Understanding PDA Spectral Doppler

Spectral Doppler interrogation is a cornerstone of echocardiographic evaluation of a patent ductus arteriosus (PDA). Beyond simply confirming ductal patency with two-dimensional and color Doppler imaging, spectral Doppler allows detailed assessment of shunt direction, velocity, and hemodynamic significance. Using pulsed-wave (PW) or continuous-wave (CW) Doppler aligned with ductal flow, clinicians can characterize shunt physiology and gain insight into the instantaneous pressure relationships between the systemic and pulmonary circulations. A restrictive PDA is one in which the effective ductal diameter is sufficiently small that flow and pressure transmission are limited by the ductal narrowing itself. In this setting, shunt magnitude is constrained by ductal resistance, and high velocities reflect a large pressure gradient across a narrow conduit. In contrast, an unrestrictive PDA is large enough that intrinsic ductal resistance is negligible, and shunt volume is determined primarily by the instantaneous systemic–pulmonary pressure gradient, as well as by the relative impedance and compliance of the two vascular beds. Importantly, the concept of ductal “restrictiveness” is not binary. Ductal shunting is governed by continuous variables—effective diameter, length, geometry, and dynamic smooth muscle tone—interacting with continuously changing systemic and pulmonary vascular resistance, impedance, and compliance. As these determinants evolve over the cardiac cycle and over postnatal time, the PDA transitions progressively from an unrestrictive to a restrictive conduit. Restrictive and unrestrictive PDAs therefore represent points along a physiologic continuum rather than discrete states. The ductus arteriosus provides a dynamic window into the interaction between the pulmonary artery and aorta throughout the cardiac cycle. During systole, when ventricular ejection raises pressures in both great arteries, ductal flow reflects the relative timing and magnitude of pressure rise, ventricular–vascular coupling, and the balance between left and right ventricular output. During diastole, differences in pressure decay dominate ductal flow behavior: pressure falls more slowly in the stiffer systemic circulation and more rapidly in the typically more compliant pulmonary vascular bed, often shaping late-diastolic and presystolic shunt velocity profiles. Because the ductus is a passive conduit, it does not generate flow but faithfully transmits moment-to-moment hydraulic relationships between the two circulations. As such, the PDA acts as a real-time integrator of ventriculo-arterial interaction, vascular compliance, impedance, and transitional physiology rather than a static anatomic lesion. The ductus itself may change caliber from beat to beat, and even within a single cardiac cycle, with a tendency toward larger diameter during systole and relative narrowing during diastole or during phases when pulmonary and systemic pressures approach equilibrium. Consequently, velocity and pressure gradients across the ductus reflect a continuously evolving balance between aortic and pulmonary arterial pressures and vascular properties, underscoring why PDA spectral Doppler patterns must be interpreted dynamically across the entire cardiac cycle rather than at a single time point.

When resistance is high and flow is maintained by adequately coping ventricle(s), pressure upstream will increase. In contrast, if resistance is high but flow begins to fall, such as in functional pulmonary valvar stenosis or atresia on the right ventricular side, or in severe left ventricular dysfunction with markedly reduced or absent output, pressure in the corresponding vessel may actually decrease. When resistance remains appropriate but flow increases, pressure rises, whereas a reduction in flow leads to a corresponding fall in pressure. In a narrow, restrictive PDA, the ductus itself provides substantial intrinsic resistance that limits pressure equalization between the aorta and pulmonary artery, thereby maintaining a pressure gradient across the vessel. When downstream resistances differ markedly, for example when systemic vascular resistance greatly exceeds pulmonary vascular resistance, flow across the ductus becomes high velocity and turbulent. If downstream resistances are similar, velocities may be very low because the pressure gradient is minimal, even if the ductus has already become significantly narrowed. Conversely, in an extremely large, free-flowing PDA, the ductus offers negligible resistance and behaves as a wide communication that allows aortic and pulmonary pressures to equalize. In this situation, because the restriction at the ductal level is absent, flow becomes predominantly laminar and is governed by the relative downstream resistances of the systemic and pulmonary circulations, often resulting in large-volume shunting. Poiseuille’s law describes flow as being proportional to the pressure gradient and inversely related to resistance, emphasizing that resistance modulates flow magnitude and waveform characteristics but does not determine flow direction. Direction remains dependent on which side has the higher pressure at any given moment, while ductal resistance and downstream vascular resistance influence how large the pressure gradient becomes and how quickly it dissipates over the cardiac cycle.

