Abstract
Background
Major adverse cardiovascular and cerebrovascular events (MACCE) pose significant challenges in hepatobiliary and pancreatic surgeries, contributing to increased morbidity, mortality, and healthcare burdens. Accurate identification of perioperative risk factors remains critical yet underexplored due to limited large-scale data. This study aimed to find intraoperative predictors of MACCE within 30 (MACCE-30) and 90 days (MACCE-90) post-surgery using the INSPIRE database.
Methods
A retrospective cohort of 6135 patients from Seoul National University Hospital was analyzed. MACCE was defined as angina, myocardial infarction, cardiac arrest, arrhythmia, heart failure, stroke, or in-hospital mortality. Variables including patient characteristics, intraoperative factors (e.g., estimated blood loss [EBL], crystalloid volume), and comorbidities were extracted. Univariate Cox proportional hazards models identified variables (p < 0.2) for multivariate analysis, with Kaplan-Meier plots visualizing survival differences.
Results
Of 6135 patients, 179 experienced MACCE-90. Multivariate analysis identified male sex (HR 1.88, 95% CI 1.27–2.78, p < 0.01), EBL (HR 1.19, 95% CI 1.03–1.37, p = 0.02), and crystalloid infusion volume (HR 1.29, 95% CI 1.00-1.65, p = 0.05) as significant predictors of MACCE-90. Mean EBL was 565 mL (MACCE) versus 357 mL (non-MACCE), and crystalloid volume was 498 mL versus 486 mL. We observed no significant difference in early survival based on sex, EBL, or infused crystaollioids.
Conclusion
Male sex, higher EBL, and crystalloid volumes predict MACCE-90. Optimized management and refined protocols are needed.
Supplementary Information
The online version contains supplementary material available at 10.1186/s12893-025-03359-w.
Keywords: Major adverse cardiovascular and cerebrovascular events, Intraoperative predictors, Blood loss, Crystalloid volume
Background
Major adverse cardiovascular and cerebrovascular events (MACCE) represent a practical challenge encountered in clinical practice within the of surgical patient care, exerting a profound impact on morbidity, mortality, and healthcare resource utilization [1]. Contemporary epidemiological studies indicate a heterogeneous prevalence of MACCE across diverse surgical interventions, contributing to elevated rates of postoperative complications and extended hospitalizations [2]. These adverse outcomes not only diminish patients’ quality of life but also impose significant burdens on healthcare systems globally.
Consequently, precise perioperative risk assessment for MACCE is imperative to enhance patient prognosis [3]. By identifying individuals at heightened risk, clinicians can implement customized preventive measures, including preoperative optimization, meticulous intraoperative hemodynamic management, and rigorous postoperative surveillance [4–6]. Nevertheless, despite extensive investigative efforts, the comprehension of MACCE risk factors specific to hepatobiliary and pancreatic surgeries remains incomplete. Prior research has often been hampered by the lack of robust, large-scale perioperative datasets [7–9], which has constrained the ability to elucidate the complex interplay among intraoperative variables, patient-specific characteristics, and the incidence of MACCE.
To address this gap, the current study leverages the INSPIRE database, a comprehensive perioperative research repository, to conduct an in-depth analysis of patients undergoing hepatobiliary and pancreatic surgeries. The objective is to delineate potential predictors of MACCE, such as intraoperative bleeding [10]. In an era increasingly characterized by precision medicine, clarifying these factors is critical to unraveling the pathophysiological mechanisms underlying MACCE in this surgical cohort and to developing targeted therapeutic strategies.
The primary objective of this investigation is to identify predictors of MACCE occurring within 30 (MACCE-30) and 90 days (MACCE-90) post-surgery following hepatobiliary and pancreatic procedures. We posit that a range of patient-specific and procedural factors is associated with an elevated risk of major adverse cardiovascular and cerebrovascular events [11]. This hypothesis is grounded in well-established clinical and physiological evidence suggesting that males may exhibit distinct cardiovascular responses to surgical stress, potentially attributable to hormonal and anatomical differences [12, 13]. Additionally, procedural factors, such as significant blood loss, may disrupt hemodynamic stability and precipitate organ dysfunction [14–16].
Through a meticulous evaluation of intraoperative and patient-related risk factors, this study aims to bridge existing knowledge deficits concerning MACCE in hepatobiliary and pancreatic surgeries. The anticipated findings are poised to enhance perioperative risk stratification methodologies and refine surgical management protocols, ultimately improving clinical outcomes for this high-risk patient population.
Methods
Study design and data collection
INSPIRE is an extensive perioperative research database that has been meticulously described in previous studies, which includes approximately 130,000 cases (50% of all surgical cases) who underwent anesthesia for surgery at an academic institution in South Korea between 2011 and 2020. For each surgical procedure, detailed records were maintained that captured critical data including surgical and anesthesia-related variables, perioperative diagnoses, continuously monitored vital signs, and comprehensive laboratory results. This rich repository of clinical information provides a unique opportunity to examine perioperative outcomes and risk factors on a large scale [17].
All data in INSPIRE are publicly accessible, supporting transparency and facilitating further research. This study received formal approval from Seoul National University Hospital (No. H-2210-078-1368), and due to the retrospective nature of the research, the Institutional Review Board waived the requirement for informed consent. Furthermore, the Institutional Data Review Board at SNUH rigorously evaluated the dataset, confirmed that all patient data had been properly de-identified, and approved its public release (BRB No. BD-R-2022-11-02). These steps ensure that while the data remains a powerful resource for scientific inquiry, patient confidentiality and ethical standards are fully maintained [18].
