0
Original Research: Critical Care |

Is a Strategy Based on Routine Endotracheal Cultures the Best Way to Prescribe Antibiotics in Ventilator-Associated Pneumonia?Surveillance Tracheal Aspirate Cultures in VAP FREE TO VIEW

Carlos M. Luna, MD, PhD, FCCP; Sergio Sarquis, MD; Michael S. Niederman, MD; Fernando A. Sosa, MD; Maria Otaola, MD; Nicolas Bailleau, MD; Carlos A. Vay, PhD; Angela Famiglietti, PhD; Célica Irrazabal, MD; Abelardo A. Capdevila, MD
Author and Funding Information

From the Pulmonary and Critical Medicine Divisions (Drs Luna, Sarquis, Sosa, Otaola, Bailleau, Irrazabal, and Capdevila), Department of Medicine, and Microbiology Section (Drs Vay and Famiglietti), Department of Clinical Biochemistry, Facultad de Farmacia y Bioquímica, Hospital de Clínicas, Universidad de Buenos Aires, Argentina; and the Department of Medicine (Dr Niederman), Winthrop University Hospital, Mineola, NY.

Correspondence to: Carlos M. Luna, MD, PhD, FCCP, Arenales 2557, Piso 1, Dto A, Buenos Aires, 1425, Capital Federal, Argentina; e- mail: dr.cm.luna@gmail.com


Reproduction of this article is prohibited without written permission from the American College of Chest Physicians. See online for more details.

Funding/Support: Funding for this study was received from Asociacion Cooperadora del Hospital de Clinicas.


Chest. 2013;144(1):63-71. doi:10.1378/chest.12-1477
Text Size: A A A
Published online

Objectives:  The objectives of this study were to evaluate if a strategy based on routine endotracheal aspirate (ETA) cultures is better than using the American Thoracic Society/Infectious Diseases Society of America (ATS/IDSA) guidelines to prescribe antimicrobials in ventilator-associated pneumonia (VAP).

Methods:  This was a prospective, observational, cohort study conducted in a 15-bed ICU and comprising 283 patients who were mechanically ventilated for ≥ 48 h. Interventions included twice-weekly ETA; BAL culture was done if VAP was suspected. BAL (collected at the time of VAP) plus ETA cultures (collected ≤ 7 days before VAP) (n = 146 different pairs) were defined. We compared two models of 10 days of empirical antimicrobials (ETA-based vs ATS/IDSA guidelines-based strategies), analyzing their impact on appropriateness of therapy and total antimicrobial-days, using the BAL result as the standard for comparison.

Results:  Complete ETA and BAL culture concordance (identical pathogens or negative result) occurred in 52 pairs; discordance (false positive or false negative) in 67, and partial concordance in two. ETA predicted the etiology in 62.4% of all pairs, in 74.0% of pairs if ETA was performed ≤ 2 days before BAL, and in 46.2% of pairs if ETA was performed 3 to 7 days before BAL (P = .016). Strategies based on the ATS/IDSA guidelines and on ETA results led to appropriate therapy in 97.9% and 77.4% of pairs, respectively (P < .001). The numbers of antimicrobial-days were 1,942 and 1,557 for therapies based on ATS/IDSA guidelines and ETA results, respectively (P < .001).

Conclusions:  The ATS/IDSA guidelines-based approach was more accurate than the ETA-based strategy for prescribing appropriate, initial, empirical antibiotics in VAP, unless a sample was available ≤ 2 days of the onset of VAP. The ETA-based strategy led to fewer days on prescribed antimicrobials.

Figures in this Article

Ventilator-associated pneumonia (VAP) occurs in patients in the ICU who are mechanically ventilated through an endotracheal tube or a tracheostomy for ≥ 48 h.1 It is the most frequent ICU-acquired infection among patients receiving mechanical ventilation and represents the most frequent cause of death among nosocomial infections.2,3 The incidence of VAP in patients who are mechanically ventilated ranges from 7% to 30% and the VAP mortality rate ranges from 20% to 75%, depending on the population studied.1 Attributable mortality due to VAP has been described and its incidence estimated between 0% and 50%.4,5 Approximately 50% of antibiotics prescribed in the ICU are administered for respiratory tract infections.3,6 Clinical outcome parameters for the therapy of VAP, like duration of mechanical ventilation, time to resolution of clinical variables or biomarkers, and length of stay, should be used to assess the response to therapy. Mistakes with antimicrobials include inappropriate therapy (no coverage for the pathogen), overuse, and prolonged exposure beyond what is necessary. Inappropriate antimicrobial therapy in VAP is associated with increased mortality,5,7,8 while overuse and prolonged exposure are related to the emergence of bacterial resistance, adverse events, and increased costs. Physicians face the challenge of treating VAP promptly with appropriate antibiotics while avoiding the risk of inducing the selection of multidrug-resistant (MDR) bacteria.9 Some studies have raised concerns about the inappropriateness of antimicrobial therapy when guidelines are followed.10,11

Results of routine endotracheal aspirate (ETA) cultures provide updated information about airway colonization by potential pathogens that could lead to VAP. Prior studies have suggested that this information could provide a rationale for prescribing appropriate antibiotics, while waiting for culture results, in up to 95% of patients in whom VAP is ultimately diagnosed by BAL culture.10 Our aim was to prospectively test the hypothesis that ETAs in patients receiving ventilation are an accurate way to select appropriate antimicrobial therapy covering most of the involved pathogens adapted to the local epidemiology, compared with the use of empirical therapy suggested by the American Thoracic Society and Infectious Diseases Society of America (ATS/IDSA) guidelines.1

Study Design and Subjects

This was a prospective, observational, cohort study conducted in the ICU of a teaching hospital in Buenos Aires, Argentina, from October 2005 to September 2007. Inclusion criteria were age ≥ 18 years and an estimated length of ventilation ≥ 48 h. Exclusion criteria were AIDS-defining conditions or neutropenia (< 500 neutrophils/mm3). The study was approved by the institutional research board and the ethics committees from the Hospital de Clínicas. All patients or their legal representatives signed an informed consent authorizing the performance of the routine ETAs.

