Diagnostic Value of Endotracheal Aspirates Sonication on Ventilator-Associated Pneumonia Microbiologic Diagnosis

Abstract

Microorganisms are able to form biofilms within respiratory secretions. Methods to disaggregate such biofilms before utilizing standard, rapid, or high throughput diagnostic technologies may aid in pathogen detection during ventilator associated pneumonia (VAP) diagnosis. Our aim was to determine if sonication of endotracheal aspirates (ETA) would increase the sensitivity of qualitative, semi-quantitative, and quantitative bacterial cultures in an animal model of pneumonia caused by Pseudomonasaeruginosa or by methicillin resistant Staphylococcusaureus (MRSA).

Material and methods: P.aeruginosa or MRSA was instilled into the lungs or the oropharynx of pigs in order to induce severe VAP. Time point assessments for qualitative and quantitative bacterial cultures of ETA and bronchoalveolar lavage (BAL) samples were performed at 24, 48, and 72 h after bacterial instillation. In addition, at 72 h (autopsy), lung tissue was harvested to perform quantitative bacterial cultures. Each ETA sample was microbiologically processed with and without applying sonication for 5 min at 40 KHz before bacterial cultures. Sensitivity and specificity were determined using BAL as a gold-standard. Correlation with BAL and lung bacterial burden was also determined before and after sonication. Assessment of biofilm clusters and planktonic bacteria was performed through both optical microscopy utilizing Gram staining and Confocal Laser Scanning Microscopy utilizing the LIVE/DEAD^®BacLight kit.

Results: 33 pigs were included, 27 and 6 from P.aeruginosa and MRSA pneumonia models, respectively. Overall, we obtained 85 ETA, 69 (81.2%) from P.aeruginosa and 16 (18.8%) from MRSA challenged pigs. Qualitative cultures did not significantly change after sonication, whereas quantitative ETA cultures did significantly increase bacterial counting. Indeed, sonication consistently increased bacterial burden in ETAs at 24, 48, and 72 h after bacterial challenge. Sonication also improved sensitivity of ETA quantitative cultures and maintained specificity at levels previously reported and accepted for VAP diagnosis.

Conclusion: The use of sonication in ETA respiratory samples needs to be clinically validated since sonication could potentially improve pathogen detection before standard, rapid, or high throughput diagnostic methods used in routine microbial diagnostics.

Keywords: Pseudomonas aeruginosa; Staphylococcus aureus; animal model; biofilm; endotracheal aspirate; sonication; ventilator associated pneumonia.

Figure 1

Bland–Altman plot showing the differences between sonicated and non-sonicated endotracheal aspirate (ETA) paired samples (Mean ± SD). Eighty-one quantitative ETA cultures were performed twice: before and after sonication. Each single dot represents the difference between bacterial burden of sonicated minus non-sonicated ETA, using non-sonicated values in the X axis. Of notice, 52 out of 81 (64%) dots were allocated above zero, 21 out of 81 (26%) at zero, and 8 out of 81 (10%) below zero. Mean difference ± SD was 0.2 ± 0.3 log CFU/mL (p < 0.001).

Figure 2

Correlation between quantitative cultures of ETA and BAL for non-sonicated (A) and sonicated samples (B). The histograms represent the number of ETA (on the top) or BAL (right) on each bacterial concentration. Moderate correlation was found between non-sonicated and sonicated ETA and BAL(r = 0.51 and r = 0.56, respectively; p = 0.75). ETA and BAL cut-off was 6 log CFU/mL and 4 log CFU/mL, respectively. After sonication the number of true positives samples increased from 6 (11%) to 7 (13%) out of 55 samples.

Figure 3

Correlation between ETA and lung tissue bacterial burden. Of note, in 15 cases a moderate correlation was found between non-sonicated or sonicated ETA and lung tissue bacterial burden (r = 0.52 and r = 0.57, respectively; p = 0.86).

Figure 4

Confocal Laser Scanning Microscopy (CLSM) image of ETA samples before and after sonication in P. aeruginosa and MRSA infection models. Of note, viable bacteria (in green) stained with the SYTO 9, a green Fluorescent Nucleic Acid Stain, were visible. Dead bacteria (in red) stained with the propidium iodide (PI) were barely detected. The nucleus and cytoplasm of eukaryotic cells from the pig were also stained nonspecifically with the PI and SYTO 9 (large red and green blotches). (A) vs. (B) correspond to non-sonicated vs. sonicated ETA (5.14 vs. 6.20 log CFU/mL) after 72 h of P. aeruginosa instillation with biofilm clusters visible (white arrow). Similarly, (C) vs. (D) correspond to non-sonicated vs. sonicated ETA (1.38 vs. 1.88 log CFU/mL) after 72 h of MRSA instillation with biofilm clusters and free-floating cocci (white arrows), respectively. (A,B) images belong to a pig treated with IV cephalosporin and (C,D) to a pig treated with IV lipoglycopeptide. (E) 3D reconstruction of a Gram-negative bacilli biofilm cluster (Imaris, Bitplane, Oxford instruments Company, Abingdon, UK).

Figure 5

Optical microscopy images of Gram stained ETA before and after sonication. Images (A) vs. (B), (C) vs. (D) and (E) vs. (F) correspond to non-sonicated vs. sonicated ETA at 24, 48, and 72 h, respectively. Biofilm aggregates (white arrow) were observed before and after sonication. In contrast, free-floating bacteria were mainly detected after sonication (black arrow).