Skip to main content
Advertisement
Browse Subject Areas
?

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Effectiveness of first-line treatment for relapsing-remitting multiple sclerosis in Brazil: A 16-year non-concurrent cohort study

  • Kathiaja Miranda Souza ,

    Roles Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Resources, Supervision, Validation, Visualization, Writing – original draft, Writing – review & editing

    kathiajamsouza@gmail.com (KMS); atallahmbe@uol.com.br (ANA)

    Affiliation Programa de Pós-Graduação em Saúde Baseada em Evidências, Universidade Federal de São Paulo, São Paulo, São Paulo, Brazil

  • Isabela Maia Diniz,

    Roles Formal analysis, Investigation, Writing – original draft, Writing – review & editing

    Affiliations SUS Collaborating Centre for Technology Assessment & Excellence in Health, Faculdade de Farmácia, Universidade Federal de Minas Gerais, Belo Horizonte, Minas Gerais, Brazil, Programa de Pós-Graduação em Medicamentos e Assistência Farmacêutica, Faculdade de Farmácia, Universidade Federal de Minas Gerais, Belo Horizonte, Minas Gerais, Brazil

  • Lívia Lovato Pires de Lemos,

    Roles Investigation, Writing – original draft, Writing – review & editing

    Affiliations SUS Collaborating Centre for Technology Assessment & Excellence in Health, Faculdade de Farmácia, Universidade Federal de Minas Gerais, Belo Horizonte, Minas Gerais, Brazil, Programa de Pós-Graduação em Saúde Pública, Faculdade de Medicina, Universidade Federal de Minas Gerais, Belo Horizonte, Minas Gerais, Brazil

  • Nélio Gomes Ribeiro Junior,

    Roles Writing – review & editing

    Affiliation SUS Collaborating Centre for Technology Assessment & Excellence in Health, Faculdade de Farmácia, Universidade Federal de Minas Gerais, Belo Horizonte, Minas Gerais, Brazil

  • Isabella de Figueiredo Zuppo,

    Roles Writing – review & editing

    Affiliation SUS Collaborating Centre for Technology Assessment & Excellence in Health, Faculdade de Farmácia, Universidade Federal de Minas Gerais, Belo Horizonte, Minas Gerais, Brazil

  • Juliana Alvares Teodoro,

    Roles Writing – review & editing

    Affiliations SUS Collaborating Centre for Technology Assessment & Excellence in Health, Faculdade de Farmácia, Universidade Federal de Minas Gerais, Belo Horizonte, Minas Gerais, Brazil, Programa de Pós-Graduação em Medicamentos e Assistência Farmacêutica, Faculdade de Farmácia, Universidade Federal de Minas Gerais, Belo Horizonte, Minas Gerais, Brazil

  • Francisco de Assis Acurcio,

    Roles Writing – review & editing

    Affiliations SUS Collaborating Centre for Technology Assessment & Excellence in Health, Faculdade de Farmácia, Universidade Federal de Minas Gerais, Belo Horizonte, Minas Gerais, Brazil, Programa de Pós-Graduação em Medicamentos e Assistência Farmacêutica, Faculdade de Farmácia, Universidade Federal de Minas Gerais, Belo Horizonte, Minas Gerais, Brazil, Programa de Pós-Graduação em Saúde Pública, Faculdade de Medicina, Universidade Federal de Minas Gerais, Belo Horizonte, Minas Gerais, Brazil

  • Álvaro Nagib Atallah ,

    Roles Supervision, Writing – review & editing

    kathiajamsouza@gmail.com (KMS); atallahmbe@uol.com.br (ANA)

    Affiliations Programa de Pós-Graduação em Saúde Baseada em Evidências, Universidade Federal de São Paulo, São Paulo, São Paulo, Brazil, Cochrane Brazil, Centro de Estudos de Saúde Baseada em Evidências e Avaliação Tecnológica em Saúde, São Paulo, São Paulo, Brazil

  • Augusto Afonso Guerra Júnior

    Roles Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Project administration, Software, Supervision, Validation, Visualization, Writing – review & editing

    Affiliations SUS Collaborating Centre for Technology Assessment & Excellence in Health, Faculdade de Farmácia, Universidade Federal de Minas Gerais, Belo Horizonte, Minas Gerais, Brazil, Programa de Pós-Graduação em Medicamentos e Assistência Farmacêutica, Faculdade de Farmácia, Universidade Federal de Minas Gerais, Belo Horizonte, Minas Gerais, Brazil

Abstract

Background

Relapsing-remitting multiple sclerosis (RRMM) is a chronic, progressive, inflammatory and immune-mediated disease that affects the central nervous system and is characterized by episodes of neurological dysfunction followed by a period of remission. The pharmacological strategy aims to delay the progression of the disease and prevent relapse. Interferon beta and glatiramer are commonly used in the Brazilian public health system and are available to patients who meet the guideline criteria. The scenario of multiple treatments available and in development brings the need for discussion and evaluation of the technologies already available before the incorporation of new drugs. This study analyses the effectiveness of first-line treatment of RRMS measured by real-world evidence data, from the Brazilian National Health System (SUS).