Data collected at admission included laboratory data, chest radiograph, volume and quality of secretions, and arterial blood gases necessary to evaluate the potential presence of VAP. After inclusion, routine, biweekly tracheal aspirate cultures were systematically performed. Patients were evaluated daily for the presence of VAP and the use of antimicrobial therapy was recorded. If a clinical diagnosis of VAP was present, a BAL culture was performed. Patients were followed until hospital discharge or death.

Procedures

ETA culture and daily clinical follow-up were performed by two investigators (S. S. and F. A. S.). The results of the ETA cultures were not routinely given to the attending physicians, as it was not part of the standard of care.

ETA cultures were collected every Monday and Thursday between 8.00 am and 9.00 am, avoiding the instillation of saline solution. A positive culture of the first ETA was termed basal colonization. BAL was done within 6 h following the clinical diagnosis of VAP; six 20-mL aliquots of sterile saline solution (0.9% sodium chloride) were instilled through the bronchoscope channel and aspirated by hand; the first aliquot was discarded.

ETA and BAL cultures were submitted for direct microscopic examination, including screening for epithelial squamous cells, expressed as percentage of cells per low-power field; Gram stain; and semiquantitative evaluation of polymorphonuclear leukocytes. BAL and those ETA samples showing < 10 epithelial cells and > 25 polymorphonuclear leukocytes per low-power field were submitted and processed by semiquantitative aerobic bacteria culture. Using a quantitative loop, 0.005 mL of the ETA and 0.01 mL of the BAL fluid were plated onto each of four culture media (sheep blood agar, Levine EMB agar, mannitol salt agar, and chocolate blood agar). After inoculation, each plate was incubated at 35° to 37°C in a CO2-enriched environment. Estimates of the number of bacteria originally in the fluid were made by colony counts and were expressed as colony-forming units (cfu) per milliliter of fluid. ETA was reported as negative if there was no growth or < 103 cfu/mL and positive if there were ≥ 103 cfu/mL. BAL culture was reported as negative if there was no growth or < 104 cfu/mL and positive if there were ≥ 104 cfu/mL. BAL and ETA cultures were processed by different microbiologists who were not aware of the results of the other bacteriologic results.

Clinical Diagnosis of VAP:

VAP was suspected in patients who were intubated and ventilated and presented with a new or progressive radiographic infiltrate plus two of the following: (1) body temperature > 38°C or < 36°C, (2) WBC count > 12,000 or < 4,000 per mL, and (3) macroscopically purulent tracheal aspirate. Those clinically defined VAP episodes occurring in patients who had had a previous VAP episode > 7 days before were considered and analyzed as a different VAP episode. Microbiologically confirmed VAP was defined as growth of ≥ 104 cfu/mL of a microorganism on bronchoscopic BAL culture, followed by the prescription of antimicrobials for pneumonia, in patients with a clinical diagnosis of VAP.

Appropriate Antibiotic Therapy:

Appropriate antibiotic therapy was defined as coverage of all the pathogens isolated in the BAL by at least one of the antimicrobials administered at the onset of VAP, determined by the sensitivity pattern in the antibiogram. For Pseudomonas aeruginosa, 5 days of combination therapy, including two antipseudomonal (one β-lactam and one non-β-lactam) antimicrobial agents, and susceptibility to at least one of them, was required.1

Theoretical Models of Antimicrobial Use:

To investigate the usefulness of ETA, all possible combinations of a BAL sample and an ETA performed from 0 to 7 days before the clinical suspicion of VAP (ETA/BAL pairs) in each of the episodes of clinically suspected VAP were defined. For each suspected VAP episode, one BAL culture was done, but multiple ETAs could have been done, creating multiple different pairs for some VAP episodes. Initial empirical antibiotic therapy that could be initiated on the day of the clinical diagnosis of VAP was analyzed considering two possible theoretical models: the ETA-based therapy strategy or an ATS/IDSA guideline-based therapy strategy (considering the presence of risk factors for MDR microorganisms, de-escalating at day 2 according to the microbiologic results), with empirical antimicrobial selection based on local epidemiologic data (Table 1). These theoretical models considered 48 h to be the estimated time for availability of BAL culture and susceptibility results. In both strategies, treatment was continued, escalated, or de-escalated to complete 10 days of appropriate therapy when a positive BAL culture was present; or discontinued when BAL culture results were negative. In the analysis of the ETA-based strategy, a negative result of the ETA was considered an indication not to prescribe antimicrobials while awaiting the BAL culture results. The antimicrobials used in a de-escalation strategy are outlined in Table 1.

Table Graphic Jump Location
Table 1 —General Principles for Prescribing Antimicrobial Therapy Under the Two Different Strategies in This Study

ATS/IDSA = American Thoracic Society/Infectious Diseases Society of America; ETA = endotracheal aspirate; MRSA = methicillin-resistant Staphylococcus aureus; VAP = ventilator-associated pneumonia.

Antimicrobial-Days:

The sum of all antibiotics per day in all pairs from the onset of VAP therapy until completion of 10 days of appropriate therapy or stopping for a negative BAL culture was calculated for each of the two theoretical models. The total sum of antimicrobial-days for each particular pair included appropriate use, inappropriate use, and overuse during all the time that antimicrobials were to be administered, including inappropriate use or overuse that could occur during the first 2 days of empirical therapy. The appropriateness of therapy and the antimicrobial-days were evaluated for the two different strategies.

Statistical Analysis

Continuous variables are expressed as mean ± SD unless stated otherwise. Statistical significance was determined using the t test to compare continuous variables and the χ2 test or the Fisher exact test to compare categorical variables. The sensitivity of ETA to predict the pathogen of VAP and the sensitivity of the ETA-based and ATS/IDSA-based strategies to prescribe appropriate antibiotics were expressed as percentages. Adjusted ORs and 95% CIs were calculated. Significance was defined as a P value < .05. The statistical software program SPSS 15.0 for Windows (IBM) was used.

