Table e-1. Details of studies selected for qualitative synthesis.
Study / Patients / Setting / Clinical diagnosis / Gold standard diagnosis / Results / PD misdiagnosis (N° of subjects and/or percentage respect to all patients diagnosed as PD or non-PD)
Rajput AH et al, Can J Neurol Sci 1991 / 59 with PS diagnosis (mean age at onset not reported). Collected from 1968 to 1990. / Clinic-based / Initial clinical diagnosis performed by a general neurologist or a movement disorder specialist.
Final clinical diagnosis performed by the same movement disorder specialist after an average of 11.7 years of illness.
No specific diagnostic criteria used, but presence of 2 of the 3 cardinal signs (bradykinesia, rigidity and resting tremor), no identifiable cause of the disease and no clinical evidence of widespread lesions of the CNS. / Pathological examination / Initial clinical diagnosis: correct PD diagnosis in 28/43. Sens.= 90%; Spec.= 46%; PPV= 65%; NPV= 81%; accuracy= 69%
Final clinical diagnosis: corrected PD diagnosis in 31/41. Sens.= 100%; Spec.= 64%; PPV= 76%; NPV= 100%; accuracy= 83% / For final diagnosis: 10 FP (24.4%). MSA (4, 9.8%); SN cell loss but not inclusion (2, 4.9%); Only NFT pathology in SN and LC (2, 4.9%); AD (1, 2.44%); DIP (1, 2.44%).
No FN
Hughes AJ et al, jnnp 1992 / 100 with clinical PD diagnosis (mean age at onset 64.5; range 31-85). Collected from 1987 to 1990. / Clinic-based / Clinical diagnosis performed by movement disorders specialist or general neurologist or geriatrician.
All cases were also re-evaluated with the UKPDSBRC clinical criteria / Pathological examination / Initial clinical diagnosis: correct PD diagnosis in 76/100. PPV= 76%
Diagnosis of PD based on UKPDSBRC criteria: correct PD diagnosis in 73/89. Sens.= 96%; Spec.= 33%; PPV= 82%; NPV= 73%; accuracy= 81% / For initial clinical diagnosis: 24 FP (24%). PSP (6, 6%); MSA (5, 5%); AD (3, 3%); Alzheimer-type pathology (3, 3%); Vascular lesions (3, 3%); Nigral atrophy without LB (2, 2%); PEP (1, 1%); ET (1, 1%)
Using UKPDSBRC criteria: 16 FP (18%) and 3 FN (27.3%) but diagnosis not specified
Hughes AJ et al, Neurology 1992 / The same patients of the above study (Hughes AJ et al jnnp 1992) / Clinic-based / Clinical diagnosis plus these three best predictors derived from logistic regression analysis: no atypical features for PD, asymmetrical onset and no suggestion of a cause for another PS / Pathological examination / Sens.= 68%; Spec.= 83%; PPV= 93%; NPV= 45%; accuracy= 72% / 4 FP (7·1%) and 24 FN (54%) but diagnosis not specified
Hughes AJ et al, Arch Neurol 1993 / 100 cases with histologically confirmed PD diagnosis (mean age at onset 62.5±9.2). Collected from 1986 to 1990. / Clinic-based / Clinical diagnosis performed by movement disorders specialist or general neurologist or geriatrician. / Pathological examination / Clinical diagnosis of PD in 90 patients.
Sens.= 90% / 10 FN (10%): atypical parkinsonism (5), vascular parkinsonism (2), MSA (1), DIP (1), post-traumatic parkinsonism (1).
Comment: the most of the FN and a part of the TP have evidence of mixed pathology at autopsy (as vascular lesions, striatal plaques, AD, AD-type lesions, diffuse Lewy body disease)
Litvan et al, Arch Neurol 1998 / 105 patients with PS diagnosis (mean age at onset 60 y) / Clinic-based / Diagnosis of primary neurologists (from clinical practice) and diagnosis of six movement disorders expert (three senior and three junior) based on clinical vignettes. An initial (based only on the clinical judgment) and a final diagnosis (with all information including laboratory and imaging data; mean follow-up of 9 years) were provided.
