Article Text


Research paper
Extended interval dosing of natalizumab in multiple sclerosis
  1. L Zhovtis Ryerson1,
  2. T C Frohman2,
  3. J Foley3,
  4. I Kister1,
  5. B Weinstock-Guttman4,
  6. C Tornatore5,
  7. K Pandey6,
  8. S Donnelly7,
  9. S Pawate8,
  10. R Bomprezzi9,
  11. D Smith10,
  12. C Kolb4,
  13. S Qureshi2,
  14. D Okuda2,
  15. J Kalina1,
  16. Z Rimler1,
  17. R Green6,
  18. N Monson2,
  19. T Hoyt3,
  20. M Bradshaw8,
  21. J Fallon1,
  22. E Chamot11,
  23. M Bucello4,
  24. S Beh2,
  25. G Cutter11,
  26. E Major12,
  27. J Herbert1,
  28. E M Frohman2,13,14,15
  1. 1Department of Neurology, Langone Medical Center, New York University, New York, New York, USA
  2. 2Departments of Neurology & Neurotherapeutics, University of Texas Southwestern Medical Center, Dallas, Texas, USA
  3. 3Rocky Mountain MS Clinic, Salt Lake City, Utah, USA
  4. 4University of Buffalo, Buffalo, New York, USA
  5. 5Georgetown University, Washington DC, USA
  6. 6Barnabas Health MS Center, Livingston, New Jersey, USA
  7. 7CUNY Graduate Center, New York, New York, USA
  8. 8Vanderbilt University Medical Center, Nashville, Tennessee, USA
  9. 9University of Massachusetts School of Medicine, Worcester, Massachusetts, USA
  10. 10Multiple Sclerosis Center of Connecticut, Norwich, Connecticut, USA
  11. 11University of Alabama School of Public Health, Birmingham, Alabama, USA
  12. 12National Institute of Neurological Disorders and Stroke, Bethesda, Maryland, USA
  13. 13Department of Ophthalmology, University of Texas Southwestern Medical Center, Dallas, Texas, USA
  14. 14Department of Bioengineering, University of Texas at Dallas, Dallas, Texas, USA
  15. 15Department of Behavioural and Brain Sciences, University of Texas at Dallas, Dallas, Texas, USA
  1. Correspondence to Professor EM Frohman, Department of Neurology and Neurotherapeutics, University of Texas Southwestern School of Medicine, 5323 Harry Hines Blvd, Dallas, TX 75390, USA; Elliot.frohman{at}


Background Natalizumab (NTZ), a monoclonal antibody to human α4β17 integrin, is an effective therapy for multiple sclerosis (MS), albeit associated with progressive multifocal leukoencephalopathy (PML). Clinicians have been extending the dose of infusions with a hypothesis of reducing PML risk. The aim of the study is to evaluate the clinical consequences of reducing NTZ frequency of infusion up to 8 weeks 5 days.

Methods A retrospective chart review in 9 MS centres was performed in order to identify patients treated with extended interval dosing (EID) regimens of NTZ. Patients were stratified into 3 groups based on EID NTZ treatment schedule in individual centres: early extended dosing (EED; n=249) every 4 weeks 3 days to 6 weeks 6 days; late extended dosing (LED; n=274) every 7 weeks to 8 weeks 5 days; variable extended dosing (n=382) alternating between EED and LED. These groups were compared with patients on standard interval dosing (SID; n=1093) every 4 weeks.

Results 17% of patients on SID had new T2 lesions compared with 14% in EID (p=0.02); 7% of patients had enhancing T1 lesions in SID compared with 9% in EID (p=0.08); annualised relapse rate was 0.14 in the SID group, and 0.09 in the EID group. No evidence of clinical or radiographic disease activity was observed in 62% of SID and 61% of EID patients (p=0.83). No cases of PML were observed in EID group compared with 4 cases in SID cohort.

Conclusions Dosing intervals up to 8 weeks 5 days did not diminish effectiveness of NTZ therapy. Further monitoring is ongoing to evaluate if the risk of PML is reduced in patients on EID.

