Objective Little is known about disease-modifying treatments (DMTs) for multiple sclerosis (MS) and infection risk in clinical practice. We examined the association between DMTs and infection-related medical encounters.
Methods Using population-based administrative data from British Columbia, Canada, we identified MS cases and followed them from their first demyelinating event (1996–2013) until emigration, death or study end (December 2013). Associations between DMT exposure (by DMT generation or class) and infection-related physician or hospital claims were assessed using recurrent time-to-events models, adjusted for age, sex, socioeconomic status, index year and comorbidity count. Results were reported as adjusted HRs (aHRs).
Results Of 6793 MS cases, followed for 8.5 years (mean), 1716 (25.3%) were DMT exposed. Relative to no DMT, exposure to any first-generation DMT (beta-interferon or glatiramer acetate) was not associated with infection-related physician claims (aHR: 0.96; 95% CI 0.89 to 1.02), nor was exposure to these drug classes when assessed separately. Exposure to any second-generation DMT (oral DMT or natalizumab) was associated with an increased hazard of an infection-related physician claim (aHR: 1.47; 95% CI 1.16 to 1.85); when assessed individually, the association was significant for natalizumab (aHR: 1.59; 95% CI 1.19 to 2.11) but not the oral DMTs (aHR: 1.17; 95% CI 0.88 to 1.56). While no DMTs were associated with infection-related hospital claims, these hospitalisations were also uncommon.
Conclusion Exposure to first-generation DMTs was not associated with an altered infection risk. However, exposure to the second-generation DMTs was, with natalizumab associated with a 59% increased risk of an infection-related physician claim. Continued pharmacovigilance is warranted, including an investigation of the DMT-associated infection burden on patient outcomes.
Statistics from Altmetric.com
Multiple sclerosis (MS) is a chronic disease characterised by inflammation, demyelination and neurodegeneration. Although there is no cure, several disease-modifying treatments (DMTs) have been introduced that have been shown to reduce the risk of relapses and have a beneficial impact on imaging outcomes.1 2 These drugs target the inflammatory process in MS by modulating the immune system or by immunosuppression. One of the safety concerns raised with the DMTs has been the potential association for an increased risk of infection.3 However, most information about the adverse effects of these DMTs stems from short-term clinical trials, which are often limited by the strict inclusion criteria for participants, such as exclusion of those with comorbid conditions.1 Safety findings from clinical trials may not represent the ‘real life’ setting where a broader range of persons are exposed to a DMT for much longer periods of time. The aim of this study was to assess the association between the DMTs approved for use in MS and risk of infections in a population-based setting.
Data sources and cohort formation
We performed a retrospective cohort study using prospectively collected population-based health administrative data in British Columbia (BC), Canada. BC has a public health insurance programme in which all residents must enrol. This insurance programme covers all required medical services, diagnostic services and prescription drugs dispensed in the community or outpatient setting. These encounters with the healthcare system are routinely collected, capturing data on >99% of all BC residents. The health administrative data were accessed through Population Data BC, which is BC’s electronic repository of administrative data held by the British Columbia Ministry of Health, British Columbia Vital Statistics Agency, Data Stewardship Committee and other provincial agencies for research.4 Datasets included: Medical Service Plan Billing Information and the Discharge Abstract Database, which provided information on physician claims and hospital admissions; PharmaNet, which captured drug prescriptions filled; Census Geodata, which provided an estimate of socioeconomic status (SES); Registration and Premium Billing files, which confirmed residency in the province via documentation of registration days with the mandatory provincial healthcare plan; and Vital Statistics Deaths, which captured death dates. We used an algorithm to identify persons with MS (PwMS) that has been validated in two Canadian provinces and applied successfully across provinces with consistent performance and results.5 The algorithm required persons to have at least three MS-specific encounters, which could include a hospitalisation or physician claim with an International Classification of Diseases (ICD) code 340/G35 or a prescription for a MS DMT (online supplementary appendix table e-1a).5 The date of the first demyelinating disease-related event (online supplementary appendix table e-1a) was the index date, and PwMS were eligible for inclusion if they were 18 years or older and resident in BC for at least 5 years before this date (to ensure incident onset MS). All persons were followed from their index date until the earliest of emigration from BC, death or study end, which was 31 December 2013. As per data availability, the earliest possible index date was 1 April 1996.
