Application of the comprehensive complication index and Clavien–Dindo complication classification in cardiac surgery: a retrospective study

Abstract

Background

Standardized tools are needed to accurately assess the severity of postoperative complications in cardiac surgery, which is critical for patient management and outcome evaluation. This study aims to evaluate the application of the Clavien Dindo Complications Classification (CDCC) and Comprehensive Complications Index (CCI) in cardiac surgery.

Results

563 patients (76.3%) had at least one complication. The mean CCI was 22.5 ± 19.3. Patients who underwent more complex surgeries had higher CDCC grades and CCI scores. Hospitalization cost and length of stay increased with increasing CDCC grade and CCI score. The ACCI (OR: 1.334 [1.109–1.606], p = 0.002) and LVEF (OR: 0.965 [0.943–0.987], p = 0.002) were associated factors for the most severe complications. The CCI may reduce the required sample size by 20.5 times and 19.6 times compared with the most severe complication or major adverse cardiovascular and cerebrovascular events, respectively, as an endpoint.

Conclusion

The CCI and CDCC can help reflect the complexity of cardiac surgery and the burden of hospitalization, providing a more accurate assessment of surgical complications. This may reduce the sample size needed for clinical trials, thus facilitating research in cardiac surgery.

Clinical trial registration

Chinese Clinical Trial Registry (No. ChiCTR2400084925) on May 28th, 2024. https://www.chictr.org.cn/showproj.html?proj=228499.

Introduction

Postoperative complications are among the most common adverse events in cardiac surgery, with some being transient and minor, whereas others are long-term and severe, potentially leading to serious harm. These complications include bleeding from the surgical incision, infection of the surgical incision, perioperative myocardial infarction, poor sternal healing, postpericardiotomy syndrome, arrhythmias, hemodynamic instability, sepsis, respiratory system complications, gastrointestinal damage, stroke, renal failure, neurocognitive dysfunction, and death, among others [1, 2]. However, in the assessment of cardiac surgery complications, there is a lack of a widely accepted classification and grading system for complications, which hinders the correct interpretation of surgical outcome data, preventing the comparison of postoperative outcome data across different periods and institutions. Moreover, most reports on cardiac surgery complications fail to provide information on the severity of complications or report only the most severe complications, neglecting other less severe events [1,2,3]. There is an urgent need for a standardized assessment and reporting tool to correctly interpret information related to postoperative complications in cardiac surgery, to better manage the perioperative period of heart surgery and to accurately assess patient surgical outcomes and prognosis.

The Clavien‒Dindo complication classification (CDCC) is a tool designed to evaluate postoperative complications in surgery by categorizing the clinical interventions required to address these complications into five grades [4]. This system facilitates a more accurate assessment of the severity of surgical complications, thereby guiding appropriate treatment measures. Originally developed on the basis of the outcomes of elective general surgery, the CDCC has been validated in a broad range of general surgical procedures and is also applicable in various surgical fields, including orthopedic surgery, head and neck surgery, thoracic surgery, plastic surgery, gynecological surgery, and emergency surgery [5,6,7,8]. However, the CDCC system considers only the highest-grade complication in the final complication grade for outcome analysis, overlooking other lower-grade complications. This limitation prevents a comprehensive reflection of the overall severity of an individual’s postoperative complications and hinders direct quantitative comparisons between patients with the same highest-grade complications but different lower-grade complications. To address this, the comprehensive complication index (CCI) calculates a score on the basis of each complication’s CDCC grade. The CCI is a scoring system that takes into account all postoperative complications and their severity following surgical procedures, serving as a valuable tool for quantifying and comparing the burden of multiple complications [9]. The CDCC and CCI assessment methods have been widely applied in the evaluation of postoperative complications during surgery, however, only the CDCC has been evaluated in cardiac surgery [10]. Based on the CDCC, CCI may provide more detailed information for the evaluation of postoperative complications in cardiac surgery.

This study retrospectively analyzed patients who underwent cardiac surgery and classified postoperative complications using the CDCC and CCI. This study compared the impact of these two methods on the burden of complications and hospital stay, aiming to provide a standardized approach for the assessment of postoperative complications in the perioperative period of cardiac surgery. It also aims to provide a more comprehensive assessment strategy for the management of postoperative complications, thus contributing to the continuous improvement in the quality of research and practice in the field of cardiac surgery. This study hypothesizes that (a) the complexity of surgery should be correlated with the severity of postoperative complications; (b) the severity of complications is positively correlated with clinically relevant outcomes, such as hospitalization costs, length of stay (LOS) and the postoperative-to-discharge interval; and (c) the CDCC and CCI indices can reduce the sample size required in clinical trials.

