Ulinastatin Reduces Postoperative Pulmonary Complications in Cardiac Surgery

Overview

This study investigates the association between intraoperative ulinastatin administration and PPCs incidence in cardiac surgery. It comprises three core components: an innovative retrospective analysis, a bioinformatics study integrating network pharmacology and molecular docking, and a clinical cohort study. Figure 1 presents a flowchart of the three-phase study, including participant screening procedures, key sample collection time points, and the analytical framework.

Step Ⅰ: Retrospective Cohort Study of Intraoperative Ulinastatin and PPCs Risk This segment of the study is designed as a retrospective, multicenter observational investigation, encompassing adult patients who underwent elective cardiac surgery at three medical centers-specifically Tongji Guanggu Hospital, Tongji Zhongfa Hospital, and Tongji Hankou Hospital, all situated in Wuhan, China-between January 2014 and December 2022. Due to the retrospective design of the study, the sample size was pre-determined\[14\]. Participants Adult patients undergoing elective cardiac surgery, including open-heart valve repair, valve replacement with cardiopulmonary bypass, and off-pump coronary artery bypass grafting, were consecutively screened for eligibility. Exclusion criteria were as follows: (1) age \< 18 years; (2) pregnant women; (3) history of prior cardiac surgery before the current hospitalization; (4) presence of acute respiratory diseases prior to surgery, such as respiratory tract infections, respiratory failure, pleural effusion, atelectasis, pneumothorax, bronchospasm, and pneumonia; (5) administration of ulinastatin therapy prior to cardiac surgery during the current hospitalization; (6) The drug ulinastatin was utilized during the surgery, with a total intraoperative dose of less than 50,000 U; (7) history of malignancy or bone marrow transplantation. Variables The primary exposure was intraoperative ulinastatin administration, defined by: (1) dosage: total intraoperative dose (≥50,000 U); (2) timing: first administration at anesthesia induction vs during cardiopulmonary bypass; (3) route: intravenous drip. Demographic characteristics included age, gender, and smoking status, all ascertained at admission. Clinical comorbidities were documented based on clinical diagnoses in medical records, including hypertension, diabetes, chronic pulmonary diseases (chronic obstructive pulmonary disease, asthma, chronic bronchitis, bronchiectasis), and chronic renal failure. Preoperative laboratory indicators for patients consisted of venous blood test results obtained within 24 hours of admission, including white blood cell count and hemoglobin concentration. Based on prior literature and biological plausibility, measured confounding factors included variables known or hypothesized to affect the risk of PPCs. These covariates comprised age, gender, body mass index (BMI), preoperative hemoglobin level, preoperative white blood cell count, preoperative blood pressure, and comorbidities. All covariates were pre-specified and incorporated into a multivariate adjustment model to address baseline imbalances and mitigate confounding bias in the assessment of clinical outcomes. Outcomes The primary outcome was the development of PPCs within 7 days postoperatively or prior to hospital discharge. The diagnostic criteria for PPCs were based on the definitions outlined in the European Perioperative Clinical Outcome definitions, published jointly by the European Society of Anaesthesiology and the European Society of Intensive Care Medicine in 2015. In the present study, PPCs included respiratory tract infections, respiratory failure, pleural effusion, atelectasis, bronchospasm, and aspiration pneumonia. Notably, given that thoracotomy is a standard component of cardiac surgery and postoperative pneumothorax is thus a common procedural sequela, postoperative pneumothorax was excluded from the definition of PPCs in this study. Secondary outcomes included postoperative delirium and in-hospital all-cause mortality. For postoperative delirium diagnosis, two criteria were applied: (1) a documented postoperative delirium diagnosis was regarded as confirmed and accepted as valid; (2) in the absence of an initial postoperative delirium diagnosis, data abstractors retrospectively reviewed clinical records for signs of delirium. The criteria for retrospective postoperative delirium diagnosis were aligned with the adapted supplementary diagnostic criteria from the Diagnostic and Statistical Manual of Mental Disorders, 5th Edition (DSM-5) for delirium\[7\]. In-hospital all-cause mortality was defined as in-hospital death from any cause following