Deep Brain Stimulation in Patients With Mutations in Parkinson's Disease–Related Genes: A Systematic Review

ABSTRACT

Background

Deep brain stimulation (DBS) is an effective treatment for Parkinson's disease (PD), and careful selection of candidates is a key component of successful therapy. Although it is recognized that factors such as age, disease duration, and levodopa responsiveness can influence outcomes, it is unclear whether genetic background should also serve as a parameter.

Objectives

The aim of this systematic review is to explore studies that have evaluated DBS in patients with mutations in PD‐related genes.

Methods

We performed a selective literature search for articles regarding the effects of DBS in autosomal dominant or recessive forms of PD or in PD patients with genetic risk factors. Data regarding changes in motor and nonmotor scores and the presence of adverse events after the stimulation were collected.

Results

A total of 25 studies were included in the systematic review, comprising 135 patients. In the shorter term, most patients showed marked or satisfactory response to subthalamic DBS, although leucine rich repeat kinase 2 carriers of R114G mutations had higher rates of unsatisfactory outcome. Longer term follow‐up data were scarce but suggested that motor benefit is sustained. Patients with the glucosidase beta acid (GBA) mutation showed higher rates of cognitive decline after surgery. Motor outcome was scarce for pallidal DBS. Few adverse events were reported.

Conclusions

Subthalamic DBS results in positive outcomes in the short term in patients with Parkin, GBA, and leucine‐rich repeat kinase 2 (non‐R144G) mutations, although the small sample size limits the interpretation of our findings. Longer and larger cohorts of follow‐up, with broader nonmotor symptom evaluations will be necessary to better customize DBS therapy in this population.

Deep brain stimulation (DBS) is a well‐established therapy ultimately resulting in an improvement of the quality of life for patients with Parkinson's disease (PD).1, 2, 3 DBS response can differ depending on several clinical characteristics, including motor PD phenotype,4 age, disease stage, and associated symptoms,5 but the influence of genetics on clinical outcomes over time has not been carefully evaluated. As the current approach on therapeutic strategies for DBS in PD claims to an individualized and tailored assessment,6 the knowledge of whether genetic background should guide specific therapeutic plans is of the utmost importance.

Different PD‐related mutations have been associated with distinct motor and nonmotor symptoms that may be relevant in the context of DBS surgery. For instance, R1441C/G/H carriers of leucine‐rich repeat kinase 2 (LRRK2) mutations have more motor fluctuations, more tremor, and postural instability when compared with carriers of other mutations within the same gene7; the Parkin gene (PRKN) mutation has been associated with higher frequency and severity of specific impulse control behaviors8, 9; and glucosidase beta acid (GBA)–associated PD patients tend to have poorer cognitive performance and are more prone to be diagnosed with mild cognitive impairment or dementia.10

Case reports and small case‐control studies have reported the effects of DBS in different genetic forms of PD with heterogeneous measures. While the present study was underway, Rizzone and colleagues11 narratively described outcomes of DBS in PD‐related genes, Kuusimäki and colleagues12 systematically reviewed monogenic PD patients including subjective improvement as an outcome measure, and Artusi and colleagues13 quantitatively reported STN‐DBS outcomes in monogenic PD patients. None of these studies addressed motor outcome over time. Furthermore, the systematic retrieving of adverse events was also not performed. Thus, the question of whether genetic status should inform DBS prognosis, risks, and target selection remains unanswered.

The aim of this systematic review is to analyze motor outcome over time and summarize nonmotor outcome and safety from recent studies that have evaluated DBS in patients with PD‐related mutations during the past 15 years.

Methods

This review follows the Preferred Reporting Items for Systematic Reviews and Meta‐Analyses statement and was registered with International prospective register of systematic reviews (PROSPERO; CRD42017067198). A selective literature search for articles published from 2000 to May 2018 using the MEDLINE, EMBASE, and Cochrane Library databases was performed. The search strategy is described in the Supporting Information.