Methods and findings

We conducted a non-concurrent national cohort between 2000 and 2015. The study population consisted of 22,722 patients with RRMS using one of the following first-line drugs of interest: glatiramer or one of three presentations of interferon beta. Kaplan–Meier analysis was used to estimate the time to treatment failure. A univariate and multivariate Cox proportional hazard model was used to evaluate factors associated with treatment failure. In addition, patients were propensity score-matched (1:1) in six groups of comparative first-line treatments to evaluate the effectiveness among them. The analysis indicated a higher risk of treatment failure in female patients (HR = 1.08; P = 0,01), those with comorbidities at baseline (HR = 1.20; P<0,0001), in patients who developed comorbidities after starting treatment (i.e., rheumatoid arthritis—HR = 1.65; P<0,0001), those exclusive SUS patients (HR = 1.31; P<0,0001) and among patients using intramuscular interferon beta (IM βINF-1a) (28% to 60% compared to the other three treatments; P<0,0001). Lower risk of treatment failure was found among patients treated with glatiramer.

Conclusions

This retrospective cohort suggests that glatiramer is associated with greater effectiveness compared to the three presentations of interferon beta. When evaluating beta interferons, the results suggest that the intramuscular presentation is not effective in the treatment of multiple sclerosis.

Introduction

Multiple sclerosis (MS) is a chronic, progressive, inflammatory and immune-mediated disease that affects the central nervous system and is characterized by loss of motor and sensory function resulting from immune-mediated inflammation, demyelination, and subsequent axonal damage. It is characterized by episodes of neurological dysfunction that can be followed by a period of remission or may progress [1].

Relapsing-remitting multiple sclerosis (RRMS) is the most common form of the disease and is identified by relapses or acute attacks (worsening of symptoms of neurological dysfunction lasting 24 hours or longer, in the absence of fever), which can enter remission spontaneously or upon treatment with corticosteroids (pulse therapy). The most commonly observed symptoms are optic neuritis; paresis or paresthesia of limbs; motor incoordination; lack of balance; myelitis; sphincter dysfunction and cognitive-behavioral disorders, either alone or in combination [13].

The prevalence rate of MS is variable worldwide, with a global average prevalence of 33 cases per 100,000 inhabitants. In Brazil, a recent systematic review of prevalence studies estimated an average rate of 8.69 cases per 100,000 inhabitants, with considerable variation among the regions of the country, with a prevalence of 1.36 cases per 100,000 inhabitants in the northeast region and 27. 2 cases per 100,000 inhabitants in the south region. The Brazilian Multiple Sclerosis Association (ABEM) estimates that there are 35,000 Brazilians currently living with the disease [46].

The treatment, based on the use of Disease Modifying Therapies (DMTs) or immunomodulators, aims to delay the progression of RRMS and prevents relapse by decreasing circulating immune cells or by preventing these cells from crossing the blood-brain barrier, thereby reducing the inflammatory response [7, 8]. Until December 2015, interferon beta (βINF) and glatiramer were the only alternatives for first-line treatment of patients who met the guideline criteria in the Brazilian public health system (SUS). According to international guidelines and the Clinical Guideline of the Brazilian Ministry of Health, patients could start with either one of the first-line treatments and switch to the other in the case of treatment failure. Treatment failure is defined as a presence of relapses, activity on magnetic resonance imaging, a progression of disability or intolerance due to adverse events [9].

The results of primary clinical studies are important sources of information for the listing and maintenance of medicines towards the treatment of RRMS in the health system. However, most studies have limitations such as the use of placebo as a control group, short follow-up protocols, specific eligibility criteria of patients, high costs, little relevant clinical outcomes and existence of competing interests [10, 11]. These limitations raise the question of whether health systems achieve the expected results when they decide to incorporate and pay for health technologies. Also, it is known that there are more than 12 different treatments commercially available, as well as new medicines that appear in the technological horizon to treat MS [12]. This scenario has a strong impact on SUS since there is an increasing demand for the incorporation of new technologies to treat this condition [13]. This fact brings forth the need for discussion and evaluation of the technologies already available in SUS for a better understanding of the use of these medicines and obtaining adequate information to guide decision-making for SUS sustainability and efficiency [14].

In Brazil, the information on the dispensing of DMT and outpatient procedures are recorded in the SUS Outpatient Information System. Hospital procedures, such as treatment for relapses and complications are recorded in the SUS Hospital Information System. The integration of these databases with the Mortality Information System allows the construction of a robust patient-centered registry for long term longitudinal follow-up of these patients [15, 16]. We believe it is important to analyze the profile of utilization and the effectiveness of first-line DMT in patients with RRMS in the Brazilian public health to guide stakeholders in future investment and disinvestment decisions. We also hope our results will be useful to aid more researches on multiple sclerosis.

Methods

Ethical statement

The Research Ethics Committee for the University Hospital at the Federal University of São Paulo (HSP/UNIFESP, processes number 2.468.332 and CAAE 81043417.6.0000.5505) approved of the study, under Brazilian national legislation.

Study design and setting

This study is a nationwide non-concurrent open cohort with follow-up from 2000 to 2015 developed by deterministic-probabilistic linkage using the patient-centered registry of the Hospital Information System (SIH), Ambulatory Information System (SIA) and Mortality Information System (SIM) [15]. SIH comprises data on hospitalization from both public and private hospitals contracted by SUS. The subsystem of High-Complexity Procedure Authorization (APAC) of the SIA database comprises all information about DMT drugs dispensing and also records the medical diagnosis for which the drug is prescribed, using the International Classification of Diseases, Tenth Revision (ICD-10) codes.

