Rational use of antibiotics in the ICU: balancing stewardship and clinical outcomes.

Clinicians working in the intensive care unit (ICU) setting encounter the dilemma of prescribing antibiotics to treat critically ill patients with serious infections while minimizing the emergence and spread of antimicrobial resistance. Delaying the administration of appropriate antibiotic therapy (ie, an antibiotic regimen active against the causative pathogen based on in vitro testing) in the ICU has been associated with an increase in hospital mortality.1 One of the most important risk factors for delayed appropriate antibiotic therapy in seriously ill infected patients is prior exposure to antibiotics, typically administered within the previous 90 days.1 Prior antibiotic exposure promotes colonization and subsequent infection with antibiotic-resistant pathogens, complicating initial choice of antibiotics, thereby increasing the likelihood that delayed administration of appropriate antibiotic therapy will occur. Moreover, more prolonged durations of exposure to antibiotics seem to be most important in promoting the emergence of antibiotic-resistant pathogens in critically ill patients.2 Thus, clinicians working in the ICU must balance the needs of the patient they are directly treating with antibiotics with the needs of the other patients in the ICU who could subsequently be exposed to antibiotic-resistant bacteria induced by the currently prescribed antibiotic regimens. The importance of the global problem of increasing antimicrobial resistance is highlighted by a recent report from the Centers for Disease Control and Prevention indicating that infections attributed to antibiotic-resistant pathogens represent one of the most serious health threats.3 In this issue of JAMA, Oostdijk and colleagues4 used a cluster randomized crossover trial to study 2 different methods of administering antibiotic prophylaxis in patients requiring ICU care: selective digestive tract decontamination (SDD) (n = 6166 patients) vs selective oropharyngeal decontamination (SOD) (n = 5881 patients). Both regimens consisted of a combination of colistin, tobramycin, and amphotericin B.4 The SDD regimen also involved 4 days of intravenous antibiotic prophylaxis with a broad-spectrum cephalosporin. Both regimens are considered “selective” in that they aim to suppress overgrowth in the gut by unwanted potentially pathogenic microorganisms. Monthly point-prevalence surveys of respiratory and perianal culture samples were performed and demonstrated that the prevalence of antibiotic-resistant gram-negative bacteria in perianal swabs and ICU-acquired bacteremia were significantly less common with SDD compared with SOD (5.6% vs 11.8%, respectively; P < .001). However, there appeared to be significantly greater perianal carriage with aminoglycoside-resistant bacteria over time with the use of SDD (7% per month) than with SOD (4% per month). No differences in hospital mortality or other clinical outcomes were observed. Several important limitations of this report should be noted. First, this study was performed in the Netherlands, a country having historically low rates of antibiotic resistance compared with other parts of Europe. It is possible that the routine use of SDD or SOD could promote greater emergence of antibiotic resistance in ICUs in which the endemic background rate of colonization with antibiotic-resistant bacteria is greater than that observed in the Netherlands. The authors acknowledge the importance of their observation of increased perianal carriage of aminoglycoside-resistant bacteria during SDD administration, a finding similar to that observed in a previous study conducted by the same group demonstrating emergence of ceftazidime resistance over time with SDD.5 Of great concern is emergence of resistance to antibiotics, such as colistin, reserved for the treatment of antibiotic-resistant infections. Indeed, colistin resistance in gramnegative bacteria is increasingly reported in many parts of the world, including Europe, and has been associated with SDD use.6 The description of carbapenem-resistant Enterobacteriaceae acquiring colistin resistance is particularly concerning because these pathogens would potentially be resistant to all available antibiotic classes.7 Another limitation is the lack of a control group in which patients received neither SDD nor SOD. The rationale for exclusion of a control group seems to be an earlier study performed by the Netherlands group suggesting a mortality reduction with both SDD and SOD compared with placebo,8 even though in that study no reduction in crude mortality was observed with SDD and SOD. A post hoc random-effects logistic regression model adjusting for age, sex, severity of illness, intubation status, and medical specialty was required to demonstrate an association of SDD and SOD with mortality. Intuitively, the most important potential limitation of the routine use of SDD and SOD is that these interventions promote greater overall use of antimicrobial agents in the ICU, despite preventing nosocomial infections that could reduce subsequent need for antimicrobial therapy. The association between increasing antibiotic use and emergence of resistance is well established, especially in the ICU setting.9 The current trial by Oostdijk et al4 was not designed to assess the effect of SDD or SOD on total antimicrobial exposure in ICU patients. Given the important linkage between antibiotic use and emergence of resistance in the ICU, nonantibiotic pharmacologic and nonpharmacologic Related article page 1429 Opinion