No CrossRef data available.
Published online by Cambridge University Press: 02 June 2017
Background: In patients with acute stroke, nasogastric (NG) tubes are commonly inserted for feeding when dysphagia is identified, and percutaneous endoscopic gastrostomy (PEG) tubes are placed for severe or persistent dysphagia. However, little is known regarding predictors of PEG insertion. Methods: We used the Ontario stroke registry from 2003-2013 to identify baseline characteristics of all patients with NG or PEG tube insertion after stroke. We used multiple logistic regression with backwards selection to determine variables that were independent predictors of PEG tube insertion during admission. Results: 4002 patients with NG and 1903 patients with PEG were included in the analysis. Independent predictors of PEG were: Age (80+ vs. <60; odds ratio [OR] 1.70), past history of stroke (OR 1.17), higher stroke severity (severe vs. mild stroke; OR 1.37), stroke unit admission (OR 1.46), and dysphagia screening (OR 1.52). Factors associated with reduced odds of PEG insertion were: Prior history of peptic ulcer disease (OR 0.70), prior independence (OR 0.78), dementia (OR 0.76), palliative status (OR 0.49), and thrombolysis (OR 0.66). *All p<0.01 Conclusions: The strongest predictors of PEG were older age, higher stroke severity, stroke unit admission and dysphagia screening. Patients with dementia had reduced odds of PEG. Thrombolysis also reduced odds of PEG and may be protective.