Meta-Analysis of Bowel Protocols in Critical Care Patients: A Word of Caution
In their recent meta-analysis pertaining to the efficacy of bowel protocols in critically ill adults, Oczkowski et al (1), in a recent issue of Critical Care Medicine, analyzed the results of four randomized controlled trials and eventually failed to demonstrate a beneficial effect of such protocols, though they stated that their results were limited by the small number and sample size of trials. We would like to underscore potential important flaws further limiting the conclusions of the meta-analysis. First, there was a wide heterogeneity in the case mix of the included studies. In one of the study (2), mean age of patients was roughly 38 years in both groups, sex ratio was rather imbalanced (up to 80% of men in both groups), and survival was 96% after a mean length of stay of about 21 days in both groups. This was very different when compared with other trials gathered in the meta-analysis including much older patients, and many more females. Furthermore, previous gastrointestinal surgery, the presence of an intestinal ostomy, chronic liver disease, or inflammatory bowel disease were unevenly declared as exclusion criteria. Second, in one study (2), more than 40% of patients in both groups underwent tracheostomy. This procedure was not reported in the other studies and might have biased the analysis of time spent under mechanical ventilation. Third, although authors stated that the consistent use of lactulose led to some homogeneity in outcomes, we would like to underscore the substantial differences among the tested protocols. The most important one is the timing of the tested interventions. As a matter of fact, timing of administration among included studies varied from a few hours to up to several days after ICU admission. Recent data plead in favor of a higher efficacy of prophylactic protocols (3), which could also have biased the analysis of the outcomes. Other differences that might have blurred the analysis are unequally found among studies and encompass: cointerventions (imbalanced administration of enemas and prokinetics), duration of the tested interventions, and the unblinded characteristics of some studies. Last, as stated by the authors, we would like to underscore the fact that the common criterion of 3 days without stool passage to define constipation seems to be inappropriate and could even explain why some previous well conducted studies in the field found no difference of outcomes according to this diagnostic criterion (4). Altogether, we shall not be surprised by the absence of effect evidenced in the meta-analysis. One might even wonder what practical conclusions were to be drawn, should it have found a positive impact.
Reduction of transit time in critical care patients seems to be the sum of the interactions between the acute disease together with its treatments, and the gut. We fully agree with the authors that the gap of knowledge, in this somewhat neglected area of critically ill patients management, should prompt us to perform well-designed studies to address this important issue.