You are on page 1of 16

Intensive Care Med (2001) 27: S 33S 48

Pierre-Yves Bochud Michel P. Glauser Thierry Calandra

Antibiotics in sepsis

P.-Y. Bochud M. P. Glauser T. Calandra Division of Infectious Diseases, Centre Hospitalier Universitaire Vaudois, Lausanne, Switzerland

Introduction
The management of patients with sepsis, severe sepsis, or septic shock requires an integrated approach combining the use of rigorous diagnostic measures and the rapid initiation of appropriate antimicrobial therapy and supportive care. Antimicrobial therapy remains the cornerstone of therapy of patients with sepsis. However, drainage of abscesses and removal of infected foreign material or necrotic tissues are also of critical importance for recovery. In recent years several review articles have been published on the treatment of patients with severe sepsis and septic shock [1, 2, 3]. However, only few have focused specifically on the antimicrobial aspect of the patient management. Therefore the aim of the present article was to use an evidence-based approach to review the literature on antimicrobial therapy for severe sepsis and septic shock. In reviewing the literature on this topic we were rapidly confronted with two difficulties. One was the lack of standard definitions of sepsis, severe sepsis and septic shock. Until the publication of the definitions of the Consensus Conference of the American College of Chest Physicians and the Society of Critical Care Medicine in 1992 [4], the terms sepsis, severe sepsis, and septic shock were ill-defined and often employed interchangeably. The other was the surprising paucity of large comparative studies on the efficacy and safety of different antimicrobial regimens in nonneutropenic patients. Whereas the initial studies on treatment of bacteremias published in the 1960s and 1970s included a majority of nonneutropenic patients, most of the recent large, clinical trials have been conducted in neutropenic

cancer patients. This emphasis is probably due to the high frequency of bloodstream infections in neutropenic cancer patients and the severely compromised host defenses in the context of neutropenia, providing stringent conditions for testing antimicrobial agents. Therefore treatment guidelines for patients with severe sepsis and septic shock have been based for the most part on the results of large, multicenter studies conducted in neutropenic cancer patients [5, 6]. Treatment guidelines for the use of antimicrobial agents in the neutropenic host have been published recently [7], but these are unlikely to apply to patients with severe sepsis or septic shock. Moreover, neutropenic cancer patients account for a minority of patients with severe sepsis or septic shock and have often been excluded from septic shock trials [8, 9]. Hence the need to review, using an evidence-based approach, the literature on antibiotic therapy for patients with severe sepsis and septic shock.

Methods
Data source Medline was used to search articles published between 1966 and October 1999. Keywords were the generic Medical Subject Heading (MeSH) terms sepsis, anti-infective agents, and clinical trials. Sepsis comprised the terms septicemia, sepsis syndrome, septic shock, bacteremia, fungemia, parasitemia, and viremia. Anti-infective agents comprised the term antibiotics which was exploded to include all classes of antibiotics, and all antibiotic names. Clinical trial was defined as a pre-planned clinical study of the safety, efficacy, or optimum dosage schedule, of one or more diagnostic, therapeutic or prophylactic drugs, devices, or techniques in humans selected according to predetermined criteria of eligibility and observed for predefined evidence of favorable and unfavorable effects. The MeSH keyword agranulocytosis was used to exclude studies of neutropenic patients. Additional articles were retrieved from review articles or from the reference list of articles identified by the Medline search. Epidemiological data were ex-

S 34

Fig. 1 A, B Etiology of infections in patients with severe sepsis and septic shock. A 19631987: Data are derived from 3 studies [2224] that included 674 patients. B 19881998: Data are derived from 18 studies [8, 9, 1114, 2536] that included 8,988 patients

tracted from articles identified by a Medline search using the keywords epidemiology and sepsis, and by a systematic review of 27 clinical trials of anti-inflammatory or mediator-targeted therapies in patients with severe sepsis and septic shock [10, 11, 12, 13, 14]. Selection of articles Abstracts of all articles meeting the selection criteria were reviewed to exclude irrelevant studies. Review articles and articles on topics such as antibiotic prophylaxis, pharmacology, microbiology, oncology, hematology, immunology, mediators of inflammation, allergy, catheter management, animal studies, chronic infections, and specific infections (AIDS, endocarditis, chronic salmonellosis, viral infections in organ transplant patients, hemorrhagic fever, viral hepatitis, parasitic infections, and malaria) were excluded if they did not satisfy the inclusion criteria. Articles were selected only if there was unequivocal evidence that patients had clinically or microbiologically documented infections, and if the study met at least one of the following criteria: (a) a definition of sepsis or severe sepsis consistent with the definition of the Consensus Conference of the American College of Chest Physicians and the Society of Critical Care Medicine [4], (b) sepsis with at least one organ dysfunction or sign of hypo-perfusion present in more than 50 % of the patients, or (c) an overall mortality greater than 10 %. This cutoff was chosen because it represents the lower end of the mortality range of patients with the sepsis syndrome [15]. Medline search and selection of articles was done by one reviewer (P. Y.B.). To ensure that articles had been properly selected a random sample of 25 % of the articles identified by the Medline search were examined by a second reviewer (T. C.). Agreement between the two reviewers was assessed using the k test [16]. There was 91 % overall agreement between the two reviewers. The k statistic was 0.75 indicating that there was substantial agreement between the two reviewers. Articles for which there was disagreement were discussed to reach consensus. Levels of evidence and graded responses to questions were assessed following the criteria proposed by Sackett [17].

Epidemiological features of severe sepsis and septic shock


Micro-organisms As shown in Fig. 1A, Gram-negative bacteria caused the majority of bloodstream infections in the 1960s and early 1970s [18, 19, 20, 21, 22, 23, 24]. This trend persisted through the middle 1980s, when the proportion of infections caused by Gram-positive bacteria began to increase. Recent data derived from three epidemiological studies and 15 clinical trials of anti-inflammatory agents conducted between 1988 and 1998 are summarized in Fig. 1B [8, 9, 11, 12, 13, 14, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36]. Standard definitions of severe sepsis and septic shock were used in all these studies. A pathogen was identified in 71 % of the patients. On average, blood cultures were positive in 34 % of the patients, ranging between 9 % and 64 %. The most striking finding was that Gram-positive infections were almost as frequent as Gram-negative infections, confirming a trend reported in many studies [9, 13, 25, 26, 37]. In fact, cases of Gram-positive bacteremia outnumbered Gram-negative bacteremia in some studies [12, 32, 33, 34]. In the present compilation of articles, Gram-positive bacteria and Gram-negative bacteria accounted for 34 % and 42 % of the infections, respectively. Mixed bacterial infections occurred in 14 % of the patients. Half of the Gram-positive infections were caused by staphylococci (Staphylococcus aureus: 12 % and coagulase-negative staphylococci: 7 %). Enterococci were isolated in 8 % of the patients and pneumococci in 4 %. Most Gram-negative infections were caused by Enterobacteriaceae (29 %) with Escherichia coli (13 %) and Klebsiella pneumoniae (8 %) being the most frequent enteric pathogens. Pseudomonas aeruginosa (8 %) was the third most frequent agent of Gram-

S 35

Fig. 2 A, B Sites of infections in patients with severe sepsis and septic shock. A 19631987: Data are derived from 2 studies [22, 23] that included 585 patients. B 19881998: Data are derived from 7 studies [9, 1114, 26, 36] that included 5,423 patients

negative sepsis. Fungi, mainly Candida species, were the causal agent of sepsis in 5 % of the patients. The number of fungal infections are also increasing. Candida was the fourth most common bloodstream pathogen in all recent studies of nosocomial bloodstream infections in the United States [38], and it outnumbered all Gramnegative bacteria [39]. Anaerobes were isolated in a minority of patients (2 %). Sites of infection The identification of the primary site of infection is a critical part of the work-up of the septic patient. Together with the Gram stain of specimens obtained from any site suspected of infection, it is probably the single most important information in guiding the choice of antibiotic therapy. A site of infection was identified in 92 % of 2803 patients included in nine studies [14, 18, 23, 27, 35, 40, 41, 42, 43]. Over the past 30 years significant changes have occurred in the relative frequencies of the site of infections in septic patients. The abdominal cavity and the urinary tract were the most frequent sites of infections (27 % and 21 %, respectively) in studies conducted in the 1960s and 1970s (Fig. 2A) [21, 23]. Data derived from seven studies performed between 1988 and 1998 revealed that lung infections were predominant, accounting for 36 % of all infections (Fig. 2B) [9, 11, 12, 13, 14, 26, 36]. Primary bloodstream infections (i.e., without any other source of infection) were recorded in 20 % of the patients. Abdominal infections and urinary tract infections were the third and fourth most common foci of infections. Similar findings have been reported recently [44].