We included published studies that examined the clinical changes of patients with PD‐related mutations after DBS. Articles were selected according to the following criteria: (i) autosomal dominant or recessive forms of PD (monogenic PD); (ii) PD patients presenting with “risk genes,”14, 15 that is, mutations associated with increased susceptibility to develop PD; and (iii) response to treatment objectively described. Objective response to treatment was defined by (i) the presence of off medication/ON stimulation and off medication/OFF stimulation Unified Parkinson's Disease Rating Scale (UPDRS)/Movement Disorder Society‐Sponsored Revision of the Unified Parkinson's Disease Rating Scale (MDS‐UPDRS) motor score (or if the latter was not available, preoperative off medication motor UPDRS/MDS‐UPDRS) or (ii) percentage change in motor UPDRS/MDS‐UPDRS score after surgery. Exclusion criteria were the following: (i) other surgical therapeutic interventions, such as pallidotomy, or (ii) X‐linked dystonia parkinsonism or rapid onset dystonia‐parkinsonism. One author screened the articles and 2 authors independently reviewed the relevant studies. Levodopa equivalent daily dose (LEDD) was calculated as per previous recommendations.16 When available, motor complications were also recorded and measured by UPDRS IV. The following data were retrieved and organized using a structured form: study design, year of publication, age at symptoms onset, age at surgery, duration of disease, genetic analysis, DBS placement, motor outcome (motor UPDRS/MDS‐UPDRS), cognitive and psychiatric outcomes (including UPDRS I when available), measures of disability (UPDRS II, Schwab and England, Brown's Disability Scale) and quality of life (39‐item PD questionnaire; PDQ‐39), and side‐effects. Data for all of the studies included were described in the tables and summarized in the article text.

Mean percentage improvement after DBS was calculated using mean off medication/ON stimulation UPDRS III score when compared with mean off medication/OFF stimulation UPDRS III score. If the latter were not available, the mean preoperative off medication UPDRS III scores were used. Mean LEDD change was calculated using the mean LEDD (mg) before and after surgery.

To synthetize the data, the authors defined mean UPDRS III change of 50% or more as marked response, mean UPDRS III change of 30% to 50% as satisfactory response, and less than 30% change as unsatisfactory response. Because of the variable postoperative follow‐up intervals adopted by different studies, we arbitrarily defined shorter follow‐up as the mean follow‐up of less than 2 years, intermediate follow‐up as the mean follow‐up between 2 and 6 years, and longer term follow‐up as the mean follow‐up of more than 6 years. LEDD and UPDRS IV were described in details in the tables, and the range of the mean percentage change for these outcomes was described in the text.

To assess the main results in terms of statistical significance, an analysis was performed of a subset of patients with individual changes in motor UPDRS up to 2 years after DBS. In this sample, we calculated age at onset, age at surgery, and change in LEDD in each genetic group.

Nonmotor outcomes are summarized in the text and described in detail in the Supporting Information tables. The report of present or absent adverse events was retrieved, and adverse events are described in detail in the Supporting Information tables.

Results

The search of databases and reference lists provided 1754 studies (Fig. 1). Using the criteria noted previously, 25 articles were included in this review. A total of 12 were case reports/series, and 13 were retrospective observational studies.

Figure 1.

Flow diagram. DBS, deep brain stimulation; PD, Parkinson's disease.

The 25 articles included reported 151 patients with PD‐related mutations (2 patients are represented twice in Angeli and colleagues17); from which 135 patients were included in the outcome analysis. The selected articles included PRKN,17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27 LRRK2,17, 23, 24, 28, 29, 30, 31, 32, 33, 34 Phosphatase and tensin homolog (PTEN)‐induced putative kinase 1 (PINK1),20, 24, 25 GBA,17, 35 alpha‐synuclein (SNCA)36, 37 and retromer complex component (VPS35) mutations38, 39 in addition to chromosome 22q11.2 microdeletion. Age at onset ranged from 1027 to 57 years old,24 and the duration of disease at DBS implant ranged from 427 to 45 years.25 Baseline characteristics, mutations, target, and number of patients are described in Supporting Information Table 1. The motor outcome by genetic status and target (subthalamic nucleus [STN], internal globus pallidus [GPi], ventral intermediate nucleus [VIM]) is described in Supporting Information Table 2. The motor effect of STN‐DBS in each genetic group is illustrated in Figure 2.

