The value of Age-adjusted Charlson and Elixhauser-Van Walraven comorbidity index in predicting prognosis for patients undergoing heart valve surgery

Abstract

Objective

This study aims to compare the efficacy of the Age-adjusted Charlson Comorbidity Index (ACCI) and the Elixhauser-Van Walraven Comorbidity Index (ECI-VW) in predicting mortality risk among patients undergoing heart valve surgery.

Methods

Clinical data were extracted from the INSPIRE Database using R language. The Receiver Operating Characteristic (ROC) Curve was employed to assess the predictive accuracy of ACCI and ECI-VW for in-hospital all-cause mortality and post-surgical all-cause mortality at 7 and 28 days. Subgroup analysis was conducted to validate the application efficacy, and the optimal cutoff value was identified.

Results

The study included 996 patients, with 931 survivors and 65 cases of in-hospital all-cause mortality. The area under the curve (AUC) for ACCI in predicting in-hospital all-cause mortality was 0.658 (95% CI: 0.584, 0.732), while the AUC for ECI-vw in predicting the same outcome was 0.663 (95% CI: 0.584, 0.741). For predicting all-cause mortality within 7 days post-surgery, the AUC of ACCI was 0.680 (95% CI: 0.04, 0.56), and for ECI-vw, it was 0.532 (95% CI: 0.353, 0.712). Regarding the prediction of all-cause mortality within 28 days after surgery, the AUC for ACCI was 0.724 (95% CI: 0.622, 0.827), and for ECI-vw, it was 0.653 (95% CI: 0.538, 0.69). Patients were categorized into two groups based on the ACCI cutoff value of 3.5, including Group 1 (ACCI < 3.5 points, 823 cases) and Group 2 (ACCI > 3.5 points, 173 cases). The overall survival rate for these two patient groups was calculated using the Kaplan-Meier method, revealing that the 28-day postoperative survival rate for patients in Group 1 was significantly higher than that for patients in Group 2 (P < 0.0001).

Conclusions

ACCI demonstrates significant predictive value for in-hospital all-cause mortality within 28 days following cardiac valve disease surgery. Patients presenting with an ACCI greater than 3.5 exhibit an increased risk of mortality within 28 days post-surgery compared to those with an ACCI less than 3.5. This finding suggests that the ACCI can serve as a preliminary tool for assessing the prognosis of patients undergoing this type of surgical intervention.

Introduction

Valvular heart disease (VHD) is a significant contributor to the decline in physical function, quality of life [1], and life expectancy, and it is a primary factor behind the increasing incidence and mortality rates of cardiovascular diseases globally [2]. Cardiac valve surgery plays a crucial role in enhancing patients’ cardiac function and overall prognosis [3].

The assessment of perioperative risks for individuals undergoing cardiac valve surgery is a multifaceted process that requires consideration of various factors. The Age-adjusted Charlson Comorbidity Index (ACCI) and the Elixhauser-Van Walraven Comorbidity Index (ECI-VW) are widely utilized tools in this context. This study aims to investigate the association between ACCI and ECI-VW scores and in-hospital all-cause mortality among patients with valvular heart disease post-surgery. Our goal is to evaluate the utility of these two scoring instruments in the prognostic assessment of patients following valvular heart disease surgery, providing clinicians with a more precise basis for risk assessment, enhancing perioperative patient management, and improving postoperative survival rates and quality of life.

Methods

Data extraction

Demographic and clinical data of patients undergoing cardiac valve surgery were extracted from the INSPIRE database [4] using R language. This included information on in-hospital all-cause mortality, all-cause mortality within 7 and 28 days post-surgery, the use of extracorporeal membrane oxygenation (ECMO), Intra-aortic Balloon Pump (IABP) assistance, and continuous renal replacement therapy (CRRT), among other indicators. Data with major variables missing, duplicate records, or illogical entries were excluded, resulting in a final effective sample size of 996. The ACCI and ECI-VW scores were calculated directly from the visual representations in the INSPIRE database using R language.

