Introduction

Pulmonary arterial hypertension (PAH) is a progressive disease characterized by an increase in mean pulmonary arterial pressure and pulmonary vascular resistance (PVR), which lead to a decrease in cardiac output (CO) and death.¹ Until now, although there are a dozen target drugs for PAH treatment, the pulmonary arterial pressure is hardly reversed to normal.² The primary goal for PAH treatment is to maintain a low-risk status with normal CO and cardiac index (CI).³ Thermodilution (TD) during right heart catheterization (RHC) is the reference standard for CO measurement in PAH patients. However, TD is invasive, and equipment and PAH expertise are in demand, constraining its use in the routine risk assessment of PAH patients.

The technique of impedance cardiography (ICG), initially introduced by Kubicek et al⁴ in 1966, has garnered widespread recognition over the past half-century in cardiovascular diseases. Impedance cardiography operates on the principles of Ohm’s law, assuming that the impedance of thoracic tissue is parallel to the blood in the thorax. Four pairs of electrodes are symmetrically attached to the root of the neck and the bilateral midaxillary line at the level of the xiphoid. The stroke volume (SV) is measured by monitoring impedance changes during each cardiac cycle.⁵ Owing to its convenience and non-invasive characteristics, ICG is predominantly employed in the hemodynamic assessment during cardiopulmonary exercise testing as well as in the management of heart failure and hypertension.^6-9 Previously, several studies discussed the correlation and agreement between CO measured by ICG (CO_ICG) and TD during RHC (CO_TD) in heart diseases and PH patients; unfortunately, the conclusions were controversial.^10-13 The difference may be attributed to very small study populations and mixed types of PH patients. Moreover, previous studies were carried out based on the old PH definition that mean pulmonary arterial pressure (mPAP) ≥ 25 mm Hg, which has been changed to mPAP > 20 mm Hg according to the latest PAH guideline. The agreement between CO_ICG and CO_TD under the new PH diagnostic criteria is unclear.

Cardiac index derived from CO is a traditional hemodynamic factor involved in PAH risk stratification with cut-off value of ≥2.5 L/min/m² and <2.0 L/min/m² to determine low-risk and high-risk, respectively. Meanwhile, the new guideline, for the first time, recommends stroke volume index (SVI) in PAH risk stratification with cut-off value of >38 mL/m² and <31 mL/m² to determine low-risk and high-risk separately.³ Both CI and SVI can be acquired during an ICG examination. Whether CI_ICG/SVI_ICG can be used for PAH risk stratification has not been reported.

In the present study, the CO_ICG and CO_TD were compared under the new PH diagnostic criteria and further explored the possibility of using CI_ICG and SVI_ICG for PAH risk stratification in a retrospective PAH cohort.

Methods

Subjects

Suspected PH patients who underwent RHC from September 2018 to December 2022 were screened. Group 1 and Group 4 pre-capillary PH patients diagnosed by RHC were included. Pulmonary arterial hypertension was diagnosed according to the new hemodynamic definition that mean pulmonary arterial pressures (mPAP) > 20 mm Hg and pulmonary arterial wedge pressure (PAWP) ≤ 15 mm Hg. The exclusion criteria were: (1) patients with an unstable condition; (2) patients with severe generalized edema and sepsis; (3) PH secondary to left heart diseases and chronic respiratory diseases; (4) sustained tachyarrhythmia; (5) patients with uncorrected congenital heart disease (with intracardiac shunts); and (6) patients refused to participate or sign an informed consent form. This study was reviewed and approved by the Institutional Ethics Committee.

Data Collection

The clinical characteristics including age, gender, body mass index (BMI), type of PH, six-minute walking distance (6MWD), World Health Organization functional class (WHO-FC), and N-terminal-prohormone B‐type natriuretic peptide (NT-proBNP) were collected. Factors that have been previously reported to affect ICG accuracy were also collected, including arrhythmia (atrial flutter, atrial fibrillation, frequent atrial/ventricular premature beats), extremity edema, and pericardial effusion.

Right Heart Catheterization and Impedance Cardiography

Seven-Fr Swan-Ganz catheter (Edwards Lifesciences, California) was used for RHC. Mean right atrial pressure (mRAP), mean pulmonary arterial pressure (mPAP), PAWP, mixed venous oxygen saturation (SvO₂) and cardiac output (CO_TD) by thermodilution technique were recorded.¹⁴ Pulmonary vascular resistance was calculated as (mPAP-PAWP)/CO_TD. CI_TD was calculated as CO_TD/body surface area. SV_TD was calculated as CO_TD/heart rate and SVI_TD was calculated as SV_TD/body surface area.

