A nomogram for the prediction of traumatic intracranial abnormalities in the elderly: Development and validation

Traumatic brain injury is a major public health issue and a leading cause of death and disability among the elderly. As a result, traumatic intracerebral abnormalities (tICA) are now frequently diagnosed using cranial computed tomography (cCT), which has led to an overuse of this investigation. The present study aimed to develop and validate the nomogram for predicting tICA in the elderly.

This was a retrospective cohort study. The study population consisted of patients who met the following inclusion criteria: (1) aged ≥60 years, as defined by the World Health Organization; (2) admitted to a tertiary referral hospital between January 2015 and December 2022; and (3) underwent cCT scanning upon admission. Exclusion criteria were (1) patients' images were unavailable or (2) patients died in the emergency department before hospitalization. The total data was divided into the development cohort (2015-2019, n=2052) to develop the nomogram and the validation cohort (2020-2022, n=310) to test the predictability as temporal validation. Binary logistic regression was used in the construction of the nomogram.

There were 2362 patients in the entire cohort. Positive CT scan rates among the elderly amount to 21.3%. Multivariable analysis identified the following factors significantly associated with traumatic intracranial hematoma in the elderly: use of aspirin (odds ratio (OR)=4.49, 95% confidence interval (CI) 3.08-6.54), warfarin usage (OR=6.88, 95% CI 2.99-15.8), traffic accidents (OR=1.98, 95% CI 1.46-2.68), Glasgow coma scale 9-12 (OR=3.24, 95% CI 1.66-6.35), amnesia (OR=1.65, 95% CI 1.17-2.33), nausea or vomiting (OR=12.5, 95% CI 5.56-28.07), seizure (OR=11.42, 95% CI 3.33-39.22), motor weakness (OR=65.45, 95% CI 18.68-229.29), and bleeding per nose or ear (OR=24.81, 95% CI 10.57-14.58). The predictability of the nomogram demonstrated sensitivity, specificity, and an area under the receiver operating characteristic curve of 0.833, 0.788, and 0.889, respectively.

According to the temporal validation, the nomogram had an outstanding performance for tICA prediction in the elderly. In the future, the prediction tool will assist clinicians in making judgments and balancing the rate of cCT in clinical practice.