Broad-Spectrum Antimicrobials and the Treatment of Serious Bacterial Infections: Getting It Right Up Front
- Reprints or correspondence: Dr. Marin H. Kollef, Washington University School of Medicine and Barnes-Jewish Hospital, 660 S. Euclid Ave., St. Louis, MO 63110 (mkollef{at}im.wustl.edu).
Abstract
The treatment of serious bacterial infections is complicated by the fact that time to initiation of effective antimicrobial therapy is a strong predictor of mortality. Therefore, therapy must be initiated before the causative pathogen is identified. However, inappropriate or inadequate initial empirical therapy is associated with increased mortality, morbidity, and length of hospital stay. Initial empirical therapy with broad-spectrum antimicrobials attempts to address this dilemma by “getting it right up front.” The goal is to provide treatment active against the most likely pathogens until culture/susceptibility test results are obtained. After the causative pathogen is identified, streamlining to more-precise therapy of the shortest acceptable duration is implemented. In this way, the risks of death, morbid complications, increased duration of hospital stay (as a result of ineffective initial treatment), and emergence of resistance (due to extended treatment with broad-spectrum agents) are lowered. Improved clinical and economic outcomes after such an approach have been demonstrated.
Serious bacterial infections, such as bacteremia, sepsis, and pneumonia, may be acquired in the community, in hospitals, and in other health care facilities and are associated with high mortality rates if not treated promptly and effectively. In most cases, it is necessary to begin treatment before the pathogen responsible for the infection can be identified. In recent years, recommendations put forth by the Infectious Diseases Society of America, the American Thoracic Society, the Society for Healthcare Epidemiology of America, and the Paul Ehrlich Society of Chemotherapy regarding appropriate treatment of serious bacterial infections have advocated the use of initial empirical therapy with broad-spectrum antimicrobials [1–3]. One goal in minimizing the development of antimicrobial resistance is to optimize the choice and duration of empirical antimicrobial therapy [4]. Thus, the type of care (outpatient, inpatient nonintensive care unit, or inpatient intensive care unit [ICU]), knowledge of local common pathogens and resistance patterns, and patient risk factors for colonization or infection with multidrug-resistant pathogens may be used to guide the selection of the empirical regimen. Subsequently obtained culture results are then used to streamline or de-escalate the initial regimen, in accordance with good antimicrobial stewardship [5]. Improved outcomes after clinical implementation of such an approach have been demonstrated, in comparison with outcomes for patients treated before institutional guideline implementation, in prospective studies of patients with severe sepsis and septic shock [6], ventilator-associated pneumonia (VAP) [7], and severe hospital-acquired pneumonia (HAP) [8].
One recent prospective study of 895 hospitalized, febrile, adult patients compared the 30-day, all-cause mortality rate and mean duration of hospital stay for patients who received inappropriate (n=319) or appropriate (n=576) initial antimicrobial treatment. A significantly higher mortality rate (20.1% vs. 11.8%; P=.001) and longer mean duration of hospitalization (at least 2 days longer at each of 3 medical centers in 3 different countries; P=.002) were observed for those who received inappropriate versus appropriate initial therapy. Inappropriate initial therapy was an independent risk factor for increased mortality and duration of hospitalization [9]. A prospective cohort study comparing ICU patients (N=655) who had community-acquired or nosocomial infections showed a higher in-hospital mortality rate (52.1% vs. 12.2%; P<.001) and infection-related mortality rate (42.0% vs. 17.7%; P<.001) associated with inadequate (n=169) versus adequate (n=486) initial therapy, and inadequate initial therapy was a strong independent determinant of in-hospital mortality [10]. In addition, data from this study demonstrated that prior antimicrobial treatment was an independent risk factor for inadequate initial treatment [10].
Data from a prospective cohort study of 3413 hospitalized patients with bloodstream infections showed a higher in-hospital mortality rate (34% vs. 20%; P=.0001) and a longer median hospital stay for survivors (11 vs. 9 days; P<.05) with inappropriate (n=432) versus appropriate (n=2158) empirical treatment, and inappropriate empirical treatment was an independent risk factor for death [11]. When patients were stratified by specific characteristics (demographic characteristics, underlying disorder, prior antimicrobial treatment, whether the infection was community or hospital acquired, presence or absence of septic shock, temperature, systolic blood pressure, serum creatinine level, serum albumin level, hospital department, site of infection, and pathogen), higher mortality rates with inappropriate therapy were found for all subgroups, except for patients with infections caused by streptococci other than group A and Streptococcus pneumoniae and patients with hypothermia [11]. A retrospective cohort analysis of medical records of 305 hospitalized patients with Pseudomonas aeruginosa bloodstream infections and of the hospital pharmacy database also found a significantly higher rate of in-hospital mortality with inappropriate (n=75) versus appropriate (n=230) initial antimicrobial treatment (30.7% vs. 17.8%; P=.018), and inappropriate initial treatment was identified as an independent determinant of in-hospital mortality [12].