The clinical data repository served as the backbone for extracting essential variables such as surgical and anesthesia-related details, diagnoses, vital signs, and laboratory results. Clinical status parameters, including weight and height, were obtained through standardized physical measurements, ensuring consistency and reliability in the data. To minimize the risk of re-identification, every diagnosis was systematically converted to ICD-10-CM codes, and all procedure names were manually mapped to ICD-10-PCS codes. Vital signs, recorded automatically every minute during anesthesia, were aggregated to a maximum interval of five minutes to further protect patient privacy. Additionally, the anesthesia records provided manual entries for key intraoperative measures such as urine output, estimated blood loss (EBL), and the volumes of fluids or blood products transfused, as well as readings from specialized monitoring devices, which collectively enrich the dataset with nuanced clinical details.
For the purposes of this study, only hepatobiliary and pancreatic surgeries were selected based on their ICD10_PCS codes, ensuring a focused analysis on these specific types of procedures. To maintain the integrity of the analysis and avoid potential confounding from repeated measures, patients who underwent multiple surgeries during the study period were excluded. This rigorous selection process resulted in a final study cohort comprising 6135 individuals and their corresponding surgical cases. This well-defined cohort enables a precise investigation of perioperative outcomes and the factors influencing them within a targeted surgical population.
Outcomes
The primary outcome was defined as the occurrence of MACCE within 30 days after surgery, while the secondary outcome was the occurrence of MACCE within 90 days post-surgery. Drawing from previous studies, MACCE was characterized by a spectrum of critical events including angina, acute myocardial infarction, cardiac arrest, arrhythmia, heart failure, stroke, and in-hospital all-cause mortality, or any combination thereof. Each event was carefully diagnosed by a physician and subsequently coded using ICD-10-CM standards to maintain consistency and reliability across cases. The interval from the end of the operation to the first occurrence of a MACCE was defined as the time to event, providing a clear temporal framework for assessing the risk and timing of these complications. This comprehensive approach not only facilitates a detailed analysis of perioperative risk factors but also contributes to the overall understanding of how intraoperative and postoperative events impact patient outcomes.
Variables
The following variables were extracted and constructed from the vitals and labs files to provide a comprehensive profile of each patient’s perioperative status: Charlson Comorbidity Index, Revised Cardiac Risk Index, Operation Duration, Anesthesia Duration, Body Mass Index, Surgical Approach, Serum Creatinine, Glucose Level, Hemoglobin A1c, Albumin Level, C-Reactive Protein, Total EBL, Total Urine Output, Crystalloid Volume, Colloid Infusion (binary), Blood Product Transfusion (binary), Vasopressor Infusion (binary), and Occurrence of Hypotension (binary). These variables were chosen based on their potential association with postoperative outcomes and to capture both the preoperative risk factors and intraoperative management details. For example, the Charlson Comorbidity Index and Revised Cardiac Risk Index provide standardized measures of underlying patient health and cardiac risk, while operation and anesthesia durations offer insight into the complexity and extent of the surgical procedure. Laboratory measurements such as serum creatinine, glucose level, and C-reactive protein were used to assess the patient’s physiological state before and after surgery, and the binary variables indicate the presence or absence of specific interventions or events during surgery.
The occurrence of hypotension was defined as either an invasive or non-invasive mean arterial pressure below 60 mmHg for two consecutive measurements or more, emphasizing the need for sustained blood pressure monitoring during surgery. This detailed extraction process ensures that both continuous and categorical data are appropriately represented, allowing for robust statistical analysis and accurate risk stratification.
Statistical analyses
Categorical variables were summarized as counts and percentages, with group differences evaluated using Pearson’s Chi-squared test or Fisher’s exact test when cell counts were small, following the default behavior of the gtsummary package. Continuous variables were summarized as mean ± standard deviation, and group differences were assessed using the Wilcoxon rank sum test, consistent with gtsummary’s default settings for two-group comparisons.
The primary endpoint for this study was the occurrence of MACCE post-surgery, with the time to MACCE serving as the time-to-event variable.
Initially, univariate Cox proportional hazards models were constructed for each variable of interest to assess its individual association with the time to MACCE. Variables yielding a p-value less than 0.2 in the univariate analyses were considered for inclusion in the subsequent multivariate Cox regression models. This threshold was chosen to ensure that potential confounders and variables of clinical interest were not prematurely excluded. The coxph function from R’s survival package was used for the Cox regression analysis, with missing values handled according to the function’s default settings, which excludes observations containing missing values.
Before finalizing the multivariate model, we assessed multicollinearity among the candidate variables using the Variance Inflation Factor (VIF) as implemented in the ‘car’ package in R. Variables with high VIF values were scrutinized, and if necessary, collinear variables were either combined or excluded (Table S1). The missing values are process by the coxph function by default.
For visualization of the survival data, Kaplan-Meier survival plots were generated, with the patient population dichotomized at the 75th percentile of the continuous predictors. This approach allowed us to explore the differences in survival functions and the impact of predictor variables on time to MACCE visually. Furthermore, forest plots were used to succinctly display the hazard ratios (HR) along with their 95% confidence intervals (CI) derived from the multivariate Cox regression analyses, providing a clear overview of the relative risk associated with each variable.
All statistical analyses were performed using R (version 4.4), and a two-sided p-value < 0.05 was considered statistically significant unless otherwise noted.
