Ventilator-Associated Tracheobronchitis in a Mixed Surgical and Medical ICU Population FREE TO VIEW

Ventilator-associated pneumonia (VAP) is common in intubated patients and is associated with increased morbidity, mortality, and health-care costs.^1,2 Ventilator-associated tracheobronchitis (VAT) has been proposed as an intermediate condition between simple colonization of the upper airways and VAP^3,4 and has been reported to occur in 2.7%⁵ to 10.6%⁶ of intubated patients. Treatment of VAT with antibiotics has been proposed as a means to decrease subsequent progression to VAP and to improve outcomes.³ Two small randomized clinical trials^7,8 evaluating the efficacy of such treatment showed positive effects, including lower rates of subsequent VAP and decreased duration of mechanical ventilation. Unfortunately, few prospective studies have been published that directly evaluate VAT and controversies concerning diagnostic criteria and true distinction from VAP exist.^3,9 The contention that VAT is associated with adverse outcomes, such as increased lengths of stay, has been refuted by some studies,¹⁰ and increased mortality with the diagnosis has not been established.¹¹ Therefore, given the paucity of North American data concerning VAT, we set out to perform a prospective cohort study of VAT in ICU patients to determine its incidence and influence on patient outcomes compared with VAP.

The study was conducted in the surgical (24 beds) and medical (19 beds) ICUs of Barnes-Jewish Hospital, a 1,200-bed urban teaching hospital in St. Louis, Missouri. During the course of 365 days, beginning on January 19, 2009, the ICU patient rosters were screened daily. Patients mechanically ventilated for > 48 h were monitored daily for the development of either VAT or VAP. The Washington University Human Research Protection Office reviewed and approved the research protocol (HRPO number 08-1233).

Patients who met the prospectively defined definitions of VAP and VAT during the study period were included in the analysis unless they met one of the following exclusion criteria: presence of another ongoing nosocomial infection requiring antimicrobial treatment, presence of a tracheostomy at the time of VAT or VAP suspicion, significant immune suppression defined as prolonged neutropenia (> 1 week), HIV-positive with an absolute CD4 cell count < 50/μL during the preceding 6 months, or chronic steroid therapy at a dosage ≥ 40 mg of prednisolone daily for a duration of > 4 weeks.

The following baseline characteristics were recorded at the time of VAT or VAP diagnosis: age, sex, race, height, weight, maximum body temperature in the previous 24 h, leukocyte count, quantity and character of pulmonary secretions, presence of comorbid conditions, modified Clinical Pulmonary Infection Score (CPIS),¹² APACHE (Acute Physiology and Chronic Health Evaluation) II¹³ score at ICU admission, ratio of Pao₂ to Fio₂, and the causative organism(s) associated with VAT and VAP. Respiratory secretions were sent for microbiologic analysis only at times of suspected infection, and there was no routine surveillance of respiratory cultures performed during the study period.

Endotracheal tube aspirates (ETAs) were considered positive if ≥ 10⁵ colony-forming units (cfu)/mL were identified. BAL samples were considered positive at ≥ 10⁴ cfu/mL. Causative pathogens were considered multidrug resistant (MDR) if they were resistant to two or more of the primary antibiotics used to treat nosocomial pneumonia at our hospital (cefepime, piperacillin-tazobactam, meropenem).

Secretion volume and character are routinely recorded every 8 h by nurses and respiratory therapists caring for intubated patients. Secretion volume is graded on the following scale: none, scant, small (< 30 mL/d), moderate (30-100 mL/d), or large (> 100 mL/d). Patients who were diagnosed with VAT or VAP were followed until hospital discharge or death and the following parameters were recorded: duration of mechanical ventilation, duration of ICU stay, duration of hospitalization, requirement for tracheostomy, total days of antibiotic usage in the ICU, and days of antibiotics used for treatment of the episode of VAT or VAP that led to study enrollment.

Some patients were judged to progress from VAT to VAP based on development of a new or progressive infiltrate in the 96 h after initial diagnosis of VAT. In order to be included in this grouping, patients had to display continued evidence of ongoing infection as evidenced by temperature > 38.3°C, temperature < 36.0°C, or leukocyte count > 12,000/μL. If a new organism associated with VAP was isolated from the respiratory cultures of a patient previously diagnosed with VAT, then they were judged not to have progressed from VAT to VAP. These patients were instead considered to have a new distinct episode of VAP, which was not included in the analysis. Subsequent distinct episodes of VAT or VAP occurring more than 96 h following the initial episode of VAT or VAP were also not included in this analysis.

VAT was defined as the presence of all of the following in a patient endotracheally intubated and receiving mechanical ventilation for > 48 h: body temperature > 38.3°C or < 36.0°C, new or increased purulent tracheal secretions, positive culture of tracheal secretions at a concentration of ≥ 10⁵ cfu/mL, and no new or progressive infiltrate on portable chest radiograph. VAP was defined as the presence of a new or progressive pulmonary infiltrate and two of the following: temperature > 38.3°C or < 36.0°C, leukocyte count > 12,000/μL or < 4,000/μL, or purulent tracheal secretions. The diagnosis of VAP was considered to be microbiologically confirmed if either BAL or ETA cultures had significant growth. The presence or absence of a new or progressive radiographic infiltrate was based on the interpretation of the chest radiograph by board-certified radiologists who were blinded to the study. All classifications, including the radiographs and laboratory data used in their determinations, were prospectively reviewed by one of the investigators (J. D.) and confirmed by a second investigator (M. H. K.). Culture-negative cases of VAP were defined as cases meeting all the clinical criteria for VAP but with nonsignificant ETA or BAL cultures in the presence of newly started antibiotics within 48 h of culture sampling.

ETAs were only obtained when there was suspicion of either VAT or VAP. Specimens were obtained by respiratory therapist or nurses using a deep tracheal suctioning technique to obtain a specimen for culture. All ETAs and BAL cultures were processed using quantitative methods as previously described.¹⁴ The tubes containing the respiratory specimens were first vortexed for 15 s. A 0.01-mL calibrated loop was placed into the respective specimens and then onto the center of three media plates (blood agar, chocolate agar, and MacConkey agar). The media plates were then streaked using the pinwheel streak method and incubated in CO₂ at 35°C. Bacterial culture growth was quantitated according to the number of colonies observed per plate: < 10 colonies per plate represented < 10³ cfu/mL; 10 to 100 colonies per plate represented 10³ to 10⁴ cfu/mL; 100 to 1,000 colonies per plate represented 10⁴ to 10⁵ cfu/mL; and > 1,000 colonies per plate represented > 10⁵ cfu/mL. All identified microorganisms were reported with their antibiotic sensitivities.

All comparisons were unpaired, and all tests of significance were two-tailed. Continuous variables were compared using the Student t test for normally distributed variables and the Mann-Whitney U test for nonnormally distributed variables. The χ² or Fisher exact tests were used to compare categorical variables. For all analyses, a two-tailed P value < .05 was considered statistically significant. Statistical analyses were performed using SPSS, Version 11.0 for Windows (SPSS, Inc; Chicago, Illinois).

During the course of 1 year, 2,060 patients admitted to the medical and surgical ICUs required mechanical ventilation for > 48 h. Among these patients, 111 patients (5.4%) were identified as having either VAT or VAP. The baseline characteristics of the patients are shown in Table 1. There were 28 (25.2%) patients with VAT and 83 (74.8%) patients with VAP. Twenty-two (78.6%) of the patients with VAT had moderate or large secretion volume, whereas six (21.4%) had small secretion volume recorded during the 24-h period when the diagnosis of VAT occurred. Either ETA or BAL fluid was culture positive in 73/83 (88.0%) of the VAP cases (39 with BAL and 34 with ETAs). In the 10 (12.0%) patients with culture-negative VAP, all 10 received new antibiotics within 48 h of having respiratory samples obtained for culture.

Among patients with VAP, 41 (49.4%) had radiographic infiltrates present prior to the diagnosis of VAP attributed to ARDS, pulmonary edema, atelectasis, or aspiration. All 41 of these patients had new infiltrates or progression of their preexisting infiltrates qualifying for the diagnosis of VAP as determined by the investigators. There was no statistically significant difference in baseline medical comorbidities or APACHE II scores between the VAT and VAP groups. The CPIS score was significantly greater in the VAP group compared with patients with VAT at the time of infection diagnosis. The distribution of VAT and VAP relative to intubation and the start of mechanical ventilation is shown in Figure 1. The mean onset of VAT was 8.3 ± 4.8 days (median 7.5 days; interquartile range, 5.25-10.0 days), and the mean onset of VAP was 6.7 ± 4.1 days (P = .052) (median 5.0 days; interquartile range, 4.0-8.0 days).

The incidences of VAT and VAP are shown in Table 2. Among patients intubated for > 48 h, the overall incidence of VAT was 1.4%, and the overall incidence of VAP was 4.0%. This corresponds to a rate of 3.2 cases of VAT per 1,000 mechanical ventilator days and a rate of 9.4 cases of VAP per 1,000 mechanical ventilator days. The incidence of VAP was significantly greater than the incidence of VAT (P < .001). There was no statistically significant difference in the incidence of VAT in medical ICU patients (1.5%) when compared with the incidence of VAT in surgical ICU patients (1.3%). VAP occurred at a greater incidence in surgical ICU patients (5.3%) than in medical ICU patients (2.3%), P < .001.

