Outcomes 成果

The primary endpoint of FIREFISH part 2 was proportion of infants sitting without support for at least 5 s at month 12, assessed by item 22 of the BSID-III gross motor subscale. The secondary endpoints in the statistical hierarchy at month 12 were: proportion of infants who achieve a score of at least 40 on the CHOP-INTEND (the scale ranges from 0–64, the higher the score the better the motor function); proportion of infants who achieve an increase of at least 4 points in CHOP-INTEND score from baseline; proportion of motor milestone responders as assessed by the HINE-2 scale (the definition of motor milestone responder is presented in the appendix [p 11]); and proportion of infants who are alive without permanent ventilation (event-free survival; patients who died or met the criteria for permanent ventilation).

FIREFISH 第 2 部分的主要终点是婴儿在第 12 个月时在没有支撑的情况下至少坐 5 秒钟的比例，该比例由 BSID-III 大运动分量表第 22 项进行评估。第 12 个月时的次要统计终点为CHOP-INTEND评分至少达到40分的婴儿比例（评分范围为0-64分，分数越高，运动功能越好）；CHOP-INTEND评分比基线至少增加4分的婴儿比例；根据HINE-2量表评估的运动里程碑反应者比例（运动里程碑反应者的定义见附录[第11页]）；以及在无永久通气的情况下存活的婴儿比例（无事件存活率；死亡或符合永久通气标准的患者）。

Secondary endpoints included in the statistical hierarchy at month 24 were assessed using selected items from the BSID-III gross motor subscale: proportion of infants sitting without support for at least 30 s (item 26), proportion of infants standing alone (item 40), and proportion of infants walking alone (item 42). These motor outcomes are clinically relevant for infants with type 1 spinal muscular atrophy and were prespecified in the statistical analysis plan.

第24个月时纳入统计等级的次要终点使用BSID-III粗大运动分量表中的选定项目进行评估：在没有支撑的情况下坐至少30秒的婴儿比例（第26项）、独自站立的婴儿比例（第40项）和独自行走的婴儿比例（第42项）。这些运动结果与 1 型脊髓性肌萎缩症婴儿的临床相关，并在统计分析计划中进行了预设。

Secondary endpoints not included in the statistical hierarchy at month 24 were not adjusted for multiplicity, and thus no definitive conclusions can be drawn for these endpoints. The endpoints were: proportion of infants who achieve head control (defined as a score of 3 or greater for item 12 of the CHOP-INTEND); change from baseline in the total raw score of the BSID-III gross motor subscale;achievement of motor milestones as measured by the HINE-2 (milestones include head control, sitting, voluntary grasp, ability to kick, rolling, crawling, standing, and walking); proportion of motor milestone responders as assessed by the HINE-2 (appendix p 11); proportion of infants who can sit without support for at least 5 s, as assessed by the BSID-III gross motor subscale; proportion of infants who are alive; proportion of infants who are alive without permanent ventilation (event-free survival; patients who died or met the criteria for permanent ventilation); proportion of infants who are without permanent ventilation (patients who died without meeting the criteria for permanent ventilation are censored); proportion of infants who do not require invasive or non- invasive respiratory support; proportion of infants able to feed orally; and highest motor milestone achieved during the 24 months of treatment, as assessed by six items of the BSID-III gross motor subscale (head control [item 9, “controls head while upright for 15 s”], rolling [item 14, “rolls from side to back”], sitting without support for 5 s [item 22, “sitting without support for 5 s”], crawling [item 30, “crawls on stomach”], standing [item 40, “stands alone”],and walking [item 42, “walks alone”]).

