Prognostic Value of Transthoracic Impedance Cardiography, Amino-Terminal Pro-B-Type Natriuretic Peptide Levels, The Six-Minute Walk Test, and Chest X-Ray in Elderly Patients with Chronic Heart Failure: A Comparative Study in Lithuania

Introduction

Chronic heart failure (CHF) is a complex clinical syndrome that develops as a result of structural and/or functional ventricular damage and is characterized by complaints and symptoms caused by left ventricular (LV) and/or right ventricular dysfunction [1,2]. In Europe in 2019, the median incidence of CHF in the general population was 3.2 (quartile 1 [Q1] 2.66-quartile 3 [Q3] 4.17) cases per 1000 person-years, while in Lithuania, the median incidence was 4.67 cases per 1000 person-years [3]. The incidence of CHF increases with age, with about 20% of people over 75 years of age having the syndrome [2].

The etiology of CHF varies in different regions of the globe and depends on geography and cultural aspects. In the developed and Western world, leading etiologies are coronary artery disease and hypertension [1]. Chronic heart failure is associated with frequent, recurrent, and long-term hospitalizations, which results in a significant economic burden for healthcare systems. The overall 1-year survival rate of CHF patients is 75.9%, the 5-year survival rate is 45.5%, the 10-year survival rate is 24.5%, and the 15-year survival rate is only 12.7% [4]. The hospitalization of a CHF patient is an event with a clear negative prognostic value [5]. Patients hospitalized for CHF decompensation have an overall 5-year survival rate of only 25% regardless of left ventricular ejection fraction (LVEF) [5]. In patients diagnosed with CHF during hospitalization, the median survival is 2.4 years shorter than in those undergoing out-patient treatment (2.9 versus 5.3 years) [4]. In women, the survival is poorer than in men, and older patients also have a higher risk of death compared to younger patients [4].

The diagnosis of CHF can be made based on characteristic complaints, symptoms, and objective signs of cardiac dysfunction. However, medical history and objective examination alone are not sufficient, and additional diagnostic tests are needed to rule out other diseases with similar symptoms [1]. As of most recent European Society of Cardiology guidelines the following tests are recommended for the assessment of patients with suspected CHF: electrocardiogram, natriuretic peptide (NP) plasma levels, basic blood workup for comorbidities, transthoracic echocardiography (TTE), and chest X-ray [1].

Natriuretic peptides such as B-type (or brain) natriuretic peptide (BNP) and amino-terminal pro-B-type natriuretic peptide (NT-proBNP) – are indispensable biomarkers used for the diagnosis and prognosis of CHF [6]. The presence of CHF increases serum NP levels [7], which is significantly associated with a more severe disease course and a worse prognosis in elderly CHF patients [8]. Natriuretic peptides have a high negative predictive value and may exclude the diagnosis of CHF. On the other hand, the positive predictive value of the NP test is low, and thus the test does not definitively confirm the diagnosis of CHF [1].

Echocardiography is recommended as the key imagery test to assess cardiac structure and function. During TTE clinicians may determine one of key parameters in CHF which is LVEF, as well as other parameters such as chamber size, type of LV hypertrophy, regional wall motion abnormalities, right ventricular function, signs of pulmonary hypertension, valvular function, and markers of diastolic function [1]. A chest X-ray is recommended to differentiate the suspected CHF from other causes of dyspnea. Additional findings such as pulmonary congestion or enlarged heart may also provide additional diagnostic benefits in the diagnosis of CHF [1].

Transthoracic impedance cardiography (ICG) is a safe, non-invasive, cheap, simple test, that does not require special training of the personnel, and it is a reliable method for evaluating the hemodynamics of the cardiovascular (CV) system, which is also valuable in the diagnosis of CHF [9]. This method measures the impedance of the tissues through which a low-strength current passes [10]. Areas of the body with higher amounts of blood or fluid will have lower impedance, while areas with denser tissue will have higher impedance [9]. This helps to identify one of the main pathological processes of CHF-fluid accumulation in the chest and assign patients to a group with a poor prognosis [10].