Statistical analysis

Statistical analysis was performed using the SPSS 29.0 software package. The Receiver Operating Characteristic (ROC) curve was employed to evaluate the predictive efficacy of ACCI and ECI-vw for in-hospital all-cause mortality, all-cause mortality within 7 days post-surgery, and all-cause mortality within 28 days post-surgery. The method demonstrating superior performance was selected, and the optimal cutoff value was identified at the point of the maximum Youden index, ultimately leading to the generation and analysis of survival curves for two groups of patients. For normally distributed measurement data, values were presented as mean ± standard deviation and compared using an independent samples t-test. For non-normally distributed quantitative data, the interquartile range [median (P25, P75)] was used for description, with non-parametric tests applied for comparisons. Count data were presented as numbers (%) and compared using the χ [2] test or Fisher’s exact test. The overall survival rates of the two groups of patients were calculated using the Kaplan-Meier method. A P-value < 0.05 was considered statistically significant.

Results

Basic characteristics of population

By extracting information on patients undergoing cardiac valve surgery from the INSPIRE database using R language, this study included a total of 996 patients (Fig. 1), of which 540 were male (54.22%) and 456 were female (45.78%). The median age was 65 years (interquartile range [IQR], 55 to 75 years). Aortic valve surgery was performed in 603 cases (60.54%), mitral valve surgery in 340 cases (34.14%), and tricuspid valve surgery in 53 cases (5.32%). During their hospital stay, 24 patients (2.41%) received ECMO support, 24 patients (2.41%) received IABP support, and 79 patients (7.93%) underwent CRRT. There were 931 survivors (93.47%), while in-hospital all-cause mortality was observed in 65 patients (6.53%), including 13 deaths (1.31%) within 7 days post-surgery and 32 deaths (3.21%) within 28 days post-surgery (Table 1).

Fig. 1.

Patient selection flow chart

Table 1.

Baseline characteristics of dosage data (overall and subgroups based on ACCI values)

Variables	All sample	group1(ACCI < 3.5) n = 823	group2(ACCI > 3.5) n = 173	Statistical metrics	P-value
Age	65(55,75)	60(50,70)	80(70,80)	16.841	< 0.001
Weight	60(51,68)	60(51,68)	57(49,63)	-3.703	< 0.001
Height	160(154,167)	162(154,168)	159(153,165)	-2.578	0.01
Operating Room Time	445(370,530)	445(370,535)	440(360,517.5)	-1.423	0.155
Surgical Duration	357.5(285,440)	360(290,445)	350(275,430)	-1.551	0.121
Hospital Stay	25,915(18715,40315)	24,475(18715,35995)	34,555(25915,59035)	7.169	< 0.001
Anesthesia Duration	425(350,510)	425(355,510)	420(337.5,490)	-1.42	0.156
ICU Stay	2920(1551,5844)	2830(1520,5605)	5190(2805,8747.5)	6.101	< 0.001
Temperature	35.4(34.6,35.6)	35.4(34.6,35.6)	35.2(34.6,35.5)	-1.233	0.217
CI	2.3(2,2.7)	2.35(2,2.7)	2.2(1.99,2.6)	-1.528	0.126
cvp	6(2,7)	6(4,7)	5(2,7)	-0.537	0.592
Estimated Blood Loss	200(125,400)	225(150,400)	150(100,325)	-2.899	0.004
Albumin	3.375(3.1,3.55)	3.4(3.2,3.6)	3.2(3,3.4)	-6.467	< 0.001
ALT	22(17,35)	23(17,35)	22(15,30)	-1.609	0.108
APTT	31.2(28.5,33.9)	31(28.5,33.8)	31.8(29.6,36.2)	3.316	< 0.001
AST	49(35,65)	49(35,65)	40(30.625,65)	-3.239	0.001
CK-MB	19.5(7.6,42.1)	19.5(8.45,42.1)	11.3(5.6,30.8)	-4.905	< 0.001
Creatinine	0.89(0.72,1.16)	0.85(0.72,1.06)	1.16(0.85,1.64)	7.635	< 0.001
CRP	6.81(3.72,9.45)	6.81(3.72,9.39)	7.58(4.6,11.86)	3.666	< 0.001
Fibrinogen	274(218.5,343)	270(214,343)	277(232.5,360)	1.212	0.226
Hb	10.7(9.9,11.2)	10.7(10.05,11.4)	10.25(9.7,10.9)	-4.303	< 0.001
PLT	114(90,134)	114(90,134)	90(75.5,114)	-5.104	< 0.001
INR	1.21(1.13,1.37)	1.21(1.145,1.37)	1.21(1.115,1.285)	-2.99	0.003
TB	1.4(1,1.9)	1.3(1,1.9)	1.4(1.1,1.9)	1.288	0.198
Troponin I	7.29(2.61,15.6)	7.29(3.07,15.6)	5.59(1.79,15.6)	-2.052	0.04
WBC	10.26(8.81,12.98)	10.26(8.86,12.98)	10.19(7.86,13.78)	-0.938	0.348
Lac	2.05(1.4,3)	2.2(1.4,3)	1.9(1.3,2.5)	-2.654	0.008
Troponin T	0.694 (0.330,0.694)	0.694 (0.421,0.694)	0.33(0.33,0.33)	-1.066	0.286

