This single-center prospective observational study was performed at the Center for Cardiac Intensive Care of Beijing Anzhen Hospital, Capital Medical University. A total of 450 patients were consecutively enrolled between March 2023 and February 2024, including 150 patients diagnosed with ATAAD, 150 with aortic aneurysm (AA), and 150 with unstable angina (UA). Patients with ATAAD and AA underwent Sun’s surgery or partial aortic arch replacement, whereas patients with UA underwent off-pump coronary artery bypass grafting (OPCABG).

All adult subjects were over 18 years of age, had been diagnosed with AA (chronic onset), UA and ATAAD (with a time from onset of under two weeks and had been scheduled for emergency surgery). The exclusion criteria were as follows: patients with neoplastic or chronic coagulation and inflammatory disorders; patients with missing specimens or clinical data; and patients who experienced a failed surgical intervention or died within 24 hours after surgery. Clinical data of these patients were collected during the hospitalization period. In addition, nine patients in each group were randomly selected and venous plasma was collected and subjected to enzyme-linked immunosorbent assays (ELISAs) to quantify various markers of coagulation and inflammation. We randomly selected 9 patients from each group by using a random number generator to ensure the reproducibility and validity of the findings.

All subjects underwent contrast-enhanced computed tomography (CT) to confirm the surgeons’ diagnoses of ATAAD [23]. Based on the European Society of Intensive Care Medicine (ESICM) definition [5], ARDS was confirmed via arterial blood gas analysis and chest radiography. These criteria included an acute onset, an oxygenation index (partial pressure of oxygen in arterial blood/fraction of inspired oxygen, PaO₂/FiO₂) $\le$300 mmHg for ARDS, regardless of the ventilator settings, the appearance of bilateral pulmonary infiltration on chest radiography, and the presence of respiratory failure that could not be fully explained by cardiac failure or fluid overload. Hypertension (HTN) is defined as having a systolic blood pressure (SBP) of $\ge$140 mmHg and/or a diastolic blood pressure (DBP) of $\ge$90 mmHg when no antihypertensive drugs are used. The diagnosis of diabetes is based on the recommended standards of the American Diabetes Association (ADA) in 2010. Acute kidney injury (AKI) was confirmed by the guideline published by the Kidney Disease Improving Global Outcomes (KIDGO) in 2012. Heart failure (HF) was defined as left ventricular ejection fraction $\le$40% accompanied by symptoms and/or signs. Chronic lung diseases (CLD) include chronic obstructive pulmonary disease, bronchiectasis, asthma or tuberculosis, etc. Neurological disease (ND) refer to ischemic/hemorrhagic stroke, brain tumors or brain trauma. Altered consciousness was defined as drowsiness, stupor or coma that occurs before surgery due to various reasons. Chest radiography was performed daily, and the results were interpreted by two radiologists. Blood gas levels were measured every 4 hours. The primary outcome variable was the development of ARDS within two days after surgery.

Radial artery pressure, dorsalis pedis artery pressure, and central venous pressure were monitored before surgery in all patients using established methods. Anesthesia was administered in accordance with institutional standards. Aortic surgical procedures were performed under general anesthesia, CPB, deep hypothermic circulatory arrest (DHCA), and selective cerebral perfusion were conducted using a modified elephant trunk technique (Sun’s procedure) as previously described [22].

Data related to clinical characteristics, demographic information, medical history, intraoperative variables, and details pertaining to the subsequent course of recovery in the intensive care unit (ICU) post-surgery were recorded. Each patient’s preoperative risk profile was evaluated using the European System for Cardiac Operative Risk Evaluation (EuroSCORE) [24]. The dosage of vasoactive drug administration and organ function were assessed during the first 48 h post-surgery using the vasoactive inotrope score (VIS) [25] and the Sequential Organ Failure Assessment (SOFA) scale [26], respectively.

Blood samples were collected in sodium citrate tubes through a central venous catheter at the following three time points: immediately prior to anesthesia induction; immediately after surgery; and 24 hours after surgery. Samples were centrifuged at 1550 $\times$ g for 15 min at 4 °C, and the supernatants were aliquoted and stored at –80 °C for later analysis. Serum levels of factor XII (FXII), factor XIIa (FXIIa), factor VIII-related antigen (FVIII-Ag), plasmin-antiplasmin complex (PAP), interleukin one beta (IL-1$\beta$), and tumor necrosis factor-alpha (TNF$\alpha$) were assayed using ELISA kits (USCN KIT INC, Wuhan, China).