未纳入第 24 个月统计层次的次要终点未进行多重性调整，因此无法就这些终点得出明确结论。这些终点是实现头部控制的婴儿比例（定义为 CHOP-INTEND 第 12 项得分达到或超过 3 分）；BSID-III 大运动分量表原始总分与基线相比的变化；通过 HINE-2 测定实现运动里程碑（里程碑包括头部控制、坐、自主抓握、踢能力、滚动、爬行、站立和行走）；通过 HINE-2 评估的运动里程碑反应者比例（附录第 11 页）；通过 BSID-III 大运动分量表评估的在没有支撑的情况下能坐至少 5 秒钟的婴儿比例；存活婴儿比例；在没有永久通气的情况下存活的婴儿比例（无事件存活率；死亡或达到永久通气标准的患者）；无永久通气的婴儿比例（未达到永久通气标准而死亡的患者将被剔除）；不需要有创或无创呼吸支持的婴儿比例；能够口服喂养的婴儿比例；治疗 24 个月期间达到的最高运动里程碑，由 BSID-III 大运动分量表的六个项目评估（头部控制[第 9 项，"直立时控制头部 15 秒"]、翻滚[第 14 项，"从一侧翻到另一侧"]、无支撑坐 5 秒[第 22 项，"无支撑坐 5 秒"]、爬行[第 30 项，"匍匐爬行"]、站立[第 40 项，"独自站立"]和行走[第 42 项，"独自行走"]）。

Safety assessments were incidence and severity of adverse events, laboratory values, ECG, vital signs (body temperature, systolic and diastolic blood pressure, pulse rate, and respiratory rate), and ophthalmological, physical, and anthropometric examinations. SMN protein levels were measured for every patient from venous blood samples.

安全性评估包括不良事件的发生率和严重程度、实验室值、心电图、生命体征（体温、收缩压和舒张压、脉搏和呼吸频率）以及眼科、体格和人体测量检查。对每位患者的静脉血样本进行了 SMN 蛋白水平测定。

Additional methodological details (including statistical methods, information on the hierarchical endpoint analysis, and a full list of safety assessments) are presented in the appendix (p 8).

其他方法细节（包括统计方法、分层终点分析信息和完整的安全性评估清单）见附录（第 8 页）。

Statistical analysis 统计分析

To assess the efficacy of risdiplam, a performance criterion was defined for the primary endpoint based on the well established natural history of type 1 spinal muscular atrophy. For secondary endpoints at month 12 included in the statistical hierarchy, performance criteria were based on the upper limit of the 90% CIs from the historical data on individuals who met each endpoint. CIs were calculated using the Clopper-Pearson method for the BSID-III, CHOP-INTEND, and HINE-2 endpoints, and the complementary log-log transformation for the proportion of infants alive without permanent ventilation. For secondary endpoints at month 24 included in the statistical hierarchy, performance criteria were based on the well defined natural history of type 1 spinal muscular atrophy. Details of the predefined performance criteria are available in the appendix (p 21). The study protocol and statistical analysis plan prespecified the use of 90% CIs for proportions and to match the one-sided statistical tests employed for the hypotheses testing, which was used for endpoints with a predefined performance criterion. An exact binomial test was used for the BSID-III, CHOP-INTEND, and HINE-2 endpoints, and a Z test was conducted for event-free survival.

为了评估利斯地普仑的疗效，根据已明确的 1 型脊髓性肌萎缩症自然病史，为主要终点定义了一个疗效标准。对于纳入统计层次结构的第 12 个月次要终点，疗效标准基于符合各终点的个体历史数据的 90% CI 的上限。对于BSID-III、CHOP-INTEND和HINE-2终点，采用Clopper-Pearson法计算CI，对于无永久通气的存活婴儿比例，采用互补对数-对数变换法计算CI。对于纳入统计层次结构的第24个月次要终点，其表现标准是基于明确定义的1型脊髓性肌萎缩症自然病史。预设绩效标准的详情见附录（第 21 页）。研究方案和统计分析计划预先规定使用 90% CIs 表示比例，并与假设检验中使用的单侧统计检验相匹配，用于具有预定义性能标准的终点。BSID-III、CHOP-INTEND和HINE-2终点采用精确二项检验，无事件生存期采用Z检验。

The proportion of infants who were event free at month 24 was estimated using Kaplan-Meier method- ology. For the other endpoints, infants were classified as non-responders if they were unable to achieve a response, had not maintained a response achieved earlier at the time of the assessment, had withdrawn from the study, had died, or were missing an assessment at a visit. Missing scores or instances recorded as “cannot test” for items on the BSID-III, CHOP-INTEND, and HINE-2 were assigned a score of 0.

