Использование подкожного дородового ухода - окончательные результаты невмешательского, наблюдательного исследования (SISTER) в Германии и Австрии.

Introduction

Natalizumab (Tysabri ® , Biogen) is a recombinant monoclonal humanized anti-α4 integrin antibody used as single disease-modifying therapy (DMT) for the treatment of patients with highly active relapsing-remitting multiple sclerosis (RRMS). 1 , 2 Natalizumab – at a dose of 300 mg every 4 weeks (Q4W) – can be administered either by intravenous (IV) infusion or subcutaneous (SC) injection. Intravenous natalizumab, licensed in 2006, was extensively investigated in both pivotal randomized clinical trials (RCTs) versus placebo 3 – 5 and in open-label, observational studies. 6 – 8 In addition, The SC route of administration (RoA) was marketed in March 2021 based on 2 RCTs that demonstrated the clinical equivalence of the SC formulation compared with the IV formulation. 9 , 10 Until 31 st July 2023, natalizumab (IV or SC) has been administered to 264,276 patients with a total of 1,117,808 patient-years of exposure, and SC natalizumab has been administered to 21,406 patients (corresponding to 24,659 patient-years).

The SC application appears to be an attractive alternative compared with the IV route, avoiding venous access and minimizing procedure time, with – according to the SmPC – the IV infusion being administered over about 1 hour. 1 The importance of a convenient drug application and the consideration of patients’ preference for an IV or SC RoA to promote treatment adherence, and subsequently the long-term treatment success, have been described previously. 11 Likewise, study objectives, design, methods, allocation to treatment cohorts, and interim results of the (then ongoing) observational SISTER study have already been published by data cut-off date 14-Feb-2023 based on 262 study patients. 11 Briefly, SISTER primarily aimed to generate real-world data concerning the RoA preference among patients who were treated de novo with natalizumab, or who were switched from continuing natalizumab IV treatment to the SC route. In addition, demographic and other baseline characteristics, effectiveness outcomes, and safety outcomes over up to 1 year of treatment were documented. This paper, following the STROBE cohort reporting guidelines, 12 summarizes the final study results seen in 310 patients with an additional focus on the adverse event (AE) experience on either IV or SC natalizumab treatment.

Design and methods

Study design, setting, and regulatory requirements

SISTER (Subcutaneous: Non-Interventional Study for Tysabri Patient Preference – Experience from Real World) was a prospective, multicenter, non-controlled, observational cohort study conducted in Germany and Austria to evaluate the RoA preference, utilization, safety, effectiveness, and patient-reported outcomes over 1 year on treatment with SC or IV natalizumab among RRMS patients treated in daily routine. The study was conducted in line with applicable laws and regulations, approved by an independent Ethics Committee and by the national Competent Authority (Paul-Ehrlich Institute; PEI), and registered ( NCT05304520 ; PEI-No. 611). All analyzed patients had provided written informed consent to participate in the study.

Patients, treatment, and criteria for evaluation

Details about the study design, methods, as well as preliminary interim results have been published elsewhere. 11 Briefly, adult patients with highly active RRMS according to the revised McDonald criteria 13 and eligible for treatment with SC or IV natalizumab according to the local label were enrolled. Patients were allocated to 1 of 3 pre-planned treatment cohorts, i.e., i) patients switching from previous treatment with IV natalizumab to SC natalizumab (referred to as SC switcher cohort), ii) patients starting with IV natalizumab de novo (IV starter cohort), or iii) patients starting SC natalizumab de novo (SC starter cohort; see Figure 1 ). A maximum of 3 SC natalizumab administrations were allowed prior to study inclusion (i.e., patients could have been treated with 1-4 administrations at the time of the baseline assessments). Patients who prematurely discontinued natalizumab treatment before completing the final scheduled study visit at Month 12 were excluded from further study participation and were therefore no longer followed up. Figure 1. Study design. RoA=IV=Intravenous; SC=Subcutaneous. *: Originally planned sample sizes.

The patients’ preference for the IV or SC RoA of natalizumab was the main study objective and was evaluated using the 2 study-specific patient preference questionnaires PPQ1 and PPQ2 (for detailed wording see Table 1 in the previous publication 11 ). In addition, the RoA preference of treating physicians was evaluated and data concerning disease activity (MS relapse rate, time to first relapse, disability improvement and progression according to Expanded Disability Status Scale [EDSS], magnetic resonance imaging [MRI], if available), patients’ treatment satisfaction (assessed using the TSQM-II 14 ), immunogenicity, safety and tolerability of treatment (occurrence of injection site reactions [ISRs] and other treatment-emergent AEs), and utilization of natalizumab (e.g., dosing intervals, duration of application, proportions of patients with treatment discontinuation, and switch of RoA during the study) were collected. Study data were documented at baseline and at 3, 6, 9, and 12 months post-baseline. Table 1. Demographic data and other baseline characteristics of the SISTER study population. Variable SC switcher (N=220) IV starter (N=44) SC starter (N=46) Total (N=310) Sex; n (%) Female Male 173 (78.6) 47 (21.4) 32 (72.7) 12 (27.3) 39 (84.8) 7 (15.2) 244 (78.7) 66 (21.3) Age (yrs) Mean ± SD Median Range 40.9 ± 9.5 40.0 20.0-70.0 33.8 ± 11.1 31.0 19.0-58.0 34.8 ± 9.8 34.5 18.0-58.0 39.0 ± 10.2 38.0 18.0-70.0 BW (kg) Mean ± SD Median Range 74.4 ± 17.5 71.0 42.0-155.0 77.7 ± 18.5 75.0 43.0-120.0 74.7 ± 19.8 72.0 50.0-140.0 74.9 ± 17.9 71.0 42.0-155.0 BMI (kg/m 2 ) Mean ± SD Median Range 25.3 ± 5.0 24.2 17.7-40.9 26.3 ± 5.2 25.8 16.8-35.9 26.1 ± 6.2 25.1 18.0-41.4 25.6 ± 5.2 24.7 16.8-41.4 EDSS (score points) Mean ± SD Median Range 2.3 ± 1.5 2.0 0.0-6.5 2.4 ± 1.4 2.5 0.0-6.0 2.2 ± 1.1 2.0 0.0-4.5 2.3 ± 1.4 2.0 0.0-6.5 Time since first symptoms (yrs.) a Mean ± SD Median 12.6 ± 7.5 11.8 5.7 ± 5.7 4.3 8.2 ± 7.1 6.9 11.0 ± 7.7 9.7 Time since diagnosis (yrs.) a Mean ± SD Median 11.4 ± 7.1 10.4 [43] 5.0 ± 5.4 2.9 [45] 6.3 ± 6.4 4.3 [306] 9.7 ± 7.2 8.7 Latency of diagnosis (yrs.) b Mean ± SD Median 1.3 ± 2.4 0.2 [42] 0.6 ± 1.3 0.1 [41] 1.8 ± 4.4 0.2 [294] 1.2 ± 2.7 0.2 No. of patients with relapses within the past 12 months; n (%) a 0 1 2 >2 unknown 188 (87.9) 22 (10.3) 3 (1.4) 1 (0.5) 6 5 (12.2) 21 (51.2) 9 (22.0) 6 (14.6) 3 9 (20.9) 18 (41.9) 14 (32.6) 2 (4.7) 3 202 (67.8) 61 (20.5) 26 (8.7) 9 (3.0) 12 No. of relapses within the past 12 months; n (%) a Mean ± SD Median Range 0.1 ± 0.5 0.0 (0.0-4.0) 1.6 ± 1.4 1.0 (0.0-8.0) 1.2 ± 0.9 1.0 (0.0-4.0) 0.5 ± 0.9 0.0 (0.0-8.0) Patients with Gd-enhancing lesions; n (%) No Yes Missing 69 (82.1) 15 (17.9) 136 8 (42.1) 11 (57.9) 25 15 (53.6) 13 (46.4) 18 92 (70.2) 39 (29.8) 179 Anti-JCV antibody status; n (%) Negative Positive Missing 185 (86.0) 30 (14.0) 5 36 (81.8) 8 (18.2) 0 41 (93.2) 3 (6.8) 2 262 (86.5) 41 (13.5) 7 BMI=Body mass index; BW=Body weight; Gd=Gadolinium; SD=Standard deviation; wks.=Weeks; yrs.=Years. Note. The numbers of patients with missing values are not included in this table. Percentages are adjusted (i.e., based on patients with available data). a Relative to time point of informed consent. b Time elapsed between the occurrence of first symptoms and first diagnosis.