Results
Baseline
The study cohort comprised 6135 patients, stratified based on the occurrence of MACCE within 90 days (Table 1, MACCE within 30 days are shown in Table S2). Among these, 5,956 patients did not experience MACCE, while 179 patients had documented events. The mean age of the cohort was higher in the MACCE group (67.88 ± 11.84 years) compared to the non-MACCE group (54.93 ± 14.62 years, p < 0.01). Regarding sex distribution, 46% of patients in the non-MACCE group were female (n = 2,737) compared to 34% (n = 61) in the MACCE group, while males were more prevalent in the MACCE group (66% vs. 54%, p < 0.01).
Table 1.
Baseline characteristics stratified by MACCE 90 outcome
| Variable | MACCE within 90 Days N = 1791 |
No MACCE within 90 Days N = 5,9561 |
p-value2 |
|---|---|---|---|
| Age (years) | 67.88 ± 11.84 | 54.93 ± 14.62 | < 0.01 |
| Sex | < 0.01 | ||
| Female | 61 (34%) | 2,737 (46%) | |
| Male | 118 (66%) | 3,219 (54%) | |
| Charlson Index | 2.49 ± 2.28 | 1.45 ± 1.99 | < 0.01 |
| RCRI Score | < 0.01 | ||
| 0 | 53 (30%) | 5,707 (96%) | |
| 1 | 105 (59%) | 229 (3.8%) | |
| 2 | 20 (11%) | 19 (0.3%) | |
| 3 | 1 (0.6%) | 1 (< 0.1%) | |
| Operation Duration (min) | 138.46 ± 122.33 | 141.98 ± 125.13 | 0.99 |
| Anesthesia Duration (min) | 172.29 ± 135.05 | 177.21 ± 140.93 | 0.85 |
| BMI (kg/m²) | 24.19 ± 9.96 | 23.91 ± 6.17 | 0.75 |
| Surgery Approach | 0.80 | ||
| endoscopic | 110 (61%) | 3,605 (61%) | |
| open | 69 (39%) | 2,351 (39%) | |
| Creatinine (mg/dL) | 1.17 ± 0.89 | 0.89 ± 0.54 | < 0.01 |
| Glucose (mg/dL) | 119.66 ± 41.62 | 111.07 ± 33.00 | 0.02 |
| HbA1c (%) | 6.57 ± 1.06 | 6.22 ± 1.00 | 0.01 |
| Albumin (g/dL) | 3.65 ± 0.62 | 3.95 ± 0.53 | < 0.01 |
| CRP (mg/L) | 3.23 ± 5.40 | 1.86 ± 3.92 | < 0.01 |
| Estimated Blood Loss (mL) | 564.69 ± 1,305.46 | 357.16 ± 1,088.56 | 0.14 |
| Urine Output (mL) | 112.35 ± 212.72 | 132.53 ± 249.02 | 0.17 |
| Crystalloid Volume (mL) | 497.60 ± 860.41 | 486.31 ± 702.42 | 0.04 |
| Colloid Infusion (mL) | 0.01 | ||
| Yes | 43 (24%) | 975 (16%) | |
| Blood Product Transfusion (units) | < 0.01 | ||
| Yes | 26 (15%) | 323 (5.4%) | |
| Vasopressor Infusion | < 0.01 | ||
| Yes | 22 (12%) | 214 (3.6%) | |
| Hypotension Occurrence | < 0.01 | ||
| Yes | 46 (26%) | 665 (11%) |
1 Mean ± SD; n (%)
2 Wilcoxon rank sum test; Pearson’s Chi-squared test; Fisher’s exact test
Clinical characteristics
The Charlson Comorbidity Index was significantly higher in patients with MACCE (mean: 2.49 ± 2.28) compared to those without (1.45 ± 1.99, p < 0.01). The Revised Cardiac Risk Index (RCRI) scores demonstrated a significant association with MACCE incidence (p < 0.01). The majority of patients in the non-MACCE group had an RCRI score of 0 (96%), whereas only 30% of the MACCE group had an RCRI score of 0. Conversely, a higher proportion of patients in the MACCE group had RCRI scores of 1 (59% vs. 3.8%), 2 (11% vs. 0.3%), and 3 (0.6% vs. <0.1%).
Operative and anesthetic characteristics, including operation duration (MACCE: 138.46 ± 122.33 min, non-MACCE: 142 ± 125 min, p = 0.99) and anesthesia duration (MACCE: 172.29 ± 135.05 min, non-MACCE: 177.21 ± 140.93 min, p = 0.85), did not differ significantly between the groups. The distribution of surgical approaches was comparable, with 61% undergoing endoscopic surgery in both groups (p = 0.80).
Biochemical parameters showed significant differences between groups. Patients in the MACCE group had higher creatinine levels (1.17 ± 0.89 mg/dL vs. 0.89 ± 0.54 mg/dL, p < 0.01), glucose levels (119.66 ± 41.62 mg/dL vs. 111.07 ± 33.00 mg/dL, p = 0.02), and HbA1c levels (6.57 ± 1.06% vs. 6.22 ± 1.00%, p = 0.01). Additionally, albumin levels were lower in the MACCE group (3.65 ± 0.62 g/dL vs. 3.95 ± 0.53 g/dL, p < 0.01), and C-reactive protein (CRP) levels were elevated (3.23 ± 5.40 mg/L vs. 1.86 ± 3.92 mg/L, p < 0.01).
Intraoperative fluid and blood management also differed between groups. Crystalloid volume was slightly higher in the MACCE group (497.60 ± 860.41 mL vs. 486.31 ± 702.42 mL, p = 0.04), and colloid infusion was more frequent (24% vs. 16%, p = 0.01). Blood product transfusion was significantly associated with MACCE occurrence, with 15% of affected patients requiring transfusion compared to 5.4% in the non-MACCE group (p < 0.01). Vasopressor infusion was administered more frequently in the MACCE group (12% vs. 3.6%, p < 0.01), and hypotension events were more common (26% vs. 11%, p < 0.01).