View Full Text

Statistics from


Natalizumab (NTZ) is a recombinant partially humanised monoclonal antibody directed against α4β1 integrin receptor on the surface of circulating mononuclear cells which in turn interferes with cellular adhesion to cerebrovascular endothelium, via vascular cell adhesion molecule-1. This mechanism has been found useful in the treatment of relapsing forms of multiple sclerosis (MS) exhibiting substantial reductions in both clinical and radiographic evidence of disease activity.1–4 Notwithstanding these impressive benefits, the application of this therapy is associated with a potentially fatal complication by reactivation of latent JC virus (JCV) causing a demyelinating infection of the central nervous system (CNS), progressive multifocal leukoencephalopathy (PML). The PML risk in NTZ-treated recipients on treatment every 4 weeks for 49–72 months, who are seropositive for antibodies against the JCV (IgG) with a titre ≥1.5 is estimated at 8.5/1000, rising to 13/1000 in those with prior immunosuppression.5 ,6

Current treatment guidelines utilise a standard NTZ 300 mg dose administered intravenously every 4 weeks, allowing the NTZ concentrations to be maintained at levels, which ensure continuous maximal α4β1 integrin receptor saturation. After administration of a single NTZ dose, serum concentrations rise rapidly followed by a prolonged terminal phase during which serum concentrations decline, reaching values in the range of 3 μg/mL over 4 weeks.7–9

By 8 weeks postsingle infusion, NTZ concentrations are still detectable at 1 μg/mL.7 By contrast, after a single 300 mg dose of intravenous NTZ, α4β1 integrin reaches maximal saturation, defined as >80% saturation of α4β1 integrin receptors on the surface of lymphocytes, and remains at near maximal saturation for 3–4 weeks.3 ,8 Between 4 and 8 weeks postadministration α4β1 integrin saturations decline to submaximal levels, ranging between 50% and 80%.

Receptor desaturation, defined as saturation <50%, has only been shown to occur when NTZ serum concentrations fall below 1 μg/mL.10 This occurs more than 8 weeks postdrug administration since detectable levels of NTZ (>1 μg/mL) are seen at 8 weeks and are expected to be even higher among patients with repeated administration due to rising concentration levels over time, as well as based on individual metabolism, and body mass.2 ,7 ,9 ,11–16 Notably, rebound of MS clinical activity after NTZ discontinuation tends to occur around 10–12 weeks following drug withdrawal, suggesting that NTZ receptor saturation may be more germane to NTZ disease-modifying mechanism of action than NTZ serum concentration.11 ,13 ,14 ,17–22

Through an analogous process, PML susceptibility may reflect NTZ-induced blockade of α4β1 integrin receptors. The blockade leads to excessive reduction in discrete tissue compartment trafficking of immune cells required for JCV surveillance, as evidenced by the reduction of CD4+, CD8+, CD19+ and CD138+ cells in the cerebrospinal fluid persisting even 6 months postcessation of NTZ, and/or by allowing increased migration of JCV infected cells out of the bone marrow.23–27 Less frequent NTZ dosing may result in submaximal α4β1 integrin receptor saturation, adequate to exclude autoreactive T cells from entry into CNS (‘MS-protective’) but nevertheless sufficiently permissive to enable normal CNS lymphocyte scavenging of JCV to occur (‘PML-protective’). Alternatively, the reduction in receptor saturation may have implications with respect to trafficking within the bone marrow, and as such reduce the mobilisation of CD34+ cells harbouring the JCV into the circulation.

Clinicians throughout the USA and the world have begun to utilise various extended interval dosing (EID) schedules. This study is an analysis of data from a cohort of patients collected in nine individual MS clinics throughout the USA on EID, with scheduled dose extension up to 8 weeks 5 days. The objective of the current study was to ascertain the therapeutic durability of the extended dosing strategy, as a pretext from which to later operationalise the hypothesis, that a longer latency between NTZ infusions will be associated with a corresponding reduction in the risk of PML.


