Impulsivity Traits in Parkinson's Disease: A Systematic Review and Meta‐Analysis

ABSTRACT Background In Parkinson's disease (PD), impulsivity as a personality trait may be linked to the risk of developing impulse control disorders (ICDs) during dopaminergic therapy. However, studies evaluating differences in trait impulsivity between patients with PD and healthy controls or between patients with PD with and without ICDs reported partly inconsistent findings. Objectives We conducted a systematic review and meta‐analysis (Preferred Reporting Items for Systematic Reviews and Meta‐Analyses) of studies comparing Barratt Impulsiveness Scale (BIS‐11) scores between patients with PD and healthy controls and between patients with PD with and without ICDs. Methods Eligible studies were identified through a systematic search in 3 databases. Mean differences with 95% confidence intervals (CIs) for BIS‐11 total and subscale scores were separately calculated for studies comparing patients with PD and healthy controls and patients with PD with and without ICDs. Meta‐regressions were performed to explore sources of heterogeneity (percentage of men, age, disease duration, and levodopa equivalent daily dose). Results A total of 40 studies were included in the quantitative analyses. BIS‐11 total scores were significantly higher in patients with PD compared with healthy controls (mean difference 2.43; 95% CI, 1.03, 3.83), and in patients with PD with active ICDs compared with patients without ICDs (6.62; 95% CI, 5.01, 8.23). No significant moderators emerged by meta‐regression analyses. Conclusions The present meta‐analysis supports that impulsivity, as a personality trait, may characterize patients with PD, even in the absence of ICDs. Moreover, these data corroborate findings of clinical studies reporting higher levels of trait impulsivity in PD patients with ICDs compared with patients without ICDs.

increased caregiver burden and represent a critical issue for the clinical management of patients with PD.ICDs may result from the interaction between predisposing factors (ie, demographic, psychological, clinical, and genetics factors) and dopaminergic medication. 7,83][14] Moreover, evidence of higher levels of trait impulsivity have been reported in patients with PD when compared with healthy controls even in the absence of ICDs. 14,15However, some studies did not report differences either from the comparison between patients with PD and healthy controls 16,17 or between patients with and without ICDs. 18,19Therefore, whether impulsivity may represent a main vulnerability factor for the development of ICDs during dopaminergic treatment still remains an open question.
The overall goal of the present study was to investigate impulsivity as a personality trait in patients with PD using a meta-analytic approach to the current literature.To this end, we performed 2 separate meta-analyses aimed to verify: (1) whether impulsivity traits are higher in patients with PD compared with agematched healthy individuals and (2) whether and to what extent impulsivity traits differ between PD patients with and without ICDs.Moreover, we explored the possible influence of demographic and clinical factors through meta-regression analyses.We used the total score of the Barratt Impulsiveness Scale (BIS-11) as a primary measure.BIS-11 is the most administered self-report questionnaire to assess impulsive personality traits. 20Recently, a relationship between BIS-11 score and awareness of motor intention has been found both in healthy subjects 21,22 and patients with PD. 23

Search Strategy and Selection Criteria
We performed a systematic and comprehensive literature search up to May 2022 using the databases PubMed (https://pubmed.ncbi.nlm.nih.gov/),Web of Science (https://webofknowledge. com), and PsychINFO (https://search.ebscohost.com/).The selected keywords were combined using the Boolean operator AND and OR.The search input was the following: ("personality trait*" OR "impulsivity trait*" OR "Barratt Impulsiveness Scale" OR "Barratt Impulsivity Scale" OR "BIS-11") AND ("Parkinson's disease" OR "Parkinson disease" OR "impulse control disorder*").Additional studies were searched from the references of all identified publications.No language restrictions were applied.Eligibility was determined by a 2-step procedure performed by 3 of the authors (F.G., C.N., and G.G.).First, the titles and abstracts of all identified articles were screened.In the second step, the full texts of studies, according to predefined eligibility criteria, were independently examined, and agreement was reached after discussion.Our study was conducted following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines. 24e included controlled studies published in peer-reviewed journals reporting impulsivity traits assessed by the BIS-11 in patients with a diagnosis of idiopathic PD and in age-matched healthy controls or comparing patients with PD with and without ICDs.Included studies had to provide mean and standard deviation (SD) values of the BIS-11 scores or data to calculate them.Case reports, conference proceedings, and publications available only in abstract form not reporting detailed data were excluded.Studies reporting impulsivity traits assessed by self-report questionnaires (eg, Eysenck Personality Questionnaire; Eysenck Impulsiveness Questionnaire-I-7; Dickman Impulsivity Inventory; Lifetime History of Impulsive Behaviors; the Impulsive/Premeditated Aggression Scale; or the more recent Urgency, Premeditation, Perseverance, Sensation Seeking impulsive behavior scale [UPPS]) other than the BIS-11 were also excluded.We opted for this conservative approach as different self-report questionnaires developed to assess impulsive traits are based on different theories and models of impulsivity, emphasizing different aspects of this multifaceted psychological construct.
Studies conducted in patients with PD undergoing deep brain stimulation (DBS) were excluded unless the assessments were clearly carried out before the implantation.