Patient Characteristics

During 2 years, 323 patients receiving mechanical ventilation were included; a total of 800 ETAs were performed at least once in 283 patients (an average of 2.48 ETAs per patient) (Fig 1A). Eighty-three patients who had an ETA performed during the 7 days prior to BAL had 102 episodes of clinically suspected VAP, whereas the remaining 200 did not have episodes of suspected pneumonia. Fifty-five patients had 65 episodes of VAP confirmed by BAL culture, whereas in 28 patients, 37 episodes of clinically suspected VAP were not confirmed (BAL culture < 104 cfu/mL). The incidence of VAP was, therefore, 17.0% in the 323 patients who received mechanical ventilation for at least 48 h (Fig 1A).

Figure Jump LinkFigure 1. A, Flow chart of the patients and VAP episodes in this study. From the 323 patients admitted to the study, 283 had performed at least one ETA; 83 of the 283 had at least one episode of clinically diagnosed VAP and had at least one ETA specimen cultured during the preceding 7 d, as well as a BAL sample. B, ETA/BAL pair was defined as the combination of an ETA culture plus a BAL culture performed during the 7 d following an episode of clinically diagnosed VAP. Some patients had more than one episode, and some episodes had more than one ETA previously performed. An episode of VAP with BAL preceded by two ETAs performed during the previous 7 d was considered two separate ETA/BAL pairs. ETA = endotracheal aspirate; VAP = ventilator-associated pneumonia.Grahic Jump Location

We defined 146 possible ETA/BAL pairs in the 102 episodes of clinically suspected VAP (Fig 1B). In 96 of 146 ETA/BAL pairs (65.8%; 95% CI, 58.1%-73.5%), BAL culture was positive; BAL culture was negative in the remaining 50 ETA/BAL pairs. Complete concordance (positive for the same isolated pathogens or a negative result in both samples) between ETA and BAL cultures was found in only 52 ETA/BAL pairs (35.6%); 67 (45.9%) of the other 94 pairs showed total discordance (false positive or false negative), and 27 (18.5%) showed partial concordance.

The characteristics of the 55 patients who developed bacteriologically confirmed VAP and the 228 patients who did not are shown in Table 2; the Pao2/Fio2 ratio on admission was lower (P = .032) and basal ETA colonization rate (≥ 103 cfu/mL) was higher (P = .014) in those patients who developed bacteriologically confirmed VAP. Among the reasons for mechanical ventilation, postoperative respiratory failure was less frequent (P = .023) and sepsis more frequent (P = .029) in the VAP-confirmed group.

Table Graphic Jump Location
Table 2 —Demographic, Clinical, ETA-Culture Characteristics and Mortality of Patients Admitted in the Study and Comparison Between Patients Developing VAP vs Those Not Developing VAP

Data given as No. unless otherwise indicated. APACHE II = Acute Physiology and Chronic Health Evaluation II; cfu = colony-forming units; F = female; M = male; NS = not significant; TISS = Therapeutic Intervention Scoring System. See Table 1 legend for expansion of other abbreviations.

a 

“All” includes all intubated patients that had performed at least one routine ETA.

b 

VAP includes all patients with clinical diagnosis of VAP confirmed by BAL culture that had performed at least one ETA during the last 5 d.

c 

No VAP includes all those intubated patients with at least one ETA who did not develop clinically or microbiologically diagnosed VAP.

d 

P value: significance of the difference comparing VAP with subjects without VAP (no VAP).

e 

Colonization at the basal ETA obtained in the standard days.

Seventy-nine microorganisms were detected at a concentration ≥ 104 cfu/mL in the 65 BAL-positive VAP episodes that comprised 96 ETA/BAL pairs (Fig 1B). Counting all the microorganisms in the 96 positive ETA/BAL pairs, the number of pathogens was 125, and we used this number to calculate the ability of ETA to predict VAP etiology, and to evaluate the appropriateness of therapy, if the ETA result had been used to guide initial therapy.

Value of ETA Cultures for Predicting VAP Etiology

Pathogens present in the patients with VAP are listed in Table 3. Comparing the etiology of the first episodes of VAP with that of subsequent episodes, a difference in the distribution of the pathogens isolated in BAL was found. In the first episode, the 61 microorganisms found in the 48 positive BAL samples (an average of 1.27 pathogens per episode) included Acinetobacter species (n = 20, 32.8%), P aeruginosa (n = 18; 29.5%), and methicillin-resistant Staphylococcus aureus (MRSA) (n = 10, 16.4%). In the subsequent episodes, among the 18 microorganisms found in the 17 positive BAL samples (an average of 1.06 pathogens per episode) were P aeruginosa (n = 11, 61.1%), in a higher proportion than in the first episode (P = .025); and MRSA (n = 1, 5.6%) and Acinetobacter species (n = 3, 16.6%), which tended to be present in a lower frequency than in the first episode (P = NS) (Fig 2). P aeruginosa,Acinetobacter species, or MRSA represented 79.7% (63 of 79) of the pathogens found at a concentration ≥ 104 cfu/mL in the 65 BAL-confirmed VAP episodes.

Table Graphic Jump Location
Table 3 —Microorganisms Detected by BALa and by Routine ETAb Culture

Organism names and data in bold type are multidrug-resistant bacteria. MSSA = methicillin-sensitive Staphylococcus aureus. See Table 1 and 2 legends for expansion of other abbreviations.

a 

Microorganisms (n = 79) detected by BAL culture at a concentration ≥ 104 cfu/mL in 65 episodes of VAP.

b 

Microorganisms (n = 184) detected by routine ETA at a concentration ≥ 103 cfu/mL during the 7 d preceding the development of VAP (n = 114, during the 2 d before, and n = 70, 3-7 d before VAP diagnosis).

