The Intracerebral Hemorrhage Grading Scale is a Promising Predictor of Subsequent Mechanical Ventilation in Initially Non-intubated Intracerebral Hemorrhage Patients
Alan Spicer1, Crystal Dixon2, Marla Hairston1, Emily Coughlin1, William Burgin1
1USF Health Morsani College of Medicine, 2University of Miami Boca
Objective:

To predict subsequent ventilation in initially non-intubated intracerebral hemorrhage (ICH) patients.

Background:

ICH often requires mechanical ventilation (MV), increasing patient risks and hospital costs. A predictive tool could reduce costs and complications from unnecessary MV and reduce risks for patients requiring non-elective intubation. We evaluated the accuracy of seven ICH scales for their ability to predict subsequent MV.

Design/Methods:

Spontaneous ICH patients were retrospectively evaluated at a comprehensive stroke center, focusing on those not intubated upon arrival or in the Emergency Department. Chart review included demographics, Glasgow Coma Scale (GCS), ICH volume, intraparenchymal and intraventricular hemorrhage characteristics, and subsequent MV rates. Predictive scales were original ICH score (oICH), GCS, modified ICH score (mICH), mICH-A, mICH-B, Intracerebral Hemorrhage Grading Scale (ICH-GS), and Essen ICH score (Essen). To assess the ability of these scales to predict subsequent MV within two weeks, we used the area under the receiver operating characteristic curve (AUC), which combines sensitivity and specificity. Higher values indicate better test performance.

Results:

Between 2011-2014, 700 consecutive ICH patients were evaluated, with 221 meeting analysis criteria. Of these, 19/221 (8.6%) received subsequent MV within two weeks. The analysis of oICH and GCS was conducted with a sample of 231 patients with 20/231 (8.66%) receiving MV within two weeks. oICH, mICH, mICH-A, mICH-B, and Essen failed to achieve statistical significance of p<0.05. GCS and ICH-GS were statistically significant for predicting MV, albeit with relatively low AUC for GCS (AUC=0.643, p=0.040 [95% CI: 0.578-0.705]) and ICH-GS (AUC=0.675, p=0.012 [95% CI: 0.547-0.803]). ICH-GS significance was further corroborated by a Mann-Whitney U Test (p=0.008).

Conclusions:

The ICH-GS shows promise as a predictor of subsequent MV of ICH patients. Though not within an acceptable AUC range (>0.7), the ICH-GS performed better than the other scales evaluated. Development of an intubation-specific tool and evaluation of a larger sample size are needed.