Association between comorbidity burden and subclinical cardiac dysfunction in cancer patients prior to chemotherapy

Abstract

Background:

Subclinical cardiac dysfunction is a growing concern in cancer patients receiving chemotherapy. Early detection before treatment initiation is crucial to prevent overt cardiotoxicity.

Objective:

This study evaluates the association between comorbidity burden and subclinical cardiac dysfunction, using echocardiographic parameters, in cancer patients prior to chemotherapy.

Method:

This retrospective observational study analysed 110 cancer patients who underwent transthoracic echocardiography before chemotherapy. Comorbidity burden was assessed using the Charlson Comorbidity Index (CCI) and Karnofsky Performance Status (KPS). Echocardiographic evaluation included Global Longitudinal Strain (GLS), the E/e^′ ratio (early transmitral inflow velocity to early diastolic mitral annular velocity) and left atrial volume index (LAVI). Subclinical dysfunction was defined as $\mathrm{\text{GLS}}<18\%$ and/or diastolic dysfunction. Spearman’s rank correlation and multivariable linear regression were performed.

Results:

Abnormal GLS ($<18\%$) was found in 68 patients (61.8%), while 42 patients (38.2%) had normal GLS values. Higher comorbidity burden (CCI) was modestly associated with worse GLS ($\rho =-0.234$, $p=0.014$) and higher E/e^′ ($\rho =0.215$, $p=0.024$), whereas better functional status (Karnofsky score) correlated with more preserved GLS ($\rho =0.217$, $p=0.023$). Higher age was independently associated with higher E/e^′ ($B=0.035$, 95% CI 0.005–0.065, $\beta =0.222$, $p=0.023$), whereas the association between CCI and E/e^′ was of borderline significance ($B=0.186$, 95% CI −0.024–0.396, $\beta =0.168$, $p=0.080$).

Conclusion:

This study showed a significant correlation between comorbidity indices and GLS as well as diastolic function parameters. This suggests that comorbidity burden and functional status are significantly associated with subclinical cardiac abnormalities before chemotherapy.

Introduction

Cardiovascular complications associated with cancer therapy, particularly chemotherapy, have emerged as a significant clinical concern.^1,2 Subclinical cardiac dysfunction can be detected using advanced echocardiographic indices before overt changes in left ventricular ejection fraction occur. Global longitudinal strain (GLS) reflects longitudinal myocardial deformation and is more sensitive than ejection fraction for detecting early systolic impairment. The E/e^′ ratio, defined as the ratio of early transmitral inflow velocity (E) to early diastolic mitral annular velocity (e^′), is commonly used as a non-invasive surrogate of left ventricular filling pressure and diastolic function. In parallel, comorbidity burden and overall clinical status are routinely assessed in oncology using tools such as the Charlson Comorbidity Index (CCI) and the Karnofsky Performance Status (KPS), which capture the cumulative impact of chronic diseases and functional limitations. Subclinical cardiac dysfunction, undetectable by conventional left ventricular ejection fraction (LVEF), often precedes overt cardiotoxicity and contributes to long-term morbidity in cancer survivors.³ With improving cancer survival rates, early identification of cardiac impairment is essential, especially in patients with multiple risk factors such as age and comorbidities.

GLS, assessed by speckle-tracking echocardiography, has shown higher sensitivity for detecting subclinical left ventricular dysfunction compared to LVEF.^4,5 Additionally, left ventricular diastolic dysfunction is often underdiagnosed but may represent an early phase of myocardial injury.⁶ These echocardiographic parameters provide valuable insights into vulnerable cancer populations, particularly the elderly and those with multiple comorbidities.⁷

The CCI is a validated tool to quantify overall comorbidity burden, while the KPS is frequently used to assess functional status in oncology.^8,9 Both have been associated with outcomes in cancer, but their relevance in predicting early cardiac dysfunction remains underexplored. Baseline echocardiographic evaluation, combined with comorbidity stratification, may enhance cardio-oncology risk assessment.