使用 Kaplan-Meier 方法估算了第 24 个月时无事件发生的婴儿比例。对于其他终点，如果婴儿无法获得应答、在评估时未保持先前获得的应答、退出研究、死亡或在某次访问中缺失评估，则将其归类为无应答者。对于 BSID-III、CHOP-INTEND 和 HINE-2 中的评分缺失或被记录为 "无法测试 "的情况，则记为 0 分。

Because the same enrolment criteria, safety and efficacy assessments, schedule, and dosing regimen were used, exploratory post-hoc safety and efficacy (at month 12 and month 24) analyses were conducted with pooled data from part 1 high-dose cohort (n=17) and part 2 (n=41) populations. No formal hypothesis testing was performed for the pooled populations. This study is registered with ClinicalTrials.gov, NCT02913482. Recruitment is closed; the open-label extension phase of the study is ongoing.

由于采用了相同的入组标准、安全性和疗效评估、时间表和给药方案，因此对第一部分高剂量队列（人数=17）和第二部分（人数=41）人群的汇总数据进行了探索性的安全性和疗效（第12个月和第24个月）事后分析。未对汇总人群进行正式的假设检验。本研究已在 ClinicalTrials.gov 登记，编号为 NCT02913482。招募工作已经结束；该研究的开放标签扩展阶段正在进行中。

Results 成果

41 infants were enrolled in FIREFISH part 2 between March 13 and Nov 19, 2018 (figure 1). Baseline character- istics are shown in table 1.

2018 年 3 月 13 日至 11 月 19 日期间，41 名婴儿参加了 FIREFISH 第 2 部分（图 1）。基线特征见表 1。

The primary and secondary endpoints included in the statistical hierarchy and assessed at month 12 (clinical cutoff date Nov 14, 2019) were met (p<0·0001 for all endpoints; table 2), and previously reported.24,25 The first secondary endpoint included in the statistical hierarchy at month 24 (clinical cutoff date Nov 12, 2020) was met, with 18 infants (44% [90% CI 31–58]) able to sit without support for at least 30 s. This score was significantly higher than the 5% performance criterion (p<0·0001). No infants could stand alone or walk alone after 24 months of treatment; these milestones were not statistically different from the predefined performance criterion of 5%.

在第 12 个月（临床截止日期为 2019 年 11 月 14 日）进行的评估中，统计层次中包括的主要终点和次要终点均已达到（所有终点的 p<0-0001；表 2），这在之前已有报道24,25。在第 24 个月（临床截止日期为 2020 年 11 月 12 日）进行的评估中，统计层次中包括的第一个次要终点已达到，18 名婴儿（44% [90% CI 31-58]）能够在无支撑的情况下至少坐 30 秒。治疗 24 个月后，没有婴儿能独自站立或独自行走；这些里程碑在统计学上与 5%的预定绩效标准没有差异。

At month 24, improvements were reported for most endpoints included in the statistical hierarchy at month 12 (table 2). An increase was observed in the proportions of infants achieving sitting without support for at least 5 s and of infants achieving a CHOP-INTEND score of at least 40 points. The proportion of infants who achieved an increase of at least 4 points from baseline on the CHOP-INTEND was maintained at month 24. Overall, three more infants were HINE-2 motor milestone responders at month 24 versus at month 12 (table 2). Additionally, over 24 months of risdiplam treatment, the infants showed continued improvement in their mean change from baseline in CHOP-INTEND score, in the exploratory analysis (appendix p 22).

第 24 个月时，第 12 个月统计等级中的大多数终点都有所改善（表 2）。可以观察到，在没有支撑的情况下坐起至少 5 秒钟的婴儿比例和 CHOP-INTEND 评分至少达到 40 分的婴儿比例均有所提高。与基线相比，CHOP-INTEND 得分至少提高 4 分的婴儿比例在第 24 个月仍保持不变。总体而言，与第12个月相比，第24个月时有HINE-2运动里程碑反应的婴儿增加了3名（表2）。此外，在利斯地普仑治疗的 24 个月中，婴儿的 CHOP-INTEND 评分与基线相比的平均变化持续改善（附录第 22 页）。

One additional infant required permanent ventilation between month 12 and the current clinical cutoff date (figure 2). Therefore, the event-free survival at month 24 was 34 of41 infants (83% [90% CI 71–90]) versus 35 infants (85% [73–92]) at month 12 (table 2). Similarly, seven more infants achieved head control (score ≥3 on item 12 of the CHOP-INTEND, not included in the statistical hierarchy) at month 24 versus month 12 (table 2).

