Introduction

Immunomodulation with interferon-beta (INFb) is widely used to treat patients with relapsing remitting multiple sclerosis (RRMS). Phase 3 experimental data indicate that INFb may delay the progression of disability. One of the measures available for monitoring disability in MS patients is the Multiple Sclerosis Functional Composite (MSFC). The MSFC consists of simple quantitative measures for ambulation, arm function and cognition combined into a single score [4]. In international phase 3 trials, the MSFC has demonstrated excellent intra-rater reliability [2, 3]. Studies demonstrating a correlation between MSFC and health-related quality of life (HR-QoL) were mostly single-centred, limited to academic hospitals or phase 3 trials [8, 10], and typically involved a care setting not representative of daily practice.

Phase 3 trials did not adequately address whether INFb can positively affect HR-QoL. As yet, evidence from phase 4 investigations is not sufficiently informative as studies were retrospective, their duration did not exceed 12 months [1, 14] or patients were treated with different types of immunomodulatory drugs [7].

For 2 years we studied 284 RRMS patients treated with INFb-1a (Avonex®) in daily practice. The study was multi-centre (N = 36) and involved neurologists in various hospital settings in four European countries. Main outcome measures were MSFC score, HR-QoL and their interrelationship. Here we report the results.

Patients and methods

Study design and study period

The present study was a prospective, multi-centre, observational phase 4 study (ClinicalTrials.gov identifier NCT00534261). The decision to inform patients about the study was preceded by and independent from the decision to start INFb-1a (Avonex®). Written informed consent was obtained prior to conduct of any study-related procedure.

The first patient was enrolled 4 November 1999. The last patient completed the study 11 February 2004.

Inclusion and exclusion criteria

Inclusion criteria were: (1) RRMS, (2) age 18–70 years, inclusive, (3) two relapses in preceding 24 months, (4) disease duration at least 12 months, (5) EDSS 5.5 or less, (6) naïve for INFb, (7) written informed consent prior to any testing and evaluations not part of routine care.

Exclusion criteria were: (1) no RRMS, (2) history of any significant cardiac, hepatic, pulmonary or renal disease, immune deficiency, or other medical conditions that preclude therapy with INFb, (3) history of severe allergic or anaphylactic reactions or history of hypersensitivity to human albumin, (4) history of seizures in the last 3 months, (5) history of intolerance to acetaminophen, naproxen or other non-steroidal-anti-inflammatory drugs, precluding use of at least one of these drugs, (6) history of intolerance to INFs, (7) pregnant or breast feeding, (8) unwillingness to practice effective contraception, unless postmenopausal or surgically sterile, (9) women considering becoming pregnant during the study, (10) previous participation.

Study drug

INFb-1a (Avonex®) was commercially available and administered intramuscularly once a week. All patients who self injected were trained and could use the Avonex® Training Program.

Study procedures

The study was investigator-initiated (PJ) and sponsored by Biogen International. Investigating neurologists practiced in general hospitals of varying size, academic hospitals or MS-clinics (see Appendix). Investigators met before the start of the study and after 1 year of enrollment. The study was conducted according to the Declaration of Helsinki, the ICH guidelines on GCP [6], and in compliance with national drug laws and approved by central and local ethical committees. Patients that withdrew before completion were not replaced once first dose of Avonex® had been administered, and were followed-up within 2 weeks of discontinuation as per month 24.

Data quality assurance

Contract research organization (CRO) Kendle International was responsible for management and monitoring. Kendle periodically monitored study sites for data quality assurance. Double data entry (TrialBase™), automated edits and manual review guaranteed data completeness and consistency. Queries were forwarded to investigators. After transfer of data from database to statistical package (SAS, version 8.2 [SAS Institute Inc. Cary NC, USA, 1997]) data correctness was validated.

Assessments

MSFC and HR-QoL were assessed at Day 1 and Months 3, 6, 12, 18, and 24; EDSS at Day 1 and Month 24. Assessments were conducted between 10:00 am and 15:00 pm.

The MSFC tests ambulation (Timed 25-Foot Walk Test [Timed 25-FWT]), arm function (9 Hole Peg Test [9-HPT]), and cognition (Paced Auditory Serial Addition Test [PASAT 3″]). The MSFC was calculated as the average of the z-scores of these three measures. Reference population was the study population at baseline. An increase of MSFC indicates improvement, decrease a worsening. The MSFC was performed three times within 1 month prior to Day 0 with at least 5 days in between evaluations, according to recommended procedures.

