Article Text
PDF
Abstract
Background Impulse control disorders/other compulsive behaviours (‘ICD behaviours’) occur in Parkinson’s disease (PD), but prospective studies are scarce, and prevalence and clinical characteristics of patients are insufficiently defined.
Objectives To assess the presence of ICD behaviours over a 2-year period, and evaluate patients’ clinical characteristics.
Methods A prospective, non-interventional, multicentre study (ICARUS (Impulse Control disorders And the association of neuRopsychiatric symptoms, cognition and qUality of life in ParkinSon disease); SP0990) in treated Italian PD outpatients. Study visits: baseline, year 1, year 2. Surrogate primary variable: presence of ICD behaviours and five ICD subtypes assessed by modified Minnesota Impulsive Disorder Interview (mMIDI).
Results 1069/1095 (97.6%) patients comprised the Full Analysis Set. Point prevalence of ICD behaviours (mMIDI; primary analysis) was stable across visits: 28.6% (306/1069) at baseline, 29.3% (292/995) at year 1, 26.5% (245/925) at year 2. The most prevalent subtype was compulsive eating, followed by punding, compulsive sexual behaviour, gambling and buying disorder. Patients who were ICD positive at baseline were more likely to be male, younger, younger at PD onset, have longer disease duration, more severe non-motor symptoms (including mood and sexual function), depressive symptoms, sleep impairment and poorer PD-related quality of life. However, they did not differ from the ICD-negative patients in their severity of PD functional disability, motor performance and cognitive function.
Conclusions Prevalence of ICD behaviours was relatively stable across the 2-year observational period. ICD-positive patients had more severe depression, poorer sleep quality and reduced quality of life.
- Parkinson’s disease
- impulse control disorders
- non-interventional study
- prospective study
Statistics from Altmetric.com
Background
Impulse control disorders (ICDs) are an increasingly recognised psychiatric complication in Parkinson’s disease (PD), and include compulsive gambling, buying, sexual behaviour and eating. Other compulsive behaviours have also been reported in PD, such as punding (stereotyped, repetitive, purposeless behaviours), hobbyism (eg, writing, repairing or dismantling things, working on projects or computer use), walkabout (excessive, aimless walking or driving) and dopamine dysregulation syndrome (compulsive PD medication overuse).1–5 ICDs and other compulsive behaviours (all together hereafter referred to as ‘ICD behaviours’ or ‘ICDs’) can interfere with life functioning; for example, severe punding and hobbyism may lead to social avoidance and disintegration of family relationships,6 and pathological gambling may have serious financial implications.7
The reported prevalence of ICDs in PD varies widely (from 3.5% to 42.8%),3 8–17 likely reflecting differences in study designs, assessment methods (ie, screening instruments, diagnostic interviews) and sociocultural backgrounds of patients.16 In addition, underdiagnosis and under-reporting is common.18 PD itself does not appear to confer an increased risk, and ICDs have been reported in drug-naïve, newly diagnosed patients at a rate similar to that seen in the general population.19 20 ICDs are likely to develop as a consequence of dopaminergic treatment,21 and are associated with specific structural and functional changes within the brain.22 23 Dopamine receptor agonist (DA) therapy is a risk factor for ICDs3 9 12 24 25; however, associations with levodopa3 14 and monoamine oxidase B inhibitors12 have also been reported. Available data also suggest important contributions of individual predisposing factors.9 19 20 26
Prospective studies assessing ICDs in PD are scarce, and are limited by relatively small sample sizes.27 28 The primary objective of this multicentre non-interventional study was to prospectively assess the presence of ICD behaviours and their subtypes in a large sample of patients with PD over a 2-year period. The secondary objective was to evaluate patients’ clinical features, including non-motor symptoms and quality of life.
Methods
Study design and patients
This non-interventional, prospective, 2-year observational study (ICARUS (Impulse Control disorders And the association of neuRopsychiatric symptoms, cognition and qUality of life in ParkinSon disease], SP0990) was conducted in a routine clinical practice setting in Italy (further details provided in the online supplementary text). The study included adult (aged ≥18 years) outpatients with a diagnosis of probable idiopathic PD (Gelb criteria), who had been treated with clinical benefit for ≥6 months with any approved pharmacological PD treatment (exclusion criteria reported in the online supplementary text). The decision to prescribe treatment (or change treatment schedule) was made by the physician independently of their decision to include the patient in the study.