Important precision by Dr. Rohit Loomba: It is important to remember that pressure is determined by the product of flow and resistance (Pressure = Flow × Resistance). As a result, a measured pressure gradient can change even when the anatomical size of a duct remains unchanged. If flow through the duct decreases, the pressure gradient across it may fall; conversely, an increase in flow through the same duct can result in a higher pressure gradient. In clinical practice, we often intuitively associate pressure with flow when interpreting systemic arterial pressure (such as mean arterial pressure), yet tend to associate pressure with resistance when evaluating shunts or obstructions. This reflects a common but inconsistent simplification of hemodynamic principles. From a physiological standpoint, the Doppler velocity–time integral, or the area under the spectral Doppler curve, more directly reflects flow and can, at least theoretically, be used to quantify changes in shunt flow over time.

• The order and timing of these events can change substantially in the presence of pulmonary hypertension due to high pulmonary vascular resistance.  In the context of a suprasystemic pulmonary hypertension throughout the cardiac cycle, where pulmonary vascular resistance is above systemic vascular resistance in systole and diastole, the PDA will have a continuous right to left profile. Many of these neonates may have accompanied altered right ventricular performance, which can prolong RV systolic ejection, a phenomenon often reflected by an increased systolic-to-diastolic time ratio derived from tricuspid regurgitation Doppler (an echocardiographic marker discussed in the pulmonary hypertension section). In significant or suprasystemic pulmonary hypertension, onset of right ventricular ejection may be slightly delayed due to the higher afterload required to open the pulmonary valve, leading to variability regarding its relationship to the onset of LV ejection. Further, peak pulmonary artery pressure may occur before, during, or after peak aortic pressure because of the high PVR status. Also, pulmonary artery pressure remains higher than aortic pressure throughout the cardiac cycle and in this context, the normal A2–P2 sequence may be lost. The pulmonary valve closure may, as such, occur simultaneously with or even before aortic valve closure, resulting in a fused or reversed second heart sound. Clinically, this manifests as a loud P2 due to forceful valve closure and often a single or narrowly split S2 with little respiratory variation. These acoustic phenomenons may not be appreciated easily at fast neonatal heart rates. All of these timing subtleties directly influence instantaneous PDA Doppler velocities by altering the moment-to-moment pressure relationship between the aorta and pulmonary artery, as well as how effectively those pressures are transmitted across the ductus depending on its size, geometry, blood viscosity and intrinsic ductal resistance to flow.

3. Electrical Systole Does Not Define Mechanical Systole

Mechanical ventricular systole does not extend to the end of the T wave. Peak ventricular pressure and peak arterial pressure typically occur during the early phase of ventricular repolarization, rather than at the apex of the T wave. Because PDA Doppler velocities are often referenced to the ECG, there is a risk of mislabeling velocities as “systolic” when they do not coincide with true peak mechanical systole or maximal arterial pressures.

4. Ductal Flow Is Impedance-Driven, Not Valve-Driven

Flow through the ductus arteriosus is governed by impedance rather than by valvular mechanics. The PDA is not a fixed orifice and does not fully satisfy the assumptions required for strict Bernoulli application. Ductal flow reflects the interaction of instantaneous pressure differences, ductal size and geometry, pulmonary vascular impedance, and the compliance of both arterial systems. The estimation of Right Ventricular Systolic Pressure (RVSP) via a PDA-derived gradient can result in both underestimation and overestimation. 

