Abstract
In an international, multicenter, open-label, randomized comparative study, adult patients in intensive care units were enrolled to receive cefpirome intravenously at 2 g twice daily or ceftazidime intravenously at 2 g three times daily for the empiric treatment of pneumonia. Randomization was performed after a double stratification according to the investigator’s initial choice of monotherapy or combination therapy and then on the basis of the severity of disease. The primary endpoint was the clinical response at the end of treatment in the intent-to-treat population. Data for all patients were reviewed by a blinded observer. Of the 400 enrolled patients, 201 received cefpirome (monotherapy, 56%) and 199 received ceftazidime (monotherapy, 51%). Pneumonia was hospital acquired for 75% of the patients. Clinical failures rates were 34 versus 36% (odds ratio = 0.922; upper bound of 90% confidence interval = 1.301) in the intent-to-treat analysis for cefpirome and ceftazidime, respectively. For the cefpirome and ceftazidime groups, there were 35 versus 30% clinical failures among monotherapy-stratified patients, respectively, and 34 versus 42% clinical failures among combination therapy-stratified patients, respectively. The rates of clinical failures in the per-protocol analysis were 38 and 42%, respectively. In the population of patients evaluable for bacteriologic efficacy, eradication or presumed eradication was obtained for 71% (172 of 241) and 70% (162 of 230) of the pathogens isolated from the patients receiving cefpirome and ceftazidime, respectively. The mortality rates within 2 weeks after the end of treatment were similar (cefpirome group, 31%; ceftazidime group, 26%), as were the percentages of patients with at least one treatment-related adverse event (17 and 19%, respectively). An empiric treatment strategy with cefpirome at 2 g twice daily is equivalent in terms of efficacy and tolerance to ceftazidime at 2 g three times daily for the treatment of pneumonia in patients in intensive care units.
Bacterial pneumonia, whether it is community acquired or nosocomial, is still a major cause of morbidity and mortality, especially in patients hospitalized in intensive care units (ICUs) (8). When community-acquired pneumonia is associated with an indication for admission to an ICU, the mortality rate may be as high as 30% (18). Available data suggest that nosocomial pneumonia occurs at a rate of 5 to 10 cases per 1,000 hospital admissions, with the incidence increasing by as much as 6- to 20-fold in patients receiving mechanical ventilation (2). Recent data suggest that, in addition to the severity of the underlying disease, nosocomial pneumonia independently contributes to ICU patient mortality (9). In addition, it has recently been demonstrated that ventilator-associated pneumonia is responsible for an increase in the duration of mechanical ventilation and the duration of hospitalization (23). A recent study suggested that episodes of ICU-acquired pneumonia with inappropriate initial antibiotic coverage had a significantly higher mortality rate attributable to the pneumonia and to the number of related complications (1). Thus, administration of appropriate antibiotics is probably a key component in the treatment of severe bacterial pneumonia. An antimicrobial regimen is usually chosen empirically, before the results of the bacterial cultures are available. Severe pneumonia in the ICU is caused by a wide range of gram-negative and gram-positive pathogens. Many factors may influence the bacterial etiology such as the venue of acquisition (i.e., community versus hospital), previous administration of antibiotics (24), number and type of comorbidities, and the duration of hospitalization before the onset of pneumonia. Moreover, pneumonia in mechanically ventilated patients is often polymicrobial (10). Thus, the use of broad-spectrum cephalosporins alone or in combination with an aminoglycoside may be an attractive choice for the initial empiric treatment of pneumonia in ICU patients. However, the role of combination therapy for the treatment of severe pneumonia still remains controversial (7).
Cefpirome is a C-3′ quaternary ammonium cephalosporin which bears a 2,3-cyclopentenopyridinium at the C-3 position of the cephem nucleus. It belongs to the parenteral 2-amino-5-thiazolyl cephalosporins (6). It displays a broad antibacterial spectrum including difficult-to-treat gram-negative bacilli, such as members of the family Enterobacteriaceae producing class I β-lactamases. This property is partly due to the zwitterionic structure of the compound, which allows for the rapid penetration through the outer membrane of gram-negative bacilli and a high affinity for penicillin-binding proteins (3, 20). Cefpirome has a low affinity for many β-lactamases of the periplasmic space. Cefpirome displays a well-balanced antibacterial spectrum including gram-positive cocci such as methicillin-susceptible Staphylococcus aureus strains and Streptococcus pneumoniae isolates resistant to penicillin G (13). Cefpirome shows an apparent elimination half-life of about 2 h, and 80% of the administered dose is mainly excreted unchanged in the urine.
A retrospective analysis of 4,180 patients entered in 15 phase II and phase III trials suggests that cefpirome is as active as ceftazidime in the treatment of severe bacteremia in ICU patients (22).
The purpose of this prospective study was to compare strategies in which a standard regimen of 2 g of ceftazidime three times daily or 2 g of cefpirome twice daily was used for the empiric parenteral treatment of pneumonia in ICU patients. A double stratification based on the choice between monotherapy or combination therapy and on the initial severity of disease in the patients was used.
(Part of this work was presented at the 36th Interscience Conference on Antimicrobial Agents and Chemotherapy, New Orleans, La., 15 to 18 September 1996 [27]).
MATERIALS AND METHODS
Study design.
This late phase III study was a prospective, randomized, open-label, multicenter trial comparing the efficacies (equivalence testing) and safety of cefpirome and ceftazidime for the empiric therapy of pneumonia in adult patients requiring intensive care. The data were analyzed both by the intent-to-treat method and by analysis of patients with pneumonia (per-protocol analysis). Two hundred patients were to be enrolled in each group. Each of the 24 participating centers was expected to enroll between 16 and 48 patients.
The study was conducted unblinded due to the difficulty of disguising dissimilar formulations and dosing frequencies. A blinded observer, independent from the investigators, had to review data for all 400 patients to confirm the diagnosis of bacterial pneumonia, to assess the causality of the pathogens isolated during the infective episode, to evaluate the clinical and bacteriological efficacies of the study medications at the end of treatment and at follow-up, and to give a possible cause(s) for all deaths that occurred during the study. For these evaluations, the blinded observer was provided with study books, copies of the initial chest X ray, summaries of serious adverse events, discharge summaries, and autopsy reports, when available. All documents were provided without any mention of the allocated treatment to ensure independent review. This review was done according to the rules detailed in the Definitions and Efficacy Analyses sections below, and the blinded observer’s assessments were those taken into account for the intent-to-treat and per-protocol efficacy analyses.
At the time of inclusion in the study participants were preferentially treated with monotherapy, but combination therapy with an aminoglycoside or metronidazole was permitted by the protocol if Pseudomonas aeruginosa or anaerobic infections were suspected. Concomitant administration of additional antibacterial agents was allowed if the patient’s clinical condition or bacteriologic results necessitated such an addition.
Randomization was performed after a double stratification of eligible patients according to the investigator’s initial choice of monotherapy or combination therapy and then on the basis of the severity of disease, as assessed by the Ambulatory Simplified Acute Physiology Score (ASAPS) (17). ASAPS was derived from the Simplified Acute Physiology Score (16) and included only clinical variables (age, systolic blood pressure, temperature, heart rate, Glasgow coma score, respiratory rate, or use of mechanical ventilation). Scores of <10 or ≥10 determined severity group classification. Thus, patients were stratified into four groups before being assigned to one of the two treatment regimens according to a computer-generated randomization schedule, using a central telephone randomization system.