Figure Jump LinkFigure 2. Distribution of the different pathogens of the first and subsequent episodes of VAP. First episodes were caused by Acinetobacter species (32.8%), Pseudomonas aeruginosa (29.5%), MRSA (16.4%), and other pathogens, including Haemophilus influenzae, Escherichia coli,Providencia species, Streptococcus viridans, Proteus mirabilis, and Klebsiella pneumoniae (16.2%). Subsequent episodes were restricted to MRSA (5.6%), Acinetobacter species (16.6%), P aeruginosa* (61.1%); and other pathogens (16.7%). *P = .025, significantly more frequent than in the first episode. MRSA = methicillin-resistant Staphylococcus aureus; spp = species. See Figure 1 legend for expansion of other abbreviations.Grahic Jump Location

Sensitivity of ETA for predicting the etiology of VAP was evaluated in those ETA/BAL pairs with microbiologically confirmed VAP. Overall, ETA predicted 78 of 125 pathogens (62.4%; 95% CI, 53.9%-70.9%). When ETA was performed 3-7 days before VAP, it predicted only 24 of 52 pathogens (46.2%; 95% CI, 32.6%-59.7%). When it was performed < 3 days preceding VAP, ETA predicted 41 of 58 pathogens (70.7%; 95% CI, 59.0%-82.4%; P = .016), compared to ETA performed ≥ 3 days vs < 3 days before. The specificity and negative predictive values were, respectively, 10.2% and 20.3% for the overall group, 2.1% and 3.6% for ETA performed 3-7 days before VAP, and 15.4% and 28.6% for ETA performed < 3 days before VAP. Interestingly, the sensitivity of ETA for predicting VAP etiology due to the four MDR bacteria—P aeruginosa, Acinetobacter species, MRSA, and Stenotrophomonas maltophilia—was higher: 69.2% overall, 79.7% for ETA performed ≤ 2 days before BAL, and 49.0 for ETA performed 3 to 7 days before BAL.

Sensitivity of ETA for the First VAP Episodes vs Subsequent Episodes

Sensitivity of ETA for predicting VAP pathogens was evaluated in the 101 ETA/BAL pairs of a first episode of suspected VAP and in 45 ETA/BAL pairs during a subsequent (second or third) episode of suspected VAP. In the first VAP episode, ETA predicted 54 of 89 pathogens (60.7%; 95% CI, 50.5%-70.8%). When ETA was performed 3-7 days before this first episode, it predicted 13 of 31 pathogens (41.9%; 95% CI, 24.5%-59.3%), and when it was performed < 3 days before the first episode of VAP, ETA predicted 41 of 58 pathogens (70.7%; 95% CI, 59.0%-82.4%; P = .009) compared to ETA performed ≥ 3 days vs < 3 days before the first episode of VAP. In those pairs representing a subsequent episode, ETA predicted 24 of 36 pathogens (66.7%; 95% CI, 51.3%-82.1%; P = not significant) compared with the rate observed in those pairs representing a first VAP episode. When ETA was performed 3-7 days before a subsequent VAP episode, it predicted 11 of 21 pathogens (52.4%; 95% CI, 31.0%-73.8%), and when it was performed <3 days before, ETA predicted 13 of 15 pathogens (86.7%; 95% CI, 69.5-103.8; P = .040), compared to ETA performed ≥ 3 days vs < 3 days before a subsequent VAP episode.

Appropriateness of Antibiotic Therapy According to Different Strategies

The sensitivity of ETA for predicting VAP etiology was defined using the BAL result as the standard. The appropriateness of initial therapy, selected based on the ETA culture results, was compared with the appropriateness of therapy that would have been prescribed based on the ATS/IDSA strategy, adapted to the local epidemiology. In > 75% of pairs, the inappropriate therapy category was dominated by reasons different from withholding antibiotics in patients with negative ETA, but positive BAL, cultures. Even though the overall sensitivity of ETA for predicting the etiology of VAP was 62.4%, if the treatment strategy were based on the results of previous ETA, it would have been appropriate in 120 of the 146 ETA/BAL pairs (82.2%; 95% CI, 76.0%-88.4%), and inappropriate therapy would have been used in 26 pairs (17.8%). Inappropriate therapy would have been no therapy for the first 2 days before BAL data were available in five pairs with a false-negative result and the wrong or incomplete antimicrobial coverage in 21 other pairs (14.4%). Escalation therapy would have been necessary in 36 ETA/BAL pairs (24.7%), due to lack of coverage in 26 of them, and in the remaining 10 to adapt the therapy to a more convenient antimicrobial for safety reasons. De-escalation would have been performed in 59 cases (40.4%). On the other hand, if an ATS/IDSA guidelines-based strategy (broad-spectrum antimicrobials adapted to local epidemiology, combination therapy during 5 days for P aeruginosa and de-escalation after 2 days of therapy when feasible) were used, appropriate therapy would have been prescribed in 143 pairs (97.9%; 95% CI, 95.6%-100.2%; P = .002 vs ETA-based strategy) (Fig 3). With the ATS/IDSA guidelines-based strategy, escalation would have been necessary in only seven (4.8%) ETA/BAL pairs (P < .001 vs ETA-based strategy), because of lack of coverage of S maltophilia resistant to piperacillin-tazobactam and colistin in three cases and the need to replace different appropriate antimicrobials for safety reasons in the remaining four cases. De-escalation would have been performed in 139 cases (95.2%) using the ATS/IDSA guidelines-based strategy (P = .001 vs ETA-based strategy).

Figure Jump LinkFigure 3. Appropriateness of therapy according to the result of ETA performed during the 2 d preceding the VAP diagnosis, during 3-7 d preceding the VAP diagnosis, and availability of ETA performed 3-7 d preceding the VAP diagnosis. The last situation is more realistic, as in the real world the ETA culture results become available ≥48 h after ETA retrieval. ATS/IDSA = American Thoracic Society/Infectious Diseases Society of America. See Figure 1 legend for expansion of other abbreviations.Grahic Jump Location

Regarding the role of the different strategies for avoiding the overuse of antimicrobials, the ETA-based strategy would have led to the use of 1,557 antimicrobial-days, while the ATS/IDSA guidelines-based approach would have led to 1,942 antimicrobial-days (P = .002). The ETA-based strategy would have led to the prescription of 24.4% fewer antimicrobial-days than the ATS/IDSA guidelines-based strategy (Fig 4). Considering separately the timing of ETA sample collection, the antimicrobial-days were fewer using the ETA-based strategy in those ETAs performed 1-2 days before VAP (P = .005), but not in those performed 3-7 days before VAP (P = NS) (Fig 4).

