This investigation selected male adult Small-Tailed Han sheep due to their physiological characteristics and cardiovascular system exhibiting certain similarities to humans, making them a well-established model for cardiovascular device evaluation (9). A total of six male adult Small-Tailed Han sheep (weight: 45–55 kg, age: 2–3 years) were included in this study. All animals completed the experimental protocol without early termination. Inter-animal variability was assessed by calculating the coefficient of variation (CV) for key hemodynamic parameters (e.g., EF, diastolic proportion) across animals under baseline conditions. The CV ranged from 8% to 15%, indicating acceptable inter-subject consistency within this acute model. All experimental animals underwent comprehensive pre-operative physical examinations, including echocardiography and blood parameter analysis (e.g., complete blood count, biochemistry), to confirm the absence of structural cardiac anomalies and systemic diseases. The experimental animal protocol used in this study was approved by the Animal Ethics Committee of the Binhai New Area Medical Experimental Animal Center and strictly adhered to the animal protection regulations of the Binhai New Area Medical Experimental Animal Center.

The essential reagents for this investigation included: succinylcholine (Shanghai Xudong Haipu Pharmaceutical Co., Ltd., anesthetic), sevoflurane (Shanghai Hengrui Pharmaceutical Co., Ltd., inhalational anesthetic), lidocaine hydrochloride injection (Tianjin Jinyao Pharmaceutical Co., Ltd., local anesthetic/antiarrhythmic), dopamine hydrochloride injection (Shanghai Hefeng Pharmaceutical Co., Ltd., vasoactive drug), epinephrine (Tianjin Jinyao Pharmaceutical Co., Ltd., vasoactive drug), norepinephrine injection (Shanghai Hefeng Pharmaceutical Co., Ltd., vasoactive drug), amiodarone hydrochloride injection (Zhejiang Chuangxin Biological Co., Ltd., antiarrhythmic), 10% glucose injection (Otsuka Pharmaceutical Co., Ltd., China, fluid replacement). The vasoactive drugs (dopamine, epinephrine, norepinephrine) were used for blood pressure regulation.

The main instruments included: the HeartCon LVAD system (Rocor Medical, China), MP150 multi-channel physiological recording system (BIOPAC Systems Inc., U.S.), flow sensor (Transonic Systems Inc., U.S.), cardiac pacemaker (Medtronic, Inc., U.S.), color Doppler ultrasound diagnostic system (GE Vingmed Ultrasound AS), and the implantable pressure probes (for LVP, AOP, and LAP measurement).

The experiment was conducted in the operating room of the Binhai New Area Medical Experimental Animal Center. After successful anesthesia, thoracotomy was performed on the experimental sheep to expose the heart and aorta. A coring needle was used to create an opening at the apex of the left ventricle, and a ventricular sewing ring was sutured in place. After inserting the HeartCon pump, a vascular graft was anastomosed to the distal ascending aorta. To accurately monitor hemodynamic parameters, an aortic pressure (AOP) probe was placed in the aorta, 1–2 cm distal to the graft anastomosis; a flow sensor was placed around the graft near the anastomosis; a left ventricular pressure (LVP) probe was inserted into the left ventricle via the apex; and a left atrial pressure (LAP) probe was placed in the left atrium via puncture of the right atrium. All probes were connected to the MP150 physiological recording system.

