Product-limit survival estimates of the proportion with no serious infection and number of patients remaining at risk by year since treatment start, for each treatment group. Shaded areas are 95% Hall-Wellner confidence intervals. General population indicates age-matched and sex-matched general population comparator participants who were free of multiple sclerosis. GA indicates glatiramer acetate.
eTable 1. Definitions of study outcomes and baseline medical history
eTable 2. Patients with RRMS and general population comparators excluded for having had the outcomes of interest in the last 6 months before treatment start
eTable 3. Proportion with non-missing data before imputation, by cohort
eTable 4. Sequential adjustment for groups of potential confounders
eTable 5. Observed serious infections ordered by assigned main diagnosis
eTable 6. All observed inpatient serious infection diagnoses during the follow-up (including secondary diagnoses and multiple diagnoses per patient)
eTable 7. Sensitivity analysis, 90 days lag time
eTable 8. Distribution of herpes antiviral use in first 18 months of DMT treatment, among those with any herpes antivirals use and at least 18 months’ follow-up
Customize your JAMA Network experience by selecting one or more topics from the list below.
Identify all potential conflicts of interest that might be relevant to your comment.
Conflicts of interest comprise financial interests, activities, and relationships within the past 3 years including but not limited to employment, affiliation, grants or funding, consultancies, honoraria or payment, speaker's bureaus, stock ownership or options, expert testimony, royalties, donation of medical equipment, or patents planned, pending, or issued.
Err on the side of full disclosure.
If you have no conflicts of interest, check "No potential conflicts of interest" in the box below. The information will be posted with your response.
Not all submitted comments are published. Please see our commenting policy for details.
Luna G, Alping P, Burman J, et al. Infection Risks Among Patients With Multiple Sclerosis Treated With Fingolimod, Natalizumab, Rituximab, and Injectable Therapies. JAMA Neurol. 2020;77(2):184–191. doi:10.1001/jamaneurol.2019.3365
What is the risk of infections in association with different disease-modifying treatments for multiple sclerosis?
This nationwide cohort study found that patients with multiple sclerosis are at a generally increased risk of infections, and this risk is partly dependent on the choice of treatment. The rate of infections was lowest with injectable therapies; among newer treatments, use of rituximab was associated with the highest rate of serious infections but less use of herpes antiviral medications compared with fingolimod and natalizumab.
Per the results of this study, physicians and patients should be aware of infection risks associated with newer multiple sclerosis treatments and perhaps particularly anti-CD20 therapies.
Although highly effective disease-modifying therapies for multiple sclerosis (MS) have been associated with an increased risk of infections vs injectable therapies interferon beta and glatiramer acetate (GA), the magnitude of potential risk increase is not well established in real-world populations. Even less is known about infection risk associated with rituximab, which is extensively used off-label to treat MS in Sweden.
To examine the risk of serious infections associated with disease-modifying treatments for MS.
Design, Setting, and Participants
This nationwide register-based cohort study was conducted in Sweden from January 1, 2011, to December 31, 2017. National registers with prospective data collection from the public health care system were used. All Swedish patients with relapsing-remitting MS whose data were recorded in the Swedish MS register as initiating treatment with rituximab, natalizumab, fingolimod, or interferon beta and GA and an age-matched and sex-matched general population comparator cohort were included.
Treatment with rituximab, natalizumab, fingolimod, and interferon beta and GA.
Main Outcomes and Measures
Serious infections were defined as all infections resulting in hospitalization. Additional outcomes included outpatient treatment with antibiotic or herpes antiviral medications. Adjusted hazard ratios (HRs) were estimated in Cox regressions.
