Outcomes and Health Care Service Use in Adults 50 Years or Older With and Without Multiple Sclerosis: A 6-Year Observational Analysis

Multiple sclerosis (MS) is an inflammatory and degenerative disease of the central nervous system typically presenting in young adulthood (mean age at onset of approximately 34 years)^1–3; however, the prevalence of MS in older people is increasing,^4–6 with recent US data estimating the highest prevalence in 55- to 64-year-olds.⁷ This increased prevalence is most likely due to increased life expectancy in people with MS, which, in turn, has been attributed to the development of effective disease-modifying therapies (DMTs), earlier diagnosis of MS, and increased overall life expectancy in the general population.^4–6

It is not fully known how age-related changes in the immune system (immunosenescence and inflammaging) affect the MS disease process.^8–11 In addition, clinical trials typically exclude older people, who often experience age-related comorbidities, longer disease duration, and accumulated disability.^1,12 Previous studies have found that people with MS have an increased risk of infection compared with the general population, potentially as a consequence of MS disease or due to treatment with DMTs that interfere with the immune response.^13–15 In addition, some cancers may have a higher incidence in people with MS than in people without MS, possibly related to altered immune responses, inflammation, physical inactivity, or comorbidities that can increase the risk of cancer.^16–18 Furthermore, rates of infection and malignancy have been found to increase with age.^{10,16,19–22} Understanding clinical characteristics, outcomes, and health care service use in older people with MS represents an important step in improving disease management and care in this population.

In this real-world study, we assessed infections and malignancies, use of skilled nursing facilities (SNFs), use of MRI, and mobility impairment in people 50 years or older, both with and without MS, to investigate whether health-related events were primarily associated with MS or with aging itself.

METHODS

Study Design

This was a real-world, retrospective, longitudinal, observational analysis of administrative US claims data obtained from IBM-Truven MarketScan commercial and Medicare databases during the study period of 2011 to 2017. These databases contain health care claims data, deidentified to maintain Health Insurance Portability and Accountability Act compliance, from more than 60 million individuals across the US compiled from approximately 100 different payers (ie, entities that pay healthcare providers for the services they administer: commercial, private, and government-funded insurance providers). Approval of this research by an independent ethics committee or an institutional review board was not required because the study involved only secondary use of deidentified data.

Study Population

People were eligible for inclusion in the MS group if they had at least 3 MS diagnosis claims (International Classification of Diseases [ICD]-9 code 340 or ICD-10 code G35) within 1 year, or at least 1 MS diagnosis claim and at least 1 DMT claim within 1 year. The index date was the first MS diagnosis or DMT claim in 2012. The types of DMTs included in this study were the US Food and Drug Administration (FDA)–approved DMTs available during the study period and include self-injectable (ie, interferon beta-1a, interferon beta-1b, glatiramer acetate, and peginterferon beta-1a), infusion (ie, natalizumab, alemtuzumab, and ocrelizumab), and oral (ie, teriflunomide, fingolimod, and dimethyl fumarate) medications. Use of DMT was assessed through National Drug Code numbers for FDA-approved DMTs for MS available during the study period.

Continuous enrollment from 1 year pre-index (baseline) to at least 3 years postindex (follow-up) was required. People were eligible for inclusion in the general population group if they were continuously enrolled from 2011 to at least 2015 and had no previous documented diagnosis of MS; July 1, 2012 was used as the index date. For the primary analysis, eligible participants who were 50 years or older in 2012 were assigned to the aging MS cohort if they had MS and to the general population group if they had never been diagnosed as having MS. A subset of the general population group was then matched in a 1:1 ratio to the aging MS cohort using propensity scores calculated from a logistic regression model including baseline covariates (age, sex, region, health plan, comorbidities) and a greedy 5- to 1-digit matching algorithm without replacement to generate a group of matched controls.²³ The general population group was sufficiently large such that matches were found for each member of the aging MS cohort. To examine how MRI use changed with age, eligible matched pairs (ie, 1 aging MS cohort and 1 control cohort) were subsequently grouped into the following age groups: 50 to 59, 60 to 69, and 70 years or older.

Outcomes

Primary outcomes included the proportion of people with infections and malignancies and the proportion of people using SNFs or MRI between the study and control groups. The annualized rate of infections per person, the most frequent infections, and the infections of interest were recorded. Data on the most frequently reported malignancies and types of MRI performed were also obtained. Association between age group and MRI use was also examined. Secondary outcomes included time to cane/walker use and time to wheelchair use in both groups.

