Introduction

Despite significant advances in pharmacological and supportive care of critically ill patients, multiple organ failure (MOF) is still one of the most common causes of mortality in surgical hospitals, especially in intensive care units (ICU) [1]. Early diagnosis and prognostic evaluation are crucial for effective treatment of these patients [2, 3].

Preoperative identification of patients with high risk of organ failure is necessary for correct management, treatment, prevention of clinical deterioration and reduction of mortality. In this regard, there is a need to search for highly specific, sensitive and easy-to-use MOF marker that will allow ICU specialists to diagnose this unfavorable event at the early stage [1, 4].

In recent years, more and more studies have been devoted to analysis of such serological markers as cytokines, procalcitonin, C-reactive protein and triggering receptor expressed on myeloid cells 1 (TREM-1) in ICU patients with high risk of mortality [5, 6]. Today, TREM-1 is one of the key markers of inflammation [7, 8]. Its soluble form (sTREM-1) is activated both in infectious and non-infectious conditions, promotes release of other inflammatory mediators and plays an important role in innate immune responses [9].

In this study, we analyzed advisability of sTREM-1 as a biomarker for predicting the risk of MOF after cardiac surgery.

The purpose of the study was to analyze significance of serum sTREM-1 in objective assessment of the patient's condition and predicting MOF after coronary artery bypass grafting.

Material and methods

General characteristics of patients

The study was based on the coronary artery bypass grafting registry of the Research Institute for Complex Issues of Cardiovascular Diseases. There were 132 patients (24 women and 108 men) aged 47-74 years (mean 62 years) who underwent elective on-pump coronary artery bypass surgery for coronary artery disease (CAD).

Inclusion criteria: chronic CAD; 2) elective coronary artery bypass surgery; 3) age >18 years; 4) informed consent.

We excluded all patients with malignancies, autoimmune, mental and infectious diseases, combined surgeries and complications associated with surgical strategy. The local ethics committee approved the study.

Dividing into groups and their characteristics

All patients underwent coronary artery bypass grafting with standardized blood warm cardioplegia and non-pulsatile perfusion. CPB and aortic cross-clamping time was 96 (79–115) and 61 (50–75) min, respectively. There were 3 (2–4) infusions of cardioplegic solution, and revascularization index was 3 (2–3).

All patients were retrospectively divided into 2 groups depending on MOF in early postoperative period. We used the SOFA scale (Sepsis-Related Organ Failure Assessment) to estimate severity of organ failure in early postoperative period [10, 11].

Thus, the control group included 102 (77.3%) patients with uneventful early postoperative period (intact function of all organs and systems) and patients with complicated early postoperative period and no obvious signs of MOF (dysfunction of 1-2 systems with fast compensation). SOFA score in patients without MOF ranged from 0 to 4 points. The main group consisted of 30 (22.7%) patients with combined dysfunction of two or more systems in early postoperative period and further progression of disorders. In these patients, SOFA score was ≥ 4 points. Comparative characteristics of study groups are presented in Table.

Characteristics of study groups

Note. CPB — cardiopulmonary bypass; ICU — intensive care unit.

Data collection

Study material was blood from the peripheral vein (9 ml) sampled on an empty stomach in sterile Vacuette serum clot activator tubes (Greiner bio-one, Austria). Blood was sampled before surgery (point 1) and 18–20 hours after surgery in the ICU (point 2). Centrifugation at 1500 rpm for 15 min was followed by serum aliquoting into labeled Eppendorf tubes. Serum samples were stored in a freezer at -70°C. We defrosted aliquot with serum on the day of the study. Biological material was not re-frozen.

Immunological study

Serum concentration of sTREM-1 was analyzed by enzyme-linked immunosorbent assay using commercial Human TREM-1 kits (R&D Systems, USA) intended for scientific research in accordance with the manufacturer's instructions. We analyzed concentration of sTREM-1 using the Multiska semi-automatic spectrophotometer (Thermo Fisher Scientific, USA).

Statistical analysis

GraphPad Prism 8.0 software was used for statistical analysis. The Kolmogorov–Smirnov test was used to assess distribution normality. Between-group comparison was performed using analysis of variance (ANOVA). Significance of differences between two independent groups was assessed using the Mann-Whitney U test. Within-group differences were analyzed by using of the Wilcoxon W test. Results are presented as median (Me) and interquartile range (25Q; 75Q). Differences were significant at p-value <0.05. We searched for prognostic predictors using ROC analysis. The area under ROC curve (AUC) was used to assess predictive accuracy. Correlation of non-parametric data was analyzed using linear regression.