Data Extraction
Data were collected independently by 3 authors (F.G., C.N., and G.G.) using a standardized data extraction form.For each study, the mean and SD of the BIS-11 total score were extracted or calculated.If available also subscale scores (ie, attentional, motor, and nonplanning) were extracted.BIS-11 score values were retrieved from text, tables, or estimated by graphs (details are given in Supplementary Material S1).In case of discrepancies, data from tables were chosen.Moreover, authors were contacted to retrieve missing or incomplete data.Other details on data extraction are given in Supplementary Material S1.
In addition, the following data were also extracted for each study: number of participants for each group, mean age, percentage of men, disease duration, Hoehn and Yahr Scale, mean score of the Unified Parkinson Disease Rating Scale (UPDRS) Part III on and/or off medication, mean levodopa equivalent daily dose (LEDD), and mean dopamine agonist LEDD.Moreover, data on the presence of ICDs in the groups were also extracted.Data were independently extracted and cross-checked by 3 review authors (F.G., C.N., and G.G.), who also independently assessed the methodological quality of each study.
Data are available from the corresponding author on request.

Primary Measure
The BIS-11 is a 30-item self-report questionnaire widely used to measure impulsive personality traits. 25

REVIEW
focus attention or concentrate on the task at hand (attentional impulsivity), (2) tendency to act on the spur of the moment without thinking (motor impulsivity), and (3) lack of planning and forethought (nonplanning impulsivity).Translations of the BIS-11 are available in several languages. 20The internal consistencies (Cronbach's α) reported for the BIS-11 total score from different translations all fall within an acceptable range (0.71-0.83). 20Similarly, the test-retest reliability was acceptable.Therefore, the BIS-11 is considered a valid tool to assess the construct of impulsiveness in both clinical and nonclinical samples. 20A short version of the scale (15 items instead of 30) has been recently validated, and normative data have been provided. 26This short version can be used as a quick screening tool to assess impulsivity in a clinical setting.However, to the best of our knowledge, the BIS short version has not been used yet in patients with PD.

Data Analysis
The meta-analysis has been conducted using the software RevMan version 5.4.1 (Review Manager, The Cochrane Collaboration, 2020).Mean differences (continuous data) with 95% confidence intervals (CIs) for BIS-11 total score were separately calculated for studies comparing (1) patients with PD and agematched healthy controls and (2) patients with PD with and without ICDs.Included studies reporting BIS-11 subscale scores (attentional, motor, and nonplanning) were used for a secondary analysis comparing patients with PD with and without ICDs.Heterogeneity between studies has been assessed by I 2 and Cochran's Q test.Given the heterogeneity among studies (see the Results), data were analyzed using a random-effects model.
A weighted least squares linear meta-regression was performed to explore sources of heterogeneity in the BIS-11 total score mean difference between patients with PD and healthy controls.The following factors were used as independent variables: percentage of men in the sample, age, disease duration, and LEDD in patients with PD.Because of the rates of missingness across studies, each potential moderator was evaluated in a separate meta-regression model.To evaluate the influence of LEDD in the mean difference of BIS-11 total score between patients with PD with and without ICDs, the mean LEDD difference was used as an independent variable in a meta-regression.The metaregression model was weighted by the inverse of variance of each study.The meta-regression analysis was performed using the software IBM (Armonk, NY) SPSS 20.0; significance was set at P < 0.05.The publication bias has been evaluated by funnel plot inspection.A symmetric funnel plot suggests no publication bias.The presence of asymmetry in the funnel plot was statistically evaluated by Egger's regression asymmetry test using the open-source software Jeffreys's Amazing Statistics Program -JASP (version 0.16.2;JASP Team 2022, University of Amsterdam, Amsterdam, The Netherlands).