HR-QoL was assessed by the MS54 Quality of Life (MS54QoL) questionnaire [15]. The MS54QoL is a psychometrically validated MS-specific multi-dimensional inventory of patient-centered health status. It consists of the 36-item Short Form (SF-36) health survey as a generic core measure to enable comparisons to other patient populations and to the general population, supplemented with 18 additional questions exploring items relevant to patients with MS in the areas of health distress (four items), sexual function (four items), satisfaction with sexual function (one item), overall quality of life (two items), cognitive function (four items), energy (one item), pain (one item) and social function (one item) [15]. The MS54QoL contains 52 items distributed into 12 scales, and two single items. A physical and a mental dimension underlie the MS54QoL: the physical and mental domains [15]. Scores for each domain range from 0 to 100, where higher values indicate better HR-QoL. For validation of the Dutch and French versions, translations were validated against the English version. Patients received questionnaires during study visits and mailed completed questionnaires to the study site. Care partners were allowed to assist in completing the questionnaires.

Statistical analyses

The statistical analyses were performed by Kendle International and the Department of Epidemiology, Biostatistics and Health Technology Assessment of the Radboud University Nijmegen Medical Centre.

Pearson’s correlation coefficient was determined between the MSFC and the physical and mental domains of MS54Qol, and between the MSFC and the EDSS. Global trends over time were analysed using a linear mixed model, with random factor patient and fixed variables time and time squared. When the latter variable was not significant, it was removed.

In order to separately assess HR-QoL during the initial follow-up period (months 3–12) and during the final period (months 18–24), the difference from baseline was calculated for each follow-up assessment and mixed model analyses with random factor patient were carried out, separately for each period. The impacts of the factors age, gender, disease duration, pre-treatment annualized relapse rate, and baseline values for MSFC, Timed 25-FWT, 9-HPT, PASAT 3″ and EDSS, were evaluated by including these factors in the model. This was done in a stepwise manner: the factors had to be significant at the 5% level to be included in the model and had to be significant at the 10% level to remain in the model.

In order to evaluate the impact of patients who dropped out we compared the baseline characteristics of the patients who completed the study and the drop outs, we did an additional analysis on the completers only, and we carried out a last observation carried forward analysis.

In additional analyses, we also investigated the possible impact of language and centre differences by including corresponding additional factors and interaction terms in the statistical models.

Results

Patients and sites

Forty-six neurologists in 36 neurological practices enrolled 284 patients in the Netherlands (17 sites, 151 patients), Belgium (16 sites, 117 patients, of which 53 used the French version of the questionnaires), U.K. (two sites, 15 patients) and Luxemburg (one site, one patient). Twenty-two practises were in general hospitals, 10 in university hospitals and 4 in independent MS clinics.

Mean age of patients was 38.6 years (standard deviation [SD], 10.1). Eighty-nine (31.3%) patients were male, 195 (68.7%) female. Mean time since onset of symptoms was 6.6 years (SD, 6.6), mean time since diagnosis was 3.5 years (SD, 5.1), and mean number of relapses in the preceding 2 years was 2.2 (SD, 1.0). Mean EDSS score was 2.4 (SD, 1.2). Two-hundred-and-four patients (71.8%) completed the study. Mean time to withdrawal was 329.7 (SD, 224.7) days. Reasons for withdrawal are given in Table 1.

Table 1 Disposition of patients
Full size table

Disability and HR-QoL during the study

Values for MSFC during the study are presented in Table 2.

Table 2 Changes in MSFC values
Full size table

MSFC scores did not change significantly (p = 0.70), neither did EDSS scores (p = 0.24) (data not shown). Changes in MS54Qol scores during the study are presented in Table 3.

Table 3 Changes in MS54Qol scores for physical and mental domains
Full size table

Mean physical score improved from 56.6 to 61.0 (p < 0.05) and mean mental score showed a trend to improvement, from 57.2 to 61.1 (p = 0.07). Improvements from baseline to Month 24 were also seen for the Dutch and French versions separately.

Correlative analysis

Correlation between the MSFC and the physical domain of MS54QoL was −0.40 (p < 0.05). A smaller correlation was demonstrated between the MSFC and the mental domain of MS54QoL: −0.25 (p < 0.05). MSFC’s correlation to the EDSS was −0.46 (p < 0.05).

