Indications for and Utility of Tracheal Aspirate Cultures for the Diagnosis of VAI

Kathleen Chiotos, Giyoung Lee, Guy I. Sydney, Heather Wolfe, Jennifer Blumenthal, Hannah Stinson, Julie Harab, Danielle Traynor, Joseph Piccione, Ashlee Doll, Garrett Keim, Charlotte Woods-Hill, Megan Jennings, Rebecca M. Harris, Jeffrey S. Gerber, Aaditya Dudhia, Nancy McGowan, Jennalyn BurkeAntimicrobial Stewardship & Healthcare Epidemiology 2021 · DOI

Abstract

Background: Tracheal aspirate bacterial cultures are routinely collected in mechanically ventilated children for the evaluation of ventilator-associated infections (VAIs). However, frequent bacterial colonization of endotracheal and tracheostomy tubes contributes to the marginal performance characteristics of the test for diagnosing VAI. Published literature characterizing drivers of culture collection and the predictive value of positive cultures are limited.

Methods: This single-center, retrospective cohort study included children admitted to the pediatric intensive care unit who were receiving mechanical ventilation for at least 48 hours and had 1 or more semiquantitative tracheal aspirate cultures collected between September 1, 2019, and August 31, 2020. Indications for culture collection were determined through medical record review. Diagnosis of VAI was based on treating-clinician documentation.

Results: In total, 625 tracheal aspirate cultures were performed in 261 unique patients. Common indications for culture collection included isolated fever or hypothermia (20%), fever with an increase in oxygen requirement or PEEP (11%), isolated increase in oxygen requirement or PEEP (11%), or isolated secretion change (9%). Overall, 230 cultures (37%) were positive and 218 (35%) Gram stains were purulent. Odds of VAI diagnosis increased with both the number of indications for culture collection and purulent Gram stain, but not with positive culture.

Conclusions: Number and type of clinical signs were not associated with tracheal aspirate culture positivity or purulence on Gram stain, but they were associated with a clinical diagnosis of VAI. These findings suggest that positive tracheal aspirate cultures may not aid clinicians in the diagnosis of VAI, and they highlight the opportunity for improved diagnostic stewardship.

Introduction

In the Pediatric Intensive Care Unit (PICU), children on mechanical ventilation are at high risk for ventilator-associated infections (VAI). To identify these infections early, clinicians frequently order tracheal aspirate cultures—sampling the fluid from the breathing tube—whenever a patient develops a fever or requires increased respiratory support.

However, a major diagnostic hurdle exists: the breathing tubes themselves quickly become colonized by bacteria that are not necessarily causing an active infection. This makes it difficult to distinguish between simple colonization and a true infection. This study investigates whether the clinical signs that trigger these tests actually correlate with laboratory results and whether those results assist clinicians in diagnosing VAI.

Methods

The researchers conducted a single-center, retrospective cohort study at a pediatric academic center, analyzing data from September 2019 through August 2020. The study population consisted of children ventilated for at least 48 hours who had at least one semiquantitative tracheal aspirate culture collected.

The study utilized a clinical dataset of 625 tracheal aspirate cultures taken from 261 unique pediatric patients. Key metrics analyzed included:

  • Clinical Indications: Triggers for culture collection, such as fever, hypothermia, changes in secretions, or increased respiratory support (oxygen or PEEP).
  • Laboratory Findings: "Positive" cultures (moderate/heavy growth of non-commensal bacteria) and "purulent" Gram stains (presence of moderate to many white blood cells).
  • Clinical Diagnosis: Whether the treating physician documented a diagnosis of VAI.

Logistic regression, accounting for patient clustering, was performed to estimate the association between indications for culture collection and laboratory findings or VAI diagnosis.

Results

Clinical Triggers and Laboratory Findings

The study found that clinical signs—the very triggers that prompt a clinician to order a culture—were not associated with whether that culture actually came back positive. The distribution of these clinical indications is shown in Figure 1.

Distribution of clinical indications for tracheal aspirate culture collection.

Overall, 37% (230/625) of cultures were positive and 35% (218/625) of Gram stains showed purulence. Notably, isolated fever was actually negatively associated with a purulent Gram stain (OR 0.49; 95% CI, 0.30–0.81; P = .005). This suggests that in many ventilated children, fever may stem from non-pulmonary sources, yet it remains a primary driver for respiratory testing.

Predictors of VAI Diagnosis

When determining a final diagnosis of VAI, clinicians appeared to weigh the number of clinical symptoms and the presence of white blood cells (purulence) heavily. However, the actual growth of bacteria in the culture did not significantly correlate with the diagnosis.