A retrospective chart review was conducted on 2004 NTZ-treated patients from nine US MS centres, examining the experiences of 905 patients on EID treatment intervals as compared with 1099 patients on standard interval dosing (SID). Institutional Review Board approval was obtained from each of the individual institutions. Patients were stratified into four groups based on NTZ treatment schedule received: (1) group 1 (n=669)—(SID) with patients receiving infusions on average, every 4 weeks 2 days; (2) group 2 (n=230)—early extended dosing (EED) with patients infusing on a schedule from every 4 weeks 3 days to 6 weeks 6 days; (3) group 3 (n=244)—late extended dosing (LED) with patients infusing on a schedule from every 7 weeks to 8 weeks 5 days; (4) group 4 (n=196)—variable extended dosing (VED)—which included patients who alternated between EED and LED, in any sequence. EID included all patients in the EED, LED and VED groups. Owing to the individual, physician-initiated retrospective observational nature of this study, the groups were created based on the various clinics’ protocols. All patients had standard dose schedule of NTZ infusions for 6 months or longer prior to start of EID. Three or greater consecutive extended dose infusions were required for stratification into one of the extended dose groups.

Statistical methods

The four groups were compared on demographics (table 1), clinical outcomes (table 2) and ‘no evidence of disease activity’ (NEDA; table 3), a combined metric that takes into account the absence of both clinical and radiographic measures of disease activity. Linear regression for continuous variables, logistic regression for binary variables, grouped binomial regression for proportions and negative binomial regression for counts were utilised.

Table 1

Comparisons of SID and EID groups on demographic variables

Table 2

Comparisons of SID and EID groups on MS activity variables

Table 3

Comparisons of standard interval dosage and extended interval dosage groups on MS activity and NEDA

All models included three indicator variables, comparing each of the extended dosage groups to the ‘standard practice’ SID group. In order to minimise the threat of type II errors (ie, false-negative designations), p values were not adjusted for multiple comparisons. The employed models included a set of indicator variables for clinic site, to control for the presence of distinct characteristics and practices at each clinic. Dosage schedule was not balanced across sites (see online supplementary table S1). Additionally, all models for clinical outcomes, except the one utilised for the adjusted annualised relapse rate (ARR), controlled for relevant covariates, including disease duration, gender, age and the number of months on current dosage treatment regimen at the end point of the retrospective ascertainment period. The adjusted ARR, defined by treating neurologists, was estimated using a negative binomial regression in conjunction with the indicator variables listed above, as well as age and disease duration as covariates, along with years on the current dosage schedule, as the defined exposure term. Poisson regression was overdispersed. The negative binomial model also exhibited evidence of overdispersion, though it was less serious than the Poisson model. Four observations with Pearson's residuals greater than 10 were dropped, and bootstrapped SEs were used for p values.

As it was necessary to control for MS disease duration for the assessment of clinical activity, the participants with incomplete data sets for this variable were dropped from all analyses (n=30). Furthermore, given the retrospective design of our study, we purposely identified patients with incomplete ascertainment data on some of the demographic and outcome variables. The online supplementary table S2 displays the number of available cases for each demographic and each of the corresponding outcome variables.


Baseline demographics and disease characteristics were obtained on 1099 patients on the SID schedule and 905 on the EID schedule (table 1). Both of the cohorts were female predominant (70% and 72% in the SID and EID groups, respectively; 79% for EED; 64% for LED; 73% VED), highly representative of the known gender prevalence rates for MS. The mean age for both groups was 45. The mean disease duration was 12 (SD±7.4) years in the SID and 13.2 (SD±7.3) years in the EID group. A higher percentage of JCV IgG-positive patients were seen in EID compared with SID group; 63% and 50%, respectively (quest diagnostics), which is expected considering clinicians are more eager to switch patients with higher risk for PML to EID dosing, predicated on the hypothesis that this treatment strategy may be germane to reducing the overall risk of developing PML, without a corresponding compromise in treatment efficacy. The median JCV antibody index was 0.23 in SID compared with 0.75 in EID group among all patients. The mean JCV index among patients with detected JCV status was 1.7 in SID and 2.03 in EID cohort. Eighteen per cent of patients in EID and 12% of SID group had received antecedent immunosuppression therapy before the inception of NTZ treatment. Seventy-four per cent of patients in EID cohort had been on NTZ treatment for over 24 months, compared with 48% in SID group. The mean total number of NTZ doses was 27.5 (SD±18.4) for SID group. EID patients had a mean of 40.1 (SD±21.7) total NTZ infusions with 25.4 (SD±15.67) mean months on EID dose scheduling (d=0.14).

