Main pulmonary arterial wall shear stress correlates with invasive hemodynamics and stiffness in pulmonary hypertension

Abstract

Pulmonary hypertension (PH) is associated with proximal pulmonary arterial remodeling characterized by increased vessel diameter, wall thickening, and stiffness. In vivo assessment of wall shear stress (WSS) may provide insights into the relationships between pulmonary hemodynamics and vascular remodeling. We investigated the relationship between main pulmonary artery (MPA) WSS and pulmonary hemodynamics as well as markers of stiffness. As part of a prospective study, 17 PH patients and 5 controls underwent same-day four-dimensional flow cardiac magnetic resonance imaging (4-D CMR) and right heart catheterization. Streamwise velocity profiles were generated in the cross-sectional MPA in 45° increments from velocity vector fields determined by 4-D CMR. WSS was calculated as the product of hematocrit-dependent viscosity and shear rate generated from the spatial gradient of the velocity profiles. In-plane average MPA WSS was significantly decreased in the PH cohort compared with that in controls (0.18 ± 0.07 vs. 0.32 ± 0.08 N/m²; P = 0.01). In-plane MPA WSS showed strong inverse correlations with multiple hemodynamic indices, including pulmonary resistance (ρ = −0.74, P < 0.001), mean pulmonary pressure (ρ = −0.64, P = 0.006), and elastance (ρ = −0.70, P < 0.001). In addition, MPA WSS had significant associations with markers of stiffness, including capacitance (ρ = 0.67, P < 0.001), distensibility (ρ = 0.52, P = 0.013), and elastic modulus (ρ = −0.54, P = 0.01). In conclusion, MPA WSS is decreased in PH and is significantly associated with invasive hemodynamic indices and markers of stiffness. 4-D CMR–based assessment of WSS may represent a novel methodology to study blood-vessel wall interactions in PH.

Pulmonary hypertension (PH) is a progressive disorder defined by pathologic elevations in pulmonary arterial pressure and resistance that imparts a universally poor prognosis despite recent advances in pharmacologic therapy.^1,2 Increases in pulmonary pressure and resistance are hypothesized to result from remodeling of the distal pulmonary circulation as characterized by progressive pulmonary arteriolar medial hypertrophy, adventitial thickening, and neointimal lesions.³ However, PH is also associated with proximal pulmonary arterial accumulation of vascular smooth muscle cells and load-bearing proteins in the extracellular matrix.⁴ These proximal histologic changes confer morphologic and functional changes, including increased vessel diameter, wall thickening, diminished compliance, reduced elastance, and increased stiffness.⁵ Pulmonary arterial stiffness can be readily assessed using a multiple indices obtained from right heart catheterization (RHC) and cardiac magnetic resonance imaging, including capacitance, compliance, distensibility, elastic modulus, and stiffness index β.⁶ Stiffness can account for up to 30%–40% of the right ventricular (RV) afterload due to the reactive component of impedance, and RV systolic function is a critical determinant of pulmonary arterial hypertension–related mortality.^7-9 Consequently, elucidation of the mechanobiological properties of the proximal pulmonary arteries is of considerable clinical relevance in PH.

Wall shear stress (WSS), a primary mechanical drag force that affects endothelial cell mechanotransduction, may serve as a noninvasive parameter to study in situ hemodynamic forces that modulate proximal pulmonary arterial remodeling and therefore vascular mechanical behavior. Preclinical studies support a link between WSS and remodeling in the pulmonary circulation: decreased main pulmonary artery WSS in a rat PH model is associated with increases in both collagen and elastin content.¹⁰ Additionally, disturbed WSS resulting from low flow pulsatility has been shown to locally attenuate pulmonary endothelial cell nitric oxide release.^11,12 The finding of reduced WSS in the proximal conduit vessels in human subjects with PH has been corroborated using multiple approaches, including computational fluid dynamics modeling, phase-contrast cardiac magnetic resonance imaging (CMR), and four-dimensional (4-D) CMR.^13-17 However, the clinical implications of decreased WSS in PH are essentially unknown.