While maintaining stable basic physiological parameters of the experimental animals, the pump rotational speed was adjusted to 2,400, 2,500, 2,600, 2,700, and 2,800 RPM, covering the typical operating range for the HeartCon pump in preclinical settings based on preliminary tests and manufacturer recommendations, and ensuring a comprehensive assessment of speed-dependent effects on hemodynamics to investigate the effects of different rotational speeds on the flow waveform and cardiac cycle. The cardiac pacemaker frequency was adjusted to stabilize the experimental animals' heart rates at 62, 76, 90, 105, and 119 BPM, representing a clinically relevant spectrum from bradycardia to tachycardia often encountered in heart failure patients managed with LVADs (6, 8) to investigate the effects of different heart rates on the flow waveform and cardiac cycle. By adjusting the infusion rate of vasoactive drugs, the experimental animals' mean arterial pressure (MAP) was regulated to 67, 75, and 85 mmHg to investigate the effects of different blood pressures on the flow waveform and cardiac cycle. Under each set condition, physiological parameters such as pump flow, LAP, LVP, AOP, and ECG were simultaneously collected, and the flow waveform was partitioned using the proposed HeartCon partitioning method to analyze the correspondence between the cardiac cycle and the flow waveform, and to calculate the time proportion of systole and diastole in the entire cardiac cycle.

Hemodynamic information such as pump flow (obtained through a flow sensor), LAP, LVP, and AOP, as well as ECG signals, pump current, and voltage, were simultaneously collected using the MP150 multi-channel physiological recording system. The sampling frequency for all signals was set to 1,000 Hz.

The relative error (δ) was used to evaluate the difference between the EF values obtained by the HeartCon and HVAD partitioning methods and the ultrasound EF values (EFUS), calculated using (Equation 3). To provide a more comprehensive statistical validation, we also performed Bland-Altman analysis to assess agreement between methods, calculated the root mean squared error (RMSE) and mean absolute error (MAE), and computed Pearson correlation coefficients. Paired t-tests were used to compare EF values derived from the HeartCon method versus ultrasound, with p < 0.05 considered statistically significant. These analyses were conducted using SPSS version 26.0 (IBM Corp., Armonk, NY, USA).δ=(EF¯−EFUS)EFUS×100%(3)where, EF¯ is the calculated average EF value using a specific partitioning method.

Comparative analysis revealed that under different rotational speeds, the absolute relative errors of EF values obtained by the HeartCon partitioning method were significantly below 5%, while the absolute relative errors in the HVAD group exceeded 40%. Furthermore, the errors in the HeartCon group showed alternating positive and negative values, suggesting predominantly random errors, whereas the HVAD group exhibited consistent negative errors, indicating a need for systematic correction. As presented in Figure 4, the EF values measured by the HeartCon partitioning method demonstrated good agreement and high repeatability with the ultrasound results. Statistical analysis demonstrated excellent agreement between the HeartCon-derived EF and ultrasound EF, with a mean bias of 0.8% (95% limits of agreement: −3.2% to 4.8%) in Bland-Altman analysis, RMSE of 2.1%, MAE of 1.7%, and a correlation coefficient of r = 0.96 (p < 0.001). In contrast, the HVAD method showed poor agreement (mean bias: −18.5%, 95% LoA: −32.1% to −4.9%, RMSE: 20.3%, MAE: 18.7%, r = 0.42, p = 0.12). These results confirm the superior accuracy and reliability of the proposed HeartCon partitioning method.

Our findings align with and extend prior reports on continuous-flow LVAD hemodynamics. Similar to HeartMate II/III studies, we observed attenuation of atrial pressure waves (a, c, v waves) in the LAP curve, consistent with ventricular unloading and reduced atrial stretch. The flat-topped flow waveform (QMAX) resembles the “flow plateau” reported in HVAD studies during high-speed operation or hypovolemia. However, unlike HVAD suction events characterized by sharp flow dips and elevated pump power, we did not observe overt suction in this acute setting, possibly due to careful speed titration. The shortened diastolic proportion is a novel observation specific to HeartCon, suggesting device-specific hemodynamic interactions that warrant comparison with HeartMate II/III data in future studies. Overall, our results reinforce that LVAD support fundamentally alters cardiac timing and pressure morphology, but device-specific features must be recognized for accurate clinical interpretation.

