A systematic review and meta-analysis of safety and efficacy of safinamide for motor fluctuations in patients with Parkinson's disease

Background: Safinamide, a recently developed drug with several mechanisms of action has been investigated as an add-on therapy for Parkinson's disease patients suffering from motor complications due to the usage of anti-Parkinson's medications such as levodopa and dopaminergic drugs. The aim of the study is to investigate the efficacy and safety of Safinamide as add-on therapy for Parkinson's disease patients. Methods: A computerized literature search was conducted of PubMed, EMBASE, ClinicalTrial.gov and Cochrane Library until August 2019. We selected relevant randomized controlled trials comparing safinamide groups to placebo groups. Relevant outcomes were pooled as mean difference (MD) and risk ratio (RR) using Review Manager 5.3. Results: We found that the overall MD of changes in “off-time” and “on time without troublesome dyskinesia” favored the safinamide group over the placebo group (MD -0.72 h, 95% CI -0.89 to -0.56 and MD 0.71 h, 95% CI 0.52 to 0.90, respectively). Additionally, the overall MD of change in Unified Parkinson's Disease Rating Scale part three (UPDRS III) favored the safinamide group (MD -1.83, 95% CI -2.43 to -1.23). In case of adverse events, the pooled meta-analysis did not favor the safinamide group over the placebo group. Conclusions: In this study, we provide class I evidence about the potential role of safinamide as an add-on therapy for Parkinson's disease patients suffering from motor fluctuations. However, a few included studies did not mention the data of important outcomes. Also, we report high risk of bias in individual studies. Future randomized controlled trials with different doses are recommended to provide more evidence for the efficacy and safety of safinamide as a treatment for motor complications of anti-Parkinson's medications.


Introduction
Parkinson's disease prevalence in the fourth decade of life is 41 people per 100,000 and increases to 1,900 people per 100,000 among those who are older than 80 1 . According to these statistics, Parkinson's disease is the second most common neurodegenerative disease after Alzheimer's disease.
The main pathology of Parkinson's disease is loss of dopaminergic innervation in the nigrostriatal pathway and spread to various regions in the brain. This loss leads to two types of symptoms; motor and non-motor. Motor symptoms include tremors, rigidity, and bradykinesia. Non-motor symptoms include depression, inability to sustain attention, and sometimes psychosis, especially hallucinations.
Despite the recent medications and updates in the field of pharmacology, there is no definitive treatment that can stop the progress of dopamine receptor loss in the nigrostriatal pathway. Therefore, we only use symptomatic medications for both the motor and non-motor symptoms.
The main symptomatic medications of Parkinson's disease are Levodopa (L-Dopa) 2,3 , dopamine agonists (DAs), and monoamine oxidase-B (MAO-B) inhibitors. Unfortunately, increasing the dose of these medications, especially L-Dopa, leads to motor side effects such as end-of-dose wearing off and dyskinesia, which can be irritating for patients 4-6 . Recently, a novel drug called safinamide was developed, which can reduce the side effects of these symptomatic medications, especially motor adverse events.
Safinamide has several dopaminergic and non-dopaminergic mechanisms of action such as sodium channel blockade, calcium channel modulation, and MAO-B inhibition 7 . The main goal of these mechanisms is inhibiting glutamate release and subsequently, improving motor symptoms 8 .
Recently published studies discussed the beneficial role of safinamide for treatment of motor complications of Parkinson's medications 9-14 . Some of them suggest that usage of safinamide improves quality of life and delays the motor deterioration of Parkinson's disease; thus, our study aims to evaluate the safety and efficacy of safinamide use for Parkinson's patients. According to our knowledge, this is the first meta-analysis that provides class I evidence for the useful usage of Safinamide for Parkinson's motor complications.

Methods
The Preferred Reporting Items of Systematic reviews and Meta-Analyses (PRISMA) guidelines were followed during the preparation of this manuscript 15 . We specified the inclusion criteria, methods of searching, and analysis in advance. The methods and analyses were conducted in strict accordance to the guidelines of the Cochrane Handbook for Systematic Reviews of Interventions and the Methods Guide for Comparative Effectiveness Reviews 16,17 .

