Accuracy of Microscopic Urine Analysis and Chest Radiography in Patients with Severe Sepsis and Septic Shock

Abstract

Background Diagnosis of source of infection in patients with septic shock and severe sepsis needs to be done rapidly and accurately to guide appropriate antibiotic therapy. Objective The purpose of this study is to evaluate the accuracy of two diagnostic studies used in the emergency department (ED) to guide diagnosis of source of infection in this patient population. Methods This was a retrospective review of ED patients admitted to an intensive care unit with the diagnosis of severe sepsis or septic shock over a 12-month period. We evaluated accuracy of initial microscopic urine analysis testing and chest radiography in the diagnosis of urinary tract infections and pneumonia, respectively. Results Of the 1400 patients admitted to intensive care units, 170 patients met criteria for severe sepsis and septic shock. There were a total of 47 patients diagnosed with urinary tract infection, and their initial microscopic urine analysis with counts > 10 white blood cells were 80% sensitive (95% confidence interval [CI] .66–.90) and 66% specific (95% CI .52–.77) for the positive final urine culture result. There were 85 patients with final diagnosis of pneumonia. The sensitivity and specificity of initial chest radiography were, respectively, 58% (95% CI .46–.68) and 91% (95% CI .81–.95) for the diagnosis of pneumonia. Conclusion In patients with severe sepsis and septic shock, the chest radiograph has low sensitivity of 58%, whereas urine analysis has a low specificity of 66%. Given the importance of appropriate antibiotic selection and optimal but not perfect test characteristics, this population may benefit from broad-spectrum antibiotics, rather than antibiotics tailored toward a particular source of infection.