Appropriate Empirical Antibacterial Therapy for Nosocomial Infections
- First Online:
- Cite this article as:
- Kollef, M.H. Drugs (2003) 63: 2157. doi:10.2165/00003495-200363200-00001
- 80 Views
The increasing presence of drug-resistant bacterial infections among hospitalised patients has resulted in greater numbers of patients receiving inappropriate antimicrobial treatment. This has led to the development of a novel paradigm guiding the administration of empirical antimicrobial therapy for patients with serious infections in the hospital setting. Antibacterial de-escalation is an approach to antibacterial utilisation that attempts to balance the need to provide appropriate, initial antibacterial treatment while limiting the emergence of antibacterial resistance. The goal of de-escalation is to prescribe an initial antibacterial regimen that will cover the most likely bacterial pathogens associated with infection while minimising the emergence of antibacterial resistance. Antibacterial resistance is minimised by narrowing the antibacterial regimen once the pathogens and their susceptibility profiles are determined, and by employing the shortest course of therapy clinically acceptable.
There is a general consensus that antimicrobial resistance in the hospital setting has emerged as an important variable that influences patient outcomes and overall resource utilisation.[1–3] Hospitals worldwide are faced with increasingly rapid emergence and spread of resistant bacteria. Both antibacterial-resistant Gram-negative bacilli and Gram-positive bacteria are reported as important causes of hospital-acquired infections.[4,5] In many cases, few antibacterial agents remain for effective treatment, particularly with methicillin-resistant Staphylococcus aureus (MRSA), vancomycin-resistant S. aureus and Gram-negative bacteria producing extended-spectrum β-lactamase (ESBL) with resistance to multiple other antibacterials.[6–8]
The increasing presence of resistant bacterial infections among hospitalised patients is probably related to the numerous pressures promoting resistance as well as the administration of inappropriate antibacterial therapy. These pressures include: (i) the frequent use of broad-spectrum antibacterials; (ii) prolonged use of antibacterials; (iii) crowding of patients with high levels of disease acuity within relatively small specialised areas of the hospital; (iv) reductions in nursing and ancillary support staff as a result of economic pressures, which increase the likelihood of person-to-person transmission of resistant bacteria; and (v) the presence of more chronically and acutely ill patients who require prolonged hospitalisations and often harbour resistant bacteria.[9–11] This review focuses on antibacterial resistance, the importance of administering initial appropriate antibacterial treatment for serious infections and antibacterial de-escalation as it applies to hospital-acquired infections, especially hospital-acquired pneumonia (HAP). Despite the limited class IA evidence (i.e. data from prospective, blinded, randomised trials) in support of these strategies, clinical practices are evolving based on the clinical experiences described in this review.
1. Importance of Appropriate Initial Antimicrobial Therapy
One of the consequences of an increase in antimicrobial resistance has been an increased recognition of inappropriate antimicrobial treatment of infections. Inappropriate antimicrobial treatment of serious infections in the hospital setting has been demonstrated to be an important determinant of hospital mortality.[13–15] Inappropriate antimicrobial treatment represents the use of antimicrobials with poor or no in vitro activity against the identified microorganisms causing infection at the tissue site of infection. Examples of inappropriate treatment include the absence of antimicrobial agents directed at a specific class of microorganisms (e.g. absence of therapy for fungaemia due to Candida species) and the administration of antimicrobial agents to which the microorganism responsible for the infection are resistant (e.g. empirical treatment with nafcillin for pneumonia subsequently attributed to MRSA). Changing antimicrobial therapy based on the subsequent available culture results and antibacterial susceptibilities may not reduce the excess risk of hospital mortality associated with inappropriate initial antibacterial treatment.[16,17] Therefore, the selection of initial appropriate therapy (i.e. getting antibacterial treatment right the first time) is an important aspect of care for hospitalised patients with serious infections.
