A Meta-Analysis of GBA-Related Clinical Symptoms in Parkinson's Disease

Background GBA gene had been proved to be a crucial gene to the risk of PD. Numerous studies had discussed about the unique clinical characteristics of PD patients with GBA carriers (GBA + PD). However, there was lack of updated comprehensive analysis on the topic. In order to clarify the association between GBA variants and the clinical phenotypes of PD, we conducted this comprehensive meta-analysis. Method Medline, Embase, and Cochrane were used to perform the searching. Strict selection criteria were followed in screening for new published articles or data. Revman 5.3 software was applied to perform the total statistical analysis, and funnel plots in the software were used to assess the publication biases. Results A total of 26 articles including 931 GBA + PD and 14861 GBA noncarriers of PD (GBA − PD) were involved in the final meta-analysis, and 14 of them were either newly added publications or related data newly analyzed compared with the version published in 2015. Then, a series of symptoms containing depression, orthostatic hypotension, motor fluctuation, wearing-off, and freezing were newly analyzed due to more articles eligible. Besides, clinical features like family history, AAO, UPDRS-III, H-Y, and dementia previously analyzed were updated with new data added. Significant statistical differences were found in wearing-off, family history, AAO, UPDRS-III, and dementia (OR: 1.14, 1.65; MD: −3.61, 2.17; OR: 2.44; p: 0.03, <0.00001, <0.00001, 0.003, and <0.00001). Depression was slightly associated with GBA + PD (OR: 1.47; p: 0.04). Clinical symptoms such as H-Y, orthostatic hypotension, motor fluctuation, and freezing did not feature GBA + PD. Conclusion Our results demonstrated that there were unique clinical features in GBA + PD which can help the management of the whole duration of PD patients.


Introduction
Parkinson's disease (PD), a common neurodegenerative disease, was featured by motor symptoms containing bradykinesia, resting tremor, rigidity, and postural instability. Nonmotor symptoms (NMS) such as cognitive impairment, olfactory dysfunction, and depression were also common in PD patients. Nowadays, the pathogenesis of PD remains elusive. Genetic factors have been demonstrated to cause PD and, to some extent, participate in modifying the phenotypes of PD [1][2][3].
In addition to the contribution of GBA to the development of PD, studies have reported PD patients with GBA carriers (GBA + PD) manifested special clinical features compared to idiopathic PD. In the year of 2015, our group conducted a study combining the results of our new original research and meta-analysis on the association between GBA variants and the clinical features of PD [11]. e data indicated that GBA + PD are more inclined to onset at early age, initially with bradykinesia, have family history and develop to dementia when compared with GBA noncarriers of PD patients (GBA − PD). However, with more published articles, there were new clinical features such as depression, motor complications, and freezing gaits focused by researchers which will help draw a full picture of clinical features of GBA + PD or GBA − PD with complete motor symptoms (MSs) and NMSs [12][13][14]. Combined with newly published articles and newly involved data of previous articles, we performed a comprehensive analysis on clinical features of GBA + PD.

Selection Criteria.
We conducted this meta-analysis based on PICOS (participants, interventions, controls, outcomes, and studies) rules.

Literature Search.
We searched articles in English using Medline database in Pubmed, Embase database in Ovid, and the Cochrane database. Key words were "GBA," "glucocerebrosidase," "Parkinso * " and "PD." e latest search was done on March 1, 2018. Overlapping articles from different databases were excluded. Two researchers (Yuan Zhang and Li Shu) performed the search independently. In case of disagreements, a third researcher (Qiying Sun) was consulted to arrive at a consensus.

Data Extraction.
Comprehensive data were retrieved including the following items: publication year, first author, ethnicity, country, number of GBA + PD and GBA − PD, and corresponding clinical information. Two researchers did the extraction independently. Another author was asked to participate in the process when confronted with problems.
e quality of all case-control studies were assessed according to the Newcastle-Ottawa Scale (NOS) [16].

