Pulmonary Artery Pulsatility Index Predicts Right Ventricular Myofilament Dysfunction in Advanced Human Heart Failure

One-Sentence Summary:

Among clinical indices of right heart failure in advanced heart failure, the pulmonary artery pulsatility index best reflects underlying defects in right ventricular myofilament contractility.

Right heart failure (RHF) severely complicates outcomes in advanced heart failure (HF). (1) Haemodynamically, RHF is characterised by reduced RV stroke work, which increases right-sided volumes and pressures that precipitate venous congestion and malperfusion. (1) Multiple haemodynamic indices like RV afterload, right atrial pressure (RAP), the RAP-to-pulmonary artery wedge pressure (RAP/PAWP) ratio, RV stroke work index (RVSWI), and the pulmonary artery pulsatility index (PAPi) (2) are used to characterise right heart physiology. (1) All correlate with RHF, but can exhibit poor sensitivity and specificity for predicting outcomes. (1) Moreover, because of the influence of preload and afterload on right heart physiology, it remains unknown if these indices correlate with intrinsic RV myocyte dysfunction, and if so, which do so best. Such knowledge could help validate certain indices while potentially guiding therapies that target RV dysfunction in RHF. (3) Accordingly, the current study prospectively examined RV and LV tissue from advanced HF patients and compared myofilament function to RHF haemodynamic indices to determine which best reflects intrinsic RV sarcomere disease.

We prospectively acquired RV septal and LV free wall myocardium, as previously described, (4) from human subjects (HF, n=18; non-failing control, n=6) enrolled under Institutional Review Board-approved protocols at Johns Hopkins, University of Pennsylvania, and Pennsylvania Gift-of-Life Donor Program. Informed consent was obtained for all. Clinical data were obtained; for HF subjects, these included medication usage and recent right heart catheterisation haemodynamic data (median 1 day, IQR[1,7]). Control hearts were declined for transplantation due to donor age.

RV/LV tissue underwent force-calcium myofilament studies using previously-detailed methods (5) Steady-state myofilament force versus log[Ca²⁺] plots were fit to the Hill equation [F=F_max×Ca^h/(EC₅₀^h+Ca^h)] to yield maximal calcium-activated force (F_max, mN/mm²), calcium sensitivity (EC₅₀, [Ca²⁺] required to achieve 50% F_max, μM), and cooperativity coefficient (h). Myofilament passive force was also ascertained. Univariable correlations were made between haemodynamic and myofilament indices. Receiver operator curve (ROC) analyses compared the ability of RHF indices to identify myofilament dysfunction. Appropriate parametric comparisons were made throughout. Data presented as mean±SEM.

HF and control subjects were matched with respect to age (52±3 vs. 56±5 years, P=0.5), sex, and race. Mean LV parameters, including ejection fraction and end-diastolic diameter, were significantly worse in HF versus controls (21±3% vs. 63±2%, 6.2±0.3 vs. 3.8±0.1 cm, respectively; P<0.001). Haemodynamically, HF subjects had elevated PAWP (21±2 mmHg), mean pulmonary pressure (30±2 mmHg), and pulmonary vascular resistance (PVR, 2.4±0.4 WU), consistent with pulmonary hypertension. Mean RAP was 12±1 mmHg; RAP/PAWP 0.56±0.07; PAPi 4.2±1.4.

RV myofilament activated force-calcium curves revealed significant RV-F_max reduction in HF (15.8±1.1 vs. 20.5±0.6 mN/mm², P=0.001; Fig-A). EC₅₀ was also pathologically reduced (1.72±0.06 vs. 1.94±0.05 μM, P=0.02; Fig-B). We found a positive correlation between RV-F_max and RV afterload measured by PVR (Fig-C) and negative correlation with pulmonary arterial compliance (r=−0.55, P=0.02). RV-F_max meanwhile correlated negatively with RV preload reflected by RAP and RAP/PAWP (Fig-C). PAPi [PAPi=(PASP–PADP)/RAP]—which declines when RV contractile force and PA afterload worsen (all lowering PA pulse pressure) and RV preload increases (6)—was significantly correlated with RV-F_max (Fig-C). Neither RV EC₅₀ nor passive stress correlated with any haemodynamic indices. RV myofilament parameters did not correlate with PAWP or cardiac output, or with echocardiographic indices of RHF.

Fig.