A total of 6421 patients (3260 taking rituximab, 1588 taking natalizumab, 1535 taking fingolimod, and 2217 taking interferon beta/GA) were included, plus a comparator cohort of 42 645 individuals. Among 6421 patients with 8600 treatment episodes, the mean (SD) age at treatment start ranged from 35.0 (10.1) years to 40.4 (10.6) years; 6186 patients were female. The crude rate of infections was higher in patients with MS taking interferon beta and GA than the general population (incidence rate, 8.9 [95% CI, 6.4-12.1] vs 5.2 [95% CI, 4.8-5.5] per 1000 person-years), and higher still in patients taking fingolimod (incidence rate, 14.3 [95% CI, 10.8-18.5] per 1000 person-years), natalizumab (incidence rate, 11.4 [95% CI, 8.3-15.3] per 1000 person-years), and rituximab (incidence rate, 19.7 [95% CI, 16.4-23.5] per 1000 person-years). After confounder adjustment, the rate remained significantly higher for rituximab (HR, 1.70 [95% CI, 1.11-2.61]) but not fingolimod (HR, 1.30 [95% CI, 0.84-2.03]) or natalizumab (HR, 1.12 [95% CI, 0.71-1.77]) compared with interferon beta and GA. In contrast, use of herpes antiviral drugs during rituximab treatment was similar to that of interferon beta and GA and lower than that of natalizumab (HR, 1.82 [1.34-2.46]) and fingolimod (HR, 1.71 [95% CI, 1.27-2.32]).
Conclusions and Relevance
Patients with MS are at a generally increased risk of infections, and this differs by treatment. The rate of infections was lowest with interferon beta and GA; among newer treatments, off-label use of rituximab was associated with the highest rate of serious infections. The different risk profiles should inform the risk-benefit assessments of these treatments.
Over the past decade, a diverse group of new disease-modifying therapies (DMTs) for multiple sclerosis (MS) treatment has entered the market.1,2 These treatments have different modes of action, but all modulate and interfere with the patient’s immune response, thereby raising concerns about adverse effects, such as an increased susceptibility to infections.
Regardless of treatment, patients with MS have been reported to have an increased risk of infections compared with the general population.3 Existing evidence suggests that treatment with first-generation, injectable DMTs interferon beta and glatiramer acetate does not increase this risk further.4,5 In contrast, the second-generation, so-called high-efficacy DMTs, including natalizumab and fingolimod, have been associated with increased risk of infections compared with placebo or interferon beta and glatiramer acetate in randomized clinical trials and postmarketing suveillance.4-7 Among specific infections, natalizumab increases the risk for life-threatening progressive multifocal leukoencephalopathy (PML).8 Herpes zoster may be a concern with several treatments but has been specifically associated with fingolimod.7,9,10 Most data on these risks stem from randomized clinical trials, however, with strict inclusion criteria, limited size, and limited follow-up time. Confirmation from large studies that include patients with MS who are representative of those treated in clinical practice are needed to evaluate whether this translates into meaningful differences in risk of infections overall. Although such studies remain rare, a recent cohort study in British Columbia, Canada, supported a difference in infection risks between first-generation and second-generation DMTs but was not powered to separate individual DMTs.11
Even fewer data are available on infection risks associated with the anti-CD20 antibody rituximab, which is extensively used off label to treat MS in Sweden. Although a retrospective review of Swedish patients with MS who were treated with rituximab concluded that there were no immediate safety concerns, patients did report a range of infections.12 Further assessment with proper comparator cohorts would be valuable, in particular since randomized clinical trials of ocrelizumab, an approved anti-CD20 therapy for MS, reported increased risk of respiratory tract infections compared with interferon beta treatment and placebo.13,14 To fill these gaps, we linked the Swedish MS register (SMSreg) to national health care and census registries to estimate and compare the infection risks among contemporary Swedish patients with MS who were treated with rituximab, natalizumab, fingolimod, and interferon beta and glatiramer acetate.
This nationwide cohort study included all patients with relapsing-remitting MS whose data were recorded in the SMSreg and who started treatment with interferon beta and glatiramer acetate, fingolimod, natalizumab, or rituximab between January 1, 2011, and December 31, 2017. The SMSreg covers approximately 80% of all patients with MS in Sweden,15 with high validity in recorded data on treatment (initiation and duration) and disease activity at the start of treatment.16
Patients could contribute to multiple treatment cohorts, entering each at their first start of each drug. To avoid immortal time bias attributable to retrospectively entered data in the MS register, treatments started 90 or more days before register inclusion were excluded. For each patient, Statistics Sweden matched 5 comparator participants who were free of MS from the general population by age, sex, and region, using risk-set sampling. Covariates and outcome data were added by linking the study population to national health and census registries using the national personal identity number.
The study was approved by the Regional Ethical Board of Stockholm. As is common for pseudonymized linkages of national registers in Sweden, the need for individual informed consent was waived.
