From the AGA Journals

Model could cut CT scans of patients with Crohn’s disease by 43%




A risk stratification model that determined whether patients with Crohn’s disease needed computed tomography cut scans of these patients in emergency departments by 43%, with a miss rate of only 0.8%, researchers reported online in Clinical Gastroenterology and Hepatology.

Computed tomography scans yield nonsignificant findings for almost one-third of patients with Crohn’s disease (CD) who present to emergency departments, said Dr. Shail Govani of the University of Michigan in Ann Arbor and his associates. By using their model to identify patients with serious gastrointestinal complications as opposed to straightforward intestinal inflammation, emergency departments could prevent more than 250 cancer cases and save more than $80 million per decade in the United States, the investigators added.

Source: American Gastroenterological Association

Patients with CD may be exposed to increasing cumulative radiation levels, and 30% of this exposure occurs in emergency departments, with 75% due to CT scans, the researchers said (Clin. Gastroenterol. Hepatol. 2014 [doi:10.1016/j.cgh.2014.02.036]).

For the study, the investigators retrospectively reviewed electronic medical records from the University of Michigan from 2000 through 2011, identifying 613 adults with CD who made 1,095 visits that included CT scans within 24 hours of presentation. The researchers then modeled associations between laboratory values and perforation, abscess, or other serious complications as opposed to intestinal inflammation.

Patients averaged 1.8 CT scans during the decade-long study period, and the overall rate of CT scans during that time rose from 63% to 87%, the investigators said. Only 16.8% of scans revealed a complication that would change clinical management, they reported.

Only C-reactive protein (CRP) and erythrocyte sedimentation rate (ESR) were significantly associated with complications (odds ratio for CRP, 1.10; 95% confidence interval, 1.05-1.15; P less than .001; odds ratio for ESR, 1.02; 95% CI, 1.01-1.03; P less than .001), the researchers said. Adding ESR to 5 x CRP and not scanning patients with a resulting value of 10 or less would avoid CT scans in 18.5% of patients, they reported. But by using the more complex logistic regression model instead of the simpler equation, scans could be avoided for 43.0% of patients, with a miss rate of 0.8%, they said.

Based on the study, patients assessed as likely to have complications should undergo a standard CT scan of the abdomen and pelvis with nonbarium contrast to avoid barium peritonitis, said Dr. Govani and associates. Other patients should forego CT scans, have a consult with a gastroenterologist prior to further imaging, or undergo lower-radiation CT enterography, depending on presenting signs and probability of inflammation, they added.

The researchers said they were unable to construct good models that included obstruction as an outcome. Patients with suspected obstructions should have abdominal x-rays and then CT if an obstruction remained likely, they said.

"These models are limited in that they are retrospective and represent data from one center," the investigators added. "Although our internal validation with 10-fold cross-validation shows that these models have good performance characteristics, further external validation studies are needed to determine whether these models are generalizable to CD patients elsewhere."

The authors are prospectively testing the algorithms and hope to continue to validate and study them in emergency departments, they said.

The Inflammatory Bowel Disease Working Group, the Department of Veterans Affairs, and UCB supported the research. The authors reported having no conflicts of interest.

Next Article: