Background

Pompe disease is a rare, progressive neuromuscular disorder caused by deficiency of acid α-glucosidase (GAA) and accumulation of lysosomal glycogen. We assessed the safety and efficacy of avalglucosidase alfa, a recombinant human GAA enzyme replacement therapy specifically designed for enhanced mannose-6-phosphate-receptor targeting and enzyme uptake aimed at increased glycogen clearance, compared with the current approved standard of care, alglucosidase alfa, in patients with late-onset Pompe disease.

Methods

We did a randomised, double-blind, phase 3 trial at 55 sites in 20 countries. We enrolled individuals (aged ≥3 years) with enzymatically confirmed late-onset Pompe disease who had never received treatment. We used a centralised treatment allocation system to randomly allocate participants to either avalglucosidase alfa or alglucosidase alfa. Participants and investigators were unaware of their treatment allocation. The primary outcome measure was change from baseline to week 49 in upright forced vital capacity percent (FVC%) predicted. We used a hierarchical fixed sequential testing strategy, whereby non-inferiority of avalglucosidase alfa compared with alglucosidase alfa was assessed first, with a non-inferiority margin of 1·1. If non-inferiority was seen, then superiority was tested with a 5% significance level. The key secondary objective was effect on functional endurance, measured by the 6-minute walk test (6MWT). Safety was assessed, including treatment-emergent adverse events and infusion-associated reactions. The modified intent-to-treat population was the primary analysis population for all efficacy analyses. The safety population was the analysis population for safety analyses. This trial is registered with ClinicalTrials.gov, NCT02782741. We report results of the 49-week primary analysis period.

Findings

Between Nov 2, 2016, and March 29, 2019, 100 participants were randomly allocated avalglucosidase alfa (n=51) or alglucosidase alfa (n=49). Treatment with avalglucosidase alfa resulted in a least-squares mean improvement in upright FVC% predicted of 2·89% (SE 0·88) compared with 0·46% (0·93) with alglucosidase alfa at week 49 (difference 2·43% [95% CI −0·13 to 4·99]). Non-inferiority was shown because the lower bound of the 95% CI for the difference far exceeded the predefined non-inferiority margin but did not exclude 0 (p=0·0074). Superiority was not reached (p=0·063), so formal testing was stopped, as per the testing hierarchy. Improvements were also seen in the 6MWT with avalglucosidase alfa compared with alglucosidase alfa, with greater increases in distance covered (difference 30·01 m [95% CI 1·33 to 58·69]) and percent predicted (4·71% [0·25 to 9·17]). Treatment-emergent adverse events potentially related to treatment were reported in 23 (45%) of 51 participants in the avalglucosidase alfa group and in 24 (49%) of 49 in the alglucosidase alfa group, and infusion-associated reactions were reported in 13 (26%) participants in the avalglucosidase alfa group and 16 (33%) in the alglucosidase alfa group. Of the five trial withdrawals, all in the alglucosidase alfa group, four were due to adverse events, including two infusion-associated reactions. Serious treatment-emergent adverse events were reported in eight (16%) participants who received avalglucosidase alfa and in 12 (25%) who received alglucosidase alfa. One participant treated with alglucosidase alfa died because of acute myocardial infarction determined to be unrelated to treatment. Antidrug antibody responses were similar in both groups. High and persistent titres (≥12 800) and neutralising antibodies were more common with alglucosidase alfa (in 16 [33%] participants) than with avalglucosidase alfa (ten [20%]).

Interpretation

We consider that this study provides evidence of clinically meaningful improvement with avalglucosidase alfa therapy over alglucosidase alfa in respiratory function, ambulation, and functional endurance, with no new safety signals reported. An open-label extended-treatment period is ongoing to confirm the long-term safety and efficacy of avalglucosidase alfa, with the aim for this therapy to become the new standard treatment in late-onset Pompe disease.

Funding

Sanofi Genzyme.