Even though the six-minute walk test (6MWT) is not a routine test, it is a widely available, inexpensive, easy-to-perform, and well-tolerated method of assessing the functional capacity of CHF patients [11]. Studies have shown that patients over 65 years of age walk a significantly shorter distance during the 6MWT than younger patients do [12,13]. Reduced 6-minute walk distance (6MWD) in CHF patients is associated with increased mortality, non-fatal CV events, and hospitalizations [11]. A 6MWD ≤300 m indicates a poor prognosis, while <200 m indicates a significantly increased risk of death. A stable 6MWT score over 12 months indicates improved survival in patients with CHF [11]. Studies have also shown a statistically significant correlation between a higher degree of LV systolic or diastolic dysfunction and a lower 6MWD [14]. The combination of the 6MWT results with stroke volume (SV) and cardiac output (CO) determined by ICG testing may help more accurately predict a patient’s functional capacity than 6MWT alone [15]. In the comparison of the 6MWT and NP plasma levels, 6MWT reflects functional capacity, disease severity, and quality of life, while NP is more related to cardiac function and reflects CHF symptoms, LVEF, and the New York Heart Association (NYHA) class [16]. The higher the NYHA class and serum NT-proBNP concentrations, the poorer the echocardiographic indices of cardiac chamber size and wall thickness [17], thus these tests should be considered together rather than in isolation to better predict disease prognosis. It can be argued that the combined use of NP and ICG testing increases specificity and sensitivity, as the systolic time ratio (STR) index complements, rather than replicates, the NP results, which helps to identify high-risk patients, who are less likely to survive and are more likely to experience CHF-related deterioration [18]. Other investigators have found a significant association between LVEF determined by TTE and pre-ejection period (PEP), LV ejection time (LVET), and STR determined by ICG [19].

Currently, interventional hemodynamic, radiological, and nuclear medicine tests are considered the gold standard for confirming etiologic diagnosis of CHF, selecting the appropriate treatment, and assessing the prognosis [20–23]. These testing methods are highly accurate in diagnosing etiology of CHF and predicting the disease outcome for the patients. However, they have some important drawbacks, such as high cost, invasiveness, and limited availability [20–23]. Thus, the search for new non-invasive, low-cost, and widely available diagnostic and prognostic tests for CHF is also driven by economic considerations, among other factors.

In this study, we hypothesized that low-cost, widely available, and rapidly performed non-invasive tests such as 6MWT, NT-proBNP, TTE, ICG, and chest X-ray would have an aggregate prognostic value in predicting poor outcomes in the elderly CHF patients.

Therefore, this study from a single center in Lithuania aimed to evaluate outcomes in 87 elderly patients hospitalized with CHF. The methods used included comparing ICG, NT-proBNP serum levels, the 6MWT, and chest X-ray.

Material and Methods

STUDIED SAMPLE:

An observational prospective study was conducted in Lithuania, in 2019–2022. All participants of the study provided written informed consent. The study was approved by the local Biomedical Research Ethics Committee (No: BE-2-17). The study included 87 patients (49 men and 38 women) hospitalized for CHF who underwent 6MWT and ICG tests in addition to the standard CHF diagnostic tests such as TTE, NT-proBNP, chest X-ray. The inclusion criterion was a diagnosis of CHF confirmed according to the European Society of Cardiology guidelines in force at the time [1,24]. The exclusion criteria were age <65 years, severe aortic valve regurgitation, severe aortic valve stenosis, acute myocardial infarction, acute heart failure, heart rate ≥180 bpm, mean arterial blood pressure >130 mmHg, an implanted cardioverter-defibrillator, or an implanted intrathoracic impedance monitor.