EBL (MACCE: 564.69 ± 1,305.46 mL, non-MACCE: 357.16 ± 1,088.56 mL, p = 0.14) and urine output (MACCE: 112.35 ± 212.72 mL, non-MACCE: 132.53 ± 249.02 mL, p = 0.17) were not significantly different.
Overall, the MACCE group demonstrated a higher burden of comorbidities, worse renal and metabolic profiles, increased intraoperative interventions, and greater hemodynamic instability compared to those without MACCE. These findings highlight potential risk factors that may contribute to postoperative adverse cardiovascular and cerebrovascular events.
Male Sex, higher EBL, and increased crystalloid infusion as significant predictors of MACCE-90 risk
The multivariable analysis identified male gender, EBL, and crystalloid infusion volume as significant predictors of MACCE-90. The results are shown in Table 2 and the forest plot is shown in Fig. 1. Additionally, no significant predictors were identified for MACCE-30, as detailed in Table S3 and Figure S1.
Table 2.
Multivariable analysis identifying predictors of MACCE-90
| Characteristic | Univariate Analysis (MACCE-90) | Multivariate Analysis (MACCE-90) | |||||
|---|---|---|---|---|---|---|---|
| HR1 | 95% CI1 | p-value | HR1 | 95% CI1 | p-value | ||
| Age (years) | 0.99 | 0.98, 1.01 | 0.47 | ||||
| Sex | |||||||
| Female | — | — | — | — | |||
| Male | 1.49 | 1.09, 2.05 | 0.01 | 1.88 | 1.27, 2.78 | < 0.01 | |
| Charlson Index | 1.01 | 0.95, 1.09 | 0.68 | 1.01 | 0.93, 1.10 | 0.80 | |
| RCRI Score | 0.88 | 0.69, 1.12 | 0.31 | 1.03 | 0.74, 1.43 | 0.90 | |
| Operation Duration (min) | 1.00 | 0.99, 1.00 | 0.28 | ||||
| Anesthesia Duration (min) | 1.00 | 0.99, 1.00 | 0.18 | 1.00 | 0.99, 1.00 | 0.23 | |
| BMI (kg/m²) | 1.02 | 1.00, 1.04 | 0.06 | 1.01 | 0.99, 1.03 | 0.15 | |
| Surgery Approach | |||||||
| Endoscopic | — | — | |||||
| Open | 1.09 | 0.80, 1.47 | 0.60 | ||||
| Creatinine (mg/dL) | 0.96 | 0.80, 1.14 | 0.62 | ||||
| Glucose (mg/dL) | 1.00 | 0.99, 1.00 | 0.77 | ||||
| HbA1c (%) | 1.00 | 0.77, 1.31 | 0.98 | ||||
| Albumin (g/dL) | 0.74 | 0.56, 0.98 | 0.03 | 0.84 | 0.57, 1.25 | 0.37 | |
| CRP (mg/L) | 0.99 | 0.96, 1.04 | > 0.90 | ||||
| Estimated Blood Loss (L) | 1.15 | 1.02, 1.29 | 0.03 | 1.19 | 1.03, 1.37 | 0.02 | |
| Urine Output (L) | 1.34 | 0.65, 2.77 | 0.43 | ||||
| Crystalloid Volume (L) | 1.17 | 0.98, 1.40 | 0.08 | 1.28 | 1.00, 1.65 | 0.05 | |
| Colloid Infusion (mL) | |||||||
| Yes | 1.44 | 1.01, 2.04 | 0.04 | 1.25 | 0.65, 2.40 | 0.53 | |
| Blood Product Transfusion | |||||||
| Yes | 1.29 | 0.85, 1.96 | 0.24 | ||||
| Vasopressor Infusion | |||||||
| Yes | 1.14 | 0.73, 1.79 | 0.56 | ||||
| Hypotension Occurrence | |||||||
| Yes | 1.24 | 0.88, 1.74 | 0.22 | ||||
1 HR = Hazard Ratio, CI = Confidence Interval
Fig. 1.
Forest Plot of Predictors of MACCE Within 90 Days Post-Surgery. HR and 95% confidence intervals (CI) for factors associated with MACCE within 90 days following hepatobiliary and pancreatic surgeries, derived from a multivariate Cox proportional hazards model. Squares represent point estimates of HRs; horizontal lines indicate 95% CIs. The vertical dashed line at HR = 1 denotes no effect. Predictors include male sex, EBL (L), and crystalloid infusion volume (L). Statistical significance is denoted by CIs not crossing the reference line
Male patients had a significantly higher risk of MACCE-90 compared to female patients, with a HR of 1.88 (95% confidence interval [CI]: 1.27–2.78, p < 0.01). This finding is consistent with the baseline characteristics, where the proportion of males was higher in the MACCE group. The increased risk associated with male sex may reflect underlying sex-related differences in cardiovascular physiology, hormonal influences, or preexisting comorbidities.
EBL was also significantly associated with MACCE-90 (HR: 1.19, 95% CI: 1.03–1.37, p = 0.02). While the HR appears numerically small, it is important to consider that the volume of blood loss is typically large in surgical settings. Given that the mean EBL was 357 mL in the non-MACCE group and 565 mL in the MACCE group, an additional 500 to 1000 mL of blood loss could meaningfully increase the risk of MACCE. Greater intraoperative blood loss may contribute to hemodynamic instability, requiring transfusions and other interventions that could predispose patients to adverse cardiovascular events.