在第 12 个月和当前临床截止日期之间，还有一名婴儿需要永久通气（图 2）。因此，第 24 个月的无事件存活率为 41 个婴儿中的 34 个（83% [90% CI 71-90]），而第 12 个月为 35 个婴儿（85% [73-92]）（表 2）。同样，与第 12 个月相比，第 24 个月达到头部控制（CHOP-INTEND 第 12 项得分≥3，未纳入统计等级）的婴儿增加了 7 个（表 2）。

Although no infants could walk or stand independently at month 24, more infants achieved a higher motor milestone category in the HINE-2 compared with month 12. For example, for the standing milestone, more infants achieved “standing with support” (six [15%] vs two [5%] at month 12; figure 3); for the walking milestone, one infant (2%) achieved “cruising”,while no infant achieved this milestone at month 12 (figure 3). Furthermore, more infants were recorded as able to achieve the highest motor milestone category. For instance, more infants were able to “pivot (rotate)” as recorded within the sitting milestone (12 [29%] vs four [10%] at month 12), more infants were able to “roll from supine to prone” (18 [44%] vs four [10%] at month 12; figure 3); and two infants (5% vs 0 at month 12) were recorded for the crawling milestone as able to “crawl on their hands and knees” at month 24 (secondary endpoint; appendix p 23).

虽然没有婴儿在第 24 个月时能够独立行走或站立，但与第 12 个月相比，有更多婴儿在 HINE-2 中达到了更高的运动里程碑类别。例如，在站立里程碑方面，更多婴儿达到了 "有支撑物的站立"（6 名婴儿[15%] 与第 12 个月时的 2 名婴儿[5%]相比；图 3）；在行走里程碑方面，1 名婴儿（2%）达到了 "巡航"，而第 12 个月时没有婴儿达到这一里程碑（图 3）。此外，有更多的婴儿能够达到最高的运动里程碑类别。例如，在坐的里程碑中，有更多婴儿能够 "转动（旋转）"（第 12 个月时有 12 名婴儿[29%] 对 4 名婴儿[10%]），有更多婴儿能够 "从仰卧翻滚到俯卧"（第 12 个月时有 18 名婴儿[44%] 对 4 名婴儿[10%]；图 3）；在爬的里程碑中，有 2 名婴儿（第 12 个月时为 5%，第 24 个月时为 0）在第 24 个月时能够 "用手和膝盖爬行"（次要终点；附录第 23 页）。

Following 24 months of risdiplam treatment, 38 of 41 infants (93% [90% CI 82–97]) were alive. 35 infants (85% [73–93]) were able to feed orally at month 24 versus 34 infants (83% [70–92]) at month 12. Moreover, at month 24,eight infants (20% [10–32]) compared with ten infants (24% [14–38]) at month 12, did not require ventilatory support and 37 infants (90% [78–95], compared with 38 infants (92% [81–97]) at month 12, were without permanent ventilation (table 2).

经过 24 个月的利地普仑治疗后，41 名婴儿中有 38 名（93% [90% CI 82-97]）存活。第 24 个月时，35 名婴儿（85% [73-93]）能够口服喂养，而第 12 个月时，34 名婴儿（83% [70-92]）不能口服喂养。此外，第 24 个月时有 8 名婴儿（20% [10-32]）不需要通气支持，而第 12 个月时有 10 名婴儿（24% [14-38]）不需要通气支持；第 12 个月时有 37 名婴儿（90% [78-95]）不需要永久通气，而第 12 个月时有 38 名婴儿（92% [81-97]）不需要永久通气（表 2）。

The median blood SMN protein concentration at month 24 was 4·76 ng/mL (IQR 4·11–5·62), with a median 1·95-fold change (1·33–2·26) from baseline (appendix p 25). Results of the exploratory efficacy endpoints at month 24 arepresented in the appendix (p 12). Up to the 24-month clinical cutoff date, a total of 356 adverse events were reported in FIREFISH part 2 (table 3). A full list of adverse events and serious adverse events can be found in the appendix (p 16). The most frequently reported adverse event was upper respiratory tract infection, in 22 infants (54%). 28 infants (68%) experienced 68 serious adverse events; the most frequently reported serious adverse event was pneumonia in 16 infants (39%).