Selection of subjects.
Four hundred patients were to be recruited at 24 hospitals in eight countries (Argentina, Canada, France, Italy, The Netherlands, South Africa, Sweden, and United Kingdom). Patients who were ages 18 years or older and who were hospitalized in ICUs with clinical signs and symptoms of community-acquired or nosocomial pneumonia were eligible for entry into the study. Patients who had received more than two doses of antibiotics for the current episode of pneumonia within 3 days before enrollment were eligible for participation in the study if they had not improved clinically or if the pathogen(s) that was known to be resistant to the previous therapy persisted. Subjects receiving antibiotics for nonpulmonary infections were also eligible for participation in the study and were assigned to the combination groups when therapy was continued after enrollment. Patients were entered into the study only if both cefpirome and ceftazidime were considered appropriate for initial treatment before bacteriologic results became available.
Patients were excluded from participation in the study if they had suspected nonbacterial pneumonia, cystic fibrosis, or pulmonary tuberculosis, were pregnant or lactating, had received any investigational drug within 2 weeks prior to entry, had concomitantly received aerosolized antibiotics or parenteral antibiotics for gut decontamination, required hemodialysis or hemofiltration, had a history of hypersensitivity to β-lactam antibiotics, or were previously enrolled in this study. Patients were also excluded if culture results, available before enrollment, showed that the patients were infected with pathogens resistant to either cefpirome or ceftazidime.
The protocol was approved by each study center’s ethics committee, and all participants (or next of kin) gave informed consent.
Definitions.
The following criteria were used to diagnose bacterial pneumonia: at least one symptom of pneumonia (cough, pleuritic pain, dyspnea, purulent sputum, or purulent tracheal secretions) or auscultatory evidence of pneumonia, with fever (temperature, >38°C) or leukocytosis (>10,000 cells/mm3) and on a chest X ray a pulmonary infiltrate thought to be due to infection. Pneumonia was diagnosed as definite when pathogens were isolated from at least one of the following samples: blood, pleural fluid, protected specimen brush (PSB) (≥103 CFU/ml), or bronchoalveolar lavage (BAL) (≥104 CFU/ml) samples. Community-acquired pneumonia was also considered definite when pathogens were isolated from sputum or tracheal aspirates containing less than 10 squamous epithelial cells and more than 25 polymorphonuclear cells per oil-immersion field with a count of ≥107 CFU/ml or a score of ≥+++. Pneumonia was diagnosed as possible in case of clinical and radiological signs with no microbiologic documentation. Nosocomial pneumonia was also considered possible when pathogens were isolated from tracheal aspirates from intubated patients or sputum from nonintubated patients without a quantitative culture.
Study treatments.
Patients were randomly assigned to receive cefpirome at 2 g intravenously (i.v.) every 12 h or ceftazidime at 2 g i.v. every 8 h. Both drugs were administered by 2-min i.v. injections or as ≤30-min infusions. The expected duration of treatment was 5 to 15 days unless either complete cure or therapeutic failure requiring another antibiotic occurred. For patients with impaired renal function, the dosages of both agents were adjusted (for cefpirome, creatinine clearance [CLCR] of 21 to 50 ml/min; 1 g every 12 h, and CLCR of ≤20 ml/min, 1 g every 24 h; for ceftazidime, CLCR of 31 to 50 ml/min, 1 g every 12 h; CLCR of 16 to 30 ml/min, 1 g every 24 h; CLCR of 6 to 15 ml/min, 0.5 g every 24 h; and CLCR of ≤5 ml/min, 0.5 g every 48 h).
Clinical assessments.
A medical history was recorded, and a physical examination and a chest X ray were performed within 24 h prior to starting therapy. The demographic features of the participants were recorded, as were an assessment of the underlying condition, the severity of infection, ASAPS (for stratification before randomization), and the Acute Physiology and Chronic Health Evaluation (APACHE II) (14). Physical examination and chest X ray were repeated after 48 h of treatment and within 24 h of stopping therapy. The patient’s temperature, pulse rate, and systolic blood pressure were monitored daily during the study period, and assessment was continued for 2 weeks following the completion of therapy. A follow-up physical examination and chest X ray were performed 2 weeks posttreatment, unless there was previous evidence of radiological resolution of pneumonia.
Bacteriologic assessments.
Pleural fluid and blood specimens (when possible) were taken for bacteriologic culture and susceptibility testing within 48 h prior to the beginning of treatment with a study drug. Specimens for quantitative culture were obtained by protected brush sampling, BAL, or transtracheal aspiration. If it was not possible to obtain such samples, specimens of tracheobronchial aspirate or expectorated sputum were examined. Gram-stained specimens of each aspirate or sputum specimen were examined to determine the number of leukocytes and epithelial cells per oil-immersion field.
All isolates were identified by standard bacteriologic methods, and susceptibilities to cefpirome and ceftazidime were determined at each center by the disk diffusion method with 30-μg disks. Isolates were considered resistant if they had zone diameters of ≤14 mm.
Subsequent samples for culture were taken during therapy, at least for all patients who failed treatment, and before modification of the initial treatment regimen. Cultures were repeated within 48 h posttreatment, unless there was evidence of a complete clinical cure, and at the follow-up assessment for patients with relapses and those with signs of reinfection.
Safety assessments.
All adverse events, whether observed by the investigator or reported by the patient, were recorded at each visit and were classified according to their severity and causal relationship to the study treatments. Deaths were recorded during and up to 14 days after the completion of therapy.
Efficacy analyses. (i) Primary endpoint.
The primary efficacy analysis was the analysis of the clinical response at the end of treatment (within 48 h posttherapy) in the intent-to-treat population (all randomized patients were analyzed as treated). Clinical response was assessed by the blinded observer and was classified as success without modification of the initial therapy, success with modification of the initial therapy (the addition of another antibiotic for any reason), and indeterminate (nonbacterial pneumonia or the presence of another disease interfering with treatment assessment). Clinical failure was defined as the presence of at least one of the following: persistence or progression of symptoms of pneumonia or worsening of chest X ray leading to replacement of study drugs by another antibiotic (switch), invasive superinfection, or death due to pneumonia or superinfection (definitions of causes of death are detailed in the Safety Analysis section).
(ii) Secondary endpoints.
In addition to the intent-to-treat analysis, patients with proven or possible pneumonia were analyzed separately by using the same definitions for clinical responses (per-protocol analysis).
Bacteriologic response at the end of treatment was expressed in terms of eradication, presumed eradication, persistence, or indeterminate for the originally isolated pathogen(s). In addition, colonization or superinfection was recorded. Eradication was defined as the absence of the original pathogen(s) from the same site (pulmonary or blood sample) during or within 48 h after the completion of therapy. Presumed eradication was considered in the absence of specimens for culture when patients had clinically improved. Persistence occurred when the original pathogen was still present during the course of therapy or within 48 h after the completion of therapy, and indeterminate meant that there was incomplete bacteriologic data, death, or withdrawal of patients before follow-up samples for culture could be taken. Colonization was classified as one or more new isolates present in a culture of a sample taken after 48 h of the start of therapy but not associated with infection. Superinfection was defined as an infection caused by an organism different from the original pathogen, with the infection developing during treatment or within 48 h of stopping therapy or with the infection diagnosed at autopsy. Superinfection was considered invasive if the new pathogen was isolated from cultures of blood or a major organ such as the lung.