In the systemic circulation, endothelial response to abnormal WSS is known to lead to vascular disease, hypertension, thrombosis, and atherosclerosis. On a cellular level, WSS is known to influence endothelial morphology and alignment, with laminar WSS as the principal component parallel to flow direction, showing an increase in endothelial nitric oxide synthesis. This could have significant implications for pressure regulation in the pulmonary circulation. From a functional perspective, the relationship between WSS and invasive pulmonary hemodynamic indices that define, stratify, and prognosticate PH has yet to be established. We hypothesized that WSS correlates with characteristics of independent biomechanical properties, including resistance, stiffness, and ventricular functional performance, and investigated the hypothesis using same-day invasive hemodynamics and 4-D CMR in a cohort of subjects with PH and matched controls.

Methods

Subjects

Using a prospective National Jewish Health (NJH) Institutional Review Board–approved study, we consecutively enrolled patients aged 18–75 years old from the NJH Pulmonary Hypertension Center. Informed consent was obtained from all subjects. Study participants underwent clinically indicated RHC for diagnosis of PH or reassessment of hemodynamics on PH therapy. RHC was performed using a 7-French Swan-Ganz catheter via right internal jugular access for measurement of mean pulmonary arterial pressure (mPAP), systolic pulmonary arterial pressure, pulmonary arterial wedge pressure (PAWP), thermodilution cardiac output (CO), cardiac index (CI), and stroke volume (SV) and for calculation of pulmonary vascular resistance (PVR). Effective arterial elastance (E_a), an index of pulmonary arterial afterload, was calculated using the equation (mPAP − PAWP)/SV.¹⁸ PH was defined by demonstration of a mPAP of ≥25 mmHg during the protocol RHC or a history of RHC-proven PH (mPAP of ≥25 mmHg) while on PH pharmacologic therapy at the time of protocol.¹⁴ Patients with a prior diagnosis of systemic-to-pulmonary shunts/congenital heart disease or chronic thromboembolic PH were excluded. Control subjects were defined by a mPAP of <25 mmHg. Biometric data, including age, sex, body mass index (BMI), and heart rate, were collected on the day of protocol performance. Hematocrit and serum creatinine levels were obtained on the day of study protocol and measured using ADVIA chemistry assays (Siemens Healthcare Diagnostics, Tarrytown, NY). Glomerular filtration rate (GFR) was calculated using the Modification of Diet in Renal Disease Study equation.¹⁹

CMR protocol

CMR was performed on the same day as RHC in the supine position using a 1.5-T clinical magnetic resonance imaging system (MAGNETOM Avanto; Siemens Healthcare Sector, Erlangen, Germany) and 8-channel phased array cardiac coil. The cardiac short axis was determined from scout images at the level of the midventricle, vertical long axis, and horizontal long axis. A cine steady-state free precession (SSFP) technique with retrospective gating was used to image from the base to apex during brief end-expiratory breath holds using contiguous short-axis slices in 8-mm increments. Additional long-axis and short-axis SSFP images were obtained at the level of the MPA 1 cm distally from the pulmonic valve. Ventricular volumetric and functional analyses were performed off-line by a blinded reader using commercially available software (Argus, MR B17; Siemens Healthcare Sector). RV end-diastolic and end-systolic contours were manually traced in the axial view for each slice. RV ejection fraction (RVEF) was determined using the modified Simpson’s rule.²⁰ Maximal and minimal MPA cross-sectional areas were measured 1 cm above the pulmonary valve in a long-axis view. 4-D flow CMR was performed with interleaved three-directional velocity encoding (voxel volume = 11.7–20.3 mm³, α = 15°, TE/TR = 2.85/48.56 ms, velocity encoding = 150 cm/s, temporal resolution = 50 ms). 4-D flow images were acquired using a radio-frequency-spoiled gradient echo pulse sequence, prospective electrocardiogram gating, and respiratory navigators using bellows, as described elsewhere.²¹ Raw 4-D CMR phase-contrast data sets were preprocessed by filtering for noise and corrected for aliasing using a custom MatLab program (Mathworks, Natick, MA). The preprocessed 4-D CMR data were converted into ParaView visualization and quantification software (Kitware, Clifton, NY) using a custom MatLab program.