Discussion: literature-based recommendations


Does appropriate antimicrobial therapy improve the outcome of patients with bloodstream infections and severe sepsis or septic shock in patients with Gramnegative bacteremia? Answer: yes, grade D. Recommendation Retrospective studies have shown that early administration of appropriate antibiotics reduces the mortality in patients with bloodstream infections caused by Gramnegative bacteria. Rationale For obvious reasons there have never been prospective, randomized, controlled trials on the impact of antibiotic treatment versus no treatment on the outcome of patients with sepsis. Several retrospective studies conducted in the 1960s and in the 1970s showed that appropriate antimicrobial therapy, defined as the use of at least one antibiotic active in vitro against the causative bacteria, leads to lower mortality among patients with Gram-negative bacteremia than among similar patients receiving inappropriate therapy (Table 1) [19, 20, 22, 24]. The landmark study by McCabe and Jackson [22] included 173 patients with Gram-negative bacillary bacteremia, who were classified in three categories based on the severity of the underlying disease categories (i.e., rapidly fatal, ultimately fatal, and nonfatal). Hypotension (i.e., a blood pressure of less than 90/60 mmHg or a decrease of more than 70 mmHg in a hypertensive patient) was present in 37 % patients and the overall mortality was

S 36

Table 1 Impact of the appropriateness of antibiotic therapy on the mortality of Gram-negative bacteremia

Mortality with appropriate Mortality without antibiotics appropriate antibiotics Category of underlying disease McCabe and Jackson [22] Rapidly fatal Ultimately fatal Nonfatal Total Freid and Vosti [20] Rapidly fatal Ultimately fatal Nonfatal Total Bryant et al. 1971 [19] Rapidly fatal Ultimately fatal Nonfatal Total Young et al. 1977 [24] Rapidly fatal Ultimately fatal Nonfatal Total Four studies combined Rapidly fatal Ultimately fatal Nonfatal Total n 8/10 10/22 0/49 18/81 21/25 33/78 14/109 68/211 12/14 19/49 11/95 42/158 41/49 62/149 25/253 128/451 82/98 124/289 50/506 256/902 % 80 45 0 22 84 42 13 32 86 39 12 27 84 42 10 28 84 42 10 28 n 2/2 10/16 3/13 15/31 10/11 9/14 9/33 28/58 5/7 13/18 10/35 28/60 17/20 32/48 22/71 71/139 34/40 64/96 44/152 142/288 % 100 63 23 48 91 64 27 48 71 72 29 47 85 67 31 51 85 67 29 49 p* NS NS 0.001 0.007 NS NS 0.05 0.02 NS 0.02 0.02 0.005 NS 0.003 < 0.001 < 0.001 NS < 0.001 < 0.001 < 0.001

*p values are based on the c2 test

30 %, suggesting that most patients probably presented with severe sepsis or septic shock. Appropriate antibiotic therapy was linked with a reduction in mortality from 48 % to 22 %. Subsequent studies gave similar results. In the study by Freid and Vosti [20] the mortality rate was 32 % in patients who had been treated with appropriate antibacterial agents, compared to 48 % in those who did not receive adequate antibiotics. Proper antibiotic therapy was associated with a reduction in mortality from 47 % to 27 % in a retrospective analysis of 218 patients with Gram-negative rod bacteremia by Bryant et al. [19]. In the study by Young et al. [24] that included 451 patients with Gram-negative rod bacteremia, appropriate antibiotic treatment was also found to reduce mortality from 51 % to 28 %. Of note, the impact of appropriate antibiotic treatment on patients' outcome was shown to be statistically significant in patients with nonfatal or ultimately fatal diseases, but not in patients with rapidly fatal diseases. Moreover, in a review of 612 episodes of Gram-negative bacteremia Kreger et al. [45] showed that prompt administration of proper empirical antimicrobial therapy reduced by half the frequency with which shock developed in patients with rapidly fatal, ultimately fatal, or nonfatal diseases.

Does appropriate antimicrobial therapy improve the outcome of patients with bloodstream infections and severe sepsis or septic shock in patients with Grampositive bacteremia? Answer: yes, grade E. Recommendation By analogy with the observations made in patients with Gram-negative sepsis and despite the lack of substantial clinical data in the literature, it is likely that appropriate antibiotic therapy reduces the morbidity and the mortality of Gram-positive sepsis. Rationale While the percentage of patients treated with inappropriate antibiotics ranged between 27 % and 38 % in the initial studies on Gram-negative bacteremias [19, 20, 22], it was less than 15 % in recent trials in which Grampositive bacteria predominated [34, 46]. This explains why there are almost no data on the impact of appropriate antibiotic therapy in patients with Gram-positive sepsis. Clinical studies comparing antibacterial therapy

S 37

to no therapy would be unethical. However, the emergence of multiresistant Gram-positive bacteria may give us an opportunity to address that question. In fact, clinical and microbiological success rates were evaluated in 20 patients with severe infections due to vancomycinresistant Enterococcus faecium [47]. That study compared patients treated with quinupristin-dalfopristin with 40 historical controls treated with other agents, mostly vancomycin. The mortality directly attributable to infection was lower in the quinupristin-dalfopristin group (5 of 20, 25 %) than in the control group (17 of 42, 40 %), suggesting that appropriate antibiotics reduced mortality (p = 0.27, two-tailed Fisher's exact test). Does appropriate antimicrobial therapy improve the outcome of patients with bloodstream infections and severe sepsis or septic shock in patients with candidemia? Answer: yes, grade D. Recommendation Antifungal therapy is recommended for patients with candidemia. An international panel of experts who participated in a consensus conference on the management and prevention of severe Candida infections made similar recommendations [48]. Whether early treatment is associated with better outcome is unknown, and additional studies are needed to evaluate this question. Rationale Candidemia may cause significant morbidity and serious long-term sequelae and is associated with mortality rates in the range of 4060 %. In a large, multicenter, prospective observational study of 427 patients with candidemia 369 patients were treated with antifungal agents, while 58 patients did not receive antifungal therapy, for unknown reasons [49]. The mortality rates after 14 days and after 30 days were 27 % (99 of 369) and 37 % (136 of 369) in patients who received antifungal therapy and 74 % (43 of 58) and 76 % (44 of 58) in those who did not (p < 0.001), suggesting that antifungal therapy reduced mortality. However, factors other than antifungal therapy may also have contributed to these differences as the proportion of critically ill and cancer patients was higher in the untreated group than in the group who benefited from antifungal therapy. Of the 369 patients 319 received early therapy (i.e., treatment started within 72 h of the first positive blood cultures), while 50 patients were treated more than 3 days after documentation of fungal sepsis. Early administration

did not improve survival, even when patients were stratified by antifungal agents or severity of illness. In contrast, in a study of 46 patients with candidemia, early antifungal therapy (i.e., treatment initiated 48 h or sooner after the onset of candidemia) was found to improve survival (p = 0.06) [50].

Monotherapy versus combination therapy


Historical background and rationale Following the demonstration that early, appropriate antibiotic treatment of Gram-negative bacteremia improves patients' outcome and prevents the development of septic shock, investigators examined whether treatment with two active antibiotics instead of one would improve outcome further. Several arguments would support the use of antibiotic combinations. First, combination therapy broadens the antibacterial spectrum, which might be important, since treatment is usually initiated empirically in the critically ill patient with sepsis. Moreover, polymicrobial infections may occur, especially in patients with intra-abdominal or pelvic infections, and two antibiotics may also help to cover a broader range of pathogens. Second, a combination of two antibiotics may exert additive or synergistic effects against the infecting organism, resulting in enhanced antibacterial activity and possibly improved clinical response [51, 52, 53]. Theoretically, synergism may also allow the use of a reduced dose of the most toxic of the two agents, but this is rarely done in practice. Third, the use of a combination of antibiotics has been shown to reduce the emergence of resistant bacteria [54] and the incidence of superinfections [55]. The potential advantages of combination therapy were first assessed in a series of retrospective studies performed in the late 1960s and early 1970s. In an analysis of 444 episodes of bacteremias Anderson et al. [56] found that treatment with two antibiotics active against the causative organism was not superior to treatment with one single antibiotic. However, in a subgroup of patients with rapidly or ultimately fatal diseases the use of synergistic as opposed to nonsynergistic antibiotic combinations reduced mortality from 78 % to 52 % (p < 0.005). In a review of 612 episodes of Gram-negative bloodstream infections, Kreger et al. [45] reported equivalent mortality rates in patients treated with antibiotic combinations or with one single agent (22 % vs. 21 %). However, in a small subset of patients with rapidly fatal disease mortality was 23 % in patients treated with two antibiotics, compared to 50 % in those treated with a single antibiotic. However, these early studies provide limited information relevant to the management of patients today. These studies were retrospective and did not include multivariate analyses to take into

S 38

Table 2 Studies comparing carbapenem monotherapy with a combination of a b-lactam and an aminoglycoside as empirical therapy of severe sepsis (IMI imipenem, Cx+A cefotaxime + amikacin, I+N Clinical success n Mouton et al. [63] (n = 140) IMI Cx+A Cometta et al. [62] (n = 280) IMI I+N Solberg and Sjursen [65] (n = 53) MER Cd  A % Overall mortality n %

imipenem + netilmicin, MER meropenem, Cd+A ceftazidime  amikacin, n.a. not available) Superinfections n % Relapse n % Eradication n % Nephrotoxicity n %

Mortality due Colonization to infection n % n %

58/70 54/70

83 77

7/70 7/70

10 10

3/70 2/70

4 3

7/44 6/45

16 13

4/44 7/45

9 16

n.a.