No specific diagnostic criteria were used / Pathological examination / Initial diagnosis of PD by expert raters (median values): Sens.= 73.3%; Spec.= 85.6%; PPV= 45.9%; NPV= 95%; accuracy= 83.8 %
Final diagnosis of PD by expert raters (median values): Sens.= 80%; Spec.= 92.2%; PPV= 64%; NPV= 96.3%; accuracy= 90.5%
Moderate interrater reliability at the first visit and substantial at the last visit. Differences in the seniority of raters were not significant.
Initial diagnosis of PD by primary neurologists: Sens.= 93.3%; Spec.= 76.7%; PPV= 40%; NPV= 98.6 %; accuracy= 79%
Final diagnosis of PD by primary neurologists: Sens.= 93.3%; Spec.= 87.8%; PPV= 56%; NPV= 98.7 %; accuracy= 88.5% / Misdiagnosis more frequent in the first visit and involved primarily DLB, MSA and PSP (exact numbers not reported)
Meara J et al, Age and Aging 1999 / 402 subjects on anti-parkinsonian medication (mean age at onset 67 y) / Community-based / Diagnosis performed by general practitioners in 59% of cases / UKPDSBRC criteria / True parkinsonism in 299 (74%)
Diagnosis of PD in 213/402. PPV= 53% / 189 FP (47%): Other type of parkinsonism (74, 39.2%); DIP (12, 6.3%); ET (50, 26.5%); gait apraxia due to VaE (37, 19.6%); Dementia preceding motor signs (16, 8.5%)
Jankovic J et al, Arch Neurol 2000 / 800 patients with clinical PD diagnosis enrolled in the multicentric DATATOP trial between 1987 and 1988 (mean age at onset 59.5 years) / Clinic-based / Diagnosis performed by movement disorders experts in the early stage of disease (H&Y I-II) / Reassessment after a mean of 6±1.4 years (based on investigator’s confidence, atypical clinical features, imaging data, levodopa response, and autopsy examinations if available / Diagnosis of PD confirmed in 735/800 patients. PPV= 92% / 65 FP (8.1%). Final diagnosis of false positive cases was reported only for 9 patients with available neuropathological examination: MSA (3); PSP (2); DLB (1); PD + AD (1); Chronic cerebral infection (1); Vascular lesions (1)
Jellinger KA, Adv Neurol 2001 / 710 patients with clinical diagnosis of PD. 110 patients collected between 1957 and 1970 (mean age at onset 60.1 years); 600 patients collected between 1971 and 1999 (age at onset not available). / Clinic-based / Clinical diagnosis without specific criteria. Expertise of the doctors who performed the diagnosis not specified. / Pathological examination / Sample 1957-1970: diagnosis of PD confirmed in 89/110 patients. PPV= 80.9%
Sample 1971-1999: diagnosis of PD confirmed in 496/600 patients. PPV= 82.6% / 21 FP (19%) in sample 1957-1970: AD (1, 0.9%); VaE (4, 3.6%); PSP (4, 3.6%); MSA (5, 4.6%); nigral atrophy, unclassified (1, 0.9%); DLB (3, 2.7%); CBD (1, 0.9%); others (1, 0.9%); PEP (1, 0.9%).
104 FP (17.3%) in sample 1971-1999: AD (25, 3.5%); VaE (13, 2.2%); PSP (17, 2.8%); MSA (14, 2.3%); nigral atrophy, unclassified (4, 0.7%); AD+VaE (3, 0.5%); DLB (23, 4.3%); CBD (3, 0.5%); normal, ET? (4, 0.7%); others (2, 0.3%).
Hughes AJ et al, Neurology 2001 / 100 patients with clinical diagnosis of PD (mean age at onset 62.2; range 29 to 82).
Collected between 1996 and 1998. / Clinic-based / Clinical diagnosis without specific criteria performed by neurologists in 86 subjects (movement disorders experts in 61), geriatricians in 7 and internal medicine specialists in 7.
