Valvular heart disease (VHD) represents a significant proportion of cardiovascular conditions necessitating surgical intervention. With the global population aging at an accelerated pace, VHD has become increasingly prevalent, contributing substantially to cardiovascular morbidity and mortality worldwide (1–3). The primary approach for treating VHD is through open-heart valve replacement or repair under cardiopulmonary bypass (CPB), which remains the cornerstone of clinical management (4, 5). Despite advancements in surgical techniques and perioperative care, a growing number of critically ill patients with complex conditions continue to face suboptimal postoperative outcomes, including severe complications that may lead to death (6).

Hyperbilirubinemia (HB) is a common complication after VHD surgery. According to incomplete statistics, the incidence of HB in VHD patients after surgery is as high as 43.3% (7, 8). It is closely associated with poor perioperative prognosis and can progress to acute liver failure, becoming an important cause of in-hospital mortality (9, 10). Consequently, understanding the factors contributing to postoperative HB and establishing effective prevention strategies are crucial for improving patient outcomes. McSweeney et al. found that patients with postoperative HB after cardiac surgery had a prolonged ICU stay of approximately one week, doubled hospitalization time, and significantly increased mortality compared to non-hyperbilirubinemic patients (11). This suggests that clinical attention should be paid to the prevention and treatment of HB in VHD patients with after surgery.

In recent years, with the increasing epidemiological research on postoperative HB in VHD patients, most studies have also analyzed the influencing factors of HB. Currently, most studies focus on analyzing the relationship between laboratory indicators and postoperative HB in VHD patients. However, there is a notable absence of comprehensive studies investigating the multifactorial influences on the occurrence of HB and the lack of predictive models specifically tailored for this context. Therefore, further exploration of the influencing factors of postoperative HB in VHD patients and the construction of a predictive model for HB based on these factors are of great significance for reducing the risk of HB and improving patient prognosis (12, 13).

Given these gaps, this study aims to fill an important void by exploring the risk factors associated with postoperative HB in VHD patients who underwent surgery at Qingdao Cardiovascular Hospital. By retrospectively examining the clinical data of these patients, we divided them into groups based on whether they developed HB within one week after surgery. This division enabled us to identify key risk factors and develop a predictive model using multivariable logistic regression analysis. Furthermore, we constructed a nomogram model to facilitate the early identification of high-risk individuals, thereby enabling more proactive prevention and treatment strategies.

The establishment of such a predictive model not only aids in identifying patients at higher risk of developing HB but also provides valuable insights into the underlying mechanisms driving this condition. Understanding these mechanisms could pave the way for novel therapeutic approaches aimed at mitigating the impact of HB on patient recovery. Additionally, the application of our predictive model in clinical settings could enhance personalized medicine efforts, allowing healthcare providers to tailor interventions based on individual risk profiles. In summary, this study seeks to contribute significantly to the field of cardiac surgery by providing a robust framework for predicting and managing postoperative HB in VHD patients, ultimately aiming to improve overall patient outcomes and reduce the burden on public health systems.

Our study identified several key risk factors for postoperative HB in patients with VHD, including the type of surgery, preoperative MELD score, CPB time, aortic cross-clamp time, total intraoperative blood transfusion volume, and mechanical ventilation time. These findings are significant because they provide insights into the mechanisms behind postoperative HB and offer potential strategies for its prevention and management.

The incidence of HB varied among different types of surgeries, with a higher risk of HB associated with more complex surgical procedures. Patients undergoing aortic valve replacement or tricuspid valve surgery were more likely to develop HB compared to those undergoing mitral valve procedures. This may be attributed to the anatomical and physiological challenges of aortic and tricuspid surgeries, which often require longer CPB and cross-clamp times. Furthermore, multiple valve replacement surgeries, which inherently carry higher complexity and longer perfusion times, are recognized to increase the risk of hyperbilirubinemia. While our study did not include “multiple valve surgery” as a separate categorical variable, its risk profile is captured within our model through the closely correlated variables of “surgery type” and prolonged “CPB time”. For instance, aortic valve surgery involves intricate manipulation of the ascending aorta, potentially increasing the risk of systemic inflammation and microvascular injury (16, 17). Similarly, tricuspid valve procedures are frequently performed in patients with advanced right heart failure, which may compromise hepatic perfusion and exacerbate postoperative liver dysfunction (18, 19). Previous studies have indicated that the number of valve replacements is an independent risk factor for postoperative HB after cardiac surgery (20, 21). But this study emphasizes the role of specific valve types. This distinction could inform preoperative risk stratification, guiding surgeons to prioritize protective strategies for high-risk procedures.

A higher MELD score, reflecting impaired baseline liver function, was strongly linked to HB. The MELD score is a scoring system used to predict the severity and prognosis of liver disease, based on serum bilirubin, serum creatinine, and prothrombin time. A higher score indicates poorer liver function and worse prognosis (22, 23). Research has shown that many patients who develop HB after cardiac surgery often have preoperative liver dysfunction (9). Patients with elevated MELD scores likely have reduced hepatic synthetic capacity and bile excretion efficiency, making them more vulnerable to the metabolic stress of surgery. This aligns with Wang et al., who noted that preoperative liver dysfunction increases the likelihood of postoperative HB due to diminished resilience to surgical trauma (24).