Statistical analysis

Generally, the statistical methods were the same as applied for the interim analysis. 11 However, only one analysis set was defined for the final analyses, where all 310 treated and otherwise valid patients were observed within their originally designated treatment cohort until (if applicable) a potential first switch to the alternative RoA. In those cases (emerging in 19 patients overall), all study data documented after the first switch were excluded and not used for further cohort analyses in order to maintain pure and undistorted cohorts with either SC or IV natalizumab administration. Nevertheless, AEs occurring after RoA switch were separately analyzed.

Still, no confirmatory tests were performed, and descriptive p-values and 95%-confidence intervals (CIs) were considered exploratory. Missing patient data were not imputed. Continuous variables were analyzed using appropriate descriptive measures (e.g., mean, standard deviation [SD], median, 1 st and 3 rd quartiles, minimum, and maximum), while categorical variables were presented in tabulated summaries showing missing values and absolute and relative frequencies. If not otherwise specified, percentages were calculated as adjusted percentages (i.e., without consideration of missing values). However, the relative frequencies of AEs were calculated relating to the total population. For selected frequencies, 2-sided exact 95%-CIs according to Clopper-Pearson were displayed. Patient preference rates were compared using Fisher’s exact test (for dichotomous variables) or Chi-Square test (for categorical variables with >2 categories). A Poisson regression model adjusted for duration of disease and EDSS baseline score was used for the calculation of the annual relapse rate (ARR). In addition, a box plot was provided, and non-parametric post hoc tests were performed for the global satisfaction score of the TSQM-II to assess the descriptive changes within cohorts (Friedman test) and differences across cohorts (Kruskal-Wallis test).

Results

Patient disposition and duration of observation

The first patient’s first visit was on 29-Sep-2021 (study start: 01-Sep-2021), and the last patient’s last visit for the final analysis was on 21-May-2024 (study end: 30-Jun-2024). A total of 37 study sites (36 sites in Germany and 1 site in Austria) had enrolled at least 1 patient (range per site: 1 to 45 patients). Overall, 315 patients were enrolled, 5 patients were excluded due to major protocol deviations (in particular 4 patients without natalizumab administration after informed consent), and 310 patients were eligible for the final analyses. Of these, 220 patients were allocated to the SC switcher cohort, 44 to the IV starter cohort, and 46 to the SC starter cohort (see Figure 2 ). The proportion of patients with premature study discontinuation was higher in the IV starter cohort (22.7%) than in the SC starter cohort (17.4%) and in the SC switcher group (7.7%), as was the proportion of patients with premature treatment discontinuation (IV starter: 18.2%, SC starter: 8.7%, SC switcher: 3.6%). “Safety aspects” (52.4%) and “effectiveness aspects” (23.5%) were the 2 most common reasons in 20/310 study patients (6.5%) with treatment discontinuation. Figure 2. Patient disposition. AS=Analysis set; IC=Informed consent; IV=Intravenous; SC=Subcutaneous. Note. Data are based on information provided on the study end page.

The mean observation period of study patients was 10.8 ± 3.1 months (median: 11.9 months; range: 0.0–13.9 months). Separated by cohorts, the mean study duration was shorter in the IV starter cohort (8.3 ± 4.4 months) than in the SC switcher cohort (11.2 ± 2.7 months) and in the SC starter cohort (11.2 ± 2.4 months). Observations at Month 12 were available for 86.8% of patients (n=191) in the SC switcher group, 47.7% (n=21) in the IV starter group, and 84.8% (n=39) in the SC starter group.

Demographic and other baseline characteristics

The demographic data of the 310 patients at baseline represent a typical RRMS population (see Table 1 ). 78.7% of study patients were female, and the mean age of the study population at baseline was 39.0 ± 10.2 years, ranging from 18 to 70 years (median: 38.0 years). Consistent with the longer disease duration, the switcher cohort was older than the 2 starter cohorts (median age 40.0 years vs. 31.0 and 34.5 years, respectively).

The most common initial symptoms of RRMS (with a prevalence of ≥10%) were sensory symptoms (57.0%), visual disturbances (31.4%), and pyramidal symptoms (18.8%). The mean time between the occurrence of first MS symptoms and study entry was 11.0 ± 7.7 years (median: 9.7 years), and the mean time since the first RRMS diagnosis was 9.7 ± 7.2 years (median: 8.7 years). These periods were clearly longer in the SC switcher cohort than in the 2 starter cohorts (see Table 1 ).

The mean EDSS at baseline was comparable among the cohorts (2.3 ± 1.5 in SC switchers, 2.4 ± 1.3 in IV starters, and 2.2 ± 1.1 in SC starters). Most study patients (67.8%, n=202) were relapse-free in the year before enrollment. The SC switcher cohort showed a higher proportion of relapse-free patients in the past year (87.9% vs. 12.2% in the IV starter cohort and 20.9% in the SC starter cohort) as well as a lower mean number of relapses in the past year (0.1 ± 0.5 vs. 1.6 ± 1.4 in IV starters and 1.2 ± 0.9 in SC starters, respectively). The proportion of patients with gadolinium (Gd)-enhancing lesions on T1w scan at baseline was smaller in the SC switcher cohort than in the 2 starter cohorts (17.9% vs. 57.9% IV starters and 46.4% SC starters, respectively; see Table 1 ).

The proportion of study patients that were negative for anti-JCV antibodies at baseline was 86.5% (n=262), 30 of the 41 positive patients had an anti-JCV-antibody-index ≤0.9; 4 patients >0.9 to ≤1.5, and 7 patients >1.5. 15

Prior MS treatments

Prior MS treatment at any time was reported for 81.9% of patients (the mean number of treatments given before study entry was 1.6 ± 1.3 treatments (median: 1.0; range: 0–7); 18.1% had not had any prior treatment (see Figure 3 ). The most recent treatments prior to study entry included the following DMTs (if administered to ≥5.0% of patients): glatiramer acetate 15.8% (n=49), fingolimod: 14.8% (n=46), dimethyl fumarate: 13.9% (n=41), teriflunomide: 7.7% (n=24), interferon beta-1a SC): 7.7% (n=22), and interferon-beta-1a intramuscular: 5.2% (n=16) ( Figure 4 ). Figure 3. Any MS pre-treatments administered before start of natalizumab. Note. Percentage base is N=310. Listed are all medications that were administered to >1.0% of patients (>3 patients) at any time before study entry. Figure 4. Mean natalizumab administration time at baseline and Month 12. IV=Intravenous; N=Number of observations at baseline/Month 12; SC=Subcutaneous. Note. Provided is the arithmetic mean. Horizontal error bars indicate the standard deviation.

Utilization of natalizumab

The mean time on natalizumab treatment before study start in the SC switcher cohort was 5.3 ± 4.1 years (median: 4.5 years), and the mean number of infusions administered in the past 12 months was 8.6 ± 2.6 infusions (median: 9.0 infusions).

The mean duration of natalizumab exposure in the total study population was 12.5 ± 7.8 months (median: 13.8 months; range: 0.0–121.8 months). Consistent with the composition of the cohorts (i.e., patients continuing natalizumab treatment in the SC switcher cohort; a higher discontinuation rate and proportion of patients changing the initial RoA in the IV starter cohort compared with the 2 other cohorts), the mean exposure time to natalizumab was longest in the SC switcher cohort (13.1 ± 8.6 months) and shortest in the IV starter cohort (9.7 ± 5.0 months; SC starter cohort: 11.9 ± 4.2 months).

Single doses of 150 mg were given on 10 occasions, while the remaining doses were administered at the full dose of 300 mg per administration. A total mean number of 11.5 ± 3.9 natalizumab injections/infusions were administered to the 310 study patients (median: 12.0; range: 1–28). The mean number of natalizumab administrations was highest in the SC starter cohort (12.2 ± 4.2 injections) and, consistent with the higher treatment discontinuation rate, lowest in the IV starter cohort (10.3 ± 4.8 infusions). A total of 11.6 ± 3.6 administrations were applied in the SC switcher cohort during the study.

In general, almost all IV/SC starters were scheduled to receive natalizumab at Q4W standard intervals, while the proportion of patients treated with extended intervals >4 weeks was higher in the SC switcher cohort (see Table 2 ). Of note, the proportion of patients treated with extended treatment intervals within the 3 months prior to the last assessment time point (Month 12) had increased in the IV starter group (23.8% Q5W and 28.6% 6QW) and in the SC starter group (7.7% Q5W, 12.8% 6QW, and 2.6% each 7QW and 8QW, respectively). Table 2. Planned treatment intervals at baseline and actual treatment intervals within the period of 3 months prior to the last study evaluation at Month 12. Variable SC switcher (N=220) IV starter (N=44) SC starter (N=46) Total (N=310) Planned treatment intervals at baseline Every 4 wks. 111 (50.5) 44 (100.0) 42 (91.3) 197 (63.5) Every 5 wks. 36 (16.4) 0 1 (2.2) 37 (11.9) Every 6 wks. 72 (32.7) 0 3 (6.5) 75 (24.2) Every 7 wks. 0 0 0 0 Every 8 wks. 1 (0.5) 0 0 1 (0.3) Treatment intervals prior to Month 12 Every 4 wks. 75 (39.3) 10 (47.6) 29 (74.4) 114 (45.4) Every 5 wks. 34 (17.8) 5 (23.8) 3 (7.7) 42 (16.7) Every 6 wks. 79 (41.4) 6 (28.6) 5 (12.8) 90 (35.9) Every 7 wks. 1 (0.5) 0 1 (2.6) 2 (0.8) Every 8 wks. 2 (1.0) 0 1 (2.6) 3 (1.2) Note. Number of patients with missing values are not included in this table. Percentages are adjusted (i.e., based on patients with available data).

The analysis of the natalizumab RoA showed that the procedure time comprising preparation, administration, and safety follow-up at baseline was significantly shorter in the 2 SC cohorts (median of 1.0 hour each) compared with the IV starter cohort (median of 3.0 hours).

Analyses of RoA preferences and changes during the study

Initial patient preference at baseline

Analysis of patients’ initial preference at baseline using PPQ2 showed that the majority of patients in either cohort preferred their current RoA. However, there was a nominal, statistically significant difference with a higher preference rate in SC starters than in IV starters (95.1% [39/41 patients] vs. 77.5% [31/40 patients] IV starters; p=0.045; see Figure 5 ). No particular preference was expressed by 12.5% (5/40 patients) IV starters and 4.9% (2/41 patients) SC starters; 10.0% (4/40 patients) in the IV starter cohort were not treated to their preference, since they had preferred the alternative RoA. Figure 5. Patient preference at baseline as measured with PPQ2. IV=Intravenous; N=Number of observations; SC=Subcutaneous. Note. The original question was: All things considered, which route of administration do you prefer? (Response options: IV/SC/None). Depicted are the proportions of patients who preferred the actual route of administration (i.e., the IV route among IV starters or the SC route among SC starters). Vertical error bars indicate the 95%-CIs for the respective preference rate. *: Result from Chi-square test.

The most common reasons for the preference in the 2 starter cohorts were a more convenient administration among IV starters (47.7%) and the aspect of minimizing the procedure time among SC starters (56.5%). A “very strong” preference for their chosen treatment was reported by 21.9% IV starters and 26.3% SC starters (see Table 3 ). Table 3. Strength of preference and main reason for preference as measured with PPQ2. Variable IV starter (N=44) n (%) SC starter (N=46) n (%) Strength of preference a Very strong 7 (21.9) 10 (26.3) Fairly strong 16 (50.0) 20 (52.6) Not very strong 9 (28.1) 8 (21.1) Missing data 12 8 Main reason for preference b Feels less emotionally distressing 4 (9.1) 6 (13.0) Requires less time in the clinic 10 (22.7) 26 (56.5) Lower level of administration-site pain 13 (29.5) 6 (13.0) Feels more comfortable during administration 21 (47.7) 21 (45.7) Other reasons 8 (18.2) 2 (4.3) IV=Intravenous; SC=Subcutaneous. Note. Multiple specifications per patient were possible. a Original question: If you had preference for one of the administration routes, how strong is this preference? b Original question: If you have a preference, what are the two main reasons for your preference?

