HQ Curve in Centrifugal Pumps #1: Durable LVAD Physiology

1. Why the HQ Curve Matters

A durable left ventricular assist device (LVAD) using a centrifugal pump does not deliver a fixed flow merely because a fixed rotational speed has been selected. Pump output is determined by the interaction between rotational speed, the pressure difference across the pump, and the resistance imposed by the circulation. The head–flow, or H-Q, curve is the fundamental framework for describing this interaction [1].

“H” represents pump head: the pressure rise generated between the pump inlet and outlet. In a conventional LV-to-aorta configuration, it is clinically approximated as:

ΔP = Aortic pressure − Left ventricular pressure

“Q” represents pump flow, usually expressed in liters per minute. At a fixed pump speed, each pressure head corresponds to a particular flow on the device-specific H-Q curve. Flow generally decreases as the pump must overcome a larger pressure difference and increases as that difference falls.

The operating point is therefore not determined by the pump alone. It reflects ventricular filling, native contraction, aortic pressure, systemic vascular resistance, pump speed, cannula geometry, and blood properties. The relevant question is not simply, “What is the displayed flow?” but “What hemodynamic conditions produced this flow?”

2. Definition of Pump Head

Pump head is the hydraulic load across the device:

Pump head = Pump outlet pressure − Pump inlet pressure

For an LVAD, outlet pressure is related to pressure within the outflow graft and systemic arterial circulation, whereas inlet pressure reflects pressure near the inflow cannula within the supported ventricle. The simplified aortic-pressure-minus-LV-pressure equation is useful, but it may become inaccurate when major local pressure losses are present.

Inflow cannula pressure may differ from pressure in the wider ventricular cavity because of flow acceleration, cannula malposition, trabecular or septal contact, ventricular collapse, or partial obstruction. Pump outlet pressure may differ from central aortic pressure when the outflow graft is kinked, compressed, thrombosed, or narrowed.

A high head means that the device must generate a large pressure rise to move blood into the systemic arterial circulation. A low head means that the pressure difference is smaller. These pressure relationships, rather than pump speed alone, determine instantaneous flow.

3. H-Q Curve Shape and Pump Design

At a fixed rotational speed, a centrifugal pump has a downward-sloping H-Q relationship. However, the gradient varies among pump designs and determines how strongly flow changes in response to a change in pressure head.

A relatively flat H-Q curve permits a large change in flow for a modest change in pressure. A steeper curve attenuates the flow response to the same pressure variation. In experimental centrifugal LVAD systems, H-Q curve gradient altered the magnitude and timing of flow pulsatility during the cardiac cycle [2]. A pump with an exceptionally flat pressure–flow curve generated high systolic peak flow at constant speed because the LV-to-aortic pressure difference fell during systole [3].

Two pumps operating at the same mean speed and flow may therefore respond differently to identical changes in preload, afterload, or ventricular pressure. Recent numerical modeling comparing centrifugal devices with different H-Q gradients found that a flatter curve produced greater pulse pressure and greater flow responsiveness during acute preload changes. These findings are model-based and should not be interpreted as proof of clinical superiority [4].

The H-Q curve is thus a device-specific physiologic property. Mean flow alone does not fully characterize heart–pump interaction.

4. Movement Along the H-Q Curve During the Cardiac Cycle

Even when pump speed is constant, aortic and ventricular pressures change continuously. The pressure head therefore varies throughout each heartbeat, moving the operating point along the fixed-speed H-Q curve.

During diastole, LV pressure falls while aortic pressure remains relatively high:

ΔP_D = Aortic diastolic pressure − LV diastolic pressure

The pump operates against a larger head, so instantaneous flow is lower. The diastolic operating point lies higher and farther to the left.

During systole, native LV pressure rises:

ΔP_S = Aortic systolic pressure − LV systolic pressure

The pressure head falls, and pump flow increases. The systolic operating point moves lower and to the right. Under usual conditions:

Q_S > Q_D when ΔP_S < ΔP_D

A centrifugal LVAD therefore provides continuous support while retaining phasic flow variation. The magnitude of this pulsatility depends on the native pressure cycle and the H-Q curve slope. Pressure head–flow loop analysis demonstrates that greater native ventricular contribution broadens the H-Q loop and that timing between ventricular contraction and pump action affects ventricular loading and unloading [5].

5. Native Ventricular Contractility and Flow Pulsatility

Native contraction raises pump inlet pressure during systole, reduces the outlet-to-inlet pressure difference, and increases instantaneous flow. A ventricle producing a larger systolic pressure rise usually creates a larger difference between systolic and diastolic pump flow.

Greater native contribution may be associated with greater pump-flow pulsatility, higher arterial pulse pressure, more frequent aortic valve opening, and greater ventricular volume variation. However, pulsatility is not a direct measurement of contractility. It is influenced by preload, afterload, pump speed, rhythm, ventricular size, aortic valve competence, and curve slope. A flat curve may generate substantial flow pulsatility from modest pressure oscillation, whereas a steeper curve may show limited flow variation despite meaningful contraction.