Methods

Patient selection

This study retrospectively collected data from patients who underwent cardiac surgery at Sir Run Run Shaw Hospital, School of Medicine, Zhejiang University, from January 2021 to December 2023. The consent to participate was not applicable. This study was approved by the Scientific Ethics Committee of Sir Run Run Shaw Hospital (No. 0257) and registered at the Chinese Clinical Trial Registry (No. ChiCTR2400084925).

The inclusion criteria were as follows: (1) aged ≥ 18 years and (2) underwent cardiac surgery, including coronary artery bypass surgery, valve surgery, and ascending aorta surgery. The exclusion criteria were as follows: (1) heart transplantation, ventricular assist device implantation, aortic dissection surgery, congenital heart surgery, cardiac trauma procedures, pericardiac windows or isolated pericardiectomies; and (2) incomplete medical records (Fig. 1).

Fig. 1

The flow chart of case inclusion study. CABG, coronary artery bypass grafting

Collection of clinical data

Patient demographic data and surgery-related information were recorded. Comorbidities were also assessed with the age-adjusted Charlson comorbidity index (ACCI) [11]. The following biochemical markers before cardiac surgery were also collected: (1) hematocrit level; (2) creatinine level; and (3) left ventricular ejection fraction (LVEF) [12].

Outcome measures

The main outcome measures of this study were postoperative complications, CDCC grade, CCI score, LOS, and hospitalization costs. On the basis of previous research results and reviews, the CDCC grade in cardiac surgery was used to grade the postoperative complications of patients [4, 13]. The CCI was calculated via an online calculation program (see website for details: https://www.cci-calculator.com/cciCalculator). Please refer to Supplementary 1 for a detailed complication grading system for patients.

Statistical methods

On the basis of the results of the normality test, parametric or nonparametric testing was applied in future analyses. Parameter testing was performed on normally distributed data. Otherwise, the Wilcoxon Mann‒Whitney test was used for group comparisons, and the Spearman rank correlation coefficient was used for nonparametric testing. The reference values for the magnitude and intensity of the correlation effect are considered “moderate” (r = 0.40–0.59), “strong” (r = 0.60–0.79), and “very strong” (r = 0.80-1.00) [14]. Owing to the small number of patients with a CDCC grade of V, these patients were combined with those with a grade of IVb for the analysis of postoperative burden indicators.

Univariate linear regression and Binomial multivariate logistic regression models were used to identify the factors associated with the most severe CDC ≥ Grade IIIb complexes. All factors that could influence the most severe complications were tested with simple logistic regression analyses. BMI was categorized as a dichotomous variable according to the presence of obesity in Chinese (< 28 kg/m²/≥ 28 kg/m²) [15]. Factors with p-value < 0.1 in univariate linear regression will be included in the multivariate analysis. Receiver operating characteristic (ROC) curve analyses were used to assess the true-positive (sensitivity) and false-positive rates (1-specificity) of the ACCI and LVEF with respect to the most severe complication binary outcome. The areas under the curve (AUCs) were calculated. The value with the highest sensitivity and specificity was selected as the cutoff value (Youden index J).

The sample size calculation for a fictive future superiority trial was based on the assumption of a 30% reduction in the incidence of complications (yes vs. no) according to the CDCC grade, mortality rate during hospitalization, incidence of major adverse cardiovascular and cerebrovascular events (MACCEs), and incidence of the most severe complications ≥ IIIb. A ten-point reduction in the CCI score reflects clinical relevance [16]. For the sample size calculation of two independence proportions, Two Sample T-Tests Assuming Equal Variance is used, while for two independence means, Tests for Two Proportions is used.

The statistical significance level was set to α < 0.05, with a confidence level of 1-β = 0.8. All the statistical analyses were performed with IBM SPSS Statistics version 26.0 (IBM Corp., Somers, NY, USA) and PASS 15.0.5 (NCSS, LLC, Kaysville, Utah, USA).

Results

Patient characteristics

Data from 738 patients were analyzed. The baseline characteristics are detailed in Table 1.