A prospective analysis of 339 critically ill patients admitted to an ICU with community-acquired bloodstream infections and sepsis or septic shock identified inappropriate antimicrobial treatment (n=49) as a strong independent predictor of death, and the effect of inappropriate therapy in terms of reduced survival increased with disease severity [13]. Survival rates were significantly lower with inappropriate versus appropriate treatment in subgroups of patients with sepsis or severe sepsis, as well as among patients with septic shock [13]. Similarly, in a separate prospective study of 904 patients with severe sepsis or early septic shock, inappropriate initial therapy (n=211) was associated with a higher mortality rate than appropriate therapy (39% vs. 24%; P<.001), and inappropriate therapy was shown to be an independent risk factor for increased mortality [14]. Inappropriate initial treatment was also identified as an independent risk factor for in-hospital mortality in a prospective observational cohort study of 102 patients with severe sepsis, with an 87% mortality rate among those receiving inappropriate treatment and a 29% mortality rate among those receiving appropriate treatment (P<.001) [15]. In a prospective study of patients with bacteremia caused by gram-negative bacilli (N=2124), inappropriate initial treatment (n=670) was also associated with a significantly higher mortality rate than appropriate initial treatment (34% vs. 18%; P<.0001) [16]. In a retrospective analysis of 63 pediatric ICU patients with bacteremia or fungemia (n=74 separate episodes), inadequate initial empirical treatment (n=11 episodes) was associated with greater mortality within 7 days of infection and was a significant predictor of infection-related mortality [17].
A single-site, prospective hospital study of 372 patients with bacteremia (n=428 episodes) that found inappropriate empirical antimicrobial treatment (n=77 patients) to be an independent risk factor for mortality also demonstrated that appropriate empirical antimicrobial treatment was more frequently prescribed by infectious diseases specialists (78%) than by other physicians (54%) [18]. Similarly, in a prospective study of 103 patients with bacteremia and systemic inflammatory response syndrome, treatment provided by a hospital infectious diseases service versus that provided by attending physicians was shown to incorporate a significantly higher incidence of optimal initial empirical therapy and therapy de-escalation after culture results were obtained [19]. All antimicrobial regimens were switched from a broad-spectrum to a narrow-spectrum agent by a hospital's infectious diseases service, versus only 50% of regimens being managed by an attending physician (P<.001) [19].
Retrospective assessment of the impact of inappropriate therapy on 133 patients with bacteremia harboring extended-spectrum β-lactamase (ESBL)-producing Escherichia coli or Klebsiella pneumoniae who received appropriate definitive antimicrobial therapy (after culture results were available) showed that inappropriate empirical therapy did not result in a significantly higher 30-day mortality rate than appropriate empirical therapy (18.9% vs. 15.5%, respectively) [20]. In contrast, in a retrospective assessment of 286 patients with bacteremia caused by antimicrobial-resistant gram-negative bacilli, the 30-day mortality rate was significantly higher with inappropriate empirical therapy (38.4%) than with appropriate empirical therapy (27.4%) (P=.049). In a subgroup of 132 patients with a high-risk source of bacteremia (lung, peritoneum, and unknown), inappropriate initial therapy was an independent risk factor for mortality [21]. Results of a retrospective cohort study of 663 patients with E. coli bacteremia showed that the frequency of inappropriate empirical treatment and the frequency of death increased significantly with the number of antimicrobials to which E. coli strains isolated from blood were resistant (P<.001 for both) [22]. In a retrospective cohort study of patients with bacteremia, a delay in appropriate therapy was more common among patients (n=99) infected with ESBL-producing Enterobacteriaceae strains than among patients (n=99) infected with non-ESBL-producing strains (66% vs. 7%, respectively), and multivariable analysis indicated that ESBL production was a significant predictor of mortality (P=.008) and delay in appropriate therapy (P<.001) [23]. In a retrospective cohort analysis of 186 patients with bloodstream infections caused by ESBL-producing Enterobacteriaceae strains, inadequate initial therapy was a significant predictor of mortality, with mortality rates of 59.5% with inadequate therapy and 18.5% with adequate therapy (P<.001) [24].