Clinical and radiographic activity for both SID and EID remained very low, and were comparable (table 2). While 17% of patients on SID had new T2 lesions compared with 14% in EID (p=0.02), with 9%, 17% and 16% seen across EED, LED and VED groups, respectively, the SID group had a lower percentage of patients exhibiting gadolinium-enhancing T1 lesions, at 7% compared with 9% in EID groups (p=0.08). Conspicuously, the VED treated group exhibited the highest rate of new active lesions at 13%, as compared with 4% in the EED and 7% in LED group. The ARR was 0.14 in SID and 0.09 in EID group (p=0.02). LED showed the lowest ARR with 0.04, compared with 0.21 and 0.10 observed in EED and VED groups, respectively. The percentage of patients requiring steroids was similar across both SID and EID groups: 22% and 20% (p=0.54) respectively, with 20%, 17% and 22% observed across EED, LED and VED groups, respectively.

The efficacy of NTZ is reflected in the high percentage of patients in both groups fulfilling the designation of ‘NEDA’, a preplanned analyses of composite clinical and radiographic activities (table 3). No statistically significant differences were observed between groups in demonstrating no evidence of radiological activity (SID 82%; EID 86%; p=0.35), no evidence of clinical activity (SID 74%; EID 72%; p=0.74), and absence of both clinical and radiographic activity (SID 62%; EID 61%; p=0.83). The VED cohort had the lowest percentage of patients meeting criteria for NEDA at 59%.


The purpose of this study was to document the clinical outcome of extending the dosing treatment schedule up to 8 weeks (±5 days) already utilised in the MS community as reflected by the high number of patients in the cohort. EID does not diminish the effectiveness of NTZ, with comparable results seen across all measured outcomes, including ARR, MRI activity, and NEDA between the SID and EID groups. Furthermore, no statistical significance was seen among the three different EID groups except the LED cohort demonstrated a significantly lower percentage of patients with relapses or clinical worsening. This may reflect, at least in part, selection bias for patients who have less aggressive disease, and therefore are more likely to be disease free on EID schedule. While the LED group's ARR was only 0.04, 17% required steroids, which may reflect the phenomenon of patients reporting the well-recognised ‘wearing off’ phenomenon with NTZ, and reporting increased MS symptoms potentially leading to higher corticosteroid use by physicians.28

A high proportion of patients demonstrating NEDA is encouraging in our ability to halt the inflammatory components of MS. However, while NEDA is increasingly being utilised as an outcome measure in MS treatment trials, it is critical to underscore the principle that this designation does not truly indicate that the patient under investigation is completely devoid of disease activity secondary to this disease. Instead, the meaning of this designation is restricted to the absence of disease activity by virtue of how we measure it. NEDA is too simplistic of a model to assess the complexity of MS, which in this paper was further reduced by the lack of expanded disability severity scale scores available.

The EID cohort is a theoretically high-risk group for PML due to previously identified risk factors including 74% of patients on treatment for over 2 years, elevated mean JCV index of 2.03 among patients with detected JCV, which constitutes 63% of the EID cohort, and 18% of patients with prior history of immunosuppression.5 ,6 However, no cases of PML have been encountered in the EID cohort with 1090 JCV antibody-positive person years observed, while SID cohort has four cases of PML in 1052 JCV antibody-positive person years. Thus far, the EID cohort is not adequately powered to achieve a statistically significant PML risk reduction. Based on the properties of the Poisson distribution, we need to observe 0 cases over a total of 1248 person years in EID cohort to demonstrate lower PML risk than the postmarketing incidence rate of 2.5 per 1000 JCV antibody-positive person years at the p=0.05 level (calculations based on postmarketing incident rate of PML is 566 cases in 415 207 total patient years of NTZ exposure worldwide as of June 2015 with estimated 55% of population being JCV antibody positive).5 ,6 Demonstration of maintenance of clinical efficacy of EID allows for continued evaluation for PML risk reduction.

The limitations of our study include a retrospective design, the inability to control for baseline disease characteristics and possible selection bias, with likely less aggressive patients with MS moved on to the EID schedule. This is reflected in statistically significant better efficacy characteristics across multiple variables explored, including ARR and percentage of new T2 lesions on MRI. Although the cumulative number of T2 lesions is one of the outcome measures, we are limited by lack of a central reading centre applying a highly stringent and systematic method on the formal scoring of individual scans. With respect to the clinical ascertainment data as it relates to disability progression, we did not utilise EDSS as an outcome measure due to inconsistency of its use across the nine MS centres.

