Accuracy of the Early Diagnosis of Parkinson’s Disease

We have read with interest the recent report by Virameteekul et al (1) on the estimated clinical diagnostic accuracy for early Parkinson’s disease (PD) at the Queen Square Brain Bank (QSSB). We have a few important concerns.

1. The QSSB’s clinical diagnostic accuracy for early PD is at the high end of the ranges reported by other centers and therefore may not be representative. This situation is reminiscent of that for reported clinical diagnostic accuracies for Alzheimer’s disease (AD), where for many years there was a very wide range between reporting centers (2), with sensitivities between 41% and 100% and specificities between 37% and 100%. Multicenter studies using curated, centralized and standardized databases would greatly help the PD research community resolve inter-center variability, as was done for AD with the National Alzheimer’s Coordinating Center (2) database, drawing on more than 30 centers. The subsequent introduction of PET amyloid imaging confirmed a lower accuracy than that reported by many reputable centers.

2. The QSBB’s study is based on a retrospective review of medical records. It would be necessary to perform analogous prospective studies in order to adequately assess the meaning of the currently-presented retrospective figures. Our own clinical diagnostic accuracy figures, while not utilizing MDS PD criteria, are based on prospective, annual and standardized research-devoted clinical examinations by movement disorders subspecialists in the Arizona Study of Aging and Neurodegenerative Disorders, a longitudinal clinicopathological study since 1997. Our initial finding of low early diagnostic accuracy for PD was supported by two other autopsy-validated studies (3) which together reported on 71 subjects with early PD. In our updated recent study (4), for those with symptom duration of less than 5 years, a probable PD diagnosis at first visit was only neuropathologically confirmed in 71% while for those with duration greater than 5 years, 89% were neuropathologically confirmed.

3. The definition of “early PD” differs between centers, varying from 2-8 years and with some centers using time since first PD symptoms and some using time since first PD diagnosis. This creates additional variability in diagnostic accuracy estimates. We have used the diagnosis given at the patient’s very first visit but it is not clear whether the QSSB group did the same. There may also be a variable interval between first visit to a primary care provider and subsequent first referral visit to a neurologist. For therapeutic trials, it would be optimal to enroll subjects at the first visit that clinical signs are detectable since dopaminergic neuron loss and striatal deafferentation is then already at about 40% to 60% or greater and proceeds rapidly at an annual rate of up to 8-10% per year to a profound 70-80% or greater by 4-5 years after symptom onset (5–9).

4. The high rate, in the current QSSB study, for neuropathologically-diagnosed multiple system atrophy, (MSA), at 50% of that for PD, is unusual, as other centers have reported much lower prevalences of neuropathologically-confirmed MSA. How this might affect their diagnostic accuracy estimates is uncertain.