Although in Brazil there is a mix of public (SUS) and private health services, the SUS covers health services for approximately 80% of citizens. Moreover, it is believed most citizens of Brazil use the SUS for high-cost medicines such as the DMT agents [17, 18]. Thus, we believe that SUS database comprehensively includes people with RRMS in Brazil. Previous studies have used the same methodology and proven accuracy of the registers [15, 18, 19].

Study population

The study population consisted of RRMS patients aged 18 years and older identified by the registry of ICD-10 code G35 and received one of the following first-line drugs of interest: Glatiramer, intramuscular interferon beta 1a (IM βINF- 1a), subcutaneous interferon beta 1a (SC βINF- 1a) and subcutaneous interferon beta 1b (SC βINF- 1b). We considered 4 exposure groups according to the initial drug dispensed at cohort entry (intention to treat). We excluded patients who started the cohort using other DMTs or were using a combination of more than one drug. Also, as an inclusion criterion, patients were required to have at least 6 months of records based on the dispensation of the drugs of interest to be considered as a truly RRMS case. The date of entry into the cohort corresponded to the date of the first dispensation registry of DMT after continuous DMT use during these 6 months. The entry period was between January 2000 to December 2014, and patients were followed up from January 2000 to December 2015, 16 years. This strategy assured a minimum follow-up of 12 months. Patients were censored if they abandoned or interrupted their treatment for more than three months or at the end of follow-up (right censoring).

Study outcomes

The effectiveness outcome was the time until treatment failure, defined as the record of a relapse, a switch to another DMT, or death, whichever occurred first. Relapse was defined by the occurrence of methylprednisolone dispensing and/or pulse steroid therapy and/or hospital relapse treatment. For the event of switch treatment, we allowed a tolerance of three consecutive months without dispensing registry until the patient initiated the other drug. A 3-month gap was selected in an attempt to ensure that patients switching DMTs were not misclassified as having discontinued therapy. It is also consistent with the threshold used in a recently published study examining persistence in patients with MS [20].

Statistical analysis

Demographic and clinical characteristics were presented as mean ± SD or median and lower and upper quartiles range (Q1–Q3) for the continuous variables and as the number (percentage) of subjects for categorical variables. We used the SUS databases to establish the presence of the comorbidity before and after initial treatment based on the Elixhauser Comorbidity, adapted from Quan et al. [21]. We compared treatment groups using univariate comparisons with Kruskal-Wallis or Chi-squared test.

In addition, the cohort patients were paired (groups βINF-1a IM vs. βINF-1b SC; βINF-1a IM vs. glatiramer; βINF-1a IM vs. βINF-1a SC; βINF-1b SC vs. glatiramer; βINF-1b SC vs. βINF-1a SC; and glatiramer vs. βINF-1a SC) based on the propensity score matched to control possible differences between treatment groups. The propensity score is defined by the probability that the individual will receive a specific treatment, conditioned to the studied covariables. Propensity scores were estimated using logistic regression models considering the following variables: age, sex, presence of comorbidities at the beginning of treatment, region of residence, exclusive SUS patient and period of entry into the cohort (baseline characteristics). The nearest neighbor matching within the caliper method was used, with a maximum fixed difference of 0.2 between the propensity scores to select patients with homogeneous baseline characteristics [22, 23].

Time to treatment failure was analyzed using the Kaplan-Meier survival curves and the log-rank test was used to evaluate the difference among drug treatments. Univariate and multivariate Cox proportional hazards models were used to investigate the impact of baseline characteristics on treatment failure. Additionally, the impact of the appearance of comorbidities after starting DMT treatment was investigated. Variables with p < 0.20 in the univariate analysis and variables considered clinically relevant were included in the multivariable model. The proportional hazards assumption was evaluated using statistical tests based on the Schoenfeld residuals [24].

We performed two sensitivity analyses. In the first one, we did not impose a time limit on the off-treatment gap for treatment failure event assignment, that is, patients were not censored for abandonment if they returned to index treatment or switched drugs at any time after interrupting first treatment. If they did not return to treatment, they were censored for loss to follow-up.

In the second sensitivity analysis we considered patients who use the public health system to access the other offered services which include outpatient and hospital procedures (exclusively SUS patients) (a), and patients who use SUS only to get high-cost drugs (b).

Results were considered significant for p-values < 0.05 (two-sided) and a Bonferroni approximation was employed to adjust for the multiple paired comparisons. The adjusted p-value that indicated significance was 0.008 and was computed as (α/number of comparisons), or (0.05/6) [25, 26]. Data analysis was performed using “R” version 3.4.1 (R Foundation for Statistical Computing).

Results

Study population

We identified 27,446 patients with RRMS who had at least one of the DMT treatments in the period 2000 to 2015. Of these, we excluded 95 (0.35%) patients whose data indicated an error during the deterministic-probabilistic record linkage; 1,060 (3.9%) patients for being younger than 18 years old at the start of treatment; 2,571 (9.4%) patients who not fulfilled of the criteria of 6 months of continuous dispensation prior to the start of follow-up and; 998 (3.6%) for having started treatment with other drugs for RRMS (azathioprine, fingolimod or natalizumab) or for using more than one drug, yielding 22,722 patients for analysis. Of the included patients, the majority started treatment with SC βINF-1a (35.6%); followed by IM βINF-1a (24.7%); SC βINF-1b (22.3%) and glatiramer (17.4%) (Fig 1).