Supplementary file 1
The primary study outcome was the hazard of infection-related physician claims following the index date, identified using the primary (most responsible) ICD-9/10 diagnostic code (online supplementary appendix table e-1b).6 Secondary outcomes included the hazard of infection-related hospital admissions and the hazard of physician claims according to the five most commonly recorded infections (online supplementary appendix table e-1c). These approaches have been used successfully before.6 7
To minimise the possible impact of repeated encounters for the same infection and overestimation of infection risk, the following steps were taken. Subsequent to an infection-related physician claim, a person could not be at risk of any other infection for 30 days. Therefore, the relevant person-time was discounted, that is, excluded from the follow-up time.8 Similarly, the time during an infection-related hospitalisation was not considered.
We identified all prescriptions filled for a DMT (online supplementary appendix table e-1a) and calculated follow-up time on each drug based on days’ supply and the dispensed quantity. Follow-up time during which no prescription for a DMT was filled was considered ‘unexposed’. A person’s DMT exposure status was allowed to change over time. No person filled a prescription for teriflunomide, alemtuzumab or the generic beta-interferon-1b (Extavia) during the study period; the former two agents were largely unavailable in Canada during the study period (eg, alemtuzumab was approved by Health Canada close to our study end date of December 2013).
We used the proportional means model for recurrent events with robust sandwich variance estimates to examine the association between DMT exposure and infection-related physician claims.9 This model allows multiple infection-related events for a person and takes into account dependence of events within individuals. The hazard of infection was assessed using three separate models, with the reference category being ‘no current DMT exposure’. Model 1 included ‘any current DMT exposure’ as the comparison group. Model 2 included current exposure by DMT generation, using two comparison groups: (1) first-generation (beta-interferon or glatiramer acetate) and (2) second generation (natalizumab, fingolimod or dimethyl fumarate). Model 3 included current exposure by DMT class, using four comparison groups: (1) any beta-interferon, (2) glatiramer acetate, (3) natalizumab and (4) an oral DMT (fingolimod or dimethyl fumarate; the limited number of cases precluded individual analyses of these drugs). Also, based on model 2, the hazard of infection for ‘exposure to any second-generation DMT’ was assessed with ‘exposure to any first-generation DMT’ as the reference category. Online supplementary appendix table e-1d shows each DMT by its designated generation. The associations between DMTs and the secondary outcomes were assessed using a similar approach (except that the oral DMTs were not assessed separately due to the limited number of outcomes and people exposed to these drugs; however, they were still included along with the second-generation DMTs, as outlined above).
All models were adjusted for sex, age (continuous) and SES (quintiles) at the index date, index year (continuous) and number of comorbidities (none, 1, 2 or more). Comorbidities were selected based on clinical relevance to MS,10 potential association with DMT use11 and possible effect on infection risk.12–16 These included: diabetes mellitus, malignancies, chronic lung disease, inflammatory bowel disease, rheumatoid arthritis and psoriasis and were identified by the relevant diagnostic code recorded at two or more physician claims or one hospitalisation (online supplementary appendix table e-1e). The number of comorbidities was included in all models as a time-dependent covariate. We opted not to adjust for corticosteroids as there is limited evidence that short courses of this drug, when used in the context of an MS relapse, alter infection risk.17 Furthermore, in other circumstances, corticosteroids may be prescribed when infections arise, such as pneumonia and septic shock.18 19 Adjusting for corticosteroids that were used to treat an infection could introduce a bias. We choose not to adjust for immunosuppressants as they can be used to treat the comorbid conditions already included in our models.
As a complementary analysis of the association between DMT and infection-related physician claims (ie, the primary outcome), we stratified by sex and by age, with age grouped as <45 and ≥45 years (based on the mean age of the cohort at the index date). Findings were reported as adjusted HRs (aHRs) and 95% CIs. Statistical analyses were performed using IBM SPSS Statistics V.22 and SAS V.9.4.