Results

Study population

Between-group analysis revealed similar course of CAD in both groups (p>0.05). However, chronic obstructive pulmonary disease was more common among patients with MOF in early postoperative period compared to patients without organ dysfunction (4 vs. 1 patient, respectively, p=0.01). Intraoperative variables were similar. However, high SOFA score and severe organ dysfunction on the 1^st postoperative day were observed in patients with MOF that increased ICU-stay.

Structure of multiple organ failure

Renal failure prevailed (77%). Damage to the central nervous system was observed in 47% of cases, gastrointestinal abnormalities — in 30% of cases. Mortality rate in patients with early postoperative MOF was 50% (Fig. 1).

Fig. 1. Incidence of damage to various organs in patients with multiple organ failure.

Enzyme-linked immunosorbent assay

Perioperative changes in sTREM-1 concentration are shown in Fig. 2. We observed significant between-group (at two time points) and within-group changes in sTREM-1 concentration.

Fig. 2. Serum sTREM-1.

Preoperative sTREM-1 level differed in patients with or without MOF in early postoperative period (307.5 (276.2–452.7) vs. 155.3 (131.2 -200.6) pg/ml, p<0.0001). Serum sTREM-1 increased on the 1^st postoperative day in all patients (p<0.0001). Nevertheless, this value was higher in patients with MOF (655.3 (556.4 — 782.2) vs. 238.9 (194.4 — 326.1) pg/ml.

Regression analysis

Regression analysis revealed moderate positive correlation between sTREM-1 concentration and SOFA score on the 1^st day after surgery (r=0.39; p≤0.0001). There was no similar correlation at the preoperative stage (r=0.03; p=0.75).

ROC analysis

We analyzed sTREM-1 as a predictor of MOF after cardiac surgery using ROC curve (Fig. 3). AUC at the preoperative stage was 0.884 (95% CI 0.825–0.943, standard error 0.030) (Fig. 3a), on the 1^st day after surgery — 0.913 (95% CI 0.860–0.966, standard error 0.027) (Fig. 3b). Thus, we can use sTREM-1 as a marker of MOF.

Fig. 3. ROC curves of MOF prediction models.

a — preoperative stage; b — the first postoperative day.

Discussion

Bouchon A. et al. [12] described TREM-1 in 2000 [12]. This type of receptor is expressed on the surface of neutrophils, mature monocytes, macrophages and non-myeloid cells such as epithelial and endothelial cells [13]. Hypoxia, dysregulation of transcription factor (NF-kB), bacterial and fungal components increase TREM-1 expression. The next events are immediate stimulation of all effector mechanisms with release of pro-inflammatory cytokines and chemokines and degranulation with oxidative burst of neutrophils [7, 14]. Thus, extracellular receptor domain (sTREM-1) is released from the surface of myeloid cells by shedding into body fluids, and we can quantify this marker. Therefore, sTREM-1 is an essential component of innate immune and inflammatory responses [8, 9, 15].

According to literature data, sTREM-1 is involved in development of infectious (viral, bacterial and fungal) and non-infectious diseases. Many authors registered high serum concentrations of sTREM-1 in patients with various infectious diseases. Indeed, sTREM-1 has prognostic significance in meningitis [16], hemorrhagic fever [17], pneumonia [18] and other lung infections [19–21]. Many authors demonstrated usefulness of sTREM-1 in diagnosis of sepsis. In these patients, this receptor is considered a potential biomarker. Oku R. et al. [22] and Rivera-Chavez F.A. et al. [23] revealed significantly higher serum sTREM-1 in patients with sepsis compared to systemic inflammatory response syndrome (SIRS). In contrast to these results, Jedynak M. et al. [24] consider sTREM-1 to be inflammatory mediator not associated with infection. They found higher concentration of sTREM-1 in patients with sepsis than in patients with SIRS, but these differences were not significant (p = 0.06). Nasr El-Din A. et al. [25] found higher sTREM-1 in patients with sepsis compared to SIRS without infection at admission (p<0.0001) and after 7 days (p<0.0001). In this study, as well as in ours, concentration of sTREM-1 increased in the following days in both groups. However, the authors found no significant differences in mean concentrations of sTREM-1 between patients with culture-confirmed septicemia on the 1^st day (p=0.98) and between both groups after 7 days (p=0.17). Thus, we can suppose sTREM-1 as an active participant of inflammation regardless of the presence of infectious agents.