Comorbidity indices such as the CCI and functional scales such as KPS are attractive tools for cardio-oncology risk stratification because they can be derived from routine clinical assessment and have been linked to cardiovascular events and survival in both oncology and cardiology cohorts. In many centres, especially in resource-limited settings, comprehensive objective evaluation of physical fitness (e.g. treadmill exercise testing or cardiopulmonary exercise testing) is not routinely available before chemotherapy. Therefore, understanding how comorbidity burden and functional status relate to subclinical cardiac abnormalities may provide a pragmatic approach to identify vulnerable patients who require closer cardiac surveillance.

Whether these simple, routinely collected measures of comorbidity and functional status are associated with subclinical cardiac dysfunction on echocardiography in chemotherapy-naïve cancer patients remains unclear, and addressing this gap could help refine risk stratification and guide baseline cardiac assessment before potentially cardiotoxic treatment. The results are expected to inform risk-adapted monitoring and preventative strategies in clinical practice.

Methods

This was an observational cross-sectional study conducted on adult, chemotherapy-naive cancer patients at a tertiary referral hospital in Indonesia. This observational study is reported in accordance with the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement. This study included 110 patients who underwent baseline transthoracic echocardiographic evaluation before receiving their first cycle of chemotherapy. Patients were included if they had a confirmed diagnosis of solid tumours or hematologic malignancies and had available complete clinical and echocardiographic data. Patients with known cardiovascular disease, prior chemotherapy, or poor image quality were excluded.

The study protocol was approved by the Ethics Committee of Dr. Soetomo General Academic Hospital (ethical certificate number: 0581/KEPK/II/2023). Given the retrospective use of routinely collected clinical and echocardiographic data, the requirement for individual informed consent was waived in accordance with institutional and national regulations.

Comorbidity burden was quantified using the CCI, a weighted index incorporating major chronic conditions to estimate long-term mortality risk. Higher CCI scores reflect greater comorbidity burden and have been consistently associated with poorer survival outcomes in general medical and oncology populations. Functional status was assessed using the KPS scale, which ranges from 0 (death) to 100 (normal, no complaints) in 10-point increments. The KPS is widely used in oncology and has shown good inter-rater reliability and construct validity, as well as prognostic value for treatment tolerance and survival. In this study, CCI and KPS were extracted from clinical records at the time of echocardiography.

All transthoracic echocardiograms were performed by experienced sonographers using GE Vivid E9 ultrasound systems, following American Society of Echocardiography (ASE) and European Association of Cardiovascular Imaging (EACVI) recommendations for cardiac chamber quantification and diastolic function assessment. Standard parasternal long- and short-axis and apical two-, three-, and four-chamber views were acquired with patients in the left lateral decubitus position. LVEF was measured using the biplane Simpson method. Mitral inflow velocities (E and A waves) were obtained by pulsed-wave Doppler at the leaflet tips, and early diastolic mitral annular velocity (e^′) was measured by tissue Doppler imaging at the septal and lateral annulus, with averaged values used for analysis. GLS was calculated by two-dimensional speckle-tracking echocardiography from apical two-, three-, and four-chamber views, using EchoPAC PC v204. GLS was reported as the average peak systolic longitudinal strain of all left ventricular segments, and three consecutive cardiac cycles were averaged for patients in sinus rhythm. GLS was analysed both as a continuous variable and dichotomized using an absolute cut-off of $<18\%$ to define reduced GLS, in line with prior meta-analyses and expert recommendations indicating that normal adult GLS values are typically more negative than −18%, with values less negative than this threshold consistent with subclinical LV systolic dysfunction.^4,5,6 The primary outcomes were abnormal GLS and/or diastolic dysfunction prior to chemotherapy.