Most inappropriate antimicrobial treatment of hospital-acquired infections appears to be related to bacteria resistant to the prescribed agents.[12–18] Although inappropriate antibacterial treatment may explain, in part, the greater mortality rates associated with resistant bacterial infections, other factors may also contribute to this excess mortality. Resistant Gram-positive bacteria such as MRSA express a number of virulence factors, which potentially contribute to their higher rates of associated mortality.[19,20] However, not all investigators have demonstrated greater mortality rates with infections due to MRSA compared with methicillin-susceptible S. aureus. Similarly, some resistant Gram-negative bacteria (e.g. Pseudomonas aeruginosa) are associated with increased virulence factors compared with susceptible pathogens.[22,23] This may explain the excess attributable mortality observed with infections due to these pathogens as well.
Hospital-acquired bloodstream infections are among the most serious infections acquired by intensive care unit (ICU) patients. Antibacterial resistance appears to have contributed to the increasing administration of inappropriate antibacterial therapy for hospital-acquired bloodstream infections, which is associated with greater hospital mortality rates.[13,14,24,25] The problem of antibacterial-resistant bacteraemia also appears to be increasing both in the hospital setting as well as in the community. Given the current trend of greater severity of illness for critically ill patients, it can be expected that infections due to resistant bacterial strains will be associated with greater morbidity and mortality, particularly when inappropriate empirical antibacterial therapy is administered. Along with greater patient mortality rates, infections due to resistant bacteria and inappropriate antibacterial treatment are associated with prolonged hospitalisation and increased healthcare costs relative to infections due to susceptible bacteria.[15,27,28] The overall US national costs of antibacterial resistance have been estimated to be between $US100 million and $US30 billion (2000 values) annually for the control and treatment of infections caused by resistant bacteria.
In addition to selecting the most appropriate antimicrobial agents for the treatment of serious infections, clinicians must ensure that antibacterial administration follows certain minimal requirements. These minimal requirements include proper dose, interval administration, optimal duration of treatment, monitoring of drug concentrations when appropriate and avoidance of unwanted drug interactions. Lack of adherence to these minimal requirements can result in unforeseen administration of suboptimal or excessive antibacterial tissue concentrations, which increases the likelihood for antibacterial resistance, patient toxicity and lack of effectiveness despite selecting an appropriate antibacterial regimen.
2. The De-Escalation Approach to Antibacterial Utilisation
An effective approach to antibacterial de-escalation necessitates that clinicians be aware of the bacteria that are most likely to be associated with infection and inappropriate antibacterial treatment in their practice setting. This requires that hospitals have updated and accurate antibiograms reflecting the bacterial pathogens and their antibacterial susceptibility encountered at the local level. Variability in the bacteria associated with hospital-acquired infections among hospitals, as well as within the wards of large hospitals, has been demonstrated to occur.[34,35] Additionally, changing temporal patterns of nosocomial pathogens and antibacterial susceptibility have been described. This suggests that hospitals may need to develop systems for updating local antibiograms on a regular basis because of the potential existence of intra-hospital and temporal variations. Using such data can improve the efficacy of antibacterial therapy by increasing the likelihood that appropriate initial antibacterial treatment will be prescribed to infected patients.[36,37]
In order to de-escalate or narrow the initial empirical antibacterial regimen, clinicians must be able to obtain culture specimens before starting antibacterial therapy. However, it is understood that prolonged delays in the administration of appropriate antibacterial treatment should not occur in seriously ill patients while waiting for specific cultures to be obtained (e.g. bronchoalveolar lavage, cerebral spinal fluid). The most common pathogens associated with the administration of inappropriate antibacterial treatment for hospital-acquired infections include potentially resistant Gram-negative bacteria (P. aeruginosa, Acinetobacter species, Klebsiella pneumoniae and Enterobacter species) and S. aureus, especially the strains with methicillin resistance.[12–18] Therefore, clinicians should consider initial empirical therapy for these pathogens, especially in patients at high risk for antibacterial-resistant infections.