Re-evaluation of diagnosis according to:
UKPDSBRC criteria
Calne criteria (clinically definite)
Gelb criteria (clinically possible and probable)
Three selected clinical features (asymmetrical onset, no atypical features, no possible etiology for another parkinsonian syndrome) / Pathological examination / Clinical diagnosis of PD confirmed in 90/100. PPV= 90%
UKPDSBRC criteria: Sens.= 90%; Spec.= 30%; PPV= 92%; NPV= 25%; accuracy= 84 %
Clinically definite PD (Calne criteria): Sens.= 90%; Spec.= 20%; PPV= 91%; NPV= 18%; accuracy= 83 %
Clinically possible PD (Gelb criteria): Sens.= 87%; Spec.= 40%; PPV= 93%; NPV= 25%; accuracy= 82 %
Clinically probable PD (Gelb criteria): Sens.= 72%; Spec.= 40%; PPV= 92%; NPV= 14%; accuracy= 69 %
Three selected clinical features: Sens.= 67%; Spec.= 30%; PPV= 90%; NPV= 9%; accuracy= 63 % / For clinical diagnosis: 10 FP (10%). MSA (6, 6%); PSP (2, 2%); PEP (1, 1%); VaE (1, 1%)
Lees AJ et al, Neurology 2001 / 782 patients with clinical PD diagnosis enrolled in the multicentric PDRGUK trial between 1985 and 1990 (mean age 62.8 years, mean disease duration 14 months, range 1-144) / Clinic-based / Clinical diagnosis mostly by movement disorders experts according to UKPDSBRC clinical criteria / Reassessment of the initial diagnosis by experts. Clinical follow-up of 10 years / Final clinical diagnosis of PD in 733/782.
PPV= 93.7% / 49 FP (6.3%). Diagnosis not specified.
Hughes AJ et al, Brain 2002 / 143 patients with PS (mean age at onset 55.5, range 5-80). Collected between 1990 and 1999. / Clinic-based / Clinical diagnosis performed by 11 movement disorders specialists (final diagnosis). Diagnostic criteria not specified. / Pathological examination / Final clinical diagnosis of PD in 73 (confirmed in 72): Sens.= 91.1%; Spec.= 98.4%; PPV= 98.6%; NPV= 90 %; accuracy= 94.4%
Projection of results on a community-based scenario (based on prevalence data): PPV= 99.8%; NPV= 49.9 %. / For final diagnosis of PD: 1 FP (PSP) (1.4%)
6 FN (8.6%): clinical diagnosis of MSA (4, 5.7%), undetermined parkinsonism (1, 1.4%), PEP (1, 1.4%)
Comment: in 44 of 122 patients where full follow-up information was available the diagnosis was revised and in 42 the initial diagnosis of PD was changed in a final diagnosis of non-PD.
Schrag A et al, jnnp 2002 / 200 patients with a diagnosis of parkinsonism, tremor with onset over age 50 y, or who had ever received antiparkinsonian drugs (who were seen and were resident in the study
area on prevalence day 1 July 1997) / Community-based / Clinical diagnosis performed by general practitioners or neurologists or geriatricians / UKPDSBRC criteria (with the exception of isolated Babinski sign in elderly patients with otherwise typical PD) with a follow-op of at least one year. / Initial diagnosis of PD by all doctors: Sens.= 88.1%; Spec.= 73%; PPV= 84.7%; NPV= 78.3 %; accuracy= 82.5%
Initial diagnosis of PD by neurologists or geriatricians: Sens.= 93.5%; Spec.= 64.5%; PPV= 88.7%; NPV= 76.9 %; accuracy= 86.2%
Initial diagnosis of PD by general practitioners: Sens.= 73.5%; Spec.= 79.1%; PPV= 73.5%; NPV= 79.1 %; accuracy= 76.6% / For initial diagnosis by all doctors: 20 FP (15.3%). Non-parkinsonian tremor (4, 3%), VaP (6, 4.6%); PSP (4, 3%); MSA (3, 2.3%); idiopathic torsion dystonia (2, 1.5%); dementia without parkinsonism (1, 0.8%).
15 FN (21.7%): previous diagnosis of atypical parkinsonism (1, 1.45%); VaP (1, 1.45%); non-parkinsonian tremor (11, 15.9%); on antiparkinsonian drugs without specific diagnosis (2, 2.9%)
Bower JH et, Mov Disord 2002 / 39 incident cases of parkinsonism
in Olmsted County, MN, for the years 1976 to 1990 (Rochester
Epidemiology Project) / Community-based / Clinical diagnosis based on revision of available medical records by a movement disorders expert / Pathological examination / Sens.= 43.8%; Spec.= 73.9; PPV= 53.8%; NPV= 65.4 %; accuracy= 61.5 / 6 FP (37.5%): PSP (4, 25%); MSA (2, 12.5%).