A reduction in pulsatility may result from deteriorating myocardial function, reduced preload, excessive unloading, increased afterload, arrhythmia, or changed device settings. The waveform identifies altered heart–pump interaction but does not establish the mechanism.

6. Effect of Preload

Preload determines the volume and pressure available at the pump inlet. Increased pulmonary venous return and ventricular filling generally raise inlet pressure, reduce pump head, and increase flow at a fixed speed. Reduced preload lowers ventricular pressure, increases head, and decreases flow.

Causes of preload limitation include hypovolemia, bleeding, right ventricular failure, tamponade, elevated pulmonary vascular resistance, excessive positive-pressure ventilation, venous obstruction, and arrhythmias impairing right-sided output. These conditions can produce similar low-flow findings because each restricts delivery to the supported ventricle.

Rotary pumps have weaker intrinsic preload responsiveness and stronger afterload responsiveness than the native heart [6]. This mismatch creates risks of underpumping during increased demand and overpumping when venous return falls. Frank–Starling-like control strategies have therefore been developed to adjust target flow according to filling pressure or flow pulsatility [7]. In vitro testing showed that adaptive Starling-like control increased preload sensitivity and reduced flow during simulated pulmonary hypertension, avoiding suction that occurred during constant-speed operation [8].

When preload becomes critically low, the ventricle may collapse around the inflow cannula. Local cannula pressure can then fall sharply, producing intermittent inflow obstruction and suction. Average ventricular pressure may no longer represent true pump inlet pressure.

7. Effect of Afterload

LVAD afterload is primarily related to aortic pressure and systemic vascular resistance. When aortic pressure rises while speed and ventricular pressure remain similar, pump head increases and the operating point moves toward lower flow.

Hypertension can therefore reduce pump flow despite unchanged speed and adequate circulating volume. It may also reduce ventricular unloading, increase filling pressure, and contribute to pulmonary congestion. Conversely, vasodilation lowers head and may increase displayed flow without improving preload or myocardial performance.

The same low-flow value therefore has different implications in hypertensive and hypotensive patients. In hypertension, increased outlet pressure may dominate. In hypotension, preload limitation, right ventricular failure, bleeding, tamponade, or mechanical obstruction becomes more likely. Rotary LVAD physiology reflects coupled preload and afterload effects rather than a single flow target [6].

8. Effect of Pump Speed

Increasing rotational speed shifts the family of H-Q curves upward and to the right. At the same pressure head, the pump can generate greater flow, usually reducing ventricular volume and pressure.

Excessive speed may produce ventricular underfilling, leftward septal shift, worsening right ventricular geometry, reduced aortic valve opening, or suction. Reducing speed shifts the curve downward and to the left, lowers flow at a given head, and permits greater ventricular filling and native ejection. If speed is too low, unloading may be inadequate and pulmonary congestion may persist.

Speed optimization is therefore not an attempt to maximize flow. The objective is adequate systemic perfusion with appropriate ventricular geometry, filling pressures, right ventricular performance, and aortic valve behavior.

Programmed speed modulation can alter pulsatility independently of the passive constant-speed H-Q response. In a bovine ischemic heart-failure model, synchronized and asynchronous speed modulation increased arterial pulsatility and produced phasic ventricular unloading compared with constant-speed support [9]. This distinguishes passive pulsatility generated by native pressure changes from active pulsatility generated by speed variation.

9. Displayed Flow and Its Limitations

In many durable rotary LVADs, flow is estimated rather than measured directly. The controller calculates flow from pump speed, power consumption, and a device-specific hydraulic model, often incorporating an assumed blood viscosity.

Displayed flow can therefore be influenced by hematocrit, viscosity, speed, motor power, pressure head, and the estimation algorithm. Controller-derived flow and pulsatility have been unreliable under some operating settings in device-specific testing [10]. Although those observations were obtained with an axial-flow system, the general principle remains relevant: displayed flow is a modeled device parameter, not an independent measurement of systemic cardiac output.

Trends are often more informative than isolated values. Flow should be interpreted with arterial pressure, waveform characteristics, power, speed, echocardiography, filling pressures, lactate, urine output, and clinical perfusion.

10. Important Clinical Exceptions

Aortic Regurgitation

Aortic regurgitation creates a recirculating loop. The pump moves blood from the ventricle into the aorta, but part returns through the incompetent valve. Pump flow may appear normal or high while effective systemic forward output is lower.

Outflow Graft Obstruction

Kinking, compression, thrombus, or anastomotic narrowing increases resistance between the pump and aorta. Flow falls despite unchanged speed, and true pump outlet pressure may exceed central aortic pressure because of the graft pressure loss.

Inflow Obstruction and Suction

Cannula malposition, wall contact, thrombus, or chamber collapse restricts blood entry. Flow may fall abruptly or vary with respiration and body position. During suction, local inlet pressure may be substantially lower than average ventricular pressure.

Pump Stoppage or Very Low Speed

A rotary device does not necessarily behave as a closed valve. If the aortic-to-ventricular gradient is sufficient, retrograde flow may occur through the pump circuit.

