Optimising the yield from bronchoalveolar lavage on human participants in infectious disease immunology research

Abstract

Bronchoalveolar lavage (BAL) is becoming a common procedure for research into infectious disease immunology. Little is known about the clinical factors which influence the main outcomes of the procedure. In research participants who underwent BAL according to guidelines, the BAL volume yield, and cell yield, concentration, viability, pellet colour and differential count were analysed for association with important participant characteristics such as active tuberculosis (TB) disease, TB exposure, HIV infection and recent SARS-CoV-2 infection. In 337 participants, BAL volume and BAL cell count were correlated in those with active TB disease, and current smokers. The right middle lobe yielded the highest volume. BAL cell and volume yields were lower in older participants, who also had more neutrophils. Current smokers yielded lower volumes and higher numbers of all cell types, and usually had a black pellet. Active TB disease was associated with higher cell yields, but this declined at the end of treatment. HIV infection was associated with more bloody pellets, and recent SARS-CoV-2 infection with a higher proportion of lymphocytes. These results allow researchers to optimise their participant and end assay selection for projects involving lung immune cells.

Figure 1

Bronchoscopy and bronchoalveolar lavage. (A) A research participant is about to have a Fujifilm flexible videoscope with 3.2 mm working channel inserted through her nose while under conscious sedation. Clinicians performing the procedure are authors STM and JAS. (B) Sequential photos from within the airways (from top to bottom: vocal cords before intubation, main carina, and right middle lobe medial segment subsegmental bronchi) taken with the bronchoscope. (C) Results of the bronchoalveolar lavage in a non-smoker, aliquots of saline with foamy surfactant on the surface. (D) An example of significant staining of the lavage fluid with black particulate matter from tobacco smoking.

Figure 2

Scatter plot with regression trends showing the relationship between BAL total cell count and BAL volume yield. (A) The correlation between the log cell count and BAL volume in the Household Contacts and Community Control groups (the two groups without any current active infection), stratified by smoking status, either ‘current’ or ‘other’ (prior smokers and never smokers). The trend line for current smokers indicates an increase of 0.61 × 10⁶ cells collected per ml increase in volume (SE 0.21 × 10⁶ cells, P = 0.004) (B) Correlation between the log cell count and BAL volume among TB groups. The cell yield increases by 0.32 × 10⁶ cells (SE 0.15 × 10⁶, P = 0.044) per ml BAL volume among participants before TB Pre-Treatment (TB pre-treatment), by 0.54 × 10⁶ cells (SE 0.14 × 10⁶, P < 0.001) per ml among TB Early Treatment participants, and only 0.10 × 10⁶ cells (SE 0.13 × 10⁶, P = 0.451) per ml BAL volume among participants at End of Treatment (TB EOT).

Figure 3

Bee swarm plots showing the effect of participant age on BAL cell yield, volume, and cell concentration. Age was binned into three categories (18–30, 30–50 and 50–80 y). (A) BAL volume yield; (B) BAL cell yield; (C) BAL cell concentration. PDF, a distance metric based on the probability density function showing distance from the mean. ANOVA on medians: (A) F = 4.2957, P = 0.0195; (B) F = 8.202, P = 0.001; (C) F = 5.2743, P = 0.004.

Figure 4

Bee swarm plots showing the relationship between BAL pellet colour, smoking status and gender.

Figure 5

Bee swarm plots showing the effect of smoking on BAL volume, cell yield and cell concentration. (A) BAL volume yield; (B) BAL cell yield; (C) BAL cell concentration. PDF, see Fig. 3. ANOVA on medians: (A) F = 6.6511, P = 0.0075; (B) F = 25.754, P < 0.001; (C) F = 31.0721, P < 0.001.

Figure 6

The effect of participant TB clinical group on BAL yields with smoking status taken into account. (A) BAL cell yield; (B) BAL cell concentration; (C) BAL volume yield. Red dots are current smokers; blue dots are prior smokers and never smokers. P values < 0.005 for interactions between groups (including both smoking categories) are shown as ****. The PLHIV group is not included in this analysis.

Figure 7

The effect of participant clinical group and smoking status on BAL cell differential counts. (A) The effect of participant smoking status on BAL cell differential counts. Orange is never-smokers; red is prior smokers; blue is current smokers. (B) The effect of participant clinical group on BAL cell differential count. Clinical groups are represented by colours, in order: orange is Community Controls; red is Household Contacts; blue is TB Pre-Treatment; forest green is TB Early treatment; olive green is TB End of Treatment; yellow is PLHIV. Figures show the relative proportions of each cell type within the total number of cells counted (which was either 100 or 200 for each sample).