Analysis

Patient baseline demographic and clinical characteristics were evaluated. A propensity score matching algorithm was used to minimize selection bias and balance the aging MS cohort and the control group. Improvement in balance between cohorts after propensity score matching was assessed for each baseline characteristic using standardized mean differences with a threshold of 0.1 or less. Multivariable models were used to compare risk of infection, malignancy, SNF use, MRI use, and cane/walker or wheelchair use between the groups, adjusting for baseline covariates on age, sex, region, health plan types, and Charlson-Quan Comorbidity Index score.²⁴ Odds ratios were estimated from logistic regression models; rate and rate ratios were estimated from generalized linear models; and risk of use of mobility aids (cane/walker or wheelchair) was estimated from Cox proportional hazards models, which included adjustment for cane/walker or wheelchair use at baseline. The P values were calculated using the Wald χ² test; a critical value of 0.05 was specified as the threshold for statistical significance. The CIs were Wald 95% confidence limits. Analyses of infection, malignancy, and SNF and MRI rates were conducted using data from the follow-up period between 2012 and 2017.

Data Availability

The proprietary database used for this study was made available to EMD Serono through a license that limits dissemination of the database. Statistical coding used to conduct the analyses is proprietary and, as such, has not been made available.

RESULTS

Patient Demographics

Details of the population selected for the aging MS cohort and the general population group are shown in FIGURE 1. Overall, 10,746 people with MS were included in the aging MS cohort, and 3,521,326 people were eligible for inclusion in the general population group (Figure 1). Subsequently, 10,746 pairs were matched between the aging MS cohort and the general population group using propensity scores, which ensured that baseline characteristics were similar between groups (TABLES S1 AND S2). The mean ± SD follow-up time for this analysis was 5.1 ± 0.8 years in the aging MS cohort and 5.0 ± 0.7 years in the control group.

Population Selection

Infections and Malignancy

In people 50 years or older, infections occurred in more people with MS (aging MS cohort, 61%; controls, 45%; P < .0001) (FIGURE 2). Furthermore, the mean annualized infection rate was 3 times greater in the aging MS cohort (0.74 per person per year; 95% CI, 0.69–0.79) than in the control group (0.25 per person per year; 95% CI, 0.23–0.26) (rate ratio, 3.0; 95% CI, 2.8–3.1; P < .0001). The 3 most common infections in the aging MS cohort and the matched control group were urinary tract infections (40% and 19%; P < .0001), acute upper respiratory tract infections (25% and 24%; P = .7154), and herpes zoster infections (9% and 5%; P < .0001). Incidences of additional infections of interest were low in the aging MS cohort and matched controls (progressive multifocal leukoencephalopathy, <1% and 0%; pneumonia, <1% and <1%). The proportion of people 50 years or older who developed a malignancy was identical between both groups (20%; P = .8167) (Figure 2). Malignancy rates were also similar in the aging MS cohort vs matched controls when individually evaluating skin (8% vs 7%), breast (4% vs 5%), and prostate (2% vs 2%) cancers, the 3 most common malignancies reported in this data set.

Infections, Malignancy, and Health Care Use

Health Care Use

In people 50 years or older, SNFs were used by 4 times as many people in the aging MS cohort vs the control group (12% vs 3%; P < .0001) (Figure 2). In addition, MRI was used by more than double the proportion of people with MS (aging MS cohort, 87%; matched controls, 37%; P < .0001) (Figure 2). The difference in MRI use between the cohorts remained significant when evaluating people aged 50 to 59 years (aging MS cohort, 90%; controls, 35%; P < .0001) and 60 to 69 years (aging MS cohort, 84%; controls, 40%; P < .0001) but not when evaluating people 70 years or older (aging MS cohort, 68%; controls, 44%; P = .8380). People with MS 50 years or older most frequently underwent brain (82%) and spine (62%) MRI.

Mobility Impairment

Time to first cane/walker or wheelchair use was shorter in people 50 years or older with MS than in those without MS (cane/walker use: HR, 2.1; 95% CI, 1.9–2.3; P < .0001; wheelchair use: HR, 6.9; 95% CI, 6.0–8.1; P < .0001) (FIGURE S1).

DISCUSSION

The analysis of US administrative claims data from more than 20,000 people found that infections and health care use related to SNFs and MRI were more common and mobility aids were used sooner in people with MS 50 years or older than in people 50 years or older without MS. The present findings suggest that measures typically associated with worse health were primarily associated with having MS rather than being a consequence of aging alone.