Table 1 Basic patient information. Means ± SDs (standard deviations) or percentages, as appropriate

Postoperative complications

Overall, 175 patients had no complications, accounting for approximately 23.7% of the patients. Among the patients with documented CDCC complications, 101 were Grade I, 261 were Grade II, 127 were Grade IIIa, 12 were Grade IIIb, 37 were Grade IVa, 22 were Grade IVb, and 3 were Grade V. The frequency of the types of complications and associated treatments exemplified by each grade and system is shown in Table 2.

Table 2 Examples of complications by CDCC and by organ system

CCI

The mean CCI for all patients was 22.5 ± 19.3. Figure 2 shows the CCI and CDCC corresponding to each patient. There was a strong correlation between CDCC and the CCI, with a correlation coefficient of r = 0.954 (p ≤ 0.001).

Fig. 2

Comparison of the highest CDCC (black bars) with the cumulative CCI (red line) for each patient

Complexity of surgery and severity of complications

There were 637 cases of valve surgery (86.3%), accounting for the highest proportion, followed by CABG surgery, with a total of 118 cases (16.0%), including isolated CABG surgery (n = 65, 8.8%), CABG and aortic valve replacement (AVR) surgery (n = 26, 3.5%), CABG and mitral valve (MV) surgery (n = 14, 1.9%), CABG and multiple valve surgery (n = 13, 1.8%), isolated AVR surgery (n = 101, 13.7%), AVR and ascending aorta surgery (n = 33, 4.5%), isolated MV surgery (n = 147, 19.9%), isolated tricuspid valve (TV) surgery (n = 48, 6.5%), multiple valve surgery (n = 251, 34.0%), and other surgeries (n = 40, 5.4%). Figure 3 shows that patients with a relatively more complex surgical type (such as CABG & multiple valve, multiple valve surgery) had a higher CCI score.

Fig. 3

Box-and-whisker plot illustrating the CCI for calculating the quantitative morbidity sustained during postoperative evolution on the basis of the severity and number of complications relative to the type of surgery. AVR, aortic valve replacement; CABG, coronary artery bypass grafting; MV, mitral valve; TV, tricuspid valve

Patients undergoing total sternotomy, cardiopulmonary bypass, aortic occlusion during surgery, or prolonged surgery had a higher risk of postoperative complications. Specifically, patients who underwent complex surgeries had higher CCI scores and CDCC scores (Table 3). However, most severe complications were only related to the surgical approach and operative duration. MACCEs were only positively correlated with the operative duration. At the same time, the older the patient, the lower the preoperative LVEF, the higher the preoperative ACCI, and the higher the risk of postoperative complications. Please refer to the Supplementary 2 for specific content.

Table 3 Relationships between surgical complexity and postoperative complications

Economic burden indicators

The higher the CDCC grade of the patient, the more severe the complications, which may lead to a heavier economic burden during hospitalization. Compared to patients with CDCC grade 0, patients of other grades had higher hospitalization cost, longer LOS and postoperative-to-discharge interval. Among patients assessed as CDCC grades 0 to IVa, the hospitalization costs, LOS, and postoperative-to-discharge interval gradually increased (Fig. 4A-C). However, in CDCC IVb and above grade, there was no significant increase in hospitalization cost, LOS, and postoperative-to-discharge interval. The daily hospitalization costs for all CDCC grades were roughly similar (Fig. 4D).

Fig. 4

Association between patient CDCC grade and postoperative burden. * p < 0.05 compared with patients without complications; * * * p < 0.001 compared with patients without complications; LOS, length of hospital stay

Hospitalization cost and the postoperative-to-discharge interval were moderately positively correlated with the CDCC grader (r = 0.408, p < 0.001; r = 0.428, p < 0.001; Fig. 5A&C). The hospitalization cost, LOS, and postoperative-to-discharge interval were moderately positively correlated with the CCI score (r = 0.464, p < 0.001; r = 0.470, p < 0.001; r = 0.492, p < 0.001; Fig. 5D-F). And the correlation between patient CCI scores and the economic burden during hospitalization seems to be stronger than the CDCC grades.

