Critical appraisal of the instantaneous end-diastolic pulmonary arterial wedge pressures.

AIMS: A substantial shift in the field of pulmonary hypertension (PH) is ongoing, as the previous practice of mean pulmonary arterial wedge pressure (PAWPM ) is no longer supported. Instead, aiming for a better estimate of end-diastolic pressures (EDP), instantaneous PAWP at mid-A-wave (PAWPmid-A ) or, in the absence of an A-wave, at 130-160 ms following QRS onset has recently been recommended. Electrocardiogram-gated PAWP (PAWPQRS ) has also been proposed. The quantitative differences as well as the diagnostic and prognostic utility of these novel PAWP measurements have not been evaluated. We set out to address these issues. METHODS AND
RESULTS: Pressure tracings of 141 patients with PH due to left heart disease (PH-LHD) and 43 with primary pulmonary arterial hypertension (PAH) were analysed. PAWP was measured as follows: (i) mean pressure (PAWPM ); (ii) per the latest consensus approach [PAWPmid-A , or in atrial fibrillation 130, 140, 150, and 160 ms following QRS onset (PAWP130-160 )]; (iii) at QRS onset (PAWPQRS ); and (iv) Z-point (PAWPZ ). For each PAWP, the corresponding pulmonary vascular resistance (PVR) and diastolic pressure gradient were calculated. The cohort comprised 45% female. Mean age was 66 ± 15. PAWPmid-A was in good agreement with PAWPZ (17.3 [14.5 to 21.2] vs. 17.6 [14.2 to 21.6] mmHg, P = 0.63), whereas PAWPQRS provided significantly lower values (15.3 [12.5 to 19.2] mmHg, P < 0.001). In atrial fibrillation, PAWP130 and PAWPQRS yielded the optimal temporal and quantitative analyses of EDPs. The ability to differentiate PAH from PH-LHD was similar for the various PAWP measurements [PAWPM : area under the curve (AUC) 0.98, confidence interval (CI) 0.96-0.99; PAWPmid-A/130 : AUC 0.94, CI 0.91-0.98; PAWPQRS : AUC 0.96, CI 0.94-0.99, P < 0.001 for all]. PVR based on instantaneous PAWP measurements failed to provide superior prognostic information in PH-LHD as compared with conventional PVR.
CONCLUSIONS: Although instantaneous PAWP measurement might better represent EDP, they nevertheless fail to yield incremental diagnostic or prognostic information in PH-LHD as compared with conventional measurements.

Introduction

The pulmonary arterial wedge pressure (PAWP) entails a fundamental role in haemodynamic evaluation in heart failure (HF) as well as for the distinction between pulmonary hypertension (PH) due left heart disease (PH‐LHD) and primary pulmonary arterial hypertension (PAH). However, despite its abundant utility in everyday practice, the lack of uniformity in the PAWP assessment remains considerable. , The mean PAWP (PAWPM) averaged over the entire cardiac cycle is used as a surrogate of left sided end‐diastolic pressure (EDP) in many instances with reasonable accuracy. However, particularly in the context of PH‐LHD and atrial fibrillation (AF), , incorporation of large V‐waves in the PAWP measurement (PAWPM) might overestimate the EDP, while measuring instantaneous end‐diastolic wedge pressure might yield more concordant results. , Recent guidelines proposed a classification system for PH‐LHD on the basis of the diastolic pressure gradient (DPG). However, in case of DPG, the PAWPM, being compared with the instantaneous end‐diastolic pulmonary artery pressure (PAP) (PAPD), would invariably yield misleading results. Moreover, augmented left atrial pressure (LAP) pulsatility, which impacts on PAWPM, , , further mitigates the potential information that DPG might otherwise provide. Recently, Wright and colleagues proposed the use of instantaneous electrocardiogram (ECG)‐gated PAWP for DPG calculation. Similarly, the latest consensus statement on PH‐LHD advocates for instantaneous instead of mean PAWP measured, in case of sinus rhythm (SR), at the mid‐A‐wave, while in AF, at 130–160 ms after the QRS onset. Additionally, with regards the controversies about the prognostic DPG value in PH‐LHD, , , , , , , , pulmonary vascular resistance (PVR)—derived from instantaneous PAWP—has been introduced as the discerning marker of pre‐capillary involvement.

