This study was a retrospective cohort analysis of patients with MS treated at a tertiary care center in Saudi Arabia. We included all adult patients (aged ≥18) who had a confirmed relapsing-remitting MS (RRMS) diagnosis who were seen during the study period (March 2016 to August 2024). This included both patients who initiated their first DMT during this period and patients who had started a DMT prior to 2016 but continued follow-up and treatment during the study period. Patients with a diagnosis of secondary progressive or primary progressive MS and who had never started a DMT were excluded from the analysis. A total of 826 MS patients met the inclusion criteria. The cohort was followed from the start of the first DMT until treatment discontinuation or the end of the study period. The study was approved by the relevant institutional ethics committee, with a waiver of informed consent due to the retrospective use of de-identified patient data.

Demographic and clinical baseline data were collected from medical records, including age, sex, BMI, comorbidities (e.g. diabetes, hypertension, thyroid disease, and depression), diagnosis date, family history, initial clinical presentation, treatment information (e.g. first DMT type, start and discontinuation dates, and reason for discontinuation). Neurologic disability was assessed using the EDSS. For analysis, EDSS scores at the last follow-up (i.e., at data cutoff) were categorized into three disability levels: mild (EDSS 0–3.5), moderate (EDSS 4.0–5.5), and severe (EDSS ≥6.0). These categories were selected to reflect clinically meaningful disability thresholds commonly used in MS research, distinguishing patients with mild disability or no impairment (scores 0–3.5), moderate disability (ambulatory without continuous assistance, EDSS in the mid-range), and severe disability (significant impairment requiring assistance or wheelchair, EDSS 6 and above) (10).

All recorded DMTs were tabulated for each patient, including the start date of the first DMT, discontinuation (stop) date, and the documented reason for discontinuation (if applicable). For the purpose of analysis, we classified all DMTs into two therapeutic categories: heDMTs vs. platform therapies. In line with prior literature, the platform therapies were defined to include the traditional first-line injectable DMTs (interferon-β1a/1b and glatiramer acetate) as well as oral agents of moderate efficacy such as dimethyl fumarate and teriflunomide. The interferon-β1a formulations used in our cohort included Avonex, Rebif, and Plegridy, while interferon-β1b was represented by Betaseron; in cases where the specific interferon type or regimen could not be determined from the medical records, patients were categorized under “interferon beta (non-specific).” heDMTs were defined to include the more potent DMTs, principally the monoclonal antibody treatments (natalizumab, alemtuzumab, ocrelizumab, ofatumumab, and the off-label anti-CD20 antibody rituximab) and other highly active agents like cladribine and fingolimod (10). For patients with multiple sequential DMTs, the first DMT initiated was identified and categorized as heDMT or platform according to the above definitions. Additionally, the calendar year of first DMT initiation was stratified into discrete time periods (“eras”) to reflect changes in therapy availability over time. Specifically, first DMT start dates were grouped into three eras (2008–2013, 2014–2018, and 2019–2024) for use in subsequent analyses, to account for the evolving treatment landscape and shifting prescribing practices across these periods.

Treatment persistence was defined as the time from initiation of the first DMT until its discontinuation for any reason. Patients were monitored for discontinuation of the first DMT through regular clinic visits and pharmacy records. If a patient's first DMT was stopped, the discontinuation date and the reason for discontinuation (as documented by the treating neurologist) were recorded. If a patient had no recorded discontinuation date for the first DMT, the patient was assumed to be still ongoing with that therapy at the time of data cut-off; such patients were considered still on their initial DMT as of the data extraction date and were censored at that date in the persistence analysis. In this study, August 2024 served as the data cut-off for censoring ongoing first DMT treatments.

As part of data cleaning, the reasons for first DMT discontinuation were standardized into broader categories. Free-text reasons in the medical records were mapped to a set of predefined groups agreed upon by the investigators. Specifically, any reason indicating lack of efficacy or disease breakthrough (e.g. “failure/non-response” or “clinical/radiological progression”) was classified under inefficacy. Reasons related to medication intolerance or adverse events (such as “side effects” or “poor tolerance”) were grouped as adverse effects. Patient-driven or elective reasons were also consolidated: instances of pregnancy or family planning leading to therapy interruption were noted as a distinct category given their clinical importance, while other patient choice reasons (e.g. “patient preference to stop or switch”) and external factors (including drug unavailability) were categorized under patient preference/other. If no specific reason was documented or the reason was unknown, it was classified as unknown. This recategorization ensured consistency in analysis of discontinuation causes.

All study data were extracted and verified by the research team, and quality checks were performed to resolve any inconsistencies (for example, clarifying ambiguous drug names or partial dates by reviewing source records). All variables were then coded according to the definitions above for use in subsequent analyses.

