DISCUSSION
Many treatments have been the subject of randomised controlled trials to decrease the rate of post-operative ileus. Nevertheless, our previous descriptive analysis of the literature did not allow us to draw clear conclusions as to the superiority of one treatment over another5. This finding led us to perform this network meta-analysis. We included in this meta-analysis only randomised controlled trials that reported the criteria commonly used in the return to normal transit. This network meta-analysis shows that prokinetics significantly reduce the duration of first gas, duration of first bowel movements and duration of post-operative hospitalisation. This treatment is ranked (SUCRA and rankogram) among the three best ones apart from food tolerance and the number of patients requiring a nasogastric tube. For food tolerance, opioid antagonists are the treatment that significantly improve the duration of food recovery.
The definition of return to normal transit is an important point to discuss. Indeed, not all segments are affected to the same extent. Small bowel motility is disturbed within 24 hours, gastric motility within 24 to 48 hours and colonic motility within 48 to 72 hours post-surgery4,50. The difference in time for recovery of motor function explains why the passage of the first stool and gas is most often used to define return to normal function. The complexity of the definition lies in the fact that the return of the migrating motor complex is not synonymous with a return to normal function, i.e. the perception of peristalsis on auscultation is not indicative of a return to normal transit. A recent literature review of 215 articles identified a total of 73 criteria defining return to normal transit9. Thus, in descending order of frequency, the criteria are: passage of first gas (140 studies out of 217, 64.5%), passage of first stool (69 studies out of 217, 31.8%) followed by first bowel movements (65 studies out of 217, 30%)9. The commonly accepted outcome for assessing the pharmacological effects of treatment for POI is the presence of first gas 9. Some studies have proposed composite scores but this is variable across studies. The COMET-registered core outcome set aims to standardise the reporting of outcomes in clinical studies of post-operative ileus8. Among recent work, the American Society for Enhanced Recovery After Surgery (ERAS) and Perioperative Joint Consensus have considered a more functional definition of POI and a classification system for post-operative gastrointestinal transit disorders51. Classification was proposed on a pathophysiological and functional basis using the following criteria: tolerance to oral ingestion, nausea, vomiting and physical signs of ileus (intake, sensation of nausea, vomiting, physical examination and duration of ”I-FEED” symptoms). A three-category classification system was therefore established. This recent score has never been evaluated in a prospective cohort of GI tract surgery patients. This score would allow a reproducible evaluation of the return to normal transit and therefore have comparable criteria for pharmacological studies51.
Adding duration of hospitalisation (a reproducible criterion) to the criteria commonly used and reported in the literature made two pharmacological principles stand out: prokinetics and opiate antagonists (as reported above). Prokinetics are made up of active principles used in clinical practice to treat nausea and vomiting. Their action on peristalsis supported an interesting approach in POI52,53. Among this class of potential active molecules, 5HT3 receptor antagonists (metoclopramide), selective 5HT4 receptor agonists (mosapride, prucalopride, cisapride) and ghrelin receptor agonists (ulimorelin) were the most evaluated. The results of our meta-analysis show that prokinetics are among the three best treatments for commonly used criteria (time for first bowel movement and for first stool) to characterise POI as well as the post-operative length of stay. Nevertheless, these results do not show a real superiority.
One way to optimise the post-operative recovery of bowel function after surgery would be to antagonise peripheral opioid receptors without negating their central analgesic action. The most commonly used drug for analgesia and anaesthesia is morphine which is a central and peripheral μ receptor agonist54. This central and peripheral action contributes to the prolongation of post-operative ileus although it is gastrointestinal receptors that have a predominant role in inhibiting post-operative gut motility. Morphine and other opioid analgesics inhibit the release of acetylcholine from the mesenteric plexus, thereby increasing colonic muscle tone and reducing propulsive activity in the gastrointestinal tract. There are several types of opioid receptors, the three main ones being μ, δ, and κ receptors with each class having several subtypes as well54. Opioid receptors are stimulated exogenously by agonists such as morphine and codeine. Both alvimopan and methylnaltrexone are the main peripheral opioid antagonists used that do not cross the blood-brain barrier55. Since the early 2000s, randomised controlled trials have been conducted in North America on cohorts of patients who have undergone bowel resection and hysterectomy23,26,28. Compared to placebo, patients treated with alvimopan had a significant reduction in time to transit recovery as evidenced by clinical functional signs such as first gas, first bowel movements or first stools. These encouraging results were not confirmed in a large clinical trial involving 70 hospitals in 10 countries on the European continent (Austria, Belgium, France, United Kingdom, Germany, Greece, Poland, Portugal, Spain and Sweden) and New Zealand22.
Despite the lack of a strict definition for POI resulting in discrepancies regarding the endpoints reported across studies, the current analysis provides the first Bayesian network analysis focused on pharmacological intervention of POI. This analysis is based on 6 robust endpoints, with nasogastric tube placement being the weakest endpoint because it is not included in all studies.
In conclusion, based on our meta-analysis, the two most consistent pharmacological treatments in terms of effectiveness for reducing POI after abdominal surgery are prokinetics and opioid antagonists. The absence of clear superiority of one treatment over another highlights the limits of the pharmacological principles available. It therefore appears necessary to act on other pathways. Indeed, there is a need to study and develop new pharmacological approaches that target the intimate mechanisms of intestinal damage involved in inflammation and/or neuroinflammation observed during post-operative ileus. New research approaches are required to help understand this phenomenon and develop new pharmacological treatments.