Biomarker profiles of coagulopathy and alveolar epithelial injury in acute respiratory distress syndrome with idiopathic/immune-related disease or common direct risk factors

The ARDS without indirect risk factors was diagnosed according to the Berlin definition with the following criteria: within 1 week of new or worsening respiratory symptoms, bilateral lung opacities were found on chest radiography, and the PaO₂/F_IO₂ ratio was ≤ 300 mmHg with a positive end-expiratory pressure of ≥ 5 cmH₂O. Additionally, no cardiac failure or fluid overload and no common indirect risk factors for ARDS, such as non-pulmonary sepsis, major trauma, or pancreatitis could be found [2]. Direct lung injury risk factors were defined as pneumonia, aspiration of gastric contents, pulmonary contusion, inhalation injury, and near drowning, based on the Berlin definitions. Patients with vasculitis were classified as having ARDS without common risk factors because vasculitis is not pathologically characterized by diffuse alveolar damage (DAD). The diagnosis of pneumonia was based on Infectious Diseases Society of America/American Thoracic Society consensus guidelines combined with clinical data and microbiological diagnostic testing (including a blood culture, sputum culture, or culture of endotracheal aspirate, and a urinary antigen test for Streptococcus pneumoniae and Legionella pneumophila) [15, 16]. Bronchoalveolar lavage (BAL) fluid for Gram staining and culture, direct fluorescence assay for Pneumocystis jirovecii, and a rapid influenza A/B diagnostic test (immunochromatographic assays for specific influenza viral antigens) were also performed, as needed.

ARDS without common risk factors were separated into four etiological groups, as described below [7]. Idiopathic ARDS was defined as the absence of any ARDS etiology including common risk factors despite a comprehensive diagnostic work-up, or acute presentation of idiopathic interstitial pneumonia [17]. Immune-related ARDS was defined as an acute presentation of CTD-ILD as defined in accordance with established CTD criteria (e.g., American College of Rheumatology criteria [18]) during hospitalization, or hypersensitive pneumonitis [19]. Malignancy-associated ARDS was defined as requiring cytological or pathological evidence of hematological or solid malignancy. Drug-induced ARDS was defined as previous exposure to a drug that is known to be a pneumonia inducer in the absence of any other risk factor for ARDS [20].

Descriptive data (including demographic, diagnostic, clinical, and laboratory data) were collected from the electronic medical records of all eligible patients. Initial severity indices, including the Acute Physiology and Chronic Health Evaluation (APACHE) II and Simplified Acute Physiology Score (SAPS) II, were calculated on the day of ICU admission [21, 22]. Sequential Organ Failure Assessment (SOFA) scores were calculated during the first 7 days [23]. Clinical outcomes were assessed according to ICU days, ventilator-free days, and all-cause 28- and 90-day mortality. For the patients with idiopathic and immune-related ARDS, BAL fluid cytological analysis and autoimmunity tests were extracted from the medical charts when available.

At our institute, the biomarkers of coagulation and type II pneumonocytes are routinely measured for the patients who are admitted to the ICU with respiratory failure and/or with suspected sepsis. Plasma biomarkers were measured at the time of ICU admission (ICU day 1) and on ICU days 2–5. Coagulation and fibrinolytic markers included global markers (platelet count, immature platelet fraction, prothrombin time–international normalized ratio [PT-INR], fibrin degradation product [FDP]), markers of thrombin generation (TAT), markers of anticoagulant activity (PC activity), and markers of fibrinolytic activity (plasmin–α₂-plasmin inhibitor complex [PIC], PAI-1).

Global markers were assayed using an XE-5000 hematology analyzer (Sysmex, Kobe, Japan) and a CS-2100i automatic coagulation analyzer (Sysmex). Berichrom assays (Siemens Healthcare Diagnostics, Tokyo, Japan) were used to assay PC activity. The TAT/PIC test F enzyme immunoassay (Sysmex) was used to measure TAT and PIC levels. The PAI-1 was measured using the tPAI test (Mitsubishi Chemical Medience, Tokyo, Japan).