Figure Jump LinkFigure 4. Comparison of the number of antimicrobial-days between the different strategies for prescribing empirical antimicrobial therapy. The ETA-based strategy would lead to the use of 1,557 antimicrobial-days, significantly fewer than the ATS/IDSA guidelines-based therapy (1,942 antimicrobial-days). Considering separately ETAs performed < 3 d and those performed 3-7 d before VAP diagnosis, there was a trend toward fewer antimicrobial-days with the ETA-guided approach (1,184 vs 918 antimicrobial-days and 758 vs 639 antimicrobial-days, respectively) (P = NS). See Figure 1 and 3 legends for expansion of abbreviations.Grahic Jump Location

Several authors have promoted the use of surveillance ETAs to predict the etiology of VAP and to guide initial antimicrobial therapy.10,12,13 In our study, the sensitivity of ETAs to detect the pathogens for VAP was 62.74%, which is similar to the findings of Declaux et al,14 who reported that protected-specimen brush culture performed every 48 to 72 h in patients with ARDS predicted the etiology in 66% of cases. However, the present study does not support the hypothesis that routine ETA performed twice a week would be a better way to prescribe antimicrobials in suspected VAP, compared with the use of an empirical therapy based on the ATS/IDSA guidelines, adapted to the local epidemiology. An ETA-based strategy leads to an appropriate therapy for BAL-confirmed microorganisms in VAP significantly less often than a therapy guided by a strategy based on the ATS/IDSA guidelines (82.2% vs 97.9%). The major goals of the ATS/IDSA guidelines were to use early, appropriate antibiotics in adequate doses and to avoid excessive antibiotic use through de-escalation of initial antibiotic therapy, based on microbiologic cultures and the clinical response of the patient, while shortening the duration of therapy to the minimum effective period.1 Specificity and negative predictive values were low, further reducing the potential use of surveillance ETA cultures to rule out the use of last-line antibiotics.

Michel et al10 observed that the sensitivity of ETAs performed every 72 to 96 h for predicting the pathogen causing VAP was 83%, but even when the ETA did not give the same results as the BAL culture, the antibiotic therapy based on the results of the ETAs was appropriate in 38 of the 40 evaluable cases (95%). Although these findings differ from ours, their observation that ETAs could be useful for prescribing appropriate antimicrobials, regardless of whether it detected the etiologic pathogen, coincided with our observation that antimicrobial therapy based on the ETAs would have been appropriate in 82.2% of ETA/BAL pairs, while its sensitivity to predict a specific pathogen was only 62.4%. Similar to our findings are those of two previous studies that found poor agreement between prior ETA cultures and cultures performed at time of VAP. These were the study by Sanders et al,15 with the limitation of being a secondary analysis, and the prospective study by Bouza et al,16 which was not comparable because ETA cultures were performed weekly.

We found that the sensitivity of ETA for predicting the pathogens causing VAP was significantly better when it was performed during the 48 h preceding the episode of VAP, compared with when it was performed earlier. ETA also was better when it was performed during a subsequent episode of VAP than during an initial episode. Both findings are not surprising because, as demonstrated by Ewig et al,17 airway colonization is dynamic: At the time of VAP, most etiologic pathogens are present in the tracheal aspirate of the patients, but they may not have been present much earlier. The higher sensitivity of ETA in the ETA/BAL pairs representing subsequent episodes of VAP could be explained because prolonged mechanical ventilation is often associated with prior broad-spectrum antimicrobial therapy that exerts a selective pressure promoting colonization by a limited number of MDR microorganisms that had become the only candidates for a subsequent episode of VAP.11

MDR microorganisms, particularly P aeruginosa,Acinetobacter species, and MRSA are present as airway colonizers with increasing frequency globally. In this study, we found that these organisms were significantly more frequent as airway colonizers than as VAP pathogens, being present in 70% of positive BAL cultures, but representing 84.5% in the positive ETA samples in the same group of patients (Table 3). Among these species of MDR organisms, we found that P aeruginosa was a significantly more common pathogen during a subsequent episode of VAP than the other microorganisms, and was responsible for 33.9% of VAPs during a first episode, but 61% of subsequent episodes of VAP (Fig 2).

Regarding the use of antimicrobials, we observed, consistent with the observations of Michel et al,10 that using ETA cultures for prescribing antimicrobials could significantly reduce (16.7%, P < .05) the unnecessary use of broad-spectrum antibiotics (compared with an ATS/IDSA guideline-based strategy) and the number of antimicrobial-days. However, the beneficial effect of reducing antimicrobial prescription was mitigated by the unacceptably high rate of inappropriate antimicrobial therapy when using an ETA-based strategy compared with the lower rate observed if the ATS/IDSA guideline-based therapy would have been used (17.8% vs 2.1%). This finding is important in choosing between the different strategies because it is well known that both inappropriate therapy and delay in the initiation of appropriate therapy are associated with increased mortality in VAP.7,8,18 We also found that de-escalation would be significantly more frequent and escalation less frequent using the ATS/IDSA guideline-based therapy than the ETA-based strategy.

Another consideration that should be taken into account when applying our observations to the real world is that the time until the ETA culture and antibiogram results are available is usually not < 48 to 72 h, increasing the difficulty in using ETA to guide accurate decisions in those patients who had an ETA obtained within 48 h before the development of VAP. In addition, we found that ETA better predict the pathogens in patients developing a subsequent VAP episode, rather than an initial episode, and that these infections often involved MDR microorganisms, such as P aeruginosa. Thus, if routine ETAs are used, they probably should be limited to patients with a subsequent, and not an initial, episode of VAP.

Michel et al10 stated that antimicrobial therapy based on routine ETA cultures would have been adequately prescribed in a significantly higher proportion of cases than antimicrobial therapy based on the ATS guidelines.10 The difference between our findings and theirs could be explained because the therapy in our study was based on the more recent ATS/IDSA guidelines,1 which established that therapy should be modified and adapted to the local microbiologic data, and a specific agent should be selected that is likely to be effective when an etiologic pathogen is suspected or identified. The ATS/IDSA guideline published in 2005 is based on different principles than earlier guidelines, including the following: (1) patients at risk for MDR pathogens should receive MRSA and dual Gram-negative coverage to increase the likelihood of initially appropriate therapy; (2) when culture data became available, de-escalation to fewer drugs with a narrower spectrum of coverage should occur; and (3) long durations of combination Gram-negative therapy could increase toxicity, antibiotic resistance, and mortality.1