The ROC curve was utilized to assess the predictive performance of ACCI and ECI-vw for in-hospital all-cause mortality, as well as mortality within 7 and 28 days post-surgery. The area under the curve (AUC) for ACCI in predicting in-hospital all-cause mortality was 0.658 (95% CI: 0.584, 0.732), while the AUC for ECI-vw was 0.663 (95% CI: 0.584, 0.741), as displayed in Fig. 2. The AUC for ACCI in predicting all-cause mortality within 7 days post-surgery was 0.680 (95% CI: 0.04, 0.56), and for ECI-vw, it was 0.532 (95% CI: 0.353, 0.712), as displayed in Fig. 3. For predicting all-cause mortality within 28 days post-surgery, the AUC for ACCI was 0.724 (95% CI: 0.622, 0.827), and for ECI-vw, it was 0.653 (95% CI: 0.538, 0.69), as displayed in Fig. 4.

Fig. 2.

The ACCI and Elixhauser-vw were utilized to conduct ROC analyses for all-cause mortality during hospitalization post-operation

Fig. 3.

The ACCI and Elixhauser-vw were utilized to conduct ROC analyses for all-cause mortality within 7 days post-operation

Fig. 4.

The ACCI and Elixhauser-vw were utilized to conduct ROC analyses for all-cause mortality within 28 days post-operation

Further analysis of ACCI for predicting all-cause mortality within 28 days post-surgery identified an optimal cutoff value of 3.5 based on the maximum Youden index, corresponding to a sensitivity of approximately 56.3% and a specificity of approximately 83.9%. Patients were divided into two groups using the ACCI cutoff value of 3.5, with Group 1 (ACCI < 3.5, 823 cases) and Group 2 (ACCI > 3.5, 173 cases). Analysis of demographic and disease characteristics between the two groups revealed that the in-hospital all-cause mortality rate for Group 1 was 4.62% (38/823), significantly lower than the 15.61% (27/173) observed in Group 2 (P < 0.001). Patients in Group 2 were older and had a higher proportion of anesthesia ASA grades III-IV (P < 0.05), as well as higher rates of ECMO support, IABP support, and CRRT (P < 0.05). Additionally, Group 2 patients had longer hospital stays, cardiopulmonary bypass times, and ICU stays compared to Group 1 (P < 0.05), as displayed in Tables 1 and 2.

Table 2.

Baseline characteristics of quantitative data (subgroups based on ACCI values)

Variables		group1(ACCI < 3.5) n = 823		group2(ACCI > 3.5) n = 173		Statistical metrics		P-value
Gender						2.935		0.087
Male		436(52.98%)		104(60.12%)
Female		387(47.02%)		69(39.88%)
Asa						13.282		0.004
1		16(1.95%)		3(1.76%)
2		218(26.55%)		29(17.06%)
3		550(66.99%)		121(71.18%)
4		37(4.51%)		17(10%)
Emergency Surgery						2.302		0.129
0		769(93.44%)		156(90.17%)
1		54(6.56%)		17(9.83%)
In-hospital Mortality						28.273		< 0.001
0		785(99.95%)		146(84.39%)
1		38(0.05%)		27(15.61%)
7-day mortality						12.199		< 0.001
0		817(99.27%)		166(95.95%)
1		6(0.73%)		7(4.05%)
IABP						10.812		0.001
0		779(94.65%)		152(87.86%)
1		44(5.35%)		21(12.14%)
CRRT						43.371		< 0.001
0		779(94.65%)		138(79.77%)
1		44(5.35%)		35(20.23%)
ECMO						6.944		0.008
0		808(98.18%)		164(94.8%)
1		15(1.82%)		9(5.2%)
Reoperation						0.133		0.715
0		816(99.15%)		172(99.42%)
1		7(0.85%)		1(0.58%)