19/44 15/45

43 33

1/105 4/106

1 4

113/142 80 119/138 86

n.a.

18/142 13 13/138 9

8/142 6 13/138 9

8/148 5 11/138 8

n.a.

n.a.

0/158 6/155

0 4*

56/61 66/70

92 94

n.a.

n.a.

n.a.

0/37 1/45

0 2

1/37 0/45

3 0

n.a.

n.a.

Mouton and Beuscart [64] (n = 237) MER 97/111 C+A 98/118 *p = 0.04

87 83

7/116 8/121

6 7

n.a.

n.a.

n.a.

11/116 14 11/121 14

13/116 11 10/121 8

n.a.

account the role played by confounding factors likely to affect mortality. Subgroups analyses also included a relatively small number of patients. Most importantly, the majority of the antibiotics used at that time would no longer be considered appropriate today. Subsequent studies performed in the late 1970s and early 1980s then evaluated the efficacy of various antibiotic combinations for the treatment of Gram-negative infections, most frequently a b-lactam and an aminoglycoside. Combinations of penicillin or carbenicillin with amikacin showed similar clinical efficacy (55 % and 63 %, respectively) as empirical therapy of severe Gram-negative infections in nonneutropenic cancer patients [57]. Likewise, ticarcillin plus sisomycin or mezlocillin plus sisomycin were found to be equally effective for the treatment of Gram-negative sepsis [58]. Clinical response rates were higher when patients were treated with synergistic antibiotic combinations or when peak serum bactericidal activity were greater than 1:8. More recently 153 adult ICU patients with nosocomial pneumonia or bacteremia were randomized to receive either low- or high-dose isepamicin or amikacin given in combination with ceftazidime [59]. An unknown number of patients were treated with imipenem instead of ceftazidime. Clinical response rates were

comparable in patients with nosocomial pneumonia and in those with bacteremia. The proportion of patients experiencing at least one adverse event was similar in the three treatment groups. Some investigators went even further, treating septic patients with three instead of two antibiotics [43, 60]. However, intensification of therapy did not improve clinical outcome, but was associated with increased liver toxicity in one study [60]. Combinations of an aminoglycoside and an antibiotic with activity against anaerobic bacteria have been used to treat patients with intra-abdominal infections. Clinical response and mortality were comparable among 93 patients with intra-abdominal sepsis who were treated either with clindamycin plus gentamicin or with chloramphenicol plus gentamicin (28 of 52, 54 % vs. 20 of 41, 49 %, p = 0.68, and 8 of 52, 15 % vs. 10 of 41, 24 % p = 0.30, respectively) [61]. With the advent of broad-spectrum and bactericidal antibiotics, such as the extended-spectrum penicillins, third- or fourth-generation cephalosporins, or the carbapenems, the need for aminoglycoside-containing antibiotic combinations has subsided. In recent years studies have compared the efficacy and toxicity of a single broad-spectrum antibiotic with those of a b-lactam paired with an aminoglycoside.

S 39

Table 3 Studies comparing monotherapy with a third or a fourthgeneration cephalosporin with a combination of a b-lactam and an aminoglycoside as empirical therapy of severe sepsis (MOX moxalactam, Conv conventional therapy, CEFO cefotaxime, C+T Regimen Clinical success n Oblinger et al. [69] (n = 97) MOX Conv Arich et al. [67] (n = 47) CEFO C+T White et al. [70] (n = 161) CEFTA D+C+T/S Extermann et al. [68] (n = 128) CEFTA Best guess McCormick et al. [66] (n = 128) CEFTA M+N % Overall mortality n % Mortality due to infection n %

cefazolin + tobramycin, CEFTA ceftazidime, D+C+T/S doxcycline + chloramphenicol + trimethoprim/sulfamethoxazole, Best guess best guess combination, M+N mezlocilin + netilmicin, n.a. not available) Superinfections n % Relapse n % Eradication n % Nephrotoxicity n %

33/38 32/40

87 80

8/33 9/32

24 28

2/33 4/32

6 13

2/33 2/32

6 6

n.a.

n.a.

3/41 11/47*

7 23

22/25 17/22

88 77

8/25 5/22

32 23

n.a.

1/25 0/22

4 0

n.a.

n.a.

0/25 0 3/22*** 14

n.a.

13/35 37 20/27** 74

n.a.

n.a.

4/20 7/19

20 37

n.a.

n.a.

38/41 28/30

93 93

6/41 4/30

15 13

3/56 0/55

5 0

n.a.

n.a.

20/22 14/16

91 88

n.a.

50/65 48/63

77 76

13/65 9/63

20 14

n.a.

n.a.

n.a.

n.a.

n.a.

2/65 8/63*

3 13

*p = 0.04, **p = 0.05, ***p = 0.06

Is monotherapy with a carbapenem as efficacious as combination therapy with b-lactam and aminoglycoside as empirical therapy of patients with severe sepsis or septic shock? Answer: yes, grade B. Recommendation Prospective, randomized controlled studies suggest that monotherapy with carbapenem antibiotics is as effective as combination therapy with a b-lactam and an aminoglycoside for the empirical treatment of nonneutropenic patients with severe sepsis. Rationale Four studies have compared the efficacy and safety of a carbapenem (i.e., imipenem-cilastatin or meropenem) to that of a b-lactam paired with an aminoglycoside as

empirical therapy of patients with severe sepsis or septic shock (Table 2). Imipenem monotherapy was compared with a combination of imipenem and netilmicin in 313 patients with severe peritonitis, nosocomial bacteremia, or pneumonia [62]. Of note, this is the only study in which the same b-lactam antibiotic was used in both treatment arms. Overall success rates were similar in the two treatment groups. Netilmicin accounted for almost one-half of the cases of nephrotoxicity that occurred in the combination arm, whereas no case of nephrotoxicity was attributed to imipenem monotherapy. Moreover, the addition of netilmicin to imipenem did not prevent the occurrence of superinfections or of Pseudomonas aeruginosa resistant to imipenem. Thus, adding an aminoglycoside to imipenem did not improve outcome or prevent the incidence of resistance, but it increased nephrotoxicity. In a randomized study of 140 ICU patients with suspected pneumonia or bacteremia, imipenem was found to be as effective as cefotaxime plus amikacin [63]. As shown in Table 2, meropenem also was as efficacious as ceftazidime given either alone or in combination with amikacin [64, 65].

S 40

Is monotherapy with a 3rd or 4th generation cephalosporin as efficacious as combination therapy with b-lactam and aminoglycoside as empirical therapy of patients with severe sepsis or septic shock? Answer: yes, grade C. Recommendation Prospective, randomized controlled studies suggest that monotherapy with third- or fourth-generation cephalosporins is as effective as combination therapy with a blactam and an aminoglycoside for the empirical treatment of nonneutropenic patients with severe sepsis. Rationale Five prospective randomized studies have compared monotherapy with a third- or fourth-generation cephalosporin with combination therapy (Table 3). Ceftazidime was compared with mezlocillin plus netilmicin for the treatment of 128 cirrhotic patients with severe sepsis [66]. Clinical success and mortality rates were similar in the two treatment groups (p = 0.9 and p = 0.4, respectively). However, renal failure occurred in 13 % of the patients treated with mezlocillin and netilmicin but in only 3 % of those who received ceftazidime monotherapy (p = 0.04). Cefotaxime was compared to cefazolin plus tobramycin in a study of 47 patients with Gramnegative bacteremia. The clinical success rate was 88 % with cefotaxime and 73 % with cefazolin plus tobramycin, a difference that was not statistically significant [67]. In a multicenter study 128 patients with severe sepsis or septic shock were randomized to receive either single-agent therapy with ceftazidime or combined therapy to be freely chosen by the investigator [68]. Bacteriological eradication rates and clinical success rates were similar in the two treatment groups. Moxalactam was reported to be as effective as several antibiotic combinations for the treatment of patients with moxalactam-sensitive organisms [69]. Finally, ceftazidime was found to more effective that combined therapy with chloramphenicol, doxycycline, trimethoprim and sulfamethoxazole in patients with severe melioidosis, reducing mortality from 74 % to 37 % (p = 0.009) [70]. Is monotherapy with an extended-spectrum penicillin as efficacious as combination therapy with b-lactam and aminoglycoside as empirical therapy of patients with severe sepsis or septic shock? Answer: uncertain, grade E.