中文翻译：

---

阿瓦葡萄糖苷酶 α 与阿葡萄糖苷酶 α 在迟发性庞贝病 (COMET) 患者中的安全性和有效性：一项 3 期、随机、多中心试验

背景

庞贝病是一种罕见的进行性神经肌肉疾病，由酸性α-葡萄糖苷酶（GAA）缺乏和溶酶体糖原积累引起。我们评估了阿法葡萄糖苷酶 α 的安全性和有效性，这是一种重组人 GAA 酶替代疗法，专门设计用于增强 6-磷酸甘露糖受体靶向和酶摄取，旨在增加糖原清除率，与目前批准的护理标准阿葡萄糖苷酶 α 相比，在患有迟发性庞贝病的患者中。

方法

我们在 20 个国家的 55 个地点进行了一项随机、双盲、3 期试验。我们招募了从未接受过治疗的酶促证实的迟发性庞贝病患者（年龄≥3 岁）。我们使用集中治疗分配系统将参与者随机分配到阿法葡萄糖苷酶α或阿糖苷酶α。参与者和调查人员不知道他们的治疗分配。主要结果测量是预测的直立用力肺活量百分比 (FVC%) 从基线到第 49 周的变化。我们使用了分层固定顺序测试策略，首先评估阿伐葡糖苷酶 α 与阿糖苷酶 α 相比的非劣效性，非劣效性边界为 1·1。如果观察到非劣效性，则以 5% 的显着性水平测试优效性。关键的次要目标是对功能耐力的影响，通过 6 分钟步行测试 (6MWT) 测量。评估了安全性，包括治疗中出现的不良事件和输液相关反应。修改后的意向治疗人群是所有疗效分析的主要分析人群。安全性人群是安全性分析的分析人群。该试验已在 ClinicalTrials.gov 注册，NCT02782741。我们报告了 49 周主要分析期的结果。该试验已在 ClinicalTrials.gov 注册，NCT02782741。我们报告了 49 周主要分析期的结果。该试验已在 ClinicalTrials.gov 注册，NCT02782741。我们报告了 49 周主要分析期的结果。

发现

在 2016 年 11 月 2 日至 2019 年 3 月 29 日期间，100 名参与者被随机分配到阿法葡糖苷酶α（n=51）或阿葡糖苷酶α（n=49）。在第 49 周，阿伐糖苷酶 α 治疗导致直立 FVC% 预测的最小二乘平均改善为 2·89% (SE 0·88)，而阿糖苷酶 α 治疗为 0·46% (0·93)（差异 2·43 % [95% CI -0·13 至 4·99]）。显示非劣效性是因为差异的 95% CI 的下限远远超过了预定义的非劣效性界限，但不排除 0 (p=0·0074)。没有达到优越性 (p=0·063)，因此根据测试层次结构停止了正式测试。与阿葡糖苷酶 α 相比，使用阿法葡糖苷酶 α 的 6MWT 也有改善，随着覆盖距离的增加（差异 30·01 m [95% CI 1·33 到 58·69]）和预测百分比（4·71% [0·25 到 9·17]）。阿伐葡萄糖苷酶 α 组 51 名参与者中有 23 名（45%）和阿糖苷酶 α 组 49 名参与者中有 24 名（49%）报告了可能与治疗相关的治疗中出现的不良事件，13 名参与者报告了输液相关反应。 26%) 参与者参加阿法葡萄糖苷酶 α 组和 16 名参与者 (33%) 参加阿糖苷酶 α 组。在全部在阿糖苷酶α组中的五项试验退出中，四项是由于不良事件，包括两项输液相关反应。8 名 (16%) 接受阿法葡萄糖苷酶的参与者和 12 名 (25%) 接受阿法葡萄糖苷酶的参与者报告了严重的治疗出现的不良事件。一名接受阿糖苷酶 α 治疗的参与者死于与治疗无关的急性心肌梗死。两组的抗药抗体反应相似。阿糖苷酶α（16名[33%]参与者）的高滴度和持续滴度（≥12 800）和中和抗体比阿阿葡糖苷酶α（10名[20%]）更常见。

解释

我们认为，这项研究提供的证据表明，阿法葡糖苷酶 α 治疗在呼吸功能、行走和功能耐力方面优于阿糖苷酶 α 治疗具有临床意义的改善，但没有报告新的安全信号。一个开放标签的延长治疗期正在进行中，以确认 avalglucosidase alfa 的长期安全性和有效性，旨在使该疗法成为迟发性庞贝病的新标准治疗方法。

资金

赛诺菲健赞。