Shape of the Pulmonary Artery Doppler‐Flow Profile Predicts the Hemodynamics of Pulmonary Hypertension Caused by Left‐Sided Heart Disease

ABSTRACT

Background

Previous studies demonstrated a relationship between pulmonary hemodynamics and shape of pulmonary artery (PA) Doppler‐flow profiles in a mixed pulmonary hypertension (PH) cohort.

Hypothesis

Shape of PA Doppler‐flow profiles could illustrate the hemodynamic characteristics of pulmonary venous hypertension (PVH), especially identifying it with or without pulmonary vascular disease (PVD).

Methods

We retrospectively analyzed hemodynamic, echocardiographic, and clinical data from 47 patients referred for PH caused by left‐sided heart disease (PH‐LHD). All patients underwent right‐sided heart catheterization within 1 week of echocardiography. We concluded a simple echocardiographic prediction rule to give hemodynamic differentiation of PVH with PVD, defined as capillary wedge pressure >15 mm Hg and pulmonary vascular resistance (PVR) >3 Wood units (WU). The PA Doppler‐flow profiles were categorized into 2 groups, no notch (NN) and MSN/LSN.

Results

The PVR was higher in the MSN/LSN group at 4.04 WU (interquartile range, 3.1–5.3) vs the NN group at 1.91 WU (interquartile range, 1.8–3.0; P < 0.001). Pulmonary artery Doppler‐flow notching (MSN and LSN) was highly associated with PVR >3 WU, whereas the NN pattern predicted a PVR ≤3 WU (odds ratio: 19.8, 95% confidence interval: 4.3‐91.3) and normal transpulmonary pressure gradient ≤12 mm Hg (odds ratio: 4.7, 95% confidence interval: 1.3‐16.2). The NN pattern had 74% specificity and 88% sensitivity for PVR ≤3 WU.

Conclusions

Absence of PA Doppler‐flow notching was highly associated with PVH, and a notching pattern indicated PVH with PVD in the PH‐LHD cohort.

Introduction

Left‐sided heart disease (LHD) is the most frequent cause of pulmonary hypertension (PH); PH is also a common complication of LHD.1 The current hemodynamic definition of PH caused by left‐sided heart disease (PH‐LHD) combines a mean pulmonary artery pressure (mPAP) ≥25 mm Hg, a pulmonary capillary wedge pressure (PCWP) >15 mm Hg, and a normal or reduced cardiac output.2, 3

In the “isolated post‐capillary” stage, PH‐LHD is merely a reflection of the increase in downstream left‐sided heart pressure (transpulmonary pressure gradient [TPG] ≤12 mm Hg), and pulmonary vascular resistance (PVR) is normal or near normal. When it progresses to a “combined post‐capillary PH and pre‐capillary PH” stage, with increased TPG (>12 mm Hg) and PVR, PH‐LHD was due to severity and chronicity of left atrial (LA) hypertension, leading to pulmonary vascular remodeling.2, 4 The 2 hemodynamic profiles in PH‐LHD are difficult to distinguish between in a noninvasive manner. The current hemodynamic definition of pulmonary venous hypertension (PVH) with pulmonary vascular disease (PVD) is defined as mPAP ≥25 mm Hg, PCWP >15 mm Hg, and PVR >3 Wood units (WU).2, 5, 6 Right‐sided heart catheterization (RHC) is the gold standard used for hemodynamic evaluation, but it is an invasive procedure.3 An understanding of the noninvasive manifestations of PH‐LHD is critically important prior to an invasive approach for patients with undifferentiated PH.

The distinction between pulmonary venous hypertension (PVH) with or without PVD is critical given the considerable differences in diagnostic and treatment strategies. Those with high PVR are especially prone to right ventricular (RV) dysfunction, a critical determinant of patient morbidity and mortality.7, 8, 9 A previous study has shown that the pulmonary artery (PA) mid‐systolic Doppler‐flow profile could identify PVD and right‐heart dysfunction in a mixed PH cohort.10 In this study, we reviewed the clinical significance of differing patterns in the PA Doppler‐flow profile in a PH‐LHD cohort. We hypothesized that the absence of PA Doppler‐flow notching could indicate PVH, whereas the presence of a Doppler notching pattern, either mid‐systolic notch (MSN) or late‐systolic notch (LSN), could indicate the presence of PVH with PVD.

