All great truths are iconoclastic: selective decontamination of the digestive tract moves from heresy to level 1 truth.

OBJECTIVE: The objective was to compare evidence of the effectiveness, costs and safety of the traditional parenteral antibiotic-only approach against that gathered from 53 randomised trials involving more than 8,500 patients and six meta-analyses on selective decontamination of the digestive tract (SDD) to control infection on the intensive care unit (ICU). PHILOSOPHY: Traditionalists believe that all infections are due to breaches of hygiene except those established in the first 2 days, and that all micro-organisms can cause death. In contrast, newer insights show that transmission via the hands of carers are responsible only for infections occurring after one week, and that only a limited range of 15 potential pathogens contribute to mortality. INTERVENTIONS TO PREVENT ICU INFECTION: The traditional approach is based on hand disinfection aiming at the prevention of transmission of all micro-organisms, to control all infections that occur after 2 days on the ICU. The second feature is the restrictive use of systemic antibiotics, only in cases of microbiologically proven infection. In contrast, SDD aims to control the three types of infection: primary, secondary endogenous and exogenous due to 15 potential pathogens. The classical SDD tetralogy comprises four components: (i) a parenteral antibiotic, cefotaxime, administered for three days to prevent primary endogenous infections typically occurring "early"; (ii) the oropharyngeal and enteral antimicrobials, polymyxin E, tobramycin and amphotericin B administered in throat and gut throughout the treatment on the ICU to prevent secondary endogenous infections tending to develop "late"; (iii) a high standard of hygiene to control transmission of potential pathogens; and (iv) surveillance samples of throat and rectum to monitor the efficacy of the treatment. ENDPOINTS: (i) Infectious morbidity; (ii) mortality; (iii) antimicrobial resistance; and (iv) costs.
RESULTS: Properly designed trials on hand disinfection have never demonstrated a reduction in either pneumonia and septicaemia, or mortality. Two randomised trials using restrictive antibiotic policies failed to show a survival benefit at 28 days. In both trials the proportion of resistant isolates obtained from the lower ways was >60% despite significantly less use of antibiotics in the test group. A formal cost effectiveness analysis of the traditional antibiotic policies has not been performed. On the other hand, two meta-analyses have shown that SDD reduces the odds ratio for lower airway infections to 0.35 (0.29-0.41) and mortality to 0.80 (0.69-0.93), with a 6% overall mortality reduction from 30% to 24%. No increase in the rate of super infections due to resistant bacteria could be demonstrated over a period of 20 years of clinical research. Four randomised trials found the cost per survivor to be substantially lower in patients receiving SDD than for those traditionally managed.
CONCLUSIONS: The traditionalists still rely on level 5 evidence, i.e. expert opinion, with a grade E recommendation, whilst the proponents of SDD are able to cite level 1 evidence allowing a grade A recommendation in their attempts to control infection on the ICU. The main reason for SDD not being widely used is the primacy of opinion over evidence.