| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Guest Access | Sign In via User Name/Password
|
|
|
|||||||
Guest Access | Sign In via User Name/Password |
|||||||||
* From the Washington University School of Medicine (Dr. Kollef), St. Louis, MO; the Department of Pulmonary and Critical Care Medicine (Dr. Morrow), Creighton University, Omaha, NE; the Department of Internal Medicine (Dr. Niederman), Winthrop University Hospital, Mineola, NY; the Department of Pulmonary and Critical Care Medicine (Dr. Leeper), Emory University, Atlanta, GA; the Department of Pulmonary and Critical Care Medicine (Dr. Anzueto), University of Texas Health Sciences Center, Houston, TX; and Rodino Healthcare (Drs. Bullard and Rodino), Millburn, NJ.
Correspondence to: Marin H. Kollef, MD, FCCP, Campus Box 8052, Washington University School of Medicine, 660 South Euclid Ave, St. Louis, MO 63110; e-mail: mkollef{at}im.wustl.edu
Abstract
Study objectives: To evaluate clinical characteristics and treatment patterns among patients with ventilator-associated pneumonia (VAP), including the implementation of and outcomes associated with deescalation therapy.
Design: Prospective, observational, cohort study.
Setting: Twenty ICUs throughout the United States.
Patients: A total of 398 ICU patients meeting predefined criteria for suspected VAP.
Interventions: Prospective, handheld, computer-based data collection regarding routine VAP management according to local institutional practices, including clinical and microbiological characteristics, treatment patterns, and outcomes.
Measurements and results: The most frequent ICU admission diagnoses in patients with VAP were postoperative care (15.6%), neurologic conditions (13.3%), sepsis (13.1%), and cardiac complications (10.8%). The mean (± SD) duration of mechanical ventilation prior to VAP diagnosis was 7.3 ± 6.9 days. Major pathogens were identified in 197 patients (49.5%) through either tracheal aspirate or BAL fluid and included primarily methicillin-resistant Staphylococcus aureus (14.8%), Pseudomonas aeruginosa (14.3%), and other Staphylococcus species (8.8%). More than 100 different antibiotic regimens were prescribed as initial VAP treatment, the majority of which included cefepime (30.4%) or a ureidopenicillin/monobactam combination (27.9%). The mean duration of therapy was 11.8 ± 5.9 days. In the majority of cases (61.6%), therapy was neither escalated nor deescalated. Escalation of therapy occurred in 15.3% of cases, and deescalation occurred in 22.1%. The overall mortality rate was 25.1%, with a mean time to death of 16.2 days (range, 0 to 49 days). The mortality rate was significantly lower among patients in whom therapy was deescalated (17.0%), compared with those experiencing therapy escalation (42.6%) and those in whom therapy was neither escalated nor deescalated (23.7%; 
2 = 13.25; p = 0.001).
Conclusions: Treatment patterns for VAP vary widely from institution to institution, and the overall mortality rate remains unacceptably high. The deescalation of therapy in VAP patients appears to be associated with a reduction in mortality, which is an association that warrants further clinical study.
Key Words: mortality • nosocomial infections • outcomes • pneumonia • treatment • ventilator
Pneumonia is the second most common nosocomial infection reported among ICU patients and is the number one cause of death from nosocomial infection in the United States.123 The estimated prevalence of ventilator-associated pneumonia (VAP) within the ICU setting ranges from 5 to 67%, with reported fatality rates ranging from 24 to 50%.24567 Despite improvements in diagnosis, treatment, and prevention of VAP, it remains a significant cause of hospital mortality. VAP is also associated with significant morbidity, including longer hospital and ICU stays and longer periods of mechanical ventilation, all of which impact medical resources and finances.68 A number of economic analyses have concluded that a single VAP episode prolongs the duration of hospital stay by 6 to 30 days or even longer, and incurs additional medical expenses ranging from $5,000 to $40,000 per patient.6910
The initial choice of antimicrobial regimen appears to be of critical importance in determining the eventual clinical outcomes in patients with VAP, particularly hospital mortality. Early, aggressive, empiric therapy with broad-spectrum agents targeted at likely pathogens has been associated with a reduction in VAP mortality rates.11121314 Luna et al13 prospectively studied 132 VAP patients. Fifty patients with positive BAL findings received empiric treatment prior to undergoing bronchoscopy. The mortality rate was significantly lower among patients whose initial therapy was considered to be adequate (n = 16), based on BAL results, compared to that among patients whose empiric regimen was considered to be inadequate (n = 34) [38% vs 91%, respectively; p < 0.001]. Alvarez-Lerma,12 in an analysis of 430 patients with VAP, reported a significantly higher mortality rate among patients given inappropriate, empiric antimicrobial coverage vs those given an appropriate empiric regimen (24.7% vs 16.2%, respectively; p = 0.039). Rello et al11 also reported a significantly higher VAP-attributable mortality rate among patients with initially inadequate antimicrobial coverage compared with those treated with adequate empiric regimens (37.0% vs 15.4%, respectively; p < 0.05). Finally, Kollef and Ward,14 in a study examining the utility of mini-BAL in patients with VAP, observed differences in mortality rates among patients whose therapy stayed the same after mini-BAL results were available (33.3%) compared with patients who either started therapy anew after undergoing mini-BAL or experienced regimen changes based on mini-BAL findings (60.8%; p < 0.001).