第 24 个月时，血液中 SMN 蛋白浓度的中位数为 4-76 纳克/毫升（IQR 4-11-5-62），与基线相比，变化中位数为 1-95 倍（1-33-2-26）（附录第 25 页）。第 24 个月的探索性疗效终点结果见附录（第 12 页）。截至 24 个月的临床截止日期，FIREFISH 第 2 部分共报告了 356 例不良事件（表 3）。不良事件和严重不良事件的完整列表见附录（第 16 页）。最常报告的不良事件是上呼吸道感染，22 名婴儿（54%）发生了上呼吸道感染。28 名婴儿（68%）发生了 68 起严重不良事件；最常见的严重不良事件是肺炎，16 名婴儿（39%）发生了肺炎。

Seven infants (17%) experienced at least one adverse event that was considered to be related to risdiplam treatment by the investigator. Treatment-related adverse events included rash maculopapular, skin discolour- ation, and constipation, each in two infants (5%), and increased aspartate aminotransferase eosinophilia, neutropenia, upper respiratory tract infection, decreased neutrophil count, and pulmonary hypertension, each in one infant (2%). No infants left the study owing to drug- related adverse events.

有 7 名婴儿（17%）出现了至少一种不良反应，研究者认为这些不良反应与利血平治疗有关。与治疗相关的不良事件包括：2 名婴儿（5%）出现皮疹、皮肤变色和便秘；1 名婴儿（2%）出现天冬氨酸氨基转移酶嗜酸性粒细胞增多、中性粒细胞减少、上呼吸道感染、中性粒细胞计数减少和肺动脉高压。没有婴儿因药物相关不良事件而退出研究。

Discussion 讨论

The primary analysis of FIREFISH part 2 found a benefit of risdiplamin infants with type 1 spinal muscular atrophy at 12 months. The first secondary endpoint in the statistical hierarchy at month 24, the proportion of infants sitting without support for at least 30 s, was markedly different from the predefined performance criterion of 5% based on natural history data. Without treatment, children with type 1 spinal muscular atrophy are never able to sit without support, and thus the ability to achieve sitting is an important motor milestone in treated type 1 spinal muscular atrophy.

FIREFISH 第 2 部分的主要分析结果表明，12 个月大的 1 型脊髓性肌萎缩症婴儿服用利迪拉明可获益。第 24 个月时，统计层次中的第一个次要终点，即在没有支撑的情况下至少坐 30 秒的婴儿比例，与根据自然病史数据预先设定的 5%的绩效标准有明显差异。在没有治疗的情况下，1 型脊髓性肌萎缩症患儿永远无法在没有支撑物的情况下坐立，因此实现坐立的能力是治疗 1 型脊髓性肌萎缩症的一个重要运动里程碑。

All infants who met the primary endpoint at month 12 continued to do so at month 24. By month 24, 13 more infants were able to sit without support for at least 5 s and 11 more infants were able to sit without support for at least 30 s. Additionally, three more infants were classified as having a motor milestone response in the HINE-2 and eight more infants achieved a CHOP-INTEND score of at least 40 points. The continuous benefit of risdiplam treatment is also reflected in the changes from baseline in the CHOP-INTEND total score (an exploratory analysis) and BSID-III gross motor subscale total score showing that motor ability continues to progress over 24 months.

所有在第 12 个月达到主要终点的婴儿在第 24 个月时仍能达到主要终点。到第 24 个月时，又有 13 名婴儿能够在没有支撑物的情况下坐立至少 5 秒钟，另有 11 名婴儿能够在没有支撑物的情况下坐立至少 30 秒钟。此外，又有 3 名婴儿在 HINE-2 中被归类为运动里程碑反应，另有 8 名婴儿的 CHOP-INTEND 得分达到至少 40 分。CHOP-INTEND总分（探索性分析）和BSID-III大运动分量表总分与基线相比的变化也反映了利斯地普仑治疗的持续益处，表明运动能力在24个月后仍在不断提高。

These findings show clinically meaningful gains in motor function and indicate a clear deviation from natural history data, where achievement of major motor mile- stones and a CHOP-INTEND score of 40 points or greater is rarely observed.