Overall, the bacteriologic response by a patient was considered satisfactory when all causative pathogens were eradicated or presumed to be eradicated with no occurrence of invasive superinfection. For all other patients the response was unsatisfactory.
Clinical and bacteriologic responses at follow-up assessment (14 days posttherapy) were also considered as secondary efficacy criteria. The bacteriologic response at follow-up was considered a failure if relapse or reinfection occurred. Relapse was classified as the recurrence of pulmonary infection with the same causative organism(s) and in which the organism(s) was isolated within 2 weeks after discontinuing treatment. Reinfection was classified as the elimination of the initial infecting pathogen followed by replacement with a new species (or a new biotype of the same organism) in pulmonary samples or blood in the presence of signs and symptoms of infection within 2 weeks of the completion of therapy.
Safety analysis.
All patients who received at least one dose of study treatment were evaluable for the safety analysis. Only deaths and adverse events occurring during therapy or up to 14 days after the completion of therapy were recorded and accounted for in the analysis. In case of death, the cause was evaluated by the blinded observer and was defined as being due to either pneumonia, invasive superinfection, underlying disease, an adverse event, or inadequate antibiotic support. Death was considered due to pneumonia if at least one of the following situations was present at the time of inclusion: severe hypoxia (partial arterial O2 pressure of <70 mm Hg, with fractional inspired O2 concentration of >50%), persistent hypotension (arterial blood pressure of <90 mm Hg, despite adequate resuscitation, and a requirement for vasopressors), dysfunction of more than one organ, or iatrogenic complication of any origin during the course of pneumonia. Death was considered due to invasive superinfection when at least one of the situations mentioned above was present at the time of invasive superinfection. Inadequate antibiotic support was considered the cause of death if antibiotic treatment was not active against the pathogen responsible for infection at the time of the patient’s inclusion in the study or the pathogen responsible for invasive superinfection. The death was considered to be related to inadequate surgical support when an extrapulmonary focus requiring a surgical procedure was not performed.
Statistical methods. (i) Sample size calculation.
The sample size calculation was based on equivalence testing for clinical success rates at the end of treatment in the intent-to-treat population. Assuming similar success rates in both treatment groups (70 to 80%) and a δ value (maximum difference between treatments to be accepted as equivalent) of 10%, 200 patients per group were needed to provide a power of 80% (β, 20%). Under these assumptions, the trial produced a 90% confidence interval (α; 5% one-sided) for the difference between the clinical success rates for the two treatments that excluded the prespecified δ value of 10%.
(ii) Baseline comparison of treatment groups.
The treatment groups at study entry were compared by using Student’s t test for continuous variables and the chi-square test for categorical variables.
(iii) Efficacy analysis.
A logistic regression approach was used. This allowed for adjustments for severity scores, type of therapy, and center. An odds ratio of 1.714, being equivalent to a 10% difference with 75% success rates, was substituted in the analysis as the limit of equivalence. The estimated odds ratio and the upper bound of its 95% confidence interval were calculated by means of the logistic regression. The two treatments were considered equivalent if the upper bound of the confidence interval was less than 1.714.
Analyses of robustness were then performed by logistic regression after adjustments for each severity score and type of therapy and after adjustment for center or country. Results from centers that recruited fewer than 16 patients were pooled with data from other centers in the same country for the logistic regression analyses. The possible interactions between treatment and severity scores or type of therapy were also explored by logistic regression. The homogeneity across centers or countries was tested by the Breslow and Day test.
(iv) Safety analysis.
The frequencies of patients with at least one adverse event possibly or probably related to study treatment were compared between treatment groups by a chi-square test. Death rates between treatment groups were compared by logistic regression with two models in which death was the dependent variable and in which type of therapy and APACHE II score were independent variables in the first model and study treatment was an additional independent variable in the second model. The deviance was obtained by a chi-square test for treatment effect, with the P value compared to 5%.
(v) Advisory board committee.
Considering the high rate of mortality expected in the population studied, interim analyses of mortality were performed for each 50 randomized patients who received at least one injection of cefpirome or ceftazidime and were reviewed by an independent advisory board committee. Its members were provided with the results of the interim analyses and with tabulated data on the distribution of patients regarding the severity scores at the time of inclusion in the study and regarding the different items included in these scores. This documentation was sent 3 weeks after all 50 recruits had finished their follow-up period (14 days after the end of treatment). Upon review of these data and within 15 days after receipt of the documents, each member had to give his written decision on the continuation or termination of the study.
It was planned in the protocol that the α level (5% significance level) be used at the end of the trial if the study was not to be prematurely terminated. All of these analyses were performed with SAS software, release 6.08 (SAS Institute Inc., Cary, N.C.), running on a Vax6000 computer.
RESULTS
Study population.
During a 26-month period (February 1993 to March 1995), 400 adult patients with clinically suspected pneumonia from ICUs in eight countries (Canada, 27%; South Africa, 20%; France, 13%; Argentina, 13%; United Kingdom, 11%; Sweden, 10%; Italy, 6%; The Netherlands, <1%) were enrolled and evaluated in the study described here. Of these, 201 were treated with cefpirome and 199 received ceftazidime. Two patients, both in the cefpirome group, stratified to a monotherapy treatment group were given combination therapy, and an additional patient, randomized to receive cefpirome, was given ceftazidime. These patients were analyzed according to the treatment that they actually received. One patient in the cefpirome group died before administration of the first dose and therefore was unevaluable for the safety analysis. Although this patient was diagnosed as having definite pneumonia, he was also unevaluable for the per-protocol efficacy analysis. However, data for this patient were considered evaluable for the intent-to-treat efficacy analysis and were analyzed according to the treatment group to which he was randomly allocated.
Overall, the 400 patients were included in the intent-to-treat clinical efficacy analysis and 399 patients (cefpirome, n = 200; ceftazidime, n = 199) were evaluable for the safety analysis. Of the 400 patients included in the intent-to-treat population, 170 (43%) were diagnosed as having definite pneumonia and 183 (46%) had a possible pneumonia. Thus, excluding the patient who did not receive the treatment, 352 patients (cefpirome, n = 180; ceftazidime, n = 172) were evaluable for the per-protocol clinical efficacy analysis. Of these patients, 277 (cefpirome, n = 139; ceftazidime, n = 138) were evaluable for the bacteriologic efficacy analysis. In addition, 3.8% of the cases of pneumonia were considered to be of nonbacterial origin and 8% were of indeterminate origin. There was no significant difference in the classification of pneumonia between the two treatment groups in the intent-to-treat population (P = 0.26). Three-quarters of the infections (298 of 400) were hospital acquired.
The demographic and background characteristics of the evaluable patients in the intent-to-treat population are summarized in Table 1. There were no differences between the study groups in the distribution of patients according to age, sex, comorbid conditions, number of patients with renal failure or hypotension, and previous systemic antibiotic therapy. At the time of inclusion in the study 275 patients (69%) were mechanically ventilated, with the mean partial arterial O2 pressure being 89 ± 34 mm Hg at a mean fractional inspired oxygen concentration of 53% ± 20%. Significantly more patients receiving cefpirome required mechanical respiratory assistance (cefpirome, n = 150 [75%]; ceftazidime, n = 125 [63%]; P = 0.01), with this item being part of the ASAPS on the basis of which the efficacy results were adjusted. The mean ± standard deviation (SD) ASAPSs at the time of inclusion in the study were 7.9 ± 2.7 and 7.6 ± 2.8 for the cefpirome and ceftazidime groups, respectively (P = 0.28). The mean ± SD APACHE II scores at the time of inclusion in the study were 16.3 ± 7.1 and 15.4 ± 6.4 for the cefpirome and ceftazidime groups, respectively (P = 0.19).
