Plasma sRAGE is independently associated with increased mortality in ARDS: a meta-analysis of individual patient data

Abstract

Purpose: The soluble receptor for advanced glycation end-products (sRAGE) is a marker of lung epithelial injury and alveolar fluid clearance (AFC), with promising values for assessing prognosis and lung injury severity in acute respiratory distress syndrome (ARDS). Because AFC is impaired in most patients with ARDS and is associated with higher mortality, we hypothesized that baseline plasma sRAGE would predict mortality, independently of two key mediators of ventilator-induced lung injury.

Methods: We conducted a meta-analysis of individual data from 746 patients enrolled in eight prospective randomized and observational studies in which plasma sRAGE was measured in ARDS articles published through March 2016. The primary outcome was 90-day mortality. Using multivariate and mediation analyses, we tested the association between baseline plasma sRAGE and mortality, independently of driving pressure and tidal volume.

Results: Higher baseline plasma sRAGE [odds ratio (OR) for each one-log increment, 1.18; 95% confidence interval (CI) 1.01-1.38; P = 0.04], driving pressure (OR for each one-point increment, 1.04; 95% CI 1.02-1.07; P = 0.002), and tidal volume (OR for each one-log increment, 1.98; 95% CI 1.07-3.64; P = 0.03) were independently associated with higher 90-day mortality in multivariate analysis. Baseline plasma sRAGE mediated a small fraction of the effect of higher ΔP on mortality but not that of higher V_T.

Conclusions: Higher baseline plasma sRAGE was associated with higher 90-day mortality in patients with ARDS, independently of driving pressure and tidal volume, thus reinforcing the likely contribution of alveolar epithelial injury as an important prognostic factor in ARDS. Registration: PROSPERO (ID: CRD42018100241).

Keywords: Acute respiratory distress syndrome; Biomarker; Lung epithelial injury; Prognosis; Receptor for advanced glycation end-products.

Fig. 1

Flow diagram of study selection

Fig. 2

Forest plot of odds ratios for death at day 90 after multivariate logistic regression in patients with acute respiratory distress syndrome (n = 700). *A risk score was calculated as a composite of available severity scores (SOFA, APACHE II, APACHE III) combined using an average z score. Study effect was taken into account as a random effect covariate. Plasma levels of sRAGE (in pg.ml⁻¹), PaO₂/FiO₂, tidal volume, and PEEP are natural log-transformed in the model to meet the assumption of linearity with log odds of outcome; the ORs presented here are for each log increase in the level of plasma sRAGE, PaO₂/FiO₂, tidal volume, and PEEP. APACHE II Acute Physiology and Chronic Health Evaluation II Score, APACHE III Acute Physiology and Chronic Health Evaluation III Score, SOFA Sequential Organ Failure Assessment Score, ΔP driving pressure, PEEP positive end-expiratory pressure, sRAGE soluble receptor for advanced glycation end-products

Fig. 3

Mediation analysis of 90-day mortality in patients with acute respiratory distress syndrome. Tested mediator: changes in baseline plasma sRAGE. Independent variable: changes in baseline ΔP. Top: the first step in our mediational analysis was the demonstration that higher ΔP had a measurable impact on mortality after accounting for baseline risk covariates. Middle: second, we checked if mediator changes correlated with higher mortality, after accounting for baseline risk covariates. Bottom: finally, a multilinear regression (mixed effects) calculated the influence of higher ΔP on the tested mediator (baseline plasma sRAGE). Subsequently, we jointly calculated the influence of the mediator on 90-day mortality, after accounting for baseline risk covariates, and the direct effects of the independent variable (higher ΔP). This last step shows that higher plasma sRAGE partially mediates [9%, P = 0.04 for the average causal mediation effect (ACME)] the original effect of baseline ΔP on mortality and, consequently, baseline ΔP remains directly associated with mortality in an independent manner (characterizing incomplete mediation). Mediator and independent variables are assessed as continuous variables. Plasma levels of sRAGE (in pg.ml⁻¹), PaO₂/FiO₂, tidal volume, and PEEP are natural log-transformed in the model to meet assumption of linearity with log odds of outcome. ARDS acute respiratory distress syndrome, sRAGE soluble receptor for advanced glycation end-products, ΔP driving pressure, PEEP positive end-expiratory pressure

Fig. 4

Mediation analysis of 90-day mortality in patients with acute respiratory distress syndrome. Tested mediator: changes in baseline plasma sRAGE. Independent variable: changes in tidal volume. Top: the first step in our mediational analysis was the demonstration that higher tidal volume had a measurable impact on mortality, after accounting for baseline risk covariates. Middle: second, we checked if mediator changes (higher baseline plasma sRAGE) correlated with higher mortality after accounting for baseline risk covariates. Bottom: finally, a multilinear regression (mixed effects) calculated the influence of higher tidal volume on the tested mediator (baseline plasma sRAGE). Subsequently, we jointly calculated the influence of the mediator on 90-day mortality, after accounting for baseline risk covariates, and the direct effects of the independent variable (higher tidal volume). This last step shows that higher plasma sRAGE does not significantly mediate [P = 0.5 for the average causal mediation effect (ACME)] the original effect of higher tidal volume, and, consequently, higher tidal volumes remain directly associated with mortality in an independent manner (characterizing lack of mediation). Mediator and independent variables are assessed as continuous variables. Plasma levels of sRAGE (in pg.ml⁻¹), PaO₂/FiO₂, tidal volume, and PEEP are natural log-transformed in the model to meet the assumption of linearity with log odds of outcome. PBW predicted body weight, ARDS acute respiratory distress syndrome, sRAGE soluble receptor for advanced glycation end-products, PEEP positive end-expiratory pressure