Factors relating to HR-QoL

Improvement of the physical score of MS54QoL was associated with lower age, lower EDSS score, less time for Timed 25-FWT, and higher PASAT3″ score at baseline. Improvement of the mental score was more likely to occur in patients with high PASAT3″ or low EDSS values.

Additional analyses: drop outs, language and centre

Completers and drop outs had similar EDSS scores at baseline: 2.4 versus 2.2. The mean MS54Qol scores for the physical domain differed significantly: 57.1 in completers versus 45.9 in drop-out patients. Scores on the mental domain differed similarly: 58.8 versus 47.8. However, the results of an analysis on completers only as well as the results of a last observation carried forward analysis were similar to the results of the primary analysis. No statistically significant impact of language and centre on the results was found.

Discussion

In a prospective study of 284 RRMS patients, treated in 36 neurological practices, we demonstrated that HR-QoL improved during 2 years of treatment with intramuscular INFb-1a. In 1999 Rice et al. [12] reported that patients with relapsing forms of MS treated with IFNb-1b had better HR-QoL than untreated patients, especially when EDSS was less than 3.0. A recent report on 210 MS patients showed that disease modifying treatment was associated with a significant and sustained increase in HR-QoL [7]. As the study, however, included patients treated with various types of INFb or glatiramer acetate, a conclusion on the effect of INFb or a given product could not be drawn. Others demonstrated that IFNb had a negative impact on HR-QoL, influencing mainly the mental domain [13]. Notably, in the first 6 months, IFNb therapy was shown to have a mixed impact by improving role-physical functioning and transiently worsening experienced bodily pain [1]. Conceivably, after start of therapy, temporary side-effects may interfere negatively with HR-QoL. A recent study was unable to detect significant changes in HR-QoL in the first 12 months of INFb therapy [16], whereas a previous report on patients treated with intramuscular INFb-1a found no negative effect on HR-QoL [14] in the first 12 months. The duration of our study of 24 months and the homogeneity of the study population—all patients were treated with intramuscular INFb-1a—may have increased the chance of detecting a significant change in HR-QoL.

We found that HR-QoL correlated with the MSFC. To validate the MSFC, Miller et al. [9] performed a cross-sectional study in 300 MS patients in four hospitals in the USA and Canada and showed that MSFC scores correlated with patient-reported physical functioning and, more weakly, with emotional functioning. Likewise, we found that the MSFC’s correlation with physical HR-QoL was stronger than with the mental domain. Miller et al. [9] also demonstrated a strong correlation between the MSFC and the EDSS of r = −0.80. We, however, found a weak correlation, which may be explained by the fact that the severity of the disease varied less in our study and even more so by the fact that only patients with an EDSS of 5.5 or less were included. For EDSS scores between 0 and 5.0, the figure in Miller’s article also suggests a much weaker correlation [9]. In addition, our study involved a larger number of neurologists, who were to varying degrees experienced in EDSS and who worked in a variety of clinical settings.

It would be helpful to identify patients at risk for worsening of HR-QoL during immunomodulating treatment. A recent single-centre, 2-year study in 240 MS patients showed that higher EDSS, presence of clinical depression, and use of support services, pain medications or antidepressants at baseline were associated with poorer HR-QoL at follow-up, while higher income and higher education were associated with better HR-QoL [5]. Another single-centre study demonstrated that EDSS contributed to both physical and mental HR-QoL, and that age had an effect on the physical, but not on the mental, dimension [11]. In our patient group, higher EDSS at baseline was associated with lower HR-QoL after 2 years of intramuscular INFb-1a treatment, as were older age (physical), higher Timed 25-FWT (physical and mental) and lower PASAT3″ (mental). These associations between higher disability/older age at baseline and poorer HR-QoL at follow-up may be considered to favour the concept of early treatment. Our finding that mean scores for physical and mental HR-QoL at baseline were lower in patients who did not complete the 24-month treatment period suggests that patients with low HR-QoL at the start of INFb treatment may be given extra care to prevent early discontinuation.

In conclusion, for up to 2 years we studied 284 RRMS patients treated with INFb-1a (Avonex®) in settings reflecting daily practice and found that after 2 years of treatment HR-QoL was increased, especially in younger patients with low disability.