The overall survival rates of the two groups were calculated using the Kaplan-Meier method, indicating that the 28-day postoperative survival rate for Group 1 patients was significantly higher than that for Group 2 patients (P < 0.0001), as displayed in Fig. 5.

Fig. 5.

28 days Kaplan-Meier curve of the two groups. Patients were divided into two groups based on an ACCI cutoff value of 3.5, with Group 1 scoring less than 3.5 and Group 2 scoring more than 3.5. The overall survival rates of the two groups were calculated using the Kaplan-Meier method. The horizontal axis represents time in days. The results indicated that the 28-day postoperative survival rate of patients in Group 1 was higher than that of patients in Group 2 (P < 0.0001)

Discussion

The Charlson Comorbidity Index (CCI) was initially proposed by Charlson et al. [5]. in 1987 and revised in 1992. It scores various chronic diseases and uses the total score to predict long-term mortality. ACCI is a variant of CCI, incorporating age as a risk factor on top of the original comorbidity index, making it more suitable for risk assessment in elderly patients. Similarly, the Elixhauser Comorbidity Index (ECI) predicts in-hospital mortality based on 30 acute and chronic comorbidities [6]. ECI-vw is a modified index based on ECI. In 2009, Van Walraven et al. [7] transformed the ECI scoring into a weighted scoring system to simplify its use.

Compared to other scoring methods such as the Acute Physiology and Chronic Health Evaluation (APACHE II) and the Sequential Organ Failure Assessment (SOFA), ACCI and ECI-vw can be calculated at the time of patient admission without the need for interpretation of laboratory and bedside clinical data. Therefore, they can be easily extracted from management databases, and their use has been increasing in critical care-related literature. Both CCI and ECI-vw have been widely used to predict survival rates of patients in ICUs [8]. Multiple studies [9–14] have demonstrated the effectiveness of these indices in predicting in-hospital mortality among ICU patients. Recent evidence suggests that ECI-vw may slightly outperform CCI in predicting mortality among hospitalized patients [15–20] but is inferior to ACCI [21].

Although ACCI and ECI-vw have been proven to have predictive value in other types of surgeries, their relevance and predictive capability in the prognosis of patients after cardiac valve surgery have not been fully explored. Moreover, patients with heart valve disease often have multiple comorbidities, such as hypertension, diabetes, and chronic kidney disease, which could affect surgical outcomes and long-term prognosis. Therefore, accurately assessing postoperative in-hospital all-cause mortality risk using ACCI and ECI-vw scores is crucial for improving patient management and enhancing surgical safety.

Our results indicate that both ACCI and ECI-vw could predict the risk of postoperative in-hospital all-cause mortality in patients undergoing cardiac valve surgery. Notably, ACCI shows good predictive performance for all-cause mortality within 28 days post-surgery. The application of these two scoring tools provides important reference information for clinicians in perioperative management. Especially today, with the advent of an aging society and the increasing average age of patients with heart valve disease, along with more complex comorbidity conditions, more precise assessment tools are needed to guide clinical decision-making. The advantage of ACCI lies in its integration of patients’ age with comorbid conditions, offering a more comprehensive perspective for assessing postoperative risks in elderly patients with heart valve disease. In this study, patients undergoing heart valve surgery with an ACCI score greater than 3.5 exhibited a higher risk of mortality within 28 days postoperatively compared to those with an ACCI score less than 3.5, further confirming its applicability in the postoperative prognosis assessment of patients with heart valve disease.