Potential confounders were assessed at the start of treatment. Demographic factors included age, sex, country of birth, educational level, and proportion who received sick leave (of 2 or more weeks) and a disability pension in the last year from census registers maintained by Statistics Sweden. Medical history was assessed through diagnoses coded via the International Statistical Classification of Diseases and Related Health Problems, Tenth Revision (ICD-10) and recorded in the patient register, covering all inpatient and nonprimary outpatient care in Sweden (positive predictive value range of inpatient diagnoses, 85%-95%)17,18 and the prescribed drug register, with complete coverage of prescription medications dispensed at pharmacies in Sweden (including those prescribed at discharge but not hospital-administered drugs).19 Medical history was defined through diagnoses given or medications claimed (complete definitions in eTable 1 in the Supplement) in the 5 years before treatment start and included serious infections, antibiotic or herpes antiviral use, invasive cancer, antidepressant and antipsychotic medication use, cardiovascular diseases (ie, arrhythmia and major adverse cardiovascular events), diabetes mellitus, and chronic obstructive pulmonary disease.
The MS register provided disease-specific variables: treating clinic, previous treatments with DMTs, disease duration, relapses in the preceding year, and scores of the Expanded Disability Status Scale (EDSS), Multiple Sclerosis Impact Scale (MSIS-29), Symbol Digit Modalities Test (SDMT), and EuroQol 5-Dimension (EQ-5D) scale. The EDSS scores recorded within 180 days before and 15 days after the treatment start were considered baseline measures; for the MSIS-29, SDMT, and EQ-5D scales, scores 30 days before and 15 days after the treatment start were used.
The main outcome was the time until the first serious infection, defined as any infection recorded as the main reason for a hospitalization (eTable 1 in the Supplement). Hospitalizations for herpes and PML were also analyzed separately. Less serious infections were identified through the Prescribed Drug Register according to filled prescriptions of any systemic antibiotic (Anatomical Therapeutic Chemical Classification System chapter J01) and antiviral medication for herpetic infections (ie, acyclovir, famciclovir, and valaciclovir).
Participants were considered at risk for infections from the start date of treatment (for the matched general population participants, the date of the matched patient’s treatment start was used) until the first of (1) 90 days after recorded discontinuation of therapy, (2) death, (3) emigration from Sweden, (4) the outcome of interest, and (5) the end of follow-up on December 31, 2017. Comparator participants were censored if they later developed MS (ie, at their date of inclusion in the SMSreg). To exclude ongoing infections, participants with each end point in the 6 months before treatment start were dropped from analyses of that end point (eTable 2 in the Supplement).
Tabulations were made of baseline characteristics, number of events, and incidence rates for each cohort. Cox proportional hazard models were fitted to estimate hazard ratios, with time since treatment start date as the time scale. Since participants could contribute data to multiple cohorts, robust (sandwich estimator) 95% CIs were calculated. Potential confounding variables adjusted for in the models were age, sex, country of birth (categorized as Nordic countries [Sweden, Norway, Denmark, Iceland, and Finland] or other countries), educational level, number of days hospitalized during the last 5 years (categorized as 0, 1-10, or >10), and other general health variables (Table 1), year of treatment start, disease duration, EDSS score, MSIS-29 score, SDMT score, and EQ-5D scale score. Continuous covariates were modeled with third-degree polynomials. Multiple imputation was applied to account for missing data, creating 25 imputed data sets using fully conditional specifications. Imputation models were made separately for each outcome and included all covariates (parametrized as in the analysis model), the event indicator, and the Nelson-Aalen estimator of the cumulative hazard.20 Parts of the data had previously been record reviewed and thus had fewer missing data16; an indicator for this was included in the imputation model. Trace plots were reviewed to ensure convergence after 10 burn-in iterations, and population statistics before and after imputation were compared. We used SAS version 9.4 (SAS Institute) for all statistical analyses.
To test for potential bias from early surveillance or prophylactic treatment among patients linked to the start of a new therapy, we tabulated the amount of herpes antiviral drugs used and ran a sensitivity analysis excluding the first 90 days after treatment start and all patients who had had an event in this period. Finally, the comparison between high-efficacy DMTs was adjusted for number of previous DMTs; interferon beta and glatiramer acetate was excluded because it was almost exclusively used as a first-line DMT.
