Association of Disease-Modifying Therapies with COVID-19 Susceptibility and Severity in Patients with Multiple Sclerosis: A Systematic Review and Network Meta-Analysis

Background We conducted this study to assess the effect of disease-modifying therapies (DMTs) on coronavirus disease (COVID-19) susceptibility and severity in people with multiple sclerosis (MS). Methods Available studies from PubMed, Scopus, EMBASE, Web of Science, and gray literature, including reference lists and conference abstracts, were searched from December 1, 2019, to July 26, 2021. We included cross-sectional, case-control, and cohort studies assessing the association of DMTs with risk of contracting COVID-19 or its outcomes in MS patients on univariate or multivariate regression analyses. We conducted a network meta-analysis (NMA) to compare the risk of COVID-19 and developing severe infection across DMTs. Results Out of the initial 3893 records and 1883 conference abstracts, a total of 10 studies were included. Pairwise comparisons showed that none of the DMTs meaningfully affect the risk of acquiring infection. There was significant total heterogeneity and inconsistency across this NMA. In comparison with no DMT, dimethyl fumarate (0.62 (0.42, 0.93)), fingolimod (0.55 (0.32, 0.94)), natalizumab (0.50 (0.31, 0.81)), and interferon (0.42 (0.22, 0.79)) were associated with a decreased risk of severe COVID-19; but, rituximab was observed to increase the risk (1.94 (1.20, 3.12)). Compared to rituximab or ocrelizumab, all DMTs were associated with a decreased risk. Pairwise comparisons showed no differences across other DMTs. Interferon and rituximab were associated with the lowest and highest risks of severe COVID-19. Conclusion Our study showed an increased risk of severe COVID-19 in patients on rituximab and ocrelizumab. No association with COVID-19 severity across other DMTs was observed.


Introduction
The outbreak of Coronavirus Disease-2019 (COVID- 19), caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), has led to a newly emerging pandemic. This globally spreading virus affects people in different ways, with manifestations ranging from no symptoms to hospitalization and death due to acute respiratory distress syndrome (ARDS) [1,2]. More than one year after the outbreak of COVID-19, the number of reported COVID-19 cases exceeds 150 million, with more than 3.5 million deaths [3].
Multiple sclerosis (MS) is one of the most common demyelinating diseases in the central nervous system (CNS), affecting generally young female adults. MS patients often receive immunosuppressive agents which put those at greater risk of developing viral and bacterial infections [4][5][6]. This raised a question regarding whether people living with MS were at higher risk of COVID-19 and were more likely to develop severe symptoms when infected than the general population. A recent systematic review has suggested a mortality rate of 3.5% among MS patients considering suspected/confirmed COVID-19 cases, which is slightly higher than the rate of 2.2% among the general population [2,7]. This study showed that patients on anti-CD20 agents had highest rates of hospitalization and mortality than those on other DMTs. Moreover, studies suggested an increased risk of developing the infection in MS patients on anti-CD20 agents [8,9].
Knowledge of the association between disease-modifying therapies (DMTs) and COVID-19 susceptibility/severity is necessary to provide the best care for patients during the pandemic and could be important for policymakers to adopt vaccine strategies. However, the current evidence is inconsistent and unclear. Therefore, this study was conducted to present the current evidence regarding the effect of DMTs on COVID-19 susceptibility and severity in people living with MS.

Method
2.1. Inclusion and Exclusion Criteria. Studies were included according to the following criteria: population (participants), outcomes, and study types. The population (participants) consists of suspected or confirmed COVID-19 patients with a previous diagnosis of MS. Outcomes are the association of each specific DMT with COVID-19 susceptibility and outcomes reported based on univariate or multivariate regression analyses. The included study types are cross-sectional, case-control, and cohort studies. Studies with the following characteristics were excluded: (a) studies did not compare DMTs with each other; (b) studies pooled DMTs based on the mechanism of action (immune cell depleting medications or immune-cell trafficking inhibitors) or risk of systemic infection (no risk, mild, risk, or high risk); (c) nonpeer-reviewed articles; (d) non-English studies; (e) review articles and systematic review; and (f) qualitative studies. ple Sclerosis, Acute Fulminating)). We also screened the reference lists of identified articles, review studies, or other relevant documents for inclusion in the study. In addition, we also searched the online library and abstracts of the following congresses: 8 th American and European Committees for Treatment and Research in Multiple Sclerosis  identify the eligible studies. Then, the full text of the potentially eligible studies was reviewed. Disagreement regarding the study selection was resolved by consulting with a third investigator (AAS).