A retrospective cohort analysis of 549 patients with methicillin-resistant Staphylococcus aureus (MRSA) sterile-site infection (diagnosed by culture from samples of blood; bronchoalveolar lavage fluid; pleural, cerebrospinal, or peritoneal fluid; or muscle) found that inappropriate initial antimicrobial treatment (n=380) was an independent risk factor for in-hospital mortality [25]. Fluoroquinolone resistance was shown to be independently associated with in-hospital mortality in a retrospective study of 193 patients with E. coli or K. pneumoniae infection; further analysis showed that patients infected with fluoroquinolone-resistant organisms were significantly less likely to have received appropriate therapy than were those infected with fluoroquinolone-susceptible organisms (P=.002 for therapy initiated within 24 h; P<.001 for therapy initiated within 48 h) [26].
For patients with VAP (N=530; 565 episodes), the mortality rate attributable to pneumonia-related complications with inappropriate initial empirical antimicrobial therapy was higher than that with appropriate initial empirical therapy (24.7% vs. 16.2%; P=.039), in a 1-year prospective study [27]. Another prospective study of 113 patients found that the pneumonia-related mortality rate with inappropriate versus appropriate initial empirical therapy was 37.0% versus 15.6%, respectively (P<.05); subsequent therapy changes (when initial therapy was found to be inappropriate by means of culture from bronchoscopically obtained samples) allowed clinical resolution in 17 of 27 patients [28]. When lower respiratory tract cultures were used to assess the appropriateness of initial empirical antimicrobial therapy in a prospective study of 130 patients, the in-hospital mortality rate attributable to pneumonia-related complications for patients who had therapy started or changed because of the findings from lower respiratory tract culture (n=51) was significantly higher than the rate for the 51 patients who had no change in their initial antimicrobial therapy (23.5% vs. 7.8%, respectively; P=.018); similar results were obtained for the overall in-hospital mortality rate (60.8% vs. 33.3%, respectively; P=.005). Therapy was discontinued for the other 28 patients [29]. A recent 115-patient prospective study that used an integrative approach to treatment, incorporating consideration of patient factors in the selection of initial empirical therapy, found that the mortality rate among patients who received inappropriate therapy (n=15) was significantly higher than that among patients who received appropriate therapy (n=100) (47% vs. 20%; P=.04) [30].
Time to initiation of effective antimicrobial therapy is a strong predictor of mortality. Studies have shown that survival rates are significantly higher in many settings if treatment is initiated sooner; therefore, appropriate initial empirical therapy is warranted (as opposed to delaying therapy, for example, until culture results are obtained). Results of a retrospective cohort analysis of 157 patients with Candida bloodstream infection demonstrated that initiation of initial empirical therapy 12 h after drawing the first positive blood sample for culture was an independent predictor of in-hospital mortality [31]. A retrospective cohort study found that hospitalized patients (N=187) infected with ESBL-producing Enterobacteriaceae strains for whom antimicrobial therapy was initiated >48 h after the culture sample was obtained had a mortality rate of 21.4%, whereas the group for whom therapy was initiated earlier had a mortality rate of 10.7%. A significant association between increased time to initiation of therapy and increased mortality rate was identified (P<.001). When the analysis was stratified by clinical site of infection, a significant association between increased time to initiation of therapy and increased mortality rate was found for patients with non-urinary tract infections (P=.02) but not for patients with urinary tract infections [32]. In a retrospective cohort study of records for 2154 patients with septic shock, increasing delays in the initiation of effective therapy after the onset of persistent or recurrent hypotension were associated with a significantly (P<.0001) increased risk of death (figure 1), and median time to initiation of effective therapy was a strong predictor of mortality. The survival rate with therapy initiated within the first hour was 79.9%; each additional hour that therapy was delayed resulted in a mean decrease in survival rate of 7.6%. The survival rate with therapy initiated in the sixth hour was 42% [33].
Figure 1
Mortality risk associated with increasing delays in initiation of effective antimicrobial therapy. Adapted from Kumar A, Roberts D, Wood KE, et al. Duration of hypotension before initiation of effective antimicrobial therapy is the critical determinant of survival in human septic shock. Crit Care Med 2006; 34:1589–96, with permission from Lippincott Williams & Wilkins [33].