Recommendation Extended-spectrum carboxypenicillins or ureidopenicillins combined with b-lactamase inhibitors have been shown to be effective for the treatment of suspected infections in febrile, neutropenic cancer patients and in patients with peritonitis or nosocomial pneumonia [71, 72, 73, 74, 75]. However, similar studies have not yet been carried out in patients with severe sepsis or shock. Rationale Extended-spectrum carboxypenicillins or ureidopenicillins combined with b-lactamase inhibitors (such as ticarcillin-clavulanate or piperacillin-tazobactam) show excellent in vitro activity against a broad range of Gramnegative and Gram-positive bacteria as well as against anaerobes making them candidates for single-agent therapy. The Medline search identified only one randomized, prospective study comparing extended-spectrum penicillin monotherapy with combination therapy. In that study 396 premature neonates at risk of early onset sepsis were treated with piperacillin or with ampicillin plus amikacin. Mortality rates were similar in both treatment groups (17 of 200, 9 % vs. 27 of 196, 14 %, p = 0.13) [76], suggesting that piperacillin is at least as efficacious as combination therapy for the empirical therapy of sepsis in premature newborns. Is monotherapy with a monobactam as efficacious as combination therapy with b-lactam and aminoglycoside as empirical therapy of patients with severe sepsis or septic shock? Answer: (a) for the treatment of patients with documented Gram-negative sepsis: yes, grade C; (b) as empirical therapy of sepsis: no, grade C. Recommendation Monotherapy with aztreonam appears to be as effective as combination of a b-lactam and an aminoglycoside for the treatment of patients with documented Gram-negative sepsis. The fact that aztreonam lacks any appreciable activity against Gram-positive or anaerobic bacteria precludes its use as empirical single-agent therapy in patients with severe sepsis. Rationale Aztreonam is a monocyclic b-lactam, hence its designation as monobactam that is active against a broad range

S 41

of Gram-negative bacteria, including Enterobacteriaceae and most Pseudomonas aeruginosa. It has no relevant activity against Gram-positive and anaerobic bacteria. Aztreonam has been used successfully for the treatment of patients with Gram-negative bacteremia [77]. It has been shown to be as efficacious as, but less nephrotoxic than aminoglycosides in patients with severe Gram-negative infections [78]. However, this study found that superinfections with Enterococcus faecalis were more frequent in patients treated with aztreonam than in those treated with aminoglycosides. A prospective randomized study investigated the role of aztreonam as empirical therapy of severe sepsis or septic shock caused by Gram-negative bacteria. In a multicenter study 157 ICU patients with Gram-negative bacillary infections (78 pneumonias, 28 urinary tract infections, 23 peritonitis and 40 bacteremias) were randomized to be treated either with aztreonam alone or with a combination of amikacin with or without a broad-spectrum b-lactam. Clinical cure was achieved in 44 of 48 patients (92 %) who received aztreonam and in 25 of 34 patients (73 %) who received amikacin and b-lactam (p = 0.01) [79]. These studies thus suggested that aztreonam is an effective treatment of documented Gram-negative infections. However, aztreonam monotherapy should not be used in patients with severe sepsis or septic shock, as it is devoid of activity against Gram-positive bacteria. Is monotherapy with a quinolone as efficacious as combination therapy with b-lactam and aminoglycoside as empirical therapy of patients with severe sepsis or septic shock Answer: uncertain, grade E. Recommendation Fluoroquinolones have been shown to be highly effective for the treatment of documented Gram-negative bloodstream infections. However, data are lacking to support their use as single-agent treatment of sepsis, especially as first-generation fluoroquinolones display suboptimal activities against Gram-positive bacteria. Furthermore, resistant strains are selected fairly rapidly. Rationale Fluoroquinolones are effective therapy for bloodstream infections caused by enteric Gram-negative bacteria [80, 81] and therefore are an excellent alternative to blactam antibiotics for the treatment of patients with documented Gram-negative sepsis. Limited data are available to evaluate the role of fluoroquinolone antibiotics

as a single agent for the treatment of patients with severe sepsis. The efficacy of ciprofloxacin monotherapy was compared with that of several b-lactams (aztreonam, ceftazidime, ticarcillin-clavulanate, or imipenem) given either alone or in combination with an aminoglycoside in a subset of patients (i.e., with an Acute Physiology and Chronic Health Evaluation II score of 20 or less) enrolled in a multicenter study of 540 patients with severe infections [82]. Ciprofloxacin treatment was given intravenously for a minimum of 23 days, and patients were then switched to oral therapy at the discretion of the investigator. Overall, clinical response was achieved in 138 of 166 patients (83 %) treated with ciprofloxacin monotherapy and in 74 of 87 (85 %) treated with aztreonam, ceftazidime, ticarcillin-clavulanate, or imipenem. Other studies are needed to assess the role of fluoroquinolones in this setting, especially as first-generation fluoroquinolones, such as norfloxacin and ciprofloxacin, have limited activity against Grampositive bacteria. Newer fluoroquinolones, such as levofloxacin, trovafloxacin, gatifloxacin, and moxifloxacin show enhanced in vitro activities against Gram-positive bacteria, but demonstration of activity awaits publication of the results of ongoing studies. Comments The data suggest that monotherapy is a safe alternative to combination therapy for the empirical treatment of critically ill septic patients. However, compared to the wealth of studies performed in febrile neutropenic cancer patients, only a small number have been conducted in patients with severe sepsis. Moreover, many of these studies included fewer than 200 patients, and their statistical power is therefore limited. The fact that monotherapy with carbapenem, third- or fourth-generation cephalosporins or ureidopenicillins plus b-lactamase inhibitors (i.e., piperacillin-tazobactam) have been shown to be as effective as b-lactam antibiotics paired with an aminoglycoside in large multicenter studies in severely compromised neutropenic patients certainly lends further support to the present recommendations [7]. However, critically ill septic patients differ markedly from neutropenic cancer patients. Capillary leak syndrome and multiorgan dysfunction are much more frequent in patients with severe sepsis or septic shock than in neutropenic cancer patients. Such conditions are likely to influence both the volume of distribution and metabolism of the antibiotics, which may result in altered pharmacokinetics and ultimately may affect drug efficacy and toxicity. One should therefore be concerned that treatment guidelines for the most severely ill septic patients tend to rely heavily on the results of trials accomplished in the neutropenic host. There is undoubtedly a need for large, prospective, randomized trials to assess

S 42

the efficacy and toxicity of any new antibiotic in patients with severe sepsis and septic shock. Monotherapy should not be regarded as a universal panacea to be used indiscriminately. Despite a lack of clearcut advantage, some clinicians may still prefer to rely on b-lactam and aminoglycoside combinations to treat patients with nosocomial pneumonia or those with infections caused by Pseudomonas, Serratia, or Enterobacter species. The potential benefit from additive or synergistic effects and the possible prevention of emerging resistant bacteria must be weighed against the risk of increased toxicity. Aminoglycoside-containing regimens have been shown repeatedly to increase the incidence of nephrotoxicity and/or ototoxicity. Furthermore, the concept that adding an aminoglycoside may prevent the emergence of resistance has been challenged by the results of a recent study in patients with severe nosocomial infections [62]. Pseudomonas aeruginosa resistant to imipenem have been found to be as frequent in patients treated with netilmicin and imipenem as in those treated with imipenem alone. Whether up-front empirical therapy should comprise a specific anti-Gram-positive agent is discussed below.