5. PDA-Based sPAP Estimates Tend to Underestimate True RVSP,  but can also Overestimate

When PDA peak velocity is used to estimate systolic pulmonary artery pressure by subtracting the Doppler-derived gradient from systolic blood pressure, the result often underestimates true right ventricular systolic pressure. This occurs because the peak PDA velocity is frequently measured at a time when aortic pressure has not yet reached its maximum or when pulmonary artery pressure has already begun to fall. The gradient therefore reflects a submaximal pressure difference rather than true peak systolic separation. Overestimation of RVSP occurs if the PDA gradient is underestimated. The most common cause is Doppler beam misalignment; if the ultrasound beam is not perfectly parallel to the blood flow, the recorded velocity will be lower than reality. A lower velocity produces a smaller gradient, which, when subtracted from the SBP, leaves a falsely high (overestimated) RVSP. Similarly, if the systemic blood pressure used in the calculation is measured at a different time than the echocardiogram and is falsely elevated, the resulting RVSP will be overestimated. In the setting of a right-to-left PDA, the Doppler-derived velocity also represents an instantaneous pulmonary artery–to–aorta pressure gradient. The peak right-to-left velocity may occur during a phase of the cardiac cycle that does not coincide necessarily with the peak aortic pressure (or peak pulmonary arterial pressure). In this situation, the measured gradient reflects a momentary "higher" dominance of pulmonary pressure rather than the gradient at peak aortic systole (used for the assumption of: sBP + right to left gradient is about sPAP). Because of this timing mismatch, using the peak right-to-left PDA velocity to infer systolic pulmonary artery pressure can lead to also overestimation of the true systolic PA pressure. 

6. Directional Flow Changes Reflect Instantaneous Dominance

Bidirectional or transient right-to-left components of PDA flow reflect instantaneous pressure dominance rather than mean or sustained pressure relationships. Early systolic right-to-left flow may occur when pulmonary vascular resistance briefly exceeds systemic resistance, even without suprasystemic pulmonary pressures at peak systole. As systole and diastole progress, left-to-right flow typically dominates as aortic pressure remains higher while pulmonary artery pressure decays more rapidly. Net shunt direction over the cardiac cycle does not capture these instantaneous dynamics.

7. Peak Velocity Does Not Equal Peak Pressure Difference

The maximal Doppler velocity across the ductus does not necessarily correspond to the maximal physiologic pressure difference between the aorta and pulmonary artery (as the peak pressure difference may occur at different moment of the cardiac cycle). Indeed, the peak pulmonar Larger gradients may occur early in systole when pressure curves diverge rapidly, while smaller gradients may be present at the time of true peak arterial pressures. End-diastolic gradients reflect diastolic pressure separation and should not be interpreted as systolic load. Timing, rather than magnitude alone, determines physiologic meaning.

Dicrotic Notch in the context of Pressure curves in a PDA

• Dicrotic Notch: The dicrotic notch is the small dip seen on the descending limb of the arterial pressure waveform that marks the transition from ventricular systole to diastole, occurring just after the peak systolic pressure when the aortic or pulmonary valve closes and flow briefly reverses toward the heart, causing a transient fall and then rebound in aortic/pulmonary pressure. This notch corresponds temporally to early isovolumetric relaxation, when ventricular pressure has fallen below the great vessel pressure and the semilunar valve snaps shut to prevent further backflow. Its shape and prominence are influenced not only by valve closure itself but also by pressure wave reflections from the peripheral arterial tree, so it can vary with changes in vascular resistance and arterial stiffness. In a normal heart, these notches occur at roughly similar timings but at vastly different pressure levels. In a restrictive PDA, the Ao and PA curves remain distinct. 

  • As the PDA becomes large and non-restrictive, in the context of low PVR/SVR ratio the pressures begin to equalize, but the "notches" may still diverge in appearance and timing (or even disappear) due to: 

    • Aortic Runoff: The aorta loses volume into the lower-resistance pulmonary bed throughout the cycle. This "stealing" of blood causes the aortic dicrotic notch to appear lower on the pressure scale and the diastolic limb to decay more steeply (collapsing pulse). Indeed, the diastolic steal will dampen or abolish the normal dicrotic notch in peripheral arterial waveforms, because of diastolic runoff into the pulmonary circulation.

    • Pulmonary Augmentation: Conversely, the PA receives a significant surge of flow. This increases the volume and pressure in the PA during late systole and early diastole, often making the pulmonic dicrotic notch appear more prominent or "elevated" compared to a normal PA tracing. Because of continuous left to right flow, the notch may even dampen. 

