Handling of the Non-Assessable Verbal Glasgow Coma Scale and Computed Illness Severity in Critically Ill Adults with Acute Brain Injury: A Retrospective Analysis
Background The Glasgow Coma Scale (GCS) contributes to illness-severity scores. The verbal component cannot be observed in intubated patients; the MIMIC Code Repository sets “No Response-ETT” rows to a total of 15. Objective To quantify non-assessable verbal GCS examinations and describe how their handling changes computed severity, the analyzed population and mortality-model performance. Materials and Methods Retrospective MIMIC-IV cohort (Boston, 2008 to 2022) of adults with acute brain injury (first intensive care unit stay per hospitalization); a “No Response-ETT” verbal entry was non-assessable. Five GCS derivation strategies were compared; four used the lowest eye-plus-motor examination of the first 24 h. The eICU Collaborative Research Database (eICU-CRD; United States, 2014 to 2015) supported replication. Results Of 14,230 stays, 45.2% had a non-assessable verbal examination; 47.5% of ventilated stays had none assessable in the first 24 h. A non-assessable lowest examination was associated with mechanical ventilation (adjusted odds ratio, 32.8; 95% CI, 27.8 to 38.6). The official first-day GCS was 15 in 44.7% of 14 215 stays; 94.1% of default-rule totals of 15 were incompatible with the eye-plus-motor sum charted at that examination. With one mortality-model specification refitted per GCS derivation on the same stays, the area under the receiver operating characteristic curve was 0.749 (95% CI, 0.739 to 0.760) with the official first-day GCS and 0.820 (0.812 to 0.828) with an eye-and-motor (Brennan) estimate. Complete-case analysis excluded 5509 of 14 192 stays and 72.8% of deaths. In eICU-CRD, stays with an unscored GCS component had higher mortality (16.6% vs 9.2%), but totals of 12 or more were no more frequent after the first documented ventilation entry (17.3%) than before (19.5%). Conclusion In this cohort, how a non-assessable verbal GCS was handled changed computed severity and the analyzed population. Studies and benchmarks that use GCS-based severity should report the handling rule.
Authors
- Oved Daniel (ORCID: https://orcid.org/0000-0002-8232-9138)
- Eyal Klang (ORCID: https://orcid.org/0000-0002-4567-3108)
- Mahmud Omar (ORCID: https://orcid.org/0009-0001-0438-0827)
- Yiftach Barash (ORCID: https://orcid.org/0000-0002-7242-1328)
- Alon Gorenshtein (ORCID: https://orcid.org/0009-0000-7542-8608)
- Yosef Adiniaev (ORCID: https://orcid.org/0009-0002-7678-633X)
Institutions
- Beth Israel Deaconess Medical Center (US)
- Harvard University (US)
- Tel Aviv Sourasky Medical Center (IL)
- Icahn School of Medicine at Mount Sinai (US)
Publication Details
- Journal
- Journal of Intensive Care Medicine
- Published
- 2026-11-14
- DOI
- https://doi.org/10.1177/08850666261508745
- Primary Topic
- Traumatic Brain Injury and Neurovascular Disturbances
- Type
- article
- Field-Weighted Citation Impact
- 0.00