这些研究结果表明，运动功能的提高具有临床意义，与自然病史数据明显不同，在自然病史数据中，很少能观察到主要运动里程碑的实现和CHOP-INTEND评分达到或超过40分。

After 24 months of treatment, infants showed continued improvement  in  motor  function  and  in  attaining milestones, demonstrating a continuum of developmental gains from month 12. Despite this progress, no infants achieved independent standing or walking, as assessed by the BSID-III gross motor subscale. This might be related to the age at disease onset, age at treatment initiation (treatment initiated 1 day after enrolment; median age at enrolment 5·3 months, IQR 4·2–6·8), and disease severity when patients started treatment (median disease duration, defined as time from symptom onset to first dose, 3·4 months, IQR 2·5–4·9). Longer treatment might lead to achievement of some or part of these milestones; a possibility supported by the greater proportion of infants achieving higher responses in the sitting, standing, and walking categories in the HINE-2 at month 24 versus month 12.

经过 24 个月的治疗后，婴儿的运动功能和达到里程碑的能力继续得到改善，显示出从第 12 个月开始的持续发展。尽管取得了这些进步，但根据 BSID-III 大运动分量表的评估，没有婴儿实现独立站立或行走。这可能与患者的发病年龄、开始治疗的年龄（入院后 1 天开始治疗；入院年龄中位数为 5-3 个月，IQR 为 4-2-6-8）以及开始治疗时的病情严重程度（病程中位数，定义为从症状出现到首次用药的时间，3-4 个月，IQR 为 2-5-4-9）有关。较长的治疗时间可能会使患者达到部分或全部里程碑；在HINE-2中，第24个月与第12个月相比，在坐、站和行走方面达到较高反应的婴儿比例更高，这也证明了这种可能性。

Most infants maintained the ability to swallow and feed orally after 24 months of risdiplam treatment. This finding is markedly different from the results of the US Paediatric Neuromuscular Clinical Research Network natural history study, in which infants with type 1 spinal muscular atrophy typically required nutritional support or  combined  ventilatory  and  feeding  support  by 11 months of age.8  Event-free survival time was greatly improved in infants treated with risdiplam compared with natural history. In FIREFISH part 2, three infants experienced fatal respiratory complications characteristic of type 1 spinal muscular atrophy which occurred early in the study  (within the first  3 months of treatment). Between the clinical cutoff date of the primary analysis and  this  clinical  cutoff date,  there  have  been  no additional deaths, and only one additional infant required permanent ventilation between month 12 and month 24.

大多数婴儿在接受利斯地普仑治疗 24 个月后仍能保持吞咽和口服喂养的能力。这一结果与美国儿科神经肌肉临床研究网络（US Paediatric Neuromuscular Clinical Research Network）的自然病史研究结果明显不同，在自然病史研究中，1 型脊髓性肌萎缩症婴儿通常在 11 个月大时需要营养支持或呼吸与喂养联合支持。在 FIREFISH 第 2 部分中，有 3 名婴儿在研究早期（治疗的头 3 个月内）出现了 1 型脊髓性肌萎缩症特有的致命性呼吸系统并发症。从主要分析的临床截止日期到本次临床截止日期之间，没有再出现死亡病例，只有一名婴儿在第 12 个月到第 24 个月之间需要永久通气。

Adverse events reported up to the 24-month clinical cutoff date were consistent with results from the previous clinical cutoff dates for part 1 and part 2 of the study.24,25 No risdiplam-related adverse events led to withdrawal or discontinuation of treatment. The serious adverse event incidence rate of pneumonia declined in the second year of treatment.  Ophthalmological monitoring did not reveal  any  findings  suggestive  of risdiplam  effects previously observed in the preclinical animal studies. SMN protein concentrations were stable over time and were consistent with the results reported for FIREFISH part 1 and at month 12 in FIREFISH part 2.