Descriptive statistics were used to summarize the baseline characteristics of the cohort. Categorical variables were presented as frequencies and percentages, while continuous variables were reported as medians with interquartile ranges (IQR). Differences between patients who initiated platform therapies vs. heDMTs were assessed using Pearson's Chi-squared test for categorical variables and the Wilcoxon rank sum test for continuous variables. Multivariable logistic regression analysis was performed to identify factors independently associated with the use of heDMT as the first-line treatment. Ordinal logistic regression was used to examine the association between baseline factors, including treatment category, and current disability status measured by the EDSS. Treatment persistence was evaluated using Kaplan-Meier survival analysis with group comparisons conducted via the log-rank test. A Cox proportional hazards model was used to estimate the hazard of treatment discontinuation, while competing-risks regression using the Fine–Gray method was applied to assess the subdistribution hazard ratios (sHR) for specific causes of discontinuation, including inefficacy, adverse events, and other reasons. Time from MS diagnosis to initiation of the first disease-modifying therapy was included as a continuous covariate in all multivariable models to account for potential confounding related to treatment timing. All analyses were conducted using RStudio (version 2024.9.1.394, Boston, MA, USA) with R version 4.4.2. A two-sided p-value of < 0.05 was considered statistically significant.

Among the 826 patients included in the study, the majority were female (67.4%), with a median age of 35.0 years (IQR = 30.0–42.0). Most patients had mild disability at the time of last follow-up, as measured by EDSS 0–3.5, accounting for 81.1% of the cohort. The distribution of BMI categories was relatively balanced, with 35.0% of patients classified as normal weight, 32.3% overweight, and 28.1% obese. Regarding comorbidities, 5.0% had diabetes mellitus and 4.7% had hypertension. A family history of MS was reported in 14.1% of cases, while 29.4% had unknown family history status. The most common initial clinical presentation was sensory symptoms (33.5%), followed by visual symptoms (23.8%). The median time from MS diagnosis to DMT initiation was 1.1 months (IQR = 0.0–12.2). Most patients initiated their first DMT in the 2019–2024 era (47.2%), followed by 2014–2018 (31.0%) and 2008–2013 (21.8%, Table 1).

Among patients under study, platform therapies were initiated in 496 patients, representing 60.0% of the total cohort, whereas heDMTs were initiated in 330 patients, accounting for 40.0% of the cohort. Among patients who received platform therapies (n = 496), the most frequently prescribed agent was interferon beta-1a, used in 321 patients (64.7%), followed by non-specific interferon beta formulations in 105 patients (21.2%). In the heDMT group (n = 330), ocrelizumab was the most used first-line agent, accounting for 186 patients (56.4%), followed by fingolimod in 65 (19.7%), and rituximab in 22 (6.7%, Figure 1).

Significant differences were observed between the platform and heDMT groups across several baseline characteristics. Patients receiving heDMTs were younger (median age 31.5 years, IQR 27.0–37.0) compared with those receiving platform therapies (median 38.0 years, IQR 32.0–44.0; p < 0.001). The prevalence of diabetes mellitus (2.7 vs. 6.5%; p = 0.015) and depression (10.6 vs. 17.0%; p = 0.011) was lower among patients on heDMTs. The time from diagnosis to therapy initiation was longer in the heDMT group (median 3.0 months, IQR 0.3–14.2) than in the platform group (median 0.0 months, IQR 0.0–11.3; p < 0.001). Differences were also noted in disability status (EDSS category; p = 0.006), with a higher proportion of patients with mild disability initiating heDMT (86.4 vs. 77.6%).

A significant shift in prescribing patterns was observed over time (p < 0.001). Most patients who received platform therapies initiated treatment in earlier periods (37.0% between 2008–2013 and 44.0% between 2014–2018), whereas only 19.0% began in the most recent era (2019–2024). In contrast, the majority of patients initiating heDMTs (81.4%) did so during 2019–2024, reflecting a clear temporal trend toward earlier adoption of heDMTs (Table 1).

In the multivariable logistic regression model, treatment initiation in more recent eras was independently associated with a higher likelihood of receiving heDMT as first-line treatment. Compared with patients who began therapy between 2008 and 2013, those initiating treatment from 2014 to 2018 had significantly greater odds of receiving heDMT (OR = 3.17; 95% CI, 1.41–8.13; p = 0.009), and the likelihood was markedly higher among patients treated between 2019 and 2024 (OR = 48.8; 95% CI, 21.4–128; p < 0.001). An initial clinical presentation with cerebellar symptoms was also independently associated with increased odds of initiating heDMT compared with sensory onset (OR = 2.51; 95% CI, 1.13–5.63; p = 0.025). No other variables, including sex, age, body mass index, comorbidities, or other clinical presentation, were significantly associated with first-line use of heDMT (Table 2).