This investigation methodically assessed the effects of changes in heart rate, blood pressure, and LVAD rotational speed on the HeartCon flow waveform and cardiac cycle by controlling these parameters in experimental animal. The results show that an increased heart rate significantly shortens the proportion of the cardiac cycle occupied by diastole, potentially leading to insufficient ventricular filling, which in turn may affect pump flow and reduce the amplitude of aortic valve opening. Under hypotensive conditions, the pump's afterload is reduced, resulting in increased pump flow. However, a decrease in the diastolic phase proportion was also observed, possibly due to enhanced pump aspiration force affecting ventricular diastolic function. When the LVAD rotational speed is increased, both the pump's output flow and aspiration force increase, further shortening the diastolic phase proportion (11), suggesting that excessively high pump speeds may excessively unload the left ventricle and interfere with its normal filling. Notably, the diastolic phase proportion was significantly lower than the 58% in a normal heart (15) under all tested conditions, which may be related to the hemodynamic changes after LVAD implantation and the continuous flow support provided by the pump. The observed shortening of diastole raises the hypothesis that it may adversely affect myocardial perfusion, since coronary perfusion occurs predominantly during diastole, driven by the aortic diastolic pressure that generates the critical perfusion gradient across the coronary vascular bed (16–20). While physiological and supraphysiological augmentation of diastolic pressure and duration are recognized mechanisms for enhancing coronary flow (20, 21–23), whether the abbreviated diastole under HeartCon support directly compromises coronary perfusion remains speculative and requires direct measurement of coronary flow in future studies. Previous studies have shown that in normal physiological conditions, the duration of ventricular systole is approximately 330 ± 42 ms, and diastole is approximately 469 ± 76 ms. Diastole is significantly longer than systole, accounting for about 58% of the cardiac cycle, which is considered crucial for maintaining sufficient cardiac preload and ensuring effective coronary artery perfusion (15). For HF patients, appropriately prolonging diastole has been associated with improved left ventricular reverse remodeling and patient prognosis (24). Our data suggest that the optimal diastolic phase proportion for patients with implanted LVADs may differ from the normal reference value of 58%; however, this remains a hypothesis that requires validation in future clinical studies. The associations between prolonged tachycardia, persistent hypotension, or indiscriminate increases in LVAD rotational speed and impaired postoperative cardiac function recovery in HF patients (6) are plausible but speculative based on our acute experimental data, warranting confirmation in chronic clinical settings.

The phenomenon of T-wave inversion observed in the ECG during the experiment is consistent with findings in previous studies using other animal models (25). T-wave inversion observed in our experiments likely resulted from multiple factors specific to our model: (1) anesthetic agents such as sevoflurane and opioids are known to affect ventricular repolarization; (2) surgical trauma from apical coring and suturing may cause local myocardial injury or edema, altering depolarization/repolarization vectors; (3) hemodynamic shifts due to LVAD unloading could change wall stress and consequently affect repolarization; (4) possible electrolyte imbalances from fluid shifts or drug infusions (e.g., dopamine) may contribute. While we did not perform detailed electrophysiological mapping, the consistent presence of T-wave inversion across animals suggests a reproducible effect of the acute implant model rather than random variation. Future studies incorporating continuous electrolyte monitoring and high-resolution mapping could elucidate the precise mechanisms (26). The specific mechanisms of T-wave inversion were not deeply investigated in this research and require further electrophysiological studies to elucidate.

The electromechanical delay (EMD), typically 25–40 ms in humans and similar large animals, represents the interval between electrical depolarization (R-wave) and mechanical contraction onset (pressure rise). In our study, we observed an EMD of 30 ms between the ECG R-wave and the rapid flow upstroke, consistent with published physiology. This delay has important implications for clinical waveform interpretation: it means that the flow waveform lags behind electrical events. Therefore, using ECG landmarks (R-wave peak, T-wave end) as temporal anchors for partitioning inherently accounts for this physiological delay, ensuring that mechanical phases are correctly aligned. Ignoring EMD could lead to misidentification of systolic onset, especially in patients with conduction abnormalities or on inotropic support. Future implementations of automated waveform analysis algorithms should incorporate EMD correction for enhanced accuracy (27). This physiological phenomenon was considered when analyzing the relationship between the ECG and the flow waveform. However, since a time interval of 30 ms is difficult to visually distinguish directly on the HeartCon monitor interface, for the convenience of clinical application and waveform interpretation, this study simplifies the point of rapid flow increase as the onset of systole.