We identified 6793 PwMS, of whom 4999 (73.6%) were women (table 1). The mean age (SD) at the index date was 45.4 (13.3) years. PwMS were followed for up to 17.7 years from the index date. During the study period, 1716 (25.3%) persons filled at least one prescription for a DMT, and there were 6066 person-years on DMT. The characteristics at the index date according to DMT exposure are shown in the supplementary appendix table e-2; a minority (n=458, 6.7%) of PwMS were exposed to two or more different DMT groups. On average, those who were unexposed during the entire study period were older (mean (SD) age: 47.8 (13.5) vs 38.2 (9.4) years) and had more comorbidities at the index date compared with those who filled at least one DMT prescription. However, the exposed and unexposed groups were similar with respect to sex and SES (online supplementary appendix table e-2).
Infection-related physician claims
Exposure to any DMT was not associated with an altered hazard for an infection-related physician claim (aHR 0.97; 95% CI 0.91 to 1.04) relative to being unexposed (table 2). Similarly, exposure to any first-generation DMT (either examined as one group or for beta-interferon or glatiramer acetate separately) was not associated with a significantly altered hazard relative to no DMT exposure (table 2). Exposure to any second-generation drug (aHR 1.47; 95% CI 1.16 to 1.85), in particular natalizumab (aHR 1.59; 95% CI 1.19 to 2.11), was associated with an elevated HR as compared with no exposure. While the hazard was not significantly increased for the oral DMTs, these most recently approved drugs had the shortest person-years of exposure (table 2). When compared with the first-generation DMTs, the second-generation DMTs exhibited a 53% greater hazard of an infection-related physician claim (aHR 1.53; 95% CI 1.21 to 1.95).
Exposure to any DMT (vs no exposure) was not associated with a significantly altered hazard for an infection-related hospitalisation (aHR 0.98; 95% CI 0.77 to 1.26; table 3). Similarly, exposure to any first-generation or second-generation DMT, whether assessed as a group or by class (beta-interferon, glatiramer acetate or natalizumab) were not associated with a significantly altered hazard of infection-related hospitalisation relative to no DMT exposure, although the 95%CIs were wide, especially for the second-generation DMTs (table 3).
Physician claims for specific infections
Infections that resulted in the highest number of physician claims per 10 000 person-years were: upper respiratory tract infections (1938), urinary tract infections (1247), bronchitis (788), skin-related infections (760) and pneumonia (245). Exposure to any DMT or any first-generation DMT was associated with a lower hazard of pneumonia as compared with no DMT exposure (figure 1 and online supplementary appendix table e-3). Correspondingly, a decreased hazard of pneumonia was found for beta-interferon and for glatiramer acetate (vs no exposure), but the association only reached significance for beta-interferon (aHR 0.74; 95% CI 0.56 to 0.98). Exposure to a second-generation drug was associated with a 58% and 68% greater hazard of an upper respiratory infection when compared with no DMT exposure or exposure to any first-generation drug. Of the second-generation DMTs, only natalizumab was assessed separately, and relative to no DMT, was associated with a 77% higher hazard of an upper respiratory infection (aHR 1.77; 95% CI 1.26 to 2.49, figure 1 and online supplementary appendix table e-3). No associations were identified between any of the DMTs and infections of the urinary system, skin infections or bronchitis (figure 1 and online supplementary appendix table e-3).
Findings regarding the association between DMTs and infection-related physician claims were in a similar direction when stratified by age and sex (online supplementary appendix table e-4). Although the aHRs for natalizumab or any second-generation DMT were higher for the men than for the women, the 95% CIs were wide, and they overlapped.
We examined the relationship between the DMTs and infection-related medical encounters in a large population-based cohort of individuals with MS. Overall, exposure to any DMT, or any first-generation DMT (beta-interferon or glatiramer acetate), was not associated with an increased infection risk. However, exposure to a second-generation DMT (natalizumab or the oral drugs) was associated with a 47% increase in the risk of an infection-related physician claim; 59% for natalizumab alone. None of the DMTs were associated with an altered risk of infection-related hospitalisations.