The primary endpoint of the study was CV death during in-patient treatment and the secondary endpoint was CV death during the whole follow-up of the study. Data about CV deaths after discharge from the hospital were gathered from the National Medical Record Database. These data were used to calculate Kaplan-Meier curves.

METHODS OF EXAMINATION:

In this study, we analyzed data from serum NT-proBNP, TTE, chest X-ray, 6MWT, and ICG tests. All these tests were performed on the day of enrollment, and NT-proBNP, 6MWT, and ICG tests were repeated on the day of discharge.

The six-minute walking test was performed in accordance with the most recent 6MWT guidelines [25] and methodological recommendations [11]. TTE was performed with an Aloka Prosound α6 (Hitachi Aloka Medical, Ltd., Tokyo, Japan) using a 3 MHz transducer UST-5299. All measurements were performed by an experienced cardiologist, and all of them were performed by the same doctor. The researcher was blinded to the patient’s data. All measurements were performed based on the recommendations of the American Society of Echocardiography [26] and the European Association for Cardiac Imaging [27]. For ICG measurements, the Niccomo® (MEDIS Medizinische Messtechnik GmbH, Ilmenau, Germany) ICG monitor was used, which uses a modification of the 8-electrode setup: 4 electrodes are attached to either side of the neck, at the level of the great vessels, and a further 4 electrodes are attached to either side of the thorax, at the midline of the axilla and at the level of the xiphoid process. The following parameters were recorded during the ICG test: SV, stroke index (SI), thoracic fluid content (TFC), TFC index (TFCI), CO, STR, PEP, LVET, left cardiac work (LCW), and LCW index (LCWI) [28,29]. The dilatation of the right pulmonary artery on chest X-ray (dRPAcXR) was identified when interlobar artery at the hilum on the right side was >15 mm in women and >16 mm in men.

STATISTICAL ANALYSIS:

The IBM Statistical Package for Social Sciences Statistics for Windows, version 29.0 (IBM Corporation, New York, United States) was used for data analysis. All the continuous data were not distributed normally. Therefore, for descriptive statistics, all continuous data are presented as the median and Q1–Q3, and non-parametric tests were used to test statistical hypotheses. For categorical variables, presented as numbers (%), the chi-square test was used for comparisons between the groups. Two independent variables were compared using the Mann–Whitney U test. Spearman’s correlation coefficient (r) was used to assess relationships between all non-invasive CHF diagnostic tests. Receiver operating characteristic (ROC) curves were used to assess the diagnostic accuracy of the test to identify lethal patient outcomes, and the area under the curve was evaluated. The sensitivity and specificity of the main diagnostic tests were calculated. Cut-off values were selected so that Youden’s index had the highest value possible. The Kaplan-Meier method, log-rank test and univariate and multivariate Cox regression models (controlling for all main covariates such as age, gender, body mass index [BMI], and comorbidities) were used to calculate the survival rates, differences in survival curves, CV death risk and survival chances. Hazard ratios (HRs) were estimated using the Cox proportional hazards model. Results were considered statistically significant when the P value was <0.05.

Results

DESCRIPTIVE STATISTICS OF THE STUDIED SAMPLE:

The median age of the study participants was 75 years (70.0–83.5). The descriptive statistics of the patients are presented in Table 1. The patients were distributed into 3 groups according to NYHA functional capacity classes: NYHA class II – 17 patients (19.5%), NYHA class III – 55 patients (63.2%), and NYHA class IV – 15 patients (17.2%). According to the cause of CHF, the patients were divided into 6 groups. In 56 patients (64.4%), the cause of CHF was found to be ischemic heart disease (IHD), in 19 patients (21.8%), arterial hypertension, in 6 patients (6.9%), atrial fibrillation or atrial flutter, in 2 patients (2.3%), cardiomyopathy, in 2 patients (2.3%), chronic obstructive pulmonary disease (COPD), and in 1 patient (1.1%), valvular heart disease.