Patients with VAP also had increased mortality with delays in the initiation of appropriate antimicrobial therapy. Adequate initial empirical therapy was associated with significantly lower mortality rates than inadequate initial empirical therapy (38% vs. 91%, respectively; P<.001), as determined by bronchoalveolar lavage data in a prospective study of 132 patients; however, switching patients receiving inadequate initial empirical therapy to adequate therapy did not have a beneficial effect on mortality rate, which demonstrated the importance of expedient initiation of effective therapy in this patient population [34]. In a separate prospective study of 107 patients, a delay in the administration of appropriate antimicrobial therapy of ⩾24 h after diagnosis of VAP was found to be an independent risk factor for in-hospital mortality, with an in-hospital mortality rate of 69.7% (vs. 28.4% with earlier administration; P<.01) and a mortality rate attributable to pneumonia-related complications of 39.4% (vs. 10.8% with earlier administration; P=.001) [35].
Treatment of Serious Infections with Broad-Spectrum Antimicrobials
An aim of aggressive initial empirical therapy with broad-spectrum antimicrobials is to provide treatment that will be active against the most likely bacterial pathogens until culture/susceptibility test results are obtained. After such results are available, streamlining to more-precise therapy is implemented. This narrowing of therapy involves not only changes in specific agents or doses but also incorporation of the shortest duration of therapy that is clinically acceptable. In this way, the risk of death (due to inappropriate initial treatment) and the potential for emergence of resistance (due to extended treatment with broad-spectrum agents) should be lowered [36]. Another strategy to minimize the emergence of resistance is to incorporate scheduled changes in antimicrobial classes used for initial empirical therapy. This approach significantly lowered the incidence of VAP when a 6-month period of ceftazidime use was followed by a 6-month period of ciprofloxacin use (11.6% vs. 6.7%, respectively; P=.028); results were attributed to a decreased incidence of VAP caused by antimicrobial-resistant gram-negative bacteria in the second 6-month period [37]. In a separate study, the use of scheduled antimicrobial class changes (from ceftazidime to ciprofloxacin to cefepime) was associated with a significant decrease in the rate of administration of inadequate therapy for nosocomial infections with gram-negative organisms (4.4% in period 1; 2.1% in period 2; 1.6% in period 3) [38].
Bloodstream Infections
>Bacteremia. S. aureus is a common cause of bacteremia, and options for treatment of S. aureus infection have been limited, in part, by the emergence of MRSA [39]; thus, empirical therapies should include agents that have activity against MRSA. Results of a recently published open-label, randomized trial showed that daptomycin, a cyclic lipopeptide antimicrobial, was statistically noninferior to a standard regimen of low-dose gentamicin plus antistaphylococcal penicillin (nafcillin, oxacillin, or flucloxacillin) or vancomycin for the treatment of patients (N=235) with S. aureus bacteremia with or without endocarditis. Rates of successful treatment at 42 days after the end of therapy were 44.2% and 41.7% with daptomycin and standard therapy, respectively. Rates of successful treatment did not differ between groups when assessed in predefined strata (baseline diagnosis of definite plus possible endocarditis, final diagnosis of right-sided endocarditis plus complicated bacteremia, and final diagnosis of uncomplicated bacteremia) [40]. Thus, daptomycin may become an alternative therapy for vancomycin. In a prospective, randomized trial of patients (N=381) with methicillin- and fluoroquinolone-susceptible S. aureus bacteremia, the addition of levofloxacin to standard treatment (cloxacillin, dicloxacillin, cefuroxime, clindamycin, vancomycin, cephalexin, or cefadroxil) provided no survival advantage over standard treatment alone [41].
Treatment of bacteremia due to Acinetobacter baumannii was investigated in a retrospective analysis of 48 patients given imipenem/cilastatin or ampicillin/sulbactam. Both regimens provided similar outcomes in terms of median days to resolution of fever or WBC count, end-of-treatment success, and median length of ICU stay [42]. In a separate retrospective analysis of 55 patients with bacteremia caused by multidrug-resistant A. baumannii (resistant to ceftazidime, cefepime, ticarcillin/clavulanate, piperacillin/tazobactam, azetreonam, imipenem, meropenem, gentamicin, amikacin, ofloxacin, and ciprofloxacin), the combination of ampicillin/sulbactam with a carbapenem was superior to the combination of amikacin with a carbapenem and to a carbapenem alone, in terms of 30-day mortality rate [43].
Data from 38 patients with P. aeruginosa bacteremia treated with an aminoglycoside, ciprofloxacin, piperacillin, ceftazidime, gentamicin, tobramycin, or a combination of antimicrobials were used in a retrospective pharmacodynamic analysis that accounted for patient factors and the pharmacokinetic characteristics of the antimicrobials used. Although the overall clinical cure rate was 58%, results of the pharmacodynamic analysis of treatment with aminoglycosides and ciprofloxacin showed that the probability of cure was ⩾90% when the peak plasma drug concentration/MIC ratio was ⩾8, demonstrating the importance of pharmacodynamic data in guiding future empirical treatment decisions [44].
Carbapenems and cefepime are generally effective for the treatment of serious infections caused by ESBL-producing bacteria [45]. In a prospective study of patients (N=71) with ESBL-producing K. pneumoniae bacteremia (N=85 episodes), the 14-day mortality rate associated with inappropriate treatment was 63.6%, and that associated with isolate-susceptible treatment was 14.1%. The 14-day mortality rate among patients who received a carbapenem within 5 days after the first blood culture result was obtained was 4.8%; in contrast, the mortality rate among patients who received noncarbapenem antimicrobials was 27.6% (P=.012). Use of a carbapenem was an independent predictor of decreased mortality [46]. For patients with ESBL-producing E. coli or K. pneumoniae bacteremia (N=133), 30-day mortality rates were lowest among those who received a carbapenem (12.9%) or ciprofloxacin (10.3%), and administration of a broad-spectrum cephalosporin (cefotaxime, ceftriaxone, ceftizoxime, or ceftazidime) as definitive antimicrobial therapy was an independent risk factor for mortality, as identified in a retrospective analysis [20]. In a prospective cohort study of 43 patients with bacteremia caused by E. coli producing ESBL, which in 70% of patients was of a CTX-M type, the mortality rate among those who received empirical treatment with a β-lactam/β-lactamase inhibitor or a carbapenem regimen was 9%, compared with a mortality rate of 35% among those who received empirical treatment with cephalosporin or a fluoroquinolone. In addition, 78% of patients who received empirical treatment with a cephalosporin or fluoroquinolone required changes in therapy after culture results became available, compared with 24% of patients who received empirical treatment with a β-lactam/β-lactamase inhibitor or a carbapenem (P=.001) [47]. For bacteremia caused by ESBL-producing bacteria other than E. coli and Klebsiella species, a retrospective evaluation of 54 patients who received carbapenem or noncarbapenem regimens found no significant differences in overall survival rates with ciprofloxacin (70%) and carbapenems (72.7%), leading the investigators to suggest that ciprofloxacin may be considered as an alternate therapy in this setting [48].
>Severe sepsis and septic shock. Results of studies investigating various regimens for the empirical treatment of sepsis with broad-spectrum antimicrobials are summarized in table 1. In general, monotherapy with a range of antimicrobials showed activity in patients with sepsis, and, in studies comparing monotherapy with combination therapy, combination therapy provided no greater benefit than monotherapy.
Figure 2
Improved survival probability after implementation of a standardized hospital order set for the emergency department. Filled circles denote patients with septic shock managed in the “before” period, and open circles denote patients with septic shock managed in the “after” period (P <0.01, log-rank test). Reprinted from Micek ST, Roubinian N, Heuring T, et al. Before-after study of a standardized hospital order set for the management of septic shock. Crit Care Med 2006; 34:2707-13 [6], with permission from Lippincott Williams & Wilkins.
Figure 3
Practice algorithm for providing appropriate initial antimicrobial treatment. HAP, hospital-acquired pneumonia; MDR, multidrug resistant; VAP, ventilator-associated pneumonia. Adapted from Kollef MH, Micek ST. Strategies to prevent antimicrobial resistance in the ICU. Crit Care Med 2005; 33:1845–53 [77], with permission from Lippincott Williams & Wilkins.
Table 1
Empirical treatment of sepsis with broad-spectrum antimicrobials.
