Article Text

Download PDFPDF

Short report
Abnormal inflammatory activity returns after natalizumab cessation in multiple sclerosis
  1. Antoine Gueguen1,
  2. Pascal Roux2,
  3. Romain Deschamps1,
  4. Antoine Moulignier1,
  5. Caroline Bensa1,
  6. Julien Savatovsky2,
  7. Françoise Heran2,
  8. Olivier Gout1
  1. 1Department of Neurology, Fondation Ophtalmologique A. de Rothschild, Paris, France
  2. 2Department of Medical Imaging, Fondation Ophtalmologique A. de Rothschild, Paris, France
  1. Correspondence to Dr Antoine Guéguen, Department of Neurology, Fondation Ophtalmologique A. de Rothschild, 25–29, rue Manin, Paris 75019, France; agueguen{at}


Objective To characterise recurrence of multiple sclerosis (MS) inflammatory activity during the year following natalizumab (NTZ) cessation.

Methods Thirty-two patients with MS were included in a monocentric cohort study. Data were collected prospectively during and after NTZ, with serial clinical and MRI evaluations. The first relapse occurring after interrupting NTZ was the primary outcome measure. The numbers of gadolinium-enhancing lesions before, during and after NTZ treatment, were compared.

Results During the year following NTZ cessation, the cumulative probability of relapses was estimated at 52.9% and an unusually high MRI inflammation was noticed. It was defined by a number of gadolinium-enhancing lesions >5 and exceeding the gadolinium lesions existing before NTZ initiation. Rebound of MS activity after NTZ cessation was characterised by association of relapses and unusual MRI inflammation. Cumulative probability of rebound was estimated at 39% and mostly occurring between 3 months and 9 months after interrupting NTZ. Risk of rebound appears related with a higher annualised relapse rate and a lower Expanded Disability Status Scale score before NTZ initiation. Rebound was associated with severe recurring relapses in 9% of the patients.

Conclusions This study identifies rebound after NTZ cessation as an association of relapses and high MRI activity.

  • Multiple Sclerosis
  • MRI
View Full Text

Statistics from

Natalizumab (NTZ), a monoclonal antibody (α-4 integrin antagonist), is mostly used as a second-line treatment of multiple sclerosis (MS). NTZ efficacy is countered by the risk of progressive multifocal leucoencephalitis. The ability to interrupt NTZ for patients at high risk of progressive multifocal leucoencephalitis has improved with the stratification of that risk and the availability of alternative therapeutic agents.

After stopping NTZ, severe relapse and unusually high inflammatory MRI activity were observed in monocentric open studies and case reports.1–6 Conversely, the analysis of a large cohort of patients included in phase III trials after discontinuing NTZ provided evidence of renewed of disease activity without rebound.7 Additional data have been provided recently by larger studies8–10 but the evaluation of MS inflammatory recurrence risk associated with NTZ interruption remains difficult in daily practice.

Therefore, we undertook a non-interventional monocentric study to characterise the recurrence of inflammatory activity during the year following NTZ interruption.


NTZ-treated patients’ data were recorded prospectively during and after NTZ treatment. Patients underwent serial evaluations with the Expanded Disability Status Scale (EDSS) score and MRI. Each patient's MS history and treatments were collected from his/her medical chart. Three alternative treatment procedures after stopping NTZ have been defined in our department: (1) no treatment, (2) interferon-β or glatiramer acetate started within 1 month or immediately, respectively, and (3) 1 month after NTZ interruption, a 1-gram methylprednisolone pulse given monthly for 6 months.

Among 122 NTZ-treated patients, 32 who discontinued NTZ fulfilled the three inclusion criteria: (1) they had been treated with NTZ for >1 year, (2) they did not have a relapse after the first 3 months of NTZ use and (3) they were followed for >12 months after stopping NTZ.

The date of the first relapse during the year following NTZ withdrawal was defined as the main clinical outcome. In addition, EDSS scores and all relapses from the year preceding the year following NTZ treatment, were taken into account. Relapse was defined as the emergence of new or worsening of existing, neurological symptom(s) lasting >24 h. A severe relapse was defined by an increase above one point of the EDSS score.