Our study has some limitations, since this is a prospective observational study and not an intervention study, and was performed at only one institution to evaluate which of the two strategies (based on routine ETA cultures or based on the ATS/IDSA guidelines) is better to prescribe initial empirical antimicrobials in VAP. A randomized, controlled trial should be done to explore which is the better strategy. Our ICU has a high incidence of MDR bacteria, including 10% to 20% of P aeruginosa infections susceptible only to colistin; 30% of Acinetobacter species infections susceptible only to colistin; and methicillin resistance in > 90% for S aureus producing VAP. On the other hand, there exists a wide variation among different studies about the prevalence of colonization in patients on mechanical ventilation. Although there are some reports indicating a colonization rate similar to the rate we found,17,19,20 and some reported lower rates,21 the complexity of our patients; the use of only basic, standard measures of prevention; and the older age of our patients could lead to a higher colonization rate. Generalization of these data is difficult because these flora are not representative of many VAP episodes elsewhere. Our definition of a subsequent VAP episode made it impossible to rule out the persistence of the microorganism causing the first episode rather than the development of a subsequent episode in one case in which the pathogen and its antibiotic susceptibility were the same in both episodes. Antimicrobial resistance is increasing worldwide, and the organisms present in our ICU may become common in other ICUs in the future.

In conclusion: We found that ATS/IDSA guidelines-based strategy was preferable to ETA-based strategy for antibiotic selection in patients with a first episode of VAP, rather than a recurrent episode. ATS/IDSA guidelines led to a significantly higher rate of early and appropriate therapy, although at the cost of more antimicrobial-days.

Author contributions: Dr Luna had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Dr Luna: contributed to the study design, data analysis, and writing of the manuscript, and served as principal author.

Dr Sarquis: contributed to the study design, recruitment of patients, data collection, and revision of the manuscript.

Dr Niederman: contributed to the study design, data analysis, and writing the manuscript.

Dr Sosa: contributed to the study design, recruitment of patients, data collection, and revision of the manuscript.

Dr Otaola: contributed to recruitment of patients, data collection, and revision of the manuscript.

Dr Bailleau: contributed to recruitment of patients, data collection, and revision of the manuscript.

Dr Vay: contributed to the study design, the microbiologic work, and revision of the manuscript.

Dr Famiglietti: contributed to the study design, the microbiologic work, and revision of the manuscript.

Dr Irrazabal: contributed to support during patient recruitment and revision of the manuscript.

Dr Capdevila: contributed to support during patient recruitment and revision of the manuscript.

Financial/nonfinancial disclosures: The authors have reported to CHEST the following conflicts of interest: Dr Luna has participated in advisory boards for Pfizer, Inc; Bayer AG; and AstraZeneca PLC, and has given talks for Pfizer, Inc. All other authors have reported to CHEST that no potential conflicts of interest exist with any companies/organizations whose products or services may be discussed in this article.

Role of sponsors: Asociacion Cooperadora del Hospital de Clinicas was the administrator of the financial funds.

Other contributions: The authors thank all the clinicians who took part in the preparation of this paper.

ATS/IDSA

American Thoracic Society/Infectious Diseases Society of America

cfu

colony-forming units

ETA

endotracheal aspirate

MDR

multidrug-resistant

VAP

ventilator-associated pneumonia

American Thoracic Society; Infectious Diseases Society of America. Guidelines for the management of adults with hospital-acquired, ventilator-associated, and healthcare-associated pneumonia. Am J Respir Crit Care Med. 2005;171(4):388-416. [CrossRef] [PubMed]
 
Vincent JL, Bihari DJ, Suter PM, et al.;. 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. 1995;274(8):639-644. [CrossRef] [PubMed]
 
Vincent JL, Rello J, Marshall J, et al; EPIC II Group of Investigators. International study of the prevalence and outcomes of infection in intensive care units. JAMA. 2009;302(21):2323-2329. [CrossRef] [PubMed]
 
Heyland DK, Cook DJ, Griffith L, Keenan SP, Brun-Buisson C; The Canadian Critical Trials Group. The attributable morbidity and mortality of ventilator-associated pneumonia in the critically ill patient. Am J Respir Crit Care Med. 1999;159(4 pt 1):1249-1256. [CrossRef] [PubMed]
 
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(4):1230-1235. [CrossRef] [PubMed]
 
Bergmans DC, Bonten MJ, Gaillard CA, et al. Indications for antibiotic use in ICU patients: a one-year prospective surveillance. J Antimicrob Chemother. 1997;39(4):527-535. [CrossRef] [PubMed]
 
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(5):387-394. [CrossRef] [PubMed]
 
Luna CM, Vujacich P, Niederman MS, et al. Impact of BAL data on the therapy and outcome of ventilator-associated pneumonia. Chest. 1997;111(3):676-685. [CrossRef] [PubMed]
 
Waterer GW, Wunderink RG. Controversies in the diagnosis of ventilator-acquired pneumonia. Med Clin North Am. 2001;85(6):1565-1581. [CrossRef] [PubMed]
 
Michel F, Franceschini B, Berger P, et al. Early antibiotic treatment for BAL-confirmed ventilator-associated pneumonia: a role for routine endotracheal aspirate cultures. Chest. 2005;127(2):589-597. [CrossRef] [PubMed]
 
Trouillet JL, Chastre J, Vuagnat A, et al. Ventilator-associated pneumonia caused by potentially drug-resistant bacteria. Am J Respir Crit Care Med. 1998;157(2):531-539. [CrossRef] [PubMed]
 
Jung B, Sebbane M, Chanques G, et al. Previous endotracheal aspirate allows guiding the initial treatment of ventilator-associated pneumonia. Intensive Care Med. 2009;35(1):101-107. [CrossRef] [PubMed]
 
Yang K, Zhuo H, Guglielmo BJ, Wiener-Kronish J. Multidrug-resistant Pseudomonas aeruginosa ventilator-associated pneumonia: the role of endotracheal aspirate surveillance cultures. Ann Pharmacother. 2009;43(1):28-35. [CrossRef] [PubMed]
 
Delclaux C, Roupie E, Blot F, Brochard L, Lemaire F, Brun-Buisson C. Lower respiratory tract colonization and infection during severe acute respiratory distress syndrome: incidence and diagnosis. Am J Respir Crit Care Med. 1997;156(4 pt 1):1092-1098. [CrossRef] [PubMed]
 