Eligibility criteria
Studies that fit all of the following criteria were included in the meta-analysis: (1) Population: Studies whose population was patients with idiopathic Parkinson's disease (diagnosed using the UK Parkinson's Disease Society Brain Bank Criteria) 18

Study selection
After removal of duplicate articles, two reviewers (M.H and R.G) screened a spreadsheet of titles and abstracts independently using Microsoft Excel 2013 (windows version). Full text studies selected were examined independently by the same two reviewers, the third reviewer (A.N) solve any disagreement by discussion with the main author before the final selection. The independent reviewers are acknowledged for their generous help in searching, screening, and data extraction processes. We did not need to contact any study investigator for further clarification.
Data collection process and data items An online data extraction sheet was constructed. One independent reviewer (M.H) extracted the data from included studies and entries were checked by the main author. The data extraction form included the following domains: 1) study ID; 2) year of publication; 3) country; 4) study design (randomized controlled trials only); 5) follow-up duration; 6) safinamide dose; 7) population definition; 8) inclusion and exclusion criteria; 9) sample size; 10) baseline characteristics; 11) available data of outcome measures (pre, post, and change from baseline); and 12) quality assessment domains. A copy of data extraction form is available as extended data 19 .

Risk of bias in individual studies
Cochrane risk of bias assessment tool was used to assess the risk of bias in randomized controlled trials. We assessed the following risks: 1) Selection bias; 2) performance bias; 3) detection bias; 4) attrition bias; 5) reporting bias; and 6) any other source of bias that might have influenced the study data 20 . One reviewer (A.N) and the main author rated each domain separately as low, high or unclear risk of bias. We used Review Manager software (RevMan 5.3) to summarize the risk of bias of included randomized controlled trials.

Efficacy measures
The efficacy of drugs treating motor complications in Parkinson's disease was assessed for the following outcomes: (1) Unified Parkinson's Disease Rating Scale part three (motor part) (UPDRS III): The unified Parkinson's disease rating scale 21 is a reliable score of four parts to assess the severity of Parkinson's symptoms. Part three indicates the motor score, which is the main measure for motor function in Parkinson's disease patients.
(2) Patient-reported diaries: Patient diaries gave information about the duration of the following motor outcomes: "on time with non-troublesome dyskinesia", which means the duration of absence of dyskinesia associated with the long term usage of Parkinson's dopaminergic drugs such as levodopa and "off-time", which is the duration of returning motor and non-motor symptoms of Parkinson's disease, even with the use of levodopa and other antiparkinsonian drugs. (

Synthesis of results
Since all the data in the study are continuous data, each efficacy measure is reported as mean difference (MD) between the two groups from the baseline to endpoint, along with its standard error (SE). Both were pooled using the DerSimonian-Laird random effect model. In the case of studies reporting data at multiple time points, the last endpoint was considered.
The overall MD was interpreted with the consideration that efficacy measures are in different directions; an improvement in "on time without troublesome dyskinesia" would be indicated by an increased MD, while an improvement in UPDRS III, "off-time", UPDRS II, DRS, PDQ-39, MMSE, and HAM-D scores would be indicated by a decreased MD.
The proportion of risk ratio (RR) was used to pool the adverse events reported in the studies to the total number in each group between the two groups in the DerSimonian-Laird random effect model. To examine heterogeneity of studies, forest plot visual inspection was used and assessed using the Cochrane Q and I2 tests using RevMan version 5.3 for windows.

Calculation of missing data
A few studies, such as as Stoochi (2004)

Risk of bias across studies
Funnel plots was used to explore the publication bias across studies and to show the relationship between effect size and precision. The evidence of publication bias was assessed using the following: 1) Egger's regression test, and 2) the Begg and Mazumdar rank correlation test (Kendall's tau).

Software
We performed all the analysis and calculations in this meta-analysis using Review Manager software version 5.3 (RevMan 5.3).

Study selection
The literature search resulted in 160 studies. After the complete screening process of titles, abstracts, and full texts, 154 studies did not meet the eligibility criteria and six articles with six randomized controlled trials remained with a total of (2556) patients included in the meta-analysis 9-14 .
A description of the flow of study selection is shown in the PRISMA flow diagram in Figure 1.

