Serum caffeine and metabolites are reliable biomarkers of early Parkinson disease

Abstract

Objective: To investigate the kinetics and metabolism of caffeine in serum from patients with Parkinson disease (PD) and controls using liquid chromatography-mass spectrometry.

Methods: Levels of caffeine and its 11 metabolites in serum from 108 patients with PD and 31 age-matched healthy controls were examined by liquid chromatography-mass spectrometry. Mutations in caffeine-associated genes were screened by direct sequencing.

Results: Serum levels of caffeine and 9 of its downstream metabolites were significantly decreased even in patients with early PD, unrelated to total caffeine intake or disease severity. No significant genetic variations in CYP1A2 or CYP2E1, encoding cytochrome P450 enzymes primarily involved in metabolizing caffeine in humans, were detected compared with controls. Likewise, caffeine concentrations in patients with PD with motor complications were significantly decreased compared with those without motor complications. No associations between disease severity and single nucleotide variants of the ADORA2A gene encoding adenosine 2A receptor were detected, implying a dissociation of receptor sensitivity changes and phenotype. The profile of serum caffeine and metabolite levels was identified as a potential diagnostic biomarker by receiver operating characteristic curve analysis.

Conclusion: Absolute lower levels of caffeine and caffeine metabolite profiles are promising diagnostic biomarkers for early PD. This is consistent with the neuroprotective effect of caffeine previously revealed by epidemiologic and experimental studies.

Classification of evidence: This study provides Class III evidence that decreased serum levels of caffeine and its metabolites identify patients with PD.

Figure 1. Receiver operating characteristic (ROC) curves

ROC curves were made by the Youden index using the maximum of sensitivity + specificity − 1 ([A] caffeine, [B] caffeine and its main metabolites, [C] all metabolites). The optimal cutoff point was calculated using the maximum sensitivity + specificity. The main metabolites were paraxanthine, theophylline, and theobromine. AUC = area under the curve.