Comparison of bronchoalveolar lavage and catheter lavage to confirm ventilator-associated lower respiratory tract infection

Abstract

Lower respiratory tract infection (LRTI) is a well recognised complication of artificial ventilation in intensive care units (ICU). Ideally, specimens for microbiological analysis should be obtained during bronchoscopy, but this is not always possible. Therefore, the microbiological diagnosis of lower respiratory tract infection by broncho-alveolar lavage (BAL) obtained during bronchoscopy was compared with catheter lavage (CL) with a balloon-tipped catheter. Adult patients with clinical evidence of lower respiratory tract infection in an adult ICU were randomly assigned to undergo BAL followed by CL or vice versa. Forty ml of normal saline 0.9% were instilled and then aspirated with a flexible bronchoscope to obtain BAL. A similar volume was instilled and aspirated with a 12-gauge Foley balloon-tipped catheter to obtain a C L sample. The number of inflammatory cells, epithelial cells and organisms seen by microscopy were quantified. Culture results were semi-quantified and classified as negative, positive, equivocal or contaminated. Seventy-nine paired specimens were obtained from 66 patients, including specimens from 10 patients taken on two or more occasions. Only 20% of BAL and 16% of CL had one or more epithelial cells and bacteria were seen in 26 BAL and 21 CL specimens, respectively; 35% of BAL and CL specimens were positive and there was a discrepancy in the culture result in only two cases. Staphylococcus aureus was the pathogen isolated most frequently and polymicrobial lower respiratory infection was diagnosed on 10 occasions (15%). CL fluid is as reliable as BAL in diagnosing lower respiratory tract infection in ICU. This approach does not require bronchoscopic expertise and utilises convenient laboratory techniques.

The diagnosis of ventilator-associated pneumonia using non-bronchoscopic, non-directed lung lavages

“…However, the diagnostic reliability of quantitative NBL was comparable with bronchoscopic techniques (Table 4.) and good qualitative concordance was found between bronchoscopic and NBL cultures, consistent with previous findings [8,16,17,18,33]. Lower positive predictive values (PPV) were found for NBL, reflecting a bias in sampling.…”

“…Our results suggest that it may be useful as a firstline technique for patients with clinical suspicion of VAP, as it has comparable diagnostic accuracy with bronchoscopic techniques [8,16,17,18] and is much more convenient to perform. Twice weekly qualitative surveillance NBLs can predict the causative organism of subsequent VAP episodes in most cases [17,37] provided that new antibiotics are not introduced in the interval between surveillance sampling and NBLs taken at onset of VAP.…”

“…However these techniques are expensive, have recognised morbidity [10] have produced conflicting results [11,12,13,14] and may not influence outcome [15]. Non-bronchoscopic lung lavages (NBL) have demonstrated comparable diagnostic accuracy [16,17,18] but have not been prospectively compared with both BAL and PSB in the same patient, ante-mortem. We evaluated diagnostic utility of: (a) quantitative culture of NBL, BAL and PSB in VAP; (b) serial NBL surveillance cultures in predicting onset of pneumonia; (c) the use of the Bacterial Index (BI) [12]; and (d) Clinical Pulmonary Infection Score (CPIS).…”

Accuracy and Precision of Quantitative Calibrated Loops in Transfer of Bronchoalveolar Lavage Fluid

“…After incubation, the colonies are counted on the plates and the number of CFU per milliliter is determined by multiplying the number of colonies by the dilution factor. Calibrated loops are widely used for quantitative BAL fluid cultures in the diagnostic and research settings (4,7,8,11,12,19,20,21,23).…”

Surveillance culture utility and safety using low‐volume blind bronchoalveolar lavage in the diagnosis of ventilator‐associated pneumonia

“…In addition, BM‐BAL show utility in guiding initial antibiotic decisions in 42% of patients, while allowing early cessation of antibiotics in a further 20% of patients where pneumonia diagnosis is doubtful 4 . Despite sensitivities and specificities of between 70% and 75% and negative predictive values above 90%, 4,11,35–38 BM‐BAL have low positive predictive values (14–17%) due to airway colonization and inconsistently defined diagnostic thresholds 11 . Where the local rates of inappropriate antibiotic prescription are greater than approximately 30%, there is potential utility in the use of surveillance cultures to support targeted antibiotic use.…”