Sanders KM, Adhikari NKJ, Friedrich JO, Day A, Jiang X, Heyland D; Canadian Critical Care Trials Group. Previous cultures are not clinically useful for guiding empiric antibiotics in suspected ventilator-associated pneumonia: secondary analysis from a randomized trial. J Crit Care. 2008;23(1):58-63. [CrossRef] [PubMed]
 
Bouza E, Pérez A, Muñoz P, et al; Cardiovascular Infection Study Group. Ventilator-associated pneumonia after heart surgery: a prospective analysis and the value of surveillance. Crit Care Med. 2003;31(7):1964-1970. [CrossRef] [PubMed]
 
Ewig S, Torres A, El-Ebiary M, et al. Bacterial colonization patterns in mechanically ventilated patients with traumatic and medical head injury. Incidence, risk factors, and association with ventilator-associated pneumonia. Am J Respir Crit Care Med. 1999;159(1):188-198. [CrossRef] [PubMed]
 
Iregui M, Ward S, Sherman G, Fraser VJ, Kollef MH. Clinical importance of delays in the initiation of appropriate antibiotic treatment for ventilator-associated pneumonia. Chest. 2002;122(1):262-268. [CrossRef] [PubMed]
 
Drakulovic MB, Bauer TT, Torres A, Gonzalez J, Rodríguez MJ, Angrill J. Initial bacterial colonization in patients admitted to a respiratory intensive care unit: bacteriological pattern and risk factors. Respiration. 2001;68(1):58-66. [CrossRef] [PubMed]
 
Durairaj L, Mohamad Z, Launspach JL, et al. Patterns and density of early tracheal colonization in intensive care unit patients. J Crit Care. 2009;24(1):114-121. [CrossRef] [PubMed]
 
Garrard CS, A’Court CD. The diagnosis of pneumonia in the critically ill. Chest. 1995;108(2)(suppl):17S-25S. [CrossRef] [PubMed]
 

Figures

Figure Jump LinkFigure 1. A, Flow chart of the patients and VAP episodes in this study. From the 323 patients admitted to the study, 283 had performed at least one ETA; 83 of the 283 had at least one episode of clinically diagnosed VAP and had at least one ETA specimen cultured during the preceding 7 d, as well as a BAL sample. B, ETA/BAL pair was defined as the combination of an ETA culture plus a BAL culture performed during the 7 d following an episode of clinically diagnosed VAP. Some patients had more than one episode, and some episodes had more than one ETA previously performed. An episode of VAP with BAL preceded by two ETAs performed during the previous 7 d was considered two separate ETA/BAL pairs. ETA = endotracheal aspirate; VAP = ventilator-associated pneumonia.Grahic Jump Location
Figure Jump LinkFigure 2. Distribution of the different pathogens of the first and subsequent episodes of VAP. First episodes were caused by Acinetobacter species (32.8%), Pseudomonas aeruginosa (29.5%), MRSA (16.4%), and other pathogens, including Haemophilus influenzae, Escherichia coli,Providencia species, Streptococcus viridans, Proteus mirabilis, and Klebsiella pneumoniae (16.2%). Subsequent episodes were restricted to MRSA (5.6%), Acinetobacter species (16.6%), P aeruginosa* (61.1%); and other pathogens (16.7%). *P = .025, significantly more frequent than in the first episode. MRSA = methicillin-resistant Staphylococcus aureus; spp = species. See Figure 1 legend for expansion of other abbreviations.Grahic Jump Location
Figure Jump LinkFigure 3. Appropriateness of therapy according to the result of ETA performed during the 2 d preceding the VAP diagnosis, during 3-7 d preceding the VAP diagnosis, and availability of ETA performed 3-7 d preceding the VAP diagnosis. The last situation is more realistic, as in the real world the ETA culture results become available ≥48 h after ETA retrieval. ATS/IDSA = American Thoracic Society/Infectious Diseases Society of America. See Figure 1 legend for expansion of other abbreviations.Grahic Jump Location
Figure Jump LinkFigure 4. Comparison of the number of antimicrobial-days between the different strategies for prescribing empirical antimicrobial therapy. The ETA-based strategy would lead to the use of 1,557 antimicrobial-days, significantly fewer than the ATS/IDSA guidelines-based therapy (1,942 antimicrobial-days). Considering separately ETAs performed < 3 d and those performed 3-7 d before VAP diagnosis, there was a trend toward fewer antimicrobial-days with the ETA-guided approach (1,184 vs 918 antimicrobial-days and 758 vs 639 antimicrobial-days, respectively) (P = NS). See Figure 1 and 3 legends for expansion of abbreviations.Grahic Jump Location

Tables

Table Graphic Jump Location
Table 1 —General Principles for Prescribing Antimicrobial Therapy Under the Two Different Strategies in This Study

ATS/IDSA = American Thoracic Society/Infectious Diseases Society of America; ETA = endotracheal aspirate; MRSA = methicillin-resistant Staphylococcus aureus; VAP = ventilator-associated pneumonia.

Table Graphic Jump Location
Table 2 —Demographic, Clinical, ETA-Culture Characteristics and Mortality of Patients Admitted in the Study and Comparison Between Patients Developing VAP vs Those Not Developing VAP

Data given as No. unless otherwise indicated. APACHE II = Acute Physiology and Chronic Health Evaluation II; cfu = colony-forming units; F = female; M = male; NS = not significant; TISS = Therapeutic Intervention Scoring System. See Table 1 legend for expansion of other abbreviations.

a 

“All” includes all intubated patients that had performed at least one routine ETA.

b 

VAP includes all patients with clinical diagnosis of VAP confirmed by BAL culture that had performed at least one ETA during the last 5 d.

c 

No VAP includes all those intubated patients with at least one ETA who did not develop clinically or microbiologically diagnosed VAP.

d 

P value: significance of the difference comparing VAP with subjects without VAP (no VAP).

e 

Colonization at the basal ETA obtained in the standard days.