The ICG examination was conducted by the same physician using CSM3100 (Shenzhen Qianfan Electronics Co. Ltd, China). There were two electrodes positioned bilaterally at the base of the neck, and another two pairs of electrodes located at the intersection of the xiphoid process level and mid-axillary level. The essential information, including sex, age, height, weight, systolic blood pressure, and diastolic blood pressure, was entered in the software. CO_ICG, CI_ICG, SV_ICG, and SVI_ICG were calculated automatically.

Patients Follow-up and Risk Stratification

Clinically deteriorated patients were assessed by an experienced PH expert if they exhibited a decline in exercise tolerance, required hospitalization due to symptom aggravation, or experienced death within 1 year of follow-up. Risk stratification was carried out according to the 2022 ESC/ERS guidelines of PAH. The cut-off values for CI_TD were >2.5 L/min/m² to determine low-risk status and <2.0 L/min/m² to determine high-risk status. The cut-off values for SVI_TD were >38 mL/m² to determine low-risk status and <31 mL/m² to determine high-risk status.

Statistical Analysis

The normality of continuous data was assessed using the Shapiro-Wilk test. Continuous variables were expressed as the mean with the SD (M ± SD) for normally distributed data, or otherwise as the median with the interquartile range (IQR). Categorical variables were described as counts and percentages. For normally distributed data, when comparing the means of continuous variables between two independent groups, an independent samples t-test was employed; for comparisons involving three or more independent groups, a one-way analysis of variance (ANOVA) was utilized. The linear relationship between two variables was assessed using Pearson’s correlation coefficient. Conversely, for non-normally distributed data, the Mann-Whitney U test was used to compare the median of variables between two groups, while the Kruskal-Wallis H test was employed for comparisons involving more than two groups. The relationship between two variables was assessed using Spearman’s rank correlation coefficient. If there was a significant difference between groups, a post-hoc test was used to compare groups (Tukey HSD test for one-way ANOVA and Bonferroni adjusted Dunn test for Kruskal–Wallis H test). When comparing categorical variables, Pearson’s chi-square test was used when all expected frequencies were ≥5. If any expected frequency was <5, Fisher’s exact test was employed. The Bland–Altman method was used to assess the agreement between TD and ICG, with the mean difference and limits of agreement (LoA) calculated. The acceptability of the new method was evaluated by calculating the percentage error, which was obtained by dividing twice the SD by the mean of the 2 methods.¹⁸ The receiver operator characteristic (ROC) curve was employed to elucidate the efficacy of CI_ICG and SVI_ICG in clinical deterioration prediction. A P-value < .05 was considered significant. Statistical analysis was performed using GraphPad Prism (v. 8.0.2, GraphPad Software, San Diego, CA, USA) and OriginPro (v. 2024 SR1, OriginLab Corp., Northampton, MA, USA).

Statement

We have not used artificial intelligence (AI)-assisted technologies in the production of submitted work.

Results

Patient Characteristics

A total of 132 patients were screened, and 21 patients were excluded with PH after RHC. There were 14 patients with uncorrected intracardiac shunts and 4 patients with left heart disease excluded. Finally, 93 pre-capillary PH patients were enrolled in the study. Fifty-four (58.06%) patients belong to Group 1 PAH, of which 31 (57.41%) patients were IPAH, 10 (18.52%) patients were CTD-PAH, and 13 (24.07%) patients were repaired CHD-PAH. The other 39 (41.94%) patients were diagnosed with chronic thromboembolic pulmonary hypertension (CTEPH). Fifty-nine (63.44%) patients were female with an average age of 47.38 ± 16.08 years. Baseline hemodynamic parameters measured by RHC and ICG were shown in Table 1.

Cardiac Output Comparison Between Thermodilution and Impedance Cardiography

The CO measured by TD and ICG were 4.93 ± 1.06 L/min and 4.41 ± 1.23 L/min, respectively. Correlation between CO_TD and CO_ICG was moderate (r = 0.49, P < .001) (Figure 1A). Among patients with Group 1 PAH, the mean CO_TD and CO_ICG were 5.13 ± 1.10 L/min and 4.57 ± 1.22 L/min, respectively. The correlation slightly stronger compared to the whole cohort (r = 0.52, P < .001) (Figure 1B). The Bland-Altman method was used to analyze the agreement of the two methods. The bias was 0.52 L/min, with LoA ranging from −1.76 to 2.80 L/min and a percentage error of 49.89% (Figure 2A). For Group 1 PAH patients, the bias was 0.55 L/min and the LoA was −1.70 to 2.80 L/min, with a percentage error (Figure 2B).