Study characteristics
The follow up duration in the studies ranged from 12 Table 1.

Risk of bias within studies
The Cochrane risk of bias assessment tool was used to assess the quality of included studies. All included studies had a low risk of bias in terms of random sequence generation, allocation concealment, blinding of participants and personnel, blinding of outcome assessment, incomplete outcome data, and selective reporting except Schapira et al. (2012) 11 , which had a high risk of bias for random sequence generation as the randomization method was not reported and for the blinding of outcome assessment as the method of blinding was not reported, and Stoochi et al. (2004) 9 , which had a high risk for incomplete outcome data because there was no intention to treat analysis for missed or withdrawn patients mentioned in the study. The summary of risk of bias domains is shown in Figure 2.

Drug efficacy
Off-time. The overall MD between the two groups from baseline to endpoint in terms of change in "off-time" favored safinamide over placebo (MD -0.72 h, 95% CI [-0.89 to -0.56], Figure 3A). Pooled studies were homogenous (P=0.42).
On time without troublesome dyskinesia. The overall MD between the two groups from baseline to endpoint in terms of change in "on time without troublesome dyskinesia" favored safinamide over placebo (MD 0.71 h, 95% CI [0.52 to 0.90], Figure 3B). Pooled studies were homogenous (P=0.54).

UPDRS II.
The overall MD between the two groups from baseline to endpoint in terms of change in "UPDRS II" favored safinamide over placebo (MD -0.69, 95% CI [-1.03 to -0.36], Figure 3D). Pooled studies were homogenous (P=0.26).  DRS score. The overall MD between the two groups from baseline to endpoint in terms of change in "DRS score" did not favor either of the two groups (MD -0.14 h, 95% CI [-0.36 to 0.08], Figure 4E). Pooled studies were homogenous (P=0.15).

CGI severity.
The overall MD between the two groups from baseline to endpoint in terms of change in "CGI severity" favored safinamide over placebo (MD -0.18 h, 95% CI [-0.24 to -0.12], Figure 4F). Pooled studies were homogenous (P=0.70).

HAM-D.
The overall MD between the two groups from baseline to endpoint in terms of change in "HAM-D" favored safinamide over placebo (MD -0.35 h, 95% CI [-0.64 to -0.06], Figure 4H). Pooled studies were homogenous (P=0.88).
(G) Worsening Parkinson's disease Seven studies reported patients with worsening of Parkinson's disease during the study. The pooled meta-analysis did not favor either of the two groups (RR 0.82, 95% CI [0.65 to 1.03], Figure 6G). Pooled studies were homogenous (P=0.54).
(H) TEAEs leading to discontinuation Nine studies reported the number of patients with TEAEs leading to discontinuation of the study. The pooled meta-analysis did not favor either of the two groups (RR 1.05, 95% CI [0.75 to 1.46], Figure 6H). Pooled studies were homogenous (P=1.00).

(I) Serious drug-related TEAEs
Three studies reported the number of patients with serious drug-related adverse events. The pooled meta-analysis did not favor either of the two groups (RR 0.72, 95% CI [0.32 to 1.62], Figure 7I). Pooled studies were homogenous (P=0.32).

(L) Any drug-related TEAEs
Five studies reported the number of patients with any drugrelated TEAEs. The pooled meta-analysis showed an increase in the number of patients with any drug-related adverse events in the placebo group compared to the safinamide group (RR 1.19, 95% CI [1.03 to 1.36], Figure 7L). Pooled studies were homogenous (P=0.36).

(M) Any TEAEs
Seven studies reported the number of the patients with any TEAEs. The pooled meta-analysis did not favor either of the two groups (RR 0.98, 95% CI [0.93 to 1.02], Figure 7M). Pooled studies were homogenous (P=0.88).

(N) Serious TEAEs
Nine studies reported the number of the patients with serious TEAEs. The pooled meta-analysis did not favor either of the two groups (RR 1.02, 95% CI [0.81 to 1.28], Figure 7N). Pooled studies were homogenous (P=0.38).

Risk of bias across studies
As showed in Figure 8, funnel plots of UPDRS III, off-time, on-time without troublesome dyskinesia, and UPDRS II show no significant publication bias across studies.