The recruitment period lasted 12 months, and the observational period was 2 years and included three time points for data collection: baseline (visit 1), year 1 (ie, after 52±4 weeks; visit 2) and year 2 (ie, after 104±4 weeks; visit 3; end of observational period). At each visit, assessments were completed over 1 or 2 days over a total period of no more than a week. At each study visit (baseline, year 1 and year 2), a modified version of the Minnesota Impulsive Disorders Interview29 (mMIDI) was used to screen for clinically relevant ICD behaviours, and included queries for buying disorder, compulsive gambling, compulsive sexual behaviour, compulsive eating and punding behaviours. Patients answered the questions based on their state at the time they were interviewed (details on defining ICD positive by mMIDI in the online supplementary text). At each study visit, patients also completed the Questionnaire for Impulsive-Compulsive Disorders in Parkinson’s Disease (QUIP), a validated questionnaire specifically assessing ICDs and related behaviours in PD. It consists of three sections: (1) ICDs: gambling, sexual, buying and eating behaviours; (2) other compulsive behaviours: punding, hobbyism and walkabout; and (3) compulsive PD medication use5 (details on defining ICD positive by QUIP in the online supplementary text). At each visit, the following assessments were also performed: Hoehn and Yahr (HY), Unified Parkinson’s Disease Rating Scale (UPDRS), Non-Motor Symptom Scale (NMSS), Parkinson’s Disease Sleep Scale-2 (PDSS-2), Parkinson’s Disease-Cognition Rating Scale (PD-CRS), Parkinson’s Disease Questionnaire-8 items (PDQ-8), Beck Depression Inventory II (BDI-II), Frontal Assessment Battery (FAB) and three items of the Neuropsychiatric Inventory (NPI-3): delusions, hallucinations and apathy/indifference. The type of PD medication (including 4 weeks prior to the visit) was recorded at each study visit.
The observational plan of the study, its amendment and Patient Data Consent Form were reviewed and approved by relevant ethics committees. The study was conducted in accordance with Italian legal requirements for non-interventional studies; all patients provided written informed consent for the use of their medical data.
Outcome measures
Prespecified primary variable was the presence of overall ICDs and ICD subtypes according to mMIDI. Diagnosis of an ICD as specified by the protocol was based on: (1) the mMIDI as a means of screening for ICD behaviours, and (2) confirmation of ICD by the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV) criteria.30 However, as confirmation via DSM-IV/other proposed diagnostic criteria31–33 was not consistently followed by the study sites, this study reports all relevant behavioural abnormalities as ICD behaviours (surrogate primary variable). Analysis of the surrogate primary variable included calculation of the point prevalence of ICD behaviours (overall and subtypes), defined as the number (and proportion) of patients who screened positive for ICD behaviours by mMIDI. Point prevalence was calculated at the three study visits (ie, three time points): baseline, year 1 and year 2. The original planned primary analysis also included incidence calculations of ICDs (see online supplementary text for details on the planned incidence calculations of ICDs).
Secondary variable was the presence of overall ICDs and ICD subtypes as measured by QUIP. Analysis of the secondary variable followed the analysis of the primary variable (calculation of prevalence at each study visit).
Other variables included baseline PD symptom severity (non-motor symptoms (NMSS total score and individual domain scores)), sleep disturbances (PDSS-2 total score), PD-related quality of life (PDQ-8 total score), depressive symptoms (BDI-II total score), cognition (Mini Mental State Examination, MMSE, total score, FAB total score and PD-CRS total score), neuropsychiatric symptoms (delusions, hallucinations and apathy/indifference items from NPI-3), motor symptoms (UPDRS III)) and are reported by baseline ICD behaviour status (ICD positive vs negative by mMIDI). In addition, baseline ICD behaviour status is also reported according to baseline demographics and clinical characteristics, lifestyle and social profile, and state of disease according to PD treatment (see online supplementary text for details on state of disease according to the PD treatment classification).
Subgroup analyses of the surrogate primary variable according to ongoing PD therapy are reported, with patients classified based on the therapy they received during the 4 weeks prior to a given visit (baseline, year 1, year 2): (1) levodopa (without DA), (2) DA (without levodopa), (3) DA + levodopa and (4) other (neither DA or levodopa). All patients could also be receiving any other approved pharmacological PD treatment. ICD behaviour prevalence by ongoing PD therapy at each visit is reported. As a patient’s PD therapy could change at any time, the PD therapy groups could comprise different patients at each of the baseline, year 1 and year 2 visits. Therefore, comparisons of ICD prevalence by PD therapy over time are not possible.
Statistics
Based on the literature,9 it was calculated that an expected ICD frequency of 6.6%, with an accepted interval limit of 4.9% (corresponding to 1.7% precision) and a 95% confidence intercal (CI), required a sample size of 818 patients; assuming a 18% drop-out rate, 1000 patients were to be recruited. Data are reported for the Full Analysis Set (FAS; all patients who completed at least the mMIDI, HY staging scale and UPDRS at baseline). In the analysis of baseline demographics and clinical characteristics by ICD status at baseline (ICD positive vs negative), the Wilcoxon/Mann-Whitney test was used for continuous variables, and the χ2 test was used for categorical variables. All analyses were exploratory in nature and p values of <0.05 do not infer statistical significance.
Results
Patients
The study was conducted between January 2011 and February 2014. Of 1095 patients enrolled at 62 sites in Italy, 97.6% (1069) comprised the FAS. In total, 13.3% (142/1069) patients discontinued: non-compliance (5 patients), withdrawal due to the use of prohibited medications (8, of whom 3 received treatment with Duodopa), lost to follow-up (91), other (38); overall, 86.7% (927/1069) patients completed the 2-year study. All patients (mean±SD age: 65.7±9.5 years (range: 24.6–88.8); 64.2% (686/1069) male) were Caucasian apart from one. The mean±SD age at PD onset was 59.6±10.7 years (range: 21–88), and PD duration was 6.1±5.0 years (range: 0–41); patients had a mean±SD UPDRS III score of 14.2±6.8 (range: 1–42) and HY stage of 2.0±0.65 (range: 0–4).
Prevalence of ICDs and ICD subtypes over the 2-year period
Primary analysis by mMIDI demonstrated a relatively stable overall ICD prevalence across the three visits of the 2-year observational period, with 28.6% (306/1069) of patients screening positive for an ICD at baseline, 29.3% (292/995) at year 1 and 26.5% (245/925) at year 2. Prevalence of individual ICD subtypes was also relatively stable at the baseline, year 1 and year 2 visits, with compulsive eating the most prevalent, followed by punding and compulsive sexual behaviour (figure 1A).
Prevalence of ICD subtypes over the 2-year observational period according to (A) mMIDI, and (B) QUIP (FAS). *Term classified under ‘Other behaviour’. FAS, Full Analysis Set; ICD, impulse control disorder; mMIDI, modified Minnesota Impulsive Disorder Interview; QUIP, Questionnaire for Impulsive-Compulsive Disorders in Parkinson’s Disease.
Analysis by QUIP also demonstrated stable point prevalence across the three visits of the 2-year period. The overall prevalence was generally higher versus mMIDI, with 34.2% (366/1069) of patients screening positive for an ICD at baseline, 34.8% (346/995) at year 1 and 32.1% (297/925) at year 2. Overall, hobbyism was the most prevalent ICD subtype, followed by binge-eating disorder, and sexual behaviour (figure 1B).
Number of patients with multiple ICD subtypes
The majority of patients with an ICD reported one type of behaviour (figure 2). In analysis by mMIDI, at baseline 9.5% (102/1069) of all patients (or 33.3% (102/306) of patients with ICDs) were positive on multiple ICD modules (ie, ≥2); this was similar at year 1 and year 2 (figure 2A). In analysis by QUIP, 15.4% (165/1069) of all patients (or 45.1% (165/366) of patients with ICDs) were positive on multiple ICD modules at baseline; again this was similar at year 1 and year 2 (figure 2B).