9 FN (39.1%): parkinsonism in dementia (5, 21.7%); DIP (3, 13%); MSA (1, 4.3%)
Stoffers D, Eur J Nucl Med Mol Imaging 2005 / 72 patients with early-stage, untreated IPD (mean age at onset 59.4 years) / Clinic-based / Clinical diagnosis by movement disorders experts according to UKPDSBRC clinical criteria (all with reduced [123I]β-CIT SPECT binding) / Reassessment of the initial diagnosis by experts. Clinical follow-up 36-80 months (mean ± SD, 62±11 months), plus MRI and/or IBZM SPECT when available / Diagnosis of PD confirmed in 62/72 patients. PPV= 86.1% / 10 FP (13.8%): MSA (7, 9.7%); DLB (2, 2.8%); PSP (1, 1.4%)
Jellinger, J Neural Transm 2007 / 740 patients with clinical diagnosis of PD. Collected between 1971 and 2006 (partially overlapping with the sample reported in Jellinger 2001): 380 patients 1971-1988 (mean age at onset 67.1 y), 260 patients 1989-2001 and 100 patients 2001-2006 (age at onset not available). / Clinic-based / Clinical diagnosis performed by movement disorders experts in sample 1989-2001, while by less specialized neurological, intern and geriatric specialists in samples 1971-1988 and 2001-2006.
No specific diagnostic criteria used / Pathological examination / Sample 1971-1988: diagnosis of PD confirmed in 322/380 patients. PPV= 84.7%
Sample 1989-2001: diagnosis of PD confirmed in 230/260 patients. PPV= 88.5%
Sample 2001-2006: diagnosis of PD confirmed in 76/100 patients. PPV= 76% / 58 FP (15.3%) in sample 1971-1988: AD (10, 2.6%); VaE (13, 3.5%); PSP (7, 1.8%); MSA (8, 2.2%); nigral atrophy, unclassified (2, 0.5%); AD+VaE (2, 0.5); DLB (14, 3.6%); CBD (1, 0.2%); normal, ET? (1, 0.5%); others (1, 0.2%).
30 FP (11.5%) in sample 1989-2001: AD (5, 1.9%); VaE (2, 0.8%); PSP (3, 1.1%); MSA (3, 1.1%); nigral atrophy, unclassified (1, 0.4%); AD+VaE (1, 0.4%); DLB (12, 4.6%); normal, ET? (2, 0.8%); others (1, 0.4%).
23 FP (23%) in sample 2001-2006: AD (4, 4%); VaE (4, 4%); PSP (3, 3%); MSA (1, 1%); AD+VaE (1, 1%); DLB (8, 8%); normal, ET? (1, 1%); others (1, 1%).
Caslake R et al, jnnp 2008 / Incident cohort of 66 patients (over 18 months from November 2002) with PS and at least one year of follow up (mean age 74.4±10.1; median disease duration 12.7 months, interquartile range 6-8-24.2; mean age at diagnosis 75.0±10.5 yrs and mean age at onset 73.5±9.01). / Community-based / Referral by general practitioners. Baseline clinical diagnosis by movement disorders specialists. No diagnostic criteria used. / Latest clinical diagnosis by experts (at least one year after baseline clinical evaluation).
Latest diagnosis on UKPDSBRC criteria (in five patients pathological examination) / Clinical diagnosis at baseline: 46 PD, 20 non-PD
Baseline clinical vs latest clinical PD diagnosis: 33 TP, 13 FP, 16 TN, 4 FN. Sens.= 89.2%; Spec.= 55.2%; PPV= 71.7%; NPV= 80 %; accuracy= 74.2%
Baseline PD clinical vs latest PD diagnosis on UKPDSBRC criteria (or autopsy): 28 TP, 18 FP, 15 TN, 5 FN. Sens.= 84.8%; Spec.= 45.5%; PPV= 60.9%; NPV= 75 %; accuracy= 65.2%