The role of bedside ultrasonography, urinalysis, and computed tomography in the diagnostic evaluation of flank pain

Abstract

Study objectives: Numerous patients can present to the emergency department (ED) with flank pain. The diagnostic evaluation of these patients is varied but commonly includes urinalysis and an imaging modality. In this study, we evaluate the safety and utility of emergency ultrasonography for the evaluation of renal colic in patients with flank pain. Methods: This was a consecutive, prospective enrollment of afebrile patients with flank pain presenting to an ED of a tertiary care center. Exclusion criteria included trauma, recent urologic surgery, known current kidney stone, or unstable vital signs. All patients underwent identical evaluations, including urinalysis, bedside emergency ultrasonography, computed tomography (CT) scan, and a follow-up call. Ultrasonographic scans were performed by experienced emergency ultrasonographers who were blinded to the CT scan results. A percentage of the CT scans (n=44) were initially read by a radiology resident and later reread by a radiology attending physician. Final diagnosis was determined at 1-month follow-up after review of all patient records and a patient interview. Sensitivity and specificity of the ultrasonographic and initial CT scans were determined using final CT reading. To evaluate a hypothetical decision tree for the evaluation of flank pain in the ED, data were also analyzed post hoc using a set of predefined assumptions: (1) all patients undergo urinalysis; (2) ultrasonography is performed on patients with hematuria; (3) patients with hematuria and unilateral hydronephrosis (by ultrasonography) are diagnosed with renal colic and discharged; and (4) CT scans were performed on all patients with nondiagnostic urinalysis and ultrasonography. The accuracy of this decision tree was evaluated using the final diagnoses as determined by analysis of data and telephone interview at 1-month follow-up. Results: A total of 82 patients were enrolled from December 2001 to December 2003. The average age was 38.8 years, and 59% were female patients. The overall sensitivity and specificity of the emergency ultrasonography (87% and 78%, respectively) closely matched that of the initial resident-read CT scan (84% and 89%, respectively). An analysis of the flank pain decision tree shows that the combination of a positive urinalysis and ultrasonography (n=39) correctly diagnosed 97% of the patients in the study, whereas CT scan (n=43) diagnosed the remainder. Final diagnoses included renal colic (62%), urinary tract infection (6%), diverticulitis (3.6%), pneumonia (2.4%), flank pain (15%), and other (11%). Conclusion: Patients with multiple diagnoses can present with flank pain. Emergency ultrasonography of the kidney has a sensitivity and specificity similar to that of CT. A flank pain decision tree using limited bedside ultrasonography would result in a reduction of 52% of the CT scans, with no loss of diagnostic sensitivity or specificity.