Evaluation of the swallowing ability by video nasal endoscopic examination of the swallow (VEES) and videofluorographic swallow study (VFSS) is recommended to safely resume oral intake after fasting patients with stroke. However, expert staff and special equipment are needed to perform these examinations. Furthermore, these are difficult to perform in patients who are not cooperative during examinations due to delirium or cognitive impairment. We sought to investigate the predictive factors of oral intake and create a model to estimate the swallowing function of patients with acute stroke.

We included 93 patients with first ever stroke, who underwent VEES or VFSS at our hospital between October 1, 2015 and October 31, 2017. Patients’ functional oral intake scale (FOIS) was recorded at the time of the procedures. A multiple linear regression analysis was performed using FOIS as a dependent variable. Independent variables were age, FIM motor, FIM cognition, Onodera's prognostic nutritional index (O-PNI), speech intelligibility (SI), and Brunnstrom recovery stage for the arm (BRS-A), leg (BRS-L) and hand (BRS-H). The ethical committee of our hospital approved the study.

An estimation model of FOIS using three factors (FIM motor, O-PNI, speech intelligibility) was derived by multiple linear regression analysis. The formula used for estimating FOIS (Y_FOIS) was Y_FOIS=exp (0.098+0.308×FIM motor+0.304×O-PNI−0.575×SI)−1. The multiple correlation coefficient was 0.76, and the adjusted R-square was 0.56.

The ability to perform activities of daily living, overall severity of speech function, and nutritional status before disease onset are effective ways to estimate swallowing function in patients with acute stroke.