The Clinical Effectiveness and Cost-Effectiveness of Rituximab Versus Natalizumab in Patients With Relapsing Remitting Multiple Sclerosis

Introduction: Multiple sclerosis (MS) is an inammatory disease in which the myelin sheaths of the nerve cells in the brain and spinal cord, which are responsible for communication, are destroyed and cause physical signs and symptoms. According to studies, anti-CD20 monoclonal antibodies have signicant results in the treatment of this disease. Thus, the aim of the present study was to determine the effectiveness and cost-effectiveness of Rituximab against Natalizumab in the patients with MS in southern Iran in 2020. Methods: This is an economic evaluation including cost-effectiveness analysis in which the Markov model with a lifetime horizon was used. The study sample consisted of 120 patients randomly selected from among those referred to the MS Association and the Special Diseases Unit of Shiraz University of Medical Sciences. In this study, the costs were collected from a societal perspective, and the outcomes were obtained in the form of Quality Adjusted Life Years (QALY) and the mean recurrence rate. The TreeAge pro 2020 and Excel 2016 software were used for data analysis. Results: The comparative study of Rituximab and Natalizumab showed that the patients receiving Rituximab had lower costs ($ 58307.931 vs. $ 354174.85) and more QALYs (7.77 vs. 7.65). In addition, the incidence of relapse by Rituximab was lower compared to Natalizumab (1.15 vs. 2.57). The probabilistic one-way sensitivity analysis showed the robustness of the results. The scatter plots also showed that Rituximab was more cost-effective for the patients in 100% of the simulations for the threshold of >$ 37,641. Discussion and Conclusion: According to the results of this study, Rituximab had higher cost-utility and cost-effectiveness than Natalizumab. Therefore, it could be a priority for MS patients compared to Natalizumab because it reduced treatment costs and increased effectiveness.

In addition, a meta-analysis research carried out in 2020 indicated that the annual prevalence of multiple sclerosis had increased by 2.3% during 1985-2018 (13).
According to studies, multiple sclerosis has a signi cant negative effect on patients' quality of life (14).
The average life expectancy of the patients is 40 years from the onset of the disease, which is 5 to 10 years less than the average life expectancy of non-infected people. About 60% of MS patients reach the age of 70 (15). Furthermore, studying the patients with reduced disability DMT showed that about 90% of the total cost of the patients with mild to moderate disability was associated to drugs, and it was found out that the costs of drugs (except DMT) and non-medical sources were higher for the patients with a more severe disease. The increased disability also had a signi cant effect on health-related quality of life and the fatigue in daily life (16).
Studies show that multiple sclerosis imposes a signi cant economic burden on patients and societies.
The ndings of a study indicated that the cost of annual health care per MS patient increased from $ 45,471 in 2011 to $ 62,500 in 2015, representing an average annual growth of 8.3%. In addition, the annual cost of purchasing medication for each MS patient increased from $ 26,772 to $ 43,606, with an average annual growth rate of 13.0% (17). A study conducted in Spain also found that the total cost of MS was € 1,395 million per year, with an average annual cost of € 30,050 per patient. In addition to the costs, the disease had a signi cant effect on patients' quality of life, and it was estimated that sclerosis imposes would cause a loss of 13,000 QALYs annually. In general, MS had a great economic impact on the Spanish society and a signi cant effect on the patients' quality of life (18).
There is currently no de nitive cure for MS, but various drugs are being used to better control the disease and better adapt the patients to the conditions, amongst which are interferon beta and Glatiramer acetate, oral drugs (Dimethyl fumarate, Teri unomide and Fingolimod) as well as Natalizumab and Alemtuzumab (19). Monoclonal antibodies are currently very popular with specialists. They act against the CD20 protein like Rituximab, Ocrelizumab, and Ofatumumab that have all shown positive results and are being studied as potential drugs (20). Rituximab is a drug used to treat some types of autoimmune diseases and cancers. It is a chimeric monoclonal antibody against the CD20 protein that is commonly found on the surface of B-lymphocytes. Rituximab binds to CD20 and causes cell death. Relatively common complications that often occur 2 hours after starting the injection include skin rash, itching, low blood pressure, and dyspnea (21).
In general, considering the increased number of multiple sclerosis patients as well as the increasing costs of the disease imposed on the patients and societies (22), and since the researchers found no study conducted to evaluate the cost-effectiveness of Rituximab and Natalizumab, the present research aimed to determine the effectiveness and cost-effectiveness of these two drugs in the patients with relapsing remitting multiple sclerosis in Fars province in 2020 in order to identify the most cost-effective one for the patients with MS and to help managers, policy makers, and specialists prescribe better drugs and use the limited resources properly.

