Introduction

Cognitive impairment is known to be a common and clinically relevant consequence of multiple sclerosis (MS), estimated to affect 40–65 % of MS patients (Rao 1995). The most frequent presentation of CI in MS is impaired information processing speed (PS); PS refers to the speed with which a person can interpret and process incoming information, to make an appropriate response (Rao et al. 1991; Benedict et al. 2006; Chiaravalloti and DeLuca 2008). Longitudinal studies demonstrate a slowly progressive, insidious course; thus, these deficits are unlikely to remit once present (Kujala et al. 1997; Amato et al. 2001; Schwid et al. 2007). CI in MS, specifically impaired PS, has a negative effect on personal relationships and activities of daily living (Benedict et al. 2001; Carone et al. 2005), and MS patients with CI are less likely to be fully employed (Beatty et al. 1995; Benedict et al. 2005, 2006; Parmenter et al. 2007b). Studies show that a decrease in PS over time predicts a decrease in employment status in MS patients (Morrow et al. 2010b). Yet, there are currently no approved therapies to treat this common symptom in MS patients.

Amphetamines, used to treat attention deficit hyperactivity disorder (ADHD), represent a candidate class of drugs to treat PS in patients with MS as they are used in other disorders for this indication. In addition to improving symptoms in adults and children with ADHD (Rhodes et al. 2004), studies have demonstrated that methylphenidate improves working memory and attention in schizophrenic patients (Barch and Carter 2005). Traumatic brain injury patients show improved accuracy on working memory and visual/spatial attention tasks as well as decreased reaction time (Kim et al. 2006). There have been a few studies in MS populations demonstrating promising preliminary results (Harel et al. 2008; Morrow et al. 2009, 2013). However, there is concern regarding psychiatric adverse events and abuse potential noted with amphetamines, especially fast-acting or immediate-release formulations (Berman et al. 2009).

A mixed amphetamine salt formulation with an extended release delivery system, under the trade name Adderall XR, is currently approved for the treatment of ADHD in children, adolescents, and adults (Pharmaceuticals 2011). Extended release (XR) formulations are thought to have less abuse-related risk and a lower potential to cause psychiatric side effects than immediate-release amphetamines, due to the lack of rapid increase in dopamine levels in the brain, as the rate of release is controlled with the XR delivery system (Heal et al. 2009). The aim of this study was to determine if this mixed amphetamine salts, extended release medication (MAS-XR) may improve PS deficits in MS patients with demonstrated impairment in this cognitive domain using a double-blind, placebo-controlled, proof of concept randomized trial comparing changes in performance following MAS-XR 5 mg, 10 mg, or placebo.

Methods

Outcomes

Our primary outcomes were the Symbol Digit Modalities Test (SDMT) (Smith 1982) and the Paced Auditory Serial Addition Test, 3 s version (PASAT) (Gronwall 1977). These two tests are the most commonly used tests to assess PS in the MS population. Both are included in the two validated neuropsychological test batteries routinely used in MS studies, the Minimal Assessment of Cognitive Function in MS (MACFIMS) (Benedict et al. 2002) and the Brief Repeatable Neuropsychological Battery (BRNB) (Rao 1991). Further, the SDMT is included in the recently recommended screening battery, Brief International Cognitive Assessment for MS (BiCAMS) (Langdon et al. 2012). The PASAT is incorporated into the Multiple Sclerosis Functional Composite (MSFC) (Fischer et al. 1999), used in MS clinical trials. Additionally, repeat assessments with the SDMT demonstrate little practice effect (Morrow et al. 2010a; Benedict et al. 2008). The SDMT is a visual test of processing speed and consists of the presentation of a standard letter-sized paper (8.5 × 11 in.) that contains the numbers and symbols to be processed. At the top of the page is a key where nine symbols are each paired with a single digit. The remainder of the page has a pseudo-randomized sequence of these symbols. The SDMT was adapted for the MS population by Rao (Rao 1991); the subject responds orally with the digit associated with each of the symbols as quickly as possible and is scored as the total number of correct responses in 90 s.

The PASAT is an auditory task of both processing speed and working memory. Subjects listen to a recording in which 61 single-digit numbers are voiced every 3 s. After each number is presented, the subject’s task is to respond with the sum of the last two digits presented. As each new number is presented, the subject must disregard previous numbers heard and his/her response, recall the last two digits, and add them together, voicing the total prior to the next number being announced on the recording [3, 5 respond 8, 7 respond 12, 1 respond 8, etc.]. There are 60 responses and it is scored as the total number of correct responses during the test.