Baseline characteristics

Table 1 provides the baseline demographic and clinical characteristics of MS patients. The median (Q1–Q3) age was 37 (29–46) years and 73.3% were women. Most patients were from the Southeast region (58.4%), followed by South (19.2%), Northeast (11.9%), Midwest (9.0%), and North (1.6%). The majority of patients (35.2%) enrolled in the cohort during the 2008–2011 period and 10% of patients had one or more comorbidities before starting treatment, in which other neurological diseases and paralysis (MS-related comorbidities) were the most frequent.

thumbnail
Table 1. Baseline characteristics of Brazilian patients with RRMS included in the study, 2000–2015.

https://doi.org/10.1371/journal.pone.0238476.t001

We observed statistical differences between groups, which were not considered clinically meaningful. Exclusive SUS patients represent 41.9% of all patients, 35.9% of IM βINF-1a group, 39.8% of glatiramer group; 42.8% of SC βINF-1a group; and 48.7% of SC βINF- 1b group.

The characteristics of the matched cohorts are shown in S1 Table. The resulting six paired groups were closely matched on all recorded demographic and clinical parameters. The mean differences between groups in the propensity scores were significantly reduced, with emphasis on the variable period of entry into the cohort, with a reduction of up to 97% of the difference, and up to 62.5% for exclusively SUS patients.

Treatment failure

After starting initial DMT treatment, the proportion of patients who experienced treatment failure during follow-up was 25.6%, and the IM βINF-1a group (30.3%) had experienced most events compared to other groups (Table 2). The median (SD) follow-up time for this cohort was 68.7 (45.9) months (range 6–192). The median time to treatment failure of all first-line DMT treatment, estimated by Kaplan-Meier survival analysis, was approximately 7 years (88 months, 95% CI 83–96). The median time to treatment failure was worse (i.e., shorter) among female patients; exclusive SUS patients (i.e., users of SUS for other health care procedures); patients that developed comorbidities after starting treatment and among patients of IM βINF-1a group (Fig 2).

thumbnail
Fig 2.

Kaplan–Meier estimates for the time to treatment failure of all (A), exclusive and non-exclusive SUS patients (B), gender (C), comorbidities after starting treatment (D) and initial DMT treatment (E). Abbreviations: DMT, disease-modifying therapy; CI, confidence interval; IM-IFNβ-1a, intramuscular interferon beta-1a; SC-IFNβ-1b, subcutaneous interferon beta-1b; SC-IFNβ-1a, subcutaneous interferon beta-1a.

https://doi.org/10.1371/journal.pone.0238476.g002

thumbnail
Table 2. Treatment failure among the patients with RRMS included in the study, 2000–2015, Brazil.

https://doi.org/10.1371/journal.pone.0238476.t002

The univariate and multivariate analysis indicated that, among all first-line DMT users, higher risk of treatment failure were observed in female patients, those enrolled in the 2000–2003 period, those with comorbidities at baseline, in patients who had comorbidities after starting treatment (MS-related comorbidities, renal failure and rheumatoid arthritis), those exclusive SUS patients and among patients of IM βINF-1a group. Region of residence and other comorbidities were not significant in univariate analysis and were not included in the multivariate analysis (Table 3).

thumbnail
Table 3. Univariate and multivariate Cox regression models with predictors of DMT treatment failure for RRMS patients included in the cohort, 2000 to 2015, Brazil.

https://doi.org/10.1371/journal.pone.0238476.t003

The Kaplan–Meier curves for the matched comparison among the first-line DMT treatment groups are shown in Fig 3. This analysis revealed a significant difference in favor of glatiramer in all comparisons and IM βINF-1a had the worst result in most of the comparisons. Also, SC βINF- 1b presented a shorter time to failure compared to SC βINF- 1a.

thumbnail
Fig 3.

Kaplan–Meier matched comparisons for the time to treatment failure of IM βINF- 1a vs. SC βINF-1b (A), IM βINF- 1a vs. glatiramer (B), IM βINF- 1a vs. SC βINF-1a (C), SC βINF- 1b vs. glatiramer (D) SC βINF- 1b vs. SC βINF-1a (E) and Glatiramer vs. SC βINF-1a (F). Bonferroni adjustment for multiple comparisons. The significance level is p ≤ 0.008 (6 comparisons). Abbreviations: IM-IFNβ-1a, intramuscular interferon beta-1a; SC-IFNβ-1b, subcutaneous interferon beta-1b; SC-IFNβ-1a, subcutaneous interferon beta-1a.

https://doi.org/10.1371/journal.pone.0238476.g003

The risk of therapeutic failure of IM βINF-1a after pairing by propensity score was 28%; 60% and 41% higher than βINF-1b SC, glatiramer and βINF-1a SC, respectively. We observed that the risk was slightly lower compared to the results of the multivariate analysis. Glatiramer had a lower risk of failure compared to βINF-1b SC and βINF-1a SC by 25% and 13% respectively; however, there was no difference after Bonferroni's adjustment (Table 4).

thumbnail
Table 4. Cox regression of DMT treatment failure for RRMS patients included in the cohort after pairing by propensity score, 2000 to 2015, Brazil.

https://doi.org/10.1371/journal.pone.0238476.t004

Sensitivity analysis

When we did not impose a time limit on the off-treatment gap for the treatment failure event assignment revealed similar results as the main analysis. The median time to treatment failure was 50 (95% IC 46–53) months for IM βINF-1a group; 56 (95% IC 52–61) months for SC βINF-1b group; 97 (95% IC 87–113) months for glatiramer group; 69 (95% IC 65–73) months for SC βINF- 1a group and; the overall median time to treatment failure was 64 (95% IC 62–67) months (log-rank p-value = <0.0001).