Overview and Model Structure:
This is a full economic evaluation of cost-effectiveness analysis type conducted on the patients using Natalizumab and Rituximab who referred to the MS Association and the Special Diseases Unit of Shiraz University of Medical Sciences in 2020. According to the results of the pilot study and considering α = 5%, SD = 1.01, and d = 0.25, the sample size was determined to be 60 in each group. The patients were selected through the random sampling method and entered the study. The inclusion criteria were the use of the mentioned drugs for at least one year and the willingness to participate in the research.
Due to the chronic and relapsing nature of multiple sclerosis, the Markov model was used in this study. Figure 1 indicates a schematic diagram of the Markov model for the disease. The time horizon in this study was the lifespan and the interval of the annual Markov cycles. For each one-year cycle in the model, the patients remained in their current states or were transferred to other ones, and any patient might experience relapse or death (23,24). All the probabilities of transition to other states, Annualized Relapse Rate, and the probability of death in each Expanded Disability Status Scale (EDSS) were extracted based on previously published studies and were presented in Table 1 1 (25)(26)(27)(28). It should be noted that some clinical trial data were reported as rates. They were rst converted to transition probabilities using the following formula, and the transition probabilities were then entered the model (29).
where p is wanted transition probability and r is the overall rate over the time of t.
Since the time horizon was over one year, the cost data and the outcomes of the model were discounted based on the discount rates of 5.8% (30) and 3% (31), respectively.
The outcomes entered this model included QALYs and the costs of any health condition and any treatment strategy. Utility scores were extracted using the Euro Quality-5 Dimension (EQ-5D) questionnaire, and the health outcomes were valued based on quality-adjusted life years (QALY) (34). To measure the utility scores, face-to-face interviews or telephone calls with 120 MS patients were done in 2020. The interviews were conducted with the outpatients referred to the hospitals and clinics a liated to Shiraz University of Medical Sciences. It is worth noting that the EQ-5D questionnaire is a standard tool used to measure health outcomes and includes 5 questions that measure mobility, self-care, usual activities, pain / discomfort, and anxiety / depression. Respondents can score each aspect from 0 to 1, and higher scores mean better utility. The questionnaire was introduced by the EuroQol group in 1990 (https://euroqol.org/). The MS patients who were willing to participate in the present study were interviewed accordingly. Once the EQ-5D questionnaire was completed, its 5-digit codes were changed into numerical utility by considering the values of Iran determined by Goodarzi et al. (35) using the time trade-off method (TTO).
The community perspective was also used to extract the costs.

Effectiveness
The mean relapse rate was used to measure effectiveness. To this end, the total number of relapses of all patients was calculated and divided by the total number of years of drug use to obtain the annual relapse rate of the patients. The obtained value was then divided by the number of patients using each drug in order to obtain the mean relapse rate of each drug (40,41).

Determining the Incremental Cost-Effectiveness Ratio
After the calculation of the costs and utilities in the previous stages, the incremental cost-utility was calculated using the following formula. To make the nal decision on the cost-effectiveness of each intended drug, the ICER level was compared with the threshold.
Due to the lack of an explicit threshold for Willingness To Pay (WTP) in Iran, the WHO's proposal for developing countries was used, i.e. the willingness to pay per QALY was one to three times the per capita Gross Domestic Product (GDP) (42). In Iran, GDP was $ 12,547 in 2019 (43) based on which the threshold for willingness to pay was $ 37,641 (3* GDP). The Excel 2016 and TreeAge Pro 2020 software were also used for data analysis.

Sensitivity Analysis
The researchers used the one-way sensitivity analysis and Probabilistic Sensitivity Analysis (PSA) to investigate the effects of parameter uncertainty on the results. In order to perform the one-way sensitivity analysis, some key parameters such as cost and utility were changed by 20% for each drug strategy and the results were presented in the form of a Tornado Diagram. In addition, since the utility and cost variables in the present study were measurable and probabilistic, PSA was performed and they were considered as distributions, so that beta distribution (ß ) was used to determine the distribution of utility values (ranged from 0 to 1) and gamma distribution was used to determine the cost distribution.
Accordingly, Second-order Monte Carlo simulation was performed using 5000 trials. The PSA results are presented using the cost-effectiveness acceptability curve and incremental cost-effectiveness scatter plot. Cost-effectiveness acceptability curve is one of the best curves for planning and policy-making that

Results
The ndings of the present research showed that most of the patients were female (74.17%) and housewives (52.50%), and all of the patients had an insurance coverage. In addition, the mean ages of the patients treated with Natalizumab and Rituximab were 33.4 ± 7.27 and 34.92 ± 5.94 years, respectively. Tables 2 and 3 show the mean cost, utility, and relapse in the multiple sclerosis patients using Natalizumab and Rituximab. According to  As shown in Table 3, the highest utility score obtained from the EQ-5D questionnaire was that of the MS patients using Rituximab and with the EDSS of 0.5-2 (0.833 ± 0.125). The lowest mean relapse was that of the patients using Rituximab with the EDSS of 0.5-2 (0.313 ± 0.528). As presented in Fig. 2 and Table 4, the results of the cost-utility analysis using the Markov model suggested that the mean cost was $ 58,307,931 in the Rituximab arm and the QALY was 7.77 as well. However, the mean predicted cost in the Natalizumab arm was $ 354,174.85 and the obtained QALY was $7.65.