Surfactant protein (SP)-D, KL-6, C-reactive protein (CRP), and procalcitonin (PCT) were measured using the SP-D kit enzyme immunoassay (Yamasa, Chiba, Japan), Presto II KL-6 chemiluminescent enzyme immunoassay (Sekisui Medical, Tokyo, Japan), CRP-HG latex immunoassay (Eiken Kagaku, Tokyo, Japan), and Brahms PCT chemiluminescent enzyme immunoassay (Roche Diagnostic, Tokyo, Japan), respectively.

Differences in clinical characteristics and laboratory data among the groups were analyzed using the χ² test or Fisher’s exact test for categorical variables and the Wilcoxon rank-sum test or Kruskal–Wallis test with/without Steel–Dwass pairwise comparisons for continuous variables, as appropriate. Changes in the biomarker concentrations over time in the groups were compared with multiple analysis of variance. A multivariate logistic regression model based on a forward stepwise method was used to identify the best combination of coagulation biomarkers to diagnose iARDS. Receiver operating characteristic (ROC) curve analysis was performed to calculate the area under the receiver operating characteristic curve (AUC) of the biomarkers at day 1 to evaluate the discriminative capacity between the two groups. All P values were two-tailed, and P < 0.05 was considered to indicate statistical significance. Data were analyzed using JMP version 12 (SAS Institute, Tokyo, Japan).

Among the 97 patients with pulmonary ARDS, 56 had been exposed to direct lung injury risk factors and 41 had not been exposed to any of the common risk factors. The direct risk factors of lung injury included pneumonia (42; 75.0%), aspiration (13; 23.2%), and drowning (1; 1.8%). The 41 ARDS patients without common risk factors were classified as idiopathic (17; 41.5%), immune-related (14; 34.1%), malignancy-associated (7; 17.1%), and drug-induced (3; 7.3%).

Table 1 shows the baseline characteristics and outcomes of the study patients with iARDS and dARDS and those with unilateral pneumonia. Patients with dARDS were more severely ill, with higher APACHE II, SAPS II, and SOFA scores on ICU admission compared with patients with iARDS. The PaO₂/F_IO₂ ratio on admission and the severity of ARDS, however, were not different between patients with dARDS and those with iARDS. Ventilator-free days, length of ICU stay, and mortality were also similar for the two groups.

The distribution of pathogens in patients with dARDS and those with pneumonia are shown in Additional file 1: Table S1. In patients with dARDS, the most common causative microorganisms were Klebsiella pneumoniae (17.9%), followed by Streptococcus pneumoniae (12.5%) and methicillin-susceptible Staphylococcus aureus (10.7%).

Among the 31 patients with iARDS, 17 (54.8%) were diagnosed with idiopathic ARDS and 14 (45.2%) with immune-related ARDS, which included the following: rheumatoid arthritis (n = 5), dermatomyositis (n = 3), systemic lupus erythematosus (n = 2), scleroderma (n = 1), microscopic polyangiitis (n = 1), granulomatosis with polyangiitis (n = 1), and hypersensitivity pneumonitis (n = 1). Table 2 shows the BAL findings and autoantibodies in patients with iARDS. In about half of these patients (idiopathic, 62.5%; immune-related, 42.9%), neutrophils and lymphocytes were both elevated in BAL fluid, showing a mixed cellular pattern (defined as neutrophil > 3% and lymphocyte > 15% on BAL differential cell counts). Antinuclear antibody was positive (with > 1:160 titers) in 64.3%, and anticyclic citrullinated peptide antibody was positive in 35.7% of the patients with immune-related ARDS. Notably, 23.5% of the patients with idiopathic ARDS were positive for autoantibodies against aminoacyl-tRNA synthetase.