Echocardiography

Because of the potential for right-sided valvular disease to confound the WSS analysis, tricuspid regurgitation (TR) and pulmonic regurgitation (PR) were graded as described elsewhere using echocardiograms that were performed as part of clinical care.²² 2-D and color Doppler data acquisition was performed using a Vivid 7 ultrasound system (General Electric Medical System, Milwaukee, WI) equipped with a 3-MHz transducer.

Stiffness analysis

The following equations were used to characterize pulmonary arterial stiffness, as described elsewhere:⁶

$$\begin{array}{}\text{(1)}& \text{distensibility }\left(\frac{\%}{\text{mmHg}}\right)=\frac{A_{\text{MPA,max}}-A_{\text{MPA,min}}}{A_{\text{MPA,min}}/(\text{sPAP}-\text{dPAP})}\times 100(\%),\text{(2)}& \text{capacitance }\left(\frac{{\text{mm}}^3}{\text{mmHg}}\right)=\frac{\text{SV}}{\text{sPAP}-\text{dPAP}},\text{(3)}& \text{compliance }\left(\frac{{\text{mm}}^2}{\text{mmHg}}\right)=\frac{A_{\text{MPA,max}}-A_{\text{MPA,min}}}{\text{sPAP}-\text{dPAP}},\text{(4)}& \text{elastic modulus }(\text{mmHg})=\frac{\text{sPAP}-\text{dPAP}}{A_{\text{MPA,min}}/(A_{\text{MPA,max}}-A_{\text{MPA,min}})},\text{(5)}& \text{stiffness index }\beta =\frac{\text{ln}(\text{sPAP}/\text{dPAP})}{(A_{\text{MPA,max}}-A_{\text{MPA,min}})/A_{\text{MPA,min}}}\mathrm{.}\end{array}$$

WSS analysis

WSS analysis was performed in parallel by two blinded readers (MS, JD). The hemodynamic regional WSS (τ) was calculated as a product of vascular axial shear rate defined as the difference of velocity u with respect to distance from the vessel wall r (du/dr) and hematocrit (Hct)–dependent viscosity (μ):

$$\begin{array}{}\text{(6)}& \tau =\mu (\text{Hct})\frac{\mathrm{d}u}{\mathrm{d}r},\end{array}$$

where the shear rate was calculated on the basis of a generated luminal velocity profile. Kinematic viscosity was obtained from Hct drawn from peripheral blood on the day of protocol performance and calculated using the logarithmic relation proposed by Barras^23,24 based on human viscosity studies:

$$\begin{array}{}\text{(7)}& \mu ={10}^{(0.01703)\text{Hct}+0.0933}\end{array}$$

The anatomic plane used for WSS calculation was selected by superimposing the 2-D CMR cross-sectional image at the time of ventricular systole into the velocity-generated field. The superimposed SSFP MPA cross-sectional image served as the navigational tool for manual segmentation within the 4-D flow domain. For every subject, four independent velocity contours determining lines were placed into the WSS plane (Fig. 1). The line termini were established at the inner lumen of the MPA with the same line-scale resolution applied to each velocity profile. The peak cardiac cycle WSS values were computed at 45° segments along the MPA circumference and were designated as MPA-1 to MPA-8 (Fig. 2), as described elsewhere.¹⁷ The in-plane systolic WSS was then derived from peak regional MPA values. Furthermore, time-averaged WSS over the cardiac cycle (WSS_TA) was calculated as described elsewhere.¹⁶

Figure 1.

Postprocessing analysis of in-plane wall shear stress in a representative control subject. a, Representative image of velocity vector field in the proximal pulmonary vasculature of a control subject. b, Superimposition of corresponding 2-D cardiac magnetic resonance imaging cross-sectional main pulmonary artery (MPA) image on velocity vector field. c, Visualization of streamlines originating in right ventricular outflow tract (RVOT) flowing through MPA cross-sectional plane and beyond through the MPA bifurcation. d, Final determination of flow profile through the MPA lumen. Ao: ascending aorta; LPA: left pulmonary artery; RPA: right pulmonary artery.

Figure 2.