One-way Sensitivity Analysis
According to the tornado diagram in Fig. 3, ICER was most sensitive to the price of Natalizumab and less sensitive to that of Rituximab. In addition, the tornado diagram shows that changes in input parameters had little effect on the result and the ICER rate remained negative.
Probabilistic sensitivity analysis (PSA): The results of uncertainty measurement are presented using the cost-effectiveness acceptability curves and ICER scatter plot as follows. The acceptability curves indicated that Rituximab was the most costeffective treatment in 100% of simulations for the thresholds lower than $ 37,641 (Fig. 4).
The scatter plot showed that Rituximab was totally in the acceptance area below the threshold in 100% of the cases and defeated Natalizumab, thus being more cost-effective. On the other hand, Natalizumab was in the rejection zone and above the threshold line in 100% of the cases and was considered as a noncost-effective (ine cient) strategy (Fig. 5).

Discussion
Along with the identi cation and approval of new drugs for the management of multiple sclerosis, there has been ever-increasing costs of treating the disease imposed on families and communities (45). Therefore, conducting economic evaluations seems necessary to identify the cheapest, most effective, and thus most cost-effective medicines and suggest them to doctors for prescription. Such studies can also help health policy makers to reduce patients' out-of-pocket payments. To the knowledge of the researchers, this is one of the rst health-economic studies conducted to compare Rituximab and NTZ in the treatment of multiple sclerosis. In fact, this is the rst study carried out in Iran on the costeffectiveness of Rituximab versus Natalizumab in MS patients.
According to the ndings of this study, the mean annual costs of treatment with Natalizumab and Rituximab were $ 36,058.08 and $ 5,399.28, respectively. Thus, the mean cost of a course of treatment per patient was lower with Rituximab than with Natalizumab. One reason for this difference could be the higher price of Natalizumab compared to Rituximab. The studies by Chisari et al. (2021) showed that the annual cost of Rituximab was signi cantly lower than other approved drugs used in the treatment of MS (20). Hartung (2017) also showed that the annual cost of MS drugs was about $ 70,000 (46), which is consistent with the results of the present study. The results of the present research showed that the highest utility rate and the lowest relapse rate in treatment with each drug was found in the patients with an EDSS of 0-2.5, and as disability increased, the life utility rate decreased and the relapse rate increased as well. This might be due to the fact that in higher EDSSs, the disease usually progresses and the patients' limitations increase; so it is natural for the relapse rate to get higher and the utility rate to decrease (53,54).  (27,(55)(56)(57).
Based on the results of the Markov model analysis, the costs of Natalizumab and Rituximab were respectively $ 58,307.93 and $ 354,174.86, and the obtained QALYs were 7.77 and 7.65 over the lifetime horizon. Therefore, since Natalizumab had higher cost and lower QALY, it was considered the recessive option in cost-effectiveness analysis. Thus, Rituximab was more cost effective than Natalizumab.
Rezaei et al. (2019) conducted a study in Iran and examined the cost-utility of Natalizumab versus Fingolimod. They found out that the mean cost per patient during life was $ 58,751 in the Fingolimod arm and the utility was 8.09 QALY, but in the Natalizumab arm, the mean cost and the obtained QALY were $ 204,264 and $ 7.37, respectively. Thus, Natalizumab had higher cost and lower QALY and was the recessive option. This is consistent with the results of the present study (47).
In their study, Taheri et al. (2019) examined the cost-effectiveness of Alemtuzumab versus Natalizumab and concluded that the total discounted costs per patient were $ 147,417 and $ 150,579, respectively. In addition, the mean discounted QALYs were estimated at 7.07 and 6.05 for Alemtuzumab and Natalizumab, respectively, over a period of 20 years. Therefore, Natalizumab was the recessive option, which is consistent with the present study (48).
robustness. In addition, the results of probabilistic sensitivity analysis showed that Rituximab regimen was more cost-effective than Natalizumab and in all cases, it was in the acceptance area and below the threshold. Thus, it had obtained the best results in terms of the mentioned prices. The results also showed that doing sensitivity analysis did not change the status of Rituximab as the most effective drug regimen, suggesting the robust study results. In this regard, the present study is in line with those of One limitation of the present study was the self-reporting of the patients or their companions about direct non-medical and indirect costs, as they were likely to forget or approximate some of the costs. In addition, intangible costs were not calculated in this study due to the inability to measure them accurately

Conclusions
The results of the present research suggested that Rituximab had higher cost-effectiveness and costutility than Natalizumab. Therefore, considering the results of the sensitivity analysis and the robustness of the results, Rituximab treatment is suggested as the rst priority (compared to Natalizumab) to treat the patients with multiple sclerosis, and health policy makers and managers should try to increase insurance coverage and reduce the patients' out-of-pocket payments. Subjects Act and no approval by a medical ethics committee was necessary. Verbal consent was obtained from respondents who completed a questionnaire anonymously, and their response expressed their willingness to participate. Written consent was obtained from participants to the cognitive interviews Consent for publication: All participants completed a consent form, stating that they were well-informed about the Content of questionnaires and that they agreed upon the publication of anonymized data.  The cost-effectiveness acceptability curves of Rituximab versus Natalizumab based on the QALY obtained through the Monte Carlo simulation