The concept of deescalation therapy is emerging as an effective strategy for the management of VAP and other serious infections.715 This concept entails the early implementation of broad-spectrum empiric coverage followed by a regimen tailored according to susceptibility findings. This strategy, while ensuring a high likelihood of adequate initial coverage, at the same time avoids the long-term use of unnecessary antibiotics, thereby minimizing resistance concerns.16 The most recent VAP treatment guidelines put forth by the American Thoracic Society (ATS) and Infectious Disease Society of America (IDSA)5 include recommendations for early, appropriate, broad-spectrum coverage and subsequent deescalation of antibiotic regimens when possible, based on culture findings.
The Assessment of Local Antimicrobial Resistance Measures study was designed as a large, observational analysis of clinical characteristics and treatment patterns among patients with VAP across the United States. The objective of this study was to evaluate clinical characteristics and treatment patterns among patients with VAP, including the implementation of and outcomes associated with deescalation therapy.
Materials and Methods
The Assessment of Local Antimicrobial Resistance Measures study was a prospective, observational, cohort study of outcomes variables for VAP. Investigators affiliated with 20 ICUs across the United States identified and enrolled eligible patients with VAP. The study was approved by the individual institutional review boards associated with the participating sites. As most of the eligible patients were sedated for mechanical ventilation and because of the purely observational nature of the study, informed consent was waived. However, the study required access to and use of protected information. To this end, all patient information was deidentified in accordance with the Health Insurance Portability and Accountability Act final privacy rule 
164.514,17 and study staff was not to retain any information linkable to specific patients.
Patient Eligibility
Patients were eligible for inclusion in the study if they met the following criteria: hospitalized > 48 h; intubated and receiving mechanical ventilation; and > 18 years of age. The diagnosis of VAP was based on American College of Chest Physicians criteria18 and was prospectively defined as the occurrence of new and persistent radiographic infiltrates following intubation. In addition, at least following criteria must have been present: (1) temperature > 38.3°C; (2) leukocytosis > 10,000 cells/mm3; and/or (3) purulent tracheobronchial secretions. When available, bronchoscopic and nonbronchoscopic BAL cultures with appropriate quantitative thresholds were used to support the diagnosis of VAP.1819 All eligible patients must have had a respiratory tract culture (tracheal aspirate or other) prior to beginning antibiotic treatment. Patients were excluded from the study if they had received a lung transplant or had been studied during a previous ICU admission. Only first-episode cases of VAP were eligible.
Data Collection
Data collected included patient demographics, hospital and ICU admission dates, ICU admitting diagnosis, chest radiograph findings, tracheal aspirate cultures, antimicrobial therapy prior to and during ICU stay, days of mechanical ventilation received prior to and after therapy initiation, and severity-of-illness indexes, including the Acute Physiology and Chronic Health Evaluation (APACHE) II score20 and the clinical pulmonary infection score (CPIS).21 Data were recorded up to the time of ICU discharge or death. Information sources included medical records, bedside flow sheets, computerized bedside nursing stations, computerized radiographic reports, and reports of microbiological studies including sputum Gram stains and cultures of sputum, tracheal aspirate, blood, and pleural and BAL fluids.
Definitions
Culture and sensitivity (C&S) testing of tracheal aspirates was performed by the laboratories of each participating institution. Organisms were classified as either sensitive or not sensitive to specific antibiotics typically used for the treatment of VAP. Organisms reported as being "susceptible" were considered to be sensitive. Both "intermediate" and "resistant" C&S reports were considered to be not sensitive. Therapy was defined as "inappropriate" if an NS result was reported for the antibiotic currently being prescribed in a particular patient. "Time to appropriate therapy," measured in 4-h intervals, was the time from when the diagnosis of VAP was established to when the first correct/appropriate antibiotic regimen was administered.
