Multiple Sclerosis (MS) is a chronic inflammatory disease of the central nervous system causing neurodegeneration, axonal injury and demyelination, and is responsible for significant physical disability and social impact. Direct and indirect costs associated with MS are high, due to its chronic management and to early onset in economically productive years [1,2].

The introduction of Disease Modifying Treatments (DMTs) during the last 20 years has been changing the natural history of MS. Clinical trials support the early initiation of chronic DMTs in order to avoid short term disease activity (i.e. relapses), and to reduce or, ideally, to prevent long term disability accrual [3–6]. However, few studies have been conducted in the long term and, apparently, the use of DMTs might be responsible for delaying the reaching of milestones of disability progression [4,7]. Notwithstanding this, DMTs are nowadays the main responsible for direct healthcare costs in MS, and are expected to have a further expansion among overall expenses as a consequence of the introduction of newer and more effective drugs characterized by high healthcare costs for administration and management [8–11].

Soaring healthcare costs for DMTs have received much attention in health policy contexts, with some studies estimating DMT costs being beyond the health care system tolerance, whereas others below the threshold [11,12]. In particular, there are no investigations on whether the long-term economic burden for administration and management of DMTs is associated with MS clinical evolution [10,13]. Therefore, the present retrospective cohort study aims to explore the relationships between the expenditure for MS treatments, and the risk of relapses and of disability progression during a 10-year observation period.

Among 685 eligible subjects, the present study included 544 newly diagnosed, drug naïve RRMS patients, who were followed-up for 10.2±3.4 years. Details of subjects assessed for eligibility and excluded or included in the present study, are reported in Fig 1.

Demographic features, clinical findings and healthcare costs of the MS cohort are reported in Table 1. Additional clinical and economic characteristics of the present population have been fully described elsewhere [18].

In agreement with the indications of the Italian regulatory agency, the first prescribed DMT was discontinued and possibly changed in 291 patients (53.4%), which underwent 2.2±1.1 different DMTs during the 10-year study period. When considering the entire observation time, subcutaneous Interferon beta-1-a was the most frequently prescribed DMT (32.1%), followed by intramuscular Interferon beta-1-a (19.7%), Interferon beta-1-b (16.6%), Natalizumab (6.4%), Fingolimod (5.4%), Glatiramer acetate (5.4%), Teriflunomide (2.3%), Dimethyl Fumarate (2,1%), and no DMT (10.0%). Specifically, a limited number of patients, after starting the first prescribed DMT, spent at least part of the study without any DMT.

Overall annual healthcare costs were associated with age (Coef = -9.561; 95%CI = -17.072–2.051), and with baseline EDSS (Coef = 3.252; 95%CI = 1.706–4.797), but not with gender (p = 0.595), and with disease duration (Coef = -0.483; 95%CI = -3.350–2.384).

After adjusting for different covariates, 10% increase in the annual healthcare costs was associated with 1.1% reduction in the rate of 1-point EDSS progression (HR = 0.897; 95%CI = 0.820–0.981; Fig 2B), with 0.7% reduction in the reaching of EDSS 6.0 (HR = 0.925; 95%CI = 0.862–0.992; Fig 2D), and with 1.0% reduction in the conversion to SP (HR = 0.902; 95%CI = 0.838–0.971; Fig 2E), but not with the occurrence of the first relapse (HR = 0.993; 95%CI = 0.890–1.109; Fig 2A), and with the reaching of EDSS 4.0 (HR = 0.929; 95%CI = 0.859–1.005; Fig 2D) (Table 2).

Overall annual healthcare costs were positively associated with the ARR (Coef = 2.770; 95%CI = 1.056–4.483; and Coef = 2.468; 95%CI = 0.629–4.307 at the adjusted model) (Fig 3).