surgery. Bias Multiple strategies were implemented throughout the study to mitigate potential bias. Clinical data extraction was performed independently by two trained researchers in accordance with a standardized protocol, utilizing information retrieved from electronic medical records. Discrepancies were adjudicated by a third senior reviewer to ensure data consistency and accuracy. Clinical outcomes were ascertained by an independent clinical research team blinded to exposure data, thereby further reducing information bias. Predefined confounding variables, identified based on biological plausibility and prior evidence, were incorporated into multivariate models to control for confounding bias. Step Ⅱ: Network Pharmacology-Based Prediction and Molecular Docking Simulations To investigate the molecular mechanisms underlying the reduction of PPCs by ulinastatin in cardiac surgery patients, this segment adopts an integrated approach combining network pharmacology and molecular docking. Disease Target Collection and Standardization These targets encompassed six key PPCs-related clinical conditions, with respective target counts as follows: Respiratory tract infections (n = 14,411); Aspiration pneumonia (n = 4,033); Pulmonary atelectasis (n = 1,107); Pleural effusion (n = 2,282); Bronchial spasm (n = 7,293); and Respiratory failure (n = 12,467). To ensure consistency in target names, all target gene names were standardized using the UniProt database. By screening for overlapping genes in the aforementioned datasets, co-expressed genes were identified, which were considered molecular mediators linking ulinastatin to potential pathophysiological mechanisms of PPCs. Functional enrichment analysis To clarify the biological functions of co-expressed genes and their participation in signaling cascades, Gene Ontology (GO) enrichment analysis and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis were performed in parallel\[17, 18\]. A statistical threshold of adjusted P-value \< 0.05 was applied to ensure the robustness and functional relevance of the enrichment results. Potential target protein localization Extracellular secreted proteins were selected because ulinastatin, as a broad-spectrum protease inhibitor, exerts its biological effects primarily in the extracellular space; thus, only extracellular targets were considered to ensure biological plausibility between ulinastatin and its targets. The extracellular secreted protein dataset was obtained from the Human Proteome Atlas, and cardiac surgery-related targets were obtained from the GeneCards database (Cardiac Surgical Procedures, n = 5,931). Potential target proteins were precisely identified by taking the intersection of postoperative pulmonary complication-related targets, cardiac surgery-related targets, and the extracellular secreted protein dataset. Protein-Protein Interaction Network Construction Concurrently with functional analysis, a protein-protein interaction (PPI) network was constructed using the STRING database, with the analytical scope restricted to the overlapping targets identified in the Venn diagram. To ensure high confidence in the interactions, the species was specified as Homo sapiens, and a stringent interaction score threshold (≥0.9) was applied. Identification of key target genes of ulinastatin A systematic approach was employed to identify the key molecular targets of ulinastatin: first, the precise name and molecular formula of the compound were retrieved from the PubChem database and imported into the Swiss-Target Prediction platform to predict and retrieve potential target molecules. Subsequently, intersection analysis was performed to integrate ulinastatin target molecules with the aforementioned co-expression gene dataset, screening for molecules that simultaneously satisfy the criteria of "regulated by ulinastatin," "extracellularly secreted in relevant biological contexts," and "co-expressed." A gene interaction network was then constructed using Friends' analysis for further optimization, with the importance of each gene evaluated based on network topological metrics such as connectivity and centrality. Finally, the key node genes mediating the potential biological effects of ulinastatin were identified. Molecular Docking and Dynamics Simulations The three-dimensional structure of the small-molecule ligand ulinastatin was retrieved from the PubChem database (PubChem CID: 105102), while the protein MMP3 structure (PDB ID: 4g9l) was obtained from the RCSB PDB database. Preference was given to the human protein conformation with the highest structural accuracy for subsequent analyses. Pre-docking preprocessing of the target protein was performed using PyMOL software: first, redundant water