Regional analysis of main pulmonary artery (MPA) wall shear stress in a representative control subject showing four velocity profile lines segmented with 45° increments and eight computational points. Anatomically, points 1–4 represent the anterior MPA from the frontal view perspective, and points 5–8 represent the posterior MPA.

Statistical analysis

Continuous values are reported as the mean ± 1 standard deviation. A two-tailed Student’s t test was performed to compare intergroup characteristics. Intervariable correlations were assessed using Spearman ρ coefficients. A P value less than 0.05 was considered to indicate statistical significance. All statistical analyses were performed in JMP (ver. 10.0; SAS, Cary, NC). The Shrout-Fleiss intraclass correlation coefficient (ICC) was used to assess for interobserver variability.²⁵

Results

Subject characteristics

Of the 22 subjects who completed the study protocol, 15 underwent RHC for the initial diagnosis of PH, and 7 had previously diagnosed PH. Etiologies for PH included idiopathic (N = 8), connective-tissue disease (N = 7), methamphetamine exposure (N = 1), and pulmonary hemangiomatosis (N = 1). All subjects with previously diagnosed PH (N = 7) were treated with targeted PH pharmacologic therapy at the time of the study protocol (sildenafil = 3, tadalafil = 2, bosentan = 1, ambrisentan = 1). Fifteen subjects met hemodynamic criteria for PH following the study RHC. Of the 7 subjects who did not meet RHC PH criteria, 2 had previously diagnosed PH, and both were treated with sildenafil. Consequently, there were 17 subjects who met diagnostic criteria for PH and 5 controls. One control subject completed only the 4-D flow portion of the CMR protocol and not the anatomic/functional imaging acquisition because of claustrophobia.

Patient characteristics and hemodynamics

There were no significant differences in age or BMI between PH and control subjects (Table 1). Although both groups were predominantly female, the control group had a greater percentage of female subjects than the PH group (80% vs. 59%). At the time of protocol enrollment, 7 of the 17 PH subjects were receiving targeted pharmacologic therapy (ambrisentan = 1, bosentan = 1, sildenafil = 3, tadalafil = 2). As anticipated, PH subjects demonstrated significant elevations relative to controls in mPAP (37.1 ± 11.1 vs. 20.1 ± 3.1 mmHg; P = 0.001), PVR (7.4 ± 7.0 vs. 2.1 ± 1.0 Wood units; P = 0.008), and E_a (0.81 ± 0.19 vs. 0.26 ± 0.09 mmHg/mL; P = 0.01), with no difference in PAWP. Although there was a trend toward reduced RVEF in PH subjects versus controls (41.2% ± 14.8% vs. 53.9% ± 8.4%; P = 0.05), there was no significant difference in CO, CI, or SV between these groups. Increased pulmonary arterial stiffness was present in the PH cohort, as demonstrated by significantly elevated elastic modulus and stiffness index β as well as reduced capacitance, compliance, and distensibility (Table 2). MPA cross-sectional area was increased in PH subjects versus controls, even when MPA area was normalized to body surface area (BSA). There was no significant difference in calculated blood viscosity. Echocardiograms were performed within 56 ± 60 days of protocol performance. TR was mild overall in the PH cohort (trace or mild = 13, mild-moderate = 3, severe = 1). Similarly, PR was trace or mild in all subjects with the exception of 1 PH subject who demonstrated mild-moderate PR.

Table 1.

Patient characteristics and hemodynamic data

	Control (N = 5)	PH (N = 17)	P
Age, years	54 ± 9	60 ± 10	NS
Sex, no.
Female	3	11
Male	2	6
BMI	29.3 ± 5.2	26.6 ± 9.4	NS
Systolic blood pressure, mmHg	127 ± 8	127 ± 16	NS
Diastolic blood pressure, mmHg	70 ± 11	79 ± 9	NS
eGFR, mL/kg/1.73 m2	60 ± 0	57 ± 6	NS
mPAP, mmHg	20.1 ± 3.1	37.1 ± 11.1	0.001
PVR, Wood units	2.1 ± 0.9	7.4 ± 7.0	0.008
PAWP, mmHg	10.4 ± 4.2	11.6 ± 4.5	NS
mRAP, mmHg	6.4 ± 2.8	7.5 ± 3.2	NS
TPG, mmHg	9 ± 4	24 ± 9	<0.0001
DPG, mmHg	4 ± 3	8 ± 7	NS
CO, L/min	5.7 ± 1.3	5.1 ± 1.5	NS
CI, L/min/m2	2.8 ± 0.8	2.8 ± 0.8	NS
SV, mL	80.2 ± 21.4	57.8 ± 19.4	NS
RVEF, %	53.9 ± 8.4	41.2 ± 14.8	0.05
Tricuspid regurgitation, no.
Trace	1	6
Mild	4	7
Mild-moderate	0	3
Moderate	0	0
Severe	0	1
Pulmonic regurgitation, no.
Less than mild	5	12
Mild	0	4
Mild-moderate	0	1
Ea, mmHg/mL	0.26 ± 0.09	0.81 ± 0.19	0.01
Viscosity, mPa	3.5 ± 0.1	3.9 ± 0.7	NS

^NoteExcept where otherwise noted, data are mean ± 1 standard deviation. BMI: body mass index; CI: cardiac index; CO: cardiac output; DPG: diastolic pulmonary vascular pressure gradient; E_a: elastance; eGFR: estimated glomerular filtration rate; mPAP: mean pulmonary arterial pressure; mRAP: mean right atrial pressure; PAWP: pulmonary arterial wedge pressure; PH: pulmonary hypertension; RVEF: right ventricular ejection fraction; PVR: pulmonary vascular resistance; SV: stroke volume; TPG: transpulmonary gradient.

Table 2.

Pulmonary arterial stiffness indices in control subjects versus those with pulmonary arterial hypertension (PAH)

	Control	PAH	P
Capacitance, mm3/mmHg	4.8 ± 1.5	1.9 ± 0.9	0.010
Compliance, mm2/mmHg	10.2 ± 2.9	5.7 ± 4.1	0.024
Distensibility, %/mmHg	1.7 ± 0.5	0.7 ± 0.4	0.004
Elastic modulus, mmHg	63 ± 17	333 ± 395	0.012
Stiffness index β	3.1 ± 0.9	9.5 ± 10.1	0.020
MPA area, cm2	5.6 ± 1.5	8.2 ± 1.5	0.013
MPA area/BSA, cm2/m2	2.9 ± 1.0	4.6 ± 1.0	0.01

^NoteData are mean ± 1 standard deviation. BSA: body surface area; MPA: main pulmonary artery.

WSS analysis

Qualitative evaluation of MPA flow patterns using the streamline visualization technique revealed unidirectional MPA flow in all controls, while PH flow patterns were characterized by a systolic recirculation complex that was typically localized to the distal right lateral and inferior aspect of the MPA (Fig. 3). Consistent with the identification of the retrograde limb of the MPA recirculation in the right side of the MPA, the PH cohort demonstrated significant decreases in average in-plane peak systolic WSS relative to controls in both the right lateral (MPA1) and the right anterior (MPA2) region (Fig. 4). Peak regional WSS values were decreased in the PH cohort in all other regions (MPA3–MPA8) but did not reach statistical significance. Overall, the in-plane systolic WSS was decreased in the PH group versus the control group (0.19 ± 0.07 vs. 0.32 ± 0.08 N/n²; P = 0.01), consistent with the trend observed in the regional analysis. The WSS_TA was also significantly decreased in the PH group (0.13 ± 0.03 vs. 0.09 ± 0.03 N/m²; P = 0.04). An evaluation for interobserver agreement for WSS calculation was performed using the Fleiss ICC method and demonstrated an ICC value of 0.77, which is considered good agreement. ICC values below 0.4 indicate low agreement, values between 0.4 and 0.75 indicate fair to good agreement, and values above 0.75 indicate very good agreement.²⁵

Figure 3.

Representative four-dimensional flow cardiac magnetic resonance imaging main pulmonary artery (MPA) flow patterns in a control subject (top) and a subject with pulmonary arterial hypertension (PAH; bottom) at different stages of the cardiac cycle. Time points represent milliseconds after the after the R wave/initiation of systole: t₁ = 0.023 ms, t₂ = 0.071 ms, t₃ = 0.118 ms, t₄ = 0.165 ms, and t₅ = 0.213 ms. The rectangular cross-sectional plane represents the region of interest for wall shear stress analysis. Note the recirculation present in the distal MPA during systole in the PAH subject that is absent in the control subject.

Figure 4.

Regional in-plane systolic wall shear stress (WSS) averaged from regional values in controls and subjects with pulmonary hypertension (PH). Data are expressed as mean ± 1 standard deviation in newtons per square meter. MPA: main pulmonary artery; NS: not significant.

MPA in-plane systolic WSS showed strong negative correlations with multiple hemodynamic indices, including same-day measured RHC metrics (Fig. 5). Specifically, with mPAP we observed negative curvilinear correlation and negative exponential trends with PVR and E_a. Numerous markers of pulmonary arterial stiffness were associated with in-plane systolic WSS, including capacitance, compliance, distensibility, and elastic modulus. However, no relationship was identified between in-plane WSS and stiffness index β. MPA normalized to BSA approached statistical significance when correlated with WSS (ρ = −0.42, P = 0.056). Although we identified a significant association between MPA WSS and SV (ρ = 0.63, P = 0.003), no relationship was found between indices of flow, including CO and CI. Furthermore, WSS failed to significantly correlate with age, viscosity, TR, PR, or RVEF.

Figure 5.

Spearman’s ρ regression analyses with associated P values demonstrating associations between in-plane average wall shear stress (WSS) and hemodynamics as well as indices of pulmonary arterial stiffness. Red indicates control, and blue indicates pulmonary hypertension (PH). CI: cardiac index; E_a: effective arterial elastance; MPA: main pulmonary artery; mPAP: mean pulmonary arterial pressure; NS: not significant; PVR: pulmonary vascular resistance; SV: stroke volume.

Discussion

To our knowledge, this is the first study to demonstrate significant correlations between WSS and both invasive pulmonary hemodynamics and indices of stiffness in a cohort of controls and PH subjects. Using 4-D CMR velocity profile mapping in the MPA, we found in-plane average MPA WSS to be decreased compared with that in controls, which appeared to be independent of changes in mean bulk flow. The uniqueness of our study relies on the performance of same-day RHC and 4-D CMR, minimizing the potential for variable loading conditions to confound associations between WSS and hemodynamics. In addition, our study used patient-specific blood viscosities, thereby decreasing the potential for assumed viscosities to introduce error into our WSS computation.

WSS significantly correlated with indices of stiffness, luminal area, and afterload but showed no statistical correlation with indices of flow, including CO and CI. It should be noted that there was no significant difference in either CO or CI between PH and control cohorts, potentially limiting the ability to test the relationship between flow and WSS. Furthermore, this similarity in mean flows, along with similarities in RVEF, could imply that the PH cohort was in the adaptive phase of RV remodeling with a coupled RV-PA axis. However, since WSS represents the friction force along the vessel wall responsible for altered endothelial arrangement, extracellular matrix composition, and altered gene expressions, it would be anticipated that vascular stiffness and SV would demonstrate stronger relationships with WSS than would indices of flow. Indeed, our results show strong correlations with both markers of stiffness and SV, supporting the thought that WSS is primarily dependent on SV representing the momentum aspect of the drag and geometrical variations. Specifically, larger SV has greater potential for frictional shear effect along the vessel wall, while gross anatomical geometry will dictate the spread of the SV momentum and velocity gradients.

The finding that average in-plane systolic and regional WSS assessed in MPA is decreased in PH subjects compared with that in controls is consistent with previous studies in PH clinical cohorts.^15,16 We found the in-plane systolic WSS to be 0.18 ± 0.07 N/m² in PH subjects and 0.32 ± 0.08 N/n² in controls. In a larger dual-center study also using 4-D CMR, Barker et al.¹⁷ reported similar MPA WSS values in PH subjects and controls (0.19 and 0.39 N/m², respectively). Decreased peak systolic right pulmonary artery (RPA) WSS is associated with pediatric PH, although RPA WSS calculated in previous studies appears to be elevated overall in both controls (2.0 ± 0.9 N/m²) and PH subjects (0.82 ± 0.5 N/m²) compared with the MPA WSS we calculated in our adult cohort.

We identified regional differences in WSS magnitude between controls and PH subjects, particularly in the right anterior and lateral MPA segments. Prior 4-D CMR studies have identified large-scale MPA vortices in PH cohorts that appear to be absent in controls, with retrograde flow typically coursing along the lesser curvature of the artery.²⁶ We suspect that retrograde flow emanating from the distal MPA vortex collides with antegrade flow in mid- to late systole, resulting in a regional decrease in WSS. The location of this vortex within the MPA—and, subsequently, the location of retrograde flow—is likely to vary depending on a multitude of factors, including the magnitude and direction of MPA inflow; the degree of MPA dilatation, curvature, and stiffness; and total afterload, including pressure, resistance, and impedance. This postulation is supported by our finding that mPAP inversely correlated with WSS and that early onset of MPA retrograde flow is associated with elevated mPAP in PH.²⁷

WSS has been implicated in the regulation of transcriptional events involved in pulmonary vascular remodeling.^11,12 Elastin and collagen appear to be significant determinants of biomechanical behavior, and their difference in stiffness creates nonlinear stress/strain behavior. Alterations in the elastin/collagen ratio leads to microscopic histopathological changes that eventually result in biomechanical alterations, such as stiffening and luminal dilation. Preclinical studies suggest that this dilation causes deformations to occur while stiff collagen is carrying the mechanical load.²⁸ The role played by the extracellular matrix in MPA enlargement is further supported by the fact that progressive MPA dilatation appears to be independent of changes in pressure and CO.^11,29 Accordingly, our PH cohort demonstrated a significant increase in MPA cross-sectional area compared with that in controls, and there was a trend toward an association between MPA dilatation and WSS (P = 0.055).

Given the numerous correlations demonstrated between WSS and invasive hemodynamic indices that are routinely used to diagnose and prognosticate PH, it is appealing to consider the application of WSS as a clinical tool for the management of pulmonary vascular disease. Because WSS also correlated with indices of stiffness, WSS may also provide incremental hemodynamic data and insights about the proximal vasculature not typically provided by commonly used measurements like mPAP and PVR. However, additional preclinical and clinical studies are required to better understand the natural history of pulmonary arterial WSS as it relates to the hemodynamic progression and vascular remodeling associated with PH.

Study limitations

Our study is limited by a number of issues, most notably a small sample size, which clearly limits our statistical power and ability to apply these findings to the larger PH population. Furthermore, our PH cohort includes both incident and prevalent cases. Specifically, 7 of the 22 PH subjects were receiving targeted PH therapy at the time of enrollment, and it is conceivable that pharmacologic treatment may alter MPA flow patterns, thereby confounding our analysis. Larger patient cohorts are required not only to corroborate these findings and but also to investigate the utility of WSS as a clinical tool for PH detection, management, and prognostication. Although our cohort could be characterized as relatively older (mean age: control, 57 ± 5 years; PH, 58 ± 10 years) compared with populations included in PH clinical intervention trials, our cohort reflects the increasing age of onset and diagnosis of PH. Indeed, the mean age in the REVEAL clinical registry on enrollment was 53 ± 14 years.³⁰ In addition, our PH cohort’s hemodynamic profile included primarily moderate PH with mildly reduced RV systolic function. Inclusion of more advanced forms of PH may reveal novel insights into the natural history of WSS in PH. Finally, our approach focused on in-plane WSS and did not include the radial and circumferential components of MPA WSS. Given the presence of large-scale vortices in the MPA that likely exert multidirectional WSS on the vessel wall, investigation of all WSS components may yield greater insights into fluid-structure interactions in PH.

Conclusion

We have demonstrated that reduced in-plane MPA WSS is associated with worsened pulmonary arterial hemodynamics, stiffness, and remodeling in PH patients and controls. 4-D CMR–based assessment of WSS may represent a novel noninvasive methodology to study fluid-structure interactions in the pulmonary vasculature.