Impact of sampling peritoneal fluid during emergency laparotomy on antimicrobial prescriptions in the ICU

Abstract

Editor—In line with recommendations from the National Emergency Laparotomy Audit, the ICU in Bath, UK attempts to admit all patients after emergency laparotomy.1Royal College of AnaesthetistsThe third patient report of the national emergency laparotomy audit. NELA Project Team, London2017Google Scholar A key tenet of ICU management after emergency laparotomy is administration of empirical antimicrobials where appropriate, with subsequent prescribing adjustment on the basis of microbiological growth patterns.2Rhodes A. Evans L.E. Alhazzani W. et al.Surviving Sepsis campaign: international guidelines for management of sepsis and septic shock.Intens Care Med. 2016; 43: 304-377Crossref Scopus (3345) Google Scholar Our hospital's empirical antibiotics for abdominal sepsis are amoxicillin, metronidazole, and gentamicin. A 3 month baseline audit by our department has shown that only 31% of patients undergoing emergency laparotomy had an intraoperative fluid sample from the peritoneal cavity sent for microbiological analysis. Subsequent antimicrobial adjustment was therefore guided by positive blood cultures or failure to thrive on empirical antibiotics after source control. Over an 18 month period (from December 2018 to May 2020), we instigated a quality improvement project in collaboration with our general surgeons, aiming to raise the frequency with which peritoneal fluid samples were sent for microbiological analysis during initial emergency laparotomy. During this time, 120 emergency laparotomies, where peritoneal fluid was noted, were performed on 110 patients who were subsequently admitted to the ICU. A sample for microbiological analysis was sent from the peritoneal cavity during 76 operations (63%); 68 of these samples were sent at initial laparotomy and eight were sent during a relook laparotomy. The details of this project were entered in the NHS Health Research Authority online decision tool, and NHS and ethical approval was not required. Positive peritoneal fluid cultures were obtained from 53 patients (70%); no organisms were grown from 22 patients (29%), and one sample (1%) was rejected on technical grounds. From the 53 patients where organisms were grown, 36 (68%) showed antibiotic resistance patterns: 22 (42%) of samples had organisms resistant to amoxicillin/co-amoxiclav, three (6%) had organisms resistant to gentamicin, and four (7%) had resistance to more than one of these antibiotics. Six samples (11%) grew Candida, one of which was resistant to voriconazole. Of the samples that yielded positive growth, a total of 31 (58%) therefore yielded results that would influence a change in prescribing from the empirical antibiotics of the hospital, either because of resistance patterns or because of unexpected discovery of Candida. For patients in whom a sample was sent for analysis, a prescription change would therefore be expected in 40%. The results are summarised in Figure 1. Five patients had positive blood cultures; two of these (3%) grew one/some, but not all of the organisms identified in the peritoneal fluid each with differing resistance patterns; one (1%) grew an organism not identified in peritoneal fluid; and two (3%) had blood cultures felt to be unrelated to abdominal sepsis. We have shown that sampling peritoneal fluid during emergency laparotomy yields results that would alter 40% of antibiotic prescriptions in these patients postoperatively in the ICU. We also found blood cultures to be far less helpful in supporting antibiotic decision-making. Further work is required to understand how any alterations in prescribing may affect patient outcome, but in the absence of this data it seems reasonable to advocate sampling peritoneal fluid during emergency laparotomy and incorporating the culture results in the antibiotic decision-making that occurs subsequently in the ICU. The authors delcare that they have no conflicts of interest.