We used the median to interpolate the missing data of continuous variables and conducted the Kolmogorov-Smirnov test. Continuous variables were expressed as the mean $\pm$ standard deviation and were analyzed using a Student’s t-test for normally distributed data; alternatively, variables with a skewed distribution were expressed as medians and quartiles and were analyzed using a Mann-Whitney U-test. Categorical variables were expressed as numbers and percentages, and a $\chi$² test or Fisher’s exact test was used to compare groups as appropriate.

The baseline characteristics of the ATAAD group were first compared with those of the AA group and UA group separately, followed by ANOVA across all three groups with Bonferroni correction for multiple comparisons. Patients with ATAAD were further stratified into ARDS and Non-ARDS subgroups, with a comparative analysis conducted on their baseline characteristics. All coagulation biomarkers were normalized and standardized using log- and z-score transformations. Logistic regression models were applied separately in the ATAAD group, AA group and UA group to assess the associations between coagulation biomarkers and postoperative ALI/ARDS, with odds ratios (ORs) and 95% confidence intervals (95% CIs) calculated. Model 1 was a univariate model and Model 2 was adjusted for traditional confounders, including sex, age, body mass index (BMI), HTN, diabetes, AKI, HF, liver dysfunction (LD), CLD, ND, smoking status, alcohol consumption, prior CS, preoperative aspartate aminotransferase (AST) level, and the preoperative oxygenation index (PaO₂/FiO₂, P/F). These covariates include medical history data and some preoperative meaningful indicators (AST and P/F) between the ARDS group and Non-ARDS group in the baseline characteristics (Supplementary Tables 1,2). Collinearity test was conducted in the multivariate analysis (Supplementary Table 3). A two-way ANOVA was used to account for both group and time as independent variables in ELISA. Spearman correlation analysis was used to analyze the associations between variables for inferring coagulation activity and the oxygenation index.

All analyses were performed using R software 4.3.1 (R Foundation, Vienna, Austria) (https://www.r-project.org/) and Prism 10.2 (GraphPad Corp, San Diego, CA, USA). p-value 0.05 indicates statistical significance (p 0.017 indicates statistical significance after Bonferroni correction).

Baseline and preoperative laboratory values of the participants are presented in Table 1. Patients in the ATAAD group were younger than those in the AA group (49.1 $\pm$ 11.5 years), whereas the patients with ATAAD exhibited a higher incidence of elevated blood pressure (84% vs 62.7%, respectively). Preoperatively, patients with ATAAD presented with significantly higher D-dimer levels than did those in either the AA group (2589 [978.5–7919] vs 180 [68.3–670.8] ng/mL; p 0.001) or the UA group (2589 [978.5–7919] vs 92.5 [56.0–157.0] ng/mL; p 0.001). The levels of FDPs in the ATAAD group increased significantly at admission (27.5 [10.9–69.3] vs 1.4 [0.5–4.6] µg/mL, p 0.001; 27.5 [10.9–69.3] vs 0.8 [0.5–1.2] µg/mL, p 0.001). However, the levels of fibrinogen (FBG) in the ATAAD group decreased significantly before surgery (2.3 [1.8–2.9] vs 2.9 [2.4–3.4] g/L, p 0.001; 2.3 [1.8–2.9] vs 3.1 [2.6–3.6] g/L, p 0.001).