The qualitative and quantitative discrepancies among the proposed PAWP measurements have not been investigated, leaving ambiguity regarding the optimal method. Also, in the recommended novel approach albeit relying on physiological rational, its diagnostic and prognostic utility have not yet been validated. We set out to address these issues.

Methods

Study population

Patients referred for right heart catheterization (RHC) at Karolinska University Hospital between February 2014 and August 2018, due to unexplained dyspnoea or suspected PH or for haemodynamic assessment before heart transplantation or left ventricular (LV) assist device listing, were enrolled prospectively. Patients with constrictive pericarditis, arrhythmogenic right ventricular cardiomyopathy, previous heart transplantation, significant valvular disease, or normal haemodynamics were excluded. Consequently, 249 patients (PAPM ≥ 20 mmHg) were included in the final analysis. A total of 43 patients were classified as PAH (PAWPM ≤ 15 mmHg and consensus clinical board opinion) and 141 as PH‐LHD, defined as elevated mean wedge at rest (PAWPREST > 15 mmHg) or during exercise (PAWPEX ≥ 25 mmHg). (Figure ).

All patients underwent transthoracic echocardiography 1 h prior to RHC as per current recommendations. The study complies with the Declaration of Helsinki and was approved by the regional ethical review board. Informed consent was provided.

Right heart catheterization

During RHC, all patients were in stable haemodynamic condition. RHC was performed through the jugular vein access using a 6F balloon‐tipped fluid‐filled Swan–Ganz catheter (Edwards Lifesciences, Irvine, CA, USA). Pressure measurements (in the right atrium, pulmonary artery, PAWP, and right ventricle) were performed under fluoroscopy after calibration with the zero‐level set at the mid‐thoracic line, at end‐expirium during spontaneous breathing, and stored in dedicated software (Xper Information Management, Philips Medical Systems, The Netherlands). Cardiac output (CO) was assessed using the Fick principle. The oxygen consumption was measured breath by breath (Jaeger Oxycon Pro, VIASYS™ Healthcare, Palm Springs, CA, USA) in mL/min. Arterio‐venous oxygen difference was calculated from oxygen concentration in arterial and mixed venous blood from the pulmonary artery. In 10 cases, thermodilution was employed.

Exercise protocol

Following the assessment of resting haemodynamics, patients with PAWPM ≤ 15 mmHg at rest but with clinical suspicion of HF with preserved ejection fraction (HFpEF) underwent supine cycle ergometry as did patients with HF with reduced ejection fraction (HFrEF) as part of the haemodynamic evaluation. Patients cycled at 60 rpm in supine position for 2 min at 20 W before the workload was incrementally increased in ~1 min interval until maximal volitional exertion. PAWPM ≥ 25 mmHg during peak exercise denoted abnormal PAWPM response.

Analysis of pulmonary arterial wedge pressure, pulmonary vascular resistance, and diastolic pressure gradient measurements

PAWP and PAP waveforms (n = 249) were analysed offline (MATLAB software; R2018b, MathWorks, MA, USA). First, the ECGs of the two measurements were synchronized manually so that despite non‐beat‐to‐beat synchronous measurements, an optimal temporal harmonization was achieved. From the PAP recordings, the peak of the ascending limb of the PAP curve (PAPS) and the EDP (PAPD) were marked manually, and then the software automatically calculated PAPS and PAPD. Subsequently, mean PAP (PAPM) was calculated by integration of PAP over the cardiac cycle. PAWP was measured in the following ways (Figure ):