Changes in patient preference over time

In contrast to PPQ2, PPQ1 also allowed the assessment of preference changes over time in terms of treatment satisfaction and the wish to change the RoA at a given time point. While the satisfaction with the current RoA remained high at Months 6 and 12 in all 3 cohorts (see Figure 6 (A)), the proportion of patients who confirmed their initially chosen RoA (i.e., no wish to change) was numerically lower in the IV starter cohort than in the 2 SC cohorts at both baseline and Month 6 (see Figure 6 (B)). Figure 6. Patient preference at baseline and Months 6 and 12. IV=Intravenous; RoA=Route of administration; SC=Subcutaneous. Note. Patients with missing data at a given observation time point are excluded from the analysis. Numbers of patients with available observations (denominator N) are shown for each observation time point. Asterisks (*) denote the exploratory p-values from Fisher’s exact test for differences across cohorts at a given observation time point. (a) Original question: Are you satisfied with the route of administration of natalizumab? (Response options: yes/no). Shown is the percentage of patients responding with “yes”. (b) Original question: If you could choose again between SC and IV administration of natalizumab, which RoA would you choose? (Response options: IV infusion/SC injection) Shown is the percentage of patients confirming their initially chosen RoA at study baseline.

An additional sensitivity analysis, in which all patients were considered (i.e., including the 19 patients who had switched their initial RoA) yielded almost similar results, indicating that the original analysis was not biased by the exclusion of these patients after their switch.

Preference assessments of health care professionals (HCPs) at natalizumab initiation

Among the 90 natalizumab starters (IV or SC), most of the HCPs considered the 2 RoAs as equal (in 93.3% of patients, n=84). The IV route was mainly preferred by the HCPs (in 81.1% [36/44 patients] vs. 18.2% [8/44 patients]), while the SC route was more of a patients’ preference (in 71.7% [33/46 patients] vs. 28.3% [13/46 of patients] who were treated as decided by the HCP).

The most common reasons for HCPs’ preference of the SC RoA (in 13 patients) were “administration less complicated/less time consuming than IV” (46.2%, n=6) and “efficacy and safety considered comparable” (30.8%, n=4). One of the most common reasons for HCPs’ preference of the IV RoA (in 36 patients) was “more experience with administration route” (41.7%, n=15).

A total of 13 sites had already experience with SC natalizumab, 12 with IV natalizumab and 11 with the administration of other IV and SC drugs. Twelve of the 13 sites (92.3%) stated that they would prefer the SC route over the IV route with regard to both preparation and administration. By Month 24 after site activation (data from 21 sites available), 85.7% of sites (n=11) still preferred the preparation and 81.0% (n=17) the administration of SC natalizumab.

Number of patients with change of initial RoA

A total of 19 study patients (6.1%) changed their originally assigned RoA during the study (9 patients [4.1%] in the SC switcher cohort, 10 patients [22.7%] in the IV starter cohort, and none in the SC starter cohort). Apart from 1 patient in the SC switcher cohort, the RoA changes were always driven by patients’ decision. In 4 patients, the decision of HCPs was also considered, and “injection fatigue” was additionally reported in 3 patients.

Global treatment satisfaction measured with TSQM-II

Mean TSQM-II global satisfaction scores at baseline were 83.7 ± 23.6 in the SC switcher cohort (N=203), 68.2 ± 25.0 in the IV starter cohort (N=33), and 70.5 ± 19.9 in the SC starter cohort (N=35). Thus, the treatment satisfaction at baseline was slightly higher in the switcher cohort than in the 2 starter cohorts. Post-baseline, global satisfaction remained stable in the SC switcher cohort and IV starter cohort (Month 12: 83.2 ± 25.0 and 72.6 ± 31.4, respectively) and increased slightly in the SC starter cohort (Month 12: 76.4 ± 21.1; see box plot in Figure 7 ). However, the post hoc tests performed tests for the descriptive changes from baseline indicated that the mean improvement from baseline in the SC starter cohort (from 70.5 to 80.6 at Month 3 and 76.6 at Month 12) was nominally significant (p=0.0429; Friedman test). Moreover, the Kruskal-Wallis test for between-cohort differences in the mean changes from baseline indicated a nominally significant difference at Month 6 (median change: 0.0 in SC switcher, 0.0 in IV starter, 8.3 in SC starter; p=0.015), but not at Month 12 (median change: 0.0 in SC switcher, 8.3 in IV starter, 12.5 in SC starter; p=0.060). No obvious post-baseline changes were observed in TSQM-II domains “effectiveness”, “side effects”, and “convenience”. Figure 7. TSQM-II global satisfaction score by visits. Note. The lower and upper horizontal box boundaries indicate the lower (Q1) and upper (Q3) quartiles, respectively, the line within the boxes shows the median, the symbol elements within the boxes show the arithmetic mean, and the lower and upper whiskers visualize the range of non-outlier observations, defined as Q1 minus 1.5 x interquartile range (IQR) and Q3 plus 1.5 x IQR, respectively (Tukey fences). Outliers are shown outside the whiskers using various symbol elements.

Effectiveness outcomes

A total of 28 clinical MS relapses were documented in the study (13 in the SC switcher cohort [13 patients, 5.9%], 4 in the IV starter cohort [4 patients, 9.1%], and 11 in the SC starter cohort [8 patients, 17.4]). None of the relapses were treated with corticosteroids, and 4 relapses required hospitalization. The corresponding ARRs were 0.05575 events/year (95%-CI: [0.03083; 0.1008]) among SC switchers, 0.1192 events/year (95%-CI: [0.04220; 0.3367]) among IV starters, and 0.2463 events/year (95%-CI: [0.1318; 0.4603]) among SC starters.

The mean post-baseline EDSS values remained almost identical through Month 12 in all cohorts. Consistently, most patients with available data showed “stable disease” up to Month 12 based on the individual changes on the EDSS over time (see Figure 8 ). By trend, individual changes in EDSS (both improvement and progression) were observed more often in the 2 starter cohorts than in the SC switcher cohort (where patients had already been on IV treatment with natalizumab prior to study entry). Figure 8. Disability improvement and progression rates over time (EDSS). IV=Intravenous; SC=Subcutaneous.