Figure 2 shows the results of multiple comparisons for coagulation biomarkers on day 1 in patients with iARDS, dARDS, and unilateral pneumonia. The changes in coagulation biomarkers over time among the three groups are shown in Fig. 3. In the iARDS and dARDS patients, the PT-INR and FDP were increased to similar levels on day 1: INR 1.28 (1.21–1.42) vs. 1.40 (1.22–1.60), P = 0.19; FDP 18.0 (13.7–29.7) vs. 21.2 (11.7–31.6) mg/mL, P = 0.99, respectively (Fig. 2), and during the observational period (overall difference: INR, P = 0.41; FDP, P = 0.36; Fig. 3), suggesting a hypercoagulable state in both groups. The TAT levels were increased in the three groups, but those levels were much lower in iARDS patients compared with dARDS patients on day 1 (9.8 [5.9–13.5] vs. 18.2 [9.4–40.5] ng/mL, P = 0.0046) and during the observational period (overall difference, P = 0.0029), and they were similar to those in patients with unilateral pneumonia. PAI-1 levels were significantly higher (251 [95.8–629] vs. 38.0 [17.5–51.0] ng/mL, P < 0.0001) and the PIC levels significantly lower (1.2 [0.75–1.7] vs. 1.7 [1.3–2.3] mg/mL, P = 0.014) on day 1 in patients with dARDS than in those with iARDS, suggesting inhibited fibrinolysis in the dARDS patients. Further, PC activity (marker of the anticoagulant system) was significantly lower in patients with dARDS compared with those with iARDS (42.9 [30.9–63.6] vs. 76.2 [66.1–95.2] %, P < 0.0001). When we compared the PC activities in patients with iARDS and unilateral pneumonia, those values were still higher in patients with iARDS. Among the coagulation biomarkers, a multivariate stepwise logistic regression analysis revealed that PAI-1 and PC activity comprised the best combination with which to identify patients with iARDS.

PCT levels (marker of infection) on day 1 were increased in the dARDS and pneumonia patients but were lower than the reference value for infection in patients with iARDS. However, levels of CRP, a widely used marker of inflammation and mechanistically downstream of IL-6, were not different among the three groups. The markers of type II pneumocyte injury, SP-D, and KL-6 were markedly increased in patients with iARDS compared with those with dARDS or pneumonia (Figs. 4 and 5).

To compare the abilities of the plasma biomarkers to distinguish between ARDS subtypes, we conducted a ROC curve analysis to calculate the AUCs of biomarkers for coagulation, infection, and pneumocytes (Table 3). We found that the AUCs for discriminating between iARDS and dARDS were high (> 0.8) for PAI-1, PC activity, PCT, SP-D, and KL-6. These AUC results showed that the ability to distinguish between iARDS and dARDS was comparable among those five biomarkers (PC, 0.86 [0.76–0.93], P = 0.33; PAI-1, 0.89 [0.74–0.96], P = 0.95; PCT, 0.89 [0.79–0.96], P = 0.66; and SP-D, 0.88 [0.77–0.94], P = 0.16; vs. KL-6, 0.90 [0.79–0.95], respectively).

There were some potential limitations to our study. First, this was a retrospective, observational study conducted at a single center with a relatively small population. A large validation study is needed to confirm our results. Second, we could not perform serological tests for non-influenza respiratory viruses, such as the respiratory syncytial virus or human metapneumovirus. Although we ruled out the common ARDS risk factors and known causes of interstitial pneumonia (e.g., drugs, environmental agents, CTDs) to diagnose idiopathic ARDS, we could not completely exclude the possibility of viral infections or environmental antigen exposures, which could subside spontaneously. Third, we could perform BAL for only about half of iARDS patients, which may not be generalizable to the whole population. Finally, we did not measure the biomarkers in the BAL fluid. Although systemic markers are easier to obtain and the BAL procedure may not always be possible because of the risk of respiratory and hemodynamic complications, the biomarkers in the BAL fluid would more specifically reflect the regional pathophysiology in the alveoli. Further studies are needed to evaluate the pathogenic processes of these biomarkers from the pulmonary compartment to the circulation.