TABLE 1.
Patient characteristics at the time of inclusion in the study
| Characteristic | Intent-to-treat population
|
|
|---|---|---|
| Cefpirome | Ceftazidime | |
| No. of patients enrolled | 201 | 199 |
| No. of males (% total) | 139 (69) | 137 (69) |
| Age (yr [mean ± SD]) | 56.0 ± 19.2 | 57.1 ± 17.8 |
| ASAPS (mean ± SD) | 7.9 ± 2.7 | 7.6 ± 2.8 |
| APACHE II score (mean ± SD) | 16.3 ± 7.1 | 15.4 ± 6.4 |
| No. (%) of patients with APACHE II score of >20 | 52 (26) | 40 (20) |
| No. (%) of patients on mechanical ventilation | 150 (75) | 125 (63)a |
| No. (%) of patients with comorbid conditions | ||
| Underlying pulmonary disease | 91 (45) | 80 (40) |
| Chronic obstructive pulmonary disease | 52 (26) | 43 (22) |
| Cancer | 7 (4) | 5 (3) |
| Cardiovascular disease | 85 (42) | 81 (41) |
| Alcoholism | 30 (15) | 36 (18) |
| Malnutrition | 20 (10) | 13 (7) |
| Diabetes mellitus | 29 (14) | 26 (13) |
| Cancer (other than pulmonary) | 17 (9) | 17 (9) |
| Systolic blood pressure of ≤90 mmHg | 25 (12) | 22 (11) |
| Serum creatinine level of ≥1.7 mg/dl | 49 (24) | 46 (23) |
| No. (%) patients with | ||
| Nosocomial pneumonia | 149 (74) | 149 (75) |
| Community-acquired pneumonia | 52 (26) | 50 (25) |
| No. (%) patients with previous systemic antibiotic therapy | 122 (61) | 110 (55) |
P = 0.01.
Study treatments.
Figure 1 presents a diagram of the stratification and randomization before treatment. At enrollment, 298 (75%) patients had an ASAPS of <10, and of these patients, 170 (57%) received monotherapy and 128 (43%) received combination therapy. Thus, a quarter (26%) of the participants had ASAPSs of ≥10, and of these 45 (44%) were treated with monotherapy and 57 (56%) received combination therapy. The median duration of treatment was 8 days for the cefpirome group (range, 1 to 24 days) and 9 days for the ceftazidime group (range, 1 to 35 days).
FIG. 1.
Overview of study population. Numbers indicate numbers of patients.
Concomitant treatments.
A total of 88 of 201 (44%) and 97 of 199 (49%) of patients treated with cefpirome and ceftazidime, respectively, received combination therapy at the time of inclusion in the study. Among them, aminoglycosides were the most commonly prescribed class of drugs, and these were given to 33 of 88 (37.5%) patients treated with cefpirome and 40 of 97 (41%) patients treated with ceftazidime.
Among the 215 patients initially treated with monotherapy, 38 (18%) patients received an additional antibiotic during the study, 44 (21%) patients had a switch of antibiotics, and 18 (8%) patients received additional antibiotics which continued beyond the study. Of the 185 patients initially treated with combination antibiotics, 26 (14%) received a supplementary antibiotic, 49 (27%) had their initial medication switched, and 20 (11%) received additional antibiotics beyond the study period. The cefpirome and ceftazidime study groups were comparable in terms of antibiotic strategy during treatment.
A total of 137 different antibiotic regimens were added to the treatment regimens for 102 patients during the study treatment. Glycopeptides, aminoglycosides, and metronidazole were added to the treatment regimens for 16 (8%) and 21 (11%), 15 (8%) and 11 (6%), and 11 (5%) and 8 (4%) cefpirome and ceftazidime recipients, respectively. There was no difference in the use of concurrent antibiotics between the two treatment groups.
Bacteriologic documentation.
A total of 471 pathogens isolated from tracheal aspiration (33%), PSB (25%), sputum (18%), BAL (15%), blood (8%), and pleural fluid (1%) specimens were identified in 277 patients (69%). Polymicrobial infections occurred in 131 patients (66 in the cefpirome group and 65 in the ceftazidime group). The proportions of pretreatment isolates in each study group were similar, being 241 in the cefpirome group and 230 in the ceftazidime group. S. aureus, Haemophilus influenzae, P. aeruginosa, S. pneumoniae, and Escherichia coli were the most frequently isolated pathogens (Table 2). Eighty-two percent (n = 343) of the isolates tested for their susceptibilities to both study drugs (421 of 471) were susceptible to cefpirome and ceftazidime. A total of 62 of 421 (15%) and 57 of 421 (14%) of the isolates were resistant to cefpirome and ceftazidime, respectively. Of these, 41 were resistant to both study drugs. Irrespective of the treatment group for the patients from whom they were isolated, the most common resistant organisms were methicillin-resistant S. aureus, Acinetobacter spp., and P. aeruginosa. There was no significant difference between treatment groups on the basis of susceptibility testing (P = 0.41).
TABLE 2.
Pretreatment isolatesa
| Organism group and organism | No. (%) of isolates
|
||
|---|---|---|---|
| Cefpirome group | Ceftazidime group | Total | |
| Gram-positive organisms | 77 | 77 | 154 (33) |
| Staphylococcus aureus | 27 | 30 | 57 (12) |
| Other staphylococci | 3 | 1 | 4 |
| Streptococcus pneumoniae | 20 | 21 | 41 (9) |
| Enterococcus faecium | 1 | 0 | 1 |
| Other streptococci | 24 | 24 | 48 |
| Others | 2 | 1 | 3 |
| Gram-negative organisms | 164 | 153 | 317 (67) |
| Haemophilus influenzae | 33 | 21 | 54 (12) |
| Pseudomonas aeruginosa | 21 | 28 | 49 (10) |
| Escherichia coli | 18 | 18 | 36 (8) |
| Acinetobacter spp. | 12 | 15 | 27 |
| Enterobacter spp. | 14 | 10 | 24 |
| Klebsiella pneumoniae | 8 | 14 | 22 |
| Proteus spp. | 6 | 5 | 11 |
| Morganella morganii | 6 | 3 | 9 |
| Moraxella catarrhalis | 8 | 6 | 14 |
| Serratia spp. | 6 | 5 | 11 |
| Klebsiella spp. | 5 | 5 | 10 |
| Neisseria spp. | 5 | 2 | 7 |
| Stenotrophomonas maltophilia | 3 | 2 | 5 |
| Haemophilus spp. | 4 | 4 | 8 |
| Bacteroides spp. | 2 | 3 | 4 |
| Others | 13 | 12 | 25 |
A total of 66 and 65 patients in the cefpirome and ceftazidime groups, respectively, had polymicrobic infections (47% of patients with documented infections). A total of 73 patients in each group had monomicrobic infections (53% of patients with documented infections).
Of the patients with bacteriologically documented infections, 223 (81%) had hospital-acquired pneumonia caused mainly by gram-negative bacteria (73%; 284 of 388). However, S. aureus was the single most prevalent pathogen in these patients, accounting for 21% of infections in cefpirome-treated patients and 23% of infections in ceftazidime-treated patients. Overall, the two groups of patients were comparable in terms of the origin of pneumonia and the distribution of pathogens (Tables 1 and 2).