In the same way, adjusted Cox regression analyses showed that IM βINF-1a had a higher risk of treatment failure at 28% (HR = 1.28, 95% CI 1.19–1.38in comparison to SC βINF-1b; 64% (HR = 1.64, 95% CI 1.52–1.79) in comparison to glatiramer and; 41% (HR = 1.41, 95% CI 1.32–1.51) in comparison to SC βINF-1a. Regarding to the sensitivity analysis restricted to exclusive SUS patients, we observed survival probabilities and HRs comparable to the main analysis. However, the median time to treatment failure were significantly lower when compared to patients who access SUS only to obtain medications (48 (95% IC 44–54) versus 82 (95% IC 72–109) months for IM βINF-1a group; 64 (95% IC 55–75) versus 97 (95% IC 86-NA) for SC βINF-1b group; 96 (95% IC 66-NA) versus 150 (95% IC 150-NA) for glatiramer group and; 69 (95% IC 59–78) versus 140 (95% IC 120-NA) for SC βINF- 1a group p<0,0001).

Discussion

To evaluate the long-term effectiveness of first-line DMT treatment in Brazil, we conducted a deterministic-probabilistic linkage of SUS administrative databases and extracted a cohort of RRMS patients. This study highlights there is an important difference among first-line treatments in terms of treatment failure.

Overall, this study reports that women, exclusive SUS patients, presence of Elixhauser comorbidities and IM βINF-1a are associated with increased treatment failure survival time. Also, IM βINF-1a had a significantly higher risk of treatment failure than patients treated with other first-line DMT and both subcutaneous presentations of interferon-beta were inferior to glatiramer, independently of age, gender, comorbidities, region of residence, period of cohort entry and exclusive SUS patients in a Cox regression model. In the paired propensity score match analysis, although a few differences were found in terms of significance, the superiority of other DMTs over IM βINF-1a was evident.

Treatment groups differed in several clinical and demographic characteristics. Usually, the allocation to therapy is determined by MS treatment centers, individual physician choice, disability grade, course of illness, and age. In the study by Kalincik et. al (2014), the authors identified that patients treated with SC βINF-1a had more severe disabilities compared to the other groups, whereas the patients receiving IM βINF-1a had fewer disabilities. Patients who started with glatiramer were older than those treated with SC or IM βINF-1a [27].

The provision of high-cost drugs is generally not included in the services provided by private insurance operators, as is the case with DMTs for the treatment of RRMS. Therefore, private insurance patients access SUS to obtain the drugs, but the registry of relapse or other medical care in SUS Information Systems for these patients is deficient [17, 18]. The data obtained in this cohort corroborate the literature of SUS coverage, as we identified 41.9% patients that are exclusively SUS users, that is, they do not have private insurance. It should be emphasized that this number may be underestimated, considering that we have adopted conservative methodology to estimate the number of these patients.

Regarding medicines, approximately 83% of patients included in the cohort began treatment with interferon beta, with SC βINF-1a being the most dispensed (35.6%), followed by IM βINF-1a (25.7%). Glatiramer represented 17.4% of the total cohort patients. Until 2017, interferons beta and glatiramer were the only first-line drugs provided by SUS and are in accordance with current clinical guidelines for MS [9, 28, 29]. This pattern of use could be observed in the cohorts of Kalincik et al. (2014), Evans et al. (2012), Halpern et al. (2011) and Reynolds et al. (2010) [20, 27, 30, 31]. The switch of medication was the event that most contributed to the composite failure outcome and is in accordance with the definition advocated by the Brazilian Clinical Guideline of MS, where the exchange between drugs is allowed through evidence of treatment failure [9].

Several evidences suggest that the presence of other comorbidities is common in individuals with MS and may have a negative impact on clinical outcomes and treatment. In this study, we identified that 10% of patients included had one or more associated comorbidities at the beginning of treatment, most of them being conditions related to MS. However, the prevalence and methods of collecting and identifying some comorbidities in MS remain underreported. In this study, we used the algorithm proposed by Quan et al. [21], which collects information from inpatient care. Also, it is known that the notification and registration of comorbidities to obtain medicines for the treatment of MS in SUS is an optional field, so we believe that this prevalence may be even higher [3234].

Furthermore, other studies had evaluated the association between comorbidities and the time to first drug change and identified that comorbidities were significantly associated with increased risk of switching due to treatment intolerance [35, 36].

Concerning to the increased risk of treatment failure among women, findings from other studies corroborate the results of this study. It is known that MS affects more women than men. Besides, the progression and severity of the disease may differ from person to person, however, in women, in addition to pregnancy and breastfeeding, which require immediate discontinuation of treatment and consequently the risk of progression, factors such as menstruation and menopause can influence symptoms and progression of MS [31, 3739].