Methods

We retrospectively analyzed consecutive patients seen by the PH center at Shanghai Pulmonary Hospital, Tongji University School of Medicine, between March 2011 and June 2013. The patients underwent a clinically indicated RHC and a transthoracic echocardiogram (TTE) within 1 week. According to the Dana Point classification, the group 2 PH data were divided into 3 distinct subcategories based on etiology: left‐heart systolic dysfunction, left‐heart diastolic dysfunction, and left‐heart valvular disease.2, 11 An mPAP ≥25 mm Hg and a PCWP >15 mm Hg, evaluated by RHC, were used as the inclusion criteria. According to a standard diagnostic algorithm, all patients underwent blood and immunological tests, pulmonary‐function testing, high‐resolution computed tomography, and ventilation and perfusion lung scan to exclude other associated PH. Patients who were administered intravenous inotropes or vasopressors, vasodilators, and loop diuretics between RHC and TTE were excluded. Patients were excluded if they had technically inadequate echocardiographic window and inappropriate PA‐flow profile. A total of 47 patients were included in the study. We compared the relationship between hemodynamics and the morphology of the PA Doppler‐flow profiles in the PH‐LHD cohort.

Echocardiography and Hemodynamics

Patients underwent a standard TTE examination, as clinically indicated, using a GE Vivid 7 ultrasound unit (GE Vingmed Ultrasound AS, Horten, Norway). Flow velocities were obtained using pulsed‐wave Doppler techniques as proposed by the American Society of Echocardiography and the European Association of Echocardiography.12 Pulsed‐wave Doppler assessment of the PA‐flow profile was performed from the basal short‐axis view, placing the pulsed‐wave Doppler sample volume at the center of the transpulmonary valve jet. Echocardiography readers, masked to the patients' clinical and hemodynamic data, categorized the shape of the PA Doppler‐flow profile as fitting 1 of 3 patterns: (1) a mid‐systolic notch (MSN); (2) a late‐systolic notch (LSN); or (3) no notch (NN). A MSN was characterized as having a distinct notch within the initial two‐thirds of the ejection‐flow velocity period, which divided the flow profile into 2 distinct peaks. An LSN pattern was defined as having deceleration or notching of transient‐flow velocity in the terminal region of the Doppler signal and lacking 2 distinct peaks. Patients without any evidence of Doppler notching were categorized as NN. Representative examples of each PA Doppler‐flow pattern are shown in Figure 1.

Figure 1.

Examples of the 3 distinct patterns of Doppler pulmonary‐flow velocity curves. Representative Doppler tracings (A) without evidence of Doppler notching (NN group); note a dome‐like contour with a maximum velocity in the middle of systole. Examples of the mid‐systolic notch group (B) are characterized by a distinct notch in the mid‐portion, dividing the flow profile into 2 distinct peaks (the MSN group). Doppler tracings of a late‐systolic notch (C), with a triangular contour and a sharp peak in early systole, with a decreased acceleration time (LSN group). Abbreviations: LSN, late‐systolic notch; MSN, mid‐systolic notch; NN, no notch.

For the analyses, the PA Doppler‐flow signal was further categorized into 2 groups, as either normal (NN) or as an MSN/LSN.

The left ventricular ejection fraction, tricuspid annular plane systolic excursion (TAPSE), LA dimension, ratio for early diastolic transmitral flow velocity vs mitral annular velocity (E/e), isovolumic relaxation time (IVRT), and pulmonary acceleration time (PAcT) were measured in accordance with previous studies.13

Right‐heart catheterization was performed using a Swan‐Ganz catheter (7‐Fr or 7.5‐Fr; Edwards Lifesciences LLC, Irvine, CA). Right atrium, RV, PA, and PCWP tracings were recorded. Cardiac output was measured by triplicate thermodilution. Pulmonary capillary wedge pressure was measured by wedging a pulmonary catheter with an inflated balloon into a small PA branch and guiding its progression to the wedge position by fluoroscopic visualization, confirmed by identifying the characteristic waveform and measured at quiet end expiration. When accurate PCWP cannot be determined, left ventricular end‐diastolic pressure should be directly measured.