Clinical efficacy.
At the end of treatment, 52% (104 of 201) of the patients in the intent-to-treat population responded successfully to an empiric cefpirome regimen, either monotherapy or combination therapy, with or without modification during treatment (Table 3). Similarly, ceftazidime regimens resulted in a combined success rate of 45% for monotherapy and combination therapy (90 of 199). Success rates for monotherapy (52 versus 44%) were comparable, as were the number of patients who responded successfully to combination therapy (51 versus 46%). A total of 28 of 201 (14%) patients receiving cefpirome and 37 of 199 (19%) patients receiving ceftazidime were classified as indeterminate for clinical efficacy, with the frequencies being not significantly different (P = 0.20). Therapeutic failure occurred in 141 patients during the treatment period: 69 patients treated with cefpirome (34%) and 72 patients receiving ceftazidime (36%). Failures were attributed to the persistence of pneumonia (n = 73), death directly related to pneumonia (n = 43), or invasive superinfection (n = 27, resulting in the deaths of six patients). The reasons for clinical failure were similar in both treatment groups, irrespective of the initial monotherapy or combination therapy strategy. The estimated odds ratio for clinical failure rates was 0.922, and because the upper bound of its 90% confidence interval (1.301) was less than the prespecified limit of equivalence (1.714), cefpirome was equivalent to ceftazidime in terms of clinical efficacy (primary endpoint).
TABLE 3.
| Population and therapy | No. (%) of patients with the indicated clinical response to the following treatment:
|
|||||||
|---|---|---|---|---|---|---|---|---|
| Cefpirome
|
Ceftazidime
|
|||||||
| Successful without modification | Successful with modification | Indeterminate | Failed | Successful without modification | Successful with modification | Indeterminate | Failed | |
| Intent-to-treatc | ||||||||
| Monotherapy | 44 (38.9) | 15 (13.3) | 15 (13.3) | 39 (34.5) | 38 (37.3) | 7 (6.9) | 26 (25.5) | 31 (30.4) |
| Combination therapy | 37 (42.0) | 8 (9.1) | 13 (14.8) | 30 (34.1) | 37 (38.1) | 8 (8.2) | 11 (11.3) | 41 (42.3) |
| All patients | 81 (40.3) | 23 (11.4) | 28 (13.9) | 69 (34.3) | 75 (37.7) | 15 (7.5) | 37 (18.6) | 72 (36.2) |
| APACHE II score | ||||||||
| ≤20 | 68 (45.6) | 19 (12.8) | 21 (14.1) | 41 (27.5) | 61 (38.4) | 12 (7.5) | 32 (20.1) | 54 (34.0) |
| >20 | 13 (25.0) | 4 (7.7) | 7 (13.5) | 28 (53.8) | 14 (35.0) | 3 (7.5) | 5 (12.5) | 18 (45.0) |
| Per protocolb,d | ||||||||
| Monotherapy | 44 (43.6) | 15 (14.9) | 4 (4.0) | 38 (37.6) | 38 (46.3) | 7 (8.5) | 6 (7.3) | 31 (37.8) |
| Combination therapy | 37 (46.8) | 8 (10.1) | 4 (5.1) | 30 (38.0) | 37 (41.1) | 8 (8.9) | 4 (4.4) | 41 (45.6) |
| All patients | 81 (45.0) | 23 (12.8) | 8 (4.4) | 68 (37.8) | 75 (43.6) | 15 (8.7) | 10 (5.8) | 72 (41.9) |
| Bacteriologically evaluableb,e | ||||||||
| Monotherapy | 34 (41.5) | 14 (17.1) | 3 (3.7) | 31 (37.8) | 31 (45.6) | 6 (8.8) | 6 (8.8) | 25 (36.8) |
| Combination therapy | 25 (43.9) | 6 (10.5) | 2 (3.5) | 24 (42.1) | 29 (41.4) | 5 (7.1) | 2 (2.9) | 34 (48.6) |
| All patients | 59 (42.4) | 20 (14.4) | 5 (3.6) | 55 (39.6) | 60 (43.5) | 11 (8.0) | 8 (5.8) | 59 (42.8) |
Within 48 h of completion of therapy.
Determined by the blinded observer.
Number of evaluable patients: cefpirome group, n = 201; ceftazidime group, n = 199.
Number of evaluable patients: cefpirome group, n = 180; ceftazidime group, n = 172.
Number of evaluable patients: cefpirome group, n = 139; ceftazidime group, n = 138.
Adjustments performed to account for possible prognostic factors showed that high APACHE II scores and ASAPS appear to be significantly predictive of treatment failure (P = 0.0001). In contrast, the choice of empiric therapy with single agents or combinations of agents was unrelated to treatment outcome (P = 0.22). The estimated odds ratio for treatment effect after adjustments for these factors was 0.8, with the upper bound of the 90% confidence interval being 1.2. In addition, no interaction between treatments and APACHE II score or type of therapy was found (P = 0.48 or P = 0.20, respectively).
In different centers or countries, the odds ratios did not differ significantly (P = 0.67 or P = 0.44, respectively). After adjustments for center and country effects, estimated odds ratios and the upper bounds of the 90% confidence intervals were 0.938 and 1.341 and 0.929 and 1.324, respectively.
An additional analysis of robustness was performed by considering as failures the responses classified as indeterminate. The results were an odds ratio of 0.924 and an upper bound of the 90% confidence interval of 1.292.
In addition, for the intent-to-treat population, the clinical failure rates for cefpirome and ceftazidime were similar for community-acquired pneumonia (42 versus 40%) and nosocomial pneumonia (32 versus 35%). Finally, the clinical failure rates for the overall population were 55% (27 of 49), 47% (27 of 57), and 18.5% (10 of 54) for P. aeruginosa, S. aureus, and H. influenzae, respectively.
In the per-protocol population, cefpirome was associated with an overall successful response in 58% (104 of 180) of the patients; among the patients in the ceftazidime group, 52% (90 of 172) responded successfully (Table 3). The responses for a total of 8 of 180 (4%) and 10 of 172 (6%) patients receiving cefpirome and ceftazidime, respectively, were classified as indeterminate. Clinical failure rates for patients receiving cefpirome and ceftazidime were 38 and 42%, respectively. The estimated odds ratio for clinical failure was 0.843, and the upper bound of its 90% confidence interval was 1.206, which is also less than the prespecified limit for equivalence.
At the follow-up visit, patients in both treatment groups sustained an overall success rate of more than 40%: 43 versus 41% for the intent-to-treat population and 48 versus 48% for the per-protocol population for cefpirome and ceftazidime, respectively.
Bacteriologic efficacy.
Eradications and presumed eradications were obtained for 172 of 241 (71%) pathogens isolated from patients in the cefpirome group and for 162 of 230 (70%) pathogens isolated from patients in the ceftazidime group; cefpirome was associated with failure due to the persistence of 9% (21 of 241) of pretreatment pathogens, and ceftazidime was associated with a failure rate of 9% (20 of 230). The distributions of persistent pathogens were similar in both groups (Table 4). Of these organisms, gram-negative bacilli predominated and P. aeruginosa was the most common. There was also no significant difference between treatment groups in either the rate of colonization or the rate of superinfection (Table 5).
TABLE 4.