Patients were scanned on a Philips 3T MRI and a contrast enhancing 3D TSE T1WI sequence was done. A neurologist and a neuroradiologist determined the numbers of gadolinium-enhancing lesions on MRI performed before NTZ introduction, at NTZ interruption and during the year thereafter. Consensus concerning differences between the two readers’ evaluations was reached by a third reader. Among the MRI obtained during the year following NTZ interruption for each patient, the one with the maximum number of gadolinium-enhancing lesions was retained for analysis. An unusual MRI pattern of gadolinium-enhancing lesions after stopping NTZ was defined by more than five gadolinium-enhancing lesions (exceeding the upper 95% CI of enhanced lesions observed before NTZ in all patients) and a number of gadolinium-enhancing lesions after NTZ interruption higher than before NTZ onset. The association of relapse to unusual high inflammatory MRI pattern defines rebound.

The population was divided into two groups based on relapse occurrence during the year following NTZ discontinuation.

Statistical analysis

Gage repeatability and reproducibility (R&R) analysis of variance was used to evaluate MRI-assessment variability between examiners (Gage R&R=0.41). Variables are expressed as means±SD or medians (range) as appropriate. Parameters were compared between groups with Fischer's test and Wilcoxon-test. Within-group and between-groups comparisons of the annualised rate of relapses (ARRs, defined as the number of confirmed relapses per year), the EDSS score and the number of gadolinium-enhancing lesions were performed using a paired Wilcoxon-test and Wilcoxon-test, respectively.

The occurrence of the first relapse was assessed by univariable survival analysis (Kaplan–Meier, Weibull) and multivariable proportional HR analysis (Cox model) to identify risk factors of relapse and unusual MRI activity.


The study population's characteristics and the alternative treatment after stopping NTZ are reported in table 1.

Table 1

Characteristics of the 32 patients studied more than 1 year after stopping NTZ

Relapse and MRI parameters

One year after stopping NTZ, the cumulative probability of relapse (Kaplan-Meier analysis; Weibull β=1.34) was 52.5% (95% CI 36% to 68%) (see online supplementary figure). The ARR before starting NTZ did not differ significantly between groups (18 patients with relapse and 14 relapse-free patients after NTZ interruption). Patients who relapsed reached an ARR during the year after NTZ interruption equal to that observed before starting NTZ.

The numbers of gadolinium-enhancing lesions were comparable between groups before NTZ treatment but increased markedly after NTZ interruption in patients with relapse. Relapse-free patients had fewer gadolinium-enhancing lesions after stopping NTZ. An unusual inflammatory MRI pattern was seen only in patients with relapse (figure 1).

Figure 1

Serial MRI assessments of gadolinium-enhancing lesions. Black circles: number of gadolinium-enhancing lesions in patients with relapse. Grey circles: numbers of gadolinium-enhancing lesions in relapse-free patients. B, before starting natalizumab (NTZ); I, at NTZ interruption; P, during the post-NTZ year; 0, relapse-free patients during the year after stopping NTZ; 1, patients who had a relapse during the year after NTZ withdrawal. **p=0.01 NS; no significance.

Relapse with an unusual MRI pattern

The cumulative probability (Kaplan-Meier analysis) of relapse associated with an unusual inflammatory MRI pattern at 1 year post-NTZ was 39.0% (95% CI 24% to 57%). The median time between NTZ interruption and the first relapse for this specific group was 5.8±2.9 months. The risk of rebound was associated with the higher ARR during the year before NTZ initiation (Cox model, HR=2.25 (95% CI 1.17 to 4.67); p=0.014) or the lower EDSS score at the time of NTZ introduction (HR=0.63 (95% CI 0.003 to 1.89); p=0.015). The univariable Cox model did not identify effect on the risk of rebound after NTZ cessation of age, age at MS onset, gender, duration of NTZ, number of NTZ infusions, number of gadolinium-enhancing lesions before NTZ, EDSS score before and at interruption of NTZ and alternative treatment.