VAT progressed to VAP in nine patients (32.1%) despite concurrent therapy with appropriate antibiotics that were predicted to be effective based on in vitro susceptibility testing of the VAT causative organisms in all nine patients. There was no statistical difference in the overall rate of appropriate initial antibiotic therapy in the VAT group (71.4%) as compared with the VAP group (71.1%).

Causative pathogens of VAT and VAP are shown in Table 3. VAT was caused by an MDR pathogen in nine (32.1%) patients and was polymicrobial in seven (25.0%) patients. VAP was caused by an MDR pathogen in 31 (37.3%) patients and was polymicrobial in 16 (19.3%) patients. Gram-positive organisms accounted for 37.5% of the pathogens isolated in patients with VAT and 27.8% of the pathogens isolated in patients with VAP. Overall, there was no significant difference among the individual causative bacterial pathogens for patients with VAT compared with patients with VAP. However, Enterobacteriaceae were more commonly associated with VAP compared with VAT (31 of 90 [34.4%] vs 4 of 32 [12.5%]; P = .022).

Outcomes of patients diagnosed with VAT and VAP are depicted in Table 4. There was no significant difference in ICU or hospital length of stay, duration of mechanical ventilation, hospital mortality, tracheostomy, or antibiotic use between the VAT and VAP groups. When the nine patients with VAT who subsequently developed VAP are removed from the analysis, there still are no significant differences between the VAT and VAP groups for any of the outcomes measured.

To our knowledge, this is the first and largest prospective study of VAT from North America. A recent meta-analysis¹¹ reviewing VAT identified only five reports^6,10,15‐17 deemed eligible for determining the incidence of VAT. The overall frequency of VAT was 11.5% based on these reports. However, most of these studies are limited by not being prospectively designed to evaluate VAT,^10,15‐17 not using the standard definition of VAT,^10,15‐17 and being conducted in very limited patient populations (head injury,¹⁶ post cardiac surgery,¹⁰ tertiary peritonitis¹⁵).

Nseir et al⁶ published the only study that prospectively evaluated the incidence of VAT based on the criteria that are most accepted currently. This investigation was performed in a French teaching hospital, included medical and surgical ICU patients, and reported an overall incidence of VAT of 10.6%. Surgical patients were more likely than medical patients to develop VAT (15.3% vs 9.9%). In our study a significantly lower overall incidence of VAT (1.4%) was noted and the incidence of VAT was similar in surgical and medical ICU patients.

Several potential factors may have contributed to the lower incidence of VAT reported in our study compared with the French study. Nseir et al⁶ reported a high incidence of COPD (58%) in their population compared with our study (15%). Patients with COPD might be diagnosed more frequently with VAT as they are more likely to produce larger quantities of purulent secretions and may be more likely to have bacterial colonization of their upper airways. Additionally, endotracheal aspirates were only sent for culture in our study when there was a clinical suspicion for the presence of VAT or VAP. Nseir et al⁶ reported that aspirates were routinely sent on admission and weekly thereafter in addition to at times of suspicion of infection. The presence of more positive cultures might lead to an increased incidence of VAT in equivocal cases.

In our study, 32.1% of patients initially diagnosed with VAT evolved to fulfill VAP criteria. This percentage is higher than that reported by Nseir et al⁶ (9.0% of patients). However, in the subsequent randomized trial⁸ of antibiotic therapy for VAT performed by these same investigators, 34% of patients with VAT developed subsequent VAP, but only three of 22 (13.6%) of these cases occurred in the group of patients who received antibiotics. These data suggest that the administration of antimicrobial therapy to patients with VAT may be an important factor influencing whether VAT progresses to VAP. However, our study suggests that such progression can still occur in some patients despite the presence of appropriate initial antimicrobial therapy for VAT.

VAT is believed to be a precursor or intermediate condition progressing to VAP.³ The lower incidence of VAT compared with VAP and the more delayed onset of VAT compared with VAP that we observed are at odds with this hypothesis and suggest that VAT is not necessarily a precursor of VAP. One potential explanation for this deviation from the hypothesis that VAT should precede VAP is the difference in diagnostic criteria used to define these infections. The diagnosis of VAT required a quantitative culture, whereas the diagnosis of VAP did not. Thus, intubated patients with purulent secretions that are culture negative would be excluded from the VAT group. This did not appear to be the case in our study, as we did not identify any patients with culture-negative purulent respiratory secretions who otherwise met the clinical criteria for VAT. An alternative explanation for the observed deviation is that VAT is not a precursor of VAP but that these are two distinct entities that can arise from the noninfected state and can either coexist or exist separately. Such a model could account for the higher incidence of VAP over VAT that we observed and could potentially explain the onset of VAT at a later time point. Additionally, VAT and VAP could be mutually exclusive, rather than steps in the progression of a single infection, when alveolar host defenses are adequate at a time when airway defenses are overwhelmed.