Statistical Analysis.
e total statistical analysis was performed in Revman 5.3 software. e final results were demonstrated by pooled odds ratio (OR) or mean difference (MD) and 95% CI (confidence interval). When the data were dichotomous variables, pooled odds ratio (OR) was calculated, otherwise when the data were continuous outcomes, pooled mean difference (MD) was expressed. Heterogeneity was reflected by Q statistic (p value) and I 2 statistic. p > 0.1, I 2 ≤ 50% indicated that the heterogeneity was not significant and suggested a fixed model (FM) be applied. Otherwise, a random model (RM) was used. e shape of funnel plot was used to reflect publication biases. Sensitivity analysis was performed by removing each original study sequentially to test the stability of the results.

Results
e complete information of searching process is shown in the flowchart (Figure 1), and the information of all included studies is demonstrated in Table 1. A total of 26 articles including 931 GBA + PD and 14861 GBA − PD were involved in the final meta-analysis, and 14 of them, which contained 582 GBA + PD and 8217 GBA − PD, were either newly added publications or related data newly analyzed compared with the version published in 2015 [11].
en, a series of symptoms containing depression, orthostatic hypotension, motor fluctuation, wearing-off, and freezing were newly analyzed due to more articles eligible. Besides, clinical features like family history, age at onset (AAO), Unified Parkinson's Disease Rating Scale Part III (UPDRS-III), Hoehn-Yahr (H-Y), and dementia previously analyzed were updated with new data added. Due to the importance of disease duration in clinical characteristics of disease, we first conducted comparison of disease duration between GBA + PD and GBA − PD. We found that there was no statistical difference between the two groups (MD: 0.17, p: 0.47) (Supplementary Table 1; Supplementary Figure 1).
For the five (depression, orthostatic hypotension, motor fluctuation, wearing-off, and freezing) newly involved clinical characteristics in this meta-analysis ( Table 2, Supplementary Figure 1), they belonged to NMS and motor complications. As can be seen from Table 2, significant statistical difference was found in wearing-off (OR: 1.14; p: 0.03). Slightly statistical significance was found in depression of GBA + PD (OR: 1.47; p: 0.04). Clinical symptoms such as orthostatic hypotension, motor fluctuation, and freezing did not feature GBA + PD in this meta-analysis.
As to the five updated clinical features of GBA + PD with newly involved data ( Table 2 We found a change in statistical differences in UPDRS-III scores from previous negative results, while we almost reached the same conclusion in analyzing family history, AAO, H-Y, and dementia. All publications included were of high quality with the NOS scores above 7. According to the funnel plots (Supplementary Figure 2), the biases were rare. By removing articles one after another, the results of the remainder did not change significantly indicating that the results of our meta-analysis were stable.

