Head CT for nontrauma patients in the emergency department: clinical predictors of abnormal findings.

PURPOSE

To identify predictors of clinically important abnormal findings in computed tomography (CT) images of the head among emergency department (ED) patients without a history of trauma.

MATERIALS AND METHODS

Approval was obtained from the institutional research ethics board, and informed consent from patients was not required. This study was a retrospective review of consecutive unenhanced head CT examinations in patients aged 18 years or older who did not have trauma or known intracranial pathologic processes in ED from January 2004 through June 2006. Multivariable logistic regression was used to identify predictors of clinically important abnormal CT findings in the derivation cohort (CT examinations from January 1, 2004, through August 15, 2005), and the reproducibility of findings in a validation cohort (all subsequent CT scans through to June 30, 2006) was assessed. The strength of association of each variable was expressed with clinically important abnormal CT findings as adjusted odds ratio (OR) and 95% confidence interval (CI).

RESULTS

Of 29 469 consecutive head CT images performed at a single institution between January 1, 2004, and June 30, 2006, 3967 were eligible for this study. Of the CT images in these patients, 548 (13.8%) revealed clinically important abnormalities. Six independent clinical predictors of important abnormal findings on head CT were identified: age (adjusted OR per 10-year increase: 1.17; 95% CI: 1.08, 1.28), focal neurologic deficit (adjusted OR: 5.39; 95% CI: 3.90, 7.47), altered mental status (adjusted OR: 2.32; 95% CI: 1.66, 3.25), history of malignancy (adjusted OR: 4.11; 95% CI: 2.28, 7.42), nausea and/or vomiting (adjusted OR: 2.22; 95% CI: 1.14, 4.33), and derangements in coagulation profile (adjusted OR: 1.91; 95% CI: 1.07, 3.41).

CONCLUSION

This study identified several potential clinical predictors of abnormal head CT findings in ED patients who did not sustain trauma. Prospective validation of a clinical prediction rule in this population is warranted.