Fig. 5

Relationships among the CDCC grade, CCI score, and hospitalization burden. ***, p < 0.001; LOS, length of hospital stay

Logistic regression analyses of the most severe complication (CDCC ≥ IIIb)

According to the logistic regression analyses, the preoperative LVEF and ACCI were significant independent predictive factors for the most severe complications during cardiac surgery (Table 4). For every one-point increase in the preoperative ACCI, the risk of the most severe complications after surgery was 1.334 times greater (95% CI: 1.109–1.606). The risk of serious complications after surgery is approximately 1.036 times greater for every one-percent decrease in preoperative LVEF (approximately 1.427 times greater for a 10-percent decrease).

Table 4 Multivariable logistic regression analyses of the associations between possible predictive factors and the presence of the most severe complication (CDCC ≥ IIIb)

We further analyzed the ACCI and LVEF via receiver operating characteristic (ROC) curve analyses to determine a cutoff value to predict the most severe complications (CDCC grade ≥ IIIb). We found an ACCI cutoff value of 2 (78.9% sensitivity, 57.4% specificity) and an LVEF cutoff value of 67.2% (39.9% sensitivity, 81.6% specificity) for accurate prediction of CDCC ≥ Grade IIIb complications (Fig. 6).

Fig. 6

ROC curves: true positive (sensitivity) and false positive (1-specificity) values of the ACCI (AUC = 0.709, 95% CI:0.648–0.770, p = 0.003) and LVEF (AUC = 0.626, 95% CI: 0.562–0.689, p < 0.001) according to the most severe complications (CDCC ≥ IIIb). ACCI, age-adjusted Charlson Comorbidity Index; AUC, area under the curve; LVEF, left ventricular ejection fraction

Sample size calculation

On the basis of previous studies, we calculated the sample size on the basis of the mortality rate during hospitalization, incidence of MACCEs, incidence of complications, incidence of most severe complications (CDCC grade ≥ IIIb), and CCI [16]. Compared to most severe complications or MACCEs, CCI need lower number of event sizes to generate adequate statistical power in sample size estimations by 20.5 times or 19.6 times (Table 5).

Table 5 Sample size calculation and assumptions for each procedure

Discussion

This study validated the significant potential of a new method for evaluating postoperative complications in cardiac surgery patients by applying the CDCC and CCI. The application of these indices has allowed for a detailed stratification of postoperative complications, providing a more granular understanding of the impact of these complications on patient outcomes. This approach is particularly relevant given the complexity of cardiac surgery, where a multitude of factors can influence postoperative recovery.

The CDCC and CCI stand out for their comprehensiveness and sensitivity in the assessment of complications. Traditional measures such as mortality rates or binary morbidity outcomes often fail to capture the full spectrum of postoperative complications, particularly those that do not result in immediate or severe consequences. Various predictive scales, such as the Society of Thoracic Surgeons (STS) score and Euro SCORE II, have been developed to predict complications after cardiac surgery [17]. However, these scores have limitations in capturing the actual occurrence and impact of complications on patients. The CDCC and CCI have been increasingly recognized for their advantages in surgical complication assessment. The CDCC provides a standardized grading system that allows for the classification of complications on the basis of the intervention needed, from minor interventions such as pharmacological treatment to major interventions, including life-support measures or readmission to the intensive care unit. The CCI, on the other hand, offers a composite score that encapsulates the total burden of complications, regardless of their nature or number. These two methods have been widely used in surgical procedures, but their application in patients undergoing cardiac surgery is insufficient. Notably, this holistic approach may be particularly beneficial in cardiac surgery, where patients may experience a range of complications that individually might not be severe enough to significantly impact outcomes but may collectively contribute to a more challenging recovery process.

Compared to CDCC, CCI provide a more nuanced view by accounting for all complications, not just the most serious one. This detailed assessment is crucial for accurately evaluating the quality of surgical care and for identifying areas for improvement in postoperative management. A prospective study reported similar conclusions to those of this study: an increase in CDCC/CCI is associated with greater comorbidities, operative duration, LOS, and surgical complexity, accurately reflecting subtle differences in the postoperative process of adult cardiac surgery [10]. However, this study did not compare CCI and CDCC. And our study is more comprehensive in evaluating postoperative complications, the economic burden of patient hospitalization and the complexity of the operation than other complication indicators, such as in-hospital mortality and MACCE incidence, highlighting the broader application prospects of CDCC and the CCI in cardiac surgery. In addition, the results of our study indicate a moderate correlation between CCI and patients’ hospitalization burden, such as LOS, hospitalization costs, and postoperative hospitalization time, which is rarely reflected in other studies.

