Clinical impression and ascites appearance do not rule out bacterial peritonitis

Previous research has demonstrated that physician clinical suspicion, determined without assessing fluid appearance, is not adequate to rule out spontaneous bacterial peritonitis (SBP) without fluid testing.

To determine the sensitivity of physician clinical suspicion, including a bedside assessment of fluid appearance, in the detection of SBP in Emergency Department (ED) patients undergoing paracentesis.

We conducted a prospective, observational study of ED patients with ascites undergoing paracentesis at three academic facilities. The enrolling physician recorded the clinical suspicion of SBP ("none," "low," "moderate," or "high"), and ascites appearance ("clear," "hazy," "cloudy," or "bloody"). SBP was defined as an absolute neutrophil count ≥ 250 cells/mm(3), or culture pathogen growth. We defined "clear" ascites fluid as negative for SBP, and "hazy," "cloudy," or "bloody" as positive. A physician clinical suspicion of "none" or "low" was considered negative for SBP, and an assessment of "moderate" or "high" was considered positive. The primary outcome measure was sensitivity of physician clinical impression and ascites appearance for SBP.

There were 348 cases enrolled, with SBP diagnosed in 43 (12%). Physician clinical suspicion had a sensitivity of 42% (95% confidence interval [CI] 29-55%) for the detection of SBP. Fluid appearance had a sensitivity of 72% (95% CI 58-83%).

Physician clinical impression, which included an assessment of fluid appearance, had poor sensitivity for the detection of SBP and cannot be used to exclude the diagnosis. Routine laboratory fluid analysis is indicated after ED paracentesis, even in patients considered to have a low degree of suspicion for SBP.