Despite obvious evidence for involvement of sTREM-1 in pathogenesis of infectious diseases, we were interested in the studies describing sTREM-1 in non-infectious inflammation. For example, Essa E.S. et al. [26] found significantly higher concentrations of sTREM-1 in patients with chronic kidney disease and hemodialysis. Soluble TPEM-1 was associated with gout [27], neutropenia after chemotherapy [28], bronchiectasis [29] and rheumatoid arthritis [30]. Moreover, high serum sTREM-1 is found in patients with chronic obstructive pulmonary disease [31]. Serum sTREM-1 is elevated in patients with inflammatory bowel disease, but its correlation with severity of inflammation is unclear [32, 33]. Hermus L. et al. [34] observed high serum sTREM-1 in patients with coronary artery disease and peripheral artery disease. These authors demonstrate the key role of sTREM-1 in pathogenesis of atherosclerosis. Other studies revealed high concentrations of sTREM-1 in patients with peripheral artery disease [35] and restenosis one year after coronary artery bypass surgery [36]. In 2007, Adib-Conquy M. et al. [37] found higher serum sTREM-1 in patients with non-specific activators of inflammatory response including on-pump surgery, blood loss and transfusion, as well as intensive care for cardiac arrest. Thus, we can consider sTREM-1 not only as diagnostic marker of microbial infections, but also as a marker for assessing clinical severity in other inflammatory and autoimmune diseases.

In our study, preoperative serum sTREM-1 was 2 times higher in patients with early postoperative MOF than in patients with uneventful postoperative period. Thus, high preoperative sTREM-1 can predict early postoperative complications and unfavorable clinical outcomes.

Higher serum sTREM-1 prior to surgery is associated with higher risk of MOF in early postoperative period. However, mean serum sTREM-1 in preoperative period was not low in patients without clinical signs of MOF (155.3 (131.2–200.6) pg/ml). This finding is explained by important role of sTREM-1 in pathogenesis of atherosclerosis per se and other comorbidities in patients scheduled for elective coronary artery bypass surgery (chronic obstructive pulmonary disease and chronic kidney disease).

We found that concentration of sTREM-1 increases on the 1^st day after surgery regardless of multiple organ failure. However, these values increased by more than 2 and 1.5 times in patients with and without MOF, respectively. Increased level of sTREM-1 in early postoperative period is associated with enhanced expression of membrane form of TREM-1 and further triggering of innate immunity reactions in response to on-pump surgery. This intervention is accompanied by blood cell contact with extracorporeal circuit, hypoxia and ischemia-reperfusion injury of various organs as a result of aortic cross-clamping. Reperfusion is associated with higher activation of key indicators of inflammatory response including sTREM-1.

Increased serum sTREM-1 on the 1^st postoperative day in patients without MOF is explained by systemic inflammatory response without signs of MOF in some patients. SIRS can regress spontaneously or require short-term vasopressor support.

Considering significant data of ROC-analysis and correlation between serum sTREM-1 and SOFA score on the 1^st postoperative day (p<0.0001), we can assume advisability of sTREM-1 as a biomarker of MOF.

Conclusion

We first determined the relationship between serum sTREM-1 and decompensated non-infectious multiple organ failure in patients with coronary artery disease after coronary artery bypass surgery. According to our data, sTREM-1 can be used as prognostic marker of multiple organ failure in ICU patients after cardiac surgery.

Author contribution:

Concept and design of the study — Khutornaya M.V., Ponasenko A.V.

Collection and analysis of data — Khutornaya M.V., Sinitskaya A.V., Sinitsky M.Yu.

Statistical analysis — Khutornaya M.V., Sinitskaya A.V. Writing the text — Khutornaya M.V.

Editing — Khutornaya M.V., Grigoriev E.V.

The study was supported by the comprehensive program of fundamental researches of the Siberian Campus of the Russian Academy of Sciences within the fundamental issue of the Research Institute for Complex Issues of Cardiovascular Diseases No. 0419-2022-0001.

The authors declare no conflicts of interest.