In this cohort, the treatment failure stratified by the initial drug indicated significant differences between the first-line treatment options. Patients treated with glatiramer, followed by SC βINF-1a and SC βINF-1b had a significantly lower risk of treatment failure (40%, 30% and 22% respectively) than those treated with IM βINF-1a. In the treatment matched comparisons, the risk of treatment failure is even greater. It is observed that glatiramer acetate is superior to SC βINF-1b and SC βINF-1a, and these are statistically more effective than intramuscular interferon beta.

These results confirm and corroborate previous non-concurrent cohorts. Zhornistky et al. (2015) evaluated the time to treatment interruption due to intolerance or inefficacy of the initial treatment and showed that patients treated with interferons had a 29% risk of discontinuing treatment when compared to patients treated with glatiramer (HR = 1.29, 95% CI, 1.11–1.50). In the study of Reynolds et al. (2010), Kaplan Meier analyses demonstrated that the initial drug switch or treatment discontinuation was significantly lower in patients treated with glatiramer [27, 31, 40]. Besides the real-world evidence observed in cohorts, the results of clinical trials and systematic reviews support the results found in this study [4143].

On the other hand, similar studies with different results were also observed [44, 45]. Although the reason for this variability is not clear, it is assumed that sample size, clinical and demographic differences in baseline characteristics among the treatments compared may have influenced the outcome. In addition, since the cohort is not randomized, other differences that are not known between the groups may have existed.

In this non-concurrent cohort, the analysis used comparisons between the treatments with uniform inclusion criteria and the current clinical protocols for MS in Brazil. The population heterogeneity is important, as it reflects how the use of medicines is in the real world and their effectiveness thereafter. In an attempt to adjust the distributions between groups and the potential confounders that could influence the effectiveness results, pairing by propensity score was performed. This analysis reinforced the main findings of effectiveness between the treatments.

The main limitation of this study is that the data collected come from the administrative systems databases developed by SUS were not designed to evaluate and follow clinical results. Consequently, there is no record of patient-level clinical variables such as disability progression (EDSS scale), RMI results, disease duration, adverse event data, and causes that led to treatment switching. Besides, there is a possibility of incorrect data caused by sub-registers that may culminate in the sub or overestimation of the analyzes performed. It is likely, that many cases of relapses and other events have not been recorded in the SUS databases, as it is known that many patients use private services.

To overcome these limitations, we chose to use the treatment failure as composed outcome and persistence, capturing the switch of treatment that, as indicated in the clinical protocol, occurs in cases of relapses, intolerance or adverse events. Additionally, to minimize possible inconsistencies, after data collection, a sample inspection was performed with subsequent cleaning and standardization, constituting important steps to guarantee the quality of the collected data.

Conclusions

This non-concurrent cohort containing more than twenty thousand patients with RRMS suggests that glatiramer acetate is associated with greater effectiveness compared to βIFN. When evaluating interferons beta, the results suggest that IM βIFN- 1a is the least effective in the treatment of RRMS.

Considering that the main objective of DMTs is to reduce the occurrence of relapses and, consequently, to reduce the progression of the disease, the inferiority of IM βIFN-1a in comparison to the other medications surpasses the convenience of one application per week. In addition, RRMS is a disease that has a large therapeutic arsenal available in the SUS and the recent updates of the Brazilian clinical guideline of Multiple Sclerosis count on new treatments, including oral medications. Therefore, considering the effectiveness of the treatment, the existence of therapeutic alternatives that present a better relation between cost, effectiveness and expected benefits for the patients is paramount for the good practices in the health system. The results of this research demonstrate the need to evaluate the real performance of DMTs and, based on these results, to encourage investment or disinvestment decisions in the health systems.

Supporting information

S1 Table. Characteristics of Brazilian patients with RRMS included in the study after matching by propensity score, 2000–2015.

https://doi.org/10.1371/journal.pone.0238476.s001

(DOCX)

