Decreased mortality rate and length of hospital stay in surgical intensive care unit patients with successful selective decontamination of the gut.

OBJECTIVE: Current studies concerning selective decontamination of the digestive tract have failed to demonstrate a decrease in the length of hospital stay and mortality rate, despite the finding of a significantly lower number of infections. To evaluate this issue in more detail, the relationship between the mortality rate and length of stay with respect to colonization and infections was studied within a group of patients receiving selective decontamination. Special attention was given to the efficacy of decontamination within each patient. The main question addressed was whether an effect on mortality rate was present, and if so, why this effect was not apparent until now.

DESIGN: Prospective observational cohort study.

SETTING: Surgical intensive care unit (ICU) in a university hospital.

PATIENTS: Ninety-seven patients primarily admitted into the surgical ICU who received selective decontamination. Transferred patients were excluded. The majority of the surgeries were elective, and all patients completed the follow-up.

INTERVENTIONS: All patients received polymyxin E, amphotericin B, and norfloxacin four times a day in a 2% solution of Orabase orally and enterally as suspensions of 200, 500, and 50 mg, respectively. Assessment of the efficacy of selective decontamination was done by identification of Gram-negative microorganisms in surveillance cultures from the oropharynx and rectum. Predicted mortality rates for each patient were calculated with a logistic regression formula.

MEASUREMENTS AND MAIN RESULTS: A possible benefit of selective decontamination of the digestive tract would be expressed by lower actual mortality rates compared to predicted mortality rates. Since we expected the efficacy of decontamination to have an influence on infection and mortality rates, we evaluated these rates in terms of successful or unsuccessful decontamination. Most patients (n = 72) were successfully decontaminated. Actual death rates in these patients were significantly lower than the expected rates (as calculated by the Acute Physiology and Chronic Health Evaluation [APACHE] II scoring system) (18% vs. 40%, p = .006), whereas no difference was found in those patients with failed decontamination (n = 25, death rate 44%). The patients with unsuccessful selective decontamination had significantly longer hospital (52 vs. 34 days) and ICU lengths of stays (23 vs. 9 days; p = .002) and higher mortality rates (44% vs. 18%, p = .020) when compared with those patients who were successfully decontaminated.

CONCLUSIONS: These results indicate that selective decontamination is beneficial in terms of mortality rate and length of stay in surgical patients only when successful decontamination has been achieved. The subgroup of patients for whom decontamination is not successful might be responsible for the obscurity in mortality effects of selective decontamination in studies until now. It is expected that identification and subsequent elimination of possible risk factors that cause a failure of selective decontamination can result in lower morbidity and mortality rates in critically ill, surgical patients admitted to the ICU.