Figure 1

Pulmonary artery wedge pressure measurements. The top panel shows the pulmonary artery wedge pressure (PAWP), the middle panel the pulmonary artery pressure (PAP) waveform, and the bottom panel the corresponding ECG traces for both measurements (ECG of the PAWP curve in red and ECG of the PAP curve in blue). In sinus rhythm (A), the instantaneous pressure values on the PAWP waveform were measured as follows: at mid‐A‐wave (PAWPmid‐A) and at the Z‐point (PAWPZ). In atrial fibrillation (B), the instantaneous pressure values on the PAWP waveform were measured at the time points 130, 140,150, and 160 ms after the QRS onset (PAWP130, PAWP140, PAWP150, and PAWP160, respectively). The following PAWP measurements were performed in both SR and AF: at QRS onset (PAWPQRS) and at the time point simultaneous with PAPD (PAWPS). On the PAP waveform, the following instantaneous pressure values were measured: peak of the ascending limb of the PAP curve (PAPS) and the point at the end of diastole (PAPD). In addition, the software provided automated calculation of mean PAP (PAPM) and PAWP (PAWPM) by integrating the PAP and PAWP, respectively, over the entire cardiac cycle (lined area under the pressure curve).

PAWPM: Mean PAWP obtained by temporal integration of the instantaneous PAWP over the entire cardiac cycle.

Consensus statement's approach : In SR, mid‐A‐wave pressure, that is, the mean value of the A‐wave amplitude, was measured from the point just prior the ascending limb to the peak of the A‐wave (PAWPmid‐A). In AF, pressures at 130, 140, 150, and 160 ms following the QRS onset were measured (four measurements: PAWP130, PAWP140, PAWP150, and PAWP160, respectively).

PAWPQRS: Instantaneous PAWP was measured at the QRS onset.

PAWPZ: Instantaneous PAWP at the Z‐point (pre‐C‐wave) was measured in cases with evident C‐wave.

PAWPS: Instantaneous PAWP obtained simultaneously with PAPD.

The corresponding PVR and DPG values were calculated using the following equations and applying the suitable PAWP value:

All pressures were averaged from a minimum of three heart cycles at end‐expiration. , To ensure the uniformity of data acquisition and analysis, the same investigator (A. M.) participated in the vast majority of RHC procedures and performed the analysis of all waveforms, blinded to patient data.

Details regarding the methodology used for MATLAB analysis are provided in .

Statistical analysis

Normality was tested by the Shapiro–Wilk test. Statistical analysis was performed using SPSS Version 26.0 (SPSS Inc., Chicago, IL, USA) Continuous variables were compared using the Mann–Whitney U test (skewed variables). All tests were performed at 95% confidence intervals. All measurements are stated as median and inter‐quartile range or mean and standard deviation (SD) based on the Kolmogorov–Smirnov test. Correlations were tested by the Pearson two‐tailed test. Intra‐observer and inter‐observer reproducibility of the instantaneous PAWP measurements was tested using the intraclass correlation coefficient in 10 randomly selected patients. Survival was analysed with Kaplan–Meier non‐parametric test and compared using a log‐rank test. For survival analysis, patients who underwent cardiac transplantation or ventricular assist device implantation were censored at the time of the latter. Receiver operating characteristic (ROC) analysis was employed for diagnostic ability assessment of the various PAWP measurements. ROC curves were compared using the DeLong test. A P‐value of <0.05 was considered statistically significant.

Results

Patient characteristics

Baseline clinical and haemodynamic characteristics of the study cohort are presented in Table . In total, 141 patients were classified as PH‐LHD. Of them, 80 patients (57%) had HFpEF (EF 63%; 57–65), and 61 (43%) had HFrEF (EF 25%; 19 to 36). In the HFrEF cohort, 21 patients (34%) had resynchronization therapy. The majority of the patients were highly symptomatic (78% in New York Heart Association III–IV). A total of 113 patients were in SR. At rest, 107 patients had PAWPM > 15 mmHg; by using PVRM, 38% of them were classified as having combined post‐capillary and pre‐capillary PH (Cpc‐PH). The same figures for PAWPQRS and PAWPmid‐A/130 were 79 and 93 patients with 49% and 48% prevalence of Cpc‐PH, respectively.