As illustrated in figure 1, for critically ill patients, the physician has very little time to determine the appropriate treatment, because delays in treatment initiation increase the risk of death [33]. Results of a “before-after” study [6] that assessed the value of implementing a standardized hospital order set for emergency department management of patients with severe sepsis and septic shock demonstrated that this type of approach to therapy can help to optimize outcomes. The institution protocol focused on hemodynamic resuscitation and appropriate initial antimicrobial therapy. Patients who received care after implementation of the standardized order set (n=60) were more likely than patients who received care before implementation (n=60) to have received antimicrobials within 3 h of arrival in the emergency department (86.7% vs. 60%, respectively; P=.001) and to have received appropriate initial antimicrobial treatment (86.7% vs. 71.7%; P=.043). In addition, patients who were seen before implementation of the standardized order set had received a greater total volume of intravenous fluids in the emergency department before transfer to the ICU (3789 vs. 2825 mL; P=.002), had a significantly shorter mean length of hospital stay (8.9 days vs. 12.1 days; P=.038), and had a significantly lower risk of 28-day mortality (30% vs. 48.3%; P=.040) [6]. A Kaplan-Meier analysis of survival for the “before” and “after” groups is shown in figure 2 [6].
>Fungemia. Fungal bloodstream infections are common ICU infections that are associated with high mortality rates [54]. In the retrospective cohort analysis of patients with Candida bloodstream infection (described above), initial empirical therapy was achieved for only 9 of 157 patients within 12 h of collection of the first blood sample for positive culture and was shown to be important in reducing the risk of in-hospital mortality [31]. Thus, more-rapid diagnostic techniques and better ways to identify patients that may have fungal bloodstream infections are needed, so that appropriate empirical treatment may be initiated sooner. In addition, a method to discriminate between patients infected with Candida albicans and patients infected with other Candida species would aid in the selection of appropriate initial empirical therapy, because non-C. albicans species of Candida have variable sensitivity to fluconazole, which is the most commonly used agent for initial therapy for fungemia [55]. Potential clinical predictors of pathogens responsible for fungal bloodstream infection were investigated in a recent retrospective case series of critically ill and noncritically ill patients; however, none of the patient demographic and clinical variables studied were able to discriminate between patients infected with C. albicans and patients infected with other Candida species [55].
Voriconazole, a derivative of fluconazole, is a broad-spectrum antifungal agent with efficacy similar to that of fluconazole for Candida infections and to that of amphotericin B for Aspergillus infections; therefore, it may be a useful alternative, particularly against organisms that are resistant to these agents [56]. Newer antifungal agents that have become available recently are the echinocandins (caspofungin, micafungin, and anidulafungin), which have fungicidal activity against yeast and fungistatic activity against mold [54].
Pneumonia
Community-acquired pneumonia (CAP) is typically diagnosed on the basis of symptoms, including fever, leukocytosis, and lung infiltrates [57]. Nosocomial pneumonia, which is defined as pneumonia that develops at least 48 h after hospital admission and is not incubating at the time of admission, includes HAP and VAP and is the second most common nosocomial infection seen in adult patients in medical ICUs (after urinary tract infection) [58]. Nosocomial pneumonia is more difficult than CAP to diagnose because of the many potential causes of fever, leukocytosis, and lung infiltrates in hospitalized patients [57]. However, because of the high associated mortality rates, individuals suspected of having HAP or VAP are candidates for empirical antimicrobial therapy, and more-invasive techniques, such as bronchoalveolar lavage, are used for diagnosis. Health care-associated pneumonia is pneumonia in any patient who was hospitalized in an acute-care hospital for at least 2 days within 90 days of infection; lived in a nursing home or long-term-care facility; received intravenous antimicrobial therapy, chemotherapy, or wound care within 30 days of infection; or attended a hospital or hemodialysis clinic. It is considered to be in the same spectrum as HAP and VAP, requiring treatment for multidrug-resistant pathogens [2].
Empirical, broad-spectrum antimicrobial treatment of nosocomial pneumonia has been investigated with a variety of regimens. Carbapenem-based regimens have shown efficacy in numerous studies [52, 59–61]. In comparisons of meropenem with imipenem/cilastatin, clinical and microbiological response rates were similar [52, 60, 61]. Addition of netilmicin to imipenem did not improve clinical or microbiological outcomes, compared with those achieved with imipenem alone [49]. A head-to-head comparison of imipenem/cilastatin and piperacillin/tazobactam found that rates of clinical failure and deaths due to infection were similar with both regimens; however, in the subset of infections due to P. aeruginosa, piperacillin/tazobactam was associated with a significantly lower rate of clinical failure [62]. No significant differences in efficacy were observed in comparisons between imipenem/cilastatin and ciprofloxacin [63] or in comparisons between imipenem/cilastatin (followed by ciprofloxacin) and levofloxacin [64]. Meropenem was shown to be superior to ceftazidime plus amikacin in a study of patients with VAP [65] and superior to ceftazidime plus tobramycin in a study of patients with HAP [66]. Compared with ceftazidime plus amikacin combination therapy, piperacillin/tazobactam plus amikacin showed similar rates of clinical cure, improvement, and failure [67]. A comparison of empirical therapy with linezolid and vancomycin, each in combination with aztreonam, showed similar rates of clinical cure and microbiological success [68].