Adverse events and other safety topics

The incidence of any AEs was 60.6% in the total population, where the smallest incidence was observed in the SC switcher cohort (57.3%) and the highest incidence in the SC starter cohort (76.1%; IV starter cohort: 61.4%; see Table 4 ). The frequency of SAEs was low and similar in the 3 cohorts (5.2% [n=16] overall; in addition, 1 patient after initial RoA switch with serious “MS pseudo relapse” that was unrelated to natalizumab and resolved). Any drug-related AEs (as per investigator’s assessment) were seen in 12.3% of patients in the SC switcher cohort, 22.7% in the IV starter cohort, and 19.6% in the SC starter cohort (overall: 14.8%), and related SAEs (i.e., serious adverse drug reactions, SADRs) were seen in only 0.6% (n=2) of patients in the SC switcher cohort (GI infection and carcinoma in situ , respectively). No deaths were reported during the study. Table 4. Overview of most common adverse events and all SAEs occurring during the study. MedDRA v27.0 preferred term SC switcher (N=220) IV starter (N=44) SC starter (N=46) Total (N=310) Patients with at least 1 AE a 126 (57.3) 27 (61.4) 35 (76.1) 188 (60.6) COVID-19 41 (18.6) 7 (15.9) 8 (17.4) 56 (18.1) Nasopharyngitis 30 (13.6) 6 (13.6) 9 (19.6) 45 (14.5) Multiple sclerosis relapse 13 (5.9) 3 (6.8) 8 (17.4) 24 (7.7) Corona virus infection 9 (4.1) 2 (4.5) 4 (8.7) 15 (4.8) Headache 3 (1.4) 3 (6.8) 2 (4.3) 8 (2.6) JC polyomavirus test positive 5 (2.3) 1 (2.3) 1 (2.2) 7 (2.3) Depression 3 (1.4) 1 (2.3) 1 (2.2) 5 (1.6) Hypertension 5 (2.3) 0 0 5 (1.6) Nausea 2 (0.9) 2 (4.5) 1 (2.2) 5 (1.6) Dizziness 4 (1.8) 0 0 4 (1.3) Fatigue 2 (0.9) 2 (4.5) 0 4 (1.3) Neutralising antibodies positive 0 1 (2.3) 3 (6.5) 4 (1.3) Urinary tract infection 1 (0.5) 1 (2.3) 2 (4.3) 4 (1.3) Patients with at least 1 SAE b 10 (4.5) 3 (6.8) 3 (6.5) 16 (5.2) Multiple sclerosis relapse 0 0 2 (4.3) 2 (0.6) Abortion induced 1 (0.5) 0 0 1 (0.3) Alcoholism 1 (0.5) 0 0 1 (0.3) Atrial fibrillation 1 (0.5) 0 0 1 (0.3) Bipolar I disorder 1 (0.5) 0 0 1 (0.3) Calculus urinary 0 0 1 (2.2) 1 (0.3) Carcinoma in situ c 1 (0.5) 0 0 1 (0.3) Cholelithiasis 0 1 (2.3) 0 1 (0.3) Chronic tonsillitis 0 1 (2.3) 0 1 (0.3) COVID-19 1 (0.5) 0 0 1 (0.3) Diabetes mellitus 0 0 1 (2.2) 1 (0.3) Dizziness 1 (0.5) 0 0 1 (0.3) Epilepsy 0 1 (2.3) 0 1 (0.3) Gastrointestinal infection c 1 (0.5) 0 0 1 (0.3) Hospitalisation 1 (0.5) 0 0 1 (0.3) Joint injury 1 (0.5) 0 0 1 (0.3) Liver injury 0 0 1 (2.2) 1 (0.3) Myelitis 1 (0.5) 0 0 1 (0.3) Oesophageal candidiasis 0 0 1 (2.2) 1 (0.3) Ovarian cyst 0 0 1 (2.2) 1 (0.3) Renal colic 0 0 1 (2.2) 1 (0.3) Note. Data are based on the number of patients with event. a Incidence cut for the subsequently listed AEs: >1.0% (i.e., AEs are listed, if at least at least 4 patients in the total population were involved at the PT event level). b All reported SAEs are listed. No fatal adverse events occurred during the study. c These 2 SAEs (in 2 patients) were regarded as drug-related by the reporting investigator. Both events had a severe intensity; outcomes were “unknown” for GI infection and “recovered/resolved” for carcinoma in situ .

Local ISRs (as assessed per PPQ1) were reported in 25/55 SC-treated patients with available documentation (45.5%, or 9.4% relating to all 266 SC-treated patients). The most common ISRs were injection site pain (24 events in 19 patients), followed by redness (17 events in 16 patients), pruritus (8 events in 8 patients), and inflammation (2 events in 2 patients). The majority of the 51 local ISRs were mild (23 events) or moderate (18 events); the remaining 10 events had a “strong” severity (see Table 5 ) as assessed by patient and confirmed by HCP. No “very strong” or even “intolerable” ISRs were documented. Table 5. Analysis of injection site reactions among the SC-treated patients. Type of local reaction n (%) patients a N events Severity of events, n b 1 2 3 4 5 Pain 19 (7.1) 24 11 11 2 0 0 Redness 16 (6.0) 17 10 3 4 0 0 Pruritus 7 (2.6) 8 2 4 2 0 0 Inflammation 2 (0.8) 2 0 0 2 0 0 Note. Original question (PPQ1, Question 3): Have you experienced reactions at the injection site? If yes: mark and indicate intensity. a Patients in the SC switcher cohort and SC starter cohort combined (denominator: N=266). b Given is the severity by number of events using the pre-specified categories 1=mild, 2=moderate, 3=strong, 4=very strong, and 5=intolerable.

One of the 170 patients with available laboratory data (in the SC switcher cohort) was ADA-positive at baseline (after 1-4 natalizumab administrations), and 98.5% (130/132 patients) with available data remained ADA-negative during the study (2 patients in the SC starter cohort developed treatment-emergent ADAs). As per (independent) AE documentation, 1 patient in the IV starter group and 3 patients in the SC starter group had developed neutralizing ADAs (see Table 4 ).

No cases of progressive multifocal leukoencephalopathy (PML) occurred during the study.

Discussion

The SISTER study was performed to evaluate the treatment preferences and the clinical course of patients with RRMS treated with IV or SC administered natalizumab for up to 12 months in a real-world setting. The originally planned sample size could not be achieved in a reasonable time frame, because the pool of de novo patients eligible for natalizumab treatment was limited and the recruitment rate into these 2 arms thus inappropriately slow. For the final study analysis, a total of 310 patients with a mean observation period of 10.8 ± 3.1 months were available. The treatment discontinuation rate (overall 6.5% patients with treatment discontinuation, 11.3% study withdrawals) was considered low in all cohorts, thereby indicating a high treatment persistence (93.5%).

Results of preference analyses

Consistent with the findings of the previous interim analysis, 11 the final study data confirm that patients treated de novo with either SC or IV natalizumab had a high preference for their initially assigned RoA. This preference was more pronounced in the SC starter cohort compared with the IV starter cohort (95.1% patient preference for SC vs. 77.5% for IV). Among the SC starters, time-saving aspects and administration convenience were the most frequently specified reasons for the SC preference, while administration convenience (and lower application-site pain level during administration) were the most common reasons reported among the IV starters.