Clinicians should also be aware that healthcare-acquired infections are similar to hospital-acquired infections in terms of the pathogens responsible for infection.[38–41] Healthcare-acquired infections are defined by a positive culture obtained within 48 hours of hospital admission and one of the following criteria: (i) received intravenous therapy at home, received wound care or specialised nursing care through a healthcare agency, family or friends or had self-administered intravenous medical therapy in the 30 days before the infection; (ii) attended a hospital or haemodialysis clinic or received intravenous chemotherapy in the 30 days before the infection; (iii) was hospitalised in an acute care hospital for ≥2 days in the 90 days before the infection; and (iv) resided in a nursing home or long-term care facility. Physicians should be aware of the factors that identify patients as having healthcare-acquired infections in order to avoid the prescription of inappropriate antibacterial treatment.
3. Examples of the De-Escalation Approach
Trouillet and co-workers identified risk factors for ventilator-associated pneumonia (VAP) caused by potentially drug-resistant bacteria such as MRSA, P. aeruginosa, Acinetobacter baumannii and/or Stenotrophomonas maltophilia in 135 consecutive episodes of VAP observed in a single ICU over a 25-month period. Seventy-seven (57.0%) episodes of VAP were caused by ‘potentially resistant’ bacteria and 58 (43.0%) episodes were caused by ‘other’ non-resistant organisms. According to logistic regression analysis, three variables were predictors for VAP due to potentially drug-resistant bacteria: (i) duration of mechanical ventilation ≥7 days (odds ratio [OR] 6.0; 95% CI 1.6–23.1); (ii) previous antibacterial use (OR 13.5; 95% CI 3.3–55.0); and (iii) previous use of broad-spectrum drugs (third-generation cephalosporin, fluoroquinolone and/or a carbapenem) [OR 4.1; 95% CI 1.2–14.2]. Differences in the potential activities (ranging from 100% to 11%) against microorganisms for 15 different antibacterial regimens were studied to determine their overall level of activity for the pathogens associated with VAP. These investigators observed that the combination of a carbapenem plus amikacin and vancomycin provided the broadest in vitro coverage against the spectrum of Gram-negative and Gram-positive bacteria that were found in their ICU. Although clinical outcomes were not assessed, this study suggests that location-specific empirical antibacterial regimens, tailored to the susceptibility patterns of the local flora, are most likely to be effective. Additionally, this study demonstrates that patients at high risk for infection with resistant bacteria can be identified based on risk factors that promote colonisation with such pathogens (e.g. previous antibacterial therapy, and exposure to high-risk environments including hospitals and long-term care facilities).
Ibrahim et al. evaluated 50 consecutive patients in the ICU setting who were receiving antibacterial therapy for VAP. They subsequently examined 52 consecutive patients with VAP whose antibacterial treatment was administered according to a unit-specific antimicrobial guideline. The main goal of the guideline was to provide initial administration of appropriate antimicrobial treatment while avoiding the emergence of antibacterial resistance. This meant providing initial coverage for P. aeruginosa and MRSA, the two most common pathogens causing VAP in that specific ICU. This was accomplished by providing initial intravenous combination treatment with vancomycin, a carbapenem and a fluoroquinolone. This combination was selected because it provided in vitro coverage for >90% of all the bacterial isolates identified based on a unit-specific antibiogram. The guideline also required that antibacterial treatment be modified after 48 hours based on the available culture results and the clinical course of the patient. In fact, 61.5% of patients had two antibacterials discontinued within 48 hours of beginning therapy based on Gram-stain results from respiratory secretions and culture data.
The studies of Trouillet et al. and Ibrahim et al. employed local bacterial susceptibility data in order to develop recommendations for empirical therapy within their respective ICUs. The use of local data should always be taken into consideration when developing empirical regimens for treatment. The American Thoracic Society (ATS) released a consensus statement in 1996 on the diagnosis, assessment, treatment and prevention of HAP that supported the de-escalation approach to antibacterial treatment. This document identifies those microorganisms commonly implicated in HAP and provides antibacterial recommendations based on the infecting pathogens. According to the consensus statement, the management of HAP should be based on the following factors: (i) severity of illness (i.e. mild-to-moderate vs severe as defined in the ATS statement); (ii) presence of risk factors suggesting infection with specific microorganisms (to include the use of local data on pathogens and susceptibility patterns); and (iii) the time of pneumonia onset (i.e. <5 days after hospital admission vs ≥5 days after admission).