Number of patients with different ICD subtypes over the 2-year period according to: (A) mMIDI, and (B) QUIP (FAS). FAS, Full Analysis Set; ICD, impulse control disorder; mMIDI, modified Minnesota Impulsive Disorder Interview; QUIP, Questionnaire for Impulsive-Compulsive Disorders in Parkinson’s Disease.
Clinical features and risk factors associated with ICDs
Baseline demographics, PD characteristics, and lifestyle and social profile by baseline ICD status
A higher proportion of males (32.5%; 223/686 males) screened ICD positive at baseline versus females (21.7%; 83/383 females). ICD-positive patients were also younger, younger at PD onset and had longer disease duration (table 1). ICD-positive and -negative patients had similar severity of PD symptoms and functional disability (HY stage), motor performance (UPDRS III) and cognitive function (MMSE, FAB and PD-CSR) (table 1). However, depressive symptoms (BDI-II), PD-related quality of life (PDQ-8), sleep impairment (PDSS-2) and overall non-motor symptoms (NMSS total score) were more severe in ICD-positive patients (table 1). ICD-positive patients had numerically higher scores in all individual NMSS domains; the largest numerical differences between scores (explorative p value of <0.0001) were observed in ‘sleep/fatigue’, ‘mood/apathy’, ‘attention/memory’, ‘sexual function’ and ‘miscellaneous’ domains (table 1). Neuropsychiatric symptoms (NPI-3) were also more common among ICD-positive patients (table 1); apathy/indifference was the most common in both groups, although less common in the ICD-negative group (19.6% (149/763)) versus ICD-positive group (33.6% (102/306)). A higher proportion of ICD-positive patients were in suboptimal treatment, or classified as complicated (table 1).
Baseline demographics and PD characteristics by ICD status at baseline (FAS)
Among ICD-positive patients, those with multiple ICDs had more severe depressive symptoms (mean±SD BDI-II score: 15.1±8.81; n=99) versus those with a single ICD (11.4±6.86; n=200).
Generally, ICD-positive patients were more likely to regularly consume alcohol or smoke, and less likely to be married or have higher education (university or postgraduate degree) (table 2).
Lifestyle habits and social profile by baseline ICD status (FAS)
Prevalence of ICDs by PD therapy over the 2-year period
At baseline, the prevalence of ICDs in patients receiving ongoing levodopa or DA was comparable (26.5% (56/211) vs 26.7% (64/240)); prevalence was numerically higher in patients receiving a combination of DA and levodopa (30.1% (179/594)) (figure 3). There were only 22 patients receiving other PD therapy (neither DA nor levodopa) at baseline; of these, five were ICD positive. A similar pattern of ICDs prevalence by ongoing PD therapy was observed at year 1 and year 2, with the highest prevalence observed in patients receiving both DA and levodopa (figure 3). Again, few patients were receiving other PD therapy at year 1 (14; of these, 4 were ICD positive), and year 2 (7; 1 ICD positive).
Prevalence of ICDs by PD therapy over the 2-year observational period (FAS). *Patients could be receiving any other approved pharmacological PD treatment (eg, monoamine oxidase B inhibitors, catechol-O-methyltransferase inhibitors, amantadine). †Ongoing by at least 4 weeks before baseline, year 1 or year 2. ‡Other PD therapy (neither DA nor levodopa) is not included because of low patient numbers (reported in the text). DA, dopamine receptor agonist; FAS, Full Analysis Set; ICD, impulse control disorder; PD, Parkinson’s disease.
Discussion
This is the largest prospective observational study assessing ICDs in patients with PD. Analysis of more than 1000 Italian outpatients demonstrated a relatively stable prevalence of overall ICD behaviours and subtypes across the 2-year period. Moreover, a comprehensive evaluation of clinical features highlighted more severe non-motor symptoms, depression and worse quality of life in ICD-positive patients.