    • From Reference: "Infants with larger PDA had the dicrotic notch displayed towards the end of the dicrotic limb which can be attributed to the lower overall arterial resistance due to higher run off from the large PDA". They also "had a smoother dicrotic limb without superimposed pressure perturbations".

  • Restrictive/small PDA: Notch generally preserved.

  • In a large unrestrictive right-to-left PDA (typically with high pulmonary vascular resistance exceeding systemic, e.g., Eisenmenger physiology), substantial flow shunts from the pulmonary artery to the descending aorta throughout the cycle. This leads to pulmonary artery volume/pressure loss ("runoff" or stealing), causing the PA diastolic limb to decay more steeply with a lower, shallower, or absent pulmonic dicrotic notch due to reduced diastolic pressure maintenance. Conversely, the aorta experiences augmented diastolic filling from the high-resistance pulmonary bed, elevating aortic diastolic pressure and raising/prominently "elevating" the aortic dicrotic notch relative to normal, with a slower diastolic decay and potentially wider pulse pressure in the descending aorta. The notches may diverge further in timing and appearance due to asymmetric wave reflections and shunt dynamics, though central aortic pressures proximal to the PDA remain relatively spared. In summary, in the right‑to‑left PDA, the dicrotic notch in both great arteries may be less distinct because of altered downstream resistances and wave reflections.

  • I have assumed the presence of a dicrotic notch in some of the theoretical examples below, in which I drew expected Aortic and PA pressure curves based on PDA velocity profile and various assumptions (note of author: Dr Gabriel Altit).

Left to right profiles

PDA tracing is an instantaneous flow between two pulsatile pressure reservoirs (Ao and PA), where the driving force is the time-varying Ao–PA pressure difference and the timing of ventricular ejection into each reservoir. Early systole: a brief moment when ΔP(Ao–PA) widens fast Right after the QRS, both ventricles are electrically activated, but mechanical ejection doesn’t start at the same instant in the Ao and PA. The LV/aortic pressure upstroke may begin slightly earlier (or rise faster initially) than the RV/PA pressure upstroke. For a short window, aortic pressure is already climbing while PA pressure is still near its end-diastolic level (or has not yet risen much). That creates a transiently larger ΔP = Ao − PA, so the PDA flow accelerates abruptly → the first narrow velocity spike. RV ejection starts (or PA pressure rises sharply) → gradient narrows → velocity drops. Then the PA “catches up”. Once pulmonary valve opening / RV ejection is underway, PA pressure rises quickly. If PA pressure rises faster than aortic pressure at that moment (or simply rises enough to reduce the gap), then ΔP(Ao–PA) shrinks. Since ductal velocity tracks that driving pressure difference, the ductal envelope decelerates, giving the dip after the spike. What we are noticing “below the baseline” - flow acceleration that starts at same time as the onset of this first spike - can actually support this: the onset of adjacent-vessel ejection flow marks the point when RV/PA ejection is happening, which is exactly when the PDA gradient transiently narrows and the PDA velocity drops. Than Δ velocity widens again → ductal velocity builds to a second, broader peak Later in systole, it’s common for the relative shapes of the pressure curves to diverge again: Aortic pressure may continue rising toward its peak, or remain relatively high. PA pressure may plateau earlier or begin to fall earlier, especially if RV ejection is shorter or pulmonary runoff is high. Net effect: Ao − PA increases again, so the ductal velocity ramps up again → the broader second systolic rise/peak. In a large ductus, the flow is less “jet-like” and more like a short conduit between two compliant pulsatile systems. That makes it especially sensitive to: small timing differences in ejection, differences in dP/dt (how fast Ao vs PA pressure rises), and late-systolic divergence in pressure decay. As such, this early separation is generated due to a short-lived increase in ductal velocity due to wideneing of the Ao-PA gradient at early systole, which would then be dampened once right ventricular ejection build-up and pulmonary artery pressure rises. As the cardiac cycle progresses, continued pressure buildup between the two vessels could lead to a second, larger, and more physiologically dominant peak in the ductal velocity gradient. 

Bidirectional profiles