Analysis of the IBM-Truven MarketScan commercial and Medicare databases between 2011 and 2017 identified 10,746 people with MS 50 years or older. Consistent with other analyses, the large number of older people with MS identified in these US databases alone demonstrates that although MS typically presents in young adulthood,^1,3 there are many people with MS 50 years or older who will require care tailored to their specific needs.^1,7,12 For example, the risk-benefit ratio of a given DMT may change in older people as the risk of adverse events may increase.^10,19,25,26 In addition, data suggest that signs of immunosenescence, such as decreased thymic output of naive T cells and increased infection rates, begin at age 50 years, factors that are unique to this population and highlight the value of characterizing MS in older people.^{8–10,19,27,28}

In the present study, we found that infections were more frequent in the aging MS cohort than in the control group, suggesting that having MS may increase the likelihood of infection in older people. A previous study of data from Swedish national registers found that people with MS had a 2.5 times increased risk of a serious infection and a 1.6 times increased risk of a nonserious infection than people without MS.²⁹ An additional study of data from electronic medical databases found that people with MS were 1.8 times more likely to have any infection in the United States and 1.3 times more likely in the United Kingdom, with a 2.0- to 2.4-fold higher rate of hospitalized infections in people with MS than in the general population.¹⁵ The increased risk of infection in people with MS might be related to the functional limitations related to long disease duration and advanced disability, such as neurogenic bladder and dysphagia.³⁰ Furthermore, reviews of data from clinical trials, the overall MS population, and administrative claims databases have found that the risk of infections may increase with age.^19–21,31 This is particularly relevant in light of the current COVID-19 pandemic because increased age is also a risk factor for worse outcomes with SARSCoV2 infection.^32,33

The higher rate of infections in people with MS in the present study was primarily driven by urinary tract and herpes zoster infections. These results are consistent with previous findings that these 2 infections are more common in people with MS than in those without MS and that bladder dysfunction and reactivation of the herpes simplex virus increase with increasing age in MS.^1,15,19 Specifically, incidences of urinary tract or kidney infections have been found to be nearly 2-fold higher, and incidences of opportunistic infection, candidiasis, and any herpes virus are 20% to 52% higher in people with MS compared with non-MS patients.¹⁵ The likely explanation for the higher rate of urinary tract infections is the higher prevalence of urinary disorders and bladder dysfunction in people with MS and is not necessarily due to treatment with DMTs.^29,34,35 At the time of data collection, the only available herpes zoster vaccine contained a live attenuated virus and, thus, although recommended for adults 60 years or older, was not suitable for people with MS³⁶; however, a recombinant herpes zoster vaccine is now approved and recommended for immunocompetent adults 50 years or older, including people with MS.³⁷ Considering the current COVID-19 pandemic, it is reassuring that this analysis showed that acute respiratory tract infections and pneumonia occurred in similar proportions of people in both cohorts. Overall, the present findings suggest that clinicians should be vigilant in monitoring older people with MS for potential infections. This is consistent with recent recommendations that suggest that clinicians should review vaccination status and screen for infection at MS diagnosis and that before initiating a new DMT, potential infection risk should be evaluated with a view to mitigating any risk.^30,35 Appropriate testing and strict monitoring of clinical signs and symptoms should be conducted throughout the MS disease course to ensure early diagnosis and prompt initiation of treatment.³⁵

In general, the risk of malignancy increases with age, with two-thirds of all cancer cases diagnosed in people 50 years or older.³⁸ Studies in people with MS have also suggested that risk of malignancy may increase with age in this population.^16,19,39 Analyses of the incidence of malignancies in people with MS compared with the general population has shown conflicting results. Most studies have found that the overall risk of malignancy is generally lower in people with MS than in the general population, with other studies identifying no difference between cohorts^16,17,39,40; however, other studies have suggested that the risk of certain malignancies (eg, respiratory, urinary, central nervous system) may be higher in people with MS than in people without MS.^16–18,40 This increased rate of malignancies in people with MS could be related to genetic or environmental factors (eg, smoking, chronic bladder and kidney irritation due to micturition problems and urinary infections, chronic neurologic inflammation) or may be a consequence of frequent MRI (eg, of the central nervous system).^17,18

The present analysis identified no overall difference in diagnosis of malignancy between people with MS and the general population, as well as for the 3 most commonly reported malignancies in the data set: skin, breast, and prostate. These findings echo the results of another recent retrospective matched cohort study that evaluated population-based administrative data in Canada.¹⁶ Marrie et al¹⁶ found that although incidences of breast and colorectal cancers increased with age, rates did not differ between people with or without MS in any age group.