Summary of evidence
The pooled meta-analysis of six studies provides a class I evidence that using safinamide as add-on therapy for Parkinson's disease is very effective and well tolerated. The meta-analysis shows that safinamide improves motor fluctuations, which is a main side effect of anti-Parkinson's medications, as reported by patient diaries and measured by "on time without troublesome dyskinesia", "off-time", and UPDRS III score. This novel drug is also improving the quality of life of Parkinson's disease patients, as measured by the UPDRS II scale, the PDQ-39 questionnaire, HAM-D, and MMSE.
Regarding tolerability, safinamide is a well-tolerated drug and despite increasing the risk of some adverse events such as dyskinesia, which was higher in the safinamide group than the placebo 10-14 , the pooled meta-analysis of RRs of adverse events did not show any statistical significance between the two groups of comparison.

Previous studies
The results obtained from the meta-analysis are consistent with the results of the previous randomized controlled trials in terms of outcomes measuring motor fluctuations and quality of life. "On time without troublesome dyskinesia" and "off-time" are the main outcomes to evaluate motor fluctuations and were mentioned in Schapira et al.  placebo group [12][13][14] . In addition, the mean difference of "off-time" in all studies favored the safinamide group and this result was consistent with the pooled meta-analysis of included studies [12][13][14] . The

Strengths of the study
The strengths of the meta-analysis are the following: 1) multiple search engines were searched and all the possible sources of studies to be included were covered; 2) clear eligibility criteria were provided; 3) multiple reviewers revised every step to ensure accuracy; 4) during the preparation of this manuscript, the PRISMA guidelines were followed; 5) the study was conducted according to the guidelines of the Cochrane Handbook for Systematic Reviews of Interventions in a strict way; 6) the randomized controlled trials included data of high validity and acceptable quality, as indicated by the risk of bias assessment.

Limitations of the study
The meta-analysis limitations are the following: A) some studies, such as as Stoochi et al. Implications for future research Based on the results of the study, future randomized controlled trials with different doses are recommended to investigate the efficacy of safinamide for Parkinson's disease patients with motor fluctuations as a side effect of anti-Parkinson's medications.

Conclusions
Despite the evidence provided by this meta-analysis, demonstrating the efficacy of safinamide as add-on therapy for treatment of motor complications of anti-Parkinson's disease medications, future studies are still needed to confirm the safety and efficacy of this novel drug.

Data availability
Underlying data All data underlying the results are available as part of the article and no additional source data are required. , which permits unrestricted use, distribution, and reproduction in any medium, provided the original Attribution License work is properly cited.

Ahmed Negida
School of Pharmacy and Biomedical Sciences, University of Portsmouth, Portsmouth, UK Zagazig University Hospitals, Zagazig University, Zagazig, Egypt The article describes a systematic review and meta-analysis of published studies on the Safenamide for the treatment of motor fluctuations in Parkinson's Disease.
First, the topic is interesting because currently, there is an unmet clinical need to find potential agents that decrease the motor fluctuations and the LED in PD midstage and late-stage PD.
Second, the methodology is clear and in accordance with the Cochrane handbook guidelines, while the manuscript is well-written in compliance with the standard reporting guidelines (PRISMA statement)endorsed by the ICMJEs.
The statistical analysis was properly done and the data analysis interpretation and discussion are sufficient.
I do not find any major concerns in the article; I believe it worth the publication. No further changes are required from my point of view.

Are the rationale for, and objectives of, the Systematic Review clearly stated? Yes
Are sufficient details of the methods and analysis provided to allow replication by others? Yes

Is the statistical analysis and its interpretation appropriate? Yes
Are the conclusions drawn adequately supported by the results presented in the review? Yes No competing interests were disclosed. Competing Interests: I confirm that I have read this submission and believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard, however I have significant reservations, as outlined above.
The benefits of publishing with F1000Research: Your article is published within days, with no editorial bias You can publish traditional articles, null/negative results, case reports, data notes and more The peer review process is transparent and collaborative Your article is indexed in PubMed after passing peer review Dedicated customer support at every stage For pre-submission enquiries, contact research@f1000.com