We included a total of 8600 treatment episodes in 6421 patients, with a mean (SD) age at treatment start of 39.0 (10.7) years. The number of female patients was 6186 (71.9%). The treatment episodes included 2217 initiations of interferon beta and glatiramer acetate, 1535 of fingolimod, 1588 of natalizumab, and 3260 of rituximab. The total time receiving drugs with interferon beta and glatiramer acetate was 4688 person-years (mean, 2.1 years); with fingolimod, it was 4129 person-years (mean, 2.7 years), with natalizumab, 3969 person-years (mean, 2.5 years), and with rituximab, 6533 person-years (mean, 2.0 years).
Baseline characteristics of all treatment groups and the general population are presented in Table 1. The mean (SD) age was similar among patients starting interferon beta and glatiramer acetate, fingolimod, and rituximab (38.8 [9.6] to 40.4 [10.6] years), while patients initiating natalizumab were younger (35.0 [10.1] years). About 70% were female across all groups (range, 1045 of 1535 [68.1%] of patients taking fingolimod; 1631 of 2217 patients taking interferon beta and glatiramer acetate [73.6%]). The cohorts with MS were similar in sociodemographic factors. Patients starting interferon beta and glatiramer acetate had lower MS disability (mean [SD]: MSIS-29 psychological score, 2.1 [0.9]; EDSS score, 1.6 [1.3]), which was likely associated with the drug’s role as a first-line DMT, while patients starting natalizumab tended to have higher disability (mean [SD]: MSIS-29 psychological score, 2.5 [1.0]; EDSS score, 2.3 [1.5]) at treatment start, which together with their younger age (mean [SD], 35.0 [10.1] years; vs 40.1 [11.3] years among those taking interferon beta and glatiramer acetate) indicated a more active disease. There were no consistent differences in medical history or in the history of infections across treatment groups.
Compared with the general population, patients with MS were more often born in Nordic countries (patients with MS, 7675 of 8600 [89.2%]; general population, 35 615 of 42 645 [83.5%]). They also had a higher rate of sick leave use (patients with MS, 2074 of 7371 [28.1%]; general population, 3958 of 36 071 [11.0%]), antidepressant use (patients with MS, 2305 of 8600 [26.8%]; general population, 6682 of 42 645 [15.7%]), and infections (patients with MS, 354 of 8600 [4.1%]; general population, 994 of 42 645 [2.3%]) in the preceding 5 years.
Data were complete for medical history (by design) and virtually complete for demographic factors. For MS disease scales, the availability of data differed by therapy and variable (eTable 3 in the Supplement). Data availability depended on the collection of certain variables as part of nationwide monitoring studies for natalizumab and fingolimod (Immunomodulation and Multiple Sclerosis Epidemiology Study 1 [IMSE1]21 and Study 2 [IMSE2]),22 resulting in most patients having MSIS-29 physical scores (natalizumab, 1297 of 1588 patients [81.7%]; fingolimod, 1295 of 1535 patients [84.3%]), MSIS-29 psychological scores (natalizumab, 1296 of 1588 patients [81.6%]; fingolimod, 1295 of 1535 patients [84.4%]), and SDMT scales (natalizumab, 1304 of 1588 patients [82.1%]; fingolimod, 1250 of 1535 patients [81.4%]); more recently, patients taking rituximab also received these assessments (MSIS-29 physical and psychological scores, 1780 of 3260 patients [54.6%]; SDMT, 1734 of 3260 patients [53.2%]). Corresponding assessments for interferon beta and glatiramer acetate were lacking, and a small proportion of patients taking this drug had data on the MS scales (MSIS-29 physical and psychological scores, 335 of 2217 patients [15.1%]; SDMT, 290 of 2217 patients [13.1%]). The EDSS score is used in clinical practice and thus had similar proportions with data across therapies (ranging from 1147 of 2217 patients [51.8%] taking interferon beta and glatiramer acetate to 1056 of 1588 patients [66.5%] taking natalizumab).