The ATS guidelines note that patients with mild-to-moderate disease presenting <5 days after hospitalisation generally are infected with Haemophilus influenzae, Streptococcus pneumoniae, methicillin-susceptible S. aureus or other ‘core microorganisms’ (i.e. susceptible bacteria). However, the presence of risk factors (e.g. recent abdominal surgery, diabetes mellitus, head trauma, history of high-dose corticosteroids) or local patterns of pathogen prevalence and antibacterial susceptibility can alter the type of predominant infecting microorganism needing coverage. The duration of hospitalisation before the onset of pneumonia is one of the most important determinants for potential infecting pathogens. The risk of infection with MRSA, P. aeruginosa and Acinetobacter species is increased if nosocomial pneumonia presents after the patient has been hospitalised for >5 days. In patients with specific risk factors, prolonged hospitalisation or severe disease, combination antibacterial therapy should initially target this broader spectrum of likely infecting bacteria. In addition, combination therapy is required in suspected infections due to P. aeruginosa to improve survival rates and reduce the development of antibacterial resistance. Monotherapy may be adequate in some cases of severe nonpseudomonal infection but few studies have evaluated the efficacy of single agents in severe disease.[44–48]
4. Timing of Antibacterial Treatment
A number of investigators have found that delays in the administration of appropriate antibacterial treatment are associated with excess hospital mortality. Alvarez-Lerma showed that among 490 episodes of pneumonia acquired in the ICU setting, 214 episodes (43.7%) required modification of the initial antibacterial regimen as a result of either isolation of a resistant microorganism (62.1%) or lack of clinical response to therapy (36.0%). Attributable mortality from VAP was significantly lower among patients receiving initial, appropriate antibacterial treatment than among patients receiving inappropriate treatment requiring a treatment change (16.2% vs 24.7%; p = 0.034).
Iregui and colleagues examined the influence of initially delayed appropriate antibacterial treatment on the outcomes of 107 patients with VAP. All 107 patients eventually received treatment with an antibacterial regimen that was shown in vitro to be active against the bacterial pathogens isolated from their respiratory secretions. Thirty-three (30.8%) patients received appropriate antibacterial treatment that was delayed for ≥24 hours after patients initially met diagnostic criteria for VAP. The most common reason for the administration of delayed treatment was a delay in physicians’ recognising the presence of VAP and writing the orders for antibacterial treatment (n = 25; 75.8%). Patients who receive delayed antibacterial treatment have a statistically greater hospital mortality rate than patients without the delay (69.7% vs 28.4%; p < 0.001).
The studies of Alvarez-Lerma and Iregui et al. suggest that delaying the administration of appropriate antibacterial treatment is an important factor contributing to the excess hospital mortality of patients with serious infections. Delays in the administration of appropriate antibacterial treatment have most recently also been associated with greater mortality for patients with severe sepsis.[50,51] These studies demonstrate that initial selection of the ‘wrong’ antibacterial regimen, because of a lack of efficacy for the identified bacteria(s) associated with infection, will delay appropriate antibacterial delivery to the tissue site of infection and result in worse clinical outcomes.
5. De-Escalation and Antibacterial Resistance
The second goal of antibacterial de-escalation is to avoid the administration of unnecessary antibacterial treatment. First, this requires narrowing the antibacterial coverage after identification of the pathogen to which the infection is attributed and the antibacterial susceptibility profile of the microorganism. Second, the duration of antibacterial treatment should be limited to the shortest effective course of therapy. The following approaches are ways in which unnecessary antibacterial therapy can be avoided in a de-escalation strategy.
Antibacterial practice guidelines or protocols have emerged as a potentially effective means of both avoiding unnecessary antibacterial administration and increasing the therapeutic effectiveness of these agents. The potential benefits of antibacterial use guidelines in the hospital setting have been well demonstrated by the experience at LDS Hospital in Salt Lake City, Utah, USA, which employs a computerised system guiding antibacterial use. This system has been successfully employed to reduce inappropriate, empirical antibacterial administration compared with individual physician prescribing practices. This automated guideline has also been shown to significantly reduce orders for drugs to which patients were allergic, overall adverse events caused by antibacterials and the total number of anti-infective doses prescribed, as well as the medical costs associated with antibacterial agents.
Nonautomated or partially automated protocols, usually driven by hospital-based quality improvement teams, have demonstrated similar results to those observed with the fully automated antibacterial utilisation systems. In two teaching hospitals, Bailey and co-workers randomised patients to have their physicians contacted by pharmacists with consensus recommendations to discontinue intravenous antibacterials versus no intervention. The intervention significantly reduced the antibacterial doses administered as well as the mean antibacterial costs but was associated with increased labour costs. Similarly, Leibovici and co-workers developed a problem-oriented database decision support system, which significantly reduced the unnecessary use of antibacterials and decreased inadequate antibacterial administration, particularly to patients infected with multiresistant Gram-negative isolates, enterococci and S. aureus. With the advent of newly available technologies, including handheld computers and portable communication devices, the ability to extend the influence of treatment protocols to the bedside is becoming increasingly possible.
5.2 Shorter Courses of Antibacterial Therapy
Recently, several groups of investigators have demonstrated that shorter courses of therapy for VAP can be efficacious while reducing the emergence of antibacterial resistance. Singh and co-workers used a scoring system to identify patients with suspected VAP who could be treated with 3 days of antibacterial therapy as opposed to the conventional practice of 10–21 days. Patients receiving the shorter course of antibacterial therapy had clinical outcomes similar to the patients receiving longer therapy, but with fewer subsequent superinfections attributed to antibacterial-resistant pathogens. Similarly, Ibrahim et al. employed a pharmacist-directed protocol in the ICU setting to reduce the empirical administration of antibacterials for suspected VAP from 14.8 ± 8.1 days to 8.1 ± 5.1 days (p < 0.001) [figure 2].
Although recommended durations of antibacterial therapy for VAP have generally ranged from 14 to 21 days in the past, these were not based on prospective studies. Dennesen and colleagues evaluated 27 patients with VAP to determine their resolution of signs and symptoms. All patients received appropriate antibacterial therapy, and the highest temperatures, leucocyte counts, oxygen tension in arterial blood to inspired oxygen ratio (PaO2 : FiO2) and semiquantitative cultures of endotracheal aspirates were recorded from the start of therapy until day 14. VAP was caused by Enterobacteriaceae (n = 14), P. aeruginosa (n = 7), S. aureus (n = 6), H. influenzae (n = 3) and S. pneumoniae (n = 1). H. influenzae and S. pneumoniae were eradicated from tracheal aspirates, whereas Enterobacteriaceae, S. aureus and P. aeruginosa persisted, despite in vitro susceptibility to the antibacterials administered. Significant improvements were observed for all clinical parameters, most apparently within the first 6 days after the start of antibacterials. Newly acquired colonisation, especially with P. aeruginosa and Enterobacteriaceae, occurred in the second week of therapy. These data support the premise that most patients with VAP receiving appropriate antibacterial therapy respond to treatment within the first 6 days and acquire colonisation with resistant bacteria usually during the second week of therapy, which frequently precedes a recurrent episode of VAP.
The available medical literature suggests that shorter courses of appropriate antibacterial therapy for VAP should be employed because of their equivalent efficacy, compared with longer courses of therapy, and reduced propensity for the emergence of bacterial resistance. A European, multicentre, randomised trial of 8 days versus 15 days of antibacterial treatment for VAP was presented at the ATS meeting; 2003 May 16–21; Seattle (WA), demonstrating equivalent clinical outcomes for both study groups. However, patients randomised to the shorter course of therapy had more antibacterial-free days and the emergence of fewer subsequent infections due to multiresistant Gram-negative bacteria. This represents the first large (n = 401), randomised clinical trial to provide grade IA evidence in support of a shorter duration of therapy for VAP.
