New Model for Predicting Surgical Feeding Tube Placement in Patients With an Acute Stroke Event
Background and Purpose—The need for surgical feeding tube placement after acute stroke can be uncertain and associated with further morbidity.
Methods—Retrospective data were recorded and compared across patients with acute ischemic stroke and intracerebral hemorrhage. We identified all feeding tubes placed as percutaneous endoscopic gastrostomy (PEG) tubes. A prediction score for PEG tube placement was developed separately for patients with acute ischemic stroke and intracerebral hemorrhage using logistic regression models of variables known by 24 hours from admission.
Results—Of 407 patients included, 51 (12.5%) underwent PEG tube placement (25 acute ischemic stroke and 26 intracerebral hemorrhage). The odds of a patient with acute ischemic stroke with PEG score ≥3 of getting a PEG are greater than those with PEG score <3 (odds ratio, 15.68; 95% confidence interval, 4.55–54.01). The odds of a patient with intracerebral hemorrhage with PEG score ≥3 of getting a PEG are greater than those with PEG score <3 (odds ratio, 12.49; 95% confidence interval, 1.54–101.29).
Conclusions—The PEG score, comprised by variables known within the first day of admission, may be a powerful predictor of PEG placement in patients with acute stroke.
Dysphagia, or difficulty swallowing, can be identified in ≤50% of patients with acute stroke.1–3 The placement of a percutaneous endoscopic gastrostomy (PEG) tube is a feasible and common medical intervention with a low complication rate4 for patients who experience dysphagia after stroke to minimize complications.5 This study aims to develop a score to assist physicians in predicting the eventual placement of a PEG in patients who experience acute ischemic stroke (AIS) and intracerebral hemorrhage (ICH).
This is a retrospective analysis of prospectively collected data from patients admitted consecutively with acute stroke (AIS and ICH) to our academic stroke center from July 2008 to December 2010. Patients who received a PEG tube or other surgical feeding tube during hospital stay were identified from our stroke registry database. Patients with an in-hospital stroke, transferred from an outside facility, with an unknown time of last seen normal, reason for dysphagia other than stroke (eg, multiple sclerosis, advanced Alzheimer dementia, or previous head and neck surgery), and who presented 24 hours after last seen normal were excluded. Admission demographic and clinical data, as well as outcome measures, were extracted from patient records. Demographic and clinical data were recorded from within 24 hours of presentation and compared across patients with AIS and ICH using χ2 and t test with nonparametric tests, as well as odds ratios, when appropriate. A prediction score was developed separately for patients with AIS and ICH. More information on data collection, statistical analyses, and prediction model building can be found in the online-only Data Supplement.
A total of 734 patients were identified for the study, and 407 patients met inclusion criteria. Of these patients, 51 (12.5%) underwent PEG placement during hospital admission (Table 1). Of those undergoing PEG placement, there were 25 patients with AIS (7.6% of patients with AIS) and 26 patients with ICH (34.2% of patients with ICH; Table 2).
Risk Prediction Model: AIS
The AIS-PEG score was developed with 1 point being awarded for each of the following: age ≥80 years, 24-hour National Institutes of Health Stroke Scale (NIHSS) 8 to 14 (with an extra point for 24-hour NIHSS score >14), 1 point for black race, and 1 point for the infarct location involving cortex with a maximum of 5 points. Severity of dysphagia documented by speech therapy, dichotomized as nil per os versus liquid or solid foods permissible, was not predictive of PEG placement, and therefore not included in the model (P=0.252). We found that for the entire cohort, the odds of patients with PEG score ≥3 getting a PEG as an inpatient were 15× higher than those with PEG score <3 (odds ratio, 15.68; 95% confidence interval, 4.55–54.01), and a score of ≥3 points is 91.7% sensitive and 62.8% specific for undergoing PEG placement during hospitalization for patients with AIS with stroke (Figure).
Risk Prediction Model: ICH
The ICH-PEG score was developed with 1 point being awarded for each of the following: black race, 24-hour NIHSS score 8 to 14 (with an extra point for 24-hour NIHSS score >14), midline shift on initial head computed tomographic scan (>3 mm), and edema on follow-up head computed tomography with a maximum of 5 points. Again, severity of dysphagia did not reach statistical significance for predicting PEG placement (P=0.523). With this risk prediction model, the odds of patients with ICH with PEG score of ≥3 undergoing PEG placement as an inpatient were nearly 12× higher than those with PEG score <3 (odds ratio, 12.49; 95% confidence interval, 1.54–101.29), and a score of ≥3 points is 96.2% sensitive and 33.3% specific for having a PEG tube placed (Figure).
Severe strokes, baseline dysarthria, and impaired level of consciousness were associated with PEG tube placement, which is consistent in the literature.2,6,7 Despite this association, dysarthria and level of consciousness were not included in the models because of nonsignificance.
The AIS risk prediction model was found to be more prognostic of future PEG placement than the ICH risk model. The PEG score could be used to demonstrate need for surgical feeding tube insertion by the day after admission, potentially minimizing length of stay.
Our study is limited by its small sample size involving only 1 academic center, making our results difficult to generalize to larger populations, and by the retrospective nature of this study. We mitigated these limitations by extensively documenting admission physical examination data, allowing us to create an effective model to predict PEG placement early during hospitalization. The study is further limited by the method of classifying the severity of dysphagia, as it does not use a validated dysphagia scoring system.
In summary, our data present a new tool for estimating who will undergo PEG placement. With the expedited placement of a PEG, enteral nutrition can be initiated and the risks of malnutrition, prolonged length of stay, and mortality could be minimized. To appreciate the full implications of such a scoring system, we recommend validating this risk prediction model prospectively and with other stroke registries.
Sources of Funding
The project described was supported by Award Numbers 5 T32 HS013852-10 from The Agency for Healthcare Research and Quality (AHRQ), 3 P60 MD000502-08S1 from The National Institute on Minority Health and Health Disparities (NIMHD), National Institutes of Health (NIH), and by award 13PRE13830003 from the American Heart Association. The content is solely the responsibility of the authors and does not necessarily represent the official views of the AHRQ, AHA, or the NIH.
This work was supported by a grant from the Doris Duke Charitable Foundation to fund Clinical Research Fellow P.H. Dubin. The other authors report no conflict.
The online-only Data Supplement is available with this article at http://stroke.ahajournals.org/lookup/suppl/doi:10.1161/STROKEAHA.113.002402/-/DC1.
- Received June 10, 2013.
- Accepted July 15, 2013.
- © 2013 American Heart Association, Inc.
- Mann G,
- Hankey GJ,
- Cameron D
- Malmgren A,
- Hede GW,
- Karlström B,
- Cederholm T,
- Lundquist P,
- Wirén M,
- et al