ASSOCIATIONS BETWEEN KEY CHRONIC HEART FAILURE MARKERS:

In our study, we looked for associations between key ICG parameters, 6MWD, and other diagnostic markers of CHF. The analysis revealed weak to moderate correlations between key ICG parameters and other CHF markers. The strongest correlations were found between TFC/TFCI and NT-proBNP (r=0.358/r=0.408, P<0.001), LVEF (r=−0.479/r=−0.326, P<0.001/P=0.002), left atrial dimension in the parasternal long axis view (LAD) (r=0.488/r=0.446, P<0.001), and NYHA class (r=0.471/r=0.434, P<0.001) (Table 2).

In addition to that, we also looked for associations between the NYHA functional class and other diagnostic markers of CHF and between 6MWD and other diagnostic markers of CHF. The strongest correlations were found between NYHA class and TTE indices LVEF (r=−0.550, P<0.001), LAD (r=0.423, P<0.001), and NT-proBNP (r=0.398, P<0.001) (Table 3). The strongest correlations were found between 6MWD and left ventricular mass (LVM) index (r=−0.353, P<0.001), NT-proBNP (r=−0.251, P=0.019) and left ventricular end-diastolic dimension index (LVEDDI) (r=−0.550, P=0.019) (Table 4).

OUTCOME ANALYSIS:

The median duration of the follow-up of the patients included in the study was 23 months (the shortest duration being 2 months, and the longest – 36 months). By the end of the study, 58 of the subjects (66.7%) survived, and 29 (33.3%) died because of a major CV event. The results of the study were analyzed by comparing the objective examination, laboratory, and instrumental data between the surviving and the deceased groups (statistically significant results are presented in Table 5). There were statistically significant differences between the surviving and the deceased groups in respiratory rate, NYHA class, NT-proBNP, glomerular filtration rate, 6MWD, ICG parameters (except for LCW and LCWI), and TTE parameters (except for the LVM).

Diagnostic accuracy to detect CV death during the whole study period was assessed for all non-invasive CHF diagnostic tests and other clinical and laboratory data. ROC curves were selected to assess the diagnostic accuracy of these tests. Baseline parameters, recorded at enrollment, were analyzed. Cut-off values were selected so that Youden’s index had the highest value possible. These cut-off values were then used in univariate and multivariate Cox regression models (statistically significant results are presented in Table 6). First, the influence of all baseline demographic, clinical, laboratory, TTE, 6MWT, and ICG parameters on the survival of patients was evaluated using a univariate Cox regression model. Next, statistically significant variables from univariate Cox regression model were included in the multivariate Cox regression model. The final Cox model of multivariate regression analysis of statistically significant variables was created using the forward stepwise procedure. The final model revealed that the following baseline variables were independently and significantly associated with CV death (all P<0.05): TFC ≥41.1 1/kΩ (HR 32.354, 95% confidence interval [CI] 2.758–379.488), NT-proBNP ≥332.0 pmol/L (HR 4.739, 95% CI 1.656–13.559), 6MWD ≤203.5 m (HR 3.975, 95% CI 1.002–15.770), and dRPAcXR (HR 5.555, 95% CI 1.714–18.005). To assess the association between the patients’ poor outcome and laboratory and instrumental findings, survival curves were calculated using the Kaplan-Meier method. Cut-off values were determined by the ROC curve analysis, data is presented in Figures 1–4.

Discussion

The major finding of the present study was that parameters such as TFC ≥41.1 1/kΩ, NT-proBNP ≥332.0 pmol/L, 6MWD ≤203.5 m, and dRPAcXR were independently and statistically significantly (all P<0.05) associated with a poorer prognosis in CHF patients. Chronic heart failure patients with TFC ≥41.1 1/kΩ had a 32.4-fold higher risk of CV death, compared to patients with TFC <41.1 1/kΩ. Patients with NT-proBNP ≥332.0 pmol/L had a 4.7-fold higher risk of CV death than patients with NT-proBNP <332.0 pmol/L did. Patients with the total 6MWD ≤203.5 m had a 3.9-fold higher risk of CV death compared to patients with the total 6MWD >203.5 m. The dRPAcXR was associated with a 5.6-fold increase in the risk of CV death.