We were also able to assess five common infection types. Interestingly, findings suggested a protective effect of the first-generation DMTs; beta-interferon in particular was associated with a 26% lower hazard of pneumonia. This might be explained by the involvement of interferon type I signalling pathway in the defence against viral infections.20 A previous study also found that beta-interferon exposure was associated with a lower risk of respiratory-related hospital events compared with no exposure.21 In contrast, the second-generation DMTs were associated with a 58% higher hazard of an upper respiratory tract infection, or 77% when natalizumab was examined alone. The underlying mechanisms for these associations are still unknown.22 No evidence of an association between the DMTs studied and urinary tract infections, skin infections or bronchitis was found. Although the pivotal clinical trials reported a greater risk of local injection site infections,23–30 we did not observe an increase in skin infections. It is possible that affected individuals may self-manage local injection site infections and do not seek direct medical care from their physician. Infections for which no physician was visited would not have contributed to the infection burden detectable in our current study.
The vast majority of individuals in this study were using a first-generation injectable DMT. Our observation of an overall lack of association between the first-generation DMTs and infection-related medical encounters concurs with that reported in the pivotal clinical trials.23–30 In contrast, although we observed a measurable increased risk of infection with the second-generation DMTs, in particular for natalizumab, none of the short-term clinical trials for these drugs (natalizumab, fingolimod or dimethyl fumarate) reported significant increased infection rates.31–36 While the number of people on these drugs in our study was modest, with the phase III randomised clinical trials often including more individuals, still, concerns regarding infection risk have been raised, including in an independent review of trial data,37 and within the relevant product monographs. Since these drugs were approved, much focus has been on the relatively rare, but serious, progressive multifocal leukoencephalopathy.38 Our study provides a much needed assessment of the risk of a broad range of infections that are managed in the primary and secondary care setting by using data derived from Canada’s comprehensive, universal healthcare system. We also greatly extend clinical trial findings23–36 and other observational studies21 39 by examining all infections during follow-up, rather than just the first infection,39 by considering a broad range of DMTs (rather than just one),21 23–26 28 31 34 35 39 by including comparative safety analyses between the first-generation and second-generation DMTs and by our extended follow-up, which averaged 8 years.23–36
In our population-based cohort, 25.3% of the people were ever exposed to a DMT. Most estimates of the proportion of people with MS on a DMT are derived from MS clinic-based cohorts, the specialised settings where the DMTs are commonly prescribed, and in a population-based cohort such as ours, this proportion can be lower.40 We assessed the potential for differences in risk by age and sex, which have often been overlooked in other studies. While reassuringly we did not observe any major (significant) effects of DMT on infection risk, further studies are warranted. Gaps in knowledge remain, including for specific age groups; for example, we observed relatively fewer older individuals receiving a DMT in our study, likely mirroring the exclusion of older people (eg, aged 55+ years) in clinical trials. Also, assessment of the risk in men was limited because of the low absolute number of men, as is typical in MS. Knowledge gaps also remain for DMT-related differences in infection severity. While we did not have a direct means of assessing infection-related severity, hospitalisations could be considered a proxy marker of severity. The risk estimates were close to one, meaning no difference between the groups, and none of them reached significance. Although it is possible that this was caused by the relatively low number of infections resulting in a hospitalisation, it may be that there truly is no association between DMT exposure and infection-related hospitalisations.
Our study has some limitations. The number of people exposed to the newer second-generation DMTs in our study was rather modest, which may have limited the power to detect differences in risk. Another potential limitation was that people exposed to a DMT, particularly a second-generation DMT, may have been followed more carefully by their physician, resulting in an overestimation of the infection risk. However, exposure to beta-interferon was associated with a 26% lower hazard of pneumonia relative to no exposure, even though individuals on beta-interferon may have undergone more surveillance. We did limit the possible impact of repeated encounters for the same infection by discounting the 30 days following an infection-related physician claim and the time during a hospital stay. In addition, people with a higher number of comorbidities may be less likely to start a DMT11 but may also have a higher risk of infection. While we adjusted for number of comorbidities using a time-dependent approach, it remains possible that residual confounding exists. Although we were unable to adjust for MS severity, our approach allowed us to assess infection risk at similar time points across individuals (ie, all relative to a person’s first demyelinating claim). Still, variation in progression between people with MS may have influenced our results. Individuals on a second-generation DMT might be expected to have a greater disease severity than those on a first-generation DMT. A greater disease severity may result in more frequent infections.
