Background: Decreased surgical site infections (SSIs) and morbidity have been reported with mechanical and oral antibiotic bowel preparation (MOABP) compared with no bowel preparation (NBP) in colonic surgery. Several societies have recommended routine use of MOABP in patients undergoing colon resection on the basis of these data. Our aim was to investigate this recommendation in a prospective randomised context.
Methods: In this multicentre, parallel, single-blinded trial, patients undergoing colon resection were randomly assigned (1:1) to either MOABP or NBP in four hospitals in Finland, using a web-based randomisation technique. Randomly varying block sizes (four, six, and eight) were used for randomisation, and stratification was done according to centre. The recruiters, treating physicians, operating surgeons, data collectors, and analysts were masked to the allocated treatment. Key exclusion criteria were need for emergency surgery; bowel obstruction; colonoscopy planned during surgery; allergy to polyethylene glycol, neomycin, or metronidazole; and age younger than 18 years or older than 95 years. Study nurses opened numbered opaque envelopes containing the patient allocated group, and instructed the patients according to the allocation group to either prepare the bowel, or not prepare the bowel. Patients allocated to MOABP prepared their bowel by drinking 2 L of polyethylene glycol and 1 L of clear fluid before 6 pm on the day before surgery and took 2 g of neomycin orally at 7 pm and 2 g of metronidazole orally at 11 pm the day before surgery. The primary outcome was SSI within 30 days after surgery, analysed in the modified intention-to-treat population (all patients who were randomly allocated to and underwent elective colon resection with an anastomosis) along with safety analyses. The trial is registered with ClinicalTrials.gov, NCT02652637, and EudraCT, 2015-004559-38, and is closed to new participants.
Findings: Between March 17, 2016, and Aug 20, 2018, 738 patients were assessed for eligibility. Of the 417 patients who were randomised (209 to MOABP and 208 to NBP), 13 in the MOABP group and eight in the NBP were excluded before undergoing colonic resection; therefore, the modified intention-to-treat analysis included 396 patients (196 for MOABP and 200 for NBP). SSI was detected in 13 (7%) of 196 patients randomised to MOABP, and in 21 (11%) of 200 patients randomised to NBP (odds ratio 1·65, 95% CI 0·80-3·40; p=0·17). Anastomotic dehiscence was reported in 7 (4%) of 196 patients in the MOABP group and in 8 (4%) of 200 in the NBP group, and reoperations were necessary in 16 (8%) of 196 compared with 13 (7%) of 200 patients. Two patients died in the NBP group and none in the MOABP group within 30 days.
Interpretation: MOABP does not reduce SSIs or the overall morbidity of colon surgery compared with NBP. We therefore propose that the current recommendations of using MOABP for colectomies to reduce SSIs or morbidity should be reconsidered.
Funding: Vatsatautien Tutkimussäätiö Foundation, Mary and Georg Ehrnrooth's Foundation, and Helsinki University Hospital research funds.
Copyright © 2019 Elsevier Ltd. All rights reserved.