Authors: Galovic M, Stauber AJ, Leisi N, et al. Development and Validation of a Prognostic Model of Swallowing Recovery and Enteral Tube Feeding After Ischemic Stroke. JAMA Neurol. Published online February 11, 2019. doi:10.1001/jamaneurol.2018.4858
Reviewed by: Lara L. Zimmermann, MD, Co-Director, Neurocritical Care Service, Assistant Professor of Neurological Surgery and Neurology, UC Davis School of Medicine
Read the article.*
*You will need to log in to US National Library of Medicine National Institutes of Health to read this article.
Stroke affects 795,000 people in the United States and 3 to 6 million individuals worldwide annually. Approximately 50% of patients suffer swallowing difficulties, or dysphagia, following their stroke. Dysphagia screening and early initiation of enteral nutrition is recommended.
For patients with dysphagia, nasogastric feeding is recommended if oral intake is anticipated to recover within seven days, whereas percutaneous gastrostomy tube placement is recommended if safe swallow is not anticipated to recover within 2-4 weeks. However, predicting swallow recovery is challenging.
In this project, authors developed (2011 – 2014) and prospectively validated (2015 – 2018) a prognostic model of swallowing recovery after stroke, named the Predictive Swallow Score (PRESS).
Following IRB approval, authors first performed a single-center prospective observational study of adults with acute ischemic stroke and severe dysphagia, which they defined as a Functional Oral Intake Scale (FOIS) score < 5 (oral diet of single consistency or worse), which suggests insufficient fluid intake, 40% reduction in energy and protein intake and need for enteral nutrition if not improved at ~ seven days. The primary end point was persistence of severe impairment at one week, discharge and 30 days. The secondary endpoint was return to pre-stroke diet.
Patients with subacute stroke (> 48 hours), late-swallow assessment (> 4 days), prior stroke or pre-existing dysphagia were excluded. Factors identified through literature search and in univariable analysis as predictors of swallowing recovery were assigned a point value and then an integer-based prediction rule was developed with scores ranging 0-10, named PRESS. Next, the prediction model was tested in a prospective multicenter validation cohort, where swallowing evaluations were performed at seven days and 30 days after stroke.
In total, 279 patients with acute ischemic stroke were included in the study, 153 in the single center-derivation cohort and 126 in the multi-center validation cohort. Median age was 77 years [interquartile range 67-84], and 47% were women. Individuals had little pre-stroke disability (baseline mRS 0-1; IQR 0-1) yet significant disability on admission (median NIHSS = 13; IQR 7-18) with severe impairment in oral intake due to dysphagia (Median FOIS 2; IQR 1-4). In the validation cohort, follow-up occurred in 120 (95%) patients at 7 days and 107 (85%) of patients at 30 days. The median time to recovery of oral intake was 12 days.
Overall, 64% of participants failed to recover functional swallow at 7 days and 30% failed at 30 days. At seven days, 64% of individuals had persistent swallowing impairment meeting guideline criteria for a NGT, and 94% had failed to return to their pre-stroke diet. At 30 days, 30% had ongoing impaired oral intake meeting guideline criteria for a PEG, and 66% failed to return to their pre-stroke diet. Longer duration of dysphagia was independently associated with poor outcomes, including aspiration pneumonia (p< .001), increased mortality (p=.01), and higher rates of dependency at discharge (p< .001) and institutionalization (p=.004).
In the single-center cohort, five variables were identified as independent predictors of swallow recovery and assigned points based on weight to create a score ranging from 0 – 10 points that estimated likelihood of swallow recovery. These predictors were age, NIHSS on admission, lesion of the frontal operculum, initial risk of aspiration and initial impairment in oral intake (measured with FOIS).
Based on these predictors, PRESS predicts the risk of persistent impairment in oral intake to range from 5% (PRESS = 0) to 96% (PRESS = 10) at day 7 and from 2% (PRESS = 0) to 62% (PRESS = 10) at day 30 following a stroke. The PRESS model performance in the validation cohort showed high agreement between predicted and observed outcomes with a discrimination statistic (C statistic) of 0.84 (95% CI, 0.76-0.91; p< .001) for predicting safe swallow recovery on day 7 and 0.77 (95% CI, 0.67-0.87; p< .001) on day 30. A smart phone app was also created.
In this project, authors both developed and prospectively validated a prognostic model of swallowing recovery after acute ischemic stroke, named the Predictive Swallow Score (PRESS).
They demonstrate that a straightforward algorithm can predict swallow recovery with good accuracy. This model includes just five prognostic factors (age, NIHSS score on admission, stroke location, initial risk of aspiration and initial impairment of oral intake) that can be determined at the bedside and predicts with good accuracy the likelihood of recovering safe swallow and adequate oral intake at seven days and 30 days following an ischemic stroke.
The strength of this model is ease of use at the bedside and good predictive accuracy. Limitations include that it is only validated for acute ischemic stroke. The PRESS model is a positive step toward personalized medicine for patients with acute stroke. Future studies are needed to evaluate if implementation of the PRESS tool reduces unnecessary PEG or G-tube placements and associated complications, reduces the incidence of aspiration, influences hospital length of stay or improves patient satisfaction measures.