For treatment of CAP, imipenem/cilastatin and meropenem have shown similar clinical and bacteriological response rates [69], and meropenem has demonstrated efficacy similar to that of ceftazidime plus amikacin [70]. Clinical and bacteriological response rates with ceftazidime monotherapy did not differ significantly from those with cefepime monotherapy [71]. Monotherapy with cefditoren pivoxil achieved clinical cure rates, microbiological cure rates, and overall eradication rates comparable to those associated with amoxicillin/clavulanate [72]. Ciprofloxacin, as monotherapy or in combination with a β-lactam, showed efficacy similar to that of standard monotherapy with a β-lactam or standard combination therapy with an aminoglycoside plus a β-lactam [73].
A recent assessment of characteristics of patients with VAP and of VAP-management practices in 20 ICUs across the United States revealed wide variation in treatment patterns (>100 different initial empirical antimicrobial treatment regimens) and a lack of adjustment of initial therapy (de-escalation or escalation) in 61.6% of cases. The average duration of mechanical ventilation before diagnosis was 7.3 days. The most common infecting pathogens were MRSA, P. aeruginosa, and other Staphylococcus species, and the most commonly prescribed regimens were cefepime and a ureidopenicillin/monobactam combination. For >25% of patients, appropriate initial therapy was initiated within 4 h of presumed diagnosis, although initiation of appropriate therapy after 24 h occurred for >25% of patients as well; the average duration of therapy was 11.8 days, and the overall mortality rate was 25.1%. For patients with a time to initiation of appropriate therapy of >24 h, the mortality rate was 30.9%. For patients with therapy de-escalation, the mortality rate was 17.1%, which was significantly lower than the rate of 23.7% for patients without therapy adjustment (P=.001) [74]. These observations emphasize the need for prompt initiation of appropriate therapy and suggest that implementation of therapy de-escalation may provide survival benefits.
Results of a “before-after” study of 102 patients with VAP [7] showed that implementation of clinical guidelines can help to optimize initial therapy. The guidelines were based on the most common pathogens causing VAP in the institution's ICU and on local susceptibility patterns, in an effort to increase the probability of providing adequate initial empirical therapy. The guidelines mandated that treatment be modified after 24–48 h (on the basis of culture results and patient clinical course) and recommended that treatment be limited to 7 days, unless persistent signs and symptoms consistent with active infection were present, in an effort to reduce unnecessary broad-spectrum antimicrobial exposure. Patients who were treated after guideline implementation (n=52) had a significantly greater likelihood of receiving adequate initial antimicrobial therapy (94.2% vs. 48%; P<.001) and a significantly shorter mean duration of therapy (8.6 vs. 14.8 days; P<.001) than did patients who were treated before guideline implementation (n=50). The incidence of a second episode of VAP was higher in the “before” period (24%) than in the “after” period (7.7%) (P=.03), although no significant differences in in-hospital mortality or length of stay were observed.
Another “before-after” study [8] showed that implementation of guidelines provided significant benefits for patients with HAP. The guidelines were developed with consideration of the institution's local antimicrobial resistance patterns, and antimicrobial use after 3 days was based on culture results and incorporated de-escalation measures. The “before” group (n=48) was given a range of initial therapies, whereas the “after” group (n=58) received an imipenem/cilastatin-based regimen. Patients treated after guideline implementation had a significantly greater likelihood of receiving adequate initial antimicrobial therapy (81% vs. 46%; P<.01) and a significantly lower 14-day mortality rate (8% vs. 23%; P=.03) than did patients who were treated before implementation.
Advantages of Initial Therapy with Broad-Spectrum Antimicrobials
As described above in detail, numerous studies of patients with a variety of serious infections have shown that initiating inappropriate empirical antimicrobial therapy or delaying the initiation of appropriate antimicrobial therapy is associated with increased mortality. In addition, studies have shown that the implementation of guidelines for early initiation of appropriate empirical antimicrobial therapy improves survival and may decrease hospital stay. Although further investigation is needed, results of studies that have assessed the potential for the emergence of antimicrobial-resistant bacteria after implementation of initial empirical therapy with broad-spectrum antimicrobials are instructive. For instance, in the “before-after” study of patients with HAP described above [8], use of imipenem/cilastatin as initial empirical therapy was not associated with an increase in the number of isolates of imipenem-resistant bacteria over the 2-year course of the study. Similarly, in surgical ICU patients with sepsis, use of imipenem/cilastatin plus gentamicin for 72 h as initial empirical therapy did not lead to the emergence of imipenem-resistant bacteria after 19 months of use [50]. Adherence to a strict policy of therapy de-escalation after culture results are available should help to reduce the risk of patients developing infection with resistant bacteria.