The number of observed events, crude incidence rates, and adjusted hazard ratios are presented in Table 2. Before adjusting for differences in patient characteristics, the incidence rate of serious infections was similar among patients treated with interferon beta and glatiramer acetate (incidence rate, 8.9 [95% CI, 6.4-12.1] per 1000 person-years) and natalizumab (11.4 [95% CI, 8.3-15.3] per 1000 person-years), which was lower than the rate among patients treated with fingolimod (14.3 [95% CI, 10.8-18.5] per 1000 person-years) or rituximab (19.7 [95% CI, 16.4-23.5] per 1000 person-years); Kaplan-Meier plot in the Figure). This comparison was clearly confounded by age, adjustment for which led to increased differences between interferon beta and glatiramer acetate (reference) and the other DMTs (hazard ratios [HRs]: fingolimod, 1.81 [95% CI, 1.21-1.27]; natalizumab, 1.53 [95% CI, 0.99-2.35]; rituximab, 2.34 [95% CI, 1.65-3.33]), but differences were attenuated by adjustments for further covariates (sequential adjustment in eTable 4 in the Supplement). In the most adjusted model, only the difference between interferon beta and glatiramer acetate and rituximab remained statistically significant (HR, 1.70 [95% CI, 1.11-2.61])), although point estimates for fingolimod and natalizumab were still greater than 1.00 (HRs, 1.30 [95% CI, 0.84-2.03] and 1.12 [95% CI, 0.71-1.77], respectively). Comparing the newer DMTs with rituximab, the rate was 34% lower receiving natalizumab (HR, 0.66 [95% CI, 0.45-0.97]) and 23% lower receiving fingolimod (HR, 0.77 [95% CI, 0.54-1.09]); only the difference between natalizumab and rituximab reached statistical significance. Rituximab and natalizumab had the highest rate of antibiotics use (the rate being 15%-18% lower with fingolimod and interferon beta and glatiramer acetate compared with rituximab), a difference mostly unchanged by adjustment for potential confounders (HR, 0.86 [95% CI, 0.76-0.97] and HR, 0.81 [95% CI, 0.71-0.92], respectively).
The pattern was different for the use of herpes antiviral drugs, where rates were significantly higher with natalizumab and fingolimod than with rituximab and interferon beta and glatiramer acetate, a difference that remained after adjusting for patient characteristics (natalizumab: HR, 1.61 [1.21-2.13] and fingolimod: HR, 1.70 [95% CI, 1.30-2.24] compared with ritixumab). After adjustment, the rate of herpetic infections was similar when comparing rituximab with interferon beta and glatiramer acetate. Hospitalizations for herpes or herpes zoster were rare, with only 1 event among patients taking interferon beta and glatiramer acetate, 2 in those taking fingolimod, and 3 events among patients taking rituximab or natalizumab.
The specific observed serious infections are tabulated by 3-digit ICD-10 codes in eTable 5 and eTable 6 in the Supplement. There were 2 cases of PML, 1 occurring in a patient receiving fingolimod therapy and 1 in a patient taking rituximab. In both cases, patients had switched from natalizumab within 6 months before the diagnosis of PML. There were no deaths with an infection recorded as the underlying cause among the patients treated for MS.
The rate of all infections was substantially higher in all MS cohorts than in the general population, a difference that remained after adjustments (HR, 0.65 [95% CI, 0.47-0.89], comparing the general population vs patients taking interferon beta and glatiramer acetate). The difference was particularly pronounced for hospitalized infections, while the relative difference in prescription medication was slighter but still significant.
Excluding the first 90 days of therapy (to exclude early prophylaxis) did not notably change any group comparisons (eTable 7 in the Supplement). Long-term prophylactic use of herpes antiviral drugs appeared uncommon, with 168 of 200 (84.0%) of those with any antiviral use filling 3 or fewer prescriptions in the first 18 months (eTable 8 in the Supplement). The number of prior DMTs used appeared not to confound the comparison between therapies (eTable 4 in the Supplement).
This nationwide cohort study is, to our knowledge, the largest observational study investigating the risks for infection associated with different DMTs in patients with MS. It is the first study to compare infection risk during rituximab treatment of MS with the risk of treatment with other highly effective DMTs. These data provide further support for previous reports that patients with MS are at a generally increased risk of infections, as well as some support for an increased risk of infections in high-efficacy DMTs compared with injectable therapies. This study also highlights the potentially different risk profiles of rituximab vs fingolimod and natalizumab, with rituximab being associated with a higher risk of serious infections but lower risk of herpetic infections.