Figure 2.

Motor effects of STN‐DBS in patients with PD‐related mutations. (I) Distribution of mean percentage improvement across patients, including cases with missing outcome. (II) Distribution of mean percentage improvement across patients, excluding cases with missing outcome. (A) Motor outcome of PRKN patients with a mean follow‐up of less than 2 years. (B) Motor outcome of PRKN patients with a mean follow‐up of 2 to 6 years. (C) Motor outcome of PRKN homozygous/compound heterozygous patients with a mean follow‐up of less than 2 years. (D) Motor outcome of PRKN homozygous/compound heterozygous patients with a mean follow‐up of 2 to 6 years. (E) Motor outcome of LRRK2 patients with a mean follow‐up of less than 2 years. (F) Motor outcome of LRRK2 patients with a mean follow‐up of 2 to 6 years. (G) Motor outcome of GBA patients with a mean follow‐up of less than 2 years. (H) Motor outcome of GBA patients with a mean follow‐up of 2 to 6 years. GBA, glucosidase beta acid; HZ/CH, homozygous/compound heterozygous; LRRK2, leucine‐rich repeat kinase 2; PRKN, parkin; STN‐DBS, subthalamic deep brain stimulation; UPDRS, Unified Parkinson's Disease Rating Scale; y, year.

The evaluation of mentation, behavior, mood (UPDRS I), and other measures of cognitive and psychiatric outcomes are described in Supporting Information Table 3. Measures of disability, other nonmotor symptoms, and adverse events are described in Supporting Information Table 4.

The main results are summarized next and include studies in which the distinct outcome was available for each genetic group (ie, excluded are studies that reported grouped outcomes of patients with mutations in different genes).

PRKN Mutation

Motor Outcome

STN. From 40 STN‐DBS PRKN patients (22 homozygotes/compound heterozygotes, 18 single heterozygotes), 34 (34/40, 85%) had shorter follow‐up data, and only 12/40 patients had intermediate outcome available.

At the shorter follow‐up, marked responses were seen in most patients (22/34, 64.7%), satisfactory responses in 7/34 (20.6%), and unsatisfactory responses in 5/34 (14.7%). At intermediate follow‐up, 4 patients each (4/12, 33%) had marked, satisfactory, and unsatisfactory responses.

Excluding single heterozygotes, 18/22 patients had shorter follow‐up data, and only 7 patients had intermediate follow‐up outcomes. At the shorter follow‐up, 12/18 (66.7%) patients had marked improvements, 3/18 (16.7%) had satisfactory responses, and the other 3/18 had unsatisfactory outcomes. At the intermediate follow‐up, 2/7 (28.6%) had marked responses, 4/7 (57.1%) had satisfactory responses, and unsatisfactory responses were seen in 1 patient.

The mean LEDD change after surgery ranged from 2.1%17 to 91.7%19 as illustrated in Figure 3, whereas improvement in UPDRS IV scores ranged from 20%17 to 100%.18

Figure 3.

Mean percentage of change in LEDD after STN‐DBS. The size of the circle represents the number of patients in each study, with each circle centering on mean percentage LEDD change. When studies provided individualized results, these were used. F/U, follow‐up; GBA, glucosidase beta acid; HZ/CH, homozygous/compound heterozygous; LEDD, levodopa equivalent daily dose; LRRK2, leucine‐rich repeat kinase 2; PRKN, parkin; SH, single heterozygous; SNCA, alpha‐synuclein; STN‐DBS, subthalamic deep brain stimulation; VPS35, retromer complex component; y, year.

GPi. GPi was targeted in 3 patients17—all homozygous or compound heterozygous—followed for 12 months, and resulted in unsatisfactory motor improvements (21% of improvement). Interestingly, these 3 GPi‐DBS patients were part of a study describing 2 PRKN patients in which STN‐DBS also produced somewhat limited motor improvements (31%). Complications of therapy (UPDRS IV) improved by 70%17 despite a mean LEDD increment after surgery of 24.7% (237 ± 315 mg).

