Background

The rather specific dopaminergic deficit in Parkinson’s disease (PD) has meant that dopaminergic replacement medications have proven to be an effective mainstay of treatment of the condition. However, later on in the course of the disease, such treatment may have increasing limitations resulting from decreasing efficacy and increasing complications such as dyskinesia, postural hypotension and hallucinations or other psychological manifestations.

Most recent developments in pharmacotherapy have therefore consisted of different formulations of or delivery systems for dopamine agonists or levodopa, as well as agents that promote the effects of dopamine.  It is rare that a new class of agent arrives on the scene for treatment of Parkinson’s disease and such an agent is therefore worthy of close attention.

Safinamide is one such agent, an alpha- aminoamide that, as well as having monoamine oxidase B inhibitory action also has a non-dopaminergic action in the form of glutamate modulation. This modulation is probably achieved by blocking N type Ca channel mobilisation and therefore reducing presynaptic glutamate vesicle release.  An action to stabilise Na channels by promoting the inactive state may also be relevant.

The MAO-B action is therefore akin to that of selegeline and rasagiline, though safinamide is reversible and more selective for MAO-B, perhaps reducing the tendency to side effects such as tyramine reactions and obviating the need for dietary restriction of cheese, etc. The action on glutamate is more akin to that of amantadine, a drug that has useful antidyskinetic properties in PD.

A number of studies on safinamide were conducted prior to its recent licence acquisition. First, as drug companies tend to do, the focus was on initiation therapy in early disease. Presumably a greater market share would be gained, and many patients would start on the drug early and remain on it longer during the long course of their disease.

There does not appear to be a major effect of safinamide when used de novo in early disease. When used in early disease as an adjunct to dopamine agonists, one trial (Stocci et al, 2012) in 270 patients found over 6 months that 100 mg had a significant UPDRS benefit versus placebo (-6 vs. -3.6). The dose of agonist was to remain stable, yet an increase was allowed if worsening symptoms! On blood analysis, drug was found in the placebo group in 26% of patients!! Someone had mixed up the bottles… Despite this, the study was published, presumably because these flaws might have negated rather than enhanced perceived benefit. An extension study in some patients failed to reach the primary end point of delay in requiring additional treatment. Another trial (Barone et al., 2013) in 679 patients failed its primary end point of change in UPDRS, but the 100mg (rather than 50mg) dose subgroup may have improved.

In more advanced disease, a study (Schapira et al., 2013) on 549 patients on any medications except MAO B inhibitors and with at least 1 ½ hours “off” time in a day showed improved “on” time without dyskinesia when safinamide was added to their regime in comparison with the addition of placebo.

The study discussed in this journal club, “Randomized trial of safinamide add-on to levodopa in Parkinson’s disease with motor fluctuations” by Borgohain et al., (2014) similarly looks at 699 patients with more advanced disease. Please refer to the Parkinson’s Disease primer for more general background information.

Study Design

This multicentre study first stabilised patients on their levodopa and then continued for 6 months, with a 18 month placebo-controlled extension study in those who had not experienced side effects or whose disease had worsened over the initial 6 months.

Enrolled patients had to have had PD lat least 3 years, be on levodopa with or without other therapies and they had to have at least 1 ½ hours of “off” time a day. Patients with severe dyskinesia or severe dose fluctuations were excluded!

Two doses of safinamide were chosen because of previous evidence that 50 mg may be sufficient for MAO–B action but 100 mg is necessary for gluatamate inhibition.

Assessments included 30 minute interval diary scores of “on” versus “off” and dyskinesia, UPDRS, clinical global impression of change, dyskinesia rating scale when “on”, % change in levodopa (the intention was to keep levodopa unchanged but it could be increased if patients deteriorated) and the PDQ-39 questionnaire. If PD therapy had to increase by 20%, their evaluation was done at this point rather than at 6 months.

A mixed model co-variate statistical analysis was used, comparing versus baseline. The 100 mg dose was analysed first and only if significant was the 50 mg dose versus placebo analysed.

The primary end point, total “on” time without troublesome dyskinesia, was a 1.36 hour improvement after 6 months on 100 mg safinamide, 1.37 hours on 50 mg safinamide and 0.97 hours for placebo. These were both significant versus placebo. There was likewise an improvement in “off” time. The disability measures, PDQ-39 and UPDRS II showed significant improvement only for 100 mg doses. In the extension study there was overall maintenance of benefits and a non significant reduction in dyskinesia. There was no significant increase in side effects.

Authors’ Conclusions

The authors concluded that the drug was successful when used as add-on therapy in improving “on” and “off” time without increasing troublesome dyskinesia, which would be a risk if increasing other types of anti PD medications. This correlates with MPTP treated monkey studies, which showed an improvement in dyskinesia as well as “off” symptoms. However, dyskinesia when reported as a side effect by patients was more common that with placebo in this study, but not more likely with 100 mg doses than with 50 mg doses.

Journal Club Comments

The study numbers were sufficiently powered to produce a meaningful result and the statistical analysis was good, making the main conclusion convincing. The issues of being allowed to change levodopa dose during the study, and then escape the study but still record the outcome if the change was 20%, were discussed. While one suspects that levodopa has stronger anti-PD action and may mask the effect of safinamide, in effect “rescuing” both placebo and safinamide groups, it would tend to decrease the observed benefit of  safinamide versus placebo. One can understand inclusion of this study design element on the grounds of ethics, and also it provides a more real-world setting.

What was strange was the exclusion of more severely dyskinetic patients from the study, given that the main novel pharmacological benefit may be in helping dyskinesia and the paper’s emphasis on measuring dyskinesia, though it was not the primary end point. It is not as if there have not been any other studies already conducted on the drug’s basic action. Perhaps it was felt another study could be got out of addressing this variable, but from the point of view of clinician prescribing, current evidence would not support its use as an antidyskinetic drug (since it was actually a side effect of the drug as reported by patients directly rather than recorded on the diary). Instead it may be a modestly beneficial drug for PD with relatively little action in provoking dyskinesia.

As is always the case, we are hampered by lack of direct comparison with a real life alternative. We would never consider offering a placebo to patients in real life. What we would like to know is does the drug work better in direct comparison with addition of rasagiline or of a dopamine agonist or entacapone. We have a clue that it may be better then simply increasing levodopa dose, but this was not really the primary comparison in the study, merely a possibility allowed by the study design.

The Queens Hospital Journal Club meeting upon which this article is based was perpared and presented  by Dr Stevan Wing, SpR in Neurology.