References

  1. 1. Howard J, Trevick S, Younger DS. Epidemiology of Multiple Sclerosis. Neurol Clin. 2016; 34(4): 919–39. pmid:27720001
  2. 2. Lublin FD, Reingold SC, Cohen J a, Cutter GR, Sorensen PS, Thompson AJ, et al. Defining the clinical course of multiple sclerosis: The 2013 revisions. Neurology. 2014; 83(3): 278–86. pmid:24871874
  3. 3. Compston A, Coles A. Multiple sclerosis. Lancet. 2008; 372(9648): 1502–17. pmid:18970977
  4. 4. Atlas of MS 2013: Mapping Multiple Sclerosis Around the World. London: Multiple Sclerosis International Federation; 2013. [Cited 2019 November 19]. Available from: http://www.msif.org/about-ms/publications-and-resources/.
  5. 5. Beatriz A, Nogueira C, Clinete M, Lacativa S, Ferreira F, Pereira C, et al. Prevalence of multiple sclerosis in Brazil: A systematic review. Mult Scler Relat Disord. 2015; 4(6): 572–9. pmid:26590664
  6. 6. ABEM. O que é Esclerose Múltipla? Associação Brasileira de Esclerose Múltipla; 2016. [Cited 2019 November 19]. Available from: http://abem.org.br/esclerose/o-que-e-esclerose-multipla/
  7. 7. Tabansky I, Stern JNH, Goldstein J, Jeganathan V, Wright P, Sands W, et al. Advancing drug delivery systems for the treatment of multiple sclerosis. Immunol Res. 2015; 63(1–3): 58–69. pmid:26475738
  8. 8. Giovannoni G, Butzkueven H, Pepper G, Sormani MP, Thalheim C, Vollmer T, et al. Brain health: time matters in multiple sclerosis. Mult Scler Relat Disord. 2016; 9: S5–48. pmid:27640924
  9. 9. Secretaria de Atenção à Saúde, Ministério da Saúde, República Federativa do Brasil. Portaria n° 391, de 05 de maio de 2015. Diário Oficial da União. Brasília, DF: Imprensa Nacional; 2015. p. 40.
  10. 10. Guerra Junior AA, Pires De Lemos LL, Godman B, Bennie M, Osorio-De-Castro CGS, Alvares J, et al. Health technology performance assessment: Real-world evidence for public healthcare sustainability. Int J Technol Assess Health Care. 2017;33(2): 279–87. pmid:28641588
  11. 11. Corrigan-Curay J, Sacks L, Woodcock J. Real-World Evidence and Real-World Data for Evaluating Drug Safety and Effectiveness. JAMA. 2018; 320(9): 867. pmid:30105359
  12. 12. Ingwersen J, Aktas O, Hartung H. Advances in and Algorithms for the Treatment of Relapsing-Remitting Multiple Sclerosis. Neurotherapeutics. 2016;13(1):47–57. pmid:26701666
  13. 13. CONITEC. Comissão Nacional de Incorporação de Tecnologias no SUS. Recomendações da CONITEC [Internet]. [cited 2019 November 2019]. Available from: http://conitec.gov.br/index.php/decisoes-sobre-incorporacoes
  14. 14. Vieira FS. Produção de Informação de Custos para a Tomada de Decisão no Sistema Único de Saúde: uma questão para a política pública. Instituto de Pesquisa Econômica Aplicada, Ministério do Planejamento, Desenvolvimento e Gestão. Rio de Janeiro: Ipea; 2017. p. 50.
  15. 15. Guerra Júnior AA, Pereira RG, Gurgel EI, Cherchiglia M, Dias LV, Ávila J, et al. Building the National Database of Health Centered on the Individual: Administrative and Epidemiological Record Linkage—Brazil, 2000–2015. Int J Popul Data Sci. 2018; 3(1).
  16. 16. Cherchiglia ML, Guerra Júnior AA, Andrade EIG, Machado CJ, Acurcio FA, Meira Júnior W, et al. A construção da base de dados nacional em Terapia Renal Substitutiva (TRS) centrada no indivíduo: aplicação do método de linkage determinístico-probabilístico. Rev Bras Estud Popul. 2007; 24(1): 163–7.
  17. 17. Paim J, Travassos C, Almeida C, Bahia L, MacInko J. The Brazilian health system: History, advances, and challenges. Lancet. 2011; 377(9779):1778–97. pmid:21561655
  18. 18. Acurcio FA, Machado MAA, Moura CS, Ferre F, Guerra Junior AA, Andrade EIG, et al. Medication Persistence of Disease-Modifying Antirheumatic Drugs and Anti–Tumor Necrosis Factor Agents in a Cohort of Patients With Rheumatoid Arthritis in Brazil. Arthritis Care Res. 2016; 68(10): 1489–96.
  19. 19. Diniz IM, Guerra Junior AA, Lovato Pires de Lemos L, Souza KM, Bennie M, Wettermark B, et al. The long-term costs for treating multiple sclerosis in a 16-year retrospective cohort study in Brazil. PLoS One. 2018; 13(6):e0199446. pmid:29928006
  20. 20. Evans C, Tam J, Kingwell E, Oger J, Tremlett H. Long-Term Persistence With the Immunomodulatory Drugs for Multiple Sclerosis: A Retrospective Database Study. Clin Ther. 2012; 34(2): 341–50. pmid:22296946
  21. 21. Quan H, Fong A, Burnand B, Saunders LD, Feasby TE, Ghali WA. Coding Algorithms for Defining Comorbidities in ICD-9-CM and ICD-10 Administrative Data. Med Care. 2005; 43(11): 1130–1139. pmid:16224307
  22. 22. Randolph JJ, Falbe K, Manuel AK, Balloun JL. A Step-by-Step Guide to Propensity Score Matching in R Information on the Dataset Used. Pract Assessment, Res Eval. 2014; 19(18): 1–6.
  23. 23. Austin PC. Optimal caliper widths for propensity-score matching when estimating differences in means and differences in proportions in observational studies. Pharm Stat. 2011; 10(2): 150–61. pmid:20925139