Stroke Volume Comparison Between Thermodilution and Impedance Cardiography

Generally, SV measured by TD and ICG were 64.63 ± 17.10 mL and 60.94 ± 18.03 mL, respectively, with a moderate correlation (r = 0.53, P < .001) (Figure 3A). For patients with Group 1 PAH, SV_TD and SV_ICG were 67.54 ± 19.04 mL and 61.88 ± 17.94 mL, respectively, with a moderate correlation (r = 0.51, P < .001) (Figure 3B).

Bland–Altman analysis showed the bias was 3.69 mL, with LoA ranging from −29.78 to 37.14 mL, and a percentage error of 54.38% (Figure 4A). For Group 1 PAH patients, the bias was 5.66 mL, LoA was −30.16 to 41.48 mL with a percentage error of 56.50% (Figure 4B).

The analysis of CO and SV among CTEPH patients revealed a poorer correlation and agreement between TD and ICG compared to the PAH patients, especially concerning agreement (Supplementary Figures 1 and 2).

Factors Which May Influence the Correlation Between the Different Methods

To further study the confounding factors which may influence CO_ICG accuracy, the patients were divided into comparable and non-comparable groups depending on whether the difference between CO_TD and CO_ICG ≥ 20%. As shown in Table 2, body weight index (BMI), heart rate, hemodynamic parameters, pericardial effusion, peripheral edema, arrhythmia, and PAH etiology do not affect ICG CO accuracy.

Impedance Cardiography in Predicting Clinical Deterioration and Classify Risk Status

Sankey diagrams were generated to visually represent the overlap of CI and SVI across the two methods for baseline risk stratification in Group 1 PAH patients. Fifty-four patients finished 1-year follow-up. Figure 5A displayed movement of PAH patients between risk categories based on CI_ICG and CI_TD. CI_ICG classified 30 (55.5%) patients as low-risk; 48.1% moved into the low-risk category of CI_TD and 7.4% low-risk moved into intermediate-risk by CI_TD risk stratification. Risk stratification between SVI_ICG and SVI_TD showed less consistence (Figure 5B). In order to determine the cut-off value of CI_ICG and SVI_ICG in predicting low-risk status, ROC curve analysis was further carried out based on follow-up data.

After a 1-year follow-up, clinical deterioration was observed in 8 out of 54 patients with Group 1 PAH. An ROC curve was constructed to study the ability of CI_ICG and SVI_ICG in predicting clinical worsening events in Group 1 PAH patients. For CI_ICG, the area under the curve (AUC) was 0.76 (95% CI 0.59-0.94), the CI_ICG cut-off value was 2.45 L/min/m², the sensitivity was 70%, and the specificity was 81.80%, respectively (Figure 6A). The AUC for SVI_ICG was 0.81 (95% CI 0.63-0.98), the cut-off value was 32.25 mL/m², the sensitivity was 76%, and the specificity was 81.80% (Figure 6B).

Discussion

Several studies have already compared ICG against TD in CO measurement in PH patients since 2004.^{10,11,13,15,16} The current study has several major differences from the previous ones. Firstly, CO_TD and CO_ICG were compared according to the latest PH definition with mPAP > 20 mm Hg, published in the 2022 ESC/ERS PH guideline, while previous studies were carried out based on the old PH definition with mPAP ≥ 25 mm Hg. Secondly, the focus was on PAH, especially those belonging to Group 1 PAH. PH patients secondary to left heart diseases (Group 2), unrepaired congenital heart disease, or lung diseases and/or hypoxia (Group 3), were excluded, because these patients exhibited different heart and lung geometry, which may be a confounding factor for ICG measurement. Thirdly, the potential of ICG-measured CI and SVI in PAH risk assessment was investigated.