11. Bedside Interpretation Framework

The literature supports these pressure–flow principles but does not provide a validated H-Q-based algorithm that reliably distinguishes hypovolemia, right ventricular failure, hypertension, mechanical obstruction, suction, and aortic regurgitation. The following is therefore a physiologic framework rather than a stand-alone diagnostic protocol.

For low flow, assess:

1. Preload delivery: volume status, RV function, pulmonary vascular resistance, venous return, rhythm, ventilation, and tamponade.
2. Afterload: arterial pressure and systemic vascular resistance.
3. Mechanical resistance: inflow cannula position, ventricular collapse, outflow graft patency, and pump thrombosis.
4. Device conditions: speed, power, alarms, and recent parameter changes.

For high flow, consider reduced afterload, increased preload, excessive speed, aortic regurgitation with recirculation, or flow-estimation error.

No single parameter establishes the diagnosis. The operating point must be reconstructed from pressure, ventricular filling, native cardiac function, and circuit resistance.

12. Application to Pediatric and Congenital Heart Disease

The hydraulic principles are unchanged in children and congenital heart disease, but application may be more complex. Small ventricular cavities, systemic right ventricular morphology, restrictive atrioventricular valves, prior intracardiac baffling, abnormal venous pathways, Fontan physiology, and nonstandard cannula orientation can alter preload and local pressure conditions.

In these settings, “LV pressure” should be interpreted more broadly as pressure within the supported systemic ventricle near the pump inlet. Particular attention is required to cavopulmonary pathway obstruction, pulmonary vascular resistance, atrioventricular valve dysfunction, ventricular geometry, and cannula alignment. Published H-Q studies remain predominantly experimental or adult-focused; direct pediatric validation is limited.

13. Core Physiologic Message

At constant speed, a centrifugal LVAD does not provide constant flow. It operates dynamically along a device-specific H-Q curve.

• A larger outlet-to-inlet pressure difference reduces pump flow.
• A smaller pressure difference increases pump flow.
• Native systole and diastole move the operating point along the curve.
• Curve slope determines how strongly pressure variation becomes flow variation.
• Preload, afterload, contractility, speed, cannula conditions, and blood properties influence the waveform.
• Displayed pump flow is not necessarily equivalent to effective systemic cardiac output.

The H-Q curve is the mechanical link between the pump and circulation. Accurate interpretation requires simultaneous consideration of device mechanics and cardiovascular physiology.

References

[1] Moazami N, Fukamachi K, Kobayashi M, Smedira NG, Hoercher KJ, Massiello A, Lee S, Horvath DJ, Starling RC. Axial and centrifugal continuous-flow rotary pumps: a translation from pump mechanics to clinical practice. J Heart Lung Transplant. 2013;32(1):1-11.

[2] Tagusari O, Yamazaki K, Litwak P, Antaki JF, Watach M, Gordon L, Kono K, Mori T, Koyanagi H, Griffith BP, Kormos RL. Effect of pressure-flow relationship of centrifugal pump on in vivo hemodynamics: a consideration for design. Artif Organs. 1998;22(5):399-404.

[3] Yamazaki K, Saito S, Kihara S, Tagusari O, Kurosawa H. Completely pulsatile high flow circulatory support with a constant-speed centrifugal blood pump: mechanisms and early clinical observations. Gen Thorac Cardiovasc Surg. 2007;55(4):158-162.

[4] Yang Y, Wang Z, Xie Z, Yu S, Zou L. Hemodynamic consequences of pressure-flow curve gradient variations in continuous-flow ventricular assist devices. Front Physiol. 2026;16:1730883.

[5] Jahren SE, Ochsner G, Shu F, Amacher R, Antaki JF, Vandenberghe S. Analysis of pressure head-flow loops of pulsatile rotodynamic blood pumps. Artif Organs. 2014;38(4):316-326.

[6] Pauls JP, Stevens MC, Bartnikowski N, Fraser JF, Gregory SD, Tansley G. Evaluation of physiological control systems for rotary left ventricular assist devices: an in-vitro study. Ann Biomed Eng. 2016;44(8):2377-2387.

[7] Stevens MC, Gaddum NR, Pearcy M, Salamonsen RF, Timms DL, Mason DG, Fraser JF. Frank-Starling control of a left ventricular assist device. Annu Int Conf IEEE Eng Med Biol Soc. 2011;2011:1335-1338.

[8] Gaddum NR, Stevens M, Lim E, Fraser J, Lovell N, Mason D, Timms D, Salamonsen R. Starling-like flow control of a left ventricular assist device: in vitro validation. Artif Organs. 2014;38(3):E46-E56.

[9] Soucy KG, Giridharan GA, Choi Y, Sobieski MA, Monreal G, Cheng A, Schumer EM, Slaughter MS, Koenig SC. Rotary pump speed modulation for generating pulsatile flow and phasic left ventricular volume unloading in a bovine model of chronic ischemic heart failure. J Heart Lung Transplant. 2015;34(1):122-131.

[10] Lund LH, Gabrielsen A, Tirén L, Hallberg AC, El Karlsson K, Eriksson MJ. Derived and displayed power consumption, flow, and pulsatility over a range of HeartMate II left ventricular assist device settings. ASAIO J. 2012;58(3):183-190.