Similar to our study, Giannitsi et al found that in patients with mild-to-moderate (NYHA functional class II–III) heart failure with reduced ejection fraction (HFrEF), reduced 6MWD was one of the strongest independent prognostic indicators associated with an increased risk of mortality, non-fatal CV events, and hospitalization [11]. Ferreira et al found that 6MWD was directly proportional to disease progression, i.e., for every 50-m reduction in 6MWD, there was an 8% higher risk of hospitalization or death and a 14% higher risk of death due to CHF [12]. Patients with a 6MWD of 241–360 m had a 1.44-fold increase in the probability of a poor outcome (95% CI 1.14–1.80, P=0.002), while patients with a 6MWD of <240 m had a 1.73-fold increase in that probability (95% CI 1.38–2.18, P<0.001) [12], which is consistent with the results of our study.

Grundtvig et al also obtained similar results: the strongest independent risk factors for lethal outcome in patients with CHF were older age (HR 1.036, 95% CI 1.025–1.047, P<0.001), lower 6MWD (HR 0.979, 95% CI 0.972–0.986, P<0.001), higher NT-proBNP (HR 1.861, 95% CI 1.450–2.389, P<0.001), and a diagnosis of cancer within the last 5 years (HR 3.179, 95% CI 2.574–3.926, P<0.001). In the group of patients with a 6MWD ≤360 m, mortality was 25%, while in groups with 6MWD 361–479 m or ≥480 m, poor outcomes were observed in 9% and 4% of patients, respectively [30].

Malfatto et al found that higher BNP and TFC values were independent predictors of a lethal outcome. In patients with BNP <450 pmol/L and TFC <40 1/kΩ, the 4-year mortality rate was only 2.1%, whereas in patients with BNP ≥450 pmol/L and TFC ≥40 1/kΩ, the 4-year mortality rate was 46.5% (P=0.001) [31]. In our study, the maximum follow-up was shorter (3 years), which may explain the lower mortality rate, with a third of the patients dying during the entire follow-up period.

Sadauskas et al found similar parameters that were independently and statistically significantly associated with a poorer prognosis in patients with CHF. The parameters that significantly increased the probability of a lethal outcome were BNP ≥350 pmol/L (odds ratio [OR] 4.4, 95% CI 1.3–15.7, P<0.05), TFC ≥34 1/kΩ (OR 4.3, 95% CI 1.0–19.7, P<0.05), and STR ≥0.55 (OR 2.9, 95% CI 1.1–7.7 P<0.05) [29]. However, similarly to our study, there was no statistically significant association between male sex, NYHA class III–IV, TFCI, SV, CO, or LCWI and the risk of death [29].

Scardovi et al. found that older age (HR 1.162, 95% CI 1.043–1.294, P=0.004), male sex (HR 2.115, 95% CI 1.257–3.560, P=0.005), grade 3 renal failure (HR 1.991, 95% CI 1.324–2.996, P=0.001), a non-obese phenotype (HR 1.905, 95% CI 1.070–3.390, P=0.028), IHD in CHF etiology (HR 1.833, 95% CI 1.264–2.660, P=0.001), NYHA class III (HR 1.958, 95% CI 1.343–2.855, P<0.001), NT-proBNP >125 pmol/L (HR 2.144, 95% CI 1.403–3.276, P=0.001), and 6MWD ≤360 m (HR 1.923, 95% CI 1.195–3.096, P=0.007) were independent risk factors in predicting patient death [32]. In our study, we did not find any association of poor outcome with age, sex, or BMI, but this could be explained by the fact that more than a half of all the participants of our study were 75 or older. The results for the NYHA functional class, NT-proBNP, and 6MWD were consistent with our findings, except for the lower NYHA class and NT-proBNP margins and the higher 6MWD margin, but this may be explained by the fact that the study by Scardovi et al. analyzed the data of outpatients with stable heart failure.

