In multiple sclerosis, treatment start or switch is prompted by evidence of disease activity. Whilst immunomodulatory therapies reduce disease activity, the time required to attain maximal effect is unclear. In this study we aimed to develop a method that allows identification of the time to manifest fully and clinically the effect of multiple sclerosis treatments (‘therapeutic lag’) on clinical disease activity represented by relapses and progression-of-disability events. Data from two multiple sclerosis registries, MSBase (multinational) and OFSEP (French), were used. Patients diagnosed with multiple sclerosis, minimum 1-year exposure to treatment, minimum 3-year pretreatment follow-up and yearly review were included in the analysis. For analysis of disability progression, all events in the subsequent 5-year period were included. Density curves, representing incidence of relapses and 6-month confirmed progression events, were separately constructed for each sufficiently represented therapy. Monte Carlo simulations were performed to identify the first local minimum of the first derivative after treatment start; this point represented the point of stabilization of treatment effect, after the maximum treatment effect was observed. The method was developed in a discovery cohort (MSBase), and externally validated in a separate, non-overlapping cohort (OFSEP). A merged MSBase-OFSEP cohort was used for all subsequent analyses. Annualized relapse rates were compared in the time before treatment start and after the stabilization of treatment effect following commencement of each therapy. We identified 11 180 eligible treatment epochs for analysis of relapses and 4088 treatment epochs for disability progression. External validation was performed in four therapies, with no significant difference in the bootstrapped mean differences in therapeutic lag duration between registries. The duration of therapeutic lag for relapses was calculated for 10 therapies and ranged between 12 and 30 weeks. The duration of therapeutic lag for disability progression was calculated for seven therapies and ranged between 30 and 70 weeks. Significant differences in the pre- versus post-treatment annualized relapse rate were present for all therapies apart from intramuscular interferon beta-1a. In conclusion we have developed, and externally validated, a method to objectively quantify the duration of therapeutic lag on relapses and disability progression in different therapies in patients more than 3 years from multiple sclerosis onset. Objectively defined periods of expected therapeutic lag allows insights into the evaluation of treatment response in randomized clinical trials and may guide clinical decision-making in patients who experience early on-treatment disease activity. This method will subsequently be applied in studies that evaluate the effect of patient and disease characteristics on therapeutic lag.

中文翻译：

---

从治疗开始延迟到对多发性硬化症的免疫治疗完全生效。

在多发性硬化症中，疾病活动的证据提示治疗开始或转换。尽管免疫调节疗法降低了疾病活性，但达到最大效果所需的时间尚不清楚。在这项研究中，我们旨在开发一种方法，该方法可以确定时间，以临床表现充分且临床上证明多种硬化症治疗（“治疗滞后”）对以复发和残疾进展为代表的临床疾病活动的影响。使用了来自两个多发性硬化症登记系统MSBase（跨国公司）和OFSEP（法国）的数据。分析包括诊断为多发性硬化，至少接受1年治疗，至少接受3年治疗前随访和每年复查的患者。为了分析残疾进展，包括了随后五年的所有事件。密度曲线 对于每种具有足够代表性的疗法，分别构建代表复发率和6个月确诊进展事件的代表。进行蒙特卡洛模拟以识别治疗开始后一阶导数的一阶局部最小值。在观察到最大治疗效果之后，该点代表治疗效果的稳定点。该方法是在发现队列（MSBase）中开发的，并在单独的非重叠队列（OFSEP）中进行了外部验证。合并的MSBase-OFSEP队列用于所有后续分析。在开始治疗之前和开始每种治疗后稳定治疗效果之后，比较年度复发率。我们确定了11180个符合条件的治疗时期用于复发分析，以及4088个治疗时期用于残疾进展分析。在四种疗法中进行了外部验证，各注册管理机构之间的治疗滞后时间的自举平均差异无显着差异。计算了10种疗法的复发治疗滞后的持续时间，范围为12至30周。针对七种疗法计算了残疾进展的治疗滞后的持续时间，范围为30至70周。除肌肉内干扰素β-1a以外，所有疗法在治疗前和治疗后的年均复发率上均存在显着差异。总之，我们已经开发并进行了外部验证，一种从多发性硬化症发作开始，对超过3年的患者在不同疗法中复发和残疾进展的治疗延迟持续时间进行客观量化的方法。客观定义预期治疗滞后的时间可以洞察随机临床试验中对治疗反应的评估，并可以指导经历早期治疗疾病活动的患者的临床决策。此方法随后将用于评估患者和疾病特征对治疗延迟的影响的研究中。客观定义预期治疗滞后的时间段有助于洞察随机临床试验中对治疗反应的评估，并可以指导经历早期治疗疾病活动的患者的临床决策。此方法随后将用于评估患者和疾病特征对治疗延迟的影响的研究中。客观定义预期治疗滞后的时间段有助于洞察随机临床试验中对治疗反应的评估，并可以指导经历早期治疗疾病活动的患者的临床决策。此方法随后将用于评估患者和疾病特征对治疗延迟的影响的研究中。