Bacteriologic responses by pathogen at the end of treatmenta
| Organism group and organism | No. of pretreatment isolates with the following response to the indicated treatment:
|
|||||
|---|---|---|---|---|---|---|
| Eradicated or presumed eradicated
|
Persisted
|
Indeterminate
|
||||
| Cefpirome | Ceftazidime | Cefpirome | Ceftazidime | Cefpirome | Ceftazidime | |
| Gram-positive organisms | 52 | 61 | 5 | 5 | 20 | 11 |
| Staphylococcus aureus | 14 | 19 | 5 | 3 | 8 | 8 |
| Streptococcus pneumoniae | 16 | 21 | 0 | 0 | 4 | 0 |
| Enterococcus faecium | 0 | 0 | 0 | 0 | 1 | 0 |
| Other streptococci | 20 | 19 | 0 | 2 | 4 | 3 |
| Others | 2 | 2 | 0 | 0 | 3 | 0 |
| Gram-negative organisms | 120 | 101 | 16 | 15 | 28 | 37 |
| Haemophilus influenzae | 33 | 20 | 0 | 0 | 0 | 1 |
| Pseudomonas aeruginosa | 8 | 13 | 7 | 8 | 6 | 7 |
| Escherichia coli | 16 | 15 | 0 | 1 | 2 | 2 |
| Acinetobacter spp. | 4 | 3 | 2 | 4 | 6 | 8 |
| Enterobacter spp. | 11 | 6 | 1 | 1 | 2 | 3 |
| Klebsiella pneumoniae | 7 | 12 | 0 | 0 | 1 | 2 |
| Proteus spp. | 3 | 4 | 2 | 0 | 1 | 1 |
| Morganella morganii | 6 | 3 | 0 | 0 | 0 | 0 |
| Moraxella catarrhalis | 8 | 5 | 0 | 0 | 0 | 1 |
| Serratia spp. | 4 | 3 | 1 | 0 | 1 | 2 |
| Stenotrophomonas maltophilia | 1 | 1 | 1 | 1 | 1 | 0 |
| Others | 19 | 16 | 2 | 0 | 8 | 10 |
| Total | 172 | 162 | 21 | 20 | 48 | 48 |
Within 48 h of completion of therapy.
TABLE 5.
Colonizations and superinfections at the end of treatmenta
| Organism group and organism | No. of pretreatment isolates causing the following for the indicated treatment group:
|
|||
|---|---|---|---|---|
| Colonization
|
Superinfection
|
|||
| Cefpirome | Ceftazidime | Cefpirome | Ceftazidime | |
| Gram-positive organisms | 30 | 38 | 10 | 6 |
| Staphylococcus aureus | 6 | 4 | 5 | 4 |
| Enterococci | 5 | 4 | 2 | 0 |
| Other streptococci | 10 | 13 | 1 | 1 |
| Others | 9 | 17 | 2 | 1 |
| Gram-negative organisms | 40 | 30 | 12 | 15 |
| Pseudomonas aeruginosa | 6 | 5 | 4 | 2 |
| Escherichia coli | 2 | 2 | 0 | 1 |
| Acinetobacter spp. | 9 | 10 | 1 | 5 |
| Enterobacter spp. | 0 | 3 | 0 | 1 |
| Klebsiella pneumoniae | 2 | 0 | 1 | 2 |
| Proteus spp. | 2 | 1 | 0 | 0 |
| Serratia spp. | 1 | 0 | 0 | 0 |
| Stenotrophomonas maltophilia | 7 | 4 | 2 | 0 |
| Others | 11 | 5 | 4 | 4 |
| Others | 14 | 18 | 4 | 1 |
| Candida albicans | 6 | 8 | 1 | 1 |
| Candida tropicalis | 1 | 1 | 1 | 0 |
| Torulopsis glabrata | 0 | 1 | 2 | 0 |
| Others | 4 | 5 | 0 | 0 |
Within 48 h of completion of therapy.
In the bacteriologic efficacy population, invasive superinfections occurred in 33 patients (17 cefpirome recipients and 16 ceftazidime recipients). The main pathogens were S. aureus (n = 9, including 5 methicillin-resistant S. aureus isolates), P. aeruginosa (n = 6), and Acinetobacter species (n = 6). Five patients developed fungemia caused by Candida albicans (n = 2), Torulopsis glabrata (n = 2), and Candida tropicalis (n = 1) (Table 5). Overall, the bacteriologic responses were considered satisfactory for 57% of the patients in both groups.
The clinical response rates for episodes of documented infection in relation to the initial treatment strategy were similar for both treatment groups (Table 3). Among patients that were evaluable for bacteriologic efficacy, 48 of 82 (59%) patients given monotherapy and 31 of 57 (54%) patients given combination therapy responded to cefpirome. Clinical response rates of 54% (37 of 68) and 49% (34 of 70) were achieved for patients who received ceftazidime as monotherapy or in combination, respectively.
At follow-up, for the cefpirome group a total of eight patients had evidence of reinfection and one patient had a relapse, whereas for the ceftazidime group three patients had relapses and two were reinfected.
Mortality.
Seven interim analyses of the mortality rates were performed during the study (after each 50 patients was recruited). No statistical result led to premature discontinuation of the study.
One hundred thirteen (28%) of the 399 patients who received at least one dose of cefpirome or ceftazidime died during the study. In total, 61 deaths (31%) occurred in the cefpirome group, with 35 of these occurring during treatment or within 2 days after completing therapy and with 26 occurring between 3 and 14 days after treatment was completed. Ninety-one possible causes of death were attributed to these 61 deaths. Underlying medical conditions contributed to 36 of the deaths, and 30 fatalities were directly related to pneumonia. In 10 patients, an inappropriate initial antibiotic regimen contributed to death. Another eight patients died from invasive superinfections, five patients had indeterminate causes of death, one patient died from an adverse event unrelated to the study drug, and one patient died as a result of failure to perform necessary surgery. Fifty-two deaths (26%) occurred in the ceftazidime group, with 34 of these occurring during treatment or within 2 days of the end of therapy and with 18 occurring between 3 and 14 days after treatment was completed. Eighty-three possible causes of death were attributed to these 52 deaths: underlying diseases (n = 29), pneumonia (n = 27), invasive superinfection (n = 10), initial antibiotic strategy (n = 8), unknown (n = 4), adverse event unrelated to study drug (n = 4), or failure to perform necessary surgery (n = 1). The treatment appeared not to significantly influence death (the P value for treatment effect by logistic regression after adjustments for APACHE II score and type of therapy was 0.56). The estimated odds ratio for death was 1.15, with the upper bound of the 90% confidence interval being 1.7135. The overall mortality rates were 37% (18 of 49), 25% (14 of 57), and 9% (5 of 54) in patients with pneumonia caused by P. aeruginosa, S. aureus, and H. influenzae, respectively.
Safety and tolerance.
Adverse events which were considered by the investigator to be possibly or probably related to the administration of cefpirome or ceftazidime were reported in 71 patients: 34 (17%) in the cefpirome group and 37 (19%) in the ceftazidime group.
The frequencies of patients with at least one adverse event reported by the investigator as possibly or probably being related to the study treatment were not significantly different between treatment groups (P = 0.67).
DISCUSSION
The purpose of the present study was to compare strategies in which cefpirome or ceftazidime is used for the empiric treatment of pneumonia in ICU patients. The demographics of the patients in both treatment groups were similar except that significantly more patients allocated to the cefpirome group were mechanically ventilated. Seventy-five percent of the episodes of pneumonia were acquired in an ICU. This study included a large number of seriously ill patients, as evidenced by relatively high APACHE II scores and the large proportion of patients requiring mechanical ventilation. Among the strengths of this study were (i) the use of a double stratification, based on the investigators’ choice of monotherapy or combination therapy and on the initial severity score, (ii) the assessment of efficacy and causes of death by a blinded observer, and (iii) the efficacy analyses performed for both the intent-to-treat and the per-protocol populations of patients.