Although the present results support most of the recent data in the literature in suggesting no difference in the incidence of malignancies between people with and without MS, it is still important to monitor people with MS for malignancies because they may experience numerous barriers to screening. Multiple sclerosis may cause severe disability that limits access to preventive care such as cancer screenings, which could delay cancer diagnosis.^41,42 Barriers to screening often include issues related to mobility impairment, such as transportation or lack of adaptive equipment within the health care setting.^41,43 Additional barriers include those related to providers, such as negative attitudes and lack of referrals, and personal barriers, such as fear, discomfort, and “having enough to handle.”⁴¹

In addition to infection and malignancy, we also evaluated mobility aid and SNF use in older people with MS to interrogate whether they were related to MS or to aging and its comorbidities. The present data demonstrated that people in the aging MS cohort were quicker to require a cane/walker or a wheelchair than matched controls. Similarly, more people in the aging MS cohort used SNFs than in the control group. The earlier use of mobility aids and the higher use of SNFs in older people with MS suggest that disability is due to MS and not merely due to aging and the associated comorbidities and highlights the need to further investigate optimal treatment strategies for this patient population; individualized assessment of treatment risk-benefit profiles will be paramount.

MRI is used to monitor disease progression and response to DMTs in people with MS. Thus, it was not unexpected that MRI was used more frequently by people with MS, or that the brain and spine were the most frequent types of MRI. However, in people 70 years or older, the proportion using MRI was no longer significantly different between people with MS and the control group. This seemed to be a result of MRI use declining with increasing age in people with MS and increasing with increasing age in matched controls. These data suggest that disease monitoring reduces as people with MS get older, possibly due to the discontinuation of DMT.²⁵ Reduced disease monitoring results in a reduced understanding of the disease pathophysiology in this oldest subset of people with MS, potentially reducing the opportunity to predict therapeutic responses and improve prognosis in this population.⁴⁴

A limitation of the present analysis is the evaluation of people with MS as a single group; due to a lack of information available in the databases, the effects of certain factors were not studied (ie, race/ethnicity, socioeconomic status, mental health status, or geographic location), all of which could influence outcomes and health care use. Similarly, we were unable to consider differences in health behaviors (ie, smoking, diet, physical activity) between the aging MS cohort and the control group, or the effect of different MS disease courses (eg, relapsing-remitting vs secondary progressive). One further limitation of claims analyses in general is that the presence of a diagnosis code on a medical claim does not guarantee the presence of disease because the diagnosis code may have been incorrectly coded or included as rule-out criteria. In addition, the initial diagnosis or claim may have happened before the patient was captured in the system, which could lead to a shorter disease duration or the incorrect categorization of patient disability stage. Finally, given the large sample size, statistically significant differences between groups should be interpreted with caution, with a focus on clinically meaningful differences.

Despite these limitations, we believe that the present analysis of more than 10,000 Americans with MS is a valuable initial contribution to the pool of knowledge available on the differential effects of MS and age on individual outcomes and health care use. Indeed, the present analysis of administrative US claims data from the IBM-Truven MarketScan commercial and Medicare databases between 2011 and 2017 identified 259,773 people with MS overall, representing a substantial proportion (approximately 35%) of the adult MS population in the United States according to recent estimates.⁷

For the scope of this article, we conducted a retrospective analysis of administrative claims data on people with and without MS 50 years and older that focused on a few key areas (infections and malignancies, health care service use, and mobility impairment); however, more data and future studies are needed to assess the long-term consequences of living with MS and the use of DMTs in older adults. Comorbidities, sedentary lifestyle, obesity, and frequent medication use are all common in people with MS⁴⁵ and are likely to have cumulative effects on health and disability that emerge with increasing age. In addition, because the goal of this analysis was to investigate whether health-related events were primarily associated with MS or with aging itself, it was not designed to evaluate the potential impact of any DMT on outcomes. Future studies through the Multiple Sclerosis Leadership and Innovation Network (MS-LINK) consortium should aim to more completely characterize disease management and care in older people with MS.

Overall, these insights into infections, malignancies, health care use, and mobility impairment in older adults with and without MS will contribute to optimizing care in the understudied and increasingly prevalent population of older people with MS.^1,7,12

PRACTICE POINTS

Measures typically associated with worse health are primarily associated with having multiple sclerosis (MS) rather than being a consequence of aging alone in people 50 years or older.

In people with MS 50 years or older, infections and use of skilled nursing facilities and MRI were more common, and mobility aids were used sooner than in peers without MS.

Insights into health-related events in older adults will contribute to optimizing care in this understudied and increasingly prevalent population of people with MS.