Bronchoalveolar bile acid and inflammatory markers to identify high-risk lung transplant recipients with reflux and microaspiration

Abstract

Background: Gastroesophageal reflux disease (GERD) is a risk factor for chronic lung allograft dysfunction. Bile acids-putative markers of gastric microaspiration-and inflammatory proteins in the bronchoalveolar lavage (BAL) have been associated with chronic lung allograft dysfunction, but their relationship with GERD remains unclear. Although GERD is thought to drive chronic microaspiration, the selection of patients for anti-reflux surgery lacks precision. This multicenter study aimed to test the association of BAL bile acids with GERD, lung inflammation, allograft function, and anti-reflux surgery.

Methods: We analyzed BAL obtained during the first post-transplant year from a retrospective cohort of patients with and without GERD, as well as BAL obtained before and after Nissen fundoplication anti-reflux surgery from a separate cohort. Levels of taurocholic acid (TCA), glycocholic acid, and cholic acid were measured using mass spectrometry. Protein markers of inflammation and injury were measured using multiplex assay and enzyme-linked immunosorbent assay.

Results: At 3 months after transplantation, TCA, IL-1β, IL-12p70, and CCL5 were higher in the BAL of patients with GERD than in that of no-GERD controls. Elevated TCA and glycocholic acid were associated with concurrent acute lung allograft dysfunction and inflammatory proteins. The BAL obtained after anti-reflux surgery contained reduced TCA and inflammatory proteins compared with that obtained before anti-reflux surgery.

Conclusions: Targeted monitoring of TCA and selected inflammatory proteins may be useful in lung transplant recipients with suspected reflux and microaspiration to support diagnosis and guide therapy. Patients with elevated biomarker levels may benefit most from anti-reflux surgery to reduce microaspiration and allograft inflammation.

Keywords: bile acid; biomarkers; gastroesophageal reflux disease; inflammation; lung transplantation.

Figure 1:. Study flow diagram for the GERD cohort

We started with a total of 285 adults who underwent bilateral or single lung transplantation at Toronto General Hospital between 2010 and 2015 and had available post-transplant 24-hour pH/impedance reflux study. GERD group: Of the 285 patients, 48 had greater or equal to 48 reflux episodes. We excluded 23 patients due to not having a reflux study within 365 days post-transplant or no BAL samples available, to get a final of 25 patients. No-GERD group: We reviewed patients with less than 48 reflux episodes, starting from those with the lowest number of episodes, until we reached 51 patients for an approximate 2:1 ratio with the GERD group. We excluded patients if they did not have a reflux study within 365 days post-transplant, reported symptoms of reflux during the study, had underlying disease commonly associated with GERD (e.g. scleroderma, hiatal hernia, Barrett’s esophagus, history of Nissen fundoplication or pyloroplasty), had no BAL samples available, and for other reasons listed. 5 patients with single lung transplants were excluded in order to match for type of transplant. In total, 98 patients, corresponding to those with greater than 23 reflux episodes, remained unmatched and were excluded.

Figure 2:. Comparison of bile acid and protein levels in GERD vs. no-GERD patients at 3 months post-transplant

Y-axes represents nM for bile acids and pg/mL for proteins. Horizontal bars represent median and 95% CI. Star indicates Wilcoxon-Mann-Whitney test p<0.05.

Figure 3:. Longitudinal trend of bile acids and proteins in BALs obtained at around 3, 6, 9, and 12 months post-transplant from GERD vs. no-GERD patients

Graphs show log-transformed bile acid (nM) and protein (pg/mL) levels on Y-axes. X-axes represent time (3, 6, 9, 12 months post-transplant). Bolded lines represent LOESS (local estimated scatterplot smoothing) smooth lines with 95% CI in shaded region. Random intercepts were used for all models, but a random time effect was only used when it improved the model significantly, which was the case for TCA, IL-1α, IL-12p70, CCL2, and S100A8. P-values and other model outputs are shown in Table S2.

Figure 4:. Comparison of BAL bile acid and protein levels before and after Nissen fundoplication

Paired BAL protein and bile acid levels pre- and post-Nissen fundoplication in lung transplant recipients are shown. Y-axes represents pg/mL for proteins and nM for bile acids. Star indicates p-value<0.05 based on the multivariate GEE analysis adjusted for rejection, infection, rejection treatment and infection treatment. P-values and other model outputs are shown in Table S3.