The present study found that higher healthcare costs for the use of more expensive DMTs from the early phases of MS, were associated with a lower risk of reaching milestones of short- (e.g. 1-point EDSS progression which occurred after 4–5 years from diagnosis, on average), and long-term disease evolution (e.g. reaching of EDSS 6.0 and conversion to SP, which took place after 8–10 years, on average). For instance, the use of treatments for highly active RRMS such as Fingolimod and Natalizumab, which are twice more expensive compared with Interferons [18], might have ideally halved the number of patients reaching EDSS 6.0 during 10 years in the present population (with a 5% overall reduction). Considering that previous studies have invariably found higher non-medical direct and indirect costs as disability progresses (i.e. rehabilitation, productivity loss) [8,31–34], it can be hypothesized that the delay or, ideally, the prevention of MS progression, also by using more expensive drugs, might determine long-term substantial rewards [2,33]. Noteworthy, in the present study, healthcare costs consisted of both the drug and the management procedures, in order to obtain a full evaluation of direct costs for treating MS, and were inflated to the most recent value to make cross-year comparisons possible.

There are different possible reasons underlying the association between healthcare costs for DMT administration and management, and MS progression. Healthcare costs in MS are driven by the use of DMTs, which are prescribed depending on the initial severity of MS and on its subsequent evolution. Indeed, the costs for DMTs are usually determined on the basis of their clinical effects as evaluated in clinical trials, and, so, they represent a surrogate marker of DMT efficacy. However, healthcare costs for DMT administration and management have never been tested in long-term real-life scenarios in relation to MS clinical outcomes. In previous observational studies trying to investigate the long-term efficacy of DMTs, Natalizumab determined more positive disease outcomes, compared with Interferon-beta or Glatiramer acetate, in a large propensity-matched cohort [28]. Similarly, the highest cumulative exposure (a parameter including both dosage and time) to Interferon beta-1a treatment had more beneficial effects on long-term outcomes, compared to the lowest, in patients with similar baseline clinical features [6]. Accordingly, the discontinuation of DMTs in patients with a more benign evolution determined a progressive disability accrual after 5 years, compared with subjects staying on continuous treatment with a more aggressive treatment approach [29]. Overall, clinical evidence support the effectiveness of highly-active, continuous and, subsequently, costly treatments seem in preventing or, at least, delaying MS progression. Accordingly, in the present study, patients who received more expensive DMTs, specifically indicated for a more aggressive disease evolution [16,17], presented better long-term outcomes, compared with subjects with relatively milder symptoms who received “low-cost” DMTs. This issue should be considered not only by physicians when profiling MS patients to prescribe the most suitable treatment [30], but also by policy makers when defining eligibility criteria for DMTs.

Among additional results, in the present cohort a higher relapse rate was associated with more expensive healthcare choices. Accordingly, previous studies showed that financial consequences of relapses depended not only on the temporary loss of productivity, but also on the subsequent change of treatment [35–37]. Indeed, overall annual healthcare costs, although not being associated with the risk of the first relapse, were positively correlated to the ARR. Conceivably, the first prescribed DMT was selected in consideration of baseline clinical features, and then changed in consideration of the subsequent disease evolution by escalating towards more expensive (and effective) treatments [18].

Limitations of the present study include the retrospective and non-randomized design, which, although generally considered acceptable for long-term studies in MS [23], might be responsible for a bias in the allocation to a first-line, second-line or no treatment. In addition, considering that in the present population the number of patients treated with highly-active treatments (e.g. Fingolimod and Natalizumab) was relatively low and DMTs more recently approved (e.g. Teriflunomide and Dimethyl Fumarate) were only marginally represented, a formal comparison among different DMTs was not performed. Moreover, the analysis did not include non-medical direct and indirect costs, and concomitant factors also affecting healthcare costs, such as comorbidities or presence of neutralising antibodies [37,38]. Finally, this is a single Centre experience and, so, different settings of care might have specific healthcare provision patterns [21].

In conclusion, the current landscape with newer and more expensive DMTs is a noticeable challenge for the economic performance of the healthcare systems. Hence, even if requiring short-term resource investments, a higher expenditure for treating MS has been associated with better disease evolution after 10 years. Current findings highlight the need for highly active and, possibly, expensive DMTs, in order to delay disease progression and its long-term burdensome economic consequences.