The duration of the CPB procedure was significantly longer in the patients with ATAAD (179 [158–206] min) than it was in the other two groups (p 0.001 for ATAAD vs. AA and for ATAAD vs. UA), as was the aortic cross-clamping time (96.5 [86–116] min; p 0.001 for ATAAD vs. AA and for ATAAD vs. UA). Immediately after surgery, the PaO₂ values (110.0 [85.5–160.5] vs. 151.5 [141.3–206.5] mmHg, ATAAD vs. AA group, p 0.001; vs. 158.0 [120.5–205.8] mmHg, ATAAD vs UA group, p 0.001) decreased and lactic acid (Lac) level increased more significantly in the ATAAD group than in the other two groups. The activated partial thromboplastin time (APTT) at the end of surgery was longer (33.9 [31.0–40.3] s) in the patients with ATAAD than that in the other groups (p = 0.001 for ATAAD vs. AA and for ATAAD vs UA). The D-dimer level in the ATAAD group increased significantly over the 24 hours post-surgery. Similarly, the trend in the levels of FDPs was consistent with that of the D-dimer levels (Table 2). More importantly, ARDS represented 13.8% of total patients and 20.7% of patients with ATAAD.

The levels of coagulation factors in the patients with ARDS from the ATAAD group are shown in Tables 3,4. Preoperative D-dimer levels were similar between the non-ARDS and ARDS groups (2403.0 [942.0–7524.5] vs 5302 [1184.0–9419.0] ng/mL, p = 0.134). The concentration of the FDPs also tended to rise in the ARDS group before surgery (22.7 [10.2–65.9] vs 49.3 [12.5–80.8] µg/mL, p = 0.098). At the end of surgery, D-dimer levels in the ARDS group were significantly higher than those in the non-ARDS group (3326 [2270.0–5989.0] vs 2281.0 [1135.0–3787.5] ng/mL, p = 0.017). Twenty-four hours after surgery, there was no significant difference between the groups, although the D-dimer levels were significantly elevated in the ARDS group 48 hours after surgery (2017 [935.0–2955.0] vs. 935.0 [869.5–2617.0] ng/mL, p = 0.046). The trend of FDPs levels was consistent with that of the D-dimer levels. In-hospital mortality was significantly higher postoperatively in the patients with ARDS (25.8% vs. 4.2%, p = 0.001).

After all coagulation biomarkers were normalized and standardized using log- and z-score transformations, the results indicated that the levels of FDPs at the end of surgery (OR: 2.26, 95% CI: 1.13–4.54; p = 0.022) was the independent risk factor for the development of ARDS among the patients with ATAAD (Table 5).

The effects of the different groups and phases on the concentrations of FXII, FXIIa, FVIII-Ag, and PAP are presented in Fig. 1. Preoperatively (immediately prior to anesthesia induction), compared with those in the other two groups, patients with ATAAD presented with significantly higher levels of FXIIa (p = 0.014 and p 0.0001 for ATAAD vs. AA and for ATAAD vs. UA groups, respectively) and FVIII-Ag (p = 0.0002 and p 0.0001 for ATAAD vs. AA and for ATAAD vs. UA groups, respectively). After surgery, FXII, FXIIa, and FVIII-Ag levels were significantly higher in the ATAAD group compared to those in the other groups (Fig. 1a–c), whereas the perioperative PAP levels were similar between the ATAAD and AA groups (p $>$ 0.05, Fig. 1d).

Fig. 1.

The serum concentration of coagulation and inflammation biomarkers among three groups. (a) FXII levels: Intergroup comparisons across time points; (b) FXIIa levels: Intergroup comparisons across time points; (c) FVIII-Ag levels: Intergroup comparisons across time points; (d) PAP levels: Intergroup comparisons across time points; (e) IL-1$\beta$ levels: Intergroup comparisons across time points; (f) TNF$\alpha$ levels: Intergroup comparisons across time points. FXII, factor XII; FXIIa, factor XIIa; FVIII-Ag, factor VIII-related antigen; PAP, plasmin-antiplasmin complex; IL-1$\beta$, interleukin one beta, and TNF$\alpha$: tumor necrosis factor-alpha. Pre, immediately prior to anesthesia induction; post 0 h, immediately after surgery; post 24 h, 24 h after surgery. ****: p 0.0001; ***: p 0.001; **: p 0.005; *: p 0.05; ns: p $>$ 0.05.

After conducting statistical analysis on the same indicators of the same group at different time points, it was found that the FXII level at 0 h after was significantly higher than that immediately prior to anesthesia induction (p 0.001, Fig. 2a) and the TNF level immediately prior to anesthesia induction was significantly higher than that after surgery (p 0.001, Fig. 2f) in patients with ATAAD. The level of PAP at immediately prior to anesthesia induction was significantly higher than that after surgery (p 0.05, Fig. 2d) in patients with AA. The levels of other biomarkers did not show significant differences at each time point (Fig. 2).