截至 24 个月临床截止日期前报告的不良事件与该研究第一部分和第二部分临床截止日期前的结果一致24,25。肺炎这一严重不良事件的发生率在治疗的第二年有所下降。 眼科监测未发现任何提示之前在临床前动物研究中观察到的利斯地平效应的结果。随着时间的推移，SMN 蛋白浓度保持稳定，与 FIREFISH 第 1 部分和 FIREFISH 第 2 部分第 12 个月的报告结果一致。

Based on studies in pubertal and adult rats and monkeys, male sperm cell division might be arrested while on treatment, thus possibly affecting male fertility. These effects are expected to be reversible upon discon- tinuation of treatment. To date, there is no clinical evidence suggesting that risdiplam causes male fertility issues in humans.

根据对青春期大鼠、成年大鼠和猴子的研究，雄性精子细胞分裂可能会在治疗期间停止，从而可能影响雄性的生育能力。这些影响预计会在停止治疗后逆转。迄今为止，还没有临床证据表明利司腺胺会导致人类男性生育问题。

Post-hoc  analysis  of the  pooled  efficacy  results including data from part 1 of the study were consistent with the results from FIREFISH part 2 (appendix p 14), showing that, in a larger cohort of infants with type 1 spinal  muscular  atrophy, prolonged treatment with risdiplam was associated with a clinically meaningful improvement in survival, motor function, and develop- mental milestones compared with natural history.

对包括第一部分研究数据在内的汇总疗效结果进行的事后分析与 FIREFISH 第二部分研究的结果一致（附录第 14 页），结果表明，在更大的 1 型脊髓性肌萎缩症婴儿队列中，与自然病史相比，延长利斯地普仑的治疗与存活率、运动功能和发育里程碑方面有临床意义的改善相关。

Despite  the  COVID-19  pandemic,  at-home  oral treatment with risdiplam was unaffected. The effect of the pandemic on the study was small and occurred due to hospital or pandemic-imposed movement restrictions that  resulted  in  patients  missing  scheduled  study assessments. Despite this, the reported deviations did not affect the conclusions and interpretation of the safety data or cumulative study results. All infants had an on- site visit at month 24.

尽管发生了 COVID-19 大流行，但利地普仑的居家口服治疗并未受到影响。大流行对研究的影响较小，是由于医院或大流行实施的行动限制导致患者错过了预定的研究评估。尽管如此，所报告的偏差并未影响安全性数据或累积研究结果的结论和解释。所有婴儿都在第 24 个月时接受了现场访视。

There were some limitations to FIREFISH part 2, particularly the use of natural history data to derive performance criteria for achieving the clinical endpoints. Specifically, these were: differences in baseline character- istics between the natural history and FIREFISH cohorts, the relatively small number of patients in the historical cohorts, and the potential for unconscious selection bias associated with the use of historical cohorts. Despite these limitations,  it  is  noteworthy  that  the  primary  and secondary endpoints are based on objective assessments and the results are clearly differentiated from available natural history data. Treatment with risdiplam over 24 months in part 2 of the FIREFISH study resulted in continued survival (no new deaths since the primary analysis at 12 months of treatment), further improvements in motor function and achievement of motor milestones. These findings show meaningful gains in motor function beyond month 12, confirming that longer-term treatment with risdiplam benefited patients with type 1 spinal muscular atrophy.

FIREFISH 第 2 部分存在一些局限性，特别是使用自然病史数据来推导实现临床终点的性能标准。具体来说，这些限制包括：自然病史队列和 FIREFISH 队列之间基线特征的差异、历史队列中患者人数相对较少，以及使用历史队列可能存在无意识选择偏差。尽管存在这些局限性，但值得注意的是，主要终点和次要终点是基于客观评估的，其结果与现有的自然病史数据有明显区别。在FIREFISH研究的第二部分中，使用利斯地普仑治疗24个月后，患者继续存活（自治疗12个月后的主要分析以来没有新的死亡病例），运动功能进一步改善，并达到了运动里程碑。这些研究结果表明，在第12个月之后，患者的运动功能得到了显著改善，这证实了利斯地平的长期治疗对1型脊髓性肌萎缩症患者有益。

Table 2: Primary and secondary efficacy endpoints at month 12 and month 24 from FIREFISH part 2