Odds ratios for VAI diagnosis based on clinical indications and laboratory findings.

As illustrated in Figure 2, the odds of a VAI diagnosis increased significantly with the number of clinical indications and the presence of a purulent Gram stain. Conversely, a positive culture (OR 1.5, 95% CI 0.9-2.3, P=0.09) did not reach statistical significance as a predictor for diagnosis. This implies that positive cultures in this setting may often represent "noise" (colonization) rather than "signal" (infection), potentially leading to unnecessary antibiotic use.

Discussion and Conclusion

This study highlights a significant "diagnostic disconnect" in the PICU. While clinical signs and Gram stain purulence drive the diagnosis of VAI, the results of tracheal aspirate cultures themselves do not appear to provide additional diagnostic utility.

Limitations

The authors acknowledge several limitations, including the single-center scope, which may limit generalizability to other institutions. Additionally, there is no "gold standard" for VAI diagnosis; the study relied on clinician documentation, which is inherently subjective. The retrospective design also limits the analysis to the quality and completeness of existing medical records.

Conclusion

These findings highlight a major opportunity for diagnostic stewardship. Future research should focus on developing stricter, evidence-based guidelines for when to order tracheal cultures to reduce the diagnostic burden and prevent the potential over-treatment of bacterial colonization. Improving the specificity of VAI diagnosis is crucial for optimizing antibiotic use and improving patient outcomes in the PICU.

Summary updated

Abstract

Background: Tracheal aspirate bacterial cultures are routinely collected in mechanically ventilated children for the evaluation of ventilator-associated infections (VAIs). However, frequent bacterial colonization of endotracheal and tracheostomy tubes contribute to the marginal performance characteristics of the test for diagnosing VAI. Published literature characterizing drivers of culture collection and the predictive value of positive cultures are limited. Methods: This single-center, retrospective cohort study included children admitted to the pediatric intensive care unit who were receiving mechanical ventilation for at least 48 hours and had 1 or more semiquantitative tracheal aspirate cultures collected between September 1, 2019, and August 31, 2020. Indications for culture collection were determined through medical record review and included fever, hypothermia, tracheal secretion changes, radiographic pneumonia, increased oxygen requirement, and/or increased positive end-expiratory pressure (PEEP). A positive culture was defined as moderate or heavy growth of a noncommensal bacterial organism. A purulent Gram stain was defined as detection of moderate or many white blood cells. Diagnosis of VAI was based on treating-clinician documentation and was ascertained through medical record review. Logistic regression accounting for clustering by patient was performed to estimate the association between indications for culture collection and (1) culture positivity, (2) purulent Gram stain, and (3) diagnosis of VAI. Results: In total, 625 tracheal aspirate cultures were performed in 261 unique patients. Common indications for culture collection included isolated fever or hypothermia (n = 124, 20%), fever with an increase in oxygen requirement or PEEP (n = 71, 11%), isolated increase in oxygen requirement or PEEP (n = 67, 11%), or isolated secretion change (n = 54, 9%) (Figure 1). Overall, 230 cultures (37%) were positive and 218 (35%) Gram stains were purulent. There were no associations between culture indications and a positive culture. Presence of isolated fever was negatively associated with a purulent Gram stain (odds ratio [OR], 0.49; 95% CI, 0.30–0.81; P = .005); otherwise, there were no associations between indication and purulent Gram stain. Finally, in a multivariable model, odds of VAI diagnosis increased with both the number of indications for culture collection and purulent Gram stain, but not with positive culture (Figure 2). Conclusions: Number and type of clinical signs were not associated with tracheal aspirate culture positivity or purulence on Gram stain, but they were associated with a clinical diagnosis of VAI. These findings suggest that positive tracheal aspirate cultures may not aid clinicians in the diagnosis of VAI, and they highlight the opportunity for improved diagnostic stewardship. Funding: No Disclosures: None
BibTeX
@inproceedings{chiotos2021indications,
  author = {Kathleen Chiotos and Giyoung Lee and Guy I. Sydney and Heather Wolfe and Jennifer Blumenthal and Hannah Stinson and Julie Harab and Danielle Traynor and Joseph Piccione and Ashlee Doll and Garrett Keim and Charlotte Woods-Hill and Megan Jennings and Rebecca M. Harris and Jeffrey S. Gerber and Aaditya Dudhia and Nancy McGowan and Jennalyn Burke},
  title = {{Indications for and Utility of Tracheal Aspirate Cultures for the Diagnosis of VAI}},
  booktitle = {Antimicrobial Stewardship \& Healthcare Epidemiology},
  year = {2021},
  doi = {10.1017/ash.2021.116},
}
Figure