ACCI, as a comprehensive assessment tool that combines the traditional CCI with age factors, is more suitable for assessing the risk in elderly patients. This was validated in our study, where patient groups with higher ACCI scores showed significantly increased risk of postoperative in-hospital all-cause mortality. This may be because elderly patients often have multiple chronic diseases, such as hypertension, diabetes, and chronic kidney disease, which not only affect their overall health condition but may also increase the risk of surgical complications and postoperative complications. Therefore, ACCI holds significant value in assessing risk in elderly patients with heart valve disease.

ECI-vw predicts perioperative risk by assessing the overall health condition of patients. The advantage of ECI-vw lies in its comprehensive consideration of the impact of more chronic diseases and complications on patients, allowing for a more accurate assessment of postoperative risk. Additionally, the universality of ECI-vw enables its application across different types of surgeries, including cardiac valve surgery.

This study offers a new perspective on the postoperative prognosis assessment of patients with heart valve disease, contributing to the advancement of personalized medicine in the field of cardiac surgery. With further research, we hope to improve the accuracy of perioperative risk assessments, provide better medical services to patients, and improve their long-term health outcomes.

Despite the promising predictive capabilities of ACCI and ECI-vw shown in this study, they still have limitations. First, these scoring tools mainly focus on patients’ comorbidities and overall health conditions, without considering specific risk factors related to the surgery itself, such as the duration of surgery, technical difficulty, and intraoperative complications. Therefore, relying solely on ACCI and ECI-vw may not fully assess postoperative risk. In clinical practice, ACCI and ECI-vw should be considered alongside other clinical information for a comprehensive assessment of postoperative risk.

However, we need to be cautious in interpreting our findings, with some limitations. First, the limitations of our study is the inability to compare the Age-Adjusted Charlson Comorbidity Index (ACCI) and the Elixhauser Comorbidity Index-Van Walraven (ECI-VW) with established risk prediction models such as the EuroSCORE and the Society of Thoracic Surgeons (STS) score. This limitation arises due to the lack of comprehensive echocardiographic data, specifically the ejection fraction (EF) values, which are essential for calculating the EuroSCORE and STS score. Consequently, our analysis focuses solely on the predictive utility of ACCI and ECI-VW in the context of our available data.Second, our study is based on the extraction of data from INSPIRE database and has not yet been validated in a clinical setting. Third, although multivariate analysis has controlled for potential confounders as much as possible, there may still be other influencing factors not included in the model. For example, patients’ lifestyle habits, nutritional status, and psychological state may also impact postoperative prognosis, but these factors were not considered in this study. Additionally, due to the retrospective nature of this analysis, there is potential for information bias and selection bias.

Future research could consider integrating more surgery-related risk factors, such as the degree of surgical trauma, intraoperative blood management, and perioperative care level, to construct a more comprehensive risk assessment model. Prospective multicenter studies could help validate the results of this study and further explore the applicability of ACCI and ECI-vw among different populations and regions.

Conclusion

This retrospective analysis found that both ACCI and ECI-VW could serve as predictive factors for the risk of postoperative in-hospital all-cause mortality in patients with heart valve disease. ACCI demonstrates superior predictive value for all-cause mortality within 28 days following heart valve surgery. The application of these two scoring tools holds significant guidance for clinicians in conducting risk assessments and developing personalized treatment plans during the perioperative period of valvular heart disease.

Author contributions

X.L. designed the study and wrote the paper. X.L. acquired data and analyzed the data. C.L. interpreted the results. Prof. F. Z. and S. M. polished the manuscript and gave valuable suggestions for revision of the manuscript. All other authors interpreted data and provided critical revision of the manuscript. The final version to be submitted was approved by all the authors.

Funding

The peak supporting clinical discipline of Shanghai health bureau (2023ZDFC0104 to L.T).

Data availability

No datasets were generated or analysed during the current study.

Declarations

Ethical approval

We confirm that ethics approval was not applicable for this study due to its use of publicly available data, in compliance with local and national guidelines.

Human ethics and consent to participate

Not applicable.

Consent for publication

All participating authors agree to publication of the article.

Competing interests

The authors declare no competing interests.