(A) RV force-calcium curves demonstrate significant reduction in RV-F_max in HF versus non-failing controls. (B) Normalised stress curves reveal a left-shifted RV-EC₅₀ in HF, indicating pathologically increased calcium sensitivity. (C) RV-F_max correlates with haemodynamic measures of global RHF. Shown here are its correlation with PVR, RAP, RAP/PAWP, and PAPi (log-transformed to achieve a Gaussian distribution). (D) Comparing receiver operator curves for predicting reveals PAPi [(PASP-PADP)/RAP] is superior among various RHF haemodynamic indices for identifying reduced RV-F_max (<2.5 SD below mean for controls). AUC of PAPi compared to those for RAP, RAP/PAWP, PVR [=(mPAP-PAWP)/CO], PAC [=(CO/HR)/(PASP-PADP)], and RVSWI [=0.136×SVI(mPAP-RAP)]. (E) PAPi was used to split HF into two subgroups: HF-RV dysfunction (HF-RVD, PAPi<1.85) and HF-RV compensated (HF-RVc, PAPi>1.85). RV-F_max was markedly lower in HF-RVD than HF-RVc, whereas RV-F_max in HF-RVc was comparable to controls. (F) In contrast to the RV, LV-EC₅₀ was significantly lower in PAPi-defined HF-RVD versus HF-RVc, whereas LV-F_max was similarly reduced in both groups. PAPi, pulmonary artery pulsatility index; PVR, pulmonary vascular resistance; PAC, pulmonary arterial compliance; RAP/PAWP, RAP-to-pulmonary artery wedge pressure ratio; PASP, pulmonary artery systolic pressure; PADP, pulmonary artery diastolic pressure; mPAP, mean pulmonary artery pressure; CO, cardiac output; HR, heart rate; SVI, stroke volume index.

We next used ROC analyses to compare the predictive utility of these indexes for reduced RV-F_max. PAPi area under the curve (AUC) was 0.89 for detecting RV-F_max<2.5 SD below mean_control), best among all RHF indices (Fig-D). Because PAPi best reflected depressed RV-F_max, we next examined if clinical RHF subgroups based on PAPi have different RV-F_max. PAPi 1.85, a clinical cut-off associated with RHF (2) was used to define RV dysfunction (PAPi<1.85, HF-RVD) and RV compensated (PAPi>1.85, HF-RVc) HF subgroups. HF-RVD had significantly lower RV-F_max (11.9±1.3 mN/mm²) versus controls (20.5±0.6 mN/mm², P<0.0001) and HF-RVc (18.3±1.0 mN/mm², P=0.0004) (Fig-E). By contrast, HF-RVc subjects had near normal RV-F_max (P=0.2). RV calcium sensitivity (RV-EC₅₀) did not differ between HF-RVD and HF-RVc. We also tested whether PAPi could reflect differential LV sarcomere disease. For all HF, LV-F_max and LV-EC₅₀ were reduced (Fig-F). In contrast to RV, LV-F_max was similar between subgroups, while LV-EC₅₀ was much less in HF-RVD than HF-RVc (0.84±0.19 vs. 1.81±0.13 μM, P=0.003) (Fig-F). Among RHF indices, low PAPi again best reflected reduced LV-EC₅₀ (<2.5 SD below mean_control) by ROC analysis, with an AUC=0.99, versus 0.72–0.92 for all others.

Our study identifies PAPi, a commonly used index of RHF, as an excellent predictor of underlying RV sarcomere contractile function. Advanced HF patients with PAPi<1.85 have depressed RV-F_max and accompanying LV calcium sensitisation, while those with PAPi>1.85 are near normal on both fronts. These findings add validity to the use of PAPi in the clinical assessment of RHF and reveal mechanistic targets for directed RHF therapeutics.

This is one of the first studies to examine RV myofilament dysfunction in advanced human HF (7) and the first to establish its relation to RHF hemodynamic indices, and most prominently, PAPi. Since all RHF indices lack precision in detecting RHF, the question of which index actually represents intrinsic RV disease is relevant. Our findings have several important implications. First, PAPi is already emerging as a useful index of RHF (6) and our findings add novel evidence for its validity. Next, low-PAPi RHF may identify a depressed RV-F_max phenotype worth targeting directly with drugs such as levosimendan or direct sarcomere enhancers. (3) Lastly, intrinsic LV diastolic dysfunction also influences PAPi, findings that raise new and important therapeutic considerations for RHF. The current study suggests that the ideal RHF agent would augment RV-F_max without adversely affecting LV calcium sensitivity. Future efforts should account for both these considerations.

Limitations to this study include sample size, the influence of catecholamines and tissue harvest on donor tissue, the possible effect of tricuspid regurgitation and pericardial restraint on RHF indices, and possible tissue heterogeneity throughout RV and LV chambers. Nevertheless, the discovery of distinct underlying RV and LV myofilament abnormalities strengthen the validity of PAPi as a RHF index, while revealing a myofilament phenotype that will hopefully serve as a blueprint for RHF-directed therapies.

ABBREVIATIONS LIST:

RHF

Right heart failure

HF

Heart failure

HFrEF

Heart failure with reduced ejection fraction

PAPi

Pulmonary artery pulsatility index

RVD

RV dysfunction

RVc

RV compensated

RAP

Right atrial pressure

PAWP

Pulmonary artery wedge pressure

F_max

Maximal calcium-activated myofilament force

EC₅₀

Calcium concentration for half-maximal activation