Quality Assessment. Two reviewers (OM and MB) inde-
pendently evaluated the quality of the included studies using the Newcastle-Ottawa scale (NOS) quality tests [11]. Different checklists were used based on the study design. The third investigator solved any discrepancies (AAS) in quality assessment. We rated the quality of included studies by giving stars to three parameters of selection, comparability, and outcome according to the NOS guidelines (Supplementary file (available here)). Cross-sectional studies were categorized to very good, good, satisfactory, and unsatisfactory quality. Cohort studies were categorized as good, fair, and poor quality.
2.6. Data Synthesis. We conducted a network meta-analysis (NMA) on the risk of developing COVID-19 and its severity to assess the relative impacts of various DMTs. Model heterogeneity was estimated by I-square (I 2 ) and tau-squared (τ 2 ). The Q statistic (Q total ) was decomposed to assess the heterogeneity (within study designs (Q within )) and inconsistency (between study designs (Q between )). League table was utilized to indicate all direct and indirect pairwise comparisons using ORs and their 95% CIs. ORs less than 1 indicated that the DMT reduced the risk of COVID-19 susceptibility or severity relative to the comparator DMT. A P score and net rank plot were also applied for ranking all DMTs based on their network estimates. A higher P score indicates a greater risk of COVID-19 susceptibility or severity. We did not perform sensitivity analysis based on the quality of studies since small number of included papers. However, only one included study had unsatisfied quality. We performed no publication bias test since less than 10 studies were        [9]. Four studies reported data from the USA [9,16,17,21], two from Spain [15,20], and one from each of Italy, [18] Iran, [13] and Sweden [19]. One study was multicentric from Europe [14].
The risk of bias judgment for each included study is presented in Supplementary Tables 1 and 2. Respecting the quality of cross-sectional studies, one was very good [17] and another one was unsatisfactory [13]. Regarding cohort studies, the qualities of 4 included studies were good and 4 were fair.

Network Meta-Analysis
3.3.1. COVID-19 Susceptibility. The network graphs and forest plots for the association of DMTs with the risk of acquiring COVID-19 are presented in Figures 2 and 3. Based on univariate analysis, three studies assessing the association of DMTs with the risk of COVID-19 were included in the NMA [13][14][15]. In comparison with no DMT, natalizumab (OR = 4:25, 95% CI: 1.34, 13.46; P score = 0.83) and anti-CD20 agents (OR = 3:17, 95% CI: 1.38, 7.25; P score = 0.72) were associated with higher risk of infection ( Figure 3(a)). Ranking of the risk of infection identified dimethyl fumarate as the best, indicating lowest risk of developing infection, and natalizumab as the worst among DMTs. No significant results were found for other comparisons (Table 2(a)). There was a disagreement between direct and indirect comparison of no DMT with platform therapies (rituximab and glatiramer acetate). In direct comparison, no DMT was associated with a decreased risk of infection compared to platform therapies (OR = 0:39, 95% CI: 0.17, 0.92); but, no significant difference in indirect model was found (OR = 0:45, 95% CI: 0.19, 1.02). The total heterogeneity in NMA was not significant (τ 2 = 0:072 and I 2 = 19:4%, Q total = 7:44, P = 0:282). There was no significant inconsistency between study designs (Q between = 7:86, P = 0:249).