Participants

Participants were recruited from October 2012 to November 2014 from the London MS clinic in London (ON) Canada. To be included in the study, participants had to demonstrate impaired PS on the SDMT or PASAT, defined by a z score of −1.5 or worse based on normative data for the MS population (Benedict et al. 2006). Participants with any type of MS (relapsing or progressive), aged 18–59, Expanded Disability Status Scale (EDSS) 6.5 or less, and, if female, not pregnant or breastfeeding, were eligible for inclusion in the study. Participants were excluded if there was a history of a relapse in the last 90 days; had untreated depression on history or a Beck Depression Inventory-Fast Screen (BDIFS) (Benedict et al. 2003) score ≥13; has heart disease; has bipolar disorder, schizophrenia or is taking an anti-psychotic medication; is currently using other stimulants or sympathomimetics, including modafinil. In addition to administering the SDMT, PASAT, and BDIFS, the Fatigue Severity Scale (FSS), a validated measure of generalized fatigue for the MS population, was also administered (Krupp et al. 1989). Demographic and MS disease characteristic data was also collected.

Intervention

MAS-XR is composed of equal amounts of the sulfate salts of dextroamphetamine, amphetamine, d-amphetamine, and d-,l-amphetamine aspartate monohydrate, resulting in a 3:1 ratio of D- to l-isomers of amphetamine (McGough et al. 2003). While the time to reach maximum plasma concentration (T max) for MAS immediate release (IR) is 3 h, it is 7 h for MAS-XR, producing a pharmacokinetic and pharmacodynamics effect similar to a twice daily dosing with the IR formulation (Biederman et al. 2002). In the ADHD population, the initial dose of MAS-XR is 5 or 10 mg daily, titrated to a maximum dosage of 30 mg/day in increments of 5 mg (McKeage and Scott 2003). Due to the extended release properties and the combined use of d- and l-amphetamine, MAS-XR is thought to have less abuse-related risk and a lower potential to cause psychiatric side effects than immediate-release amphetamines (Heal et al. 2009).

Participants were randomized to one of three treatment groups: placebo, 5 mg MAS-XR, or 10 mg MAS-XR. Randomization was done in a 1:1:1 pattern for the three treatment groups, generated by a computer program by the pharmacist; only the pharmacist was unblinded to the group assignment. The MAS-XR and placebo were encapsulated to conceal the assigned medication and the vials were labeled in an identical manner for all three subject groups. Participants were given a prescription vial with a single dose of medication and instructed to take it the morning of the second visit. This second visit was scheduled such that the subject ingested the medication 7 h (±30 min) prior to the scheduled visit time to ensure testing occurred in conjunction with the T max of MAS-XR (Biederman et al. 2002). The second visit was scheduled within 28 days of the first test administration. At the second visit, prior to administration of the SDMT, PASAT, BDIFS, and FSS, potential adverse events were reviewed.

Statistical analysis

Pearson’s correlation was used to determine any potential confounding relationship between SDMT or PASAT scores and continuous baseline characteristics while one-way ANOVA was used for categorical baseline characteristics. One-way ANOVA was used to compare the change on the SDMT and PASAT between visit 1 and visit 2 in each of the treatment groups compared to the placebo group. If significant relationships were found between baseline SDMT or PASAT scores and baseline characteristics, ANCOVA was performed with these covariates included in the model. Cohen’s d was used to calculate the effect size of the treatment if the results were found to be significant. Repeated measures ANOVA was used to examine change on FSS and BDIFS between groups between visit 1 and 2.

Results

Of the 95 participants screened for study eligibility, 62 subjects were eligible and enrolled in the study. Fifty-two of the enrolled participants completed the study; the most common reason for not completing the study was withdrawal of consent. Please see Fig. 1 for details regarding reasons for exclusion and the randomization process. Participant characteristics are presented in Table 1. The mean age was 45.9 years (standard deviation (SD) 9.3); 37 (71.2 %) were female and 50 (96.2 %) were Caucasian. The cohort achieved a mean number of years of education of 13.8 (SD 2.4), but only 18 (34.6 %) were employed, although 16/18 were working full time. The cohort was on average 10.9 years (SD 8.4) since diagnosis, the median EDSS was 3.5 (range 0–6.5), and 43 (82.7 %) had a relapsing remitting (RR) MS course. Approximately half (29, 55.8 %) were taking a disease-modifying therapy: interferon 12, glatiramer acetate 5, dimethyl fumarate 3, natalizumab 3, fingolimod 2, and teriflunomide 2, not significantly different between the three groups (treated with disease-modifying therapy vs. no disease-modifying therapy).