Fig. 2.

The serum concentration of coagulation and inflammation biomarkers at different time points for the three groups. (a) FXII levels: Intragroup temporal variations; (b) FXIIa levels: Intragroup temporal variations; (c) FVIII-Ag levels: Intragroup temporal variations; (d) PAP levels: Intragroup temporal variations; (e) IL-1$\beta$ levels: Intragroup temporal variations; (f) TNF$\alpha$ levels: Intragroup temporal variations. FXII, factor XII; FXIIa, factor XIIa; FVIII-Ag, factor VIII-related antigen; PAP, plasmin-antiplasmin complex; IL-1$\beta$, interleukin one beta, and TNF$\alpha$, tumor necrosis factor-alpha. Pre, immediately prior to anesthesia induction; post 0 h, immediately after surgery; post 24 h, 24 h after surgery. ***: p 0.001; *: p 0.05; ns: p $>$ 0.05.

There was a negative correlation between FXII levels and the oxygenation index 24 h post-surgery (Pearson correlation coefficient r = –0.682, p = 0.043), whereas no significant correlations were observed for the other markers or at other time points (Fig. 3). However, a trend was observed in which the oxygenation index decreased as FVIII-Ag levels increased, suggesting that FVIII-Ag may affect lung function.

Fig. 3.

The correlation between coagulation factors and oxygenation index in patients with ATAAD. (a) Association of FXII concentration with PaO₂/FiO₂; (b) Association of FVIII-Ag concentration with PaO₂/FiO₂. PaO₂/FiO₂, oxygenation index; FXII, factor XII; FVIII-Ag, factor VIII-related antigen.

The present study, which used a control group of patients with AA, verified that ATAAD itself can significantly increase the activation of the coagulation system. ATAAD is caused by injury to the aortic intima, where blood gains access to the false lumen formed between the intima and media. The activation of the coagulation system could be induced by contact between blood and the false lumen and the release of coagulant material from the aortic wall into the systemic circulation [26, 27].

Previous studies [28, 29] have demonstrated that ATAAD and the surgical interventions to treat it can cause an increase in the levels of biomarkers that induce the activation of coagulation. The preoperative levels of FDPs and D-dimer become sharply elevated in patients with ATAAD [27, 30]. An increased D-dimer level reflects both the augmented formation of fibrin as well as the degree of subsequent fibrinolysis, and it has been shown to be associated with postoperative adverse events in patients with ATAAD [31, 32]. However, a small number of studies [33, 34] have compared the perioperative coagulation function of patients with ATAAD undergoing emergency surgery and that of patients with AA undergoing elective aortic surgery.

In the present study, intergroup comparisons were performed between patients with ATAAD and those with AA or UA; the results showed that the preoperative (immediately prior to anesthesia induction) levels of FXIIa and FVIII-Ag were significantly higher in the ATAAD group than in the other two groups, whereas the preoperative fibrinogen level was significantly lower, indicating that patients with ATAAD consumed more fibrinogen before surgery. Although patients experiencing chronic aneurysms also presented with higher D-dimer levels, the upregulation was more pronounced in patients with ATAAD. However, whether there is a causal relationship between ATAAD itself and ARDS still requires further research to verify.

In this study, the levels of FXII, FXIIa, FVIII-Ag, D-dimer, and FDPs were elevated in patients with ATAAD and AA, and the elevation was more pronounced in the former group. Meanwhile, the CPB duration was significantly prolonged in ATAAD group. We hypothesize that the coagulation disorder after CPB is related to the activation of the “contact system”.