Data were analysed using SPSS version 25.0 (IBM Corp., Armonk, NY). Spearman’s rank correlation coefficient ($\rho$) was used to assess associations between echocardiographic parameters (GLS, E/e^′, E/A, LAVI) and clinical variables (age, BMI, CCI, KPS). Because CCI and KPS are ordinal scales and several variables showed non-normal distributions, Spearman’s $\rho$ was preferred over Pearson’s correlation as it does not assume normality and is appropriate for monotonic relationships involving ordinal or non-normally distributed data. A multivariable linear regression model was constructed with E/e^′ as the dependent variable and age and CCI score as independent variables. Model fit was summarized using R, $R^2$, adjusted $R^2$ and the F-statistic, and regression coefficients are reported with standard errors, 95% confidence intervals, and p-values.

Results

A total of 110 cancer patients were included in the study, consisting of 62 females (56.3%) and 48 males (43.7%). The cancer types were diverse, including 9 patients (8.2%) with metastatic solid tumours, 66 patients (60%) with non-metastatic solid tumours, 4 patients (3.6%) with leukemia, and 31 patients (28.2%) with lymphoma. Abnormal GLS ($<18\%$) was found in 68 patients (61.8%), while 42 patients (38.2%) had normal GLS values.

Spearman correlation analyses are summarised in Table 1. Higher comorbidity burden (CCI) was modestly associated with worse GLS ($\rho =-0.234$, $p=0.014$) and higher E/e^′ ($\rho =0.215$, $p=0.024$), whereas better functional status (Karnofsky score) correlated with more preserved GLS ($\rho =0.217$, $p=0.023$). Age showed a positive correlation with E/e^′ ($\rho =0.249$, $p=0.009$) and a negative correlation with E/A ($\rho =-0.404$, $p<0.001$). BMI was not significantly correlated with these indices.

Table 1.

Correlation between clinical variables and echocardiographic parameters.

	GLS		E/e′		E/A
Clinical variable	$\rho$	p	$\rho$	p	$\rho$	p
Age (years)	—	—	0.249	0.009	−0.404	$<0.001$
BMI (kg/m2)	—	—	0.067	0.486	−0.096	0.319
CCI score	−0.234	0.014	0.215	0.024	−0.108	0.260
KPS score	0.217	0.023	—	—	—	—

Note: Spearman correlation coefficients ($\rho$) and corresponding p-values (p) for the relationships between clinical characteristics and cardiac function markers. GLS: Global Longitudinal Strain; E/ e^′: Ratio of early diastolic mitral inflow velocity to early diastolic mitral annular velocity; E/A: Ratio of early to late diastolic mitral inflow velocity; BMI: Body Mass Index; CCI: Charlson Comorbidity Index; KPS: Karnofsky Performance Status. A $p<0.05$ is considered statistically significant. Blank cells indicate correlations that were not assessed.

In multivariable linear regression with E/e^′ as the dependent variable and age and CCI as independent variables, the overall model was statistically significant ($R=0.316$, $R^2=0.100$, adjusted $R^2=0.083$, $F(2,107)=5.924$, $p=0.004$; Table 2). Higher age was independently associated with higher E/e^′ ($B=0.035$, 95% CI 0.005–0.065, $\beta =0.222$, $p=0.023$), whereas the association between CCI and E/e^′ was of borderline significance ($B=0.186$, 95% CI −0.024–0.396, $\beta =0.168$, $p=0.080$). Cancer type and sex were not significantly associated with any echocardiographic parameters, indicating subclinical dysfunction.

Table 2.

Multivariable linear regression with E/e^′ as the dependent variable.

Predictor	B	SE	$\beta$	95% CI for B	p
Intercept	5.551	0.765	—	4.033–7.069	$<0.001$
CCI score	0.186	0.106	0.168	−0.024–0.396	0.080
Age (years)	0.035	0.015	0.222	0.005–0.065	0.023

Notes: Results of the multivariable linear regression analysis. Model statistics: $R=0.316$; $R^2=0.100$; adjusted $R^2=0.083$; $\mathrm{\text{F}}(2107)=5.924$; $p=0.004$. B: Unstandardized coefficient; SE: Standard error; $\beta$: Standardized coefficient; CI: Confidence Interval; CCI: Charlson Comorbidity Index. The intercept represents the constant value in the regression equation. Statistical significance is indicated by $p<0.05$.