The role of combination therapy for the treatment of severe pneumonia, especially those acquired in the ICU, remains unclear. A recent study showed that imipenem alone was as effective as a combination of imipenem plus netilmicin (7). In contrast, among 200 patients with Pseudomonas aeruginosa bacteremia, mortality was significantly higher among patients receiving monotherapy, including the subgroup of patients with pneumonia. However, in that study patients were not randomized to monotherapy or combination therapy (12). Likewise, in another nonrandomized study, the mortality rate due to Klebsiella pneumoniae bacteremia complicated by septic shock was lower in patients receiving combination therapy than in patients receiving monotherapy (15).
The clinical and bacteriologic responses of cefpirome- and ceftazidime-treated patients who received monotherapy or combination therapy as the empiric treatment were comparable. In addition, the choice of empiric therapy with single agents or combinations of agents was unrelated to outcome, suggesting an adequate therapeutic strategy.
A total of 69% of patients had microbiologically documented infections, which is relatively high compared to the rates in similar studies, in which approximately 50% of patients had documented infections (11, 21, 25). As in other studies, the main pathogens isolated from patients with nosocomial pneumonia were S. aureus, P. aeruginosa, H. influenzae, E. coli, Acinetobacter spp., and K. pneumoniae (5, 8, 11, 25). In this study, 47% of the patients with documented infection had polymicrobial infections. A high frequency of polymicrobial nosocomial pneumonia has been found in other studies by different diagnostic methods. Indeed, despite the use of specific techniques such as the use of PSB samples, Fagon et al. (10) found that 40% of episodes of ventilator-acquired pneumonia were polymicrobial.
The diagnosis of pneumonia in mechanically ventilated patients remains confusing and difficult. The use of diagnostic methods requiring invasive procedures is advocated by several investigators since results obtained by culturing tracheal aspirates may lack specificity. In this study, as many as 40% of the pathogens were isolated from BAL or PSB specimens, and 8% of the pathogens were isolated from blood specimens. However, a relatively high number of patients did not undergo invasive procedures. The present study was designed to include a large number of patients in an international trial. In many countries invasive techniques are not currently accepted as a standard for the management of mechanically ventilated patients with suspected pneumonia, and therefore, they are not considered necessary for antibiotic treatment trials (19). In two recent well-designed trials of severe pneumonia with mainly mechanically ventilated patients, only noninvasive methods were used to diagnose pneumonia (7, 11). In our study, noninvasive methods were predominantly used in Canada and the United Kingdom, whereas invasive techniques were predominantly performed in Italy and France.
The primary endpoint of this study was the clinical response of the intent-to-treat population at the end of treatment. The results indicate that the efficacy of cefpirome given at 2 g twice daily is equivalent to that of ceftazidime given at 2 g three times daily for the empiric treatment of pneumonia in ICU patients. However, both treatments were associated with clinical failure rates (34 and 36%, respectively) higher than those observed in previous studies of the efficacy of ceftazidime (21, 25). These findings reflect the severity of infection and comorbidity in the population of patients studied. Indeed, in a trial comparing ceftazidime and imipenem, clinical failure rates were 22 and 26%, respectively, among patients with pneumonia, but the mean APACHE II scores were lower than those observed in the current study. As expected, in our study, high severity scores were significantly predictive of clinical failure. The mortality rate due directly to pneumonia in our patients may be considered relatively low. Considerable controversy remains in the literature regarding the weight of nosocomially acquired pneumonia versus that of the underlying disease on the prognosis (23). The main concern is to know if patients die from or with nosocomially acquired pneumonia. Although it may be difficult to assess the immediate cause of death, it should be stressed that as in a previous study of ventilator-associated pneumonia due to P. aeruginosa (5), precise definitions were used in order to review specifically the mechanism of death. Moreover, the causes of death were assessed blindly.
The response rates for episodes of documented infection in relation to the initial treatment strategy were similar in both treatment groups. Ceftazidime provided a satisfactory response against S. aureus, despite the concern over its relatively weak in vitro activity. One explanation for these results could be the high dosages used in this study (2 g three times a day), leading to concentrations in the blood and lungs higher than the expected MICs for most S. aureus strains. Despite a slightly lower in vitro activity against P. aeruginosa, the efficacy of cefpirome was similar to that of ceftazidime. However, the highest bacteriologic failure rate was observed in patients with P. aeruginosa pneumonia, and this was found for both drugs. Persistent isolation of P. aeruginosa from respiratory secretions from patients with ventilator-associated pneumonia, despite treatment with appropriate antibiotics, is common and was observed in one-third of patients in a recent study (5). Since the ratio between the concentrations of β-lactams in the parenchyma and MICs for P. aeruginosa may be low (4), the organism may persist, even though it retains antibiotic sensitivity. Therefore, high dosages of both β-lactams and aminoglycosides should be used to treat patients with nosocomial pneumonia caused by P. aeruginosa.
The numbers of patients who experienced at least one adverse event considered by the investigator to be treatment related were not statistically different between treatment groups. There were also no differences between the treatment groups in terms of mortality or causes of death.
In summary, this multicenter, prospective, randomized, controlled trial shows that strategies that include cefpirome or ceftazidime for the empiric treatment of severe pneumonia are equivalent in terms of efficacy and are well tolerated. Thus, cefpirome should be considered a new, alternative therapy for the treatment of such infections.
ACKNOWLEDGMENTS
This work was supported by a grant from Hoechst Marion Roussel.
We thank Muriel Merle and Sarah Knott for preparation of the manuscript.