Table Graphic Jump Location
Table 3 —Microorganisms Detected by BALa and by Routine ETAb Culture

Organism names and data in bold type are multidrug-resistant bacteria. MSSA = methicillin-sensitive Staphylococcus aureus. See Table 1 and 2 legends for expansion of other abbreviations.

a 

Microorganisms (n = 79) detected by BAL culture at a concentration ≥ 104 cfu/mL in 65 episodes of VAP.

b 

Microorganisms (n = 184) detected by routine ETA at a concentration ≥ 103 cfu/mL during the 7 d preceding the development of VAP (n = 114, during the 2 d before, and n = 70, 3-7 d before VAP diagnosis).

References

American Thoracic Society; Infectious Diseases Society of America. Guidelines for the management of adults with hospital-acquired, ventilator-associated, and healthcare-associated pneumonia. Am J Respir Crit Care Med. 2005;171(4):388-416. [CrossRef] [PubMed]
 
Vincent JL, Bihari DJ, Suter PM, et al.;. 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. 1995;274(8):639-644. [CrossRef] [PubMed]
 
Vincent JL, Rello J, Marshall J, et al; EPIC II Group of Investigators. International study of the prevalence and outcomes of infection in intensive care units. JAMA. 2009;302(21):2323-2329. [CrossRef] [PubMed]
 
Heyland DK, Cook DJ, Griffith L, Keenan SP, Brun-Buisson C; The Canadian Critical Trials Group. The attributable morbidity and mortality of ventilator-associated pneumonia in the critically ill patient. Am J Respir Crit Care Med. 1999;159(4 pt 1):1249-1256. [CrossRef] [PubMed]
 
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(4):1230-1235. [CrossRef] [PubMed]
 
Bergmans DC, Bonten MJ, Gaillard CA, et al. Indications for antibiotic use in ICU patients: a one-year prospective surveillance. J Antimicrob Chemother. 1997;39(4):527-535. [CrossRef] [PubMed]
 
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(5):387-394. [CrossRef] [PubMed]
 
Luna CM, Vujacich P, Niederman MS, et al. Impact of BAL data on the therapy and outcome of ventilator-associated pneumonia. Chest. 1997;111(3):676-685. [CrossRef] [PubMed]
 
Waterer GW, Wunderink RG. Controversies in the diagnosis of ventilator-acquired pneumonia. Med Clin North Am. 2001;85(6):1565-1581. [CrossRef] [PubMed]
 
Michel F, Franceschini B, Berger P, et al. Early antibiotic treatment for BAL-confirmed ventilator-associated pneumonia: a role for routine endotracheal aspirate cultures. Chest. 2005;127(2):589-597. [CrossRef] [PubMed]
 
Trouillet JL, Chastre J, Vuagnat A, et al. Ventilator-associated pneumonia caused by potentially drug-resistant bacteria. Am J Respir Crit Care Med. 1998;157(2):531-539. [CrossRef] [PubMed]
 
Jung B, Sebbane M, Chanques G, et al. Previous endotracheal aspirate allows guiding the initial treatment of ventilator-associated pneumonia. Intensive Care Med. 2009;35(1):101-107. [CrossRef] [PubMed]
 
Yang K, Zhuo H, Guglielmo BJ, Wiener-Kronish J. Multidrug-resistant Pseudomonas aeruginosa ventilator-associated pneumonia: the role of endotracheal aspirate surveillance cultures. Ann Pharmacother. 2009;43(1):28-35. [CrossRef] [PubMed]
 
Delclaux C, Roupie E, Blot F, Brochard L, Lemaire F, Brun-Buisson C. Lower respiratory tract colonization and infection during severe acute respiratory distress syndrome: incidence and diagnosis. Am J Respir Crit Care Med. 1997;156(4 pt 1):1092-1098. [CrossRef] [PubMed]
 
Sanders KM, Adhikari NKJ, Friedrich JO, Day A, Jiang X, Heyland D; Canadian Critical Care Trials Group. Previous cultures are not clinically useful for guiding empiric antibiotics in suspected ventilator-associated pneumonia: secondary analysis from a randomized trial. J Crit Care. 2008;23(1):58-63. [CrossRef] [PubMed]
 
Bouza E, Pérez A, Muñoz P, et al; Cardiovascular Infection Study Group. Ventilator-associated pneumonia after heart surgery: a prospective analysis and the value of surveillance. Crit Care Med. 2003;31(7):1964-1970. [CrossRef] [PubMed]
 
Ewig S, Torres A, El-Ebiary M, et al. Bacterial colonization patterns in mechanically ventilated patients with traumatic and medical head injury. Incidence, risk factors, and association with ventilator-associated pneumonia. Am J Respir Crit Care Med. 1999;159(1):188-198. [CrossRef] [PubMed]
 
Iregui M, Ward S, Sherman G, Fraser VJ, Kollef MH. Clinical importance of delays in the initiation of appropriate antibiotic treatment for ventilator-associated pneumonia. Chest. 2002;122(1):262-268. [CrossRef] [PubMed]
 
Drakulovic MB, Bauer TT, Torres A, Gonzalez J, Rodríguez MJ, Angrill J. Initial bacterial colonization in patients admitted to a respiratory intensive care unit: bacteriological pattern and risk factors. Respiration. 2001;68(1):58-66. [CrossRef] [PubMed]
 
Durairaj L, Mohamad Z, Launspach JL, et al. Patterns and density of early tracheal colonization in intensive care unit patients. J Crit Care. 2009;24(1):114-121. [CrossRef] [PubMed]
 
Garrard CS, A’Court CD. The diagnosis of pneumonia in the critically ill. Chest. 1995;108(2)(suppl):17S-25S. [CrossRef] [PubMed]
 
NOTE:
Citing articles are presented as examples only. In non-demo SCM6 implementation, integration with CrossRef’s "Cited By" API will populate this tab (http://www.crossref.org/citedby.html).

Some tools below are only available to our subscribers or users with an online account.

Related Content

Customize your page view by dragging & repositioning the boxes below.

Find Similar Articles
CHEST Journal Articles
Guidelines
Feverish illness in children: assessment and initial management in children younger than 5 years.
National Collaborating Centre for Women's and Children's Health | 8/28/2009
Blepharitis.
American Academy of Ophthalmology | 6/5/2009
  • CHEST Journal
    Print ISSN: 0012-3692
    Online ISSN: 1931-3543