However, following patient preferences over time indicated that the initial satisfaction with the current RoA was maintained from baseline through Month 12 in all 3 treatment cohorts (always >90% of patients were satisfied), whereas the (hypothetical) wish for a change of the current RoA at baseline and Month 6 was more frequently expressed by patients in the IV starter group than in the 2 SC cohorts. At Month 12, none of the patients in either the SC or IV cohorts expressed a desire to switch their RoA, but this finding should be interpreted with caution due to the small sample sizes and a potential attrition/censoring bias (particularly within the IV starter cohort (n=14 at Month 12), where many patients had transitioned from IV to SC prior to Month 12 and thus were censored from the Month 12 analyses). In fact, 22.7% of patients (n=10) in the IV starter cohort had changed their initial IV RoA to the SC route, where these changes were solely triggered by patient’s wish. No SC starters had changed the RoA until the end of the study, while 4.1% (n=9) in the SC switcher cohort had switched back to IV during the study period. Overall, there was a trend towards a change from the IV to the SC administration post-baseline among those patients who were initially treated with the IV route. This patient-reported trend in favor of the SC route was also reflected in the statements of the HCPs, where >90% of the documenting sites (at the time of site activation) and >80% of the documenting sites (24 months after site activation) reported that they would prefer the SC route over the IV route in terms of both preparation and administration of natalizumab.

The important issue of patient preference for a specific natalizumab RoA was also addressed in the open-label extension Part 2 of the NOVA RCT ( NCT 03689972 ). 16 , 17 The primary endpoint was the proportion of patients who indicated a preference for natalizumab SC administration vs. IV administration on the PPQ at the end of the crossover period. At that time, 87.8% of patients preferred the SC route of natalizumab over the IV route, and this high patient preference for the SC route – as well as the underlying reasons – were supported by the results of the observational SISTER study.

Disease activity and patient satisfaction

For the descriptive analyses of disease activity in the SISTER study, it should be considered that patients in the SC switcher cohort had already been on IV treatment with natalizumab and thus represented a rather stable selection of patients, while the treatment response of the de novo starters with either IV or SC natalizumab was uncertain and disease activity thus more dynamic after treatment initiation. This imbalance is likely to explain the descriptive post-baseline differences between these cohorts in effectiveness outcomes. In general, relapse rates were low in all treatment cohorts and descriptive lowest in the SC switcher cohort, suggesting a smooth switch from IV to SC treatment (285/310 patients overall [91.9%] remained relapse-free). The ARRs and proportions of patients with non-stable disease (i.e., progression on the EDSS) were slightly higher in the de novo starter cohorts than in the switcher cohort. However, different patient characteristics, small sample sizes particularly in the starter cohorts, and the fact that documentation of a relapse was not confirmed, need to be considered for the interpretation of these results. Overall, no clear differences in clinical effectiveness were seen between the IV and SC starter cohorts, consistent with the pivotal studies DELIVER 9 and REFINE. 10 Moreover, the disease activity reported in NOVA Part 2 was overall low during the crossover period of the study regardless of the SC or IV route of dosing. 17

The analysis of the TSQM-II in the SISTER study revealed neither obvious changes from baseline nor clinically relevant differences across cohorts at Months 6 and 12 in any cohort. Similar experience was made in NOVA Part 2, where (apart from a small but significant difference in EQ-5D-5L) no significant differences between SC and IV administration on the used PROs were observed. 17

Treatment safety and tolerability

ISRs were reported from 25 patients (i.e., 9.4% relating to the 266 patients on SC treatment). The majority of these ISRs were mild to moderate, and no “very strong” or even “intolerable” ISRs were documented, suggesting that SC natalizumab was locally well tolerable. The observed AE pattern was overall consistent with the established safety profiles of natalizumab SC or IV (with nasopharyngitis being the most common event in each cohort) and additionally included 18.1% of patients (n=56) who were diagnosed with COVID-19. The relatively small differences in AE and SAE incidences observed in SISTER and NOVA Part 2 suggest that underreporting of AEs in SISTER may have been less pronounced than is often anticipated in observational studies. Also in NOVA Part 2, the incidence of AEs and SAEs was comparable between the SC and IV groups, and most AEs were mild or moderate in severity. 17

Adverse events indicating the development of neutralizing antibodies to natalizumab were reported in 1 patient (2.3%) in the IV starter cohort and 3 patients (6.5%) in the SC starter cohort. In the targeted analysis of anti-natalizumab antibodies, (where many patients with missing data limited the value of the analysis), only 2 patients (both in the SC starter cohort) showed a conversion from “antibody-negative” at baseline to “positive” at least at one post-baseline visit. Overall, the maximum incidence of any anti-natalizumab antibodies in this study was 1.9% (4+2=6/310 patients). This finding suggested low immunogenic potential of natalizumab in general; however, the low incidence precludes any conclusions regarding differences in immunogenicity between the IV and SC administration.

No PML cases occurred in the SISTER study. Overall, no new or unexpected safety issues were identified with either route of administration.

The use of extended dose intervals in real world

Generally, natalizumab was dosed as instructed in the label. However, extended dosing intervals of >4 weeks were more frequently planned at baseline in the (pre-treated) SC switcher group (49.5%), while extended intervals occurred less frequently in the 2 de novo treated starter cohorts (no patients in the IV starter cohort and 6.7% in the SC starter cohort). This difference reflected the more stable conditions in the pre-treated SC switcher group, where patients were already on established therapy and thus had a longer exposure compared with the de novo starters. Interestingly, the proportions of patients with extended infusion intervals of >4 weeks had increased by Month 12 in all 3 cohorts (see Table 2 ), consistent with the results of the NOVA trial, where the switch from 4-weekly to 6-weekly IV dosing provided a high level of efficacy in controlling MS disease activity in clinically stable patients. 16

The reduction of procedure time and implications on cost savings

The analysis of the total procedure time at baseline (i.e., the time for preparation, administration, and post-administration safety surveillance) showed a clear difference in favor of the SC regimens. Post-baseline data showed similar relative results; however, the mean procedure times through Month 12 slightly declined in all 3 cohorts compared with baseline, thereby suggesting a certain familiarity with the treatment procedure and/or shorter/terminated safety monitoring over time. In NOVA Part 2, IV natalizumab took about 15 times longer to prepare and administer than SC natalizumab. 17

This advantage in time saving seems to translate into verifiable economic benefits. The EASIER study, a multicenter, observational, cross-sectional study to evaluate the economic impact associated with the use of natalizumab SC versus IV in RRMS patients in Italy, found that the SC route would result in 50% reduction in patient time and 55% reduction in active working time of healthcare providers, translating into a 63% cost reduction for the MS center per natalizumab administration procedure. 18 A cost analysis of SC vs. IV natalizumab based on RRMS patient care pathway over 2 years in Spain showed that at Years 1 and 2 combined, total time and cost reductions (excluding drug acquisition cost) per patient, driven by saving on administration and patient and caregiver productivity for SC at a reference hospital vs. IV at a reference hospital, were by 54.6% and 66.2%, respectively. In the case of natalizumab SC at a regional hospital, the total time and cost reductions were by 60.6% and 69.8%, respectively. 19 In conclusion, these 2 studies illustrate the potential to save health care expenditure through SC treatment and (ideally) self-administration at home.

