and up to 6 months | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 12 Fatigue Show forest plot | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | Totals not selected | |
| 12.1 At 24 months | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 13 Adverse event: rigors Show forest plot | 2 | Risk Ratio (M‐H, Fixed, 95% CI) | Totals not selected | |
| 13.1 Over 3 and up to 6 months | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 13.2 At 24 months | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 14 Adverse event: fever Show forest plot | 2 | Risk Ratio (M‐H, Fixed, 95% CI) | Totals not selected | |
| 14.1 Over 3 and up to 6 months | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 14.2 At 24 months | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 15 Adverse event: chills Show forest plot | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | Totals not selected | |
| 15.1 Over 3 and up to 6 months | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 16 Adverse event: hypertension Show forest plot | 2 | Risk Ratio (M‐H, Fixed, 95% CI) | Totals not selected | |
| 16.1 Over 3 and up to 6 months | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 16.2 At 24 months | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 17 Adverse event: vomiting Show forest plot | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | Totals not selected | |
| 17.1 At 24 months | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 18 Adverse event: headache Show forest plot | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | Totals not selected | |
| 18.1 Over 3 and up to 6 months | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 1 Death Show forest plot | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | Subtotals only | |
| 2 Cardiac events Show forest plot | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | Subtotals only | |
| 3 Any adverse event Show forest plot | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | Subtotals only | |
| 4 Any serious adverse events Show forest plot | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | Totals not selected | |
| 4.1 New Subgroup | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 1 Pain Show forest plot | 1 | Mean Difference (IV, Fixed, 95% CI) | Subtotals only | |
| 2 Self‐assessed health state Show forest plot | 1 | Mean Difference (IV, Fixed, 95% CI) | Subtotals only | |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 1 Pain Show forest plot | 1 | Mean Difference (IV, Fixed, 95% CI) | Subtotals only | |
| 2 Self‐assessed health state Show forest plot | 1 | Mean Difference (IV, Fixed, 95% CI) | Subtotals only | |