Crystalloid infusion volume was another significant predictor of MACCE-90 (HR: 1.29, 95% CI: 1.00–1.65, p = 0.05). Similar to EBL, the numerical HR value is small because fluid administration volumes are generally large. The mean crystalloid infusion volume was 486 mL in the non-MACCE group and 498 mL in the MACCE group. A higher volume of crystalloid infusion may reflect attempts to maintain intravascular volume during surgery, but excessive administration could contribute to fluid overload, hemodilution, and increased cardiac stress, which may elevate the risk of postoperative cardiovascular complications.
Other variables, including albumin level (HR: 0.84, 95% CI: 0.58–1.25, p = 0.37) and colloid infusion (HR: 1.25, 95% CI: 0.65–2.40, p = 0.53), did not show statistically significant associations with MACCE-90 in the multivariable analysis.
These findings suggest that male sex, higher EBL, and increased crystalloid infusion volume are independent predictors of MACCE-90 risk in this cohort.
Discussion
To clarify the clinical relevance of the associations between EBL and crystalloid infusion volume with the risk of MACCE-90, it is necessary to translate the HRs into practical clinical scenarios. Multivariate analysis revealed that the HR for EBL was 1.19 per liter (95% CI: 1.03–1.37, p = 0.02), meaning each additional 1 L of blood loss was associated with an approximate 19% increase in MACCE-90 risk. Furthermore, the 208 mL difference in EBL between the MACCE group (565 mL) and the non-MACCE group (357 mL) could lead to an additional approximately 3.5% increase in risk. For crystalloid infusion volume, the HR was 1.29 per liter (95% CI: 1.00–1.65, p = 0.05), with each additional 1 L of infusion linked to a roughly 29% increase in risk. Although the mean difference in crystalloid volume between the MACCE group (498 mL) and the non-MACCE group (486 mL) was only 12 mL (resulting in a 0.3% risk increase), an infusion volume of 2000–3000 mL in complex surgeries would raise the HR to approximately 1.64 (a 64% risk increase).
In this study, male patients were found to exhibit a significantly increased risk of MACCE-90, with a HR of 1.88 (95% confidence interval [CI]: 1.27–2.78, p < 0.01). The robust association of male sex with MACCE-90 in the multivariate model underscores its status as an independent risk factor. This disparity likely stems from multifactorial origins, including hormonal influences such as androgen-mediated effects on cardiovascular function [19]. Additionally, men may have a higher baseline prevalence of cardiovascular risk factors, and anatomical differences between sexes could further contribute to divergent surgical responses, elevating MACCE risk in male patients [11, 20, 21].
EBL also emerged as a significant predictor of MACCE-90, with an HR of 1.19 (95% CI: 1.03–1.37, p = 0.02) in the multivariate analysis. Though the HR appears modest, its clinical relevance is substantial given the considerable blood volumes often lost during surgery. For instance, the mean EBL was 357 mL in the non-MACCE group compared to 565 mL in the MACCE group. The link between increased EBL and MACCE-90 likely reflects hemodynamic instability precipitated by hemorrhage, often necessitating interventions such as blood transfusions, which carry risks including transfusion-related acute lung injury and infection [22–24]. Furthermore, diminished oxygen delivery due to blood loss may induce tissue hypoxia, placing additional strain on the cardiovascular system and heightening the likelihood of adverse events [25, 26].
Crystalloid infusion volume was similarly associated with MACCE-90, yielding an HR of 1.29 (95% CI: 1.00–1.65, p = 0.05) in the multivariate analysis. The mean crystalloid infusion volume was 486 mL in the non-MACCE group versus 498 mL in the MACCE group. While crystalloids are routinely administered to maintain intravascular volume during surgery, excessive use may lead to fluid overload, resulting in hemodilution that impairs oxygen-carrying capacity and increases cardiac workload due to elevated circulating volume. These factors collectively contribute to a heightened risk of postoperative cardiovascular complications [27–29].
When contextualized within the broader literature on MACCE predictors in surgical patients, our results exhibit both alignment and innovation. Previous studies have similarly identified male sex as a risk factor for postoperative cardiovascular events, corroborating our findings [2, 30–32]. However, while prior research has linked EBL and crystalloid infusion to postoperative complications, their specific associations with MACCE in hepatobiliary and pancreatic surgeries have remained underexplored. By leveraging a large, well-characterized cohort from the INSPIRE database and employing robust statistical methodologies, this study addresses this gap, offering nuanced and precise insights into these predictors within this surgical population [33–35].
The clinical implications of these findings are important. The significant associations between EBL, crystalloid infusion, and MACCE-90 underscore the critical need to optimize intraoperative blood and fluid management. Surgeons and anesthesiologists should prioritize strategies to minimize blood loss, such as meticulous hemostatic techniques, and judiciously evaluate transfusion thresholds when hemorrhage occurs. Likewise, a balanced approach to fluid administration is essential to prevent both over- and under-hydration, with monitoring of parameters such as central venous pressure and urine output guiding individualized therapy in accordance with evidence-based guidelines [36, 37].
Moreover, these results highlight the necessity of revising perioperative protocols to mitigate MACCE risk. Updated guidelines should incorporate stringent standards for blood and fluid management, alongside enhanced preoperative risk stratification based on identified predictors. For example, patients with anticipated high blood loss or male patients could be flagged preoperatively for tailored management plans, including intensified monitoring and preventive measures. The pronounced sex-based disparity in MACCE-90 risk further emphasizes the value of integrating gender-specific considerations into risk assessment frameworks, enabling personalized care that optimizes outcomes through targeted monitoring and interventions [3, 38, 39].
