Lung Fluid Biomarkers for Acute Respiratory Distress Syndrome

A Systematic Review and Meta-analysis

Yishan Wang; Huijuan Wang; Chunfang Zhang; Chao Zhang; Huqin Yang; Ruiyue Gao; Zhaohui Tong


Crit Care. 2019;23(43) 

In This Article


In this systematic review and meta-analysis, we summarized the biomarkers related to ARDS diagnosis in the at-risk population and those related to ARDS mortality.

By searching several databases and screening for related articles, 49 studies involving 2189 patients were identified.

We discovered that total protein, albumin, PAI-1, PAF-ACH, sTNFα-RII, HGF, IL-8, PCP I, PCP III, soluble receptor for advanced glycation end products (SRAGE), and IL-6 were significantly increased in the lung fluid of patients with ARDS. Although total PLA2 activity, soluble intercellular adhesion molecule-1 (SICAM-1), IL-2, CC16, and MMP-9 were also significantly different between patients with ARDS and at-risk patients, few studies were included on each of these biomarkers, so the results are unreliable.

IL-1β, IL-6, IL-8, TNF-α, PCPIII, and total protein were significantly increased in lung fluid of patients who died in the ARDS cohort. Furthermore, few studies for KL-6, plasminogen activator inhibitor-1 (PAI-1), IL-2, and CC16 were included, although these biomarkers were significantly different between survivors and non-survivors.

Pioneering work by Terpstra et al. reported on plasma biomarkers for ARDS diagnosis and prognosis. They reported that KL-6, lactate dehydrogenase (LDH), SRAGE, von Willebrand factor (vWF), and IL-8 displayed the highest effect size for ARDS diagnosis, and interleukin-4 (IL-4), IL-2, Angiopoietin-2 (Ang-2), and KL-6 had the highest effect size for ARDS prognosis (assessed by pooled odds ratio). These biomarkers represent pathophysiological processes, which led to the hypothesis that ARDS diagnosis is correlated with tissue damage, whereas ARDS mortality is correlated with systemic inflammation (4). In our meta-analysis, biomarkers for ARDS diagnosis were related to inflammation (IL-8 and IL-6), endothelial injury (SICAM), epithelial injury (SRAGE and HGF), lung fibroproliferation (PCPI and PCPIII), and coagulopathy (PAF-ACH). With regard to ARDS mortality, biomarkers related to inflammation (IL-8, IL-6, and IL-1β), epithelial injury (KL-6), and lung fibroproliferation (PCPIII) were elevated in the lung fluid of patients with ARDS presenting the worst outcomes. Therefore, we assume that, aside from tissue damage and systemic inflammation, lung fibroproliferation is vital in both ARDS diagnosis and prognosis. Several studies on lung biopsy of patients with ARDS showed a strong relationship between fibrosis activity and ARDS mortality.[64,65]

To our knowledge, this is the first meta-analysis of biomarkers in lung fluid for ARDS. Since lower respiratory tract specimens are now easily obtained, BALF and other lung fluids have become frequently used clinical samples for pulmonary disease diagnosis, secondary to plasma/serum. Since the most intensive physiological processes in ARDS occur in the lung, theoretically, lung fluid can reflect the pathophysiological process differently from other body fluids.

In this analysis, we applied RoM to assess the effect size and to attempt to eliminate the bias caused by dilution of different kinds of lung fluid. For example, edema fluid was completely undiluted, whereas BALF might be quite diluted. This methodology has been proved to be robust and is widely used.[7–9,66,67]

This systematic review and meta-analysis may prompt further research on ARDS diagnosis and prognosis in many different ways. First, it demonstrates the research priorities and indicates that research on ARDS biomarkers in lung fluid and other compartments is needed. Second, it establishes ARDS biomarkers and their performance in lung fluid as an innovative research field. Third, it identifies numerous novel translational approaches for biomarker measurement in different compartments, such as chromatography for metabolomics separation and mass spectrometry or nuclear magnetic resonance spectroscopy for biomarker detection.[68] Although some non-quantitative methods were not included in this meta-analysis, they are still worth exploring.

There were limitations in this meta-analysis as well. First, although we performed a subgroup analysis of the biomarkers related to ARDS diagnosis and mortality, heterogeneity was not explainable for every biomarker. We assume that this could be related to the different etiologies of ARDS, variation in BALF procedures between studies, multiple control types used in the studies, wide range of intervals between study inclusion and biomarker measurement, and different treatments for ARDS. We were not able to conduct further analysis due to the limited information.

Second, a limited number of studies were included for each biomarker, which impedes the reproducibility of the results. The number of studies for each biomarker should be taken into consideration while assessing the performance in the ranking system.

Third, only the biomarkers addressed by two or more studies were included. Because of this, promising biomarkers evaluated in a single study were not considered, which may limit the view on lung fluid biomarker research as a whole.

Finally, the use of lung fluid as a study object itself had some limitations. Due to the lack of information on specific BALF procedures, we could only assume the recovery rates between subgroups in one study were within an acceptable range, which may have caused some of the heterogeneity between studies.