Nonmotor Outcome

Nonmotor outcomes were only described for STN‐DBS PRKN patients.

Postoperative evaluations of mentation, behavior, and mood (UPDRS I) were described in 4 patients, and mean scores slightly worsened after surgery in 3 patients and improved by 67% in 1 patient (Supporting Information Table 3).

Cognitive outcome was reported after 6 months to 5 years of STN‐DBS.19, 21, 22, 24 Three studies described stable cognitive performance,19, 22, 24 and 1 study21 showed that the scores in the Mattis Dementia Rating Scale after surgery in PRKN patients were slightly but significantly worse than those of patients with no mutations.

With regard to psychiatric symptoms, 1 single heterozygous PRKN patient developed mania and hypersexuality following DBS,19 and 1 homozygous PRKN patient had improvement of depression.18

Postoperative measures of disability were available in 4 studies,18, 20, 25, 26 all revealing positive outcomes.

LRRK2 Mutation

Motor Outcome

From the 50 LRRK2 patients,17, 23, 24, 28, 29, 30, 31, 32, 33, 34 35 (70%) had shorter follow‐up data and 29 (58%) patients had intermediate follow‐up outcomes. The STN was the target in all patients.

STN

Marked or satisfactory responses were seen in most patients. At the shorter follow‐up, marked improvements were seen in 16/35 (45.7%) patients, satisfactory responses also in 16/35 (45.7%) patients, and unsatisfactory outcomes in 3/35 (8.6%) patients.

At intermediate follow‐up, marked improvements were present in 16/29 (55.2%) patients and satisfactory responses in 13/29 (44.8%) patients. Only 1 patient had longer term data, showing sustained marked improvement (55.8%) after 8 years.

Most LRRK2 patients presented with the G2019S mutation (n = 44; 1 of these also had GBA heterozygous E326K). Notably, from the 4 patients who presented with the R144G mutation, 3 had unsatisfactory outcomes (less than 30% improvement) and only 1 had 30% to 50% improvement. Other mutations (2 patients with the T2031S mutation, 1 patient with the Y1699 mutation, and 1 patient with the R793M mutation) were associated with either satisfactory or marked improvements.

In 4 studies,23, 28, 30, 32 the mean LEDD decreased from 17.5%30 to 75%28 as illustrated in Figure 3. Motor complications improved from 33.3%30 to 75%.23

Nonmotor Outcome

The STN was targeted in all LRRK2 patients. Postoperative UPDRS I was described in 11 patients, mostly without major changes.28, 31 One patient had initial improvement at 1 year, followed by a return close to baseline levels after 8 years.23

Cognitive performance was stable in 27 LRRK2 patients followed from 3 months to 5.1 ± 1.4 years.28, 30, 31, 33

With regard to psychiatric complications, 2 patients with the T2031S variant experienced behavioral disorders with visual hallucinations.28, 34 In 1 LRRK2 patient, mild hypersexual behavior present before surgery persisted after STN‐DBS.

Postoperative measures of disability were available in 7 studies, most describing improvement or minor changes.23, 28, 30, 31, 33 One study revealed that the 4 carriers of the R1441G mutation had significantly less improvements in UPDRS II (22%) when compared with the control LRRK2 negative group (45%; P = 0.001), whereas the 33% improvement in PDQ‐39 score was also lower but not significantly different than controls.29

GBA Mutation

Motor Outcome

A total of 36 patients with GBA mutations were described in 3 studies.17, 35, 40 Two patients also had a LRRK2 mutation,17, 40 and 1 patient had a heterozygous PRKN mutation.40 Motor outcome was described for 28 patients (16 STN‐DBS, 2 GPi‐DBS, and 1 Vim‐DBS), but in 9/28 patients target information was not clear.40 Follow‐up ranged from 1 year to 10 years.