Pulmonary arterial hypertension is a devastating, occultly progressive disease resulting in death. The treatment strategy is a risk-based, goal-oriented treatment approach, where achieving and/or maintaining a low-risk status is recommended. Currently, the 2022 ESC/ERS risk stratification tool includes both CI and SVI measured by RHC or cardiac magnetic resonance imaging (cMRI),¹⁷ either invasive or resource-demanding. Guidelines on the diagnosis and treatment of PH have advocated a goal-oriented treatment approach since 2009,¹⁸ not only at baseline but also at follow-up. Patients who reached a low-risk profile at follow-up had similar survival as those who remained low risk from baseline to follow-up. Patients who worsened to an intermediate or high-risk profile during follow-up had as poor survival as those who remained intermediate or high-risk during follow-up.¹⁹ Here, a moderate correlation between CO_TD and CO_ICG was shown; the ICG method tends to underestimate both CO and SV, with a wide LoA. Another study also proved that ICG accuracy decreases for extreme CO values.¹³ Although ICG tends to overestimate the disease severity in risk stratification, it is more accurate in distinguishing patients in low-risk status. The main advantage of ICG over RHC is its non-invasive and reproducible characteristics, which guarantees its utilization during PAH follow-up.

The cut-off value to identify low-risk status varies among different methods. Another guideline-recommended parameter in PAH risk stratification is SVI measured by cMRI (SVI_cMRI). The cut-off value was > 40 mL/m² for SVI_{cMRI ,} ²⁰ close to > 38 mL/m² for SVI_TD. The cut-off value of ICG-measured CI and SVI to distinguish low-risk patients in the follow-up cohort was also verified. The cut-off value was > 2.45 L/min/m² for CI_ICG and > 32.25 mL/m² for SVI_ICG, close to the TD method recommended by the guideline.

The correlation between CO_ICG and CO_TD varies among different studies. Impedance cardiography accuracy may be influenced by various factors, including the skin condition, subcutaneous fat thickness, thoracic morphology, blood resistivity, and tachyarrhythmia.^13,21 However, in the current study with a larger sample size, no valuable indicators to predict CO_ICG accuracy were found. The difference between the two methods may be attributed to systematic error. CO_ICG is determined by multiplying SV_ICG and HR, while SV_ICG was influenced by blood resistivity, the distance between the voltage-measuring electrodes, basal thoracic impedance, maximal impedance alteration, and the duration of left ventricular ejection. In patients with PAH, further investigation is necessary to determine whether basal thoracic impedance is affected by increased right heart blood volume or dilation of the dilation pulmonary artery.

Study Limitations

There are several limitations of this study: (1) simultaneous execution of the two methods was not feasible, despite efforts to minimize the time interval between them; (2) the reproducibility of the measurement of CO by ICG in the same patient at different time points at baseline was not tested; (3) clinical worsening events were used to define CI_ICG and SVI_ICG cut-off values for low-risk status instead of 1-year mortality due to the improved survival rate after PAH treatment, thus extended follow-up is necessary for further convention.

Conclusion

Impedance cardiographymeasured CO and SV presented an acceptable correlation with RHC in PAH patients. SVI and CI measured by ICG is potent to identify the low-risk status and predict clinical deterioration in PAH patients.

Supplementary Materials

Footnotes

Ethics Committee Approval: This study was reviewed and approved by the Institutional Ethics Committee, Shandong University, Qilu Hospital, Shandong, China, with the approval number: ER No. 2019 (119), dated 2019.7.4.

Informed Consent: Written informed consent was obtained from the patients who agreed to take part in the study.

Peer-review: Externally peer-reviewed.

Author Contributions: Concept – Y.Z., F.B., Y.Q., D.Z., H.Z., X.C.; Design – Y.Z., F.B., Y.Q., D.Z., H.Z., X.C.; Supervision – Y.Z., F.B., Y.Q., D.Z., H.Z., X.C.; Resource – Y.Z., F.B., Y.Q., D.Z., H.Z., X.C.; Materials – Y.Z., F.B., Y.Q., D.Z., H.Z., X.C.; Data Collection and/or Processing – Y.Z., F.B., Y.Q., D.Z., H.Z., X.C.; Analysis and/or Interpretation – Y.Z., F.B., Y.Q., D.Z., H.Z., X.C.; Literature Search – Y.Z., F.B., Y.Q., D.Z., H.Z., X.C.; Writing – Y.Z., F.B., Y.Q., D.Z., H.Z., X.C.; Critical Reviews – Y.Z., F.B., X.C.; Hongyu Zhang and Xiaopei Cui jointly directed the project, contributed equally to this work and share the corresponding authorship.

Declaration of Interests: The authors have no conflicts of interest to declare.

