Background: Using standard techniques, only 10% of inflammatory bowel disease (IBD) patients with symptoms have detectable gastrointestinal (GI) tract infections. Detection is limited by the sensitivity of these tests. Clostridium difficile infection can worsen the course of IBD, but it is not known if other infections can also. Our primary aim was to determine the risk of GI infections in IBD patients during symptomatic flares using the BioFire FilmArray GI Panel test. This multiplex PCR test detects 22 organisms, including bacteria, viruses, and protozoa. Our secondary aim was to compare the IBD course between the patients with positive and negative tests for infection.
Methods: We reviewed BioFire GI stool results for all adult IBD patients from April to October 2016 at our center. We excluded patients on antibiotics in the last 2 weeks, UC patients status post colectomy, and patients with symptoms from other causes, e.g. pancreatitis or infections outside GI tract. We compared the risk of GI infection between those with active and inactive IBD defined by inflammatory markers (using QuantumBlue fecal calprotectin), imaging findings, or endoscopic /biopsy findings. Among actively inflamed IBD patients, we compared clinical characteristics, medication use, and disease course (need for escalation of immunosuppressive agents and need for surgery) between those with positive and negative tests.
Results: 131 IBD patients (60 active CD, 14 inactive CD, 54 active UC, and 3 inactive UC) were included. Most (88.6%) were inpatients. Thirty-nine had positive results; common types of infection included Norovirus, C. difficile, and E. coli. The risk in the IBD group with active inflammation was significantly higher than the inactive IBD group [38/114 (33.3%) vs 1/17 (5.9%), p=0.02]. In the inflammation group, shorter duration of presenting symptoms and use of biologics were significant predictors of a positive test. The proportions of escalation of immunosuppressive therapy and surgery were significantly lower in the group with positive tests for infection vs. those with negative tests [escalation, 23/38 (60.5%) vs 68/76 (89.5%), p<0.01], and [surgery 4/38 (10.5%) vs 21/76 (28%), p=0.034], respectively.
Conclusions: A surprisingly large fraction, one-third of IBD patients with symptoms and objective evidence of gut inflammation, had infectious agents detected in their stool. The course of IBD in patients with these infections was more benign than those with negative tests. It appears that more sensitive testing for acute infections can be prognostically helpful, particularly in IBD patients with acute worsening of symptoms, and the use of steroids and escalation of IBD therapy could be reduced if stool infections are identified early.