STN. In the short term, mean motor improvements were marked in most patients (15/16, 93.8%) and satisfactory in 1 patient. At intermediate follow‐up, 1 patient each (1/3, 33.3%) had marked, satisfactory, and unsatisfactory responses. At long‐term follow‐up, only 2 patients had outcomes available and still showed satisfactory improvement (6–10 years). One study described 9 GBA patients (7 to 9 underwent STN‐DBS, mean follow‐up 7.5 years) in which mean UPDRS III presurgery was 52.4 off medication and 18.4 on medication, and postoperative mean MDS‐UPDRS III was 50 in the on medication/ON stimulation state.

LEDD decreased in the 3 studies (Supporting Information Table 2). Motor complications improved from 37%17 to 100%.35

Other Targets. GPi was targeted in 4 patients17, 40 and VIM in 1 patient,17 but only 1 study described distinct motor outcome according to target. In this study,17 2 GPi‐DBS patients had 24.8% improvement and 1 VIM‐DBS patient showed 42.9% improvement (both at 1‐year follow‐up). Surprisingly, in this study the LEDD reduction (mg) in the GBA group was greater in patients who underwent GPi stimulation (1005 ± 77 mg reduction) compared to STN (146 ± 510 mg reduction).17 Motor complications were reduced by 94% in the 2 patients who underwent GPi‐DBS.17

Nonmotor Outcome

None of the studies described postoperative UPDRS I score. With regard to cognitive performance, all of the studies showed declines in GBA patients who underwent STN or GPI‐DBS (specific target information not available for all patients). In Lythe and colleagues,40 the cognitive performance of the GBA‐positive patients was significantly worse than the GBA‐negative patients (P = 0.006) after 7.5 years: 7 of 10 (70%) GBA‐positive patients were cognitively impaired compared to 3 of 16 GBA‐negative patients (P = 0.009). In Angeli and colleagues,17 GBA‐positive patients (n = 6) showed a higher decline in cognitive scores after STN‐DBS (4.4 ± 7.3 points per year in Mattis Dementia Rating Scale–2 scores) than mutation‐negative patients (n = 29; 0.5 ± 0.9 points per year). In Weiss and colleagues,35 all 3 GBA carriers developed cognitive impairment at final follow‐up (6 to 10 years after STN‐DBS), whereas 2 of 6 noncarriers presented with cognitive impairment.

Psychiatric symptoms after surgery were not studied in detail in the GBA patients. One cohort study only showed neuropsychiatric symptoms merged with cognitive symptoms as part of the Non‐Motor Symptom Assessment Scale for PD score.40 Another study described the presence of depression in the 3 GBA patients, anxiety developing in 1 patient (and improving in another patient), and onset of hallucinations at 8 to 10 years follow‐up in 1 patient.35

Apart from cognitive and behavioral symptoms, GBA patients composed the group in which general nonmotor performance was more often investigated. In a cohort study, the mean total score in the Non‐Motor Symptom Assessment Scale for PD was significantly higher in the GBA‐positive patients (90.6 ± 21.5) when compared with noncarriers (57.7 ± 27.5; P = 0.004).40 In a case series of 3 GBA patients and 6 noncarrier controls, although both groups exhibited severe nonmotor symptoms during the 6‐year to 10‐year follow‐up, orthostatic dysregulation appeared more pronounced in the GBA carriers.35

Two studies described postoperative measures of disability. In 1 cohort,40 the quality of life (PDQ‐39) did not differ preoperatively between patients with GBA mutations and noncarriers, but at the 7.5‐year follow‐up the GBA patients reported significantly worse quality of life (P = 0.040). A case series with 3 GBA patients showed a mixed quality of life outcome when compared with matched noncarrier controls.35

SNCA Mutation

Motor Outcome

Two patients with SNCA duplication underwent STN‐DBS and were followed for 1 to 2 years.36, 37 The improvement in UPDRS III was 43%36 and 52%.37 LEDD decreased by 29.7%,37 and motor complications improved by 87.5%.36

Nonmotor Outcome

One patient had a UPDRS I score improvement of 65% and a UPDRS II score reduction from 8 to 5 after 1 year.36 Cognitive outcome was overall stable in the 2 STN‐DBS SNCA patients followed for 12 and 48 months.36, 37 Depression scores improved or remained stable.36, 37 Impulse control disorder (ICD) resolved after surgery in 1 patient.36

