Matching-Adjusted Indirect Comparisons (MAICs) and Network Meta-Analyses (NMAs) of the Oral Small-Molecule Chaperone Migalastat vs. Intravenous Enzyme Replacement Therapies (ERTs) for Clinical Measures in Fabry Disease
Author(s)
Noemi Hummel, PhD1, Agnieszka Kopiec, MSc2, Matthew Crabtree, MPH3, Alasdair MacCulloch, BSc, PhD3, Charlotte Sharples, MD3, Biliana O. Veleva-Rotse, PhD4, Neil Johnson, BSc3, Vera Gielen, PhD3.
1Certara GmbH, Lörrach, Germany, 2Certara, Krakow, Poland, 3Amicus Therapeutics UK Ltd, Marlow, United Kingdom, 4Amicus Therapeutics, Inc., Princeton, NJ, USA.
1Certara GmbH, Lörrach, Germany, 2Certara, Krakow, Poland, 3Amicus Therapeutics UK Ltd, Marlow, United Kingdom, 4Amicus Therapeutics, Inc., Princeton, NJ, USA.
OBJECTIVES: ATTRACT (NCT01218659) compared migalastat with ERT (agalsidase alfa/beta [AGAL-α/β]) in patients with Fabry disease and amenable GLA variants. We indirectly compared migalastat with pegunigalsidase alfa (PEG) for key cardiac/renal measures and with any ERT for long-term risk of Fabry-associated clinical events (FACEs; specific cardiac/renal/cerebrovascular events/death).
METHODS: Systematic/targeted literature reviews identified publications/studies reporting left ventricular mass index (LVMi), annualised change in estimated glomerular filtration rate (eGFR slope) and/or FACEs. FACE data were extracted from Kaplan-Meier plots. Indirect treatment comparisons were feasible, including population-adjusted methods (anchored/unanchored MAICs using patient-level data from ATTRACT [LVMi/eGFR slope/FACEs] and its open-label extension [FACEs], and aggregated data from comparator studies) and NMAs. For MAICs, age, sex, either eGFR or previous ERT duration (scenarios to increase effective sample size [ESS]), and angiotensin-converting enzyme inhibitor/angiotensin II receptor blocker (ACEi/ARB) use were matched whenever reported.
RESULTS: In unanchored MAICs mean differences (95% confidence intervals [CIs]) were as follows. LVMi change from baseline (negative values favour migalastat): migalastat (echocardiogram) versus PEG from BALANCE (magnetic resonance imaging) −5.94 (−13.38-1.50) ESS=7. eGFR slope (positive values favour migalastat), migalastat versus PEG from: BALANCE 1.02 (−3.55-5.59) ESS=7; BRIGHT 0.17 (−2.19-2.53) ESS=13; BRIGHT51 1.37 (−0.60-3.33) ESS=14; NCT01981720 0.22 (−2.34-2.77) ESS=9. LVMi and eGFR from NMAs/anchored MAICs showed similar numerical trends. For FACEs in unanchored MAICs, hazard ratios (95% CIs) were (values <1 favour migalastat): migalastat versus AGAL-α from Giugliani 2023/FOS (0.08 [0.02-0.33] ESS=12); versus AGAL-α and AGAL-β from Arends 2018 (0.59 [0.23-1.54] ESS=39 and 0.53 [0.21-1.39] ESS=33, respectively); versus AGAL-α/β from Sirrs 2014/CFDI (ERT-experienced: 0.32 [0.10-1.07] ESS=22, ERT-naïve: 0.76 [0.28-2.08] ESS=28). No PEG studies reported FACEs. A limitation was the unfeasibility of matching on GLA variant amenability.
CONCLUSIONS: In populations matched for age, sex, eGFR/previous ERT duration and ACEi/ARB, LVMi change and eGFR slope were similar for migalastat and PEG; long-term FACE risk was similar for migalastat and AGAL-α/β.
METHODS: Systematic/targeted literature reviews identified publications/studies reporting left ventricular mass index (LVMi), annualised change in estimated glomerular filtration rate (eGFR slope) and/or FACEs. FACE data were extracted from Kaplan-Meier plots. Indirect treatment comparisons were feasible, including population-adjusted methods (anchored/unanchored MAICs using patient-level data from ATTRACT [LVMi/eGFR slope/FACEs] and its open-label extension [FACEs], and aggregated data from comparator studies) and NMAs. For MAICs, age, sex, either eGFR or previous ERT duration (scenarios to increase effective sample size [ESS]), and angiotensin-converting enzyme inhibitor/angiotensin II receptor blocker (ACEi/ARB) use were matched whenever reported.
RESULTS: In unanchored MAICs mean differences (95% confidence intervals [CIs]) were as follows. LVMi change from baseline (negative values favour migalastat): migalastat (echocardiogram) versus PEG from BALANCE (magnetic resonance imaging) −5.94 (−13.38-1.50) ESS=7. eGFR slope (positive values favour migalastat), migalastat versus PEG from: BALANCE 1.02 (−3.55-5.59) ESS=7; BRIGHT 0.17 (−2.19-2.53) ESS=13; BRIGHT51 1.37 (−0.60-3.33) ESS=14; NCT01981720 0.22 (−2.34-2.77) ESS=9. LVMi and eGFR from NMAs/anchored MAICs showed similar numerical trends. For FACEs in unanchored MAICs, hazard ratios (95% CIs) were (values <1 favour migalastat): migalastat versus AGAL-α from Giugliani 2023/FOS (0.08 [0.02-0.33] ESS=12); versus AGAL-α and AGAL-β from Arends 2018 (0.59 [0.23-1.54] ESS=39 and 0.53 [0.21-1.39] ESS=33, respectively); versus AGAL-α/β from Sirrs 2014/CFDI (ERT-experienced: 0.32 [0.10-1.07] ESS=22, ERT-naïve: 0.76 [0.28-2.08] ESS=28). No PEG studies reported FACEs. A limitation was the unfeasibility of matching on GLA variant amenability.
CONCLUSIONS: In populations matched for age, sex, eGFR/previous ERT duration and ACEi/ARB, LVMi change and eGFR slope were similar for migalastat and PEG; long-term FACE risk was similar for migalastat and AGAL-α/β.
Conference/Value in Health Info
2025-11, ISPOR Europe 2025, Glasgow, Scotland
Value in Health, Volume 28, Issue S2
Code
SA65
Topic
Study Approaches
Topic Subcategory
Meta-Analysis & Indirect Comparisons
Disease
Cardiovascular Disorders (including MI, Stroke, Circulatory), Neurological Disorders, Rare & Orphan Diseases, Urinary/Kidney Disorders