An observation of this study is the lack of significant predictors for MACCE-30, which requires addressing three methodological concerns using the study’s design and data: first, post-hoc power analysis for the MACCE-30 model confirms no sample size inadequacy—with a large retrospective cohort of 6135 hepatobiliary and pancreatic surgery patients (substantially larger than the typical 1000–3000 in perioperative MACCE studies), sufficient statistical power is ensured even if MACCE-30 events are fewer than MACCE-90’s 179 cases, and the same rigorous statistical protocol as MACCE-90 further rules out sample size as a barrier; second, clinically, early (≤ 30 days) MACCE is more likely influenced by short-term fluctuations during the surgical stress period (e.g., transient infections, acute physiological stress responses, or unmeasured acute variables like short-term medication adherence changes or transient hemodynamic instability not captured in the INSPIRE database), resulting in inherent randomness—unlike MACCE-90, which is predictable via cumulative, modifiable intraoperative factors (male sex, higher EBL, increased crystalloid volume), MACCE-30 shows no associations with intraoperative or patient-related variables, supporting this randomness; third, the sampling strategy (based on the INSPIRE database’s hepatobiliary and pancreatic surgery cohort) has no bias toward late-onset events: patient selection followed clear pre-specified criteria (including surgeries via ICD-10-PCS codes, excluding multiple-surgery patients to avoid confounding) without relying on MACCE onset timing, and the INSPIRE database provides comprehensive perioperative data (detailed intraoperative records, complete 90-day follow-up for both MACCE-30 and MACCE-90) with systematic event recording (via ICD-10-CM coding) ensuring no undercounting of early events. In summary, the absence of MACCE-30 predictors stems not from methodological flaws but from early MACCE’s inherent randomness (driven by short-term surgical stress fluctuations) and the unbiased sampling strategy, highlighting the clinical relevance of focusing on modifiable MACCE-90 predictors (male sex, EBL, crystalloid volume) for perioperative risk stratification.
This study’s strengths including its use of the high-quality INSPIRE database, which provides comprehensive perioperative data on a large patient cohort, ensuring a robust foundation for analysis. The application of sophisticated statistical techniques—such as Pearson’s Chi-squared test, Fisher’s exact test, one-way ANOVA, Wilcoxon rank sum test, Kruskal-Wallis rank sum test, and Cox regression models—enhances the reliability and validity of the findings. Furthermore, its focused examination of hepatobiliary and pancreatic surgeries facilitates a detailed exploration of the interplay between intraoperative factors and MACCE in this specific context.
Nonetheless, limitations must be acknowledged. The retrospective design introduces potential selection bias, as reliance on historical data may not fully represent the broader population. Unmeasured confounders could also influence the observed associations, and the generalizability of these findings to other surgical disciplines remains uncertain due to variations in physiological and procedural characteristics [40, 41]. Fourth, due to database limitations, this study could not further subdivide surgical types, making it impossible to incorporate information about the surgical procedures themselves into the model, which has affected the model’s generalizability.
Looking forward, prospective validation of these findings across diverse surgical populations is a priority to confirm their broader applicability and identify additional relevant predictors. Furthermore, research into interventions targeting these risk factors—such as advanced blood conservation techniques or optimized fluid management protocols—warrants investigation through clinical trials [42, 43]. Additionally, exploring sex-specific perioperative strategies could yield tailored approaches to preoperative optimization, intraoperative care, and postoperative management, potentially reducing MACCE incidence and improving outcomes.
In conclusion, this study successfully delineates three key predictors of MACCE-90 risk in hepatobiliary and pancreatic surgeries, offering valuable insights for clinical practice. However, further research is essential to elucidate underlying mechanisms and develop effective risk-reduction strategies, ultimately advancing perioperative care and enhancing outcomes for this high-risk patient population.
Conclusion
In conclusion, this study successfully delineates three key predictors of MACCE-90 risk in hepatobiliary and pancreatic surgeries, offering valuable insights for clinical practice. However, further research is essential to elucidate underlying mechanisms and develop effective risk-reduction strategies, ultimately advancing perioperative care and enhancing outcomes for this high-risk patient population.
Supplementary Information
Acknowledgements
Not applicable.
Clinical trial number: Not applicable.
Abbreviations
- CI
confidence intervals
- CRP
C-reactive protein
- EBL
estimated blood loss
- HR
Hazard Ratio
- IQR
interquartile ranges
- MACCE
Major adverse cardiovascular and cerebrovascular events
- MACCE-30/90
MACCE occurring within 30/90 days
- RCRI
Revised Cardiac Risk Index
- SNUH
Seoul National University Hospital
- VIF
Variance Inflation Factor
Authors’ contributions
DW & FL were responsible for conceptualization, data curation, writing-original draft. PY，YY，YL & SL were responsible for formal analysis, methodology, validation. ZH，TJ，CH & WC were responsible for investigation, writing- review & editing. XW was responsible for supervision, visualization, writing-review & editing. All authors read and approved the final manuscript.
Funding
Not applicable.
Data availability
All data generated or analysed during this study are included in this published article and its supplementary information files.
Declarations
Ethics approval and consent to participate
Not applicable.
Consent for publication
Not applicable.
Competing interests
The authors declare no competing interests.
Footnotes
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Contributor Information
Dongxu Wu, Email: wdx444613055@163.com.
Xujie Wang, Email: 13566560137@163.com.
References
- 1.Nathaniel RS, Navdeep G, Harish R, Yu G, Jeffrey SB, Sripal B. Perioperative major adverse cardiovascular and cerebrovascular events associated with noncardiac surgery. JAMA Cardiol. 2016;2(2):181–187. [DOI] [PMC free article] [PubMed]
- 2.Venkataraghavan R, Kelvin C, Sandeep A, Zhenwei Z, Md Ashfaq A, Peter M et al. Prevalence and trends of perioperative major adverse cardiovascular and cerebrovascular events during cancer surgeries. Sci Rep. 2023;13(1):2410. [DOI] [PMC free article] [PubMed]
- 3.Aman M, Stephen E, Thien Htay O, Jeffery M, Michael G, Jamie A et al. Development and validation of a machine learning model to identify patients before surgery at high risk for postoperative adverse events. JAMA Netw Open. 2023;6(7):e2322285. [DOI] [PMC free article] [PubMed]
- 4.Alina B, Agnes SM, Moritz F, Linda K, Eik V, Antonia Z et al. Effect of personalized perioperative blood pressure management on postoperative complications and mortality in high-risk patients having major abdominal surgery: protocol for a multicenter randomized trial (IMPROVE-multi). Trials. 2022;23(1):946. [DOI] [PMC free article] [PubMed]
- 5.Kieran C, Pádraig ÓS, Jeroen H, Donal JB. Current practice in the perioperative management of patients with diabetes mellitus: a narrative review. Br J Anaesth. 2023;131(2):242–252. [DOI] [PMC free article] [PubMed]
- 6.Fausto A, Irene M, Marta B, Simona B, Valentina M, Vittorio C et al. Improving cardiovascular prevention in general practice: results of a comprehensive personalized strategy in subjects at high risk. Eur J Prev Cardiol. 2015;23(9):947–55. [DOI] [PubMed]
- 7.Yoshiko T, Yoshio T, Masami F, Sumiyo T, Kaoru I, Jiro F et al. Risk factors for surgical site infection after major hepatobiliary and pancreatic surgery. J Infect Chemother. 2018;24(9):739–43. [DOI] [PubMed]
- 8.Jia Z, Ying-Hui S, Yu-Chen Q, Ou L, Guo-Yi X, Meng-Jun M et al. Full laparoscopic anatomical liver segment VII resection with preferred glissonean pedicle and dorsal hepatic approach. Hepatobiliary Pancreat Dis Int. 2023;22(5):532–36. [DOI] [PubMed]
- 9.Jong Hun K, Seog Ki M, Huisong L, Geun H, Hyeon Kook L. The safety and risk factors of major hepatobiliary pancreatic surgery in patients older than 80 years. Ann Surg Treat Res. 2016;91(6):288–94. [DOI] [PMC free article] [PubMed]
- 10.Lisa D, Shazia A, Natalia C, René MP, Saif Z, Carolin H et al. Risk prediction of bleeding and MACCE by PRECISE-DAPT score post-PCI. Int J Cardiol Heart Vasc. 2021;33(0):100750. [DOI] [PMC free article] [PubMed]
- 11.Xiao W, Jingyao F, Ruifeng G, Wen H, Wei G, Yan Y et al. Association of obstructive sleep Apnoea with cardiovascular events in women and men with acute coronary syndrome. Eur Respir J. 2022;61(1):2201110. [DOI] [PMC free article] [PubMed]
- 12.Virginia MM. Universality of sex differences in cardiovascular outcomes: where do we go from here? Eur Heart J. 2020;41(17):1697–99. [DOI] [PMC free article] [PubMed]
- 13.Abdel AA-R. Influence of sex on cardiovascular drug responses: role of Estrogen. Curr Opin Pharmacol. 2017;33(0):1–5. [DOI] [PMC free article] [PubMed]
- 14.Jeremy WC. Hemorrhagic shock. N Engl J Med. 2018;378(4):370–79. [DOI] [PubMed]
- 15.Vijaya PP, Bindiya GS. Fluid overload and acute kidney injury. Indian J Crit Care Med. 2020;24(0):S94-S97. [DOI] [PMC free article] [PubMed]
- 16.So I, Masanobu I, Yasuhiro O, Taishi N, Kenichi T, Tetsuya M et al. Impact of heart failure severity and major bleeding events after percutaneous coronary intervention on subsequent major adverse cardiac events. Int J Cardiol Cardiovasc Risk Prev. 2023;18(0):200193. [DOI] [PMC free article] [PubMed]
- 17.Lim L, Lee H, Jung C, Sim D, Borrat X, Pollard T, et al. INSPIRE, a publicly available research dataset for perioperative medicine. Sci Data. 2024;11(1):655. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Rosas-Salazar C, Chirkova T, Gebretsadik T, Chappell J, Peebles R, Dupont W, et al. Respiratory syncytial virus infection during infancy and asthma during childhood in the USA (INSPIRE): a population-based, prospective birth cohort study. Lancet (London England). 2023;401(10389):1669–80. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Cai J, Wen J, Jiang W, Lin J, Hong Y, Zhu Y. Androgen actions on endothelium functions and cardiovascular diseases. J Geriatr Cardiol. 2016;13(2):183–96. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Nyström A, Strömberg S, Jansson K, Faresjö Å, Faresjö T. Cardiovascular risks before myocardial infarction differences between men and women. BMC Cardiovasc Disord. 2022;22(1):110. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Lloyd-Jones D, Hong Y, Labarthe D, Mozaffarian D, Appel L, Van Horn L, et al. Defining and setting National goals for cardiovascular health promotion and disease reduction: the American heart association’s strategic impact goal through 2020 and beyond. Circulation. 2010;121(4):586–613. [DOI] [PubMed] [Google Scholar]