Escalation/Deescalation of Therapy
Antibiotic regimens were ranked according to the activity spectrum against Gram-negative bacteria (5, highest; 1, lowest). For combination regimens, rank was assigned according to the most potent drug. (Table 1
). Therapy escalation was defined as the switch to or addition of a drug class or classes with a broader spectrum (using definitions in Table 1) or additional coverage. Deescalation was defined as a switch to or discontinuation of a drug class resulting in a less broad spectrum of coverage.
|
2 test and Fisher exact test for categoric data. Continuous variables were compared using the Student t test for normally distributed variables and the Wilcoxon rank-sum test for nonnormally distributed variables. Comparisons were unpaired, and all tests of significance were two-tailed. Results
A total of 398 eligible patients were identified among the 20 sites between May 2003 and December 2004. Two sites were responsible for 43% of all enrollments (111 and 62 patients, respectively). Nine sites enrolled between 19 and 30 patients, and the remaining nine sites enrolled 
6 patients each.
Clinical Characteristics
Table 2
provides a summary of baseline demographics and clinical characteristics. The most frequent ICU admission diagnoses included general postoperative care (15.6%), neurologic conditions (13.3%), sepsis (13.1%), and cardiac complications (10.8%). The mean (± SD) CPIS score at baseline was 8.44 ± 2.34, and the mean APACHE II score was 22.8 ± 8.3. The mean duration of mechanical ventilation prior to VAP diagnosis was 7.3 days (range, 0 to 44 days).
|
|
VAP Treatment Patterns
There were > 100 different antibiotic regimens/combinations prescribed as initial therapy for VAP. For initial treatment, most patients received therapy with one (27.9%), two (46.2%), or three (22.6%) different antibiotics. Among the 11 study sites that enrolled at least 10 patients each, there was a wide range in initial therapies. The percentage of patients per institution who were initially prescribed carbapenem ranged from 0 to 25%. For other spectrum categories, the ranges in initial usage per institution were as follows: cefepime (0 to 73%); ureidopenicillin/monobactam (0 to 70%); quinolone (2 to 38%); and other/none (0 to 65%). Seven percent of patients finished therapy receiving a greater number of antibiotics than when they started therapy for VAP, while 22.6% of patients finished therapy receiving a fewer number of antibiotics than initially prescribed. The change in the mean number of drugs prescribed from initial therapy (1.95) to final therapy (1.75) was –0.20.
Figure 1 illustrates the breakdown of initial therapies according to the spectrum of Gram-negative coverage. The majority of patients were prescribed regimens with a spectrum equivalent to cefepime (30.4%) or a ureidopenicillin/monobactam combination (27.9%). In addition, 51.7% of patients were prescribed vancomycin as part of their initial therapy. The mean baseline APACHE II scores correlated increasingly with the spectrum of antibiotic coverage chosen for initial therapy, ranging from 19.78 ± 8.08 among patients in the lowest therapy spectrum category to 26.40 ± 8.67 among patients in the highest therapy spectrum category (p = 0.0001; Spearman correlation coefficient, 0.27).
|
In 27.6% of cases, initial therapy for VAP was instituted within 4 h of the presumed diagnosis (Fig 2 ). Another 40.2% of patients received initial therapy within 4 to 12 h. For 10.8% of patients, initial therapy was instituted > 24 h after the presumed diagnosis of VAP. As illustrated in Figure 2, the time to appropriate therapy lagged slightly behind.
|
|
2 = 6.15; p = 0.013). According to the culture technique used, deescalation occurred in 27.7% of patients in whom cultures were performed with BAL fluid (2 = 3.59; p = 0.06) and in 20.5% of patients in whom cultures were performed with tracheal aspirates (2 = 0.84; p = 0.36), compared with 8.3% of patients in whom cultures were not performed with either. The two most commonly isolated pathogens were MRSA and P aeruginosa. Among the 59 patients in whom MRSA was isolated as the major pathogen, the initial therapy included vancomycin-based regimens in 40 patients (67.8%). Thirty of the 59 patients continued to received the same therapy following the isolation of MRSA. Among the 29 patients in whom therapy was changed, 17 patients were switched to vancomycin monotherapy or combination regimens, 7 patients were switched to a cefepime/linezolid combination therapy, 3 patients were switched to linezolid therapy alone or another combination regimen, 1 patient was switched to cefepime monotherapy, and 1 final patient was categorized as receiving "other" therapy. Therapy was deescalated in 17 of 59 MRSA cases (29%) and escalated in only 3 cases (5%).
There were a total of 57 patients in whom P aeruginosa was isolated as the major pathogen. The initial therapy in these patients, by spectrum, included carbapenem (15.8%), cefepime (33.3%), ureidopenicillin/monobactam (36.8%), quinolone (3.5%), or other (10.5%). Twenty-six of these 57 patients (45.6%) completed treatment for VAP with their initial therapy. Among 29 patients in whom therapy was changed following the isolation of P aeruginosa (data were missing on 2 patients), therapy was switched to reflect the following spectrum categories: carbapenem (n = 11); cefepime (n = 8); ureidopenicillin/monobactam (n = 6); quinolone (n = 2); or other (n = 2). Among patients with P aeruginosa isolates, deescalation and escalation of therapy were reported with nearly equal frequency (13 cases deescalated; 14 cases escalated).
Outcomes
One hundred of the 398 patients died during this study period, reflecting a mortality rate of 25.1%. The mean time to death was 16.2 days (range, 0 to 49 days). Among the surviving 298 patients during a 30-day follow-up period, 89 (22.4%) were discharged to a non-ICU floor, 71 (17.8%) were discharged to an extended care facility, 56 (14.1%) were discharged to home, and 82 (20.6%) had not been discharged from the ICU.
According to the initial therapy, mortality rates were fairly similar among patients in the categories therapy with carbapenem (31.1%), cefepime (30.6%), ureidopenicillin/monobactam (25.2%), and quinolone (26.7%) [Fig 1]. Mortality was significantly lower among patients in the initial treatment category of other/none, which included vancomycin therapy (19.1%). According to escalation/deescalation patterns, the mortality rate was lowest among patients in whom therapy had been deescalated (17.0%), compared with the categories of "no change" (23.7%) or "escalation" (42.6%; 2 = 13.25; p = 0.001) [Fig 4
]. Table 4
provides a breakdown of initial therapy categories according to spectrum and subsequent patterns of escalation/deescalation within each.
|
|
2 = 5.51; p = 0.019), and 32.6% of those whose BAL fluid was cultured died (2 = 6.42; p = 0.011). A mortality rate of 36.0% was observed among patients in whom neither tracheal aspiration nor BAL was performed.
When mortality was stratified by time to adequate therapy, the mortality rate was highest among patients in whom the time to adequate therapy was > 24 h (30.9%) [Fig 5
]. This compares with a mortality rate of 22.7% for all patients combined for whom adequate therapy was initiated in 24 h, a difference that is quantitatively lower but not statistically significantly different. Among patients who began receiving adequate therapy within the first 12 h, the mortality rate was 23.1%, a rate just slightly lower than that for patients who received any initial therapy during the same time frame (25.2%).
|
Discussion
In this prospective observational study, we found that VAP remains a significant cause of mortality in the ICU setting, and that treatment patterns for this disease vary widely across institutions. The distribution of identified pathogens was similar to that observed in other VAP studies, including high frequencies of MRSA and P aeruginosa.
There were > 100 different antibiotic regimens prescribed as initial therapy among the patients that were studied. The majority of patients experienced no degree of therapy escalation or deescalation. The mean APACHE II score at baseline was 22.8, which is a score that is associated with a predicted death rate of approximately 40%.20 The actual mortality rate in the study population was 25.1%. The mortality rate was significantly lower among patients in whom therapy was deescalated, compared with those who experienced therapy escalation or "no change."
Several clinical studies evaluating the prevalence and clinical impact of VAP have been published, although many have involved single-center populations and/or non-US sites.222232425 Rello et al6 have published the largest US study to date of VAP epidemiology and outcomes, encompassing a cohort of 842 VAP patients identified through a national medical database. In this analysis of 842 VAP cases, the mortality rate was 30.5%, which was not significantly lower than that reported among 2,243 matched control subjects without VAP (30.4%). With regard to other clinical and economic outcomes, patients with VAP demonstrated a significantly longer duration of mechanical ventilation (14.3 vs 4.7 days, respectively; p < 0.001), and longer ICU and hospital stays (additional 6.1 and 11.5 days, respectively; both p < 0.001).
Few VAP studies to date have provided detailed analyses of antibiotic usage patterns, although some investigators have addressed issues regarding the adequacy of early antibiotic therapy, as discussed earlier.11121314 A delay in the initiation of appropriate therapy has clearly been shown to have measurable consequences on VAP-attributable mortality. Iregui et al26 noted significantly higher hospital and VAP-attributable mortality rates among VAP patients in whom appropriate therapy was delayed for 
24 h, compared with those in whom such therapy was initiated earlier (p < 0.01 and p = 0.001, respectively). Our findings are consistent with those previously reported, including an overall trend, although not statistically significant, for higher mortality among patients in whom appropriate therapy was initiated > 24 h after diagnosis.
The concept of therapy deescalation is increasingly being advocated as an appropriate strategy for managing VAP. The deescalation strategy provides clinical balance between one extreme of using broad-spectrum, empiric antimicrobial agents as the sole treatment strategy and the other extreme of delaying the initiation of targeted therapy pending bacteriologic results. The latter approach has clearly been associated with adverse outcomes,11121314 and the former has potential implications for fostering organism resistance. In our study, the association between therapy deescalation and lower VAP mortality rate was an important finding.
The newest, evidence-based treatment guidelines for VAP put forth by the ATS-IDSA5 emphasize the need for early and appropriate antibiotic therapy followed by deescalation whenever possible, based on culture results and patient response. The guidelines also stress the importance of maintaining local, frequently updated antibiograms within individual hospitals and ICUs to ensure the appropriateness of antibiotic coverage. Initially inappropriate therapy is associated with worse outcomes, even if therapy is subsequently changed to reflect bacteriologic findings. This assumption is supported by our finding that comparative mortality was highest among patients in whom therapy was escalated compared to those who experienced deescalation or no change in therapy.
A key factor in empiric therapy selection is a consideration of patient risk for multiple-drug-resistant organisms, including those with a recent history of hospitalization or residence in another health-care facility (eg, nursing home), and patients whose current hospital stay has exceeded 5 days. The ATS-IDSA guidelines5 include suggestions for specific agents that are likely to provide appropriate coverage for initial therapy for VAP in patients with suspected risk factors for multiple-drug-resistant disease. In our cohort, in whom the mean duration of mechanical ventilation prior to diagnosis was 7.3 days, the majority of patients were initially placed on seemingly appropriate antibiotic regimens having a Gram-negative spectrum of activity equivalent to that of cefepime (30.4%) or piperacillin-tazobactam (27.9%). However, in almost one fifth of our patients, initial therapy fell within a Gram-negative spectrum category of "other or none." Of further concern is the finding that there were > 100 different antibiotic regimens or combination regimens prescribed as initial coverage for the 398 patients studied.
The treatment of VAP has traditionally consisted of antibiotic administration for 14 to 21 days.5 Based on clinical evidence, the new ATS-IDSA guidelines5 advocate that attempts should be made to shorten the duration of treatment to as few as 7 days in appropriate circumstances, and that prolonged treatment can lead to colonization with resistant organisms. Attitudes toward shorter VAP treatment durations may be gaining acceptance in clinical practice, as reflected by a mean treatment duration in our study of 11.8 days.
Our data should be interpreted in light of certain limitations. Despite the involvement of 20 separate ICUs across the United States, the majority of patients were enrolled at only two sites, and the results may be biased somewhat toward the practices at those particular institutions. In addition, the lack of a standardized approach to the diagnosis of VAP may have impacted patient inclusion and outcomes. Regardless, our results appear to be generally consistent with findings reported elsewhere and represent one of the larger VAP patient populations that have been studied to date.
In conclusion, VAP continues to pose a therapeutic challenge in the ICU setting. Our observations confirm that treatment patterns for VAP vary widely from institution to institution, and that the overall mortality rate remains unacceptably high. The practice of deescalation therapy appears to be associated with a reduction in patient mortality due to this disease. Additional study is warranted to investigate the reasons for this finding and to further define the potential impact of deescalation strategies in improving VAP outcomes.
Footnotes
Abbreviations: APACHE = acute physiology and chronic health evaluation; ATS = American Thoracic Society; CPIS = clinical pulmonary infection score; C&S = culture and sensitivity; IDSA = Infectious Disease Society of America; MRSA = methicillin-resistant Staphylococcus aureus; VAP = ventilator-associated pneumonia
Received for publication July 18, 2005. Accepted for publication November 5, 2005.
References
This article has been cited by other articles:
|
|
 |
|
  L. Vidaur, K. Planas, R. Sierra, G. Dimopoulos, A. Ramirez, T. Lisboa, and J. Rello Ventilator-Associated Pneumonia: Impact of Organisms on Clinical Resolution and Medical Resources Utilization Chest, March 1, 2008; 133(3): 625 - 632. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 VAP -- A "Real Life" View Journal Watch Infectious Diseases, June 9, 2006; 2006(609): 3 - 3. [Full Text]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
