In this independent, multicenter, post-marketing study, we directly compare induction immunosuppression versus escalation strategies on the risk of reaching the disability milestone of Expanded Disability Status Scale (EDSS) ≥ 6.0 over 10 years in previously untreated patients with relapsing-remitting multiple sclerosis. We collected data of patients who started interferon beta (escalation) versus mitoxantrone or cyclophosphamide (induction) as initial treatment. Main eligibility criteria included an EDSS score ≤ 4.0 at treatment start and either ≥ 2 relapses or 1 disabling relapse with evidence of ≥ 1 gadolinium-enhancing lesion at magnetic resonance imaging scan in the pre-treatment year. Since patients were not randomized to treatment group, we performed a propensity score (PS)–based matching procedure to select individuals with homogeneous baseline characteristics. Comparisons were then conducted using Cox models stratified by matched pairs. Overall, 75 and 738 patients started with induction and escalation, respectively. Patients in the induction group were older and more disabled than those in the escalation group (p < 0.05). The PS-matching procedure retained 75 patients per group. In the re-sampled population, a lower proportion of patients reached the outcome after induction (21/75, 28.0%) than escalation (29/75, 38.7%) (hazard ratio = 0.48; p = 0.024). Considering the whole sample, serious adverse events occurred more frequently after induction (8/75, 10.7%) than escalation (18/738, 2.4%) (odds ratio = 3.36, p = 0.015). These findings suggest that, in patients with poor prognostic factors, induction was more effective than escalation in reducing the risk of reaching the disability milestone, albeit with a worse safety profile. Future studies are warranted to explore if newer induction agents may provide a more advantageous long-lasting risk:benefit profile.

在这项独立、多中心、上市后研究中，我们直接比较了诱导免疫抑制与升级策略对先前未经治疗的复发缓解型多发性硬化症患者在 10 年内达到扩展残疾状态量表 (EDSS) ≥ 6.0 残疾里程碑的风险。我们收集了开始使用干扰素β（升级）与米托蒽醌或环磷酰胺（诱导）作为初始治疗的患者的数据。主要资格标准包括治疗开始时 EDSS 评分 ≤ 4.0，以及在治疗前一年中，有 ≥ 2 次复发或 1 次致残复发，并且在磁共振成像扫描中显示有 ≥ 1 个钆增强病变的证据。由于患者没有被随机分配到治疗组，因此我们进行了基于倾向评分（PS）的匹配程序，以选择具有同质基线特征的个体。然后使用按配对分层的 Cox 模型进行比较。总体而言，分别有 75 名和 738 名患者开始诱导和升级。与升级组相比，诱导组患者年龄较大且残疾程度更高 ( p < 0.05)。 PS 匹配程序每组保留 75 名患者。在重新抽样的人群中，诱导后达到结果的患者比例 (21/75, 28.0%) 低于升级后达到结果的患者比例 (29/75, 38.7%)（风险比 = 0.48； p = 0.024）。考虑到整个样本，诱导后严重不良事件的发生频率 (8/75, 10.7%) 比升级后发生的频率更高 (18/738, 2.4%)（比值比 = 3.36， p = 0.015）。 这些研究结果表明，对于预后因素较差的患者，诱导治疗比升级治疗更能有效降低达到残疾里程碑的风险，尽管安全性较差。未来的研究有必要探索更新的诱导剂是否可以提供更有利的长期风险：效益概况。