Discussion

This study investigated the association between comorbidity burden, functional status, and subclinical cardiac dysfunction among adult cancer patients prior to chemotherapy. The findings demonstrated significant, albeit modest, correlations between the CCI, KPS, and specific echocardiographic parameters, particularly GLS and diastolic function indicators.

Patient characteristics

The study population consisted of 110 adult cancer patients, with a mean age of 50.33 years. Age-related changes in cardiac compliance and ventricular relaxation may partly explain the observed echocardiographic findings.^10,11,12 Most subjects were diagnosed with non-metastatic solid tumours. The lack of a significant association between cancer type and subclinical cardiac dysfunction aligns with existing cardio-oncology guidelines, which emphasize that age and comorbidities—rather than cancer histology—are key determinants of cardiotoxic risk.^1,2,13

The baseline characteristics of our cohort illustrate that many chemotherapy-naïve cancer patients already carry a considerable burden of cardiovascular risk factors and comorbid conditions before treatment initiation. This pre-existing vulnerability may partly explain the detection of subclinical cardiac dysfunction on echocardiography even in the absence of prior exposure to cardiotoxic agents. Clinically, these findings support the concept that cardio-oncology assessment should start at the time of cancer diagnosis, with systematic attention to comorbidity profiling and functional status rather than focusing solely on the chemotherapy regimen. Future prospective studies are needed to determine whether earlier optimisation of cardiovascular risk factors and comorbidities prior to chemotherapy can mitigate the development of overt cardiotoxicity and improve long-term outcomes.

Association between Charlson comorbidity index (CCI) and echocardiographic parameters

In our study, higher comorbidity burden as measured by the CCI showed weak but statistically significant correlations with both GLS and E/e^′ (absolute $\rho$ values around 0.2). According to commonly used benchmarks, these effect sizes would be considered small; however, for complex, multifactorial physiological measures such as myocardial function, even weak correlations can reflect clinically relevant trends at the population level. In practical terms, our findings suggest that patients with a higher comorbidity load tend to have slightly more impaired systolic and diastolic indices before chemotherapy. This does not imply that CCI should be used as a stand-alone screening test for subclinical dysfunction, but rather that comorbidity burden can serve as a simple, readily available indicator of lower cardiovascular reserve.

There was a significant negative correlation between CCI and GLS, indicating that a higher comorbidity burden is associated with impaired myocardial deformation, even before chemotherapy initiation. This supports the hypothesis that comorbidities such as hypertension, diabetes, and chronic kidney disease may predispose myocardial tissue to early subclinical injury through mechanisms like endothelial dysfunction, oxidative stress, and microvascular disease.^{14,15,16,17,18} CCI was also significantly correlated with diastolic function markers—the ratio of E/e^′ and left atrial volume index (LAVI). These associations suggest that comorbid conditions contribute not only to systolic subclinical dysfunction but also to diastolic abnormalities, which may increase vulnerability to overt heart failure during or after chemotherapy.^{19,20,21,22,23}

In real-world settings where resources for comprehensive imaging are limited, patients with higher CCI scores could be prioritised for baseline echocardiography (including strain imaging where available) and more intensive cardiovascular risk optimisation. Future research should examine whether combining CCI with other markers—such as biomarkers, fitness measures, or more detailed imaging parameters—can improve risk stratification and guide preventive interventions.