VPS 35 Mutation

Motor Outcome

A total of 3 patients with the VPS35 mutation underwent STN‐DBS and were followed for 1 to 8 years.38, 39 At the shorter follow‐up (n = 2), the improvement was marked (76%) or satisfactory (36%).38 Satisfactory improvement was still seen at intermediate and longer term follow‐up: 37.1% at 5 years39 and 43.8% at 8 years,38 respectively. LEDD was largely reduced in 1 study38 by around 70% (including at longer term), and decreased by 30.3% (at 1 year) and 19.2% (at 5 years) in the other study.39

Nonmotor Outcome

Nonmotor symptoms and measures of disability were not described in the 3 VPS35 mutation patients.

Other Mutations

One study41 described 3 patients with 22q11.2 deletion syndrome who underwent DBS. GPi was targeted in 2 patients and resulted in marked (>70%) or unsatisfactory motor improvements (<30%). STN was targeted in the remaining case and resulted in more than 50% UPDRS III improvement. Psychiatric and cognitive abnormalities were present in all patients, and levodopa‐induced psychiatric symptoms (hallucinations and ICD) were described in 2 patients, but it is unclear if these were present before or after surgery.

One PINK1 homozygous patient had satisfactory motor improvement at the short‐term (46.5%) and intermediate follow‐ups (43.7%) after STN‐DBS. Nonmotor outcome was not described.20

One study24 described the combined motor outcome of a group of LRRK2, PRKN, and PINK1 patients who underwent STN‐DBS (see Supporting Information Table 2). One of these mutation carriers developed hypomania; none had cognitive impairment, depression, or anxiety following DBS (see Supporting Information Table 3).

Individual Analysis

Individual quantitative data regarding the changes in UPDRS III up to 2 years after STN‐DBS were available for 20 patients. Seven had mutation in the PRKN gene (1 also with a heterozygous PINK1 mutation), 8 in the LRRK2 gene, 3 had a GBA mutation, 1 had a VPS35 mutation, and 1 had an SNCA duplication. A total of 4 PRKN patients were homozygous/compound heterozygous. Four LRRK2 patients were carriers of the R1441G mutation, the other 4 were carriers of R793M, Y1699C, T2031S, and G2019S mutations. The age at PD onset in this sample was 39.4 ± 12.3 years old (range 10–55), age at surgery was 53.8 ± 12.8 years old (range 14–69), and disease duration was 14.8 ± 9.2 years (range 4–45). Age at onset was lower in the PRKN patients when compared with the LRRK2 and GBA patients (PRKN = 28.1 ± 13.0 years old; LRRK2 = 44.0 ± 7.3 years old; GBA = 49.6 ± 3.7 years old; P < 0.05), whereas at the time of surgery the PRKN patients were significantly younger than the GBA group (PRKN, 46 ± 16.2 years old; GBA, 67.0 ± 2.0 years old; P = 0.017), but not significantly different from the LRRK2 group (55.8 ± 7.7 years old; P = 0.281). At shorter term, STN‐DBS significantly improved UPDRS III scores in the PRKN (36.2 ± 15.2 points; P = 0.001; n = 7), LRRK2 (16.8 ± 13.1 points; P = 0.008; n = 8), and GBA groups (17.6 ± 6.6 points; P = 0.044; n = 3). The difference between groups was significant (P = 0.031, analysis of variance); as the PRKN group had significantly higher improvement when compared with the LRRK2 (P = 0.020) and GBA (P = 0.028) groups. Significant LEDD changes were seen in the PRKN group (P = 0.001) and LRRK2 group (P = 0.023), without significant difference between groups (P = 0.169). Because of the small sample of patients, the impact of the single heterozygous PRKN mutations compared to compound heterozygous/homozygous mutations was not assessed.

Side Effects

Side effects are described in the supplemental material (Supporting Information Table 4).