The “contact system” is a plasma protease cascade that plays an important role in the activation of the coagulation systems; it includes FXII, FXIIa, factor XI (FXI), high-molecular-weight kininogen (HMWK), and prekallikrein, which are associated proteins in the plasma [17]. Contact system activity increases dramatically as blood comes into contact with the artificial materials comprising CPB circuits [35]. Mechanistically, FXII cleaves itself upon contact with various anionic surfaces, resulting in the production of FXIIa, which subsequently converts prekallikrein into active kallikrein. In the plasma, kallikrein generates a positive feedback loop by cleaving additional FXII, which produces more bradykinin from HMWK [17]. Previous studies have shown that active thrombin can induce the release of tissue plasminogen activator in vitro, with bradykinin potentially serving as the predominant stimulus [36, 37]. Meanwhile, the binding of soluble fibrin to CPB circuits promotes plasminogen activation [38, 39].

Previous studies have reported that coagulation disorders can be deteriorated by use of CPB during surgery, especially in conditions that promote deep hypothermia [20, 33, 40, 41]. For example, some groups have reported that CPB causes an increase in the levels of FXIIa and FDPs [17, 20]. A study conducted by Boisclair et al. [42] suggested that FXIIa can serve as a marker of contact system activation in patients undergoing CPB procedures. In addition to producing kallikrein, FXIIa has been reported to convert FXI into FXIa, thereby initiating the activation of the intrinsic pathway [20]. Therefore, the contact system exacerbates the development of coagulation system disorders when CPB is performed, and this effect may be related to the length of time required to complete CPB procedures. FXIIa, D-dimer and FDPs may serve as early biomarkers for identifying high-risk ARDS patients. We should also pay attention to the coagulation system of patients with prolonged CPB duration.

In the ATAAD group, the concentrations of D-dimer and FDPs increased immediately after surgery in the patients who developed ARDS, indicating that coagulation activity was elevated; however, the PAP concentration did not differ between the ATAAD and AA groups and was lower in patients with hypoxemia. The pathophysiology of ARDS is complex, and the underlying mechanisms have yet to be fully elucidated, especially during acute critical illness [43].

Matthay et al. [44] have speculated that one of the mechanisms underlying ARDS pathogenesis involves the activation and dysregulation of coagulation, both in the lungs and systemically. As a result, increased fibrinolytic activity may be insufficient to counteract the amplified coagulation activity in patients with ARDS. This disequilibrium may lead to fibrin deposition along the denuded alveolar basement membrane, prompting hyaline membrane formation; such lesions can, in turn, decrease lung compliance and increase inspiratory pressures [43]. In addition, the activation of procoagulant pathways may cause microvascular thromboses in the lungs; the increased amount of dead space reduces blood flow and subsequently influences gas exchange in those with ARDS [43, 45, 46]. In this study, the oxygenation index was negatively correlated with the concentration of FXII and was positively correlated with the concentration of PAP 24 hours post-surgery, manifesting as a change in coagulation activity that may be associated with ARDS. However, given the limited sample size, additional research is warranted to validate these findings.

The concentrations of the pro-inflammatory cytokines IL-1$\beta$ and TNF$\alpha$ were also significantly elevated in the ATAAD group. Correlations between ALI and inflammatory responses have been extensively investigated [22, 47]. Both local and systemic acute inflammation are prominent features of ARDS [43], and previous studies have reported that contact with the components of bypass circuits and ischemia–reperfusion injury may upregulate the expression of inflammatory cytokines, thereby leading to the release of coagulation factors [48, 49]. Meanwhile, both FXII and contact activation may induce bradykinin release and complement system activation, further promoting the systemic inflammatory response [48, 49, 50]. These findings indicated that specific interactions between coagulation and inflammation pathways contribute to lung injury.

To date, there has been little evidence to support the possibility that coagulation activation at an early stage following cardiac surgery can increase the likelihood of ARDS development, especially in patients with ATAAD. One of the strengths of this study was the detailed comparison that was performed of the relationships between coagulation disorders and postoperative ARDS between the three groups; however, there were some limitations. Firstly, the study population was derived from a single center, and it was not possible to exclude other factors that may have affected oxygenation; further validation is required with a larger cohort. Secondly, ELISAs were used to test serum samples from just nine randomly selected patients in each group. Future investigations should examine other markers of coagulation and fibrinolysis in larger cohorts. Thirdly, local coagulation and inflammatory activity are also important, as they can influence the occurrence of hypoxemia; further analysis of the coagulation biomarkers will be performed based on bronchoalveolar lavage fluid.