By Kenneth P. Steinberg, MD, FACP, Editor
Professor of Medicine, University of Washington School of Medicine, Seattle, WA
Dr. Steinberg reports no financial relationships in this field of study.
SOURCE: Theisen-Toupal J, et al. Diagnostic yield of head computed tomography for the hospitalized medical patient with delirium. J Hosp Med 2014; 9:497-501.
This study was a retrospective review of medical records of hospitalized general medicine patients with head CT imaging performed for the evaluation of delirium. It was conducted at a large academic medical center in Boston, MA. All patients admitted to the medical services (general medicine, nephrology, hepatology, cardiology, or oncology) with a head CT done over a period of nearly two years were included. Patients in the ICU were excluded. Patients were included if a chart review revealed the indication for the head CT to be delirium, altered mental status, confusion, encephalopathy, somnolence, or unresponsiveness. Scans were excluded if there was a documented fall or trauma, or a new focal neurological deficit, or an admitting diagnosis of an intracranial lesion (e.g., stroke or subdural hematoma). A positive head CT was defined as an intracranial process that could explain delirium. An equivocal head CT was defined as the presence of a finding of unclear significance in relation to the delirium.
There were 1,714 head CT studies done on hospitalized medical patients during the study period. 398 were done for an indication of delirium, altered mental status, confusion, encephalopathy, somnolence, or unresponsiveness. 178 studies were excluded and thus there were 220 scans on 210 patients included in the study. Of these 220 scans, only 6 (2.7%) were positive and 4 (1.8%) were equivocal. Thus, less than 5% of the scans revealed findings that might have explained the new-onset delirium. However, all 10 of the positive or equivocal head CT scans resulted in a change in management. The 4 patients with equivocal scans all had a repeat scan, none of which identified a cause of delirium. The 6 patients with positive scans had a change in management ranging from a higher platelet transfusion threshold, reversal of anticoagulation, repeat advanced head imaging, neurosurgical consultation, and a change in goals of care. None of the patients underwent neurosurgical intervention. Because of the small number of positive scans, no risk factor associations could be made from this study.
This study demonstrated that the diagnostic yield of head CT scans in hospitalized medical patients with new onset delirium is very low and may be unnecessary. It is important to remember that these patients did not have a history of a recent fall or trauma and did not have a new focal neurological deficit. While in my opinion this is a good study, it does have some weaknesses. It is a retrospective chart review study and is prone to the errors inherent in relying on the medical record. There may have been clinically relevant information not completely captured in the chart. Also, the study took place a single, large, academic institution and the results might not be relevant to other clinical settings.
The authors point out that since not all hospitalized patients with delirium get a head CT, the true rate of positive findings on head CT imaging is likely to be even lower than they observed. If that’s true, it strengthens the argument that head CT scans in this population are over-utilized. However, in the very small number of patients with positive or equivocal findings (<5%), there were potentially important changes in the clinical management of those few patients. One of the real challenges in today’s clinical environment is to practice high-value, cost-conscious care. In this patient population, clinicians need to know that the diagnostic yield of a head CT is very low. They need to balance that low diagnostic yield and the potential harms thereof (unnecessary cost, resource utilization, false positive results, and radiation exposure) with the risk of a delayed or missed diagnosis of an intracranial process that might require a change in management. The authors point out that the routine use of head CT in this population seems unwarranted, yet in some high-risk patients, the test may well be justified. I agree with their conclusion. While minimizing the routine use of head CT scans for hospitalized medical patients with acute delirium, clinicians still need to use their assessment and judgment to ferret out the high-risk patients until further research illuminates a better approach.