PRKN

One homozygous/compound heterozygous patient in Lohmann and colleagues21 and 2 patients in Romito and colleagues19 experienced ballistic dyskinesias after surgery, which also occurred in some non‐PRKN patients. In the latter study, stimulation‐dependent paresthesia was observed in 2 PRKN patients, and worsening of parkinsonian hypophonia was observed in 1 PRKN patient (and 12 non‐PRKN).

LRRK2

One LRRK2 patient from a cohort of 15 LRRK2 had moderate pneumocephalus (as well as 2 noncarriers) and another LRRK2 carrier had stimulator infection.33 A disabling dystonia‐like pattern (including neck pain and harmful contraction of left shoulder and proximal arm) was seen in 1 patient and was possibly related to the association of medication and stimulation.31

GBA

One study reported that a GBA patient had DBS removed following erosion of the hardware.40

SNCA

No permanent side‐effects or surgical complications were reported. In 1 patient, stimulation‐induced right‐foot dystonia was relieved by modulating stimulation parameters.36

VPS35

In 1 study,38 no side‐effects were observed, and the increased frequency of freezing episodes and falls after surgery were considered likely related to LEDD reduction rather than a deleterious effect of STN‐DBS in gait.38

One study24 described the combined outcome of a group of LRRK2, PRKN, and PINK1 patients who underwent STN‐DBS, and 1 of these mutation carriers developed hypomania after STN‐DBS (as well as 2 nonmutation patients).

Discussion

The aim of this study was to systematically explore and describe studies that have evaluated the effect of DBS in patients with PD‐related mutations. PRKN, LRRK2, and GBA were the most frequent mutations in these patients. Mutations in these genes have been suggested to account for up to 29% of the patients who undergo DBS,17 which is considerably higher than the frequency of mutation‐positive patients in population‐representative cohorts of PD.24

With regard to motor improvement, shorter term follow‐up was frequently reported in PRKN STN‐DBS patients, but not as often in LRRK2 and GBA STN‐DBS patients (around 30% of missing outcome). With this caveat in mind, overall most patients with the PRKN, LRRK2 (except for R144G), and GBA mutations had positive outcomes in the shorter term, with marked (equal or more than 50% improvement) or satisfactory (30%–50% improvement) responses to STN‐DBS. Improvement in the PRKN group was similar when we excluded single heterozygous PRKN carriers. At the intermediate follow‐up, when we excluded single heterozygous PRKN carriers, although most patients with PRKN and LRRK2 mutations had positive outcomes (marked or satisfactory responses) after STN‐DBS, in GBA patients the motor outcome varied equally among marked, satisfactory, and unsatisfactory responses. Nonsystematic report of axial score and other baseline characteristics limit the analysis of outcome predictors in the PRKN group. Nevertheless, possible reasons for PRKN patients with unsatisfactory response were limited levodopa response prior to surgery (patient 3 in Romito and colleagues19), more advanced axial symptoms at a relatively early disease stage,20 and possibly target selection in Angeli and colleagues.17 In the LRRK2 group, the presence of the R144G mutation was associated with higher rates of unsatisfactory motor outcome. These are all preliminary and exploratory conclusions, as the number of patients varied widely among groups. Motor outcome was scarce for GPi‐DBS, precluding evidence‐based conclusions with regard to target selection based on genetic status. There is scarce data on the effect of DBS on dyskinesias in patients with genetic mutations.

Aiming to investigate the effect of different PD‐related mutations in motor response, an individual analysis was performed in a subset of patients with individual outcomes available. In these 20 patients followed up to 2 years after DBS, STN‐DBS resulted in significant motor improvements in the PRKN, LRRK2, and GBA groups, with significant higher improvements in the PRKN group when compared with the LRRK2 (P = 0.020) and GBA (P = 0.028) groups. The different outcomes between the PRKN and LRRK2 patients was possibly because of the inclusion of R144G carriers in the LRRK2 group, which showed poorer outcomes.