Table 1

Demographic, echocardiographic, and haemodynamic data of the study population

	PAH (n = 43)	PH‐LHD (n = 141)
Demographics
Age (years)	56 [43–69]	66 [55–74]
HFpEF (n, %)	—	80 (57%)
Sinus rhythm	43 (100%)	113 (80%)
Female (n, %)	23 (54%)	62 (44%)
Diabetes mellitus (n, %)	2 (5%)	24 (17%)
Hypertension (n, %)	5 (12%)	80 (57%)
Hypercholesterolaemia (n, %)	3 (7%)	42 (30%)
BMI (kg/m2)	24 [22–28]	27 [23–30]
Echocardiographic data
EF (%)	61 [59–65]	54 [28–65]
E/e′	8 [7–10]	13.1 [10–19]
LAVi (mL/m2)	27 [21–35]	48 [38–66]
TAPSE (mm)	19 [15–20]	16 [12–21]
Haemodynamic data
SBP (mmHg)	119 [104–130]	119 [96–138]
DBP (mmHg)	70 [50–76]	65 [56–74]
CI (L/m2)	2.4 [2.1–3.2]	2.3 [1.8–2.7]
HR (min−1)	71 [65–90]	67 [60–78]
PAPM (mmHg)	36.6 [29.6–45.7]	30.8 [24.9–37.3]
PAPD (mmHg)	23.8 [16.8–30.7]	20.1 [16–24.2]
PAWPM (mmHg)	10.8 [7.1–11.3]	18.7 [15.4–24.2]
PVRM (WU)	5.5 [3.8–9.2]	2.6 [1.7–4.0]
DPGM (mmHg)	−13.8 [8.9–20]	0.8 [−1.8 to 4.1]

BMI, body mass index; CI, cardiac index; DBP, diastolic blood pressure; DPGM, diastolic pressure gradient calculated from the PAWPM; EF, left ventricular ejection fraction; HFpEF, heart failure with preserved ejection fraction; HR, heart rate; LAVi, left atrial volume index; PAH, pulmonary arterial hypertension; PAPD, pulmonary arterial diastolic pressure; PAPM, pulmonary arterial mean pressure; PAWPM, mean value of the pulmonary arterial wedge pressure; PH‐LHD, pulmonary hypertension due to left heart disease; PVRM, pulmonary vascular resistance calculated from the PAWPM; SBP, systolic blood pressure; TAPSE, tricuspid annular systolic excursion.

Continuous variables are given as median values, followed by the 25th and 75th percentiles in square brackets.

Pulmonary arterial wedge pressure and pulmonary vascular resistance measurements

In PH‐LHD patients, PAWPQRS was lower than PAWPM (P < 0.001) (Table ).

Table 2

Instantaneous pulmonary arterial wedge pressure measurements and derived diastolic pressure gradient values in pulmonary hypertension due to left heart disease patients with sinus rhythm vs. atrial fibrillation

	Total (n = 141)	Sinus rhythm (n = 113)	Atrial fibrillation (n = 28)
HR (min−1)	67 (61–78)	66 (60–76)	73 (63–80)
PAWPM (mmHg)	18.7  (15.4–24.2)SQR	18.5 (15.0–23.4)SQC	21 (17.7–25.7)SQC
PAWPS (mmHg)	16.7 (13.4–21.1)MQR	15.6 (12.9–20.6)MQ	18.3 (14.4–22.6)MQ
PAWPQRS (mmHg)	15.8 (13.3–20.4)MSR	15.2 (12.8–19.9)MC	19.1 (15.5–23.6)MCS
PAWPmid‐A (mmHg)	—	17.2 (14.4–22.2)MSQ	—
PAWP130 (mmHg)	—	—	17.9 (14.6–21.8)M
PAWP140 (mmHg)	—	—	17.1 (14.3–20.7)M
PAWP150 (mmHg)	—	—	17.8 (13.4 – 20.8)MQ
PAWP160 (mmHg)	—	—	14.9 (10.8 – 20.6)MQ
PVRM (WU)	2.6 (1.7 – 4.0)QC	2.5 (1.7 – 4.1)QC	2.6(1.8 – 3.5)QC
PVRQRS (WU)	3.1 (2.2 – 4.6)MC	3.0 (2.2 – 4.8)MC	3.2 (2.7 – 4.3)M
PVRmid‐A/130 (WU)	3.0 (1.9 – 4.4)MQ	2.9 (1.7 – 4.4)MQ	—
PVR130 (WU)	—		3.2 (2.7 – 4.5)M
PVR140 (WU)	—	—	3.8 (2.7 – 4.9) a
PVR150 (WU)	—	—	3.5 (2.8 – 4.8) a
PVR160 (WU)	—	—	4.7 (2.7 – 6.2) a
PAPD (mmHg)	20.2 (16.0 – 24.2)	19.7 (15.9 – 23.4)	23.0 (18.7 – 28.7)
PAPM (mmHg)	31.2 (25.1 – 38.0)	30.8 (25.0 – 36.6)	33.3 (26.1 – 43.7)

