724 - Gastrointestinal (GI) PCR vs Stool Cultures: Impact on Length of Hospital Stay (LOS) and Antibiotic Use

Background: GI PCR can detect 22 pathogens (bacteria, parasites and viruses) from a single stool sample. Stool cultures are labor intensive and only target the most common diarrheal pathogens (such as Campylobacter, E. coli and a few parasites). We hypothesized that implementation of GI PCR would result in decreased LOS and lower antibiotic use.

Methods: This retrospective study utilized data from review of electronic medical records and included patients aged > 18 years old who were admitted with diarrhea over a 3-year period from 2016 to 2019. LOS and antibiotic use data was collected for patients who had GI PCR from 2017–2019 (GIP arm) and compared with data from patients who had stool cultures from 2016–2017 (SC arm). Differences were assessed using Chi-square or Fisher’s exact test for categorical variables and the Mann Whitney Rank Sum test for continuous variables.

Results: The analysis included a total of 338 patients, 225 (66.6%) in the GI PCR arm and 113 (33.4%) in the SC arm. A significantly higher proportion of patients in the GIP arm had a positive result compared with the SC arm (26.2% vs. 9.7%, P < .0001; Table 1). Table 2 shows the most frequently isolated organisms. Median LOS was 6 days (IQR: 4–13) for the GIP arm and 5 days (IQR: 3–7) for the SC arm (p=.060); 8 patients in the GIP arm had average LOS of 75 days due to comorbidities and disposition issues. However, within the GIP arm, median LOS was much shorter for patients detected with viruses by PCR vs. those with non-viral pathogens (3.5 days (IQR: 3-7) vs. 6 days (3-12)) There was no difference in antibiotic use between the GIP and SC arms (84.9% vs. 84.1%, P=.844). Patients in GIP arm were more commonly given Piperacillin-tazobactam and Carbapenems, whereas patients in the SC arm received metronidazole more often. Within the GIP arm, antibiotic use was lower among patients detected with viruses vs. those detected with non-viral pathogens (73.1% vs. 81.8%).Table 1
Table 2


Conclusion: LOS was longer in patients in GIP arm vs SC arm, which may have been influenced by the presence of outliers in the GIP arm. No differences in antibiotic use was observed between the two groups.  However, within the GIP arm, detection of viruses by GI PCR significantly shortened LOS and lowered antibiotic use.