Long-Term Adherence to IFN Beta-1a Treatment when Using RebiSmart® Device in Patients with Relapsing-Remitting Multiple Sclerosis

The effectiveness of disease-modifying drugs in the treatment of multiple sclerosis is associated with adherence. RebiSmart® electronic device provides useful information about adherence to the treatment with subcutaneous (sc) interferon (IFN) β-1a (Rebif®). The aim of the study was to determine long-term adherence to this treatment in patients with relapsing-remitting multiple sclerosis (RRMS). This retrospective multicentre observational study analysed 258 patients with RRMS who were receiving sc IFN β-1a (Rebif®) treatment by using RebiSmart® until replacement (36 months maximum lifetime) or treatment discontinuation. Adherence was calculated with data (injection dosage, time, and date) automatically recorded by RebiSmart®. Patients in the study had a mean age of 41 years with a female proportion of 68%. Mean EDSS score at start of treatment was 1.8 (95% CI, 1.6–1.9). Overall adherence was 92.6% (95% CI, 90.6–94.5%). A total of 30.2% of patients achieved an adherence rate of 100%, 80.6% at least 90%, and only 13.2% of patients showed a suboptimal adherence (<80%). A total of 59.9% of subjects were relapse-free after treatment initiation. Among 106 subjects (41.1%) who experienced, on average, 1.4 relapses, the majority were mild (40.6%) or moderate (47.2%). Having experienced relapses from the beginning of the treatment was the only variable significantly related to achieving an adherence of at least 80% (OR = 3.06, 1.28–7.31). Results of this study indicate that sc IFN β-1a administration facilitated by RebiSmart® could lead to high rates of adherence to a prescribed dose regimen over 36 months.


Introduction
Multiple sclerosis (MS) is an autoimmune and chronic disease characterized by inflammation, demyelination and axonal degeneration occurring in the central nervous system [1,2]. Most patients are between 25-35 years of age when diagnosed, and they are mainly women. The prevalence in Spain ranges around 70-80 to 125 cases per 100,000 inhabitants, depending on the geographical area, and the used methodology [3,4]. Approximately 70-80% of patients develops the relapsing-remitting form (RRMS), defined by recurrent episodes of neurological dysfunction, and followed by partial or full recovery [5]. Although to date there is no cure for MS, first-line treatment with disease-modifying drugs (DMDs), such as interferon (IFN) β-1a, IFN β-1b, and glatiramer acetate, have demonstrated to reduce the incidence and severity of relapses and disability progression [6][7][8][9]. Apart from this partial effect, the efficacy of treatments is being compromised by the lack of persistence and adherence to the prescribed duration, interval and dosing [10,11]. Patients with poor adherence, or who discontinued the treatment, have reported higher incidence of relapses and experienced worse quality of life than adherent patients [12][13][14][15]. Indeed, adherence to injectable DMD therapies is generally suboptimal [11].
The Global Adherence Project (GAP), a multicentre, multinational phase IV study that involved 2,566 patients with RRMS, revealed a non-adherence rate of 25% to prescribed regimens (intramuscular IFN β-1a, subcutaneous, sc, β-1a 22 and 44μg, IFN β-1b, and glatiramer acetate) [15]. In another study performed in Spain [16], the proportion of adherent patients to sc IFN β-1b, assessed by the Morisky-Green test, was 68.3%, being indicative of poor adherence. Primary reasons described for non-adherence include adverse reactions (AEs), 'flu-like symptoms', injection anxiety, perceived lack of efficacy, or forgetfulness [15][16][17][18][19]. Among the strategies aimed at improving the adherence to sc IFN β-1a treatments is the use of autoinjection devices, such as RebiSmart 1 (Merck Serono SA-Geneva, Switzerland), which also increase significantly the satisfaction of patients [20,21]. In BRIDGE (RebiSmart 1 to self-inject Rebif 1 serum-free formulation in a multidose cartridge), a 12-week study involving 119 patients with RRMS, the adherence rate was 88.2% [22]. Despite these encouraging short-term results, and with the unique exception of a recent published study [23], to date there is no study evaluating the long-term adherence to sc IFN β-1a treatments in clinical practice. Furthermore, no study has been conducted in Spain with this drug. Therefore, the aim of the present study was to determine long-term adherence to sc IFN β-1a treatment administered with the RebiSmart 1 device in patients with RRMS.