Letters in superscript indicate statistically significant (P < 0.05) differences: S, simultaneous measurement; M, mean value; QRS, ECG‐gated measurement; and C, according to the consensus paper's recommendation, i.e. mid‐A‐wave in sinus rhythm or 130 ms from QRS onset in atrial fibrillation.

HR, heart rate; PAPD, pulmonary arterial diastolic pressure; PAPM, pulmonary arterial mean pressure; PAWP, pulmonary artery wedge pressure; PAWP130, instantaneous PAWP measured 130 ms after the QRS onset; PAWP140, PAWP150, PAWP160, PAWP measured 140, 150 and 160 ms after the QRS onset, respectively; PAWPM, mean PAWP; PAWPmid‐A, instantaneous PAWP measured at the mean of the A‐wave; PAWPQRS, instantaneous PAWP measured at the QRS onset; PAWPs, instantaneous value of the PAWP pressure measured simultaneously with the time point of PAPD; PVR, pulmonary vascular resistance; PVR130, PVR calculated from the PAWP130; PVR140, PVR calculated from the PAWP140; PVR150, PVR calculated from the PAWP150; PVR160, PVR calculated from the PAWP160; PVRM, PVR calculated from the PAWPM, PVRQRS, PVR calculated from the PAWPQRS; PVRmid‐A, PVR calculated from the PAWPmid‐A.

Indicates that PVR140–160 were significantly different from all the other reported PVR values, as well as from each other.

Sinus rhythm

As illustrated in Figure , PAWPmid‐A was significantly lower than PAWPM (P = 0.009), whereas both measurements yielded higher values than PAWPQRS (P < 0.001) and PAWPS (P < 0.001). The latter two measurements provided essentially similar results (P = 0.81).

Figure 2

Comparison of the various PAWP measurements among patients (A) in sinus rhythm, (B) in atrial fibrillation, and (C) where C‐wave was identifiable. PAWP130, PAWP140, PAWP150, PAWP160, and PAWP measured 130, 140, 150, and 160 ms after the QRS onset, respectively; PAWPM, mean value of PAWP over the cardiac cycle; PAWPmid‐A, mid‐A‐wave pressure; PAWPQRS, PAWP at QRS onset; PAWPS, instantaneous PAWP measured at the time point simultaneous with PAPD; PAWPZ, PAWP at the Z‐point.

Along with a lower mean value, at an individual level, PAWPmid‐A was lower than PAWPM in 62% of the cases. Compared with the subgroup with PAWPmid‐A < PAWPM, the cohort with PAWPmid‐A > PAWPM demonstrated significantly lower V‐wave amplitude (18.7 [15.1 to 27–4] vs. 30 [22.3 to 38.3] mmHg, P < 0.001). The difference between PAWPmid‐A and PAWPM was strongly related to the V‐wave amplitude (r = 0.69, P < 0.001). In contrast, the A‐wave amplitude did not differ between the two groups (20.6 [17.2 to 25.1] vs. 19 [15.7 to 26.0] mmHg; P > 0.23).

Accordingly, PVRmid‐A was higher than PVRM (P < 0.001), while PVRQRS yielded higher values than PVRmid‐A and PVRM (P < 0.001 in both).

Atrial fibrillation

At the time of RHC, 28 patients were in AF. As illustrated in Figure , in AF patients, PAWPQRS yielded similar figures to PAWP130 and PAWPS (P = 0.124 and P = 0.35, respectively) but significantly higher compared with PAWP140, PAWP150, and particularly PAWP160 (P = 0.04, P = 0.002 and P < 0.001, respectively). These relationships did not alter when the groups were dichotomized based on median hazard ratio (HR).