5.3 Scheduled Antibacterial Changes/Antibacterial Cycling/Antibacterial Rotation
Rahal and co-workers introduced a new antibacterial guideline into their hospital that significantly restricted the use of the cephalosporin class. This was done to combat an outbreak of ESBL-producing Klebsiella infection. The restriction of cephalosporins was successful, with an 80% reduction in their hospital-wide use accompanied by a 44% reduction in infection and colonisation with the ESBL-producing Klebsiella spp. However, the use of imipenem increased by 140% during the intervention year and was associated with a 69% increase in the incidence of imipenem-resistant P. aeruginosa throughout the medical centre. Therefore, this represented a change or switch from one predominant antibacterial class to another. Although the number of multipli-resistant pathogens decreased with this formulary restriction, its overall effectiveness can be questioned.
Kollef and co-workers examined the influence of a scheduled antibacterial change on the incidence of nosocomial infections among patients undergoing cardiac surgery. A 6-month ‘before-period’, during which the traditional practice of prescribing a third-generation cephalosporin (ceftazidime) for the empirical treatment of Gram-negative bacterial infections, was followed by a 6-month ‘after-period’, during which a fluoroquinolone (ciprofloxacin) was employed. Unexpectedly, the overall incidence of VAP was significantly reduced in the after-period compared with the before-period. This was primarily a result of a significant reduction in the incidence of VAP attributed to resistant Gram-negative bacteria. Similarly, a lower incidence of Gram-negative bacteraemia due to resistant pathogens was observed in the after-period. This experience was followed by a series of scheduled antibacterial changes for the treatment of suspected Gram-negative bacterial infections among patients admitted to the medical and surgical ICUs. The overall prescription of appropriate antibacterial therapy was statistically increased for Gram-negative bacterial infections during this experience.
The concept of antibacterial class cycling or rotation has been advocated as a potential strategy for reducing the emergence of antibacterial resistance. In theory, a class of antibacterials or a specific antibacterial drug is withdrawn from use for a defined time period and reintroduced at a later point in time in an attempt to limit bacterial resistance to the cycled antibacterial agents. Gerding and colleagues evaluated cycling of aminoglycosides over 10 years at the Minneapolis Veterans Affairs Medical Center, Minneapolis, Minnesota, USA, cycling amikacin and gentamicin. Resistance to gentamicin had emerged as a clinical problem limiting the use of that specific aminoglycoside at this hospital. Using cycle times of 12–51 months, these investigators found significantly reduced resistance to gentamicin when amikacin was used, but a return of resistance with the rapid reintroduction of gentamicin, followed by more gradual reintroduction of gentamicin a second time without increased levels of resistance recurring. This experience suggested that the cycling of antibacterials within the same drug class, in some circumstances, could be an effective strategy for curbing antibacterial resistance.
Gruson and colleagues observed a reduction in the incidence of VAP after introducing an antibacterial programme that consisted of supervised rotation and restricted use of ceftazidime and ciprofloxacin, which were widely prescribed before the institution of the antibacterial programme. The antibacterial selection was based on monthly reviews of the pathogens isolated from the ICU and their antibacterial susceptibility patterns. Therefore, these clinicians were rotating antibacterial agents based on ‘real-time’ information, which allowed potentially more effective antibacterials to be prescribed to their patients. They observed a decrease in the incidence of VAP, which was primarily due to a reduction in the number of episodes attributed to potentially resistant Gram-negative bacteria including P. aeruginosa, Burkholderia cepacia, S. maltophilia and A. baumannii. These same investigators have now demonstrated that their initial results could be sustained over a 5-year time period.
Clinicians caring for patients in the hospital setting must champion and employ strategies for more effective antimicrobial utilisation. The most successful strategies will be multidisciplinary involving co-operation from the pharmacy, nursing staff, treating physicians and infectious disease consultants. Antibacterial utilisation programmes should also be linked to infection control practices in order to minimise the future emergence and spread of antibacterial-resistant microorganisms. The de-escalation approach to antibacterial therapy is a method of providing appropriate initial treatment to patients with serious bacterial infections (‘getting it right the first time’) while avoiding the unnecessary use of antibacterials in order to prevent the emergence of resistance.
This work was supported in part by the Barnes-Jewish Hospital foundation. The author has no conflicts of interest that are directly relevant to the content of this manuscript.