Discussion
In our meta-analysis, we analyzed five new clinical features (depression, orthostatic hypotension, motor fluctuation, wearing-off, and freezing) and updated data of five previous analyzed clinical features (family history, AAO, UPDRS-III, H-Y, and dementia). We made the conclusion that GBA + PD patients had unique clinical features such as were more likely to have family history, earlier onset age, higher UPDRS-III scores, and develop dementia, depression, and wearing-off phenomena after adjusting disease duration.
Our meta-analysis about the demographic information of GBA + PD suggested that the carriers were more likely to have earlier age at onset with a mean of 3.6 years. Previous studies have shown that GBA + PD developed PD 1.7-6.0 years earlier than GBA − PD which were similar to our analysis [17]. Additionally, GBA + PD were more likely to have family history. ese basic features of GBA carriers will contribute to the targeted screening of the gene in researches.
As to other clinical features such as MSs and NMSs, our analysis demonstrated severe MSs reflected by higher UPDRS-III scores accompanied by motor complications like wearing-off phenomena and high possibilities to develop NMSs such as dementia and depression. Previous studies [18][19][20] have suggested deteriorative manifestations of GBA + PD such as higher UPDRS-III scores, easily presenting dementia, and motor complications. Since separating different subcategories of PD is crucial to better understand disease mechanisms, predict disease progression, or design clinical researches, recently, Fereshtehnejad et al. [21] reported that important clinical features and scales such as UPDRS I-III scores, NMSs-related scales such as Montreal Cognitive Assessment (MoCA) evaluating cognitive functions, and Epworth Sleepiness Scale (ESS) evaluation sleep disturbances were key factors defining clinical subtypes of PD. Our analysis found unique clinical features of GBA + PD which almost matched a diffuse malignant subtype in PD in the previous classifications which needed a more active treatment strategy for the deleterious prognosis. e mechanism underlying GBA + PD prominent clinical features remains elusive. Some studies suggested that GBA mutations can cause dysfunctional GCase which finally led to α-synuclein aggregations in PD brains and in dopaminergic neurons [22]. As α-synuclein was vital pathological feature in PD brains, the promotion of GBA mutations to α-synuclein aggregations may explain the deleterious clinical features of GBA + PD. e dysfunctions in pathways outside classic basal ganglia may explain the NMS features of GBA + PD. e cortex dysfunction caused   Scale; H-Y, Hoehn-Yahr Rating Scale. AAO, UPDRS-III, and H-Y were presented as mean and standard deviation; others were shown as count data. n, total number of patients whose clinical information was available in each group.
by global brain degeneration can damage functions of specific areas of brains and cause dementia or depression [23]. However, the number of researches was limited, and more mechanism studies were needed in the future. Previously, our comprehensive meta-analysis in GBA variants had proved the importance of GBA mutations to PD risks [10]. To clarify the role of GBA in PD clinical features more clearly, we did this meta-analysis. Our results demonstrated a clear phenotype-genotype correlation in GBA + PD. Knowing the unique features of GBA carriers will contribute to predicting the clinical course of GBA + PD and be benefit for the symptomatic treatments. e results of this meta-analysis can do a contribution to the precise treatments based on genetic screening and help delay the progression of the disease with more active and effective therapeutic strategies.
To evaluate the meta-analysis more objectively, there were some limitations which cannot be ignored. First, possible biases were inevitable because the included original studies were cross-sectional and possible biases existed in pooled analyses of these studies such as age, gender, or other correlated clinical phenotypes. Further longitudinal designed studies will be needed to confirm these results. Second, because most of these researches included mixed different specific variants together as GBA + PD or GBA − PD (Supplementary Table 1), we could not separate each variant with corresponding phenotype data and were not able to conduct pooled analysis based on specific variants of GBA. With more original articles conducted based on specific variants of GBA and phenotype, we may be able to do a more accurate analysis to help understand the relationship between the genotype and phenotype better. ird, although our updated meta-analysis was a comprehensive pooled analysis of GBA-associated clinical presentations, for the limited articles, we failed to prove the relationship of other clinical features such as rapid eye movement sleep behavior disorder (RBD) or freezing which were demonstrated to be associated with GBA previously [24].

Conclusion
Our meta-analysis suggested an increased risk of having family history, dementia, depression, wearing-off, earlier onset age, and higher UPDRS-III scores with GBA + PD. However, variants in GBA had no relationship with H-Y, orthostatic hypotension, motor fluctuations, and freezing in PD. More data were needed to do complete analysis on different variants and different ethnics of GBA, and the corresponding clinical manifestations which can help the management of the whole duration of PD patients.

Conflicts of Interest
e authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as potential conflicts of interest.

Supplementary Materials
Supplementary Figure 1: forest plots of the association between phenotypes and PD risks in total. (A)-(J) respond to the phenotypes of family history, age at onset, UPDRS-III, H-Y, dementia, depression, orthostatic hypotension, motor fluctuation, wearing-off, and freezing individually. Supplementary Figure 2: funnel plots of the association between phenotypes and PD risks in total. (A)-(J) respond to the phenotypes of family history, age at onset, UPDRS-III, H-Y, dementia, depression, orthostatic hypotension, motor fluctuation, wearing-off, and freezing individually. Supplementary