Fig. 1
figure 1

Randomization process. EDSS Expanded Disability Status Scale, FSS Fatigue Severity Scale, MAS-XR mixed amphetamine salts, extended release

Full size image
Table 1 Demographics and disease characteristics in the treatment and placebo groups
Full size table

At baseline, the mean SDMT score was 43.3 ± 7.2 and the mean PASAT was 34.8 ± 13.4, with 47 (90.4 %) and 25 (48.1 %) categorized as impaired on the SDMT and PASAT, respectively; 22 (42.3 %) were impaired on both the SDMT and PASAT. SDMT scores were significantly but weakly correlated with years of education (r = 0.359, p = 0.009) and EDSS (r = −0.310, p = 0.025), while PASAT scores were significantly but weakly correlated with EDSS only (r = −0.318, p = 0.022). The mean FSS score was 4.9 ± 1.6 and the mean BDIFS score was 3.6 ± 2.9. There was no significant relationship between the FSS with the SDMT (r = −0.240, p = 0.117) or PASAT (r = −0.036, p = 0.819), nor any significant relationship between the BDIFS and SDMT (r = −0.014, p = 0.928) or PASAT (r = 0.138, p = 0.378).

Participants completing the study were randomized as follows: 5 mg MAS-XR (n = 18), 10 mg MAS-XR (n = 20), and placebo (n = 14). Baseline scores of the three groups on the SDMT (5 mg—43.5 ± 6.1, 10 mg—43.4 ± 6.2, placebo—42.7 ± 6.2) were not significantly different from each other (p = 0.929). There were no significant differences on PASAT scores at baseline between the three groups (5 mg—38.2 ± 14.4, 10 mg—38.2 ± 13.5, placebo—34.0 ± 11.7; p = 0.407) and there were no significant difference between each other. Similarly, there were no significant differences between the three groups on any baseline characteristic.

Time between visit 1 and visit 2 was 10.6 ± 14.4 days. At the second assessment, the SDMT improved by 3.3 ± 5.8 points in the 5-mg group and by 5.2 ± 4.5 in the 10-mg group. In contrast, the placebo group increased by only 0.6 ± 4.4. The change in the treatment group was significant when comparing 10 mg to placebo (p = 0.043) but not when comparing 5 mg to placebo (p = 0.150). The effect size for 10 mg of MAS-XR on the SDMT compared to placebo was 0.47, a moderate effect size. For the PASAT, the placebo group increased by 5.9 ± 6.9 points, 5 mg by 5.2 ± 7.3, and 10 mg by 6.8 ± 7.9. The improvement in the treatment groups compared to placebo was not statistically significant (5 mg—p = 0.804, 10 mg—p = 0.735) (Fig. 2). The analysis for the PASAT was repeated, limiting to only those subjects who were impaired on this test at baseline; the findings continued to not be statistically significant.

Fig. 2
figure 2

a Raw score change between the first assessment (pre-dose) to the second assessment (post-dose) on the raw SDMT score on placebo, 5 mg MAS-XR, and 10 mg MAS-XR. SDMT Symbol Digit Modalities Test, MAS-XR mixed amphetamine salts, extended release. b Raw score change between the first assessment (pre-dose) to the second assessment (post-dose) on the raw PASAT score on placebo, 5 mg MAS-XR, and 10 mg MAS-XR. PASAT Paced Auditory Serial Addition Test, MAS-XR mixed amphetamine salts, extended release

Full size image

Adverse effects were reported in two subjects in the placebo group (arm tingling; headache), one subject in the 5-mg group (nausea) and three subjects in the 10-mg group (palpitations; tremor; nausea). At visit 2, the FSS score was 4.9 ± 1.5 demonstrating no significant change from visit 1 (p = 0.057) for the whole cohort or when comparing between groups (p = 0.247). The BDIFS for the whole cohort decreased to 2.7 ± 2.8, a significant decrease (p = 0.013), but there was no significant difference when comparing between the three treatment groups (p = 0.891).

The analysis was repeated controlling for years of education and EDSS for the SDMT and controlling for EDSS alone for the PASAT. The difference in change on the SDMT between the placebo group and MAS-XR 10-mg group remained significant (p = 0.047); all other analyses remained non-significant.

Discussion

This pilot, proof of concept study investigating the potential benefit of treatment with a slow-release amphetamine medication, MAS-XR, demonstrates objective improvement on the SDMT, a reliable and valid measure of PS in MS patients demonstrating impaired PS at baseline.