Severity of relapse

Three (9%) patients presented rebound and severe neurological worsening. They suffered several relapses (three to four) and EDSS score increase (1.5–3.5), led to reinitiate NTZ in a John Cunningham virus (JC)-negative patient and use of mitoxantrone for the others. The number of gadolinium-enhancing lesions of these patients ranged from 20 to 80. At 1 year after relapse, one patient had a two-point EDSS score increase, while the other two had returned to their prerelapse levels.


We observed an unusually high inflammatory MRI pattern associated with relapse defined as rebound after NTZ cessation in 39% of our patients and severe relapses in 9%. The return of inflammatory activity occurred mostly between the 3rd month and the 9th month after NTZ interruption, independently of the alternative treatment prescribed. Conversely, the risk of relapse appeared higher for patients with low EDSS scores and high ARR before starting NTZ, but the small number of patients included in the study limited our ability to identify risk factors accurately.

MS relapse after NTZ discontinuation was analysed at the end of a phase III trial that had followed 1866 patients for 8 months.7 Those authors found no effect of treatment on the risk of relapse for their patients, with no statistically significant difference in terms of ARR before and after NTZ. Our results are in line with these findings.

Conversely, the percentages of patients who relapsed in our study were higher, as were the numbers of gadolinium-enhancing lesions after stopping NTZ. A higher relapse rate was also observed in the ENIGM study8 and another monocentric study.1–6 This difference might be explained by our patients’ higher ARR before NTZ initiation and their longer follow-up. Moreover the design of the study and the assessment of gadolinium-enhancement lesions differed. We retained the MRI with the highest numbers of gadolinium-enhancing lesions because it appears more appropriate to assess abnormal inflammatory activity reappearance that varies in time among patients (ie, between 3 months and 9 months post NTZ interruption). Nevertheless the non-interventional design of the study induces MRI procedure limitation.

The risk of MS inflammatory activity reappearance following NTZ interruption seems higher at the beginning of the disease in patients with low EDSS and high inflammatory activity before NTZ initiation. The high ARR before NTZ introduction is an important predictor that has been identified as a risk factor in other studies.8 ,9 Alternative treatments used in the study have no significant effects in prevention of inflammation recurrence as observed in the RESTORE study.10 However a benefit of introducing fingolimod treatment less than 3 months after NTZ interruption and before the reappearance of inflammatory activity has recently been shown.8 ,9 In view of our study, a MRI and clinical assessment could be helpful before introducing fingolimod and in case of relapse to identify rebound and ensure better patient safety after NTZ interruption.


The authors thank Faroudy Boufassa for helpful support with statistical analysis and Janet Jacobson for helpful feedback.


View Abstract

Supplementary materials

  • Supplementary Data

    This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

    Files in this Data Supplement:


  • Contributors AG: designed the study, analysed, collected and interpreted data, performed the statistical analyses and drafted the manuscript. PR and JS analysed and collected data. RD, AM, CB and FH: collected data. OG: designed the study, analysed and interpreted data, and edited the manuscript.

  • Competing interests AG serves as an editorial board member of Biogen Idec, Novartis and Teva, and received honoraria from Biogen Idec, Novartis, Teva. He received funding for travel from Biogen Idec, Novartis, Teva Pharma. A close relative receives revenue from Genzyme. PG reports no disclosures. RD reports no disclosures. AM received funding for lectures from Abbott, Biogen Idec, Gilead Sciences, MSD France, and for travel to meetings from Biogen Idec, Novartis, Teva Pharma. CB reports disclosures for scientific advisory tasks from Biogen Idec, Novartis, Teva Pharma. JS received payments for lectures from Bayer Pharmaceuticals, MSD and Gilead Science. OG received consulting and lecture fees fram Allergan, Almitall, Bayer Schering Pharma, Biogen Idec, Genzyme, Novartis, Merck Serono, Sanofi and Teva Pharma.

  • Ethics approval The study has been ethically approved by the CNIL ‘Commission Nationale de l'Informatique et des Libertés’.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.