Background
The X-linked disorder Fabry disease (FD) is caused by deficient activity of the glycolipid-degrading enzyme α-galactosidase A, which leads to progressive accumulation of the substrate globotriaosylceramide (Gb
3) in multiple body organs, resulting in potentially severe and ultimately premature fatality [
1]. Initial signs and symptoms of FD are manifested during early childhood in both sexes, but especially in affected boys, with symptoms including angiokeratomas, neuropathic pain/acroparesthesia, hypohidrosis, gastrointestinal symptoms, cornea verticillata, and less commonly, cardiac, renal, and cerebrovascular involvement [
2].
Enzyme replacement therapy (ERT) is available as a treatment for patients with FD. However, the published literature of ERT safety and efficacy in children with FD is not as robust when compared with adults. Previously, the clinical efficacy and safety of agalsidase alfa was evaluated in 17 children with FD in studies TKT023 and TKT029, conducted over 4 years [
3]. Agalsidase alfa was generally well tolerated, and improvements were observed in levels of urine and plasma Gb
3 concentration, pain severity, and heart rate variability (HRV), while estimated glomerular filtration rate (eGFR) and left ventricular mass indexed to height (LVMI) remained stable.
This study reports on the long-term (6.5 year), open-label, follow-up of patients who qualified for and opted to transition from study TKT023 to an extension trial (TKT029). The objective of extension study TKT029 was to evaluate the safety and efficacy of agalsidase alfa in pediatric patients with FD treated for up to 7 cumulative years.
Methods
Study design, patient selection, and treatment
This open-label, multicenter extension study (TKT029; June 10, 2004–June 15, 2011; ClinicalTrials.gov identifier NCT00084084) was designed for pediatric FD patients (7–17 years of age at study enrollment) who had received 6 months of 0.2 mg/kg agalsidase alfa in study TKT023 (August 12, 2002–October 20, 2004) and were within 30 (±7) days of study completion [
3]. Together, studies TKT023 (0.5 years) and TKT029 (6.5 years) comprised patients treated for up to 7 years with agalsidase alfa. Study TKT029 was divided into two phases: before (phase 1) and after (phase 2) a change in the agalsidase alfa manufacturing process [
3]. Patients were included in this report only if they participated in both phases (transition safety population [TSP]). Phase 2 began ~197 to 223 (mean, 210) weeks after phase 1 baseline agalsidase alfa treatment.
Patients were included if they were determined to be of adequate general health, without potential safety issues or medical contraindications, and had written informed consent provided by a parent or legal guardian. Patients were excluded if they or their legal guardian were deemed unable to understand the study requirements and potential outcomes, or if they were determined by the local investigator as unlikely to follow the study protocol. In both studies TKT023 and TKT029, patients received 0.2 mg/kg body weight of agalsidase alfa every other week, with each intravenous infusion delivered over a 40-minute period.
Safety and efficacy endpoints
The primary study TKT029 endpoints were safety and tolerability of agalsidase alfa and its effect on HRV. Secondary objectives were to determine the pharmacokinetics of agalsidase alfa at baseline and after treatment initiation, as well as exploratory measurements of renal function (i.e., eGFR), LVMI, and other clinical and patient-reported outcomes (e.g., plasma and urine Gb3, pain, health-related quality of life [HRQoL]). As a post hoc analysis, urine protein: creatinine ratios were evaluated.
Safety assessments
Adverse events (AEs) were characterized by severity, potential relationship to study drug and/or infusion, and whether they were classifiable as a serious AE (SAE; an AE that resulted in death, was life-threatening, caused new or prolonged hospitalization, led to persistent disability or congenital abnormality, or was considered an SAE by the treating investigator). Additional safety monitoring included clinical laboratory parameters, vital signs, physical and neurologic examinations, 12-lead electrocardiograms, and potential anti-agalsidase alfa antibody activity evaluated from blood samples screened with an enzyme-linked immunosorbent assay and confirmed by a titration assay. Antibody-positive samples were isotyped (immunoglobulin [Ig] G, IgA, IgM, or IgE) and tested for agalsidase alfa neutralizing activity using an
in vitro assay [
4].
Efficacy and pharmacodynamic assessments
Cardiac function and structure were assessed through HRV and LVMI, respectively. HRV was assessed via 2-hour Holter monitoring. The time-domain HRV parameters assessed in this study were SDNN (standard deviation [SD] of all filtered RR intervals of the length of the analysis), r-MSSD (square root of the sum of squares of differences between adjacent filtered RR intervals over the length of the analysis), and pNN50 (percentage of differences between adjacent filtered RR intervals that are greater than 50 msec for the whole analysis). SDNN was used as an overall index of HRV, as it encompasses both long- (sympathetic) and short-term (parasympathetic) variability and thus reflects the overall autonomic nervous system activity in the heart. Both r-MSSD and pNN50 describe short-term variability and thus reflect primarily parasympathetic influences on heart rate. A reduction in parasympathetic stimulation of the heart has been reported in male pediatric patients with FD [
5]. In addition, heart rate was obtained from 12-lead electrocardiogram measurements. Echocardiogram and the Devereux equations were used to calculate LVMI [
6]. LVMI baseline for study TKT029 was week 25/26 of TKT023 (i.e., prior to the first dose of study drug in TKT029, but after the patients had received treatment in TKT023).