The effect of amphetamines on attention and PS in ADHD is well known, with multiple studies demonstrating an increase in speed and accuracy (Solanto 1998). There have been many studies demonstrating an improvement of attention and PS in other neurological conditions as well. In schizophrenic men, Pietrzak et al. (2010) found an improvement on PS in a single-dose crossover study with d-amphetamine. Similarly, Barch and Carter (2005) performed a single-dose crossover study in schizophrenic patients, but also compared the results to healthy controls (Barch and Carter 2005). They found d-amphetamine improved accuracy and selective attention in schizophrenic patients and healthy controls compared to placebo. Kim et al. (2012) found traumatic brain injury patients were more accurate and had a faster response time after one dose of methylphenidate compared to placebo (Kim et al. 2012). A longer term study also found benefit in patients with a primary brain tumor. Gehring et al. (2012) compared methylphenidate to modafinil in a 4-week randomized study (Gehring et al. 2012). Subjects in both groups demonstrated significant improvement on tests of PS and other tests requiring divided attention, with methylphenidate demonstrating a greater benefit than modafinil on measures of PS.

Studies in MS have produced mixed results, which may be due to different formulations of the stimulant drug or different study designs. Harel et al. (2008) completed a study very similar to this study with methylphenidate, a psychostimulant, testing MS patients with demonstrated PS impairment on the PASAT before and after a single dose of 10 mg methylphenidate or placebo. This study found improved scores on the PASAT by 22.8 % (Harel et al. 2008). In contrast, a multi-center placebo-controlled study in MS patients of l-amphetamine, the isomer of d-amphetamine, over a 4-week period found no difference on the SDMT or PASAT in MS patients (Morrow et al. 2009). Yet, a randomized placebo-controlled pilot study of lisdexamfetamine, a pro-drug that is metabolized to lysine and d-amphetamine, demonstrated an improvement on the SDMT of 4.6 points in the treatment group compared to 1.3 points in the placebo group after 8 weeks (Morrow et al. 2013).

In this current study, a single dose of MAS-XR 10 mg significantly improved performance on the SDMT but not the PASAT. It is not clear why there was a difference between these two tests. It may be due to the fact that the SDMT is visually mediated while the PASAT is mainly an auditory test. The PASAT also stresses working memory in addition to PS, yet previous studies have demonstrated improvement on measures of working memory, including on the PASAT, in the ADHD population (Schweitzer et al. 2004). Interestingly, the study by Morrow et al. (2013) with lisdexamfetamine also found similar results, such that there was an improvement on the SDMT but not the PASAT (Morrow et al. 2013). The difference between the two tests may also be due to the significant practice effect known to occur with the PASAT, while the SDMT has been shown to be reliable over repeated assessments (Parmenter et al. 2007a; Morrow et al. 2010a). Finally, recruitment for this study required that subjects be impaired on either the SDMT or the PASAT; while 47 subjects were impaired on the SDMT, only 25 were impaired on the PASAT at baseline, less than half of the subjects in the study. There may not have been enough power in this study to detect an improvement on the PASAT due to the small sample of those documented as impaired on the PASAT at baseline.

Finally, it is interesting to note that there was an improvement on the BDIFS overall, although this improvement was not significantly different between the three groups. It is not known, however, if this finding is clinically significant. Depression at baseline, either as reported by the subject or based on the BDIFS, was a reason for exclusion from the study. Further, the subjects’ average BDIFS score pre-treatment was 3.6 which decreased to 2.7, both within the normal or “no depression” range. It is possible this improvement may reflect overall improvement in outlook due to perceived benefit of the drug or from participation in a study or the feeling of engagement in the treatment of his/her CI.

There are several weaknesses to this study. To begin, it is a single-dose study, and thus, it is not known if the improvement noted here would be sustained over the long term. Secondly, it is not known if this improvement noted on an objective measure of processing speed translates into a clinically meaningful effect on MS patients’ day to day functioning. Thirdly, although few side effects were noted in this study, this may again be a function of only receiving one dose of the medication; this study cannot rule out the possibility of significant psychiatric side effects occurring in MS patients using MAS-XR. A longer term study including measures of quality of life and functional measures in addition to the objective measures of processing speed is warranted to better answer these questions.

In conclusion, this pilot, pre-post-dose, proof of concept study found MAS-XR 10 mg to significantly improved performance on the SDMT, a measure of PS, in MS patients impaired on PS at baseline. The results of this study demonstrate the need for further studies to confirm if the long-term use of MAS-XR conveys the same benefits.