aminoglycoside-induced nephrotoxicity and ototoxicity. These serious adverse events are a major concern, especially in critically ill septic patients, who are already at high risk of developing multiple organ dysfunction. Today aminoglycosides should not be used as single-agent empirical therapy of severe sepsis, as antibiotics such as extended-spectrum penicillins, third- or fourth-generation cephalosporins, or carbapenems have a broader spectrum of activity and are less toxic than aminoglycosides. Are there differences between third-generation and fourth-generation cephalosporins and carbapenem antibiotics as empirical therapy of patients with severe sepsis or septic shock? Answer: no, grade C. Recommendation Third-generation and fourth-generation cephalosporins and carbapenem antibiotics are equally effective as empirical therapy in patients with severe sepsis. Rationale Four studies have compared the efficacy and safety of a third-generation cephalosporin to that of a fourth-generation cephalosporin or a carbapenem. An analysis was carried out on 226 patients with microbiologically confirmed septicemia selected from a pool of 15 phase II and phase III studies of cefpirome [85]; 176 patients were treated with cefpirome and 50 with ceftazidime. A satisfactory clinical response was found in 131 of 176 (74 %) cefpirome recipients and in 34 of 50 (68 %) ceftazidime recipients (p = 0.47). A subsequent randomized multicenter study treated 372 patients with severe sepsis and suspected bacteremias empirically with cefpirome or with ceftazidime [86]. The study protocol allowed the addition of metronidazole, an aminoglycoside, or a glycopeptide antibiotic whenever indicated. Of the patients in the cefpirome group 62 % received monotherapy and of those the ceftazidime group 63 %. Clinical success rates in the intent-to-treat analysis were 65 % (123 of 188) and 70 % (132 of 188), respectively. These two studies showed that cefpirome and ceftazidime are equally effective for the treatment of patients with severe sepsis. Ceftazidime and cefepime have also been found to be equally effective in a small study of 28 severely ill patients with suspected Gramnegative bacteremia [87]. Lastly, 45 patients with pneumonias or bacteremias were randomized to receive either imipenem or ceftazi-

Studies comparing single-agent therapeutic strategies


Aminoglycosides were among the first antibiotics to be studied as single-agents for the treatment of Gram-negative infections. In the late 1970s and early 1980s several studies compared the efficacy and safety of various aminoglycosides. In a prospective, double-blind study, amikacin or gentamicin was used to treat 174 patients with suspected severe Gram-negative infections [83]. Favorable clinical response rates were 77 % (30 of 39) and 78 % (25 of 32), respectively. Both aminoglycosides were accompanied with a high rate of serious adverse events. Definite nephrotoxicity and ototoxicity occurred in 5 of 62 (8 %) and 2 of 32 (6 %) patients treated with amikacin and in 7 of 62 (11 %) and 3 of 30 (10 %) patients treated with gentamicin. Netilmicin and amikacin were used to treat 80 patients with severe sepsis due to Gram-negative bacteria, including bacteremias, genitourinary tract infections, and pneumonias [84]. A favorable clinical response was observed in 30 of 34 patients treated with netilmicin (88 %) and in 26 of 33 treated with amikacin (79 %, p = 0.34). The frequency of nephrotoxicity was 38 % with netilmicin (13 of 34) and 28 % with amikacin (8 of 29) and that of ototoxicity was 9 % (3 of 34) and 24 % (7 of 29) respectively, but the difference was not statistically significant (p = 0.42 and 0.16, respectively). Overall, the various aminoglycosides (i.e., gentamicin, tobramycin, netilmicin, and amikacin) have been shown to be equally effective as empirical monotherapy of Gram-negative sepsis. A constant finding in all these studies was the high incidence of

S 43

dime therapy. A favorable clinical response was observed in 17 of 21 (81 %) patients receiving ceftazidime and in 16 of 24 (67 %) receiving imipenem (p = 0.33) [88]. Despite limited statistical power due to the small number of patients enrolled, these four studies suggest that third- and fourth-generation cephalosporins and carbapenem antibiotics are equally effective as empirical therapy in patients with severe sepsis. Are there clinical conditions justifying the use of empirical anti-Gram-positive therapy in patients with severe sepsis? Answer: yes, grade E. Recommendation The indiscriminate use of vancomycin or teicoplanin for presumed Gram-positive infections in patients with severe sepsis and septic shock should be avoided. Glycopeptides are appropriate in severely ill patients with catheter-related infections or in centers in which methicillin-resistant staphylococci predominate. However, the possible clinical benefit associated with the empirical use of vancomycin or teicoplanin should be weighed against the risks of selecting resistant organisms and of increased toxicity, especially when vancomycin is given in combination with an aminoglycoside [89] or other nephrotoxic agent. To further reduce the risk of the emergence of vancomycin-resistant staphylococci or enterococci, empirical vancomycin therapy should also be rapidly discontinued in patients in whom Gram-positive infections has been ruled out. Finally, it is rarely, if ever, appropriate to use vancomycin alone as empirical therapy since most cases require additional Gram-negative coverage, at least until microbiological results are available. Rationale In recent years many institutions have experienced a major change in the cause of bacterial infections occurring in ICU patients with sepsis. While Gram-negative bacteria predominated until the middle 1980s, Grampositive bacteria now account for approximately onehalf of the infections occurring in patients with severe sepsis and septic shock [20]. Moreover, methicillin-resistant S. aureus and methicillin-resistant coagulase-negative staphylococci are responsible for a majority of staphylococcal infections in some institutions. The frequency of penicillin-resistant S. pneumoniae is also increasing in many areas of the world. Does this epidemiological context justify the empirical use of glycopeptide

antibiotics (i.e., vancomycin and teicoplanin) on a routine basis in all patients with severe sepsis and septic shock? The literature does not offer an answer to that question. To our knowledge, no study has prospectively examined the role of glycopeptide antibiotics in the management of nonneutropenic patients with severe sepsis. All recent studies have been performed in neutropenic cancer patients, in whom viridans streptococci and coagulase-negative staphylococci are a frequent cause of infection [7]. Controversy still exists as to the need for empirical specific anti-Gram-positive therapy at the onset of fever in neutropenic cancer patients [90, 91, 92]. Although some studies have suggested that the empirical use of vancomycin or teicoplanin at the initiation of empirical therapy is preferable, others have yielded data that do not support that concept. Prospective studies are ongoing to further address that question. Are antifungal agents indicated as empirical therapy of patients with severe sepsis or septic shock? Answer: no, grade E. Recommendation Antifungal agents, such as fluconazole, should not be used on a routine basis as empirical therapy in patients with severe sepsis and septic shock. Rationale Since the middle 1980s fungi have emerged worldwide as an increasingly frequent cause of nosocomial infections in critically ill patients and are associated with significant morbidity and high mortality [93, 94, 95, 96]. A recent survey of bloodstream infections found ICUs to have the highest incidence of candidemia, accounting for 45 % of all episodes of fungemia [97]. A 1-day point prevalence study carried out in 1417 ICUs in western European countries isolated fungi in 17 % of all ICU-acquired infections [98]. However, it is unclear whether all these fungal isolates were the causal agent of infections rather than colonizing micro-organisms. Clinical manifestations of candidiasis are usually not specific, and standard culture techniques and tests for detection of Candida antigens or metabolites lack sensitivity. Taken together these facts might support the empirical use of antifungal agents for the treatment of the critically ill ICU patient with severe sepsis or septic shock. However, recent epidemiological studies and multicenter trials have shown that fungi account for only 5 % of all cases of severe sepsis or septic shock (Fig. 1B), which would not justify the use of antifungal therapy on a routine ba-

S 44

sis. Since Candida infections from species other than C. albicans are increasing, along with the resistance to azoles, one should even be more cautious with the use of this class of antifungal agents. Are azoles as effective as amphotericin B for the treatment of patients with candidemia? Answer: yes, grade A. Recommendation Fluconazole is as effective as and less toxic than amphotericin B for the treatment of candidemia in nonneutropenic patients. However, if the patient has been treated previously with fluconazole, it might be prudent to begin therapy with amphotericin B while waiting for the identification of the Candida species and for the results of susceptibility testing. Infections caused by Candida strains other than C. albicans with reduced susceptibility to azoles are more commonly seen in patients who have received previous antifungal therapy, including azoles, than in those who did not. Among the other Candida strains, C. krusei is intrinsically resistant to fluconazole, and C. glabrata is relatively resistant to fluconazole. Many investigators consider amphothericin B the agent of choice to treat unstable patients with candidemia [48]. Whether 5-fluorocytosine should be combined with amphotericin B in these patients is debatable. No data have shown that combining these agents improve outcome of patients with candidemia. Rationale A recent study has shown that Candida is the fourth most frequent cause of bloodstream infections [99]. Between one-fourth and one-half of all episodes of candidemia occur in ICUs. Seven comparative studies compared the efficacy of fluconazole and amphotericin B for the treatment of candidemia. A large multicenter study treated 206 nonneutropenic patients with candidemia with fluconazole (400 mg per day) or with amphotericin B (0.50.6 mg/kg per day). Most of these patients had catheter-related candidemia. Fluconazole was as effective as amphotericin B (success rate were 72 % and 79 %, respectively) [100]. The proportion of patients with hypokalemia and elevated blood urea nitrogen or serum creatinine were significantly lower in the fluconazole group than in the amphotericin B group (p < 0.001 and p = 0.006). In a large multicenter, prospective observational study of the morbidity and mortality of candidemia, intravenous amphotericin B was given to 227 patients and fluconazole to 67 patients

