Can lab data be used to reduce abdominal computed tomography (CT) usage in young adults presenting to the emergency department with nontraumatic abdominal pain?

Abstract

We sought to determine whether laboratory parameters could be found, predictive of a negative abdominal CT scan in young adults with nontraumatic abdominal pain. Following institutional review board approval, we evaluated CT reports of 522 patients, aged 21-35 years old, who presented to the Emergency Department with nontraumatic abdominal pain. Bivariate analyses relating ten laboratory parameters to whether the CT detected a cause for abdominal pain were conducted. A multivariate logistic regression model was then derived, with all variables in the final model significant at p < 0.05. Variables were dichotomized to yield odds ratios and 95% confidence intervals. Of the 522 patients meeting inclusion criteria, 45% had a cause for pain demonstrated by CT. Predictors of a negative CT in men were normal hematocrit and negative urine blood (p = 0.045, p = 0.016, respectively), and in women normal hematocrit, granulocyte percent, and alkaline phosphatase (p = 0.023, p = 0.039, p < 0.0001, respectively). When standard normal values were used to calculate descriptive statistics, only granulocyte percent in women had a significant confidence interval (odds ratio 2.5, confidence interval 1.6-4.0). Among the 208 women with normal granulocyte percent, the final clinical diagnosis was appendicitis, cholecystitis, and diverticulitis, in three, three, and two cases, respectively (4% combined). In summary, no laboratory test was sufficient to offer reassurance that a CT is not necessary in a young adult patient with nontraumatic abdominal pain. Alternative strategies should be considered to decrease the use of CT, and its associated radiation exposure, in young adults with nontraumatic abdominal pain.