Study limitations

Strengths and limitations of the observation SISTER study, such as selection bias, confounder bias, and the potential impact of missing data, have been discussed elsewhere. 11 Inherently, all statistical tests performed in this study are to be regarded purely descriptive due to the exploratory nature of the study and missing adjustment for multiple testing. Moreover, the originally planned samples sizes were not achieved, and particularly the 2 starter cohorts were small from the beginning and showed further relevant attrition due to drop-out and treatment discontinuation over time. Thus, the generalizability of the results obtained from the 2 starter cohorts was likely limited, and it cannot be ruled out that a “positive” selection of on-study patients had biased the results at Month 6 and particularly Month 12.

However, and despite these – mostly inherent – limitations, real world data from observational studies – as a supplement to highly designed RCTs – may provide reliable complementary information for clinical decision making. 20 – 22

Overall conclusions

The final results of the SISTER study, which was the first observational study to collect real-word data on treatment with SC natalizumab in Germany and Austria, were in line with the results of the previously published interim analysis. The preference rate for the SC route of 95.1% at baseline was high and supported the results of the Phase 3b NOVA extension study Part 2, where 87.2% of the patients preferred the SC route over the IV route after 48 weeks of crossover treatment on both routes.

Disease activity after start of study treatment was low in all 3 cohorts, injection-site reactions were transient and mostly mild or moderate, and the documented adverse event pattern was consistent with the known safety profiles of SC and IV natalizumab. Due to the high acceptance of SC natalizumab, its good safety and tolerability, and the similar effectiveness in combination with a reduced procedure time when compared with the IV route of administration, a home-based self-administration after appropriate education and a certain treatment customization under professional surveillance might be a suitable approach to further improve convenience of treatment and save healthcare costs at the same time.

Acknowledgements

Statistical analysis (Elisabeth Grünewald) and medical writing support (Dr. Bernd Graulich) was provided by the Contract Research Organization Winicker Norimed GmbH, Nuremberg, Germany. Manuscript submission was supported by Winicker Norimed GmbH, Nuremberg, Germany. This assistance was funded by Biogen GmbH Munich, Germany. All listed authors have authorized the submission of their manuscript via a third party and approved any statements or declarations, e.g., conflicting interests, funding, etc.

Author contributions: Gold, Ralf (RG) has made a substantial contribution to the concept or design of the article; or the acquisition, analysis, or interpretation of data for the article; drafted the article or revised it critically for important intellectual content; approved the version to be published; agreed to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. Schmidt, Stephan (SS) has made a substantial contribution to the concept or design of the article; or the acquisition, analysis, or interpretation of data for the article; drafted the article or revised it critically for important intellectual content; approved the version to be published; agreed to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. Deisenhammer, Florian (FD) has made a substantial contribution to the concept or design of the article; or the acquisition, analysis, or interpretation of data for the article; drafted the article or revised it critically for important intellectual content; approved the version to be published; agreed to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. Motte, Jeremias (JM) has made a substantial contribution to the concept or design of the article; or the acquisition, analysis, or interpretation of data for the article; drafted the article or revised it critically for important intellectual content; approved the version to be published; agreed to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. Richter, Nils (NR) has made a substantial contribution to the concept or design of the article; or the acquisition, analysis, or interpretation of data for the article; drafted the article or revised it critically for important intellectual content; approved the version to be published; agreed to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. Taipale, Kirsi (KT) has made a substantial contribution to the concept or design of the article; or the acquisition, analysis, or interpretation of data for the article; drafted the article or revised it critically for important intellectual content; approved the version to be published; agreed to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. Salmen, Hans Christian (HCS) has made a substantial contribution to the concept or design of the article; or the acquisition, analysis, or interpretation of data for the article; drafted the article or revised it critically for important intellectual content; approved the version to be published; agreed to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. Bohland, Christian (CB) has made a substantial contribution to the concept or design of the article; or the acquisition, analysis, or interpretation of data for the article; drafted the article or revised it critically for important intellectual content; approved the version to be published; agreed to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. Kahrer, Roman (RK) has made a substantial contribution to the concept or design of the article; or the acquisition, analysis, or interpretation of data for the article; drafted the article or revised it critically for important intellectual content; approved the version to be published; agreed to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. Schirduan, Ksenija (KS) has made a substantial contribution to the concept or design of the article; or the acquisition, analysis, or interpretation of data for the article; drafted the article or revised it critically for important intellectual content; approved the version to be published; agreed to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.

Funding: The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study was funded by Biogen GmbH Munich, Germany.cThe following author(s) disclosed receipt of financial support for the research, authorship, and/or publication of this article by Biogen GmbH Munich, Germany: Gold, Ralf; Schmidt, Stephan; Deisenhammer, Florian; Motte, Jeremias. The following author(s) are full-time employees at Biogen GmbH Munich, Germany: Taipale, Kirsi; Salmen, Hans Christian; Bohland, Christian; Schirduan, Ksenija. Kahrer, Roman is full-time employee at Biogen Austria GmbH, Vienna, Austria.

The authors declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: RG reports consulting and/or speaker fees from Baxter, Bayer, Biogen, Genzyme, Novartis, Roche, Sanofi, TEVA, Merck and Janssen-Cilag. RG is EIC of Therapeutic Advances in Neurological Disorders, therefore, the peer review process was managed by alternative members of the Board and the submitting Editor was not involved in the decision-making process. SS received personal compensations for serving on scientific advisory boards for Hoffmann-La Roche, Novartis, Merck Serono, Bayer Vital, Biogen, Genzyme, and Teva; and received travel funding and/or speaker honoraria from Hoffmann-La Roche, Novartis, Merck Serono, Bayer Vital, Biogen, Genzyme, and Teva. FD reports consulting and/or speaker fees from Alexion, Almirall, Biogen, Celgene, Genzyme-Sanofi, Horizon, Janssen, Merck, Novartis Pharma, and Roche. JM reports research support from Biogen. NR did not receive any consulting and/or speaker fees. CB, HCS, KT, RK, and KS report support from Biogen as employees and may own stock in Biogen.

Trial Registration: This observational (non-interventional) study was registered in the local German PEI register for non-interventional studies (NIS-No. 611) and in the international CTgov register ( NCT05304520 ).

Ethical considerations

This observational study was conducted according to the World Medical Association Declaration of Helsinki and all local regulatory requirements. Prior to study start, the study was notified to and approved (on 29-Jul-2021) by the Ethics Committee responsible for the Scientific Lead of the study, Prof. Dr. med Ralf Gold: Ethik-Kommission der Med. Fakultät der Ruhr Universität Bochum. Gesundheitscampus 33, 44801 Bochum, Germany. Approval/Register Number: 1-7297_3-BR.

Consent to participate

Informed consent to participate in the study was obtained from all participants in writing.

Consent for publication

Since all results were provided at a population level and no individual data were reported.

References

1 Biogen . TYSABRI 300 mg concentrate for solution for infusion . Electronic Medicines Compendium , Available at:Last updated on emc: 22 Nov 2024 (last accessed on 29-Jan-2026). https://www.medicines.org.uk/emc/product/222/smpc

2 Biogen . Tysabri 150 mg solution for injection in pre-filled syringe . Electronic Medicines Compendium . Available at: Last updated on emc: 16 Oct 2025(last accessed on 29-Jan-2026). https://www.medicines.org.uk/emc/product/12443/smpc

3 Miller DH Khan OA Sheremata WA , et al. A controlled trial of natalizumab for relapsing multiple sclerosis . N Engl J Med 2003 ; 348 : 15 – 23 . 10.1056/NEJMoa020696 12510038

4 Miller DH Soon D Fernando KT , et al. MRI outcomes in a placebo-controlled trial of natalizumab in relapsing MS . Neurology 2007 ; 68 : 1390 – 1401 . 10.1212/01.wnl.0000260064.77700.fd 17452584

5 Polman CH O'Connor PW Havrdova E , et al. A Randomized, Placebo-Controlled Trial of Natalizumab for Relapsing Multiple Sclerosis . New England Journal of Medicine 2006 ; 354 : 899 – 910 . 10.1056/NEJMoa044397 16510744

6 Butzkueven H Kappos L Wiendl H , et al. Long-term safety and effectiveness of natalizumab treatment in clinical practice: 10 years of real-world data from the Tysabri Observational Program (TOP) . J Neurol Neurosurg Psychiatry 2020 ; 91 : 660 – 668 . 10.1136/jnnp-2019-322326 32234967 PMC7279201

7 Perumal J Balabanov R Balcer L , et al. Long-Term Effectiveness and Safety of Natalizumab in African American and Hispanic/Latino Patients with Early Relapsing–Remitting Multiple Sclerosis: STRIVE Data Analysis . Neurology and Therapy 2023 ; 12 : 833 – 848 . 10.1007/s40120-023-00461-0 36966440 PMC10195921

8 Perumal J Balabanov R Su R , et al. Natalizumab in Early Relapsing-Remitting Multiple Sclerosis: A 4-Year, Open-Label Study . Adv Ther 2021 ; 38 : 3724 – 3742 . 10.1007/s12325-021-01722-w 34014549 PMC8279996

9 Plavina T Fox EJ Lucas N , et al. A Randomized Trial Evaluating Various Administration Routes of Natalizumab in Multiple Sclerosis . J Clin Pharmacol 2016 ; 56 : 1254 – 1262 . 10.1002/jcph.707 26835603

10 Trojano M Ramió-Torrentà L Grimaldi LM , et al. A randomized study of natalizumab dosing regimens for relapsing-remitting multiple sclerosis . Mult Scler 2021 ; 27 : 2240 – 2253 . 10.1177/13524585211003020 33821693 PMC8597184

11 Gold R Schmidt S Deisenhammer F , et al. Real-world evidence and patient preference for subcutaneous versus intravenous natalizumab in the treatment of relapsing-remitting multiple sclerosis - initial results from the observational SISTER study . Ther Adv Neurol Disord 2024 ; 17 : 1 – 19 . 10.1177/17562864241241382 PMC11015759 38616781

12 von EE Altman DG Egger M , et al. The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: guidelines for reporting observational studies . J Clin Epidemiol 2008 ; 61 : 344 – 349 . 10.1016/j.jclinepi.2007.11.008 18313558

13 Thompson AJ Banwell BL Barkhof F , et al. Diagnosis of multiple sclerosis: 2017 revisions of the McDonald criteria . The Lancet Neurology 2018 ; 17 : 162 – 173 . 10.1016/S1474-4422(17)30470-2 29275977

14 Atkinson MJ Kumar R Cappelleri JC , et al. Hierarchical construct validity of the treatment satisfaction questionnaire for medication (TSQM version II) among outpatient pharmacy consumers . Value Health 2005 ; 8 ( Suppl 1 ): S9 – s24 . 10.1111/j.1524-4733.2005.00066.x 16336491

15 Plavina T Subramanyam M Bloomgren G , et al. Anti-JC virus antibody levels in serum or plasma further define risk of natalizumab-associated progressive multifocal leukoencephalopathy . Ann Neurol 2014 ; 76 : 802 – 812 . 10.1002/ana.24286 25273271 PMC4282070

16 Foley JF Defer G Ryerson LZ , et al. Comparison of switching to 6-week dosing of natalizumab versus continuing with 4-week dosing in patients with relapsing-remitting multiple sclerosis (NOVA): a randomised, controlled, open-label, phase 3b trial . Lancet Neurol 2022 ; 21 : 608 – 619 . 10.1016/s1474-4422(22)00143-0 35483387

17 Wiendl H Foley J Defer G , et al. Patient Preference for Subcutaneous Versus Intravenous Administration with Every-6-Week Natalizumab (Tysabri(®)) Dosing: NOVA Phase IIIb Extension Study (Part 2) . Neurol Ther 2024 ; 13 : 1385 – 1401 . 10.1007/s40120-024-00647-0 39046635 PMC11393236

18 Filippi M Grimaldi L Conte A , et al. Intravenous or subcutaneous natalizumab in patients with relapsing–remitting multiple sclerosis: investigation on efficiency and savings—the EASIER study . Journal of Neurology 2024 ; 271 : 340 – 354 . 10.1007/s00415-023-11955-0 37715789 PMC10769988

19 Alonso Torres AM Arévalo Bernabé AG Becerril Ríos N , et al. Cost-Analysis of Subcutaneous vs Intravenous Administration of Natalizumab Based on Patient Care Pathway in Multiple Sclerosis in Spain . Pharmacoecon Open 2023 ; 7 : 431 – 441 . 10.1007/s41669-023-00394-2 36802327 PMC10169937

20 Benson K Hartz AJ . A comparison of observational studies and randomized, controlled trials . N Engl J Med 2000 ; 342 : 1878 – 1886 . 10.1056/nejm200006223422506 10861324

21 Concato J Shah N Horwitz RI . Randomized, controlled trials, observational studies, and the hierarchy of research designs . N Engl J Med 2000 ; 342 : 1887 – 1892 . 10.1056/nejm200006223422507 10861325 PMC1557642

22 Hong YD Jansen JP Guerino J , et al. Comparative effectiveness and safety of pharmaceuticals assessed in observational studies compared with randomized controlled trials . BMC Medicine 2021 ; 19 : 307 . 10.1186/s12916-021-02176-1 34865623 PMC8647453