Statistical Analysis

Statistical analyses were performed using SPSS version 20.0 software (IBM Corp., Armonk, NY). Continuous data are presented as mean ± SD or median (interquartile range [IQR]), whereas categorical data are expressed as the numbers of patients and percentages, as appropriate. Differences between the 2 groups were assessed by Student t test or a nonparametric test for continuous variables and the Fisher exact test for categorical variables. Logistic regression analyses were used to estimate the odds ratio (OR) and 95% confidence interval (CI) for the association between notch pattern and hemodynamics. Sensitivity and specificity values were assessed according to standard definitions. A P value of <0.05 was considered statistically significant.

Results

Demographic Characteristics

The study included 47 patients with PH‐LHD. The patients' basic characteristics and hemodynamic and echocardiographic parameters are presented in tables 1 and 2. The causes of PH‐LHD were left‐sided heart systolic dysfunction in 29.8%, left‐sided heart diastolic dysfunction in 46.8%, and left‐sided heart valvular disease in 23.4%. The average age was 65.2 ± 10.9 years; 23 patients (48.9%) were female. The patients were classified according to their Doppler‐flow notching pattern: 27 patients were classified as NN and 20 patients were either LSN or MSN (17 LSN and 3 MSN).

Table 1.

Baseline Demographic Characteristics of the Overall Cohort

Characteristics	Overall, N = 47	NN Group, n = 27	MSN/LSN Group, n = 20	P Value
Age, y	65.2 ± 10.9	63.7 ± 11.6	67.2 ± 9.7	0.282
Female sex	23 (48.9)	16 (59.3)	7 (35.0)	0.142
BMI, kg/m2	23.2 (22.0–24.8)	23.2 (22.5–24.8)	22.9 (20.7–25.2)	0.636
HR, bpm	78.4 ± 14.4	80.2 ± 12.1	76.6 ± 17.1	0.342
SBP, mm Hg	131.8 ± 21.0	131.5 ± 22.3	132.2 ± 19.8	0.914
DBP, mm Hg	73.9 ± 11.3	73.9 ± 11.9	73.8 ± 10.7	0.970
mBP, mm Hg	91.0 (85.0–105.0)	90.0 (85.0–104.0)	93.0 (85.0–108.0)	0.659
Group 2 subcategories
LH systolic dysfunction	14 (29.8)	9 (33.3)	5 (25.0)	0.525
LH diastolic dysfunction	22 (46.8)	11 (40.7)	11 (55.0)	0.386
LH valvular disease	11 (23.4)	7 (25.9)	4 (20.0)	0.737

Abbreviations: BMI, body mass index; DBP, diastolic blood pressure; HR, heart rate; IQR, interquartile range; LH, left heart; LSN, late‐systolic notch; mBP, mean blood pressure; MSN, mid‐systolic notch; NN, no notch; SBP, systolic blood pressure; SD, standard deviation.

Data are presented as n (%), median (IQR), or mean ± SD.

Table 2.

Hemodynamic and Echocardiographic Data for the Overall Cohort and Both Doppler Notching Groups