Use of a standardized approach can increase the likelihood of providing appropriate empirical therapy, which is particularly important when rapid initiation of treatment is crucial to survival. In addition, the use of a standardized approach may aid in the improvement of future outcomes by identifying other risk factors for mortality. For example, as shown in table 2, the “before-after” study of 120 patients with severe sepsis and septic shock identified an inability to achieve 20 mL/kg intravenous fluid administration before initiation of vasopressor therapy as a significant risk factor for 28-day mortality [6]. This information may benefit future patients who present with sepsis.
Table 2
Independent risk factors for 28-day mortality.
A retrospective analysis of the previously discussed “before-after” study [6] of patients with severe sepsis was performed to explore the possible economic benefits of guideline implementation. In the “before” period, the median per-patient total hospitalization cost was $21,985, whereas the cost was $16,103 in the “after” period, representing a significant cost savings (P=.008). Much of this difference was attributed to ICU bed-day and ward bed-day costs. When only survivors were assessed, the cost advantage in the “after” period remained; median costs for the “before” and “after” periods were $21,926 and $13,663, respectively (P=.002) [75]. Economic advantages were also observed with implementation of initial empirical therapy with broad-spectrum agents in accordance with institutional guidelines in a retrospective analysis of 214 patients with bacteremia. The guidelines included requirements for the adjustment of therapy after culture results became available. Therapy was adjusted in 88% of the cases that required adjustment on the basis of culture results, and a 7-day course of adjusted therapy provided a cost savings of 23%, compared with a 7-day course of unadjusted empirical therapy [76].
Summary
Initial empirical therapy with broad-spectrum antimicrobials is a treatment strategy for severe antimicrobial infections that is aimed at “getting it right up front.” With an awareness of local pathogen prevalence and resistance profiles, as well as a consideration of patient clinical characteristics, the physician can implement an initial empirical antimicrobial regimen that is likely to be active against the probable causative pathogen, thereby decreasing the risk of death and the potential for complications and longer hospital stay (and associated costs) that are associated with inappropriate initial therapy. Modification of broad-spectrum coverage when the causative pathogen is identified decreases the duration of broad-spectrum antimicrobial exposure, minimizes the potential for the emergence of resistance, and provides additional cost savings. A practice algorithm incorporating these concepts is shown in figure 3 [77] and may be used as the basis for developing local institutional guidelines.
Acknowledgments
I thank Stephanie M. Leinbach, who provided freelance medical writing support, and Judy E. Fallon from Scientific Connexions, who provided editing assistance, both funded by AstraZeneca.
>Supplement sponsorship. This article was published as part of a supplement entitled “Update on the Appropriate Use of Meropenem for the Treatment of Serious Bacterial Infections,” sponsored by AstraZeneca LP.
>Potential conflicts of interest. M.H.K. is on the speakers' bureau for Pfizer, Merck, Johnson & Johnson, and AstraZeneca; has received recent research funding from Pfizer and Bard Medical; has received grants from Merck, AstraZeneca, and Pfizer; and is a consultant to Bard Medical.
- © 2008 by the Infectious Diseases Society of America
References
This Article
- AbstractFree
- » Full Text (HTML)Free
- Full Text (PDF)Free
Classifications
Services
Citing Articles
Google Scholar
PubMed
Related Content
Share
- Email this article
-
CiteULike
Connotea
Delicious
Facebook
Google+
Mendeley
Twitter
Published on behalf of
IDSA Members: For your free access to this journal, log in via the IDSA members area.
The Journal
Impact Factor: 8.186
5-Yr impact factor: 7.898
For The Media
For Authors
- Services for authors
- Self-archiving Policy
- IDSA Conflict of Interest policy
- Instructions for Authors
- ICMJE Form
- Submit Now!
Open access options for authors visit Oxford Open
This journal enables compliance with the NIH Public Access Policy
Corporate Services
Alerting Services
Related IDSA Publications