References

1. Simonneau G, Montani D, Celermajer DS. Haemodynamic definitions and updated clinical classification of pulmonary hypertension. Eur Respir J. 2019;53(1):1801913-.
2. Leber L, Beaudet A, Muller A. Epidemiology of pulmonary arterial hypertension and chronic thromboembolic pulmonary hypertension: identification of the most accurate estimates from a systematic literature review. Pulm Circ. 2021;11(1):2045894020977300-.
3. Humbert M, Kovacs G, Hoeper MM. 2022 ESC/ERS Guidelines for the diagnosis and treatment of pulmonary hypertension. Eur Respir J. 2023;61(1):2200879-.
4. Kubicek WG, Karnegis JN, Patterson RP, Witsoe DA, Mattson RH. Development and evaluation of an impedance cardiac output system. Aerosp Med. 1966;37(12):1208-1212.
5. Woltjer HH, Bogaard HJ, de Vries PM. The technique of impedance cardiography. Eur Heart J. 1997;18(9):1396-1403.
6. Mattavelli I, Vignati C, Farina S. Beyond VO2: the complex cardiopulmonary exercise test. Eur J Prev Cardiol. 2023;30(Suppl):ii34-ii39.
7. Delfrate J, Girard-Bock C, Curnier D. Cardiopulmonary response to exercise in adults born very preterm. Eur Respir J. 2023;62(5):2300503-.
8. Myers J, Christle JW, Tun A. Cardiopulmonary exercise testing, impedance cardiography, and reclassification of risk in patients referred for heart failure evaluation. J Card Fail. 2019;25(12):961-968.
9. Greco B, Chait Y, Nathanson BH, Germain MJ. A novel hypertension management algorithm guided by hemodynamic data. Kidney Int Rep. 2022;7(2):330-333.
10. Yung GL, Fedullo PF, Kinninger K, Johnson W, Channick RN. Comparison of impedance cardiography to direct Fick and thermodilution cardiac output determination in pulmonary arterial hypertension. Congest Heart Fail. 2004;10(2 ):7-10.
11. Tonelli AR, Alnuaimat H, Li N, Carrie R, Mubarak KK. Value of impedance cardiography in patients studied for pulmonary hypertension. Lung. 2011;189(5):369-375.
12. Taniguchi Y, Emoto N, Miyagawa K. Noninvasive and simple assessment of cardiac output and pulmonary vascular resistance with whole-body impedance cardiography is useful for monitoring patients with pulmonary hypertension. Circ J. 2013;77(9):2383-2389.
13. Dupuis M, Noel-Savina E, Prévot G. Determination of cardiac output in pulmonary hypertension using impedance cardiography. Respiration. 2018;96(6):500-506.
14. Ganz W, Donoso R, Marcus HS, Forrester JS, Swan HJ. A new technique for measurement of cardiac output by thermodilution in man. Am J Cardiol. 1971;27(4):392-396.
15. Panagiotou M, Vogiatzis I, Jayasekera G. Validation of impedance cardiography in pulmonary arterial hypertension. Clin Physiol Funct Imaging. 2018;38(2):254-260.
16. Yağmur B, Şimşek E, Kayıkçıoğlu M. Could impedance cardiography be a non-invasive alternative method of measuring cardiac output in patients with pulmonary hypertension?. Anatol J Cardiol. 2023;27(11):650-656.
17. Humbert M, Kovacs G, Hoeper MM. 2022 ESC/ERS Guidelines for the diagnosis and treatment of pulmonary hypertension. Eur Heart J. 2022;43(38):3618-3731.
18. . Guidelines for the diagnosis and treatment of pulmonary hypertension. Eur Respir J. 2009;34(6):1219-1263.
19. Kylhammar D, Kjellström B, Hjalmarsson C. A comprehensive risk stratification at early follow-up determines prognosis in pulmonary arterial hypertension. Eur Heart J. 2018;39(47):4175-4181.
20. van der Bruggen CE, Handoko ML, Bogaard HJ. The value of hemodynamic measurements or cardiac MRI in the follow-up of patients with idiopathic pulmonary arterial hypertension. Chest. 2021;159(4):1575-1585.
21. Charloux A, Lonsdorfer-Wolf E, Richard R. A new impedance cardiograph device for the non-invasive evaluation of cardiac output at rest and during exercise: comparison with the “direct” Fick method. Eur J Appl Physiol. 2000;82(4):313-320.