**1. Introduction**

Multiple Sclerosis (MS) is an autoimmune inflammatory demyelinating condition that affects more than 2 million people worldwide [1,2]. A recent study estimates that in 2017, nearly 1 million adults had MS in the United States [1]. MS leads to an accumulation of disability over time, although disease-modifying therapies (DMT) may lessen long-term disability severity in most

patients [3]. MS is considered a heterogeneous disease thought to result from a complex interaction among genetic predisposition, sex, and environment [4]. Increasing evidence suggests that racial disparities are important factors that may explain di fferences in the disease course, prevalence, incidence, and outcomes [5–8]. Despite comprising 13.4% and 18.3% of the American population, African-Americans (AA) and Hispanics, respectively, remain largely underrepresented and understudied in clinical trials [9–11]. Fortunately, an accumulating body of work characterizes MS in diverse populations. This development could improve our understanding of disease course and epidemiology and uncover disparities across various racial/ethnic groups. Better understanding disparities in MS clinical course and outcomes will allow for the development of more e ffective disease managemen<sup>t</sup> in patients of diverse backgrounds.

Historically, it had been widely accepted that MS incidence was higher in the White population compared to the AA population [12]. However, population-based cohort studies have challenged this paradigm. A 2013 retrospective cohort study found that AA had a 47% increased risk of MS compared to Whites [13]. Disparities in MS clinical course in minority populations also encompass disability progression, disease burden, symptom presentation, and relapse rates. AA and Hispanics with MS have a higher disease burden and more severe disability in earlier stages of disease than White patients [10,14–16]. Additionally, AA patients commonly have multi-symptomatic presentation and early motor system involvement [14,17]. AA also experience inadequate recovery from symptoms and have shorter intervals between clinical attacks [7,8]. Furthermore, amongs<sup>t</sup> MS individuals admitted to US nursing homes, AA patients are younger and more disabled than White patients [18]. Studies comparing MRI findings between AA and White patients revealed that the former show an increased degree of T2 hyperintense lesions and T1 hypointense lesions, which correlate with greater MS-related disability [19].

Clinical data for MS in the Hispanic population is comparatively limited. The few studies on Hispanics sugges<sup>t</sup> a more rapid disability accumulation over time compared to White patients [20–22]. Interestingly, Hispanics were found to have a 50% decreased risk of developing MS compared to Whites [13]. However, several studies concur that Hispanics may have an earlier age of disease onset compared with other patient cohorts [13,20]. Hispanics and AA with MS are less likely than their White counterparts to visit a neurologist or MS specialist for disease managemen<sup>t</sup> and have decreased rates of DMT usage due to noncompliance or inappropriate understanding of the treatment plan [23,24]. DMTs are critical for e ffective managemen<sup>t</sup> and reduction of long-term disability in MS patients. In assessing these data, it is essential to consider that the Hispanic population is multiethnic and diverse. Other compounding factors that should be considered include socioeconomic status, place of birth, age of migration to the US, health literacy, systemic biases and systematic racism in healthcare, and access to care [5,20,25].

Much of our understanding of MS manifestation and clinical course in minority populations have come from a limited set of studies. Clinical trials on DMTs mostly lack data for minorities despite mounting evidence that these groups are at higher risk for a more aggressive disease course [26]. Approximately only 1% of the MS literature focuses on minority populations [10]. The purpose of this study was to address this lack of information by describing the clinical presentation, MRI findings, treatment regimens, disability progression, and relapse patterns in a racially and ethnically diverse population of MS patients in Houston, Texas. Given that the data for this study were collected from a clinic that predominantly serves patients of low socioeconomic status (SES), this study captures ethnic and racial disparities in MS among patients with a similar SES, potentially decreasing the possible effects of confounding factors. This study is critical and timely because it adds to an emerging literature that explores disparity in MS disease progression in AA and Hispanic MS patients compared to their White counterparts.

### **2. Patients and Methods**

### *2.1. Study Design and Setting*

Subjects were identified by a retrospective chart review of patients treated at the Smith Clinic Multiple Sclerosis Center. Smith clinic is a unique center that is part of a network that specifically cares for underserved and low socio-economic groups in Harris County, which includes the city of Houston. Additionally, Harris County is the third most populous county in the US. The majority of the patient population seen in the clinic are Non-Hispanic Black (NH-Black) or of Hispanic descent, and Mexicans constitute the majority of the Hispanic population served at the clinic. There is also a small percentage of Non-Hispanic White (NH-White) patients seen in the clinic. For the purposes of this study we are using the terms NH-Black and NH-White to account for the racial diversity of Hispanics seen in our clinic. Patients are attended to irrespective of insurance status or ability to pay.

### *2.2. Cohort Identification and Selection*

Information from all patients who visited Smith clinic from March 2019 to March 2020 was identified through chart review and included in this retrospective study. All patients with a diagnosis of Relapsing Remitting MS (RRMS), Secondary Progressive MS (SPMS), or Primary Progressive MS (PPMS) were included.

### *2.3. Outcome Measurements*

The following pre-selected information was abstracted for each patient: year and age of first symptoms, age at diagnosis, the amount of time that elapsed between onset of symptoms and diagnosis, disease subtype, estimated Expanded Disability Status Scale (EDSS) at diagnosis and last encounter, Disease Modifying Therapy (DMT) history (adverse reactions, relapses, and changes in immunomodulatory therapy), radiological findings, number of clinical relapses, smoking status, and autoimmune comorbidities. Escalation therapies included Glatiramer Acetate, Interferons, Teriflunomide, Dimethyl Fumarate and Fingolimod. High efficacy therapies included Rituximab, Ocrelizumab, Alemtuzumab and Natalizumab. Symptoms at disease onset were recorded and included motor, sensory, cerebellar, brainstem, bowel, and bladder function among others.

### *2.4. Data Collection and Management*

Two neurologists extracted patient data from medical records and the study protocol was approved by an Institutional Review Board. Information from the most recent clinical encounter and from the clinical encounter at diagnosis was included. The EDSS at presentation was estimated based on the first documented neurologic examination by a neurologist and was not indicative of the maximal neurologic deficit during the demyelinating episode that led to the diagnosis. Severe disability was defined as an estimated EDSS score > 4.5. MRI interpretations were collected from radiology reports. Lesion quantification and atrophy scoring were extracted directly from radiology reports and raw images were not independently interpreted by the neurologists gathering the data. A relapse was defined as a new, documented, neurological complaint lasting more than 24 h with objective findings in the documented neurological exam, or a follow-up MRI showing new enhancing lesions.

### *2.5. Statistical Analysis*

The statistical analyses were performed using R (version 3·6·1, Vienna, Austria) and RStudio (Version 1·2·5001, Boston, MA, USA). Based on the race and ethnicity information of the patients, we created a composite variable called 'race/ethnicity' and categorized the responses as Non-Hispanic (NH) White, NH-Black, Hispanic and 'others'. We conducted descriptive statistics on patient socio-demographic and disease characteristics stratified by race/ethnicity. We conducted Fisher's exact tests (for categorical variables) and ANOVA (for continuous variables). We examined the usage

and impact of DMTs across racial/ethnic groups. We also examined various markers of disease progression including lesions and atrophy in the brain as well as the thoracic and cervical spine stratified by race/ethnicity using Fisher's exact test. Applying adjusted Exact logistic regression models, we evaluated the association between various patient characteristics and a high EDSS score (EDSS > 4.5). Models were adjusted for di fferent covariates based on the literature and context, along with experts' recommendations. All analyses were based on two-tailed probabilities with a type 1 error rate set at 5%.