Several important limitations of our study should be noted. First, there was potential for substantial overlap between patients with VAT and those with VAP based on the definitions we used. The absence of new or progressive infiltrates observed on portable chest radiographs was the main differentiator between VAT and VAP. Given the nonspecificity of radiographic findings for VAP,⁹ as well as the lack of sensitivity for portable chest radiographs to identify new infiltrates, we cannot be certain that all our patients with VAP actually had pneumonia or that some of our patients with VAT did not actually have VAP. The overall similarity in median time to onset of VAT and VAP also suggests that there is considerable overlap between these two infections. Additionally, we required the same organism to be present in patients with VAT who progressed to VAP. Because we used quantitative culture methods, it is possible that we may have missed pathogens in low quantities that were present at the time of VAT and subsequently not associated with VAP. The second major limitation of our study is that we only examined patients with active infections, comparing those with VAT to those with VAP. Therefore, we cannot estimate the attributable mortality from VAT in our population. However, the lack of outcome differences between patients with VAT and those with VAP suggests that the attributable outcomes associated with VAT are similar to those of VAP. Third, as noted above, we cannot with certainty exclude the possibility that some of our patients with VAT were misclassified because we did not routinely perform CT scans searching for occult infiltrates. Fourth, we only obtained respiratory samples for culture when patients fulfilled the clinical criteria suggesting the presence of either VAT or VAP. This may have resulted in an underestimation of the occurrence of VAT.

Another limitation of our study is that we measured CPIS scores in all patients, although the CPIS was originally developed for use in patients with VAP.¹² Although the CPIS scores were significantly greater in patients with VAP compared with patients with VAT, there was overlap in the scores between them. This suggests that CPIS may be a poor diagnostic tool for VAP or that some of the patients with VAT may have actually had VAP. Similarly, 41 patients with VAP had infiltrates present prior to establishing the diagnosis of VAP. Therefore, it is possible that some of these patients developed VAT and not VAP given the prior presence of infiltrates. We also observed that the duration of antibiotic therapy for VAT was similar to that for patients with VAP. This also supports the possibility that the patients with VAT may actually have had VAP that was not clinically identified. We also did not perform any sensitivity analyses for our definition of VAT. For example, if we eliminated the requirement for either hyperthermia or hypothermia then a larger group of patients would have met the criteria for VAT. This could have increased the likelihood of identifying patients with VAT that preceded VAP. However, we did not identify any patient who was treated with antibiotics for VAT without the temperature criteria or the purulent secretion criteria. Additionally, loosening the diagnostic criteria for VAT, without demonstrating any outcome benefit in doing so, has the potential to increase antibiotic use, which could further drive increases in antimicrobial resistance. Finally, cultures for nonbacterial pathogens, such as viruses and fungi, were not routinely performed on the ETA or BAL samples. Therefore, other potential pathogens may have been present, accounting for the findings in some patients.

In conclusion, we demonstrated a significantly lower rate of VAT than previously reported.⁶ However, we identified patients with VAT that did progress to VAP, and patients diagnosed with VAT had similar outcomes to those with VAP. VAT does not appear to be a necessary precursor for VAP. This does not negate the role of upper airway colonization in the pathogenesis of VAP. Indeed, local and systemic host factors, bacterial virulence properties, and therapeutic interventions may influence whether upper airway colonization progresses to clinically recognizable VAP. Further studies are needed to identify patients with VAT who would benefit the most from antimicrobial therapy and those who could safely have antimicrobial therapy withheld or limited. Additionally, the optimal duration of antimicrobial therapy and route of therapy (parenteral, aerosolized) needs to be determined for VAT.

Author contributions: Drs Dallas and Kollef had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Dr Dallas: contributed to the study concept and design, statistical analysis, and the drafting of the manuscript.

Dr Skrupky: contributed to the study concept and design, the analysis and interpretation of data, and the drafting of the manuscript and critical revision for important intellectual content.

Dr Abebe: contributed to the study concept and design, formation of the study database, and critical revision of the manuscript for important intellectual content.

Dr Boyle: contributed to the study concept and design, the analysis and interpretation of data, and critical revision of the manuscript for important intellectual content.

Dr Kollef: contributed to the study concept and design and manuscript review.

Financial/nonfinancial disclosures: The 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.