In the multivariable ordinal logistic regression model evaluating factors associated with current EDSS disability status, starting on a heDMT was not associated with better or worse EDSS disability levels compared with starting on a platform therapy (OR = 1.08; 95% CI, 0.64–1.84; p = 0.767). Increasing age was independently associated with higher EDSS scores (OR = 1.08; 95% CI, 1.04–1.11; p < 0.001). Compared with underweight individuals, patients with normal (OR = 0.27; 95% CI, 0.09–0.86; p = 0.019), overweight (OR = 0.20; 95% CI, 0.07–0.66; p = 0.005), and obese BMI (OR = 0.18; 95% CI, 0.06–0.61; p = 0.003) categories had significantly lower odds of higher EDSS scores. Regarding clinical presentation, motor (OR = 3.93; 95% CI, 2.01–7.88; p < 0.001) and cerebellar symptoms (OR = 3.89; 95% CI, 1.71–8.78; p = 0.001) were strongly associated with higher disability categories compared with sensory onset (Table 3).

Kaplan–Meier analysis demonstrated a significantly higher probability of treatment persistence among patients who initiated heDMT compared to those who received platform therapy (log-rank p < 0.0001). At 12 months, the estimated treatment persistence rate was 91.7% (95% CI, 88.4%−95.2%) in the heDMT group vs. 83.4% (95% CI, 79.4%−87.5%) in the platform group. The survival curves diverged early and continued to separate over time, indicating sustained superior persistence among heDMT users throughout the follow-up period (Figure 2).

In the multivariable Cox proportional hazards model, initiation on heDMT was associated with a significantly lower hazard of treatment discontinuation compared to platform therapy (HR = 0.41, 95% CI, 0.30–0.57, p < 0.001). Increasing age was also associated with a lower risk of discontinuation (HR = 0.97, 95% CI, 0.96–0.99, p = 0.003). Additionally, the presence of thyroid disease was associated with a higher risk of discontinuation (HR = 1.71, 95% CI, 1.09–2.69, p = 0.020). No other variables were significantly associated with discontinuation risk (Table 4).

In the Fine–Gray competing-risks regression model evaluating specific causes of treatment discontinuation, initiation with heDMT was associated with significantly lower sub distribution hazard ratios for discontinuation due to inefficacy (sHR = 0.12, 95% CI, 0.05–0.32, p < 0.001) and adverse events (sHR = 0.22, 95% CI, 0.12–0.38, p < 0.001) when compared to platform therapies. In contrast, heDMT was associated with a significantly higher hazard of discontinuation due to other reasons, such as pregnancy, preference, or supply issues (sHR = 3.46, 95% CI, 2.68–4.46, p < 0.001). Increasing age was independently associated with a lower hazard of discontinuation due to adverse events (sHR = 0.97; 95% CI, 0.95–1.00; p = 0.030) and other causes (sHR = 0.99; 95% CI, 0.97–1.00; p = 0.042). Additionally, patients with an initial motor presentation were less likely to discontinue therapy for other reasons (sHR = 0.69; 95% CI, 0.50–0.96; p = 0.026; Table 5).

Despite the strengths of a large, well-defined cohort, several limitations should be acknowledged. First, the retrospective, single-center design limits generalizability, as patients treated in tertiary referral settings often have higher disease activity and/or prior treatment failure. This may have inflated the observed rate of first-line heDMT use relative to the broader population. Second, reliance on routine clinical documentation introduces potential variability in data quality, and we lacked complete information on adherence and socioeconomic factors, which may have influenced treatment selection and outcomes. Disability was assessed using the EDSS; although standardized and widely used, the EDSS is weighted toward ambulatory function and is less sensitive to cognitive impairment, fatigue, and quality-of-life domains. As with all observational studies, causal relationships between treatment category and outcomes cannot be established, and residual confounding remains possible. In addition, we did not perform a prespecified subgroup analysis comparing early vs. delayed initiation of heDMT, and confounding related to treatment timing cannot be fully excluded. Pregnancy-related discontinuation patterns may also have been affected by evolving evidence and changing guidance regarding the safety of certain high-efficacy agents during pregnancy over the study period, limiting direct comparisons with contemporary international practice. Finally, treatment choice may have been shaped by system-level factors, including centralized drug procurement in a publicly funded healthcare system. The predominance of ocrelizumab within the heDMT group may also have introduced treatment-mix imbalance, such that observed outcomes largely reflect one agent rather than the entire heDMT class. Accordingly, findings should be interpreted at the class level and may not be generalizable to settings with different prescribing practices or for agent-specific comparisons.

Future research should extend these findings through multicenter prospective designs and stronger integration with national MS registry platforms. Broader geographic sampling would help capture variations in prescribing behavior between tertiary, secondary, and peripheral centers, as well as between public and private health systems. Longitudinal followup- is essential to evaluate durability of efficacy, cognitive trajectories, and the long-term safety of early heDMT use, including the impact of early vs. delayed initiation on disability progression and treatment persistence. Incorporating biomarkers such as neurofilament light chain (NfL) and MRI-based parameters could refine response prediction and guide therapy sequencing. Further, mixed methods work that captures patient perceptions, physician attitudes, and institutional barriers would add context to quantitative trends, helping design targeted interventions that encourage equitable- access to heDMTs nationwide.