The majority (87%) of patients completed the full 2 years of observation, and the overall ICD prevalence as assessed by mMIDI remained relatively stable at ~28% across the three visits. Stable prevalence was also observed over the 2-year period for each of the five ICD subtypes, with compulsive eating the most common (11.8% to 12.9%). Two or more types of ICDs were recorded in 8.1% to 9.5% of patients across the three study visits, supporting previous studies suggesting different ICDs can occur simultaneously in patients with PD.3 In addition to the mMIDI, the current study also employed the QUIP, allowing for direct comparison of the two screening tools for the first time. Prevalence by QUIP was slightly higher versus the mMIDI, with ~5% more patients screening positive for an ICD at each visit. As the QUIP assesses three additional modules (hobbyism, walkabout and PD medication use) than mMIDI, this, at least in part, may explain the higher overall prevalence of ICDs by QUIP. Altogether, assessment of ICDs using mMIDI and QUIP yielded generally comparable results, suggesting that both are suitable tools to screen for ICDs in patients with PD, even if some differences in ICD subtype prevalence have been observed between the two tools. For example, the prevalence of compulsive sexual behaviour by mMIDI was slightly lower than by QUIP (by ~2%) at baseline and year 2, and almost twice lower at year 1 (7.9% vs 14.6%). Why such a large difference should occur at year 1 only is unknown, and additional research with a longer follow-up may be warranted to further explore the difference in the prevalence of compulsive sexual behaviour between mMIDI and QUIP. Of note, it has been suggested that people may be more likely to report an abnormal sexual behaviour in a private setting (eg, on a self-administered questionnaire) than when interviewed by an investigator.34 Furthermore, in the validation study of QUIP,5 out of the total 157 patients, only 7 (4.5%) had binge-eating disorder, which may possibly limit validation of this section of the questionnaire. Although the prevalence of binge-eating disorder (QUIP) was similar and only slightly higher than compulsive eating (mMIDI) – ~1%–2% depending on the visit, which we deemed not clinically relevant – the possible limitation of this section of the QUIP needs to be considered. Finally, the QUIP is rated at a 12th grade reading level, and it is not fully understood how this would affect the screening of patients with PD with less than 12 years of formal education.5 In the ICARUS study, a higher proportion of ICD-positive (43.1%) versus ICD-negative (30.7%) patients had at least 12 years of formal education (high school or university/postgraduate), and it is unknown whether the education level could have influenced the ability of some patients to respond on QUIP.
This observation of more than 1000 patients represents a broad sample of patients with PD across different disease stages and with a wide range of symptom severity. Analysis of baseline demographics, lifestyle and social profile indicated that ICD-positive patients were more likely to be male, younger, smoke and consume alcohol regularly, but less likely to be married than ICD-negative patients; this is in line with previous reports.3 8–12 14–17 27 There are no comprehensive studies comparing risk factors for ICD development between individuals with and without PD, and studies in the general population typically focus on one particular ICD behaviour, rather than the overall ICD spectrum. For example, a cross-sectional study in Korean individuals found associations between pathological gambling and male gender, alcohol use, nicotine dependence and being unmarried.35 Therefore, certain demographic and lifestyle/social variables likely increase risk of ICD development, and are independent of PD itself.
In terms of PD characteristics themselves, patients who reported ICDs in this study were younger at PD onset and had longer disease duration. This is in line with previous studies,8–10 13 15 16 36 and may reflect an increased duration of dopaminergic treatment in these patients. However, the actual severity of PD cardinal symptoms as assessed by HY staging or motor performance (UPDRS III) did not differ between ICD-positive and ICD-negative patients, supporting the notion that PD itself does not confer an increased ICD risk.19 20 This is consistent with previous reports for HY staging, although mixed results have been reported for UPDRS III.15 19 27 36 37 The current study also provides a comprehensive analysis of the range of non-motor symptoms associated with PD and their relationships with ICDs. There was no difference in cognitive function between patients with and without ICDs; previous studies have reported varied results.36 38 39 ICD-positive patients had more severe depressive symptoms (with the severity increasing with the number of ICDs), worse sleep quality and poorer quality of life, in line with previous reports.3 7–11 13 15–17 20 40 Of note, a previous case-control study reported an association between increasing severity of depression and ICDs in both patients with PD and healthy controls, and the authors concluded that this association did not seem to be specific to PD.19 In line with more severe depressive symptoms, patients with ICDs in our study were also more commonly affected by apathy/indifference (as assessed by NPI-3) than those without. In addition, a higher proportion of ICD-positive versus -negative patients had hallucinations. This contradicts a previous study,11 but may be related to longer disease duration in ICD-positive versus -negative patients. Furthermore, we assessed a broad range of non-motor symptoms using the NMSS, and observed a higher severity of overall non-motor symptoms (NMSS total score) in ICD-positive patients at baseline. This was also apparent for all individual NMSS domains, with greatest numerical differences observed for the ‘sleep/fatigue’, ‘mood/apathy’, ‘attention/memory’, ‘sexual function’ and ‘miscellaneous’ domains. Greater deficits in the ‘mood/apathy’ and ‘sleep/fatigue’ domains (together with larger reductions in sleep quality (PDSS-2)) may relate to more severe depressive symptoms (higher BDI-II scores), highlighting depression as a condition that may be particularly important when monitoring ICDs. The causality of any association between the increased severity of non-motor symptoms and the increased frequency of ICDs (including the direction of any causality) is not known; however, the data suggest that monitoring specific conditions like depression may be important in at-risk patients.
The data consistently show that at the three study visits over the 2-year period, there were no apparent differences in the ICD prevalence between patients receiving either ongoing levodopa or ongoing DAs. However, at all three visits, ICD prevalence in patients receiving a levodopa and DA combination was higher than in patients receiving a DA or levodopa alone. Assessments of ICD prevalence by PD therapy over time were not possible, as patients were permitted to switch medication at any time. In addition, as patients were classified according to PD therapy that was ongoing by a minimum of 4 weeks prior to a given visit, it is possible that some of the recorded ICDs may have been reflective of a previous therapy (ie, received for ≥4 weeks prior to the visit). For example, a prospective cohort study reported a median time of 23.0 months after DA initiation to ICD onset.27 In a recent ‘post hoc’ analysis of ICD behaviours reported as adverse events in long-term studies of DA rotigotine, the number of incident cases during the first 30 months was relatively low and increased afterwards, suggesting a lag time to ICD onset in many patients.41 Furthermore, medication doses were not recorded in this non-interventional study, and thus it was not possible to calculate the levodopa equivalent daily dose (LEDD) for each PD therapy group. Therefore, it cannot be excluded that in some patients with ICDs in the DA group (eg, those that were younger), the LEDD were similar to those in the levodopa group. Despite these limitations, the analysis supports the emerging body of evidence suggesting that development of ICDs in PD is far more complex than the use of DAs. For example, the DOMINION study reported an association between levodopa and ICDs, although at a lower prevalence than with DAs; the same study found that in patients receiving DAs, concomitant use of levodopa increased the odds of an ICD by ~50%.3 Regarding DAs themselves, recent studies of non-ergoline DAs in PD suggested higher rates of ICDs in patients treated with oral versus transdermal DAs,39 and oral short-acting versus oral long-acting/transdermal DAs,40 suggesting that DA administration route and formulation may play a role in ICD development. Furthermore, pre-PD history of an ICD has been reported as the main risk factor for ICD development during treatment with a DA,9 and specific personality traits were reported to play a role in the development of pathological gambling in patients with PD.26 Finally, because the role of DAs in ICD development in PD is increasingly recognised, patients considered at risk for ICDs development may not have been prescribed a DA.17 This may have contributed to the relatively low prevalence of ICDs with DAs versus levodopa in our study. Altogether, individual predisposing risk factors coupled with PD therapy are likely to increase risk for the development of ICDs. Therefore, continuous active monitoring is important for early ICD identification and management.
This study has several limitations. First, because of the non-interventional, observational study design the results are only exploratory in nature. Second, the sample size calculation was based on confirmation of an ICD by DSM-IV (or other ICD diagnostic criteria). As the ICD confirmation was not performed consistently, the resulting threshold for detection of an ICD by mMIDI only was lower than the originally planned threshold of detection by mMIDI followed by a confirmation by DSM-IV/other diagnostic criteria. Thus, the study’s precision may have been affected (as ICDs of marginal clinical relevance may have been captured). Third, the study assessed point prevalence, and does not inform on ICD prevalence between the study visits. Finally, there was no matched control group to compare ICD presence in individuals without PD or patients with PD not receiving any PD therapy.
