Introduction

Infectious complications of acute pancreatitis include bacteremia and infected (peri-) pancreatic necrosis (IPN). These are thought to originate from the gastro-intestinal tract due to inflammation and increased intestinal permeability. Gram-negative rod-shaped bacteria (GNRB) and Enterococcus species are most frequently cultured in IPN, and recent evidence shows that the latter is associated with increased mortality.¹ We hypothesized that these infections are preceded by an increase of abundance of pathogenic bacteria in the gastrointestinal microbiota.

Aims & Methods

Our aim was to early identify patients at risk for IPN through microbiota surveillance of the upper and lower gastrointestinal tract during the first 30 days of a first episode of acute pancreatitis. We prospectively included patients from 20 hospitals collaborating in the Dutch Pancreatitis Study Group. Saliva and rectal swabs were collected twice weekly until IPN or discharge, and the microbiota profiled using 16S rDNA sequencing. For the current study, as predefined, we first analyzed the subgroup of patients who developed (peri-) pancreatic necrosis. Further analyses including the entire cohort and focuses on other clinical endpoints will follow. Community State Type (CST) analysis was used for microbiota clustering without the input of clinical outcomes. Univariate time-to-event (survival) analyses with serial, time-dependent predictors were used for the prediction of IPN.

Results

Overall, 319 patients were included of whom 57 (17.8%) developed necrotizing pancreatitis and formed the study population of the current analysis. In this group the rate of IPN was 36.8% (21/57). CSTs associated with IPN showed at Genus-level increased abundance of Enterococcus (rectal) and Staphylococcus (saliva), but not GNRB. Clinical factors such as age, gender, biliary etiology and bacteremia, as well as alpha diversity (Inverse Simpsons Index) were not associated with IPN (see table). Enterococcus colonization in rectal swabs and Staphylococcus in saliva at optimal cut-offs were predictive for IPN (respectively, HR 4.23, 95% CI 1.61–11.10, p=0.004; HR 10.80, 95% CI 2.13-55.09, p=0.004), which occurred for the first time at median 10 (range 4 – 17) days after hospital admission and median 23 days (range 2 - 96) before IPN. Antibiotic use (>24 hours within the week before sample collection) was not associated with Enterococcus or Staphylococcus colonization. Two of five patients with Enterococcus colonization had Enterococcus species in the first pancreatic culture, as compared to one in seven patients without colonization. Initial colonization of these taxa in samples taken at the time of hospital admission did not differ, highlighting the importance of longitudinal sampling in these patients.

Conclusion

Colonization of the gut microbiota early in the course of necrotizing pancreatitis with Enterococcus or Staphylococcus was a predictive factor for IPN. The time interval between colonization and IPN appeared long enough to allow for prophylaxis. Future clinical trials should focus on prophylactic, targeted antibiotics guided by microbiota surveillance.

References

1. Timmerhuis HC, van den Berg FF, Noorda PC, et al. Over- and Misuse of Antibiotics and the Clinical Consequence in Necrotizing Pancreatitis: An Observational Multicenter Study. Ann Surg 2023; Publish Ahead of Print:10.1097/SLA.0000000000005790.

Disclosure

Timmerhuis HC, van den Berg FF, Noorda PC, et al. Over- and Misuse of Antibiotics and the Clinical Consequence in Necrotizing Pancreatitis: An Observational Multicenter Study. Ann Surg 2023; Publish Ahead of Print:10.1097/SLA.0000000000005790.