A previous Swedish study (covering 1969-2005) reported a rate of serious infections increased more than 4 times among patients with MS,3 whereas we found a risk increase of about 50% among patients with MS who were treated with interferon beta and glatiramer acetate compared with the matched general population sample. This difference in results is likely explained by changes in clinical practice and different age distribution in the studies. More studies are needed on the risk for infections linked to the disease of MS itself, and since the difference we observed between treatments was of the same magnitude as the difference between MS and the general population, it seems essential that such studies take the role of treatment into account.
Among the DMTs, rituximab was associated with the highest rate of serious infections and was the only DMT with a significantly increased rate of infections (compared with interferon beta and glatiramer acetate) in the most adjusted model. Previous data from randomized clinical trials5 and a recent large observational study11 have been consistent with an increased rate of serious infections in patients taking natalizumab and fingolimod as well. Although the differences did not reach significance, we note that the point estimates in our data for fingolimod and natalizumab were also in the direction of an increased rate of infection compared with interferon beta and glatiramer acetate. This was also true for outpatient use of antibiotics, where the difference between natalizumab and interferon beta and glatiramer acetate did reach significance. In light of previous evidence, we suggest that our data offer limited additional support for a difference between the first-generation and second-generation DMTs in infection risks, but carrying out further studies of risks in real-world patient populations would be valuable.
Results for rituximab are in line with studies in rheumatoid arthritis, where evidence suggests increased infection rate associated with rituximab and other biologic therapies.23,24 The limited magnitude of the risk increase may still be considered somewhat reassuring with respect to the off-label use of rituximab in MS, considering data on the anti-CD20 antibody ocrelizumab. The development of ocrelizumab was discontinued for rheumatoid arthritis, citing uncertainties regarding its infection safety profile,5,25 and 2 deaths attributable to pneumonia were recorded in the ocrelizumab trial in primary progressive MS.14
Use of antibiotics was much more common than serious infections, but the pattern of rates was similar: highest with rituximab, followed by natalizumab and fingolimod, and with patients with MS having an increased rate of antibiotic use compared with the general population. It should be recognized that the prescription of antibiotics could be sensitive to awareness and surveillance (eg, association with frequency of visits or overprescription for patients believed to be at particular risk). This is likely to bias the comparison to the general population, and we cannot exclude that it may influence high-efficacy DMTs differently.
We found very few herpetic infections (including varicella zoster reactivations) severe enough to require hospitalization, regardless of therapy. Given that this has previously been reported in association with fingolimod, including 2 fatal cases,7,9 one plausible explanation is that current treatment guidelines work well in mitigating these risks, rather than a true absence of difference in risk between therapies. There was, however, a substantial difference between treatment groups in prescribed herpes antivirals, where the rate was about 70% higher among patients taking fingolimod and natalizumab than taking rituximab or interferon beta and glatiramer acetate. These prescription patterns may be influenced by the prescriber or patient awareness and expectations of risks, but as shown in our sensitivity analyses, it is not likely that this pattern is attributable to differences in prophylactic use (eTables 7 and 8 in the Supplement). A biological explanation for the difference between fingolimod and natalizumab vs rituximab may be the degree by which the respective DMT affects antiviral T-cell responses.
Two cases of PML were recorded, one in a patient receiving treatment with fingolimod and the other in patient receiving rituximab. In both cases, the respective treatment episode was recently started and patients had switched from natalizumab. This is consistent with a subclinical carryover infection, diagnosed at occurrence of an immune reconstitution syndrome when natalizumab concentrations sink below the threshold for impeding T-cell migration across the blood-brain barrier, as previously reported.26 The low rate of PML infections is also consistent with the clinical practice in Sweden to regularly test human polyomavirus 2 (John Cunningham virus) serology on patients treated with natalizumab and switch treatment if patients convert to positive serostatus.27
A major strength of this study is the use of national registries, which enabled inclusion of almost all patients with MS in Sweden (avoiding the risk for selection bias). Swedish health care registries are mandatory, and they have a high level of completeness and validity. However, the national registries do not cover primary care, and they lack clinical data beyond assigned diagnoses or collected drugs. We will thus miss most minor infections and are limited in what data we can show for the identified serious infections. Although the national health care registries are virtually complete, there were nonnegligible missing data on MS-specific clinical characteristics at the time of treatment start in the SMSreg. In particular, the rate of missing data was high for the interferon beta and glatiramer acetate group, in which less than 20% had data on MSIS-29 and SDMT scores. This incompleteness was expected, since collection of these variables for that treatment group has not been part of clinical practice to date. Although multiple imputation made it possible to include interferon beta and glatiramer acetate in analyses adjusted for these variables, the adjustment is based on extrapolations, and comparisons with this group may suffer residual confounding from these factors. We note, however, that results were virtually unchanged in analysis unadjusted for these factors and that comparison between other groups does not have the same limitations. We also lacked data on several potential confounders (in particular, body mass index, smoking status, and whether patients underwent varicella vaccination).
