Pro/con debate: Should antimicrobial stewardship programs be adopted universally in the intensive care unit?

You are director of a large multi-disciplinary ICU. You have recently read that hospital-wide antibiotic stewardship programs have the potential to improve the quality and safety of care, and to reduce the emergence of multi-drug resistant organisms and overall costs. You are considering starting one of these programs in your ICU, but are concerned about the associated infrastructure costs. You are debating whether it is worth bringing the concept forward to your hospital's administration to consider investing in.


Introduction
Antibiotic stewardship programs are multidisciplinary initiatives whose primary aim is to optimize antibiotic usage. Th e Infectious Disease Society of America (IDSA) and the Society for Health Care Epidemiology of America (SHEA) published guidelines for antimicrobial stewardship in 2007 aimed at providing information on how to establish such programs within health care institutions [1]. Because antibiotics are used heavily in the ICU, stewardship programs appear particularly applicable to this setting. Antimicrobial stewardship is broadly defi ned as a practice that ensures the optimal selection, dose and duration of antimicrobials and leads to the best clinical outcome for the treatment or prevention of infection while producing the fewest possible side eff ects and the lowest risk for subsequent resistance [2]. Antimicrobial stewardship programs may contain a variety of interventions that are complementary to eff ective infection prevention and control programs.
Inappropriate antimicrobial usage is a signifi cant problem, with approximately 50% of antimicrobial usage being unnecessary or suboptimal in hospital, community or ambulatory settings [3,4]. A recent study showed that approximately 20% of patients admitted to the ICU with Clostridium diffi cile-associated diarrhoea were receiving antibiotics without any obvious evidence of infection, with an accompanying 28% in-hospital mortality [5]. As a consequence of indiscriminate antibiotic use, there are reported increases in the incidence of infections caused by resistant organisms. A signifi cant correlation was demonstrated between the increase in fl uoroquinolone prescriptions in Canada from 0.8 to 5.5 per 100 persons per year and increased ciprofl oxacin-resistant Streptococcus pneumoniae from 0% to 1.7% [6]. Twelve percent of patients previously exposed to piperacillin-tazobactam were colonized with strains of enterobacteriaceae resistant to this antibiotic [7] and the use of third generation cephalosporins is associated with higher rates of vancomycin-resistant enterococci and extended-spectrum β-lactamase-producing organisms [8]. Anti microbial resistance emerging in response to the selective pressure exerted by antibiotics is also a clinical phenomenon, with outbreaks of antibiotic-resistant Pseudomonas aeuroginosa and Acinetobacter baumanii-calcoaceticus occurring in ICUs, where a huge antimicrobial pressure is present [9][10][11].
Although they are often life-saving, antibiotics can also cause serious harm to patients, including Clostridium diffi cile -associated diarrhoea, antibiotic-resistant infections and invasive candidiasis [12][13][14]. Antibiotics also result in dangerous drug interactions, life-threatening hypersensitivity reactions, nephrotoxicity, and QT prolongation, to name a few. Inappropriate antibiotic use also contributes to rising drug and hospitalisation costs, and the need to preserve our current antibiotic arsenal Pro: There is justifi cation for implementing antibiotic stewardship programs in the ICU Clinicians have long been aware of the risks of antibiotic resistance associated with inappropriate antibiotic use, but nonetheless very few eff ective antibiotic policies have been implemented, and the problem appears to be even worsening [16]. Th e costs associated with antibiotic usage are also escalating, with systemic antibiotics being the single most costly drug class over the past decade in non-federal hospitals in the United States. In 2007, systemic antibiotics accounted for 11.2% of the pharmacy budget of non-federal hospitals [17]. In addition to direct pharmacy costs, hospitalisation and other infrastructure costs are also increased, ultimately resulting in a greater strain on the healthcare system. Saving antibiotics will save money, and there are a variety of methods to do so.
Education is the cornerstone of any antibiotic stewardship program, with prescriber education and imple mentation of guidelines and clinical pathways improving antimicrobial prescribing behaviour. For example, studies using algorithms to shorten the course of antimicrobial therapy in ventilator-associated pneumonia led to significantly lower antimicrobial therapy usage with reduction in costs, antimicrobial resistance, and super-infections without adversely aff ecting the length of stay or mortality [18,19]. Th e absence of formal antimicrobial stewardship training programs for infectious diseases fellows, boardcertifi ed physicians, and pharmacists has recently been a challenge to the education imperative, however [20].
Preauthorisation (also known as formulary restriction) requires approval by a pharmacist or physician prior to clinical use of an antimicrobial. Although preauthori zation is thought to be the most eff ective method of controlling antimicrobial use, it does not alter the duration of therapy or the decision to give or withhold antibiotics. Th e main benefi ts of this strategy are the supervision of antibiotic use by experts and substantial cost savings (with some studies demonstrating cost savings upwards of US$800,000) [21,22].
Th rough prospective audit with interaction and feedback, antimicrobial use is reviewed after antimicrobial therapy has been initiated and recommendations are made with regard to their appropriateness in terms of selection, dose, route and duration. Prospective audit with feedback avoids delays in initiation of therapy and maintenance of prescribers' autonomy, and can be implemented in health care facilities of varying sizes [23,24]. A large teaching hospital reported a 37% reduction in the number of days of unnecessary antibiotics use by decreasing the duration of treatment and by reducing new starts [25]. In another study, antimicrobial suggestions from an infectious disease fellow and a clinical pharmacist resulted in 1.6 fewer days of parenteral therapy and cost savings with no adverse eff ects on clinical response [23]. Another study demonstrated a sustained decrease in parenteral antibiotics over a 7-year period following introduction of a prospective audit with interaction and feedback [26].
Multiple studies using healthcare information technology, such as computer-assisted decision support designed to provide treatment recommendations, have shown signifi cant reductions in the use of antibiotics and greater de-escalation to narrow-spectrum antimicrobials. Improvements in cost and effi ciency of existing stewardship programs, and improved physician knowledge regard ing treatment and pathogen prediction were also noted [27][28][29]. In addition to improving antimicrobial use and patient care (including tracking of antibiotic resistance patterns), such systems can improve surveil lance of hospital-acquired infections and adverse drug events when compared to manual surveillance methods [30,31]. In a 15-month study using a web-based antimicrobial approval system linked to national antibiotic guidelines, a sustained reduction in third-generation cephalosporin prescriptions were accompanied by increased concordance with antibiotic guidelines [32]. Th ese benefi ts have also been noted in an ICU-based study, where investigators used computerised anti-infective programs and were able to document signifi cant reductions in the use of excessive drug dosage, adverse drug events and length of hospital stay and costs [33].
Standardized pre-printed or computer-generated physician order sets can improve the effi ciency of antibiotic stewardship programs. In a study looking into their benefi ts in the management of patients with septic shock in an emergency department, order sets were found to improve initial fl uid resuscitation, use of appropriate antibiotics and 28-day mortality [34]. A recent study to evaluate the hospital-wide impact of a standardized order set for the management of severe bacteraemic sepsis has shown that a greater number of patients received appropriate initial antibiotic therapy with decreased incidence of organ failure and improved survival [35].
A survey of 670 US hospitals found that implementation of guideline-recommended practices to control antimicrobial use and optimize the duration of empirical therapy was associated with less antimicrobial resistance, including methicillin-resistant Staphylococcus aureus, vanco mycin-resistant enterococci, fl uoroquinoloneresistant Escherichia coli and ceftazidime-resistant Klebsiella species [36]. Given the relationship between antimicrobial use and antimicrobial resistance, antimicrobial stewardship appears to be a logical fi rst step in the eff ort to control antimicrobial resistance.
Th e effi cacy of antimicrobial stewardship programs has been the subject of a recent Cochrane systematic review, examining 66 studies from 1980 to 2003 [37]. Th e main interventions analyzed in the review were targeted to decrease treatment (57 studies), increase treatment (6 studies) or both (3 studies). Th e interventions addressed the antibiotic regimen (61 studies), the duration of treatment (10 studies), the timing of fi rst dose (6 studies), or the decision to prescribe antibiotics (1 study). Optimization of antibiotic use was seen in 81% of the studies aimed at improving antimicrobial utilization. Signifi cant improvements in microbiological outcome (for example, prevalence of antibiotic-resistant bacteria) and clinical outcomes (for example, mortality and length of hospital stay) were also noted in some studies. Recent observational studies (subsequent to the Cochrane review) have demonstrated that reducing antimicrobial pressure correlates with improved antimicrobial susceptibility of pathogens [38,39].
Antimicrobial stewardship programs using the methods described above will promote the optimal use of antimicrobial therapy, leading to the best clinical outcome for patients. Th e relative paucity of outcome data demonstrating the benefi ts of antimicrobial stewardship is likely due to its infancy: antimicrobial stewardship programs today are where infection control programs were roughly 30 years ago [40,41]. Because antimicrobials are widely prescribed in the ICU, with an apparent mortality benefi t with appropriate therapy [42], using the best available methods to optimize their use through antimicrobial stewardship is crucial.

Con: The evidence for eff ectiveness of antimicrobial stewardship is lacking
Despite the publication of guidelines for improving the use of antimicrobial agents in the United States, a great deal of scepticism about the eff ectiveness and acceptability of antimicrobial stewardship programs persists. In a survey conducted by the United States Centers for Disease Control and Prevention's National Nosocomial Infections Surveillance Systems, only 40% of selected hospitals had antibiotic restriction policies and 60% used stop orders [43]. Antimicrobial stewardship programs are also 50% less likely to be implemented in community hospitals compared to academic hospitals [44]. Two years after the publication of the IDSA/SHEA antibiotic steward ship guidelines [1] only 48% of survey respondents stated that their hospital had a program [41].
Reduction in the incidence of bacterial resistance is touted as the main advantage of antimicrobial stewardship programs, but lacks scientifi c evidence to support it. In a recent survey of 33 US hospitals, there was no signifi cant correlation between antibiotic guideline adherence by physicians and resistance rates [45].
Antibiotic use in ICUs may be the consequence rather than the cause of resistance, and there is a risk that stewardship, with its emphasis on decreased antibiotic use, could lead to a substantial increase in patient risk. It is also important to note that neither the published guidelines nor the important stewardship articles identify safety as an endpoint.
Another potentially adverse consequence of antibiotic restriction is the emergence of new resistance patterns replacing the old ones. A study documenting the introduction of new guidelines that restricted cephalosporin use was primarily aimed at reducing the incidence of cephalosporin-resistant Klebsiella spp. Even though the primary aim was achieved, this occurred at the expense of increased imipenem usage with the subsequent increase in incidence of imipenem-resistant P. aeuroginosa by about 69% [46]. Th us, formulary restriction does not necessarily prevent the potential overuse of available broad spectrum antibiotics in routine practice [47]. Rather, a signifi cant change in clinical thinking to reduce our dependence on and abuse of antibiotics is needed.
Antimicrobial stewardship programs form only one strategy for minimizing the incidence of resistance, and must partner with infection control measures, including surveillance, outbreak investigation, disinfection and sterilization, and environmental hygiene. Of the studies reported to be benefi cial, it remains unclear as to whether the reported improvements in resistance rates are related to antimicrobial stewardship programs, infection control measures or both.
Although healthcare information technology is believed to be a key component of antimicrobial stewardship programs, detailed information on the resources required to implement and maintain these sophisticated computer programs is not widely available. It is also not clear whether the reported cost-eff ectiveness of many of these stewardship programs takes into account the overall cost of these interventions above and beyond the pharmacyrelated costs and expenses associated with development and distribution of educational materials.
Another challenge to implementing antimicrobial steward ship in the ICU deals with the confi dence intensivists have in the clinical judgement of the stewardship physician. A junior physician might be a less eff ective antimicrobial stewardship team member because of a perceived or real lack of knowledge and experience [48], but may be utilized because the 'price is right' . In the survey by Pope and colleagues [41], personnel shortages (55%), fi nancial considerations (36%), and resistance from administration (14%) were frequent barriers to establishing antimicrobial stewardship programs. Oppo si tion from prescribing physicians was a barrier to establishing an antimicrobial stewardship program in about 27% of cases.
While antimicrobial stewardship programs have rather consistently shown signifi cant improvement in antimicro bial utilization, there are very few studies examining meaningful clinical outcome measures such as duration of hospitalization, mortality rates, or even quality indicators such as patient satisfaction. In the systematic review by the Cochrane Collaboration on antibiotic stewardship programs, clinical outcomes such as mortality and length of hospital stay were reported in only 15% of the studies [37]. In the 2008 survey by Pope and colleagues [41], only 25% of respondents reported clinical outcomes. Also, none of the studies report any signifi cant reduction in antimicrobial side eff ects as a result of these interventions.

Conclusion
Hospitals are increasingly implementing antimicrobial steward ship programs in response to increasing antimicro bial resistance (despite aggressive infection control practices), coupled with fewer novel antimicrobials and increasing antimicrobial costs. Th ere is little question that antimicrobial use is causally related to antimicrobial resistance, and there is growing evidence that stewardship measures aimed at optimizing antimicrobial use can reduce antimicrobial resistance while reducing associated costs. Being major foci of antimicrobial resistance and the largest consumers of antimicrobials in most hospitals, ICUs can expect to benefi t most from antimicrobial stewardship programs.
Full implementation of antibiotic stewardship programs requires signifi cant investment, however. In the present economic climate, barriers to implementing such programs include personnel shortages, fi nancial cut backs, and resistance from administration who are reluctant to assume economic risk. Focusing on patient safety initiatives and the benefi ts of cost savings and cost avoidance may enable hospital administrators to look upon antibiotic stewardship programs favourably [20]. Supplemental strategies such as consultations provided by specialists in infectious diseases might also be used in lieu of clinical decision support systems. Such expertise has been shown to improve antimicrobial use, shorten duration of mechanical ventilation and ICU stay, and to reduce in-hospital and ICU mortality [49], although it is unlikely that a clinical-decision support system would be entirely replaced. In addition to pre-authorization and/or audit-and-feedback approaches, ICUs should consider other strategies to improve antimicrobial utilization. In short, stewardship programs should be adapted according to the individual needs of institutions, but should be adequately resourced to achieve their intended aims.
ICUs are complicated systems, and implementing a complex program into another complex structure raises the potential of unintended (and often unmeasured) adverse consequences. All ICUs should have an antimicrobial stewardship program accompanied by a system to monitor clinically meaningful outcomes such as mortality and length of stay. Monitoring such outcomes presents an excellent opportunity for infection control and other patient quality and safety initiatives, whose aims include prevention of healthcare-associated infections and control of antibiotic-resistant organisms. In the absence of such monitoring, antimicrobial stewardship programs are nothing more than programs to reduce antimicrobial use with a largely unproven eff ect on patient care. Close collaboration between critical care, infectious disease, infection control, medical informatics, microbiology, and pharmacy staff are needed for the success of an antimicrobial stewardship program. From our experience, leadership and a culture that embraces change is critical to implementation of a successful antimicrobial stewardship program.