Physicians treating patients with MS are consistently attempting to balance the risks and benefits of this highly effective medication.29 The high number of patients on EID therapy reflects the clinical change in practice taken by US clinicians, diverging from the monthly SID. The retrospective study reflects the investigators interest in determining what clinical effect this is having on managing MS. Presently, we are continuing to follow the reported cohort, while expanding the scope of our study to include patients on EID NTZ treatment schedules throughout the world, via the extended NTZ dosing (EXTEND) registry through MSBase. These steps are aimed at increasing the number of JCV antibody-positive patients on EID NTZ schedules to ascertain the true risk of PML in those patients treated with EID. These data will ultimately lead to better structured prospective as well as non-inferiority studies. Ongoing pharmacokinetic characterisation of NTZ concentrations, α4β1 integrin receptor saturation and monitoring of CD34+ cell mobilisation from the bone marrow within the extended dose patient population may help in confirming our hypotheses on potential for PML risk mitigation.


The initiator and guiding light of this project for many years, JH, passed away in 2015. The authors would like to dedicate this work to the memory of this visionary and courageous man, innovative clinician, generous mentor and dear friend; he is greatly missed.


View Abstract


  • Contributors LZR, TCF, IK, JF, BW-G, SD, JK, TH, EC, JH, EMF were involved in collection and characterisation of patient data, including dosing intervals, examination and assessments of imaging investigations, drafting and editing of the manuscript. CT, KP, SP, RB, DS, CK, SQ, SB, DO, ZR, RG, NM, MBu, MBr, JF were involved in collection and characterisation of patient data, including dosing intervals, examination and assessments of imaging investigations. GC, EM were involved in drafting and editing of the manuscript.

  • Competing interests LZR has received research support from Biogen Idec. She has received compensation for advisory board and speaker activities from Biogen Idec and Teva. TCF has received speaker and consultant fees from Novartis, Genzyme and Acorda. IK is the member of scientific advisory board for Biogen Idec and received research support from Guthy-Jackson Charitable Foundation, National Multiple Sclerosis Society, Biogen-Idec, Serono and Navartis. JF has served as a consultant for and receives honoraria from Biogen, Genzyme, Teva, Novartis and Avanir. BW-G has participated in speaker’s bureaus and served as a consultant for Biogen Idec, Teva Neuroscience, EMD Serono, Novartis, Genzyme & Sanofi, Acorda Therapeutics, Inc and Genentech. BW-G also has received grant/research support from the agencies listed in the previous sentence as well as Questcor Pharmaceuticals, Inc and Shire. She serves in the editorial board for BMJ Neurology, Journal of International MS and CNS Drugs. CT has received speaker and consulting fees from Biogen Idec. KP has received speaker and consulting fees from Acorda, TEVA and Biogen. DS has received speaker and consulting fees from Biogen Idec. CK has received speaker and consulting fees from Biogen, Teva, EMD Serono and Acorda. DO received lecture fees from Acorda Therapeutics, Genzyme and TEVA Neuroscience, consulting and advisory board fees from Genzyme, Novartis and TEVA Neuroscience, and research support from Biogen. EC reports grants from Genzyme Corp, grants from Consortium of Multiple Sclerosis Centers, grants from Biogen Idec, grants from National Multiple Sclerosis Society, outside the submitted work. MBu has received speaker and consulting fees from Genzyme, Teva and EMD Serono. GC has received consulting, speaking fees and advisory boards: Cerespir Inc, Consortium of MS Centers (grant), D3 (Drug Discovery and Development), Genzyme, Genentech, Innate Therapeutics, Jannsen Pharmaceuticals, Klein-Buendel Incorporated, Medimmune, Novartis, Opexa Therapeutics, Receptos, Roche, Savara Inc, Spiniflex Pharmaceuticals, Somahlution, Teva pharmaceuticals, Transparency Life Sciences. EM reports personal fees from PML Consortium, personal fees from Takeda/Millennium Pharma, personal fees from Glaxo Smith Klein, personal fees from Genentech Roche, personal fees from Sanofi Genzyme, outside the submitted work. JH has received research support from Biogen Idec. EMF has received speaker and consultant fees from Novartis, Genzyme, TEVA and Acorda.

  • Ethics approval Investigational Review Board for each institution.

  • 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.