Accuracy of clinical diagnosis of Parkinson’s disease: a systematic review and meta-analysis

Objective To evaluate the diagnostic accuracy of clinical diagnosis of Parkinson’s disease (PD) reported in the last 25 years by a systematic review and meta-analysis. Methods We searched for papers published between 1988 and August 2014. Studies were included if reporting diagnostic parameters regarding clinical diagnosis of PD, or crude data. The selected studies were sub-classified based on different study setting, type of test diagnosis and gold standard. Bayesian meta-analyses of available data were performed. Results We selected 20 studies, including 11 using pathological examination as gold standard. Considering only these 11 studies the pooled diagnostic accuracy was 80.6% [95% credible interval (CrI), 75.2% - 85.3%]. Accuracy was 73.8% [95% CrI, 67.8% - 79.6%] for clinical diagnosis performed mainly by non-experts. Accuracy of clinical diagnosis performed by movement disorders experts rising from 79.6% [95% CrI, 46% - 95.1%] of initial assessment to 83.9% [95% CrI, 69.7% - 92.6%] of refined diagnosis after follow-up. Using UKPDSBRC criteria the pooled diagnostic accuracy was 82·7% [95% CrI, 62.6% - 93%]. Conclusion The overall validity of clinical diagnosis of PD is not satisfying. The accuracy did not significantly improve in the last 25 years, particularly in the early stages of disease, where response to dopaminergic treatment is less defined and hallmarks of alternative diagnoses such as atypical parkinsonism may not have been emerged. Misclassification rate should be considered to calculate the sample size both in observational studies and RCT. Imaging and biomarkers are urgently needed to improve the accuracy of clinical diagnosis in vivo.