Parameter	Overall	NN Group	MSN/LSN Group	P Valuea
Hemodynamic parameters
mRAP, mm Hg	9.0 (6.0–14.0)	8.0 (5.0–11.0)	10.5 (8.3–15.8)	0.04
sPAP, mm Hg	54.0 (45.0–71.0)	46.0 (41.0–60.0)	68.0(54.3–85.5)	<0.001
mPAP, mm Hg	33.0 (29.0–44.0)	31.0 (27.0–34.0)	42.0 (34.3–49.8)	0.001
mPCWP, mm Hg	20.0 (17.0–25.0)	18.0 (17.0–21.0)	23.0 (18.5–28.8)	0.02
CO, L/min	4.73 (3.7–5.5)	4.87 (4.0–5.8)	4.0 (3.2–5.2)	0.072
CI, L/min/m2	2.93 ± 0.99	3.13 ± 1.01	2.67 ± 0.91	0.11
PVR, WU	2.96 (1.9–4.1)	1.91 (1.8–3.0)	4.04 (3.1–5.3)	<0.001
TPG, mm Hg	9.0 (7.0–17.0)	10.0 (9.0–16.0)	16.5 (12.0–22.0)	0.002
Echocardiographic parameters
LA dimension, cm	4.50 ± 0.77	4.32 ± 0.62	4.71 ± 0.89	0.065
sPAP, mm Hg	55.8 ± 17.7	48.3 ± 12.5	65.3 ± 18.8	0.001
PAcT, ms	89.2 ± 23.5	98.1 ± 23.5	77.9 ± 18.5	0.003
IVRT, ms	74.6 ± 22.3	65.9 ± 20.6	85.3 ± 20.1	0.002
TAPSE, cm	1.77 ± 0.50	1.85 ± 0.59	1.67 ± 0.35	0.183
E/e	12.7 (9.1–18.4)	11.9 (8.8–16.5)	14.5 (9.3–21.8)	0.458
EI	1.1 (1.0–1.2)	1.0 (1.0–1.1)	1.2 (1.1–1.2)	0.001
LVEF, %	62.0 (54.5–68.0)	63.0 (51.5–72.0)	60.0 (54.3–66.8)	0.263

Abbreviations: CI, cardiac index; CO, cardiac output; E/e, ratio of velocity of early diastolic transmitral flow to that of mitral annular velocity; EI, eccentricity index; IQR, interquartile range; IVRT, isovolumic relaxation time; LA, left atrium; LSN, late‐systolic notch; LVEF, left ventricular ejection fraction; mPAP, mean pulmonary artery pressure; mPCWP, mean pulmonary capillary wedge pressure; mRAP, mean right atrial pressure; MSN, mid‐systolic notch; NN, no notch; PAcT, pulmonary acceleration time; PVR, pulmonary vascular resistance; RA, right atrium; SD, standard deviation; sPAP, systolic pulmonary artery pressure; TAPSE, tricuspid annular plane systolic excursion; TPG, transpulmonary gradient.

Values are presented as median (IQR) or mean ± SD.

^a P values represent the results of an independent Student t test or using a nonparametric test (Mann–Whitney U test) to compare NN and MSN/LSN, with a P < 0.05 indicating the level of significance.

Pulmonary Artery Doppler‐Flow Profile and Hemodynamics

In the overall PH‐LHD cohort, PA pressures were mild to moderately elevated. Mean PAP was 33.0 mm Hg (IQR, 29.0–44.0 mm Hg), PCWP was 20.0 mm Hg (IQR, 17.0–25.0 mm Hg), TPG was 9.0 mm Hg (IQR, 7.0–17.0 mm Hg), and PVR was 2.96 WU (IQR, 1.90–4.1 WU). Twenty‐four patients (51.06%) had a PVR ≤3 WU, whereas 23 patients (48.94%) had a PVR >3 WU.

Compared with the NN group, the MSN/LSN group had worse hemodynamic parameters, with a higher PVR (4.04 WU [IQR, 3.1–5.3 WU] vs 1.91 WU [IQR, 1.8–3.0] WU, respectively; P < 0.001), a higher TPG (16.5 mm Hg [IQR, 12.0–22.0 mm Hg] vs 10.0 mm Hg [IQR, 9.0–16.0 mm Hg], respectively; P = 0.002), and higher mPAP (42.0 mm Hg [IQR, 34.3–49.8 mm Hg] vs 31.0 mm Hg [IQR, 27.0–34.0 mm Hg], respectively; P = 0.001; Figure 2).

Figure 2.

Box plots demonstrating hemodynamic differences for (A) mean PAP, (B) mean PCWP, (C) TPG, and (D) PVR in the patients in the PH‐LHD cohort with NN and MSN/LSN patterns. Abbreviations: LSN, late‐systolic notch; MSN, mid‐systolic notch; NN, no notch; PAP, pulmonary artery pressure; PCWP, pulmonary capillary wedge pressure; PH‐LHD, pulmonary hypertension caused by left‐sided heart disease; PVR, pulmonary vascular resistance; TPG, transpulmonary pressure gradient.

Table 3 shows that 77.8% of patients in the NN group and 15% of patients in the MSN/LSN group had a PVR ≤3 WU (P < 0.001). In contrast, 22.2% of patients in the NN group and 85% of patients in both the MSN and LSN groups had a PVR >3 WU (P < 0.001). Also, 66.7% of patients in the NN group and 30% patients in the MSN/LSN group had a TPG ≤12 mm Hg, whereas 33.3% of patients in the NN group and 70% of patients in the MSN/LSN group had a TPG >12 mm Hg (P = 0.019).

Table 3.

The Percentages of Notch Groups With a PVR of Either >3 WU or ≤3 WU and a TPG of >12 or ≤12 mm Hg

Variable		NN, n = 27	MSN or LSN, n = 20	P Valuea
PVR	>3 WU	6 (22.2)	17 (85.0)	<0.001
	≤3 WU	21 (77.8)	3 (15.0)
TPG	>12 mm Hg	9 (33.3)	14 (70.0)	0.019
	≤12 mm Hg	18 (66.7)	6 (30.0)

Abbreviations: LSN, late‐systolic notch; MSN, mid‐systolic notch; NN, no notch; PVR, pulmonary vascular resistance; TPG, transpulmonary gradient; WU, Wood units.

Data are presented as n (%).

^a P values represent the results of a χ² test to compare NN and MSN/LSN groups, with P < 0.05 indicating the level of significance.

In the PH‐LHD cohort, the presence of a notched Doppler pattern (either MSN or LSN) was highly associated with a PVR >3 WU and a higher TPG (>12 mm Hg). In contrast, the NN Doppler pattern was strongly associated with PH in the context of a PVR ≤3 WU (OR: 19.83, 95% CI: 4.31‐91.26) and TPG ≤12 mm Hg (OR: 4.67, 95% CI: 1.34‐16.24). The MSN/LSN pattern had 88% specificity and 74% sensitivity for a PVR >3 WU. The NN pattern had 74% specificity and 88% sensitivity for a PVR ≤3 WU.

Other Echocardiographic Parameters

The echocardiographic data are summarized in Table 2. More than 50% of patients in the NN group had a normal or near‐normal limit of TAPSE, and the interventricular septum was not displaced to the left side (as measured by the eccentricity index), showing nearly normal RV function.

Patients in the NN group had higher PAcT (98.12 ± 23.53 vs 77.95 ± 18.46 ms; P = 0.003) and lower IVRT (65.96 ± 20.55 vs 85.30 ± 20.05 ms; P = 0.002) compared with the LSN/MSN group. Patients with notching had a higher, but not significantly larger, LA diameter compared with those in the NN group (4.71 ± 0.89 vs 4.32 ± 0.62 cm; P = 0.065). There was no significant difference in TAPSE, E/e, or left ventricular ejection fraction between the NN and MSN/LSN groups.

Discussion

Although previous studies have demonstrated a relationship between pulmonary hemodynamics and the shape of PA Doppler‐flow profiles in a mixed PH cohort, our study provide an insight into the hemodynamic basis of PH in a PH‐LHD cohort. The absence of Doppler‐flow notching was strongly associated with a PVR ≤3 WU and normal TPG (≤12 mm Hg). In contrast, the MSN and LSN patterns were highly sensitive and specific for a PVR >3 WU and higher TPG (>12 mm Hg) in the PH‐LHD cohort. Thus, this method provided a simple, noninvasive tool to immediately recognize the fundamental hemodynamic differences in PH within a referral PH‐LHD cohort.

In our study, a significant number of subjects with PH‐LHD (mPAP ≥25 mm Hg and PCWP >15 mm Hg) had no evidence of Doppler notching. The strong association between the absence of Doppler notching (NN pattern) and PH‐LHD could relate to the fact that an elevated LA pressure is the predominant cause of PH in the absence of PVD. It has long been known that Doppler notching in the PA Doppler‐flow profile is associated with PH.14, 15 Published data from a study on PA Doppler‐flow profiles showed that a MSN group had higher PA pressure, with a 2‐fold higher PVR and half of the large artery compliance (stroke volume/pulse pressure ratio) compared with the NN/LSN group. In contrast, the NN/LSN group had a lower PVR, a higher PA compliance, a higher PCWP, and much better RV function in a mixed PH cohort.8 Most (21 of 27) of our patients with NN had a PVR of ≤3 WU, which indicated PVH in our study. Previous work has shown that pulmonary venous congestion alone (without elevated PVR) does not lead to premature wave reflection in the pulmonary vasculature. In part, this is because of the distance between the site of impedance (the pulmonary veins) and the right side of the heart.16 The association between the NN pattern and PH‐LHD is of considerable practical value because it provides rapid visual evidence to strongly suggest PVH or an “isolated post‐capillary” type of group 2 PH.

In the presence of increased large‐artery stiffness and a high PVR, reflected waves return to the RV during systole and cause “pathological pulmonary arterial wave reflection” of the Doppler profile.17 The reason why PH‐PVD had “notch” may be due to transmission of pressure impulse from the LA across the pathological pulmonary vascular bed to the PA. Most (17 of 20) of our subjects in the MSN/LSN group with a higher PVR and TPG had PVH with PVD. The notch pattern was not evident in the normal pulmonary circulation but may be secondary to transient elevation of PA load at a time when the RV is still ejecting. This is caused by a decrease in PA compliance and an increase in the size and impedance of the main PA.18 Thus, the patterns of MSN and LSN in left‐sided heart disease identified PH disproportionately as left‐sided heart congestion and thus provided visual evidence that PVH was associated with PVD or combined post‐capillary and pre‐capillary type of group 2 PH.

In accordance with these basic hemodynamic principles, the presence of Doppler notching (either MSN or LSN) was highly associated with PVH and PVD in our cohort. In sharp contrast, the absence of Doppler notching was strongly associated with PVH. Given the high sensitivity for detecting PVH, future work should investigate whether visual inspection of Doppler notching of the PA Doppler‐flow profile can enhance the accuracy of noninvasive hemodynamic screening in PH‐LHD cohorts. Not only the Doppler notching of the PA Doppler‐flow profile, but also the notch in the M‐mode echocardiogram is related to the hemodynamics. Previous studies have reported the presence of an 'a'‐wave notch in the pulmonary valve M‐mode tracing in normal and its absence in PH.19, 20

In our current study, patients with LSN/MSN had higher IVRT and shorter PAcT compared with those in the NN group. As pulmonary pressure increases, peak velocity occurs earlier in systole and late systolic notching is then often present. Tossavainen et al reported that a PAcT ≤90 ms had an 84% sensitivity and an 85% specificity to identify patients with a PVR ≥3 WU in a mixed PH cohort.13 Our results suggest that a simple measure of PAcT could be an important tool to assess PVR. In keeping with previously documented findings, we found that the shorter the acceleration time, the higher the PVR. Echocardiographic variables, such as Doppler notching pattern, IVRT, and PAcT, may improve hemodynamic assessment.

Study Limitations

Our results were based on retrospective analyses of clinical data from a single center and a small cohort. Thus, a larger multicenter study with independent validation could conduct multivariate analyses on the other variables, such as PAcT, IVRT, and LA diameter to predict PVD in the cohort. The TTEs and invasive hemodynamic data were not obtained simultaneously. Although this delay should weaken our study outcomes, such a delay between obtaining data from TTE and RHC is common in clinical practice.

Conclusion

Simple visual assessment of the pattern of the PA Doppler‐flow profile provided a powerful insight into the hemodynamics of the PH‐LHD cohort. A notched Doppler signal suggested PVH with PVD, whereas the absence of a notched pattern suggested PVH without PVD in PH‐LHD patients.