Clinical Scoring and Ultrasound-Based Diaphragm Assessment in Predicting Extubation Failure in Neurointensive Care Unit: A Single-Center Observational Study

Patients in the neurointensive care unit (NICU) fail extubation despite successful weaning from mechanical ventilation. Parameters currently used in the general intensive care unit do not accurately predict extubation success in the NICU. In this study, peak cough expiratory flow rate, ultrasound-based diaphragm function assessment, and comprehensive clinical scoring systems were measured to determine whether these new variables, in isolation or combination, could predict extubation failure successfully in the NICU.

All adult patients extubated after 48 h of mechanical ventilation in the NICU of a single tertiary care center were recruited into the prospective cohort. The patient’s cough peak expiratory flow rate (C-PEFR), diaphragm function, and clinical scores were measured before extubation. C-PEFR was measured using a hand-held spirometer, diaphragm function (excursion, thickness fraction, and diaphragm contraction velocity on coughing) was assessed using ultrasound, and the clinical scores included the visual pursuit, swallowing, age, Glasgow Coma Scale for extubation (VISAGE) and respiratory insufficiency scale–intubated (RIS-i) scores. The patients requiring reintubation within 48 h were considered as extubation failure. Univariate and multivariate logistic regression analyses were done to identify predictors of extubation failure.

Of the 193 patients screened, 43 were recruited, and 15 had extubation failure (20.9%). Patients with extubation failure had higher RIS-i scores (p < 0.001) and lower VISAGE scores (p = 0.043). The C-PEFR and diaphragm function (excursions and contraction velocity on coughing) were lower in patients with extubation failure but not statistically significant. The variables with p < 0.2 in univariate analysis (RIS-i, VISAGE, and diaphragm cough velocity) were subjected to multivariate regression analysis. RIS-I score remained an independent predictor (odds ratio 3.691, 95% confidence interval 1.5–8.67, p = 0.004). In a receiver operating characteristic analysis, the area under the curve for RIS-i was 0.963. An RIS-i score of 2 or more had 94% specificity and 89% sensitivity for predicting extubation failure.

The RIS-i score predicts extubation failure in NICU patients. The addition of ultrasound-based diaphragm measurements to the RIS-i score to improve prediction accuracy needs further study.