Selective decontamination of the oropharynx and the digestive tract in ICU patients

Background: Selective Digestive tract Decontamination (SDD) and Selective Oropharyngeal Decontamination (SOD) are effective infection prevention measures in intensive care patients, but reported effects on patient outcome are conflicting and the effects on infections and respiratory tract colonization with Highly-Resistant Microorganisms (HRMO, Dutch Guidelines) are unknown. Methods: SDD and SOD were evaluated in a pragmatic open clustered group-randomized cross-over study in 13 ICUs. Patients with expected duration of intubation >48 hours or expected ICU-stay >72 hours were eligible. In each unit, the three regimens (SDD, SOD and standard care (SC)) were applied during 6 months in random order. Mortality at day 28 was the primary endpoint. A post-hoc subgroup analysis was undertaken to quantify effects of SDD and SOD among surgical and non-surgical patients. SDD consisted of four days of intravenous cefotaxime and topical application of tobramycin, colistin and amphotericin B in the oropharynx and stomach. SOD consisted only of oropharyngeal application of the same antibiotics. Cultures of respiratory tract were obtained twice weekly during SDD and SOD, and on clinical indication only during SC. All blood and respiratory tract culture results were evaluated. Monthly point prevalence studies were performed to analyze antibiotic resistance. Results: 5939 patients were enrolled (1990 SC, 1904 SOD, 2045 SDD). Crude mortality at day 28 was 27.5%, 26.6% and 26.9% for patients in the SC, SOD and SDD groups, respectively. In a random effects logistic regression model with age, gender, APACHE II scores, intubation status and medical specialty as covariates, odds ratios (OR) for mortality at day 28 were 0.86(CI950.74-0.99) for SOD and 0.83(CI950.72-0.97) for SDD. Compared to SC, adjusted ORs of mortality at day 28 were comparable in SDD-treated surgical and non-surgical patients (0.86(CI950.69-1.09) and 0.85(CI950.70-1.03)), respectively, but durations of mechanical ventilation, ICU-stay and hospital stay were significantly reduced in surgical patients only. In contrast, SOD did not influence mortality at day 28 0.97(CI950.77-1.22) among surgical patients, whereas in non-surgical patients adjusted OR was 0.77(CI950.63-0.94). Compared to SC, ORs for ICU-acquired bacteremia were 0.48(CI950.38-0.60) during SDD and 0.66(CI950.53-0.82) during SOD. The OR for ICU-acquired bacteremia caused by HRMO during SDD was 0.41(CI950.18-0.94) as compared to SC and when comparing SDD to SOD 0.37(CI950.16-0.85). ICU-acquired respiratory tract colonization of Gram negative bacteria was highest among patients receiving SC. ORs for acquiring HRMO colonization, as compared to SC, were 0.58(CI950.43-0.78) and 0.65(CI950.49-0.87) for SDD and SOD respectively. Acquired colonization with cefotaxime-resistant or colistin-resistant pathogens was lowest during SDD. Conclusions: In an ICU population with a baseline mortality rate of 27.5%, ARRs at day 28 were, after adjustment for covariates, 3.5% and 2.9% for SDD and SOD, respectively. In the subgroup analysis SDD was equally effective in reducing 28 day mortality at day 28 in surgical and non-surgical patients, whereas SOD was only effective in non-surgical patients with an ARR of 5.5%. As compared to SC, without adjustments for covariates, ICU-acquired bacteremia and respiratory tract colonization with HRMO were 48% and 59% lower during SDD and acquired respiratory tract colonization with HRMO was 38% lower during SOD