J Neural Transm (2007) 114: 1161–1165 DOI 10.1007/s00702-007-0746-0 Printed in the Netherlands Features of probable multiple system atrophy patients identified among 4770 patients with parkinsonism enrolled in the multicentre registry of the German Competence Network on Parkinson’s disease U. Wu ¨ llner 1 , T. Schmitz-Hu ¨ bsch 1 , M. Abele 1 , G. Antony 2 , P. Bauer 3 , K. Eggert 2 1 Department of Neurology, UKB, Bonn, Germany 2 Department of Neurology, Philipps-University of Marburg, Marburg, Germany 3 Department of Medical Genetics, University of Tu ¨bingen, Tu ¨bingen, Germany Received 17 January 2007; Accepted 12 April 2007; Published online 18 May 2007 # Springer-Verlag 2007 Summary We identified 221 patients with probable multiple system atro- phy (MSA) among 4770 patients enrolled in the multicentre registry of the German Competence Network on Parkinson’s disease (PD) according to the established consensus criteria to characterize their clinical presentation. Analyses of more than 100 recorded clinical items revealed several specif- ics: I) 50% of patients with probable MSA had asymmetry of symptoms at disease onset and tremor at rest was present in 25%; II) a positive response to levodopa was recorded in 51% of patients identified initially with severe autonomic failure and cerebellar ataxia; III) a positive family history was recorded in 11% (n ¼ 23), two of these patients were identified with spino- cerebellar ataxia type 3 (SCA3). Thus asymmetry of symptoms, tremor at rest and a positive response to levodopa are not as specific for idiopathic PD as believed previously. Patients with SCA3 may present with the clinical features of MSA. Keywords: Multiple system atrophy, autonomic dysfunction, patient reg- istry, German Competence Network on Parkinson’s disease (KNP e.V.) Introduction The German Competence Network on Parkinson’s disease (KNP e.V.) is a nation-wide network of 30 movement dis- order clinics, founded to promote information and research on Parkinson’s disease (PD) and related disorders, includ- ing multiple system atrophy (MSA) and progressive supra- nuclear palsy (PSP) in Germany (Eggert et al., 2005). A central internet-accessible database which records more than 100 clinical items and detailed information on diag- nostics and medication is used to collect and store the pseudonymised clinical data of the participating patients. We aimed to characterize the clinical symptoms and evolu- tion of symptoms over time in patients with probable MSA. To this end, we used algorithms corresponding to the es- tablished consensus criteria for MSA to verify the clinicians’ diagnoses which due to the network system may vary in accuracy (Gilman et al., 1998). Subjects and methods The multicentre registry data collected up to 30.06.2005 were used for the present analysis; at that time 4770 patients had been enrolled. All patients consented to the data analyses and appropriate ethics approval had been obtained by all participating institutions. The data capture template (minimal data set) records a detailed medical and pharmacological history including present drug treatment and diagnostic procedures performed in the past (www.kompetenznetz-parkinson.de). Probable MSA was defined according to the consensus criteria as (severe orthostatic hypotension OR {persistent urinary incontinence AND erectile dysfunction in men}) AND (either {poor levodopa-response AND bradykinesia AND any other parkinsonian fea- ture} (corresponding to MSA-P) or {gait ataxia AND any other cerebellar dysfunction feature} (corresponding to MSA-C)). Severe orthostatic hy- potension was defined according to the consensus criteria as a drop in RRsys > 30 mmHg or RRdia > 15 mmHg after standing up; in the following text, orthostatic dysfunction is used synonymously. For the data base search for probable MSA cases, missing data were treated as ‘‘feature not present’’ to avoid false positives. The same data set was retrieved twice if patients fulfilled the criterion for autonomic failure and both the criterion for par- kinsonism and the criterion for cerebellar dysfunction. Duplicate data sets retrieved by both algorithms were considered only once for further analysis of the clinical characteristics of patients with probable MSA. For analysis of clinical symptoms in PD patients we similarly used search algorithms to identify patients with probable PD: ({sustained response to levodopa AND no feature suggestive of alternative diagnosis} AND {three out of: resting Correspondence: Prof. Dr. Ullrich Wu ¨llner, Department of Neurology, University of Bonn, Sigmund-Freud-Str. 25, D-53105 Bonn, Germany e-mail: ullrich.wuellner@ukb.uni-bonn.de