Results of the Study Search
The flowchart of the article selection is illustrated in Figure 1.Our search yielded 124 potentially eligible studies.After full-text assessment of these articles, 40 studies (reference list is provided in Supplementary Material S1) from 2007 to 2022 were included in our quantitative analyses (28 studies for the comparison between patients with PD and age-matched controls and 18 studies for the comparison between patients with PD with and without ICDs).[29] Quantitative Analysis: PD Versus Healthy Controls The 28 selected studies included 1061 patients and 1000 healthy age-matched control subjects.The meta-analysis revealed a statistically significant mean difference (2.43; 95% CI, 1.03, 3.83), with higher BIS-11 total scores in patients with PD compared with control subjects (Fig. 2).The heterogeneity was high (I 2 = 77%, Cochran's Q test P < 0.001).
The symmetry of the plots as well as Egger's test suggested no evidence of publication bias for BIS-11 score mean difference (z = À0.371;P = 0.711) (Supplementary Material S1).

Quantitative Analysis: Patients with PD with Versus without ICDs
The 18 selected studies included 655 and 976 patients with and without ICDs, respectively.The meta-analysis revealed a statistically significant mean difference (6.62; 95% CI, 5.01, 8.23), with higher BIS-11 total scores in PD patients with ICDs compared with patients without ICDs (Fig. 3).The heterogeneity was moderate (I 2 = 52%, Cochran's Q test P < 0.001).
The symmetry of the plots as well as Egger's test suggested no evidence of publication bias for BIS-11 score mean difference (z = 1.222;P = 0.222) (Supplementary Material S1).
When mean LEDD difference between patients with and without ICDs was evaluated as moderator in the meta-regression analysis, no statistically significant effect was observed (β = 0.163, F 1,15 = 0.381, P = 0.547).