Association between Karnofsky performance status (KPS) and echocardiographic parameters

Karnofsky Performance Status showed a positive, weak but statistically significant, correlation with GLS, suggesting that better functional status is associated with preserved myocardial contractility. The observed association between higher KPS and more preserved GLS in our cohort suggests that functional status, as captured by this simple clinical scale, may reflect underlying cardiovascular reserve, even in the absence of formal exercise testing. Although no significant association was found between KPS and diastolic function parameters, the trend indicates that lower physical performance may reflect underlying cardiovascular impairment, possibly due to reduced physiological reserve or undiagnosed cardiac involvement.^{24,25,26,27,28}

Similarly, the association between KPS and GLS was weak but statistically significant, with higher KPS values correlating with more preserved longitudinal strain. This pattern is consistent with the concept that functional performance scales capture, at least in part, the cumulative impact of comorbidities, frailty, and subclinical organ dysfunction. Although the correlation magnitude is modest, in a real-world oncology population, even small differences in GLS may be relevant when layered onto multiple other risk factors. From a clinical perspective, patients with lower KPS scores may represent a subgroup with reduced physiological reserve—including cardiac reserve—who could benefit from more detailed baseline cardiac assessment and closer surveillance during chemotherapy. Future work should combine KPS with objective measures of exercise capacity (e.g. 6 min walk test, cardiopulmonary exercise testing) and biomarkers to refine how functional status relates to subclinical myocardial dysfunction and to evaluate whether targeted interventions (such as exercise prehabilitation) can modify this risk.

Influence of demographic and clinical factors

Multivariable regression analysis revealed that age was an independent predictor of E/e^′, whereas CCI showed a borderline association, with the model explaining about 8–10% of the variance ($R^2=0.10$; adjusted $R^2=0.083$). This further highlights the contribution of clinical frailty and comorbidity to early cardiac dysfunction.^{7,17,24,29,30,31,32} In our cohort, age and measures of clinical status (such as comorbidity/functional indices) were more strongly associated with subclinical cardiac dysfunction than other demographic variables (e.g. sex) or tumour-related characteristics (e.g. cancer type or stage). Thus, while age remains an important demographic predictor, our findings suggest that readily available indices of overall clinical status provide additional, complementary information for risk stratification.^1,2,33,34

The large amount of unexplained variance likely reflects the influence of multiple unmeasured factors, including lifetime exposure and control of cardiovascular risk factors (e.g. blood pressure, glycaemic control, lipid levels), subclinical coronary or microvascular disease, myocardial fibrosis and remodelling, loading conditions at the time of echocardiography, cancer-related systemic inflammation or cachexia, genetic susceptibility, and random measurement variability. Future prospective studies should incorporate more granular cardiovascular phenotyping (e.g. advanced imaging, biomarkers), objective measures of physical fitness, and longitudinal treatment data to develop more comprehensive risk models. From a clinical standpoint, our results support using age and comorbidity burden as pragmatic entry points into a broader cardio-oncology risk assessment, rather than as definitive predictive tools, and highlight the need for a multimodal approach to risk stratification before chemotherapy.

Clinical implications

Our findings have several potential implications for clinical practice. First, the association between higher comorbidity burden, lower functional status, and subclinical cardiac dysfunction suggests that simple, routinely collected indices such as CCI and KPS may help identify cancer patients with lower cardiovascular reserve prior to chemotherapy. In centres where universal advanced cardiac imaging before treatment is not feasible, these patients could be prioritized for baseline echocardiography, including assessment of GLS where available, to detect early myocardial impairment. Second, the identification of subclinical dysfunction at baseline in patients with multiple comorbidities or impaired performance status may justify closer longitudinal cardiac surveillance during and after chemotherapy and, when appropriate, initiation of cardioprotective strategies or optimization of cardiovascular risk factors. Finally, because comorbidity burden and functional status are already integral to oncologic decision-making, integrating these measures into a structured cardio-oncology risk assessment pathway may provide a pragmatic approach to target limited imaging resources and supportive interventions to those most likely to benefit. Nevertheless, our results should be viewed as hypothesis-generating, and prospective studies are needed to confirm whether such risk-stratified strategies improve clinical outcomes.

Beyond risk stratification and surveillance, there is growing interest in exercise-based and multimodal prehabilitation for patients scheduled to receive chemotherapy. An integrative review by Pietrakiewicz et al. synthesized the limited but emerging evidence on prehabilitation prior to chemotherapy and suggested that interventions delivered between diagnosis and treatment initiation—particularly structured physical activity combined with nutritional and psychosocial support—may improve functional capacity, reduce treatment-related complications, and potentially enhance survival.³⁵ In this context, simple measures such as comorbidity burden and functional status may help to identify patients with lower physiological reserve who could derive particular benefit from prehabilitation strategies, in addition to closer cardio-oncology follow-up.