The primary significance of constructing the HeartCon cardiac Wiggers diagram lies in establishing a preliminary temporal relationship between the flow waveform and key hemodynamic parameters, clearly demonstrating the significant differences between cardiac function after LVAD implantation and that of a normal heart. This lays the foundation for subsequent research on the characteristic patterns of flow waveforms in LVAD patients under different physiological or pathological conditions. These characteristics may manifest as specific waveform morphologies, changes in key time points, or abnormalities in certain characteristic parameters.

Notably, the presence of a flat-topped QMAX characteristic in the flow waveform may indicate that the pump is operating near its maximum output under the tested conditions; however, this interpretation is preliminary and requires further validation under varied loading states. The rotational speeds used (2,400–2,800 RPM) in this acute animal model were selected based on preliminary tests and manufacturer recommendations to cover the typical operating range for the HeartCon pump in preclinical settings. While the ovine cardiovascular system shares similarities with humans (6), the acute nature of this study and potential species-specific differences mean that the absolute speed values and their precise hemodynamic effects may not directly and fully replicate the chronic adaptive physiology of the human heart with an implanted LVAD. Future chronic studies in appropriate models are needed to better simulate long-term human physiology. Further analysis indicates that the peak of AOP occurs in the middle of the flow waveform peak, implying that the rapid rise in pump flow precedes the peak of ventricular systolic energy; during the sustained rise of AOP, the pump flow remains relatively stable, further supporting the inference that the pump is operating at maximum output. In situations where myocardial contractility is still adequate but circulatory volume is insufficient, Q_MAX is expected to terminate at the peak of the AOP waveform, with a narrower width compared to that observed under volume-replete conditions. Future research could focus on investigating the flow waveform characteristics under different combinations of pump operating parameters and various physiological states to establish quantitative assessment criteria for parameters such as Q_MAX width, providing clinicians with a reference for preliminary judgment of patients' volume status based on the flow waveform. Regarding the potential application of this method to assess ejection fraction in normal hearts without LVADs using pulse pressure information, the current partitioning method is specifically designed for and validated in the context of continuous-flow LVAD support, where the pump flow waveform morphology is fundamentally different from the arterial pressure waveform or native cardiac output patterns. Therefore, direct extrapolation to non-LVAD hearts is not supported by the present data and would require separate methodological development and validation.

However, this research also has certain limitations. First, the observation period in the animal experiments was relatively limited, which may not fully simulate the adaptive changes of the heart after long-term LVAD implantation. Second, the strong aspiration force generated by the pump may have masked some subtle changes in the flow waveform, limiting the observation and analysis of more nuanced physiological events. As shown in Figure 8, compared with the normal cardiac Wiggers diagram (11), the characteristic a, c, and v waves reflecting atrial pressure changes are markedly attenuated in the HeartCon cardiac Wiggers diagram. On the current flow waveform, only the feature point corresponding to mitral valve closure and the associated abrupt rise in flow can be identified, while the opening and closing points of the other valves cannot be clearly annotated. Furthermore, key parameters in heart failure management such as central venous pressure (CVP) and pulmonary capillary wedge pressure (PCWP), which reflect right heart function and left-sided filling pressures respectively, were not extensively analyzed in relation to the flow waveform in this initial study. These limitations suggest that certain abnormal physiological conditions, including those related to venous return and biventricular interactions, may not be directly evident in the current flow waveform, and a comprehensive assessment combining other monitoring methods, such as simultaneous CVP and PCWP measurement, will be necessary in the future.