The present study found a weak to moderate correlation between the ICG data, 6MWD, and other non-invasive tests for CHF. The strongest correlations were found between: NT-proBNP and TFCI; 6MWD and PEP; LVEF and STR and PEP; LAD and SRT and PEP; NYHA class and STR, TFC, TFCI and SV. However, no statistically significant correlations were found between ICG parameters and LVEDDI, while associations with other TTE parameters were weak. The associations of the NYHA functional class with the results of the other analyzed non-invasive diagnostic tests for CHF were found to be weak to moderate, with the strongest associations found between the NYHA functional class and LVEF, LAD, serum NT-proBNP levels, and LVEDDI.

The findings of researchers who have evaluated the association of ICG data and 6MWT data with other invasive and non-invasive tests for the diagnosis of CHF are contradictory. Some find moderate to very strong associations, others do not find statistically significant associations at all. For example, Sadauskas et al [29,33] and Galas et al [28] also found that TFC correlates best with serum NP concentration, with correlation coefficients of 0.425 (P=0.001), 0.4 (P<0.001), and 0.46 (P=0.000001), respectively. Ling Ding et al [34] also found that 6MWD correlated best with CO (r=0.494, P=0.023), SI (r=0.633, P=0.002), LCW (r=0.615, P=0.003), and LCWI (r=0.491, P=0.024). Fang Liu et al [15] also found that 6MWD correlated best with SI (r=0.42, P<0.05) and CO (r=0.66, P<0.01). Parot et al [35] also found that the change in LVEF determined by TTE during patient treatment correlated well with the change in STR (r=−0.73). Sadauskas et al found a similar association between LVEF and STR: r=−0.31, P=0.001 [29]; r=−0.4, P=0.002 [33]. However, one study did not find even a satisfactory association between CO determined by thermodilution and CO determined by ICG [36]. In one study, a weak association between the 2 parameters (area under the curve 0.537) was found, predicting a 15% change in CO measured by thermodilution, whereas the ICG method had a sensitivity of only 55.56%, a specificity of 33.33% and a diagnostic accuracy of 46.67% [36]. In summary, the associations between ICG data and 6MWD and other non-invasive diagnostic tests examined in this study were consistent with those of other investigators, but the associations were not strong, possibly influenced by the more severe condition of the patients in the study.

This study has several limitations. First, the study included patients from only one hospital, and thus the specifics of the hospital might have influenced the results. Second, the study included elderly patients with a median age of 75 years, thus the results of this study may only be applicable to older patients. Third, the study included 87 patients, thus a larger sample size would be required to validate our study results. Fourth, only hospitalized patients with a previously established diagnosis of CHF were included in the study, while outpatients were excluded, which might have led to a poorer prognosis for the patients. Fifth, most of the study participants belonged to NYHA functional class III and IV, and therefore the results of this study should not be generalized to patients with mild and stable CHF. Sixth, ICG is not used routinely in many in-patient and out-patient clinics, so, it is also a certain limitation for the implementation of the results of this study in the clinical practice. Seventh, the major limitation of the study is the lack of control group. And last, 6MWT is not without limitations as well. Its results can be affected by myriad factors with no direct relation to cardiopulmonary status, including age, sex, height, and weight. It can also be affected by comorbidities such as cognitive impairment, peripheral arterial disease, musculoskeletal disorders. Therefore, the authors acknowledge that a larger multicenter study with control group should be performed to confirm the results of this study.

Conclusions

In this study, TFC ≥41.1 1/kΩ, NT-proBNP ≥332.0 pmol/L, 6MWD ≤203.5 m, and dRPAcXR had a combined prognostic value in predicting cardiovascular death in patients with CHF. Therefore, these parameters may be of value in assessment of the diagnosis and prognosis in this patient cohort.