Renal function was evaluated through eGFR. The Counahan-Barratt equation was used to calculate eGFR for patients younger than 18 years of age, using height and serum creatinine [
7]. For patients who aged to 18 years and over during the study, the Modification of Diet in Renal Disease (MDRD) equation was utilized instead, which factored-in serum creatinine, age, race, and gender [
8]. To assess the impact of the equation change at the age of 18 years on the eGFR values, sensitivity analyses were performed using continued eGFR calculated from the Counahan-Barratt equation even after patients turned 18 years old, or the Chronic Kidney Disease Epidemiology Collaboration equation once patients turned 18 years old. As a
post hoc analysis, 8-hour urine collection was used to estimate urine protein level and proteinuria (defined as urine protein:creatinine ratio ≥0.2).
Plasma and urine Gb
3 were measured at 8-week intervals in study TKT023 and every 6 months in study TKT029, using a high-performance liquid chromatography assay [
9]. Other measurements included the Brief Pain Inventory (severity and interference of pain) [
10], Health Utility Index Mark 2 and 3 (HUI; HRQoL) [
11],[
12], and Child Health Questionnaire (CHQ; HRQoL) [
13].
Statistical analysis
The analyses of data are descriptive in nature and no formal inferential statistical tests were performed. For variables following a continuous distribution, tabular summaries consist of descriptive statistics (e.g., means, SDs, medians, minimums, and maximums) and observed and change from baseline values are presented. For categorical and ordinal variables, tabular summaries consist of the number and percentage. For the clinical outcome endpoints (eGFR, plasma and urine sediment Gb3, LVMI, and HRV), the percentage change from baseline for these parameters were additionally calculated. As post hoc analyses, annualized rate of change (slopes) were estimated for LVMI, eGFR, and urine protein:creatinine ratio based on the random coefficient model using time as a regressor to fit each patient’s data; then, the slopes were estimated by averaging the rate of change across all patients analyzed. This analysis took into account repeated measurements over time and included data from studies TKT023 and TKT029 (phases 1 and 2).
Discussion
This study represents the longest assessment of ERT for the treatment of children with FD in a clinical trial setting. Agalsidase alfa was generally well tolerated over approximately 7 years of cumulative treatment with most treatment-emergent AEs being mild or moderate in intensity. No treatment-related SAEs were detected; no patients discontinued study participation due to AEs and there were no deaths. The incidence of drug-related and infusion-related AEs decreased over time. No other safety concerns were identified, based on other assessed clinical parameters. Anti-agalsidase alfa antibody formation was low (3/11, 27.3% of the TSP) and with no apparent impact on clinical outcomes. No new safety concerns were identified from all assessed clinical parameters.
HRV improvements originally observed in study TKT023 appeared to be maintained in this extension trial, with an upward trend in SDNN (milliseconds) during phase 2 (SDNN HRV at baseline phase 1: 91.96 msec, by week 130 phase 2: 155.81 msec). A previous study found that normal mean SDNN HRV (by 24-hour Holter monitor) in children increases with age (boys: 57 ± 62 msec [aged 3.4 ± 1.6 years] to 187 ± 38 msec [aged 16.4 ± 0.8 years]; girls: 87 ± 16 msec [aged 2.7 ± 1.8 years] to 201 ± 24 msec [aged 17.4 ± 1.7 years]) [
14]. Several patients experienced a reduction in heart rate to below 60 beats per minute based on the 12-lead electrocardiogram, suggestive of bradycardia (Additional file
1: Figure S1). However, bradycardia was not specifically tested for as an endpoint in this study, and no investigators reported an instance of bradycardia as a treatment-emergent AE. Future research would be required to evaluate whether bradycardia is potentially a bradyarrhythmia associated with FD [
15],[
16].
In addition, LVMI remained generally stable and none of the patients reached the adult criteria for left ventricular hypertrophy. Although no studies have been published evaluating annualized slope in LVMI in a population comprising only children with FD, a cross-sectional echocardiographic study of FD in untreated patients (including adults and children) found that males without LVH had a mean (± standard error of the mean) annualized LVMI change of +4.07 ± 1.03 g/m
2.7 (
n =39) and females had 2.31 ± 0.81 g/m
2.7 (
n =39) [
17]. In patients aged <20 years, males experienced a median annualized rate of change of +2.00 g/m
2.7 (
n =5) and females changed by +1.36 g/m
2.7 (
n =9). In study TKT029 (pediatric, predominantly male patients), LVMI annualized slope was −0.48 (95% CI: −1.33, 0.38) g/m
2.7. The level of LVMI directly correlates with increasing age [
17], so future research would be required for a comparative analysis of current results and the natural history of LVMI in children with FD.