In conclusion, the risk of serious infections among patients with MS who were treated with rituximab was higher than among patients with MS who were taking natalizumab and fingolimod, and this increased even more in comparisons with those taking interferon beta and glatiramer acetate. In contrast, the use of herpes antivirals in patients taking rituximab was similar to those taking interferon beta and glatiramer acetate and lower than in those taking natalizumab and fingolimod. However, no fatal cases were recorded, which in part may be attributed to the use of clinical guidelines to mitigate risks associated with varicella for fingolimod and John Cunningham virus with natalizumab, respectively. These findings should be considered in the risk-benefit assessment of MS therapies, and further monitoring is important to better assess long-term risks.
Accepted for Publication: August 16, 2019.
Corresponding Author: Thomas Frisell, PhD, Clinical Epidemiology Division, Eugeniahemmet T2, Karolinska University Hospital, Stockholm 171 76, Sweden (firstname.lastname@example.org).
Published Online: October 7, 2019. doi:10.1001/jamaneurol.2019.3365
Author Contributions: Dr Frisell had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Concept and design: Luna, Alping, Fogdell-Hahn, Langer-Gould, Salzer, Vrethem, Piehl, Frisell.
Acquisition, analysis, or interpretation of data: Luna, Alping, Burman, Fink, Gunnarsson, Hillert, Langer-Gould, Lycke, Nilsson, Salzer, Svenningsson, Vrethem, Olsson, Piehl, Frisell.
Drafting of the manuscript: Luna, Vrethem, Piehl, Frisell.
Critical revision of the manuscript for important intellectual content: All authors.
Statistical analysis: Luna, Langer-Gould, Frisell.
Obtained funding: Fogdell-Hahn, Langer-Gould, Olsson, Piehl, Frisell.
Administrative, technical, or material support: Luna, Alping, Fogdell-Hahn, Gunnarsson, Hillert, Lycke, Nilsson, Olsson, Piehl, Frisell.
Supervision: Fink, Lycke, Piehl, Frisell.
Conflict of Interest Disclosures: Dr Langer-Gould served as site principal investigator for 2 industry-sponsored randomized clinical trials (Roche and Biogen Idec). Dr Lycke has received travel support and/or lecture honoraria from Biogen, Novartis, Teva, Merck, and Genzyme/SanofiAventis; has served on scientific advisory boards for Almirall, Teva, Biogen, Novartis, Merck, and Genzyme/SanofiAventis; serves on the editorial board of the Acta Neurologica Scandinavica; and has received unconditional research grants from Biogen, Novartis, and Teva. Dr Nilsson has received travel support from Bayer Schering Pharma, Merck Serono, Biogen, and Genzyme (a Sanofi Company), honoraria for lectures and advisory board participation from Merck Serono and Genzyme, advisory board participation for Novartis and Roche, and lectures and advisory board participation for Biogen; she has also received unconditional grants from Biogen. Dr Salzer have received research support from SYNAPSYS and Interacoustics. Dr Vrethem has received unrestricted research grants from Novartis, honoraria for lectures from Genzyme, and honoraria for advisory boards from Roche and Novartis. Dr Piehl has received research grants from Biogen, Genzyme, Merck KGaA, and Novartis, and fees for serving as chair of data monitoring committees in clinical trials with Parexel. No other disclosures were reported.
Funding/Support: Research reported in this publication was funded through a Patient-Centered Outcomes Research Institute Award (grant MS-1511–33196) and funds from The Swedish Foundation for MS Research (Dr Frisell).
Role of the Funder/Sponsor: The funders had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.
Disclaimer: The statements presented in this publication are solely the responsibility of the authors and do not necessarily represent the views of the Patient-Centered Outcomes Research Institute (PCORI), its Board of Governors, or its Methodology Committee.
Create a personal account or sign in to: