Objective: To determine the accuracy of clinical diagnosis and research diagnostic criteria in a population-representative cohort of Parkinson’s disease (PD) against a neuropathological gold standard diagnosis.

Background: Accurate diagnosis of PD is important to guide treatment, inform prognosis and to identify research participants. Research diagnostic criteria have been developed to improve diagnostic accuracy. Previous autopsy series have been highly selected, leading to a high risk of selection bias. One source of bias is under-representation of older and frailer patients, who may be more difficult to diagnose because of effects of comorbid illness and dual brain pathology. We therefore investigated diagnostic accuracy in an autopsy series of PD derived from a population-based incidence cohort.

Method: The PINE study identified all new diagnoses of parkinsonism in North-East Scotland over 4.5 years. Consenting patients were seen annually from diagnosis to death, which included a detailed clinical examination. A movement disorders neurologist reviewed diagnoses annually. All participants were invited to consent to post-mortem examination. Using all data from clinical follow-up in life, we calculated the sensitivity and specificity of UK PD Brain Bank (UKPDBB) criteria, MDS criteria and clinical diagnosis, using post-mortem diagnosis as the gold standard.

Results: Of 377 people with suspected incident parkinsonism, 355 (94%) consented to follow-up. Of these, 316 were clinically parkinsonian on follow-up. 279 (88%) died after median follow-up of 12 years. 71 (25%) had autopsy performed (mean age at diagnosis 73.7 years, 72% male). Of these, 52 (73%) had a pathological diagnosis of PD. Sensitivity and specificity of the UKPDBB criteria were 57% (95% CI 42–71) and 79% (54–94); of the MDS criteria were 59% (44–72) and 90% (67–99); and of clinical diagnosis was 94% (84-99) and 79% (54–95). The most common reasons for false-negative UKPDBB diagnoses were early severe autonomic involvement (N=9), Babinski sign (N=6), and insufficient supportive criteria (N=11); and for false-negative MDS diagnoses slow vertical saccades (N=10) and cortical sensory loss (N=9, mostly in PD dementia).

Conclusion: Rigid application of formal diagnostic criteria led to lower sensitivity than clinical diagnosis. Accuracy of these criteria was lower than in previous studies in younger patients. Diagnosis of PD remains difficult and accurate diagnostic biomarkers are needed.

« Back to MDS Virtual Congress 2021

MDS Abstracts - https://www.mdsabstracts.org/abstract/diagnostic-accuracy-in-an-incidence-cohort-of-parkinsons-disease/