[49]. Mortality at day 30 was not significantly different in patients treated with amphotericin B and in those treated with fluconazole (31 % vs. 27 %, p = 0.6). Another randomized study treated 50 patients with fluconazole and 53 with amphotericin B. Clinical success rate (24 of 42, 57 % vs. 26 of 42, 62 %, respectively, p = 0.82) and mortality at day 14 (13 of 50, 26 % vs. 11 of 53, 21 %, respectively, p = 0.69) were comparable in the two treatment groups [101]. Two smaller studies yielded analogous results. Fluconazole was found to be as effective as amphotericin B in 90 ICU patients with cancer and systemic Candida infections (i.e., bloodstream infections, pneumonia or peritonitis; clinical response rates: 33 of 45, 73 % vs. 32 of 45, 71 %, p = 0.8) [102]. The efficacy and safety of fluconazole was compared to that of amphotercin B for the treatment of disseminated fungal sepsis in neonates [103]. Case fatality rates were 33 % (4 of 12) in the fluconazole group and 45 % (5 of 11) in the amphotericin B group (p = 0.68). Adverse effects were less frequent in neonates treated with fluconazole than in those who received amphotericin B. In all studies there were fewer adverse events, especially nephrotoxicity, hypokalemia, fever, and chills, with fluconazole than with amphotericin B. There is a lack of dose-finding studies on which to base a recommendation on what dose of antifungal agents to use for the treatment of patients with candidemia. A dose of 400 mg/d has been used in most fluconazole studies. Whether higher daily doses of fluconazole would be more effective is unclear. One study of 65 patients with fungemia due to C. albicans found a clinical response rate of 60 % in patients treated with 5 mg/kg fluconazole once daily and one of 83 % in those who received a daily dose of 10 mg/kg [104]. Of note, 18 of 20 international experts who participated in a consensus conference on the management of Candida infections would use a daily dose of 400 mg fluconazole to treat stable patients with candidemia; in patients who are unstable or deteriorating, one-half of the experts would use a dose of 800 mg fluconazole while the other one-half would use amphotericin B [48]. However, these opinions were based primarily on personal experience. Prospective randomized trials are needed to address these issues.
Acknowledgements We are grateful to Dr. Jacques Cornuz for assistance in the literature search and statistical analysis, and to Prof. Jonathan Cohen for critical reading of the manuscript. This work was supported by grants from the Swiss National Science Foundation to T. C. (32-49129.96). T. C. is recipient of a career award from the Swiss National Science Foundation (32-48916.96). M. P.G. is recipient of a career award from the Bristol-Myers Squibb Foundation.

S 45

References
1. Astiz ME, Rackow EC (1998) Septic shock. Lancet 351: 15011505 2. Wheeler AP, Bernard GR (1999) Treating patients with severe sepsis. N Engl J Med 340: 207214 3. Shenep JL (1996) Septic shock. Adv Pediatr Infect Dis 12: 209241 4. Bone RC, Sibbald WJ, Sprung CL (1992) The ACCP-SCCM consensus conference on sepsis and organ failure. Chest 101: 14811483 5. Young LS (1995) Sepsis syndrome. In: Mandell GL, Bennett JE, Dolin E (ed) Principles and practice of infectious diseases. Churchill Livingstone, New York, pp 690705 6. Calandra T, Cometta A (1991) Antibiotic therapy for gram-negative bacteremia. Infect Dis Clin North Am 5: 817834 7. Hughes WT, Armstrong D, Bodey GP, et al (1997) 1997 guidelines for the use of antimicrobial agents in neutropenic patients with unexplained fever. Infectious Diseases Society of America. Clin Infect Dis 25: 551573 8. Fisher CJ, Slotman GJ, Opal SM, et al (1994) Initial evaluation of human recombinant interleukin-1 receptor antagonist in the treatment of sepsis syndrome: a randomized, open-label, placebo-controlled multicenter trial. Crit Care Med 22: 1221 9. Abraham E, Wunderink R, Silverman H, et al (1995) Efficacy and safety of monoclonal antibody to human tumor necrosis factor a in patients with sepsis syndrome. A randomized, controlled, double-blind, multicenter clinical trial. JAMA 273: 934941 10. Zeni F, Freeman B, Natanson C (1997) Anti-inflammatory therapies to treat sepsis and septic shock: a reassessment. Crit Care Med 25: 10951100 11. Brun-Buisson C, Doyon F, Carlet J, et al (1995) Incidence, risk factors, and outcome of severe sepsis and septic shock in adults. A multicenter prospective study in intensive care units. French ICU Group for Severe Sepsis. JAMA 274: 968974 12. Brun-Buisson C, Doyon F, Carlet J (1996) Bacteremia and severe sepsis in adults: a multicenter prospective survey in ICUs and wards of 24 hospitals. French Bacteremia-Sepsis Study Group. Am J Respir Crit Care Med 154: 617624 13. Cohen J, Carlet J, for the INTERSEPT Study Group (1996) INTERSEPT: an international, multicenter, placebo-controlled trial of monoclonal antibody to human tumor necrosis factor-a in patients with sepsis. Crit Care Med 24: 14311440 14. Sands KE, Bates DW, Lanken PN, et al (1997) Epidemiology of sepsis syndrome in 8 academic medical centers. Academic Medical Center Consortium Sepsis Project Working Group. JAMA 278: 234240 15. Muckart DJ, Bhagwanjee S (1997) American College of Chest Physicians/Society of Critical Care Medicine Consensus Conference definitions of the systemic inflammatory response syndrome and allied disorders in relation to critically injured patients. Crit Care Med 25: 17891795 16. Sackett DL, Haynes RB, Guyatt GH, Tugwell P (1991) Clinical epidemiology. Lippincott-Raven, Philadelphia 17. Sackett DL (1989) Rules of evidence and clinical recommendations on the use of antithrombotic agents. Chest 95: 2S-4S 18. The Veterans Administration Systemic Sepsis Cooperative Study Group (1987) Effect of high-dose glucocorticoid therapy on mortality in patients with clinical signs of systemic sepsis. N Engl J Med 317: 659665 19. Bryant RE, Hood AF, Hood CE, Koenig MG (1971) Factors affecting mortality of gram-negative rod bacteremia. Arch Intern Med 127: 120128 20. Freid MA, Vosti KL (1968) The importance of underlying disease in patients with gram-negative bacteremia. Arch Intern Med 121: 418423 21. Klastersky J, Cappel R, Debusscher L (1971) Effectiveness of betamethasone in management of severe infections. A double-blind study. N Engl J Med 284: 12481250 22. McCabe WR, Jackson GG (1962) Gram negative bacteremia. Arch Intern Med 110: 92100 23. Schummer W (1976) Steroids in the treatment of clinical septic shock. Ann Surg 184: 333341 24. Young LS, Martin WJ, Meyer RD, Weinstein RJ, Anderson ET (1977) Gram-negative rod bacteremia: microbiologic, immunologic, and therapeutic considerations. Ann Intern Med 86: 456471 25. Abraham E, Glauser MP, Butler T, et al (1997) p55 tumor necrosis factor receptor fusion protein in the treatment of patients with severe sepsis and septic shock. A randomized controlled multicenter trial. JAMA 277: 15311538 26. Abraham E, Anzueto A, Gutierrez G, et al (1998) Double-blind randomised controlled trial of monoclonal antibody to human tumour necrosis factor in treatment of septic shock. NORASEPT II Study Group. Lancet 351: 929933 27. Boillot A, Capellier G, Racadot E, et al (1995) Pilot clinical trial of an antiTNF alpha monoclonal antibody for the treatment of septic shock. Clin Intensive Care 6: 5256 28. Dhainaut JF, Tenaillon A, Le Tulzo Y, et al (1994) Platelet-activating factor receptor antagonist BN 52021 in the treatment of severe sepsis: a randomized, double-blind, placebo-controlled, multicenter clinical trial. BN 52021 Sepsis Study Group. Crit Care Med 22: 17201728 29. Dhainaut JF, Vincent JL, Richard C, et al (1995) CDP571, a humanized antibody to human tumor necrosis factoralpha: safety, pharmacokinetics, immune response, and influence of the antibody on cytokine concentrations in patients with septic shock. CPD571 Sepsis Study Group. Crit Care Med 23: 14611469 30. Dhainaut JF, Tenaillon A, Hemmer M, et al (1998) Confirmatory platelet-activating factor receptor antagonist trial in patients with severe gram-negative bacterial sepsis: a phase III, randomized, double-blind, placebo-controlled, multicenter trial. BN 52021 Sepsis Investigator Group. Crit Care Med 26: 19631971 31. Fisher CJ, Dhainaut JF, Opal SM, et al (1994) Recombinant human interleukin 1 receptor antagonist in the treatment of patients with sepsis syndrome: results from a randomized, doubleblind, placebo-controlled trial. JAMA 271: 18361843 32. Fisher CJ, Agosti JM, Opal SM, et al (1996) Treatment of septic shock with the tumor necrosis factor receptor: Fc fusion protein. N Engl J Med 334: 16971702

