We Dutch, we love gut decontamination. Not only in critically ill patients, but also in those undergoing elective colorectal surgery. A decontaminated gut is a safe gut, and that feeling was based on data from Dutch studies. A new study from Finland, published in Lancet, now questions whether our gut feeling was correct.
Colorectal surgery is associated with a considerable risk of post-operative infectious complications, ranging from superficial wound infections to intra-abdominal infections, often caused by anastomotic dehiscence (leakage), requiring subsequent surgical interventions, leading to a protracted disease course, or even death. And this can happen despite per-operative systemic antibiotic prophylaxis. Therefore, additional preventive measures have been proposed, such as;
Mechanical Bowel Preparation (MBP; rinsing the gut as done in preparation for CT- scanning);
Oral Antibiotic Prophylaxis (OAB; non-absorbable antibiotics targeting Gram- negative and/or anaerobic gut flora);
Both MBP and OAB, for simplicity called = MOABP
Just do nothing: No Bowel Preparation (NBP=no MBP, no OAB, no MOABP)
8 years ago we got the results of a single-centre RCT comparing MOABP (n=143) to MBP (n=146); so, all guts were flushed. The MOABP group had less postoperative complications: 32,9% versus 45,9% (relative risk (RR) 0,72 (95% CI 0,53-0,96)) and infectious complications (RR 0,64 (95% CI 0,42-0,96)). The incidence of anastomotic leakage 6,3% and 15,1% in the MOABP and MBP group, respectively (RR 0,42 (95% CI 0,20-0,88).
Recently followed by the SELECT trial, a multicentre RCT of the same interventions (MOABP (n=228) and MBP (n=227)) in patients undergoing elective colorectal cancer surgery (again, all guts flushed). Anastomotic dehiscence was the primary outcome, and was observed in 14 patients (6,1%) in MOABP and in 22 (9,7%) MBP patients (OR 0,61 ( 95% CI 0,30-1,22)). This was at the time of interim analysis and the study was prematurely ended as superiority was considered out of reach. Yet, the OR for infectious complications with MOABP was 0.48 (95% CI 0.30-0.76).
More or less at the same time the results of a single-centre before-after study, comparing NBP (n=352) to OAP (n=1,048) was published. So, in this study, no gut was flushed. Here the OR of OAP for deep SSI was 0.54 (95% 0.37-0.78) and 0.57 (95% CI 0.35-0.90) for anastomotic leakage.
With all that accumulating evidence from Dutch studies it became impossible to successfully complete another multi-centre study comparing OAP to NBP, as many hospitals decided to implement MOABP as standard of care.
The new Lancet study is a multicentre, randomised, parallel, single-blinded trial comparing MOABP to NBP in patients undergoing elective colectomy. 396 patients were included in the modified intention-to-treat analysis (196 for MOABP and 200 for NBP). Surgical Site Infection was detected in 13 (7%) MOABP and 21 (11%) NBP patients; anastomotic dehiscence was in 7 (4%) MOABP and 8 (4%) NBP patients and reoperations were necessary in 16 (8%) MOABP and 13 (7%) NBP patients. The odds ratios for these outcome for patients pre-treated with antibiotics were 0.63 (95% CI 0.31-1.30) for SSI, 0.89 (95% CI 0.32-2.5) for anastomotic dehiscence and 1.26 (95% CI 0.59-2.78). The authors logically conclude that the current recommendations of using MOABP for colectomies to reduce SSIs or morbidity should be reconsidered.
So, where does that leave all those that had gotten used to the idea that pre-operative gut decontamination is a good thing for patients undergoing elective colorectal surgery? Sure thing is that the confidence intervals in the Finnish study do not rule out a clinically relevant beneficial effect for each of these outcomes. Another fact, in the Dutch patients received the classical SDD approach (colistin and tobramycin) starting several days before surgery, as it takes on average 3 days to decontaminate the gut. The Finnish variant of that was a single dosage of neomycin and of metronidazole on the night before surgery. If that fails to successfully eradicate gut flora, it acted as a placebo, and that would also explain the apparent discrepant findings.
The definitive trial on this topic is still needed, I’m afraid.