Discussion
This study is aimed at summarizing the existing evidence on association of DMTs with COVID-19 susceptibility and severity in patients with MS. The finding of this network meta-analysis showed that patients on rituximab and ocrelizumab, and no DMT was at greater risk of severe COVID-19 infection compared to other MS patients. We observed no substantial difference across DMTs in the risk of developing severe infection.
When we ranked DMTs, interferon was associated with the lowest risk of acquiring COVID-19 and developing severe infection. This finding was also reported by Sormani et al. [22] that Italian MS patients on interferon were less likely to develop severe COVID-19 than those on other DMTs. These results were expected since interferon is not immunosuppressive and has anti-inflammatory and antiviral effects [23][24][25][26]. Protective effect of interferon against the SARS-CoV and MERS-CoV [27,28], discovering autoantibodies against type I interferons in critically ill COVID-19 patients [29], and inhabitation effect of this agent on SARS-CoV-2 replication [30] suggested interferon as therapeutic candidate for COVID-19 [29,31,32]. However, the effectiveness of interferon on COVID-19 severity among general population in clinical trials remains unclear [33][34][35].
The harmful and beneficial effects of moderate and high effective DMTs on COVID-19 severity are still in dispute. Dimethyl fumarate, teriflunomide, and fingolimod decrease lymphocyte counts resulting in reduced viral clearance which may theoretically increase risk of severe COVID-19 infection [36][37][38]. Moreover, natalizumab limits viral clearance from the central nervous system [39] which could negatively affect the outcome of COVID-19 infection. However, experts and international recommendations suggested that these medications would not increase the risk of severe infection and may even have beneficial effects [40][41][42]. This network meta-analysis showed that none of the interferon, glatiramer acetate, dimethyl fumarate, teriflunomide, and natalizumab had a worse outcome compared to another one. All DMTs were also independently associated with a reduced risk of severe infection compared to no DMT, except anti-CD20 agents. This finding suggests that these medications are not likely to increase the risk of severe COVID-19 and are safe for using within the pandemic. Because of a lack of data, we could not examine the effect of alemtuzumab and cladribine on COVID-19 severity.
In the comparison of each specific DMT with no DMT, rituximab was associated with the highest risk of developing severe COVID-19 infection, followed by ocrelizumab. Observed increased risk of severe illness in patients treated with rituximab and ocrelizumab goes in line with studies on other autoimmune diseases [43][44][45]. Although the exact reason for this association is elusive, it is suggested that patients who treated with anti-CD20 monoclonal antibodies experience decreased antibody production, which can lead to an impaired immune response to SARS-CoV-2 [46][47][48]. Rituximab can also cause a decrease in CD4+ and CD8+ counts [49], which play a substantial role in response to SAR-CoV2 [50].
The results of primary studies showed a stronger association between rituximab and COVID-19 severity than ocrelizumab [17,20,22]. This difference could be related to the antibody-dependent cell-mediated cytotoxic effects and immunogenicity of these drugs [51,52] or some confounders such as characteristics of patients and duration of treatment. The NMA on both univariate and multivariate results showed a decreased risk of developing severe COVID-19 in patients on ocrelizumab compared to rituximab. However, the differences were not substantially significant.
The NMA on univariate results identified lowest risk of developing COVID-19 in MS patients who received no 10 Multiple Sclerosis International DMT. However, the NMA on adjusted or multivariate results showed that platform therapies, fingolimod, dimethl fumarate, and teriflunomide had better outcome than no DMT. This inconsistency may be due the patients' characteristics. Most MS patients who received no DMTs are elderly and have advanced terminal stage. These patients are less involved in high-risk activities such as traveling, working outside the home, and spending a long time in social interaction. As a result, they may stay at home and not be in close contact with COVID-19 cases, which could decrease the risk of developing COVID-19. One major issue in early research concerned the risk of acquiring COVID-19 in those who treated with anti-CD20 agents. Epidemiological and pharmacovigilance data suggested a higher risk of developing COVID-19 in MS patients on these agents [8,9,13]. However, some studies found no association between anti-CD20 medications and risk of the infection [14,16,53]. The suggested reasons for increased risk of acquiring infection in these agents are similar with those mentioned for increased risk of developing severe COVID-19. Although the pooled univariate results showed a higher risk of infection in patients treated with anti-CD20 agents than patients receiving no DMT, no notable difference between DMTs was detected after pooling multivariate analyses. These findings should be interpreted with caution since there was a high level of heterogeneity in NMA on multivariate analyses. Further work needs to be done to investigate the effect of DMTs on the risk of COVID-19 infection.
Our study has some limitations. First, we excluded non-English studies from the study. Second, there are differences in primary studies' health policies and medical care practices, which can affect our results. Third, we combined the quantitative findings of primary studies that used different adjustment methods. Fourth, the definition of COVID-19 susceptibility and severity varied among primary studies. Fifth, a limited number of studies included in quantitative analyses could dominate the estimates. Sixth, the primary used a different primary comparator (no therapy and no DMT). Seventh, this review is based on the current published articles, some of which were relatively small or did not have the necessary statistical power. Therefore, caution must be used when interpreting the association of DMTs with COVID-19 susceptibility or severity.
In conclusion, our study showed that MS patients on anti-CD20 agents are at greater risk of developing severe COVID-19 infection compared to those who received other DMTs and no DMT. It seems that other DMTs did not increase the risk of severe infection and are safe to continue during COVID-19 pandemic. We believed that our results are helpful to design appropriate programs to identify high-risk patients early and adapt vaccination strategies.

Data Availability
All supporting the results of this study can be found within the tables, figures, and manuscript of the present study.

Disclosure
A preprint of this study is available via the link below and is added to the reference list of this manuscript as reference number 2: https://www.medrxiv.org/content/10.1101/2021 .06.11.21258765v1