Prolonged CPB and aortic cross-clamp times were independent risk factors for HB. The underlying mechanisms include ischemia-reperfusion injury, systemic inflammation, and hemolysis. During CPB, the liver experiences intermittent hypoperfusion, leading to oxidative stress and impaired bilirubin conjugation. Additionally, the mechanical shear forces of CPB can damage red blood cells, releasing unconjugated bilirubin into circulation. Aortic cross-clamp time further exacerbates this by prolonging myocardial ischemia and increasing the release of inflammatory mediators. These findings are consistent with studies by Pasternack et al. and Wang et al., which indicated that the longer CPB and aortic cross-clamp time, the higher the incidence of postoperative HB (25, 26).

Increased intraoperative blood transfusion volume was a critical predictor of HB. Previous studies have found that increased intraoperative blood transfusion volume may lead to hemolysis, resulting in increased bilirubin load, more severe HB, and delayed peak bilirubin levels. Additionally, transfused red blood cells, particularly when stored for extended periods, are prone to hemolysis, releasing free hemoglobin and bilirubin precursors, leading to increased bilirubin concentration. And its osmotic fragility may be more easily disrupted under the impact of CPB. Once the bilirubin produced during this process exceeds the liver's maximum metabolic capacity, it may lead to bilirubin accumulation, resulting in HB occurrence. Stored blood also contains pro-inflammatory cytokines that may worsen liver injury (27, 28).

The mechanical ventilation time has been found to be positively correlated with the incidence of postoperative HB. This may be because patients with severely compromised respiratory function or extensive surgical trauma require longer mechanical ventilation support after surgery. Prolonged positive pressure ventilation can increase right atrial pressure, inhibit venous return, cause hepatic congestion, further aggravate liver injury, and lead to the occurrence of HB. This mechanism is supported by Lyu et al., who noted that ventilatory support duration correlates with HB severity (29). Additionally, prolonged ventilation may delay recovery of spontaneous breathing, prolonging ICU stays and increasing exposure to nephrotoxic medications, which can synergistically worsen liver function.

Based on the identified risk factors, a predictive model was constructed using nomogram, which is an intuitive data visualization tool that presents the quantities between different categories simultaneously, making it easier to compare the relative quantities between different categories. This helps clinical doctors better understand the dynamic changes in the data (30). Although some variables are intraoperative or postoperative, the model still allows for preliminary risk assessment preoperatively based on surgery type and MELD score, which aids in preoperative communication and resource allocation. The ROC analysis results in this study showed that the nomogram had good accuracy and efficacy in predicting the occurrence of HB in the modeling and validation groups. This indicates that nomogram has good predictive accuracy and can be used by clinicians to assess the risk of postoperative HB in VHD patients based on the aforementioned risk factors, thereby providing individualized and precise interventions to reduce the risk of HB (31).

This model helps identify patients at high risk for postoperative hyperbilirubinemia, enabling intensified liver function monitoring, optimized fluid management, and timely hepatoprotective interventions to improve outcomes. Risk stratification using this model could guide the frequency of postoperative monitoring and intensity of interventions, such as dynamic bilirubin surveillance and early treatment in high-risk patients. The model could be integrated into electronic medical record systems to generate individualized risk scores preoperatively, guiding perioperative liver protection strategies and facilitating precision medicine. The model demonstrates good clinical applicability, assisting clinicians in identifying high-risk patients preoperatively and formulating individualized management plans, which may improve patient outcomes.

While this study confirms known risk factors like CPB time and transfusion volume, it expands the understanding of HB etiology by emphasizing the role of surgical type and MELD score. For example, McSweeney et al. focused on general cardiac surgery but did not differentiate valve-specific risks (11). The current findings suggest that valve type should be considered in risk models for VHD patients. Additionally, the integration of MELD score into predictive modeling offers a practical tool for preoperative counseling and resource allocation. However, the study's retrospective design limits causal inference, and further prospective validation is needed.

This study has several limitations. First, its retrospective, single-center design inherently carries risks of selection bias and unmeasured confounding, such as variations in surgical technique, anesthesia protocols, or specific medication use that could influence liver function. Second, the sample size, while sufficient for initial model development, may limit the generalizability of our findings and the model's stability for predicting very rare outcomes. Third, as noted, we did not explicitly include multiple valve surgery as a variable, and other potentially relevant factors like detailed data on stored blood product age or specific markers of hemolysis were not available. Moreover, we did not differentiate between direct and indirect bilirubin, which could provide further pathophysiological insights into the type of hyperbilirubinemia. Future research should focus on multi-center prospective trials to validate the nomogram in diverse populations. Additionally, mechanistic studies exploring the role of oxidative stress, inflammatory cytokines, and red blood cell metabolism in HB pathogenesis are warranted. Advances in machine learning and wearable monitoring technologies could further refine risk prediction by integrating dynamic physiological parameters.

In conclusion, the type of surgery, preoperative MELD score, CPB time, aortic cross-clamp time, total intraoperative blood transfusion volume, and mechanical ventilation time are risk factors for postoperative HB in VHD patients. The nomogram constructed based on these risk factors has good predictive value and accuracy. However, this study had limited variables and was retrospective in nature, which has certain limitations. Future studies will include more variables and further validate the constructed predictive model. They also should prioritize translational research to bridge the gap between risk prediction and clinical implementation.