S0866 Stool Molecular Testing and Antibiotic Use for Acute Diarrheal Illness in Immunosuppressed Patients

Abstract

INTRODUCTION: Conventional stool culture has low diagnostic yield in the evaluation of acute diarrheal illness (ADI). Thus, the decision to prescribe antibiotics for ADI is usually guided by a patient's medical history, severity of presenting symptoms, and immune-competence. PCR-based multiplex stool antigen panels (PCR panel) now allow for rapid and specific identification of causative pathogens and can help guide management of ADI, but its role in antibiotic use is unknown. This study compares rates of antibiotics prescribed for PCR panel-positive ADI in immunosuppressed and non-immunosuppressed patients. METHODS: We identified patients who presented from April 2018 to April 2019 with PCR panel-positive ADI. Those with C. difficile co-infection were excluded as these results were not available to providers. Patients were categorized as immunosuppressed or non-immunosuppressed, with immunosuppressed defined by active malignancy, history of organ transplant, primary immunodeficiency, HIV, or on immunosuppressive medications. Charts were reviewed for baseline characteristics and antibiotic management of the ADI. Baseline characteristics were compared using Student t-test, Wilcoxon rank-sum, Chi-squared, or Fishers exact test. Multinomial logistic regression was used to compare likelihood of antibiotic use. RESULTS: 749 patients presented with PCR panel-positive ADI, of whom 207 (27.6%) were immunosuppressed. A greater percentage of the immunosuppressed had received steroids (45.1% vs 4.1%) or antibiotics (47.3% vs 18.7%) in the 6 months prior to PCR testing. There were no clinically significant differences in vital signs or labs on presentation between the groups. The immunosuppressed had a higher odds of receiving antibiotics in response to a positive PCR panel vs no antibiotics (OR 1.85, P = 0.0016) than the non-immunosuppressed. This remained true after adjusting for potential confounding factors including age, sex, and recent steroid or antibiotic use (AOR 2.30, P = 0.0033). There was no significant difference in the odds of receiving empiric antibiotics prior to PCR results vs no antibiotics (OR 1.26, P = 0.2606) between the groups. CONCLUSION: We show that immunosuppressed patients are more likely to be prescribed antibiotics in response to a positive PCR panel than non-immunosuppressed patients, but there is no difference in the likelihood of being treated empirically prior to PCR results. This shows that PCR panels may be guiding management of ADI, especially in immunosuppressed patients.Table 1.: Comparison of baseline and clinical characteristics by immunosuppressed statusTable 2.: Antibiotic Management of ADITable 3.: Multinomial logistic regression models of treatment outcome