RESUMO
BACKGROUND: Acute abdominal pain accounts for about 10% of emergency department visits and has progressively become the primary indication for CT scanning in most centers. The goal of our study is to identify biological or clinical variables able to predict or rule out significant pathology (conditions requiring urgent medical or surgical treatment) on abdominal CT in patients presenting to an emergency department with acute abdominal pain. METHODS: This was a retrospective cohort study performed in the emergency department of an academic center with an annual census of 60'000 patients. One hundred and-nine consecutive patients presenting with an acute non-traumatic abdominal pain, not suspected of appendicitis or renal colic, during the first semester of 2013, who underwent an abdominal CT were included. Two medical students, completing their last year of medical school, extracted the data from patients' electronic health record. Ambiguities in the formulations of clinical symptoms and signs in the patients' records were solved by consulting a board certified emergency physician. Nine clinical and biological variables were extracted: shock index, peritonism, abnormal bowel sounds, fever (> 38 °C), intensity and duration of the pain, leukocytosis (white blood cell count >11G/L), relative lymphopenia (< 15% of total leukocytes), and C-reactive Protein (CRP). These variables were compared to the CT results (reference standard) to determine their ability to predict a significant pathology. RESULTS: Significant pathology was detected on CT in 71 (65%) patients. Only leukocytosis (odds ratio 3.3, p = 0.008) and relative lymphopenia (odds ratio 3.8, p = 0.002) were associated with significant pathology on CT. The joint presence of these two anomalies was strongly associated with significant pathology on CT (odds ratio 8.2, p = 0.033). Leukocytosis with relative lymphopenia had a specificity of 89% (33/37) and sensitivity of 48% (33/69) for the detection of significant pathology on CT. CONCLUSION: The high specificity of the association between leukocytosis and relative lymphopenia amongst the study population suggests that these parameters would be sufficient to justify an emergency CT. However, none of the parameters could be used to rule out a significant pathology.