The ACCI and LVEF were independent predictive factors for most severe complications. The ACCI has good predictive value for evaluating postoperative complications during radical cystectomy and the survival rate during lung cancer surgery and cardiac arrest [18,19,20]. A retrospective study conducted by Minol JP et al. on minimally invasive mitral valve surgery with 1-year postoperative mortality and early postoperative adverse events as endpoints revealed no significant difference between the predicted values of the ACCI and Euro SCORE II or STS scores [21]. However, it should be noted that patients with multiple illnesses or frailty may benefit from the expansion of objective evaluation parameters. A low preoperative LVEF may make performing the procedure difficult and is associated with a greater risk of mortality and complications after surgery [22, 23]. However, further improvement is needed to enhance the accuracy of the model due to the relatively low AUC.

The implications of the CCI for research are substantial. The CCI can reflect the severity of all complications in a single patient, so the CCI is more sensitive than traditional mortality, incidence rate and other outcome endpoints and can be used as a standardized and widely applicable primary endpoint, thus reducing the sample size required for clinical trials [9]. Among the three randomized controlled trials (pancreatic cancer, colectomy and esophageal cancer surgery) published in Europe, the CCI has greatly improved the sensitivity, treatment effect and sample size calculation of traditional outcome indicators (such as incidence rate and the most serious complications), and the sample size required for the CCI as an outcome indicator is 9 times lower than that required for the traditional incidence rate as an outcome indicator [16]. Its sensitivity in capturing the weighted severity of complications makes it an attractive primary outcome measure for clinical trials. By providing a more accurate reflection of the true burden of postoperative complications, the CCI allows for more efficient sample size calculations, potentially reducing the number of patients required for trials by an order of magnitude. This has significant implications for the feasibility and cost-effectiveness of conducting clinical research in cardiac surgery. Furthermore, the use of the CCI as a primary outcome measure could lead to more robust and reliable conclusions about the efficacy of different surgical techniques or postoperative management strategies.

Previous studies have shown that, owing to the unique death attitudes of local residents, half of the elderly do not choose advanced life support [24]. This may pose challenges in evaluating postoperative complications in cardiac surgery, where patients are assessed as having multiple complications of IVa or IVb grade, and patients and their families may choose to be discharged or return home. If viewed solely from the level of complications or the mortality rate of patients, the severity of complications may be underestimated. The CCI better addresses this issue by weighing the severity of various complications and providing a more accurate and comprehensive assessment of postoperative complications in patients.

Both CDCC and CCI should be used simultaneously in clinical practice. Patients with higher grades of CDCC currently have more severe conditions and poorer short-term and long-term prognosis. However, patients with multiple low CDCC grade complications may have a higher CCI score than those with a single high CDCC grade complication, which may indicate a trend of deterioration and a higher likelihood of developing higher CDCC grade postoperative complications in the later stages. This requires timely clinical intervention to prevent the condition from worsening. In future prospective studies, attention should be paid to recording the time points of different complications to explore the mutual influence relationship between low-grade and high-grade complications, and use CCI and CDCC as the basis for recording and managing patient complications.

Consistent with previous studies, this study did not include patients who underwent aortic dissection surgery and cannot confirm the effectiveness of these methods under these specific conditions [17]. The retrospective design limits our ability to establish causal relationships between surgical complexity and postoperative complications. Additionally, the study was conducted at a single institution, which may limit the generalizability of the findings. Future research should aim to validate our findings in a prospective, multicenter setting with a more diverse patient population. This would help to confirm the utility of the CDCC and CCI across different surgical practices and patient demographics. Furthermore, while the CCI has shown promise in reducing sample sizes for clinical trials, its impact on patient care pathways and healthcare economic outcomes requires further investigation. Future studies should explore how the integration of the CCI into routine clinical practice affects patient management decisions, resource allocation, and, ultimately, patient outcomes. Additionally, the potential for these indices to inform policy and guide the development of clinical guidelines should be explored.

Conclusion

This study demonstrated the utility of the CDCC and CCI in assessing postoperative complications in cardiac surgery patients. These indices offer a more detailed and sensitive approach to assessing complications, which can inform clinical management and guide future research. Despite these limitations, the study’s contributions to the field are significant and highlight the need for ongoing research to further refine these tools and expand their application in cardiac surgery and beyond. The integration of these indices into broader surgical practice could lead to more consistent and standardized outcome measurements and enhanced quality of care across various surgical disciplines.