Study limitations

This study has several limitations warranting consideration. Clinical heterogeneity, particularly regarding volume status, concomitant supportive treatments, and cancer staging, may have introduced variability in echocardiographic parameters that could not be adequately controlled within the constraints of the study’s retrospective design. The inclusion of a heterogeneous patient population, with diverse cancer types and therapeutic regimens, may have attenuated the ability to detect cancer-specific associations.

This study did not include objective measures of physical fitness, such as treadmill exercise testing, 6 min walk test, or cardiopulmonary exercise testing, which are known to be associated with cardiovascular prognosis. The retrospective design limited our analysis to variables that were consistently documented in the medical records, namely comorbidity burden (CCI), functional status (KPS), and echocardiographic parameters. Future prospective studies should incorporate standardized exercise capacity testing and cardiac biomarkers in addition to comorbidity to refine risk stratification for subclinical cardiac dysfunction in cancer patients.

Moreover, some clinically relevant associations may not have reached statistical significance due to small effect sizes and limited statistical power. The multivariable linear regression model explained only a small proportion of the variance in E/e^′ ($R^2=0.10$; adjusted $R^2=0.083$), indicating limited explanatory power. This likely reflects the multifactorial nature of diastolic function and the influence of unmeasured factors, including lifetime cardiovascular risk exposure, myocardial fibrosis, microvascular disease, loading conditions at the time of echocardiography, and genetic susceptibility. Our findings should therefore be interpreted as demonstrating independent associations rather than providing a clinically applicable prediction tool. Future prospective studies incorporating standardized measures of physical fitness, cardiac biomarkers, and more detailed echocardiographic parameters are needed to develop more robust predictive models.

To address these limitations, future research should employ a prospective study design with a larger and more homogeneous sample size, a more balanced distribution of cancer types, and the inclusion of additional predictive variables—such as right atrial strain and cardiac biomarkers. These enhancements may contribute to the development of more accurate risk stratification models and a deeper understanding of subclinical cardiac dysfunction in oncology patients.

Conclusion

Comorbidity burden and functional status are significantly associated with subclinical cardiac dysfunction in cancer patients prior to chemotherapy. These findings support the inclusion of GLS and diastolic function assessment in baseline cardiovascular evaluation for cardio-oncology risk stratification.

Acknowledgments

The author would like to express sincere gratitude to all those who contributed to the completion of this research. Appreciation is also extended to the Department of Cardiology and Vascular Medicine, Faculty of Medicine, Universitas Airlangga, and Dr. Soetomo General Hospital for providing the facilities and data necessary for this research.

Conflict of Interest

The authors declare no conflict of interest.

Funding/Support

This study is internally funded by Universitas Airlangga.

Author Contributions

Conceptualisation: DPK, AL, BSP

Methodology: DPK, AL, AEP, AOA

Data Collection: DPK, EA, DGA

Data Analysis: DPK, AOA, AEP

Writing–Original Draft: DPK, EA

Writing–Review & Editing: DPK, BSP, TW, DGA

Supervision: BSP, TW, DGA, AL

ORCID

Dian Paramita Kartikasari https://orcid.org/0000-0002-1971-7816

Budi Susetio Pikir https://orcid.org/0000-0003-0705-9462

Achmad Lefi https://orcid.org/0000-0002-0209-5491

Tri Widiandani https://orcid.org/0000-0002-0156-6095

Desak Gede Agung Suprabawati https://orcid.org/0009-0007-3533-0412

Aditea Etnawati Putri https://orcid.org/0000-0002-6350-9619

Azril Okta Ardhiansyah https://orcid.org/0000-0001-5885-0356

Emildha Dwi Agustina https://orcid.org/0009-0000-1067-5958