molecules and endogenous ligands were removed from the structure to eliminate interference from non-specific interactions; subsequently, hydrogen atoms were added to the protein backbone and side chains to ensure the completeness of the protonation state at the binding site, thereby improving the accuracy of docking calculations. Molecular docking simulations were conducted using the CB-Dock2 online platform (https://cadd.labshare.cn/cb-dock2/index.php). Following docking, LIGPLOT software was utilized to generate protein-ligand interaction schematics, which visually illustrate key binding residues and intermolecular forces (including hydrogen bonds, hydrophobic interactions, and van der Waals forces). This process clarifies the specific binding mode between the ligand and target protein. Step III: Validation in clinical samples Clinical validation cohort study design and subjects This is a prospective observational study. Plasma samples from the clinical validation cohort were collected from adult patients who underwent elective cardiac surgery between April 2023 and September 2024, including open-heart valve repair, cardiopulmonary bypass-assisted valve replacement, and off-pump coronary artery bypass grafting. Apply the same exclusion criteria as used in the first part of this study. The primary outcome of this study is the occurrence of PPCs within 7 days postoperatively or prior to discharge. Sample Size The minimum sample size was calculated based on the incidence rates of PPCs in the ulinastatin and non-ulinastatin groups. As derived from the first part of the study, the PPCs incidence rates were 22.8% in the ulinastatin group and 2.7% in the non-ulinastatin group. Sample size calculation was performed using PASS software with parameters set as follows: significance level (α) = 0.05 and statistical power (1-β) = 0.80. The sample size initially consisted of 168 cases, with an additional 15% dropout rate accounted for, leading to a minimum required sample size of 193 cases. Blood Sample Collection and Testing Venous blood samples were collected from patients at three predefined perioperative time points: (1) preoperatively (before surgical incision); (2) immediately postoperatively (upon admission to the intensive care unit); and (3) 24 hours postoperatively. Sample processing procedure: Blood was collected in EDTA anticoagulant tubes, gently inverted for mixing, and transported on ice. Plasma was separated via rapid centrifugation under the following conditions: \[3000 × g, 4℃, 10 min\]. Aliquoted plasma samples were immediately stored at -80°C in a dedicated biobank for subsequent batch testing. Quantified using an enzyme-linked immunosorbent assay (ELISA) with kit HM10736 (Bio-swamp Company, Wuhan, China), following the manufacturer's protocols strictly. Statistical Methods The Shapiro-Wilk test was used to evaluate the normality of continuous variables. Normally distributed variables were expressed as mean ± standard deviation (x̄ ± s), with between-group comparisons performed using one-way analysis of variance (ANOVA). If the Levene test indicated unequal variances, Welch's ANOVA was applied instead. Skewed continuous variables were presented as median (interquartile range) \[M (P25, P75)\], and intergroup comparisons were conducted using the Kruskal-Wallis test. Categorical variables were expressed as counts (percentages) \[n (%)\], with between-group comparisons using the chi-square test or Fisher's exact test, depending on sample size and expected frequency distribution. Given that PPCs predominantly occur within 7 days postoperatively and all patients were followed up through discharge, logistic regression was employed. Multivariate logistic regression models were employed to examine the association between changes in plasma MMP3 concentration and clinical outcomes. Predefined covariates, selected based on clinical relevance and prior literature, were included in the models for adjustment. Results were reported as odds ratios (OR) with 95% confidence intervals (95% CI), reflecting the strength of association and statistical precision. A two-tailed P \< 0.05 was set as the threshold for statistical significance. Interaction terms were incorporated into the logistic regression models to explore the modifying effects of factors such as sex and comorbidities on the primary association. The likelihood ratio test was used to compare models with and without interaction terms, and interaction P-values were calculated to assess the statistical significance of these effects. Further subgroup analyses were performed based on the aforementioned factors to identify differential association effects across subgroups.

Conditions

Eligibility

Locations (1)

Outcomes