Conclusions
The prevalence of ICD behaviours and their subtypes was relatively stable throughout the 2-year observational period, and no meaningful differences were observed between patients receiving DAs and those receiving levodopa. However, higher ICD behaviour prevalence was seen in those receiving both DAs and levodopa. In addition to the previously reported clinical characterisation of ICD behaviours in PD, our results suggest that the severity of motor symptoms and cognitive function do not differ between patients with and without ICD behaviours. However, particular attention should be paid to the presence of specific non-motor symptoms, for example, depression, sleep quality and sexual function.
Supplementary file
Acknowledgments
The authors report this study on behalf of the ICARUS study group (participating sites: A.O. Universitaria Ospedale Policlinico Consorziale, Bari; Istituto Neurologico Mediterraneo NEUROMED, Pozzilli; UOC Neurologia Ospedaliera, Azienda Ospedaliero-Universitaria OO.RR., Foggia; Azienda Ospedaliera Cardinale Giovanni Panico, U. O. di Neurologia, Tricase; ASL MT P.O. Madonna delle Grazie, Matera; A.O. Universitaria Policlinico Tor Vergata, Roma; Dipartimento Scienze Neurologiche Università Degli Studi Federico II, Napoli; Università degli Studi di Roma ‘La Sapienza’ Dipartimento di Scienze Neurologiche, Roma; Policlinico Universitario Gemelli, Roma; IRCCS S. Raffaele Pisana, Roma; UO di Neurologia, Azienda Ospedaliera di Rilievo Nazionale A.Cardarelli, Napoli; A.O. Universitaria Sant`Andrea, Roma; Università di Modena e Reggio Emilia, Ospedale C. Sant'Agostino e Estense, Baggiovara, Modena; Università di Bologna, Dipartimento di Scienze Neurologiche, Bologna; A.O. Universitaria di Parma, Dipartimento di Neuroscienze, Sezione Di Neurologia,
Parma; A.O. Universitaria S. Giovanni Battista-Molinette Di Torino, Torino; Universita' Degli Studi Di Genova, Genova; A.O. Universitaria Policlinico Monserrato Di Cagliari, Cagliari; Ospedale Maria Vittoria, Divisione di Neurologia, Torino; A.O. Universitaria Policlinico Di Sassari, Clinica Neurologica, Sassari; Ospedale Villa Sofia, Palermo; Azienda Sanitaria 8 Di Vibo Valentia, Vibo Valentia; Azienda Ospedaliera Cannizzaro, Catania; ASP Enna - P.O. Umberto I, Enna; Azienda Ospedaliera Universitaria Integrata Verona, Neurologia B, Verona; Azienda ULSS 3 Serenissima, Ospedale dell'Angelo, Neurologia, Venezia Mestre, Venezia; U.O. Neurologia, Spedali Civili Brescia, Brescia; Casa Di Cura Villa Margherita, Arcugnano Vicenza; Universita' Degli Studi Di Padova, Palazzina Neuroscienze, Clinica Neurologica I, Padova; U.O. Neurologia, Ospedale Versilia, Lido di Camaiore, Viareggio; Universita' Degli Studi G. D'annunzio Di Chieti, Clinica Neurologica, Chieti Scalo, Chieti; Azienda Ospedaliera Di Perugia, Ospedale S.Maria Misericordia, Clinica Neurologica, Perugia; A.O. Universitaria ‘Ospedali Riuniti’ di Ancona, Ospedale Torrette, Clinica di Neuroriabilitazione, Dipartimento di Scienze Neurologiche, Ancona; A.O. Universitaria Careggi Di Firenze, Neurologia I, Ambulatorio Parkinson, Firenze; PO 'SS. Filippo E Nicola' Avezzano, Aquila; Ospedale Nuovo S. Giovanni Di Dio Torregalli, Neurologia, Firenze; Nuovo Ospedale Di Prato Santo Stefano, Neurologia Prato; P.O. Di Summa, Perrino U.O. Neurologia, Brindisi; Ospedale San Giovanni Battista, Roma; IRCCS Istituto C. Besta Disturbi del Movimento, Milano; Centro Parkinson CTO, ASST Nord Milano, Milano; IRCCS Fondazione Istituto Neurologico Casimiro Mondino Di Pavia, Pavia; Ospedale Mauriziano Umberto I Di Torino, Torino; Ospedale Sant`Andrea La Spezia, La Spezia; Presidio Ospedaliero Di Savona, Cairo Montenotte, Savona; Ospedale di Circolo e Fondazione Macchi Varese, Varese; Azienda Sanitaria Universitaria Integrata di Udine, Ospedale Santa Maria della