Avoidance of antibiotic administration to Campylobacter enterocolitis mimicking severe salmonellosis by clinical and laboratory features.

Abstract

To compare the clinical and laboratory features of non-typhoid Salmonella (NTS) and Campylobacter jejuni enterocolitis in children and formulate a risk scoring system (with receiver-operating characteristic curve) to facilitate early decision making and avoid antibiotic overuse in C. jejuni enterocolitis. Between January 2008 and December 2011, children (age <18 years) diagnosed as having C. jejuni enterocolitis and NTS enterocolitis in Kaohsiung Chang Gung Memorial Hospital were retrospectively enrolled. Clinical features and laboratory data were collected for analysis and a risk calculation score is created for the identification of Campylobacter infections. A total of 309 cases of C. jejuni enterocolitis and 496 cases of NTS enterocolitis were enrolled. Compared with Salmonella group clinically, the Campylobacter group had older age (81.06 ± 50.65 vs. 32.70 ± 34.88 months, p <; 0.001), more abdominal pain (69.26% vs. 37.5%, p <; 0.001) and more watery diarrhea (79.94% vs. 20.77%, p <; 0.001). In laboratory data, the Campylobacter group had higher level of white blood cell count (11 208 ± 4380 vs. 9095 ± 3598 cell/mm3, p <; 0.001). Four criteria including age (≥5 years), leukocytosis (≥10 000 cell/mm3), abdominal pain and watery diarrhea were identified as good predictors of Campylobacter enterocolitis. When three criteria were fulfilled, Campylobacter enterocolitis was highly suspected and antibiotic could be withheld even when C-reactive protein is high and before stool culture results are known. When four criteria were fulfilled, antibiotic usage was absolutely unnecessary.