S 46

33. Fisher CJ Jr, Opal SM, Dhainaut JF, et al (1993) Influence of an anti-tumor necrosis factor monoclonal antibody on cytokine levels in patients with sepsis. The CB0006 Sepsis Syndrome Study Group. Crit Care Med 21: 318327 34. Opal SM, Fisher CJ, Dhainaut JF, et al (1997) Confirmatory interleukin-1 receptor antagonist trial in severe sepsis: a phase III, randomized, double-blind, placebo-controlled, multicenter trial. Crit Care Med 25: 11151124 35. Reinhart K, Wiegand-Lhnert C, Grimminger F, et al (1996) Assessment of the safety and efficacy of the monoclonal anti-tumor necrosis factor antibody-fragment, MAK 195F, in patients with sepsis and septic shock: a multicenter, randomized, placebo-controlled, dose-ranging study. Crit Care Med 24: 733742 36. Saravolatz LD, Wherry JC, Spooner C, et al (1994) Clinical safety, tolerability, and pharmacokinetics of murine monoclonal antibody to human tumor necrosis factor-a. J Infect Dis 169: 214217 37. Cohen J, Abraham E (1999) Microbiologic findings and correlations with serum tumor necrosis factor-alpha in patients with severe sepsis and septic shock. J Infect Dis 180: 116121 38. Edmond MB, Wallace SE, McClish DK, et al (1999) Nosocomial bloodstream infections in United States hospitals: a three-year analysis. Clin Infect Dis 29: 239244 39. Pfaller MA, Jones RN, Messer SA, Edmond MB, Wenzel RP (1998) National surveillance of nosocomial blood stream infection due to species of Candida other than Candida albicans: frequency of occurrence and antifungal susceptibility in the SCOPE Program. SCOPE Participant Group. Surveillance and Control of Pathogens of Epidemiologic. Diagn Microbiol Infect Dis 30: 121129 40. Bernard GR, Reines HD, Halushka PV, et al (1991) Prostacyclin and thromboxane A2 formation is increased in human sepsis syndrome. Effects of cyclooxygenase inhibition. Am Rev Respir Dis 144: 10951101 41. Bernard GR, Wheeler AP, Russell JA, et al (1997) The effects of ibuprofen on the physiology and survival of patients with sepsis. N Engl J Med 336: 912918

42. Haupt MT, Jastremski MS, Clemmer TP, Metz CA, Goris GB (1991) Effect of ibuprofen in patients with severe sepsis: a randomized, double-blind, multicenter study. The Ibuprofen Study Group. Crit Care Med 19: 13391347 43. Klastersky J, Coppens L, MeunierCarpentier F, Menday AP (1980) Carbenicillin plus cefazolin with or without mecillinam as an early treatment of bacteremia caused by gram-negative organisms: randomized doubleblind study. Antimicrob Agents Chemother 18: 437442 44. Friedman G, Silva E, Vincent JL (1998) Has the mortality of septic shock changed with time. Crit Care Med 26: 20782086 45. Kreger BE, Craven DE, McCabe WR (1980) Gram-negative bacteremia. IV. Re-evaluation of clinical features and treatment in 612 patients. Am J Med 68: 344355 46. Pittet D, Rangel-Frausto S, Li N, et al (1995) Systemic inflammatory response syndrome, sepsis, severe sepsis and septic shock: incidence, morbidities and outcomes in surgical ICU patients. Intensive Care Med 21: 302309 47. Linden PK, Pasculle AW, McDevitt D, Kramer DJ (1997) Effect of quinupristin/dalfopristin on the outcome of vancomycin-resistant Enterococcus faecium bacteraemia: comparison with a control cohort. J Antimicrob Chemother 39 [Suppl A]:145151 48. Edwards JE Jr, Bodey GP, Bowden RA, et al (1997) International Conference for the Development of a Consensus on the Management and Prevention of Severe Candidal Infections. Clin Infect Dis 25: 4359 49. Nguyen MH, Peacock JE Jr, Tanner DC, et al (1995) Therapeutic approaches in patients with candidemia. Evaluation in a multicenter, prospective, observational study. Arch Intern Med 155: 24292435 50. Nolla-Salas J, Sitges-Serra A, LeonGil C, et al (1997) Candidemia in nonneutropenic critically ill patients: analysis of prognostic factors and assessment of systemic antifungal therapy. Study Group of Fungal Infection in the ICU. Intensive Care Med 23: 2330 51. Calandra T, Glauser MP (1986) Immunocompromised animal models for the study of antibiotic combinations. Am J Med 80: 4552 52. Klastersky J, Zinner SH (1982) Synergistic combinations of antibiotics in gram-negative bacillary infections. Rev Infect Dis 4: 294301

53. De Jongh CA, Joshi JH, Thompson BW, et al (1986) A double beta-lactam combination versus an aminoglycoside-containing regimen as empiric antibiotic therapy for febrile granulocytopenic cancer patients. Am J Med 80: 101111 54. Gribble MJ, Chow AW, Naiman SC, et al (1983) Prospective randomized trial of piperacillin monotherapy versus carboxypenicillin-aminoglycoside combination regimens in the empirical treatment of serious bacterial infections. Antimicrob Agents Chemother 24: 388393 55. Ceftazidime combined with a short or long course of amikacin for empirical therapy of gram-negative bacteremia in cancer patients with granulocytopenia. The EORTC International Antimicrobial Therapy Cooperative Group (1987) N Engl J Med 317: 16921698 56. Anderson ET, Young LS, Hewitt WL (1978) Antimicrobial synergism in the therapy of gram-negative rod bacteremia. Chemotherapy 24: 4554 57. Klastersky J, Meunier-Carpentier F, Prevost JM (1977) Significance of antimicrobial synergism for the outcome of gram negative sepsis. Am J Med Sci 273: 157167 58. Hanson B, Coppens L, Klastersky J (1982) Comparative studies of ticarcillin and mezlocillin plus sisomicin in Gram-negative bacillary bacteraemia and bronchopneumonia. J Antimicrob Chemother 10: 335341 59. Beaucaire G (1995) Evaluation of the efficacy and safety of isepamicin compared with amikacin in the treatment of nosocomial pneumonia and septicaemia. J Chemother 7 Suppl 2: 165173 60. Korvick JA, Peacock JE Jr, Muder RR, Wheeler RR, Yu VL (1992) Addition of rifampin to combination antibiotic therapy for Pseudomonas aeruginosa bacteremia: prospective trial using the Zelen protocol. Antimicrob Agents Chemother 36: 620-625 61. Lennard ES, Minshew BH, Dellinger EP, et al (1985) Stratified outcome comparison of clindamycin-gentamicin vs chloramphenicol-gentamicin for treatment of intra-abdominal sepsis. Arch Surg 120: 889898 62. Cometta A, Baumgartner JD, Lew D, et al (1994) Prospective randomized comparison of imipenem monotherapy with imipenem plus netilmicin for treatment of severe infections in nonneutropenic patients. Antimicrob Agents Chemother 38: 13091313