Materials and Methods
This retrospective, multicentre, observational nationwide study analysed 258 patients with RRMS who were receiving sc IFN β-1a treatment (Rebif 1 ) by using RebiSmart 1 device. This study was carried out in 29 hospitals from Spain (Fig 1).
Criteria for study inclusion were as follows: aged 18 or over; diagnosed with RRMS; receiving sc treatment with IFN β-1a by means of RebiSmart 1 ; using RebiSmart 1 device until replacement (36 months maximum lifetime) or treatment discontinuation, by any cause; electronic download of adherence data stored in the RebiSmart 1 device (using the Mitra 1 software version 1.5) from start of treatment to replacement/return of the RebiSmart 1 device; and signed informed consent. Exclusion criteria were: having been diagnosed with clinically isolated syndrome, or having primary or secondary progressive MS with no relapses. Patients were invited to participate in the study at the time of a scheduled visit for device replacement. The recruitment window for collecting RebiSmart 1 devices was 6 months, from June 2013 to December 2013. Moreover, patients who discontinued the treatment and returned the device between October 2009 and August 2013 were also invited to participate in the study. Procedures were performed in accordance with guidelines established by the Ethics Committee of each participating center, and the Declaration of Helsinki.

Study endpoints
The primary assessment was adherence to the treatment over the retrospective period of observation, i.e. from the start of sc IFN β-1a treatment to the time of device replacement or treatment discontinuation. Overall adherence was calculated as follows: 100 X total number of sc IFN bÀ1a administrations total number of days X 7 days 3 administrations Adherence was quantified by using the data (dosage, time, and date) automatically recorded by RebiSmart 1 . The primary endpoint was overall adherence (as continuous outcome) doi:10.1371/journal.pone.0160313.g001 Long-Term Adherence Using RebiSmart1 in RRMS supported by the secondary endpoint, percentage of subjects with at least !60%, !70%, !80%, !90%, and 100% adherence.
Data from adherence were analysed by using the overall period of study and by quarters. Secondary endpoints included clinical and demographic characteristics of subjects treated with sc IFN β-1a by using RebiSmart 1 , the evaluation of disease relapses during the study period, treatment discontinuation including reasons, and subsequent treatments after discontinuation. Disease relapse was defined as the reappearance of neurological symptoms during at least 24 hours after a period of time with stable neurological status or after at least 30 days of improvement. Relapses were classified according to the severity (mild, moderate, or sever) of the episodes. The Expanded Disability Status Scale (EDSS) score was estimated at the time of starting sc IFN β-1a treatment and when included in the study at the time of device replacement or treatment discontinuation.

Statistical Analysis
Qualitative variables were expressed as absolute and relative frequencies, and quantitative variables as the mean with the standard deviation (SD) or the 95% confidence interval (95% CI). Confidence intervals for rates of adherence were calculated by using Wilson score and Clopper-Pearson (Exact) method. A repeated measures analysis of variance (ANOVA) was performed to evaluate differences in overall adherence over time. Statistical significance was established when p 0.05. A logistic regression model was estimated to identify demographic and clinical characteristics of patients (including age, sex, geographic location, using a different IFN β-1a as first DMD treatment, number and severity of relapses, and duration of the treatment) related to achieving an adherence !80%. All statistical analyses were performed by using SAS software 9.2.

Results
A total of 276 patients were recruited in the study, but only 258 fulfilled all inclusion criteria and were thus included in the analysis. The study population consisted of mainly women (67.8% of patients), Caucasian patients (98.8%), with a mean age of 40.7 years and mean body mass index of 23.9 Kg/m 2 . Demographic and clinical characteristics of subjects are shown in Table 1.
Half of patients (51.9%) were employed at the time of inclusion. A total of 34.9% completed the high school and 29.1% the university. Only 3.9% of patients required a caregiver for assistance. The mean time since diagnosis of RRMS and from onset of symptoms was 8.9 years and 11.3 years respectively. Most of subjects (81.8%) had been using sc IFN β-1a as first DMD treatment. Patients have been receiving Rebif 1 with RebiSmart 1 device for 3.1 years. Since the beginning of the IFN β-1a treatment, 106 patients (41.1%) had experienced on average 1.4 relapses (SD: 1.0), with maximum severity being mainly moderate (47.2% of subjects with relapses) or mild (40.6%). Mean EDSS score was 1.8 (SD: 1.2) at the time of sc IFN β-1a treatment onset and 2.0 (SD: 1.6) at the study inclusion visit (time of device replacement or treatment discontinuation).
From there, overall adherence decreased slightly until 90.4; 95% CI: 87.4-93.3) at the time of device replacement (n = 150). No significant differences were found between rates of adherence through the study period. The percentage of subjects with suboptimal adherence by 3 months periods was about 10% throughout the study period (range: 8.1-13.2%).
According to the logistic regression model, having experienced relapses from the beginning of the sc IFN β-1a treatment was the only variable associated with achieving an adherence of at least 80% (OR: 3.06, 95% CI: 1.28-7.31). The incidence of relapses decreased over time, from 15 cases at the beginning of the treatment (5.8%, n = 258 at 0-3 months) to 6 cases at the time of device replacement (4.0%, n = 150 at 33-36 months; Fig 4). Relapses were mainly mild or moderate in severity.

Discussion
Despite the demonstrated clinical effectiveness of DMDs by reducing the incidence and severity of relapses and disability progression in RRMS [4][5][6][7]24], adherence to injectable DMD therapies is generally suboptimal [10], which compromises their potential beneficial effects. Data from the GAP project revealed an overall non-adherence rate of 25% to intramuscular IFN β-1a, sc IFN β-1a, IFN β-1b, and glatiramer acetate first-line DMD treatments [15]. Moreover, adherent patients showed a significant better quality of life, fewer neuropsychological issues, shorter duration of disease, and shorter duration of therapy than non-adherent patients. These results highlighted the importance of adherence on the effectiveness of the DMD treatments and the need of identify the causes that lead to non-adherence. The most common reported reasons for non-adherence included forgetting to administer the injection, followed by tiredness of taking injections, fatigue, 'flu-like symptoms', and pain.
Maintaining the adherence to injectable treatments may be a challenge for some patients when managing a chronic disease. Electronic injection devices, such as RebiSmart 1 , have been developed with the purpose to overcome specific problems associated with the injection (needle phobias, anxiety, or pain at the injection site) [21,25,26]. When comparing with manual injections, autoinjectors have demonstrated to improve the injection tolerability and thus the satisfaction of the subject [27][28][29][30]. For this reason, the advance of these electronic injection devices has led to the improvement of adherence to sc IFN β-1a treatment, as demonstrated in the 12-week BRIDGE study, with a rate of adherence of 88.2% [22]. Subjects reacted favourably to those features of the device associated with handling, while being easy to use. Similarly, it has been recently published a study of adherence over 24 months in 225 patients with RRMS from the United Kingdom (UK) and Ireland who were receiving sc IFN β-1a treatment by RebiSmart 1 [23]. In this population, the rate of overall adherence was 95.0%, showing 92.0% of subjects with an adherence !80% at 12 months and 91.1% at 24 months. In our study, the overall rate over the course of 36 months (92.6%) was in concordance with previous studies using the same RebiSmart 1 device, corroborating its beneficial effect on improving the adherence to the treatment. In our study, the only demographic and clinical variable from subjects related with an adherence of at least 80% was having (or not) experienced relapses from the beginning of the sc IFN β-1a treatment. Indeed, suboptimal adherence was about 3 times higher in subjects who had suffered relapses than in those who had not. Furthermore, 58.9% of the subjects did not experience any relapse since the beginning of the treatment, thus corroborating its effectiveness in reducing the incidence of relapses [4][5][6][7][8]. A total of 34.5% of subjects discontinued the treatment before device replacement because of the occurrence of adverse events, perceived lack of efficacy, voluntary decision, or non-adherence. In our study, discontinuation remained at a constant rate through the study period, in contrast to what has been reported in the literature, occurring more likely in the first 6 months of treatment [10]. Those subjects who discontinued the treatment, mainly switched to glatiramer acetate, or to the second-line agents fingolimod or natalizumab. RebiSmart 1 device also can record the dose history, allowing the capture of accurate and objective information, useful to detect suboptimal adherence in patients and to avoid potential lack of efficacy. This feature becomes especially important for patients with memory or attention impairment. The present study has several strengths. It evaluated the effect of using an electronic device which may inform about the adherence to disease-modifying therapies. This fact allows avoiding healthcare database as a source of information. Therefore, this device is able to monitor adherence objectively in contrast to database studies, using medication possession ratio. The main limitation of the study was its retrospective nature. Furthermore, there is the possibility of a selection bias derived from the fact that patients aware of their non-adherence may tend to not participate voluntarily in the study. However, and unfortunately, this potential bias is intrinsically present in all studies in this field. Another issue, is the fact that although the patients were selected by the time of device replacement or treatment discontinuation, the age range was relatively small, and therefore we must accept the possibility that a hidden age selection bias could have been inadvertently occurred, when prescribing the RebiSmart1 device, as younger patients are more prone to use efficiently new technologies.
In conclusion, results of this study indicate that sc IFN β-1a administration facilitated by RebiSmart 1 in patients that accepted it, could lead to high rates of adherence to a prescribed dose regimen over 36 months. This adherence was above 90% on average over the three years of use. Advances in electronic injection devices such as RebiSmart 1 permit adherence to be recorded and monitored by the device itself, thereby improving comfort, subject satisfaction, and the adherence to the treatment of patients in this chronic disease.