Similarly, renal function was generally stable over the course of the study for most of the patients, albeit with considerable variability during follow-up (e.g., some periods of hyperfiltration with an eGFR >130 mL/min/1.73 m
2 were observed). Although we are not aware of a study evaluating annualized slope change in eGFR in children with FD, eGFR slopes were assessed in a retrospective chart review study of natural history in a mixed population of adults and children (median [range] age 41.0 years [5.0–77.1]) [
18]. In the latter study, males and females (who did not progress to end-stage renal disease) experienced annualized mean eGFR changes of −2.93 mL/min/1.73 m
2 (
n =128) and −1.02 mL/min/1.73 m
2 (
n =51), respectively. The overall annualized slope in study TKT029 showed a small upward trend of +0.22 (95% CI: −2.84, 3.28) mL/min/1.73 m
2 through phase 2. As decline in eGFR would be expected to be higher in adult patients compared with children, additional research would be required to determine if early initiation of agalsidase alfa therapy during childhood could have preventive or ameliorating effects on decline in eGFR with aging. The annualized slope of protein:creatinine ratio was relatively stable (+0.02 [95% CI: 0.00, 0.03]), no patients entered the proteinuria range (protein:creatinine ratio ≥0.2) during treatment for more than three assessments, and none had proteinuria during their final visit.
Both plasma and urine Gb
3 showed decreases from the baseline to the end of study TKT023, and these reductions in Gb
3 were essentially maintained in study TKT029. The presence of neutralizing anti-agalsidase alfa antibodies has been reported to have a negative effect on the reduction of urine Gb
3 levels [
19]. Here, only one patient developed anti-agalsidase alfa IgG antibodies with neutralizing activity. This patient had a moderately increased level of urine Gb
3 at baseline that fluctuated above baseline throughout the study and was slightly higher than baseline level at study completion. However, another patient who was anti-agalsidase alfa antibody negative with a moderately elevated level of urine Gb
3 at baseline, showed fluctuations at various time points, and at study completion had a level similar to that at baseline. It is thus not possible to draw any definitive conclusions regarding the effect of neutralizing anti-agalsidase alfa antibodies on urine Gb
3 levels in our study.
This study has several limitations. First, TKT029 was not a comparative or placebo-controlled clinical trial, which limits the ability to draw conclusions relative to a control population. In addition, the number of patients in the TSP was small, limiting the power to detect within-patient effects. Because of the lack of a comparator group and low patient number, conducting inferential statistical analyses was impractical, so the results are descriptive only. Furthermore, some patients had transitioned from childhood to adolescence and adulthood. The inclusion of adult patients made evaluation of certain endpoints challenging, as different validated equations needed to be used (e.g., Counahan-Barratt equation for pediatric eGFR versus MDRD for adult eGFR). Nevertheless, we made every effort to conduct sensitivity analyses to verify the results. Urine protein data were assessed from 8-hour urine collection; the accuracy of the estimates may be limited because of inconsistent reporting by individual laboratories as well as missing values. Finally, the prognostic value of HRV in pediatric patients has not been well established, although one study found a reduction in HRV in boys, but not girls [
5]. Future studies will be necessary to evaluate proteinuria effects in detail.
Acknowledgments
We thank Drs. Lynda Bideau, Yoko Broussard, Michael Cohen, Joe TR Clarke, Karen Johnson, Christoph Kampmann, Li-Wen Lai, Manju Thomas, Ray Pais, Tyler Reimschisel, Markus Ries, Caren Swift, and Alison Whelan, who also treated patients involved in study TKT029. Studies TKT023 and TKT029 were funded by Shire. Medical writing support was provided by Ray Beck, Jr., PhD, at Excel Scientific Solutions and was funded by Shire.
Competing interests
RS has received honoraria, travel reimbursement and research support from Shire, Amicus Therapeutics, and Genzyme in the past 5 years. GMP has received research grants or support from Actelion, Amicus/GSK, Biomarin, Genzyme/Sanofi, Protalix/Pfizer, and Shire. YHL has received research support from Shire. VC declares no potential competing interests. PC, RM, and AW are employees of Shire. Medical writing support, including writing the first draft under the guidance of the entire author group, was provided by Ray Beck, Jr., PhD, at Excel Scientific Solutions and was funded by Shire. No authors received any form of payment for the development of the manuscript.
Authors’ contributions
All authors conducted the study and collected the data. PC performed statistical analyses. All authors analyzed and interpreted the data, and contributed to development of the manuscript. All authors read and approved the final manuscript.