Abstract
Background:
Gastric aspiration is of particular concern in critically ill patients, which is characterized as acute diffuse, inflammatory lung injury. This article aimed to develop and validate a multivariable gastric aspiration risk prediction rule in Neurosurgery ICU patients with artificial airways.
Methods:
We conducted a prospective cohort study collecting 270 patients with artificial airways in two Neurosurgery ICU of two university-affiliated hospital in Beijing from August 2016 to February 2018. Patients were excluded with admitted to ICU or extubated less than one day, without the consent of the surgeons, unwilling to participate in research or without informed consent and unable to take airway secretion specimens. Gastric aspiration be defined as the concentration of pepsin A in airway secretion specimens is not less than 25 ng/mL. Multivariable logistic regressions was used to explore the predictors of gastric aspiration. Ethical approval was obtained from the ethics committee of our university.
Results:
Overall, we included 148 patients to develop prediction rule and 122 patients to validate prediction rule. 50% of the patients appeared gastric aspiration in the development cohort and 44.3% in the validation cohort. A multivariable gastric aspiration risk prediction model based on logistic regression analysis showed that gastric residual volumes [GRVs]>40ml (OR 4.809; 95% CI 2.049-11.284), head-of-bed [HOB] elevation <30° (OR 3.134; 95% CI 1.148-8.553), repeated intubation (OR 3.344; 95% CI 1.394-8.021), Glasgow Coma Scale [GCS] score<9 (OR 5.054; 95% CI 2.005-13.223), cuff pressure<20cm H2O (OR 3.253; 95%CI 1.306-8.103) are independent risk factors. The area under the curve (AUC) is 0.832. Using bootstrapping for internal validation the AUC is 0.842. To facilitate the use of the model for clinical practice, the model was converted to a prediction rule with three risk categories: low-risk group (0-1 score), moderate-risk group (2-5 score), and high-risk group (6-7 score). After that, we evaluate the discrimination and calibration of prediction rule in new samples. The external validation showed that the AUC for discrimination was 0.710 and the calibration slope for calibration was 0.788.
Conclusions:
Gastric residual volumes, repeated intubation, head-of-bed elevation, GCS score, cuff pressure are independent predictive factors of gastric aspiration in Neurosurgery ICU patients with artificial airways.
Figure 1 The ROC of prediction rule in development cohort (A) and validation cohort (B)
Get full access to this article
View all access options for this article.