- 22.Vlaar A, Juffermans N. Transfusion-related acute lung injury: a clinical review. Lancet (London England). 2013;382(9896):984–94. [DOI] [PubMed] [Google Scholar]
- 23.Timler D, Klepaczka J, Kasielska-Trojan A, Bogusiak K. Analysis of complications after blood components’ transfusions. Pol Przegl Chir. 2015;87(4):166–73. [DOI] [PubMed] [Google Scholar]
- 24.Sevransky J. Clinical assessment of hemodynamically unstable patients. Curr Opin Crit Care. 2009;15(3):234–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Sen Gupta A. Hemoglobin-based oxygen carriers: current state-of-the-art and novel molecules. Shock. 2019;52:70–83. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Alayash A. Setbacks in blood substitutes research and development: a biochemical perspective. Clin Lab Med. 2010;30(2):381–9. [DOI] [PubMed] [Google Scholar]
- 27.Sigurdsson G. Perioperative fluid management in microvascular surgery. J Reconstr Microsurg. 1995;11(1):57–65. [DOI] [PubMed] [Google Scholar]
- 28.Lobo D, Macafee D, Allison S. How perioperative fluid balance influences postoperative outcomes. Best Pract Res Clin Anaesthesiol. 2006;20(3):439–55. [DOI] [PubMed] [Google Scholar]
- 29.Lee E. Optimal fluid therapy for thoracic surgery. J Thorac Dis. 2019;11(5):1753–5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Padhy S, Kar P, Ramachandran G. Prediction of major adverse cardiovascular and cerebrovascular events (MACCE) after thoracic surgery- the role of estimated GFR. J Clin Diagn Res. 2017;11(9):UC13-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Budtz-Lilly J, Petersen C, Pedersen T, Eldrup N. Male sex associated with increased long-term cardiovascular mortality after peripheral vascular surgery for atherosclerosis despite optimal medical treatment. Eur J Vasc Endovasc Surg. 2015;50(6):767–73. [DOI] [PubMed] [Google Scholar]
- 32.Wu X, Hu M, Zhang J, Li K, Yang X. A new predictive model for in-hospital major adverse cardiac and cerebrovascular events in Chinese patients after major noncardiac surgery. Am J Cardiol. 2023;186:196–202. [DOI] [PubMed] [Google Scholar]
- 33.Mases A, de Beltrán Heredia S, Gallart L, Román L, Bosch L, Núñez M, et al. Prediction of acute myocardial injury in noncardiac surgery in patients at risk for major adverse cardiovascular and cerebrovascular events: a multivariable risk model. Anesth Analg. 2023;137(6):1116–26. [DOI] [PubMed] [Google Scholar]
- 34.Iwasaki Y, Ono Y, Inokuchi R, Ishida T, Kumada Y, Shinohara K. Intraoperative fluid management in hepato-biliary-pancreatic operation using stroke volume variation monitoring: a single-center, open-label, randomized pilot study. Med. 2020;99(50):e23617. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Morris F, Åhman R, Craswell A, Didriksson H, Jonsson C, Gisselgård M, et al. Effect of perioperative blood transfusion on preoperative haemoglobin levels as a risk factor for long-term outcomes in patients undergoing major noncardiac surgery: a prospective multicentre observational study. Br J Anaesth. 2024;133(6):1183–91. [DOI] [PubMed] [Google Scholar]
- 36.Elia J, Diwan M, Deshpande R, Brainard J, Karamchandani K. Perioperative fluid management and volume assessment. Anesthesiol Clin. 2023;41(1):191–209. [DOI] [PubMed] [Google Scholar]
- 37.Perner A, Haase N, Guttormsen A, Tenhunen J, Klemenzson G, Åneman A, et al. Hydroxyethyl starch 130/0.42 versus Ringer’s acetate in severe sepsis. N Engl J Med. 2012;367(2):124–34. [DOI] [PubMed] [Google Scholar]
- 38.Hung L, Alshareef A, Al-Ahdal T, Anh P, Huan D, Van Trang D, et al. Predicting atrial fibrillation after cardiac surgery using a simplified risk index. J Electrocardiol. 2021;67:45–9. [DOI] [PubMed] [Google Scholar]
- 39.Sun Y, Feng L, Li X, Gao R, Wu Y. The sex difference in 6-month maces and its explaining variables in acute myocardial infarction survivors: data from CPACS-3 study. Int J Cardiol. 2020;311:1–6. [DOI] [PubMed] [Google Scholar]
- 40.Tofthagen C. Threats to validity in retrospective studies. J Adv Pract Oncol. 2012;3(3):181–3. [PMC free article] [PubMed] [Google Scholar]
- 41.Papis D, Vagliasindi A, Maida P. Hepatobiliary and pancreatic surgery in the elderly: current status. Ann Hepatobiliary Pancreat Surg. 2020;24(1):1–5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42.Kluyts H, le Manach Y, Munlemvo D, Madzimbamuto F, Basenero A, Coulibaly Y, et al. The ASOS surgical risk calculator: development and validation of a tool for identifying African surgical patients at risk of severe postoperative complications. Br J Anaesth. 2018;121(6):1357–63. [DOI] [PubMed] [Google Scholar]
- 43.Simmons J, Dobyns J, Paiste J. Enhanced recovery after surgery: intraoperative fluid management strategies. Surg Clin North Am. 2018;98(6):1185–200. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
All data generated or analysed during this study are included in this published article and its supplementary information files.