Misericordia, Dipartimento di Neuroscienze, Neurologia, Udine; Ospedale Maggiore Policlinico, Dipartimento di Scienze Neurologiche, Università degli Studi di Milano, Milano; Osp. Maggiore di Modica U.O. Di Neurologia, Ragusa; Azienda Ospedaliera S.Giuseppe Moscati Di Avellin, U.O.C. di Neurologia, A.O.R.N. ‘S.G. Moscati’, Avellino; AOU Pisana, Stabilimento di Santa Chiara, Pisa; Casa di Cura Villa dei Gerani, Catania; Azienda ULSS 3 Serenissima, Ospedale Civile SS Giovanni e Paolo, Neurologia, Venezia; Fondazione Opera San Camillo, Casa di Cura San Pio X, Milano; L’Azienda Ospedaliera di Melegnano-Presidio di 'Vizzolo Predabissi’, Milano; Azienda Ospedaliera Universitaria OO.RR. S.Giovanni di Dio e Ruggi d'Aragona, Salerno; Ospedale Sant'Eugenio, Roma; Ospedale San Filippo Neri, Roma; Istituto Scientifico di Riabilitazione di Veruno, Veruno; Ospedale SS.Annunziata, Cosenza; Università degli Studi di Palermo, Palermo). The authors thank the patients and their caregivers in addition to the investigators and their teams who contributed to the study. The authors also acknowledge Karolina Rzeniewicz, PhD, CMPP (Evidence Scientific Solutions, London, UK) for writing assistance which was funded by UCB Pharma, Brussels, Belgium, and Cédric Laloyaux, PhD, (Strategic Publication Lead Neurology, UCB Pharma, Brussels, Belgium) for publication coordination. The authors also acknowledge Lars Bauer, MD (UCB Pharma, Monheim am Rhein, Germany), and Elisabeth Dohin, MD (UCB Pharma, Brussels, Belgium) for scientific and medical input into the data analyses and interpretation.
References
Footnotes
Correction notice Since this paper was first published Mirko Avesani and Francesco Palandin have been added as authors. A new affiliation was also included for these authors.
Contributors AA: research project conception, organisation, and execution; review and critique of statistical analysis; review and critique of manuscript. PB: research project conception, review and critique of statistical analysis, review and critique of manuscript. UB: research project conception and execution; review and critique of statistical analysis, review and critique of manuscript. KA: research project conception, organisation, and execution; review and critique of statistical analysis; review and critique of manuscript. MA: research project conception, organisation, and execution; review and critique of statistical analysis, review and critique of manuscript. PS: research project conception and execution, review and critique of statistical analysis, review and critique of manuscript.
Funding This study was supported by UCB Pharma, Monheim am Rhein, Germany.
Competing interests AA, PB, UB and PS were study investigators on this UCB Pharma sponsored study. AA has received consultancy fees/honoraria from AbbVie, UCB Pharma, Zambon, Angelini, Lundbeck, Mundipharma and Medtronic; has served on advisory boards for AbbVie and Acadia, provided expert testimony for Boehringer Ingelheim (pathological gambling cases); and received grants from Neureca Foundation, Gossweiler Foundation, Mundipharma, Italian National Research (project no RF-2009-1530177, RF- 2010-2319551) and Horizon2020 (project no 643706). PB has received personal fees from Acorda, Union Chimique Belge and Zambon, and grants from AbbVie, Biotie and Zambon. UB has received personal compensation for serving on scientific advisory boards from UCB Pharma and Zambon. PS has received consultancy fees/honoraria from UCB Pharma, Zambon and Chiesi; has served on advisory boards for UCB Pharma and Italian National Health Minister Research (project no RF-2013-2719661). KA is a salaried employee of UCB Pharma. MA is a former employee of UCB Pharma and received stock options from her employment.
Provenance and peer review Not commissioned; externally peer reviewed.