Effectiveness of Interventions for Dysphagia in Parkinson Disease: A Systematic Review.

22 of 27 included study assessments were rated high risk of bias; none of the 10 studies that reported on participant selection used random selection. Mild baseline swallowing impairment was common, which may have created a ceiling effect that hid real improvements. Several studies tested only a single bolus or a single consistency, limiting how well they capture real-world swallowing. 10 of 20 studies with multiple raters did not report interrater reliability, making it hard to trust the scoring. Heterogeneity in VFSS, FEES, and EMG protocols across studies made pooling impossible, so no overall effect estimate exists. All included studies were single-center, and the review excluded manometry, qualitative data, and single-arm studies. The Evidence Project risk-of-bias tool does not capture some domains (participant blinding, selective reporting) that the Cochrane tool assesses.

Funded by NIDCD Grant R01DC011020 (NIH extramural), awarded to Catriona M. Steele. No industry funding or conflicts of interest declared.

Do not read the trend toward faster swallowing or less residue as evidence that any specific treatment works. The review rates 22 of 27 included studies as high risk of bias, found no study used random participant selection, and concludes the evidence is too low in certainty to guide practice.