REFERENCES
- 1.Alvarez-Lerma F ICU-Acquired Pneumonia Study Group. Modification of empiric antibiotic treatment in patients with pneumonia acquired in the intensive care unit. Intensive Care Med. 1996;22:387–394. doi: 10.1007/BF01712153. [DOI] [PubMed] [Google Scholar]
- 2.Anonymous. Hospital-acquired pneumonia in adults: diagnosis, assessment of severity, initial antimicrobial therapy, and preventative strategies. A consensus statement. Am J Respir Crit Care Med. 1996;153:1711–1725. doi: 10.1164/ajrccm.153.5.8630626. [DOI] [PubMed] [Google Scholar]
- 3.Bellido F, Pechere J C, Hancock R E W. Novel method for measurement of outer membrane permeability to new beta-lactams in intact Enterobacter cloacae cells. Antimicrob Agents Chemother. 1991;35:68–72. doi: 10.1128/aac.35.1.68. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Bergogne-Berezin E, Vallee E. Pharmacokinetics of antibiotics in respiratory issues and fluids. In: Pennington J E, editor. Respiratory infections: diagnosis and management. 3rd ed. New York, N.Y: Raven Press, Ltd.; 1994. pp. 715–740. [Google Scholar]
- 5.Brewer S C, Wunderink R G, Jones C B, Leeper K V. Clinical investigations in critical care: ventilator-associated pneumonia due to Pseudomonas aeruginosa. Chest. 1996;109:1019–1029. doi: 10.1378/chest.109.4.1019. [DOI] [PubMed] [Google Scholar]
- 6.Bryskier A, Aszodi J, Chantot J-F. Parenteral cephalosporin classification. Expert Opin Invest Drugs. 1994;3:145–171. [Google Scholar]
- 7.Cometta A, Baumgartner J D, Lew D, Zimmerli W, Pittet D, Chopart P, Schaad U, Herter C, Eggimann P, Huber O, Ricou B, Suter P, Auckenthaler R, Chiolero R, Bille J, Scheidegger C, Frei R, Glauser M P. Prospective randomized comparison of imipenem monotherapy with imipenem plus netilmicin for treatment of severe infections in nonneutropenic patients. Antimicrob Agents Chemother. 1994;38:1309–1313. doi: 10.1128/aac.38.6.1309. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Craven D E, Steger K A. Epidemiology of nosocomial pneumonia. New perspectives of an old disease. Chest. 1995;108:1S–16S. doi: 10.1378/chest.108.2_supplement.1s. [DOI] [PubMed] [Google Scholar]
- 9.Fagon J-Y, Chastre J, Vuagnat A, Trouillet J-L, Novara A, Gibert C. Nosocomial pneumonia and mortality among patients in intensive care units. JAMA. 1996;275:866–869. [PubMed] [Google Scholar]
- 10.Fagon J-Y, Chastre J, Domart Y, Trouillet J-L, Pierre J, Darne C, Gibert C. Nosocomial pneumonia in patients receiving continuous mechanical ventilation. Prospective analysis of 52 episodes with use of a protected specimen brush and quantitative culture techniques. Am Rev Respir Dis. 1989;139:877–884. doi: 10.1164/ajrccm/139.4.877. [DOI] [PubMed] [Google Scholar]
- 11.Fink M P, Snydman D R, Niederman M S, Leeper K V, Jr, Johnson R H, Heard S O, Wunderink R G, Caldwell J W, Shentag J J, Siami G A, Zameck R L, Haverstock D C, Reinhart H H, Echols R M the Severe Pneumonia Study Group. Treatment of severe pneumonia in hospitalized patients: results of a multicenter, randomized, double-blind trial comparing intravenous ciprofloxacin with imipenem-cilastatin. Antimicrob Agents Chemother. 1994;38:547–557. doi: 10.1128/aac.38.3.547. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Hilf M, Yu V L, Sharp J, Zuravleff J J, Korvick J A, Muder R R. Antibiotic therapy for Pseudomonas aeruginosa bacteremia: outcome correlations in a prospective study of 200 patients. Am J Med. 1989;87:540–546. doi: 10.1016/s0002-9343(89)80611-4. [DOI] [PubMed] [Google Scholar]
- 13.Jones R N, Pfaller M A, Allen S D, Gerlach E H, Fuchs P C, Aldridge K E. Antimicrobial activity of cefpirome: an update compared to third-generation cephalosporins against nearly 6000 recent clinical isolates from five medical centers. Diagn Microbiol Infect Dis. 1991;14:361–364. doi: 10.1016/0732-8893(91)90029-f. [DOI] [PubMed] [Google Scholar]
- 14.Knaus W A, Draper E A, Wagner D P, Zimmerman J E. APACHE II: a severity of disease classification system. Crit Care Med. 1985;13:818–829. [PubMed] [Google Scholar]
- 15.Korvick J E, Bryan C S, Farber B, Beam T R, Jr, Schenfeld L, Muder R R, Weinbaum D, Lumish R, Gerding D N, Wagener M M, Yu V L. Prospective study of Klebsiella bacteremia in 230 patients: outcome for antibiotic combinations versus monotherapy. Antimicrob Agents Chemother. 1992;36:2639–2644. doi: 10.1128/aac.36.12.2639. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Le Gall J R, Loirat P, Alperovitch A, Glaser P, Granthil C, Mathieu D, Mercier P, Thomas R, Villers D. A simplified acute physiology score for ICU patients. Crit Care Med. 1984;12:975–977. doi: 10.1097/00003246-198411000-00012. [DOI] [PubMed] [Google Scholar]
- 17.Le Gall J R. Les facteurs pronostiques des maladies infectieuses en réanimation. Lett Infect. 1990;5:185–190. [Google Scholar]
- 18.Leroy O, Georges H, Beuscart C, Guery B, Coffinier C, Vandenbussche C, Thevenin D, Beaucaire G. Severe community-acquired pneumonia in ICUs: prospective validation of a prognostic score. Intensive Care Med. 1996;22:1307–1314. doi: 10.1007/BF01709543. [DOI] [PubMed] [Google Scholar]
- 19.Niederman M S, Torres A, Summer W. Invasive diagnosis testing is not needed routinely to manage suspected ventilator-associated pneumonia. Am J Respir Crit Care Med. 1994;150:565–569. doi: 10.1164/ajrccm.150.2.8049849. [DOI] [PubMed] [Google Scholar]
- 20.Nikaido H, Liu W, Rosenberg E Y. Outer membrane permeability and beta-lactamase stability of dipolar ionic cephalosporins containing methoxyimino substituents. Antimicrob Agents Chemother. 1990;34:337–342. doi: 10.1128/aac.34.2.337. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Norrby S R, Finch R G, Glauser M an European Study Group. Monotherapy in serious hospital-acquired infections: a clinical trial of ceftazidime versus imipenem/cilastatin. J Antimicrob Chemother. 1993;31:927–937. doi: 10.1093/jac/31.6.927. [DOI] [PubMed] [Google Scholar]
- 22.Norrby S R, Geddes A M. Efficacy of cefpirome in the treatment of septicaemia. Scand J Infect Dis. 1993;91:S51–S59. [PubMed] [Google Scholar]
- 23.Papazian L, Bregeon F, Thirion X, Gregoire R, Saux P, Denis J-P, Perrin G, Charrel J, Dumon J-F, Affray J-P, Gouin F. Effect of ventilator-associated pneumonia on mortality and morbidity. Am J Respir Crit Care Med. 1996;154:91–97. doi: 10.1164/ajrccm.154.1.8680705. [DOI] [PubMed] [Google Scholar]
- 24.Rello J, Ausina V, Ricart M, Castella J, Prats G. Impact of previous antimicrobial therapy on the etiology and outcome of ventilator-associated pneumonia. Chest. 1993;104:1230–1235. doi: 10.1378/chest.104.4.1230. [DOI] [PubMed] [Google Scholar]
- 25.Rubinstein E, Lode H, Grassi C an Antibiotic Study Group. Ceftazidime monotherapy vs ceftriaxone/tobramycin for serious hospital-acquired gram-negative infections. Clin Infect Dis. 1995;20:1217–1228. doi: 10.1093/clinids/20.5.1217. [DOI] [PubMed] [Google Scholar]
- 26.Vincent J-L, Bihari D J, Suter P M, Bruining H A, White J, Nicolas-Chanoine M H, Wolff M, Spencer R C, Hemmer M for the EPIC International Advisory Committee. The prevalence of nosocomial infection in intensive care units in Europe. Results of the European prevalence of infection in intensive care (EPIC) study. JAMA. 1995;274:639–644. [PubMed] [Google Scholar]
- 27.Wolff M, Carbon C, Falissard B, Geddes A, Verhoef J, Wilson W. Program and abstracts of the 36th Interscience Conference on Antimicrobial Agents and Chemotherapy. Washington, D.C: American Society for Microbiology; 1996. Comparative study of cefpirome (CPO) and ceftazidime (CAZ) in ICU patients with pneumonia, abstr. LM21; p. 284. [Google Scholar]