Discussion
The 2-fold aim of the present meta-analysis was (1) verifying whether impulsivity trait, as assessed by the BIS-11, is higher in patients with PD with respect to age-matched healthy individuals and (2) evaluating whether and to what extent the impulsivity traits differ between patients with PD with and without ICDs.The results showed a significantly higher level of impulsivity in patients with PD compared with healthy controls.In addition, BIS-11 total scores are significantly higher in patients with PD with active ICDs compared with patients without ICDs.5]34 Although the BIS-11 mean difference between patients with PD and controls was quite small, such a difference remained significant even after excluding studies in which a proportion of patients with PD reported ICDs.It is noteworthy here that in most of the reviewed studies, ICDs are not systematically assessed in the healthy group and, as a consequence, the difference between patients with PD and healthy controls could be underestimated; even more so if we assume that also healthy subjects may have ICDs and related behaviors. 9ifferences between patients with PD and controls were not significantly influenced by any demographic or clinical factor (ie, percentage of men in the sample, age, disease duration, and LEDD) within the patient groups as revealed by meta-regression analyses.Particularly noteworthy is that the mean LEDD did not emerge as a significant moderator of the relationship between impulsivity traits and ICDs.All BIS-11 domains were higher in patients with PD who were ICD positive compared with ICD negative.In keeping with Aumann et al, 14 the larger difference between the 2 groups emerged for the nonplanning impulsivity domain.In a recent study, 15 patients with PD showed elevated scores in all domains of the BIS-11 compared with healthy controls, whereas patients with cerebellar ataxia exhibited differences in specific domains.Conversely, in the present meta-analysis, significant differences emerged in the attentional and nonplanning domains, but not in the motor subscale.It must be said, however, that few studies reported the subscale (attentional, motor, and nonplanning) scores.Hence, no robust conclusion can be drawn from the present meta-analysis on the impulsivity profile characterizing patients with PD.
8][29] In the study by Antonini et al, 27 a large sample of drug-naïve patients with PD were screened for the presence of ICDs and assessed for levels of impulsivity and obsessive-compulsive symptoms.The proportion of patients who reported at least 1 ICD was 17.5%, a frequency similar to that reported in age-matched healthy controls. 9In patients with PD, the mean BIS-11 total scores (63.7 AE 9.5; range, 45-91) was below the normative mean values in the agematched healthy population. 27Patients with PD who were ICD positive showed higher scores in the attentional impulsiveness subscale of the BIS-11 compared with patients with PD who were ICD negative, with no differences in the total score.Similarly, no differences between patients with de novo PD and healthy controls were reported by Poletti et al 28 in the BIS-11 total score (64.8 AE 9.2 vs. 64.7 AE 9.6, respectively).van der Vegt et al 29 evaluated 13 drug-naïve patients with PD and 12 healthy age-matched control subjects who underwent functional magnetic resonance imaging recording during a 2-choice gambling task.The BIS-11 total score did not differ between the groups (66.0 AE 5.6 vs. 65.3AE 5.8 in patients with PD and in healthy controls, respectively).Recently, Lee et al 19 conducted a multicenter, open-label trial in which the baseline characteristics of 50 patients with PD with ICD were compared with those of 60 medicated and 40 drug-naïve PD control groups.The BIS-11 total score did not differ between the 3 groups of patients.Hence, available data on drug-naïve patients seem to downsize the role of impulsive personality trait in predicting the risk of developing ICDs.Interestingly, a recent study conducted in patients with de novo PD identified 3 phenotypes based on personality traits and their relationships with motor and neuropsychiatric symptoms. 35Impulsivity was observed in the "neuropsychiatric phenotype" characterized by high harm avoidance, low novelty seeking, hypodopaminergic neuropsychiatric symptoms, and higher impulsivity trait.Given the heterogeneity of PD in the early stages, it is conceivable that specific phenotypes FIG. 2. Barratt Impulsiveness Scale total scores in the 28 studies selected for the meta-analysis comparing patients with Parkinson's disease (PD) to age-matched healthy controls.CI, confidence interval; df, degree of freedom; SD, standard deviation.FIG. 3. Barratt Impulsiveness Scale total scores in the 18 studies selected for the meta-analysis comparing patients with Parkinson's disease (PD) with and without impulse control disorders (ICDs).CI, confidence interval; df, degree of freedom; SD, standard deviation.
MOVEMENT DISORDERS CLINICAL PRACTICE 2023; 10(10): 1448-1458.doi: 10.1002/mdc3.138391455 may be more associated with the risk of developing ICDs.Moreover, it has been hypothesized that the level of impulsivity may be involved in boosting the severity of ICDs rather than increasing their risk of occurrence. 18Reasoning on the results of the present meta-analysis study, it is evident that a clear definition of the role of impulsivity traits as predisposing factors for the development of ICDs can only be drawn from longitudinal studies.Such studies should aim to assess impulsivity personality profile in patients with de novo PD before starting dopaminergic treatment and to verify longitudinally the incidence of ICDs in individuals with baseline levels of impulsivity exceeding normative values.To the best of our knowledge, no such longitudinal studies have been conducted yet.
A limitation of the present meta-analysis is that only studies assessing impulsivity traits by the BIS-11 were selected.This choice may limit the generalizability of the results.However, it should be noted that the literature based on different self-report tools assessing impulsivity in patients with PD is quite limited for some questionnaires and absent for others.
In a relatively small number of studies, the UPPS was used to assess impulsivity traits in patients with PD instead of or in addition to the BIS-11.In the study by Bayard et al, 34 patients with PD without ICDs had greater levels of urgency, lack of premeditation, and lack of perseverance with respect to healthy controls, whereas levels of sensation seeking were higher in patients with ICDs compared with patients without ICDs.Similarly, in some dimensions of the UPPS, higher scores were also reported by Dawson et al 36 and Olley et al 37 in patients with ICDs.In contrast, some studies did not observe significant differences between patients with PD and healthy controls in the UPPS scores. 38,39nterestingly, Hlavat a et al 38 reported significant group differences in the BIS-11 scores but not in the UPPS subscale scores, confirming that different questionnaires evaluate different dimensions of impulsivity.
There is broad consensus that impulsivity is a multidimensional and heterogeneous concept that should not be considered as a unitary construct, instead consisting of a series of independent subtypes reflecting a variety of behaviors and processes. 40Accordingly, using voxel-based morphometry analyses, Marín-Lahoz et al 41 showed that different self-report and behavioral impulsivity measures reflect distinct brain structural correlates.Namely, the impulsivity traits appeared to be associated with lower gray matter volume in the dorsolateral prefrontal cortices.In a recent study conducted in patients who underwent bilateral DBS of the subthalamic nucleus, Ricciardi et al 42 showed a positive correlation between the oscillatory activity in the α band and the impulsivity traits (BIS-11 score) in patients with PD, irrespective of the presence and severity of active ICDs.The authors proposed that this spectral feature may represent a neural biomarker associated with impulsive behavior.
In conclusion, the results of the present study support the view that impulsivity as a personality trait may characterize patients with PD even in the absence of ICDs.Moreover, our meta-analysis corroborates findings of clinical studies reporting higher levels of impulsivity in PD patients with ICDs compared with patients without ICDs.Although the present results broaden the knowledge on the personality profiles of patients with PD, 35,43 they are currently not exhaustive.Thus, the complex relationship between impulsivity traits and ICDs in PD warrants further investigation.