To the best of our knowledge, this represents one of the first studies to assess the association between DMTs and infection-related healthcare utilisation over time in a large population-based cohort of individuals with MS. Our findings showed that first-generation DMTs were not associated with an increased risk of infection-related healthcare use in PwMS, while exposure to the second-generation DMTs was associated with an increased risk of infection in a ‘real world’ setting. Our findings complement those of clinical trials and provide additional insight into the safety profile of the MS DMTs. Future studies are needed to assess sex-related and age-related differences, the potential for an individual’s disease trajectory (MS severity) to impact risk and the burden and long-term consequences of the observed altered risk of infection.
We would like to thank the BC Ministry of Health, BC Vital Statistics Agency and BC PharmaNet for approval and support with accessing provincial data and Population Data BC for facilitating approval and use of the data.
Contributors The corresponding author HT takes responsibility for the integrity of the data and the accuracy of the data analysis. The analyst JW and principal investigator HT had full access to the data. All authors contributed to the interpretation of the data. JW drafted the manuscript. All authors revised the manuscript and approved the final version to be published.
Funding This study was supported by the National Multiple Sclerosis Society (RG5063A4/1). The funding source had no involvement in the study design, the collection, analysis and interpretation of the data or in the decision to submit this article for publication.
Disclaimer All inferences, opinions and conclusions drawn in this publication are those of the authors, and do not reflect the opinions or policies of the data steward(s).
Competing interests FZ, EK and YZ report no disclosures. JMAW receives research funding from the Michael Smith Foundation for Health Research/The Koehle Family Foundation. CE receives research funding from the Canadian Institutes of Health Research, Saskatchewan Health Research Foundation and the Multiple Sclerosis Society of Canada. JDF receives research funding from the Canadian Institutes of Health Research, Multiple Sclerosis Society of Canada, National Multiple Sclerosis Society and the Dalhousie Medical Research Foundation, consultation and distribution royalties from MAPI Research Trust, as well as speaker honoraria and travel expenses from EMD Serono (2013 and 2014). RAM receives research funding from the Canadian Institutes of Health Research, Multiple Sclerosis Society of Canada, National Multiple Sclerosis Society, Research Manitoba, the Consortium of MS Centers, Crohn’s and Colitis Canada and the Waugh Family Chair in Multiple Sclerosis and has conducted clinical trials funded by Sanofi-Aventis. HT is the Canada Research Chair for Neuroepidemiology and Multiple Sclerosis. She has received research funding from the Multiple Sclerosis Society of Canada, the Michael Smith Foundation for Health Research Scholar, the National Multiple Sclerosis Society, the Canadian Institutes of Health Research and the UK MS Trust; speaker honoraria and/or travel expenses to attend conferences from the Consortium of MS Centres (2013), the National MS Society (2012, 2014, 2015 and 2016), Bayer Pharmaceuticals (2010), Teva Pharmaceuticals (2011), ECTRIMS (2011, 2012, 2013, 2014 and 2015), UK MS Trust (2011), the Chesapeake Health Education Program, US Veterans Affairs (2012), Novartis Canada (2012), Biogen Idec (2014) and American Academy of Neurology (2013, 2014, 2015 and 2016). Unless otherwise stated, all speaker honoraria are either declined or donated to an MS charity or used as an unrestricted grant for use by her research group.
Ethics approval The study was approved by the University of British Columbia’s Clinical Research Ethics Board (#H14-00448) and the provincial health agencies regulating the use of administrative data.
Provenance and peer review Not commissioned; externally peer reviewed.
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.