S 47

63. Mouton Y, Deboscker Y, Bazin C, et al (1990) Prospective, randomized, controlled study of imipenem-cilastatin versus cefotaxime-amikacin in the treatment of lower respiratory tract infection and septicemia at intensive care units. Presse Med 19: 607612 64. Mouton YJ, Beuscart C (1995) Empirical monotherapy with meropenem in serious bacterial infections. Meropenem Study Group. J Antimicrob Chemother 36 [Suppl A]:145156 65. Solberg CO, Sjursen H (1995) Safety and efficacy of meropenem in patients with septicaemia: a randomised comparison with ceftazidime, alone or combined with amikacin. J Antimicrob Chemother 36 [Suppl A]:157166 66. McCormick PA, Greenslade L, Kibbler CC, et al (1997) A prospective randomized trial of ceftazidime versus netilmicin plus mezlocillin in the empirical therapy of presumed sepsis in cirrhotic patients. Hepatology 25: 833836 67. Arich C, Gouby A, Bengler C, et al (1987) Comparison of the efficacy of cefotaxime alone and the combination cefazolin-tobramycin in the treatment of enterobacterial septicemia. Pathol Biol (Paris) 35: 613615 68. Extermann M, Regamey C, Humair L, et al (1995) Initial treatment of sepsis in non-neutropenic patients: ceftazidime alone versus 'best guess' combined antibiotic therapy. Chemotherapy 41: 306315 69. Oblinger MJ, Bowers JT, Sande MA, Mandell GL (1982) Moxalactam therapy vs. standard antimicrobial therapy for selected serious infections. Rev Infect Dis 4 [Suppl]:S639S649 70. White NJ, Dance DA, Chaowagul W, et al (1989) Halving of mortality of severe melioidosis by ceftazidime. Lancet II:697701 71. Brismar B, Malmborg AS, Tunevall G, et al (1992) Piperacillin-tazobactam versus imipenem-cilastatin for treatment of intra-abdominal infections. Antimicrob Agents Chemother 36: 27662773 72. Eklund AE, Nord CE (1993) A randomized multicenter trial of piperacillin/tazobactam versus imipenem/cilastatin in the treatment of severe intra-abdominal infections. Swedish Study Group. J Antimicrob Chemother 31 [Suppl A]:7985 73. Jaccard C, Troillet N, Harbarth S, et al (1998) Prospective randomized comparison of imipenem-cilastatin and piperacillin-tazobactam in nosocomial pneumonia or peritonitis. Antimicrob Agents Chemother 42: 29662972

74. Mouton Y, Leroy O, Beuscart C, et al (1993) Efficacy, safety and tolerance of parenteral piperacillin/tazobactam in the treatment of patients with lower respiratory tract infections. J Antimicrob Chemother 31 [Suppl A]:8795 75. Polk HC Jr, Fink MP, Laverdiere M, et al (1993) Prospective randomized study of piperacillin/tazobactam therapy of surgically treated intra-abdominal infection. The Piperacillin/Tazobactam Intra-Abdominal Infection Study Group. Am Surg 59: 598605 76. Hammerberg O, Kurnitzki C, Watts J, Rosenbloom D (1989) Randomized trial using piperacillin versus ampicillin and amikacin for treatment of premature neonates with risk factors for sepsis. Eur J Clin Microbiol Infect Dis 8: 241244 77. Scully BE, Henry SA (1985) Clinical experience with aztreonam in the treatment of gram-negative bacteremia. Rev Infect Dis 7 [Suppl 4]: S789S793 78. Gudiol F, Pallares R, Ariza X, et al (1986) Comparative clinical evaluation of aztreonam versus aminoglycosides in gram-negative septicaemia. J Antimicrob Chemother 17: 661671 79. Rabinad E, Bosch-Perez A (1989) A multicenter comparative trial of aztreonam in the treatment of gram-negative infections in compromised intensive-care patients. Chemotherapy 35 [Suppl 1]:17 80. Bouza P, Garcia FT, Perez FM, et al (1996) Treatment of CAPD-related peritonitis with ciprofloxacin: results after seven years. Adv Perit Dial 12: 185188 81. Extermann M, Regamey C (1994) Empirical antibiotic treatment of sepsis in non-neutropenic patients: single agent or combination therapy? Infection 22: 13 82. Krumpe PE, Cohn S, Garreltes J, et al (1999) Intravenous and oral mono- or combination-therapy in the treatment of severe infections: ciprofloxacin versus standard antibiotic therapy. Ciprofloxacin Study Group. J Antimicrob Chemother 43 [Suppl A]:117128 83. Smith CR, Baughman KL, Edwards CQ, Rogers JF, Lietman PS (1977) Controlled comparison of amikacin and gentamicin. N Engl J Med 296: 349353 84. Bock BV, Edelstein PH, Meyer RD (1980) Prospective comparative study of efficacy and toxicity of netilmicin and amikacin. Antimicrob Agents Chemother 17: 217225

85. Norrby SR, Geddes AM (1993) Efficacy of cefpirome in the treatment of septicaemia. Scand J Infect Dis Suppl 91: 5159 86. Norrby SR, Geddes AM, Shah PM (1998) Randomized comparative trial of cefpirome versus ceftazidime in the empirical treatment of suspected bacteraemia or sepsis. Multicentre Study Group. J Antimicrob Chemother 42: 503509 87. Schrank JH Jr, Kelly JW, McAllister CK (1995) Randomized comparison of cefepime and ceftazidime for treatment of hospitalized patients with gram-negative bacteremia. Clin Infect Dis 20: 5658 88. Hartenauer U, Weilemann LS, Bodmann KF, et al (1990) Comparative clinical trial of ceftazidime and imipenem/cilastatin in patients with severe nosocomial pneumonias and septicaemias. J Hosp Infect 15 [Suppl A]:6164 89. European Organization for Research and Treatment of Cancer (EORTC) International Antimicrobial Therapy Cooperative Group and the National Cancer Institute of Canada-Clinical Trials Group (1991) Vancomycin added to empirical combination antibiotic therapy for fever in granulocytopenic cancer patients. J Infect Dis 163: 951958 90. Giamarellou H (1995) Empiric therapy for infections in the febrile, neutropenic, compromised host. Med Clin North Am 79: 559580 91. Karp JE, Dick JD, Angelopulos C, et al (1986) Empiric use of vancomycin during prolonged treatment-induced granulocytopenia. Randomized, double-blind, placebo-controlled clinical trial in patients with acute leukemia. Am J Med 81: 237242 92. Rubin M, Hathorn JW, Marshall D, et al (1988) Gram-positive infections and the use of vancomycin in 550 episodes of fever and neutropenia. Ann Intern Med 108: 3035 93. Banerjee SN, Emori TG, Culver DH, et al (1991) Secular trends in nosocomial primary bloodstream infections in the United States, 19801989. National Nosocomial Infections Surveillance System. Am J Med 91: 86S-89S 94. Beck-Sagu CM, Jarvis WR, and the National Nosocomial Infections Surveillance System (1993) Secular trends in the epidemiology of nosocomial fungal infections in the United States, 19801990. J Infect Dis 167: 12471251 95. Fisher-Hoch SP, Hutwagner L (1995) Opportunistic candidiasis: an epidemic of the 1980s. Clin Infect Dis 21: 897904

S 48

96. Fraser VJ, Jones M, Dunkel J, et al (1992) Candidemia in a tertiary care hospital: epidemiology, risk factors, and predictors of mortality. Clin Infect Dis 15: 414421 97. Wenzel RP (1995) Nosocomial candidemia: risk factors and attributable mortality. Clin Infect Dis 20: 15311534 98. Vincent JL, Bihari DJ, Suter PM, et al (1995) The prevalence of nosocomial infection in intensive care units in Europe. Results of the European Prevalence of Infection in Intensive Care (EPIC) Study. EPIC International Advisory Committee. JAMA 274: 639644

99. Rex JH, Bennett JE, Sugar AM, et al (1994) A randomized trial comparing fluconazole with amphotericin B for the treatment of candidemia in patients without neutropenia. Candidemia Study Group and the National Institute. N Engl J Med 331: 13251330 100. Rangel-Frausto MS, Wiblin T, Blumberg HM, et al (1999) National epidemiology of mycoses survey (NEMIS): variations in rates of bloodstream infections due to Candida species in seven surgical intensive care units and six neonatal intensive care units. Clin Infect Dis 29: 253258 101. Phillips P, Shafran S, Garber G, et al (1997) Multicenter randomized trial of fluconazole versus amphotericin B for treatment of candidemia in nonneutropenic patients. Canadian Candidemia Study Group. Eur J Clin Microbiol Infect Dis 16: 337345

102. Anaissie EJ, Vartivarian SE, Abi-Said D, et al (1996) Fluconazole versus amphotericin B in the treatment of hematogenous candidiasis: a matched cohort study. Am J Med 101: 170176 103. Driessen M, Ellis JB, Cooper PA, et al (1996) Fluconazole vs. amphotericin B for the treatment of neonatal fungal septicemia: a prospective randomized trial. Pediatr Infect Dis J 15: 11071112 104. Graninger W, Presteril E, Schneeweiss B, Teleky B, Georgopoulos A (1993) Treatment of Candida albicans fungaemia with fluconazole. J Infect 26: 133146

You might also like