With regard to nonmotor symptoms, nonsystematic reporting and small sample size limit interpretation of the results, such as the impact of the T2031S LRRK2 mutation in behavior and psychosis. Despite these limitations, the worsening of cognition was a consistent finding in the GBA patients.17, 35, 40 It is important to highlight that none of the studies describing the GBA‐DBS patients compared the cognitive outcome with GBA‐PD patients not subjected to surgery. PD patients with GBA mutations have been shown to exhibit poorer cognitive performance at baseline,10 and whether STN‐DBS inputs an additional risk and GPi would be a safer target in these patients has not yet been proven. Similarly, hemizygous 22q11.2 deletion—recently determined as a genetic risk factor for PD42—can be associated with cognitive decline,43, 44 and the effect of DBS and target choice needs to be better assessed in these patients. More important, neuropsychiatric symptoms are common features of 22q11.2 deletion syndrome, and it is unknown to what extent they could be affected by DBS. ICDs were not frequently or adequately reported, challenging the interpretation of the relative frequency of ICD after surgery in the PRKN gene mutation patients.

The main limitation of this study was the limited availability of data. Although we only selected studies with available objective outcome, the studies collected report outcome heterogeneously (eg, variable follow‐up intervals), challenging data analysis. Publication bias is also a major limitation of this review. Data on GPi‐DBS was limited, hindering whether genetic status should inform on target selection. Nevertheless, we hypothesize that in PD patients with GBA mutations, GPi would be a safer target, as STN‐DBS could input an additional risk to a population that has poorer cognitive performance at baseline. In addition, patients with genetic mutations with “on” dystonia and severe dyskinesia may benefit from GPi rather than STN stimulation. On the other hand, in patients with hyperdopaminergic syndrome (more prevalent in younger patients), STN would be preferable, as the reduction of dopaminergic medications after surgery is higher with this target.

Conclusion

Our findings show that DBS results in positive outcomes in the shorter term in patients with PRKN, GBA, and LRRK2 (non‐R144G) mutations. Despite limitations concerning small sample size, it is possible that patients carrying GBA mutations may be associated with less sustained motor improvement at intermediate follow‐up (2–6 years) or higher frequency of nonmotor symptoms after surgery, and appropriate counseling would be of utmost importance in these cases. Limitations from this study include the small sample of studies and heterogeneity of baseline and outcome measures, limiting statistical analysis. Future studies with larger samples assessing long‐term effect, nonmotor symptoms, and incidence of side‐effects will shed a light on the management of this complex group of PD patients.

Author Roles

1. Research project: A. Conception, B. Organization, C. Execution

2. Statistical analysis: A. Design, B. Execution, C. Review and Critique

3. Manuscript preparation: A. Writing of the first draft, B. Review and Critique

L.M.O.: 1A, 1B, 1C, 2B, 3A

E.R.B.: 1A, 3B

C.H.A.: 3B

R.P.M.: 3B

A.F.: 1B, 3B

R.G.C.: 1A, 1B, 2A, 2B, 2C, 3B

Disclosures

Ethical Compliance Statement: We hereby confirm that the present study conforms to the ethical standards and guidelines of the journal. The authors confirm that the approval of an institutional review board was not required for this work.

Funding Sources and Conflict of Interest: No specific funding was received for this work. The authors declare that there are no conflicts of interest relevant to this work.

Financial Disclosures for previous 12 months: L.M.O., E.R.B., C.H.A., and R.P.M. report no disclosures. A.F. is funded by grants from the University of Toronto, Weston Foundation, Abbvie, Medtronic, and Boston Scientific; is a consultant for Abbvie, Medtronic, Boston Scientific, Sunovion, Chiesi farmaceutici, UCB, and Ipsen; serves on the advisory boards of Abbvie, Boston Scientific, and Ipsen; and receives honoraria from Abbvie, Medtronic, Boston Scientific, Sunovion, Chiesi farmaceutici, UCB, and Ipsen. R.G.C. has received honoraria from Medtronic, TEVA, UCB, and Roche for lecturing and scientific board services.

Supporting information

Table S1. Baseline characteristics, mutations, and targets.

Table S2. Motor outcome.

Table S3. Cognition, behavior, and mood.

Table S4. Disability, nonmotor symptoms, and side‐effects.