  24. 24. Carvalho MS, Andreozzi VL, Codeço CT, Campos DP, Barbosa MTS, Shimakura SE. Análise de sobrevivência: teoria e aplicações em saúde. 2nd. ed. Fiocruz: Fiocruz press; 2011. 432 p.
  25. 25. Benjamini Y, Hochberg Y. Controlling the false discovery date: A practical and powerful approach to multiple testing. J R Stat Soc. 1995; 57(1): 289–300.
  26. 26. Tripathi A, Pandey A. Post-Hoc Comparison in Survival Analysis: An Easy Approach. J Biosci Med. 2017;5(3):112–9.
  27. 27. Kalincik T, Jokubaitis V, Izquierdo G, Duquette P, Girard M, Grammond P, et al. Comparative effectiveness of glatiramer acetate and interferon beta formulations in relapsing–remitting multiple sclerosis. Mult Scler J. 2014; 21(9): 1159–71.
  28. 28. Secretaria de Atenção à Saúde, Ministério da Saúde, República Federativa do Brasil. Portaria n° 97, de 22 de março de 2001. Diário Oficial da União. Brasília, DF: Imprensa Nacional; 2001. p. 114.
  29. 29. Secretaria de Atenção à Saúde, Ministério da Saúde, República Federativa do Brasil. Portaria n° 493, de 23 de setembro de 2010. Diário Oficial da União. Brasília, DF: Imprensa Nacional; 2010. p. 153.
  30. 30. Halpern R, Agarwal S, Dembek C, Borton L, Lopez-Bresnahan M. Comparison of adherence and persistence among multiple sclerosis patients treated with disease-modifying therapies: A retrospective administrative claims analysis. Patient Prefer Adherence. 2011; 5:73–84. pmid:21423591
  31. 31. Reynolds MW, Stephen R, Seaman C, Rajagopalan K. Persistence and adherence to disease modifying drugs among patients with multiple sclerosis. Curr Med Res Opin. 2010; 26(3):663–74. pmid:20070144
  32. 32. Marrie RA, Cohen J, Stuve O, Trojano M, Sørensen PS, Reingold S, et al. A systematic review of the incidence and prevalence of comorbidity in multiple sclerosis: Overview. Mult Scler J. 2015; 21(3):263–81.
  33. 33. Marrie RA, Horwitz R, Cutter G, Tyry T, Campagnolo D, Vollmer T. Comorbidity, socioeconomic status and multiple sclerosis. Mult Scler. 2008;14(8):1091–8. pmid:18728060
  34. 34. Simpson RJ, McLean G, Guthrie B, Mair F, Mercer SW. Physical and mental health comorbidity is common in people with multiple sclerosis: Nationally representative cross-sectional population database analysis. BMC Neurol. 2014;14(1):1–8.
  35. 35. Laroni A, Signori A, Maniscalco GT, Lanzillo R, Russo CV, Binello E, et al. Assessing association of comorbidities with treatment choice and persistence in MS: A real-life multicenter study. Neurology. 2017; 89(22):2222–9. pmid:29093064
  36. 36. Lanzillo R, Saccà F, Signori A, Maniscalco GT, Signoriello E, Lo Fermo S, et al. Determinants of therapy switch in multiple sclerosis treatment-naïve patients: A real-life study. Mult Scler J. 2018; 1–10.
  37. 37. Gajofatto A, Benedetti MD. Treatment strategies for multiple sclerosis: When to start, when to change, when to stop? World J Clin Cases. 2015; 3(7):545. pmid:26244148
  38. 38. Pinkerton JA V, Guico-Pabia CJ, Taylor HS. Menstrual cycle-related exacerbation of disease. Am J Obstet Gynecol. 2010; 202(3):221–31. pmid:20207238
  39. 39. Bove R, Healy BC, Musallam A, Glanz BI, De Jager PL, Chitnis T. Exploration of changes in disability after menopause in a longitudinal multiple sclerosis cohort. Mult Scler. 2016; 22(7):935–43. pmid:26447063
  40. 40. Zhornitsky S, Greenfield J, Koch MW, Patten SB, Harris C, Wall W, et al. Long-term persistence with injectable therapy in relapsing-remitting multiple sclerosis: An 18-year observational cohort study. PLoS One. 2015; 10(4):1–13.
  41. 41. Melendez-Torres GJ, Auguste P, Armoiry X, Maheswaran H, Court R, Madan J, et al. Clinical effectiveness and cost-effectiveness of beta-interferon and glatiramer acetate for treating multiple sclerosis: Systematic review and economic evaluation. Health Technol Assess (Rockv). 2017; 21(52).
  42. 42. Lucchetta RC, Tonin FS, Borba HHL, et al. Disease-Modifying Therapies for Relapsing-Remitting Multiple Sclerosis: A Network Meta-Analysis. CNS Drugs. 2018; 32(9):813–826. pmid:30014314
  43. 43. Tramacere I, Del Giovane C, Salanti G, D'Amico R, Filippini G. Immunomodulators and immunosuppressants for relapsing-remitting multiple sclerosis: a network meta-analysis. Cochrane Database of Systematic Reviews 2015, Issue 9. pmid:26384035
  44. 44. Jokubaitis VG, Spelman T, Lechner-Scott J, Barnett M, Shaw C, Vucic S, et al. The Australian Multiple Sclerosis (MS) Immunotherapy Study: A Prospective, Multicentre Study of Drug Utilization Using the MSBase Platform. Oreja-Guevara C, organizador. PLoS One. 2013; 8(3):e59694. pmid:23527252
  45. 45. Kleinman NL, Beren IA, Rajagopalan K, Brook RA. Medication adherence with disease modifying treatments for multiple sclerosis among US employees. J Med Econ. 2010; 13(4):633–40. pmid:20958113