表 2：FIREFISH 第 2 部分第 12 个月和第 24 个月的主要和次要疗效终点

除非另有说明，否则数据均为 n（% [90%CI]）或中位数（IQR）。

BiPAP=双水平气道正压。BSID-III=贝利婴幼儿发育量表第三版。CHOP-INTEND=费城儿童医院神经肌肉疾病婴儿测试。HINE-2=Hammersmith 婴儿神经系统检查第 2 部分。P 值用于比较每个终点历史数据中符合表现标准的婴儿比例。假设检验按层次进行，每个终点的单侧显著性水平为 5%，如果层次中前一个终点的 p≤0∙05 则进行假设检验。如果数据中没有显示 p 值，则表示这些结果不是第 24 个月时的预设终点；第 12 个月时属于统计层次的 CHOP-INTEND 次要终点在第 24 个月时未进行预设分析。*如果运动里程碑改善的人数多于恶化的人数，则婴儿被归类为应答者。改善的定义是：踢腿能力（或最大得分）提高≥2 分，或头部控制、翻滚、坐、爬、站或走能力提高≥1 分。恶化的定义是踢腿能力下降≥2分（或最低分），或头部控制、滚动、坐、爬行、站立或行走能力下降≥1分。

†定义为无永久通气（即连续 21 天每天气管插管或 BiPAP 时间≥16 小时，或连续 21 天插管，且无急性可逆事件或急性可逆事件已缓解）的存活婴儿。

‡p 值无显著性；在常设终点打破了等级制度。

§定义为得分≥3分，患者在躯干直立和肩部支撑的情况下坐着时保持头部直立>15秒。38名婴儿有该终点数据。||6名婴儿（15%）在第12个月时未达到六项里程碑中的任何一项，4名婴儿（10%）在第24个月时未达到六项里程碑中的任何一项。然而，由于该终点仅包括六个项目，因此并不能反映婴儿在 BSID-III 大运动分量表中达到的总体最高里程碑。

**第 12 个月时，根据 BSID-III 大运动分量表的评估，18 名婴儿（44%）能够 "控制头部直立 15 秒"（第 9 项），34 名婴儿（83%）能够 "从一侧滚到另一侧"（第 14 项）。第 24 个月时，根据 BSID-III 大运动分量表的评估，30 名婴儿（73%）能够 "控制头部直立 15 秒"（第 9 项），35 名婴儿（85%）能够 "从一侧滚到另一侧"（第 14 项）。

‡在第 24 个月达到 "手膝着地爬行 "运动里程碑的婴儿（附录第 23 页）是使用 HINE-2 量表进行评估的。§§采用 Kaplan-Meier 方法进行估计。

¶包括68%（41例中的28例）能够完全口服喂养的婴儿，以及15%（41例中的6例）在口服喂养的同时使用喂食管喂养的婴儿。||||，包括71%（41例中的29例）完全通过口腔喂养的婴儿，7%（41例中的3例）完全通过喂食管喂养的婴儿，以及15%（41例中的6例）口腔喂养与喂食管喂养相结合的婴儿。

图 3：哈默史密斯婴儿神经系统检查第 2 部分，第 12 个月和第 24 个月的运动里程碑（次要终点）

3 名婴儿在注册后的头 3 个月内死亡。对于每个运动里程碑类别，所显示的数值格式为达到运动里程碑的婴儿人数和百分比。第 12 个月的临床截止日期为 2019 年 11 月 14 日；第 24 个月的临床截止日期为 2020 年 11 月 12 日。*就 "站立 "里程碑而言，没有婴儿在第 12 个月和第 24 个月被记录为 "无助站立"；4 名婴儿在第 12 个月和 1 名婴儿在第 24 个月被记录为 "无法测试或未完成"。在行走里程碑方面，没有婴儿在第 12 个月时达到 "巡游（扶着行走）"，也没有婴儿在第 12 个月或第 24 个月时达到 "独立行走"；第 12 个月时有 37 名婴儿（90%）和第 24 个月时有 36 名婴儿（88%）被记录为 "无法测试或未完成"。有一名婴儿在第 24 个月时在 "头部控制"、"坐 "和 "翻滚 "方面被记录为 "无法测试或未完成"。