In line with the aforementioned observations, there was no significant difference between PVR130 and PVRQRS (P = 0.442), whereas PVR140–160 significantly overestimated PVRQRS (P < 0.001).

Accordingly, for comparative analyses based on the consensus proposal in AF, we used the 130 ms time point. When patients with SR and AF were analysed together, these measurements are referred to as PAWPmid‐A/130.

Reproducibility measures of the instantaneous PAWP measurements have been published previously (intra‐observer and inter‐observer intraclass correlation coefficient: 0.98 and 0.97, respectively).

Event timing

PAPD occurred roughly 100 ms following the QRS onset, similarly in the SR and AF cohorts (99 ± 17 and 99 ± 30 ms, respectively) and was significantly associated with HR in both groups (r = −0.75, P < 0.001, and r = 0.38, P = 0.043).

The A‐wave onset occurred 52 ± 7 ms, while the A‐wave peak 150 ± 6 ms after the QRS onset. In all cases, the time point for PAPD befell during the ascending limb of the A‐wave in the corresponding PAWP curve.

Z‐point‐derived measurements

A distinct Z‐point was identified in 48 cases (43%) in the SR cohort, occurring around 180 ms (179 ± 7 ms) after the QRS onset, the time delay being inversely associated with the V‐wave amplitude (−0.47, P = 0.002). As illustrated in Figure , PAWPQRS (15.3 [12.5 to 19.2] mmHg) was significantly lower than all the other pressure measurements (P < 0.001). In contrast, there was no significant difference between PAWPZ (17.6 [14.2 to 21.6] mmHg), PAWPmid‐A (17.3 [14.5 to 21.2] mmHg; P = 0.63), and PAWPM (17.3 [14.7 to 21.2] mmHg; P = 0.49).

The mean difference between PAWPmid‐A and PAWPZ was barely discernible with rather narrow limits of agreement and the in‐between difference significantly associated with the rate of A‐wave pressure decay (r = 0.5, P = 0.002), but not related to the A‐wave or V‐wave amplitude or the PAWPM (P > 0.05, for all).

Diastolic pressure gradient calculations

DPGmid‐A was higher than DPGM (P < 0.001) and significantly associated with the A‐wave pressure decay rate (r = 0.43, P < 0.001) while inversely related to the rate of the A‐wave pressure rise (r = −0.39, P < 0.001). DPGQRS yielded higher values than both DPGmid‐A and DPGM, and it was also significantly associated with A‐wave pressure rise and fall (r = 0.36 P = 0.003 and r = −0.25, P < 0.02; respectively). DPGZ was slightly, yet not significantly, higher than DPGM and DPGmid‐A, whereas DPGQRS was significantly higher than all three other DPG measurements (Table ). In patients with AF, DPG130 was similar to DPGQRS, whereas DPG140–160 was significantly higher (P < 0.001).

Table 3

Diastolic pressure gradient values based on various instantaneous pulmonary arterial wedge pressure measurements in pulmonary hypertension due to left heart disease patients with sinus rhythm vs. atrial fibrillation

	Total (n = 141)	Sinus rhythm (n = 113)	Atrial fibrillation (n = 28)
DPGM (mmHg)	0.8 (‐1.8 – 4.1)	0.8 (‐1.8 – 4.2)SQC	0.9 (‐1.7 – 4.1)MQ
DPGS (mmHg)	3.7 (1.1– 6.8)	3.7 (0.8 – 6.6)MC	3.7 (1.9 – 7.4)MC
DPGQRS (mmHg)	3.6 (1.6– 6.7)	3.5 (1.6 – 7.4)MC	3.9 (0.8 – 6.0)
DPGmid‐A (mmHg)		1.8 (‐0.6 – 5.8)MSQ	—
DPG130 (mmHg)	—	—	5.0 (1.2 – 10.3)
DPG140 (mmHg)	—	—	6.2 (1.6 – 9.9)
DPG150 (mmHg)	—	—	5.6 (2.0 – 9.7)
DPG160 (mmHg)	—	—	8.5 (1.1– 12.5)

Letters in superscript indicate statistically significant (P < 0.05) differences: S, simultaneous measurement; M, mean value; QRS, ECG‐gated measurement; C, according to the consensus paper's recommendation, i.e. mid‐A‐wave in sinus rhythm or 130 ms from QRS onset in atrial fibrillation.

DPG, diastolic pressure gradient; DPGS, DPGQRS, DPGmid‐A, DPG130, DPG140, DPG150, and DPG160 stand for DPG calculated from the PAWPS, PAWPQRS, PAWPmid‐A, PAWP130, PAWP140, PAWP150, and PAWP160, respectively; PAPD, pulmonary arterial diastolic pressure; PAPM, pulmonary arterial mean pressure; PAWP, pulmonary arterial wedge pressure; PVR, pulmonary vascular resistance.

Regarding DPGNEG, among DPGM, a rather high prevalence of 42% was observed, whereas the corresponding numbers for DPGQRS and DPGmid‐A/130 were significantly lower (16% and 28%, respectively).

Differentiation between pulmonary arterial hypertension and pulmonary hypertension due to left heart disease

We tested the diagnostic ability of PAWPM, PAWPmid‐A/130, and PAWPQRS measurements for distinguishing between PAH and PH‐LHD. ROC analysis revealed similar diagnostic ability for the three variables [PAWPM: area under the curve (AUC) 0.98, CI 0.96–0.99, SE 0.009, optimal cut‐off limit at ≥13 mmHg, sensitivity 92%, and specificity 95%; PAWPmid‐A/130: AUC 0.94, CI 0.91–0.98, SE 0.017, optimal cut‐off limit at ≥13 mmHg, sensitivity 89%, and specificity 86%; PAWPQRS: AUC 0.96, CI 0.94–0.99, SE 0.012, optimal cut‐off limit at ≥12 mmHg, sensitivity 87%, and specificity 93%; P < 0.001 for all] (Figure ).

Figure 3

Receiver operating characteristic (ROC) curve comparing the diagnostic ability of the mean pulmonary artery wedge pressure (PAWPM) with the two proposed instantaneous PAWP measurements (PAWPQRS at QRS onset and PAWPmid‐A/130 at mid‐A‐wave in sinus rhythm or 130 ms after QRS onset in atrial fibrillation) for the differential diagnosis of primary pulmonary arterial hypertension and pulmonary hypertension due to left heart disease.

Prognostic value

PH‐LHD patients were followed up over 17 (8–27) months. During this period, 25 deaths occurred, and 27 patients were transplanted. ROC analysis for the prediction of death revealed similar AUC for the three PVR measurements (PVRM: AUC 0.67, CI 0.54–0.80, SE 0.068 P = 0.008; PVRmid‐A/130: AUC 0.66, CI 0.53–0.78, SE 0.064, P = 0.016; PVRQRS: AUC 0.64, CI 0.51–0.77, SE 0.067, P = 0.029). At the cut‐off limit of 3 WU, by using Kaplan–Meier analysis, only PVRM was significantly prognostic for death (P = 0.016), while PVRQRS and PVRmid‐A/130 did not reach statistical significance (P = 0.151 and P = 0.061, respectively). However, when the ROC curves of PVRM and PVRmid‐A for the prediction of death were compared, no statistically significant difference was found in the discriminating power between PVRM and PVRmid‐A/130 (P = 0.36).

In case of DPG, none of the DPGM, DPGQRS, and DPGmid‐A;130 was prognostic at any of the tested cut‐off values (5, 6, 7, 8, and 9 mmHg) (P > 0.05 for all).

Discussion

Our study provides important insights into the various approaches for PAWP assessment. Firstly, we show that the mid‐A‐wave pressure comprises a more accurate EDP estimate as compared with PAWPQRS. In regard to AF, our results advocate for PAWP measurements at the lower end of the time interval proposed by the consensus statement, and in this particular setting, even ECG‐gated measurements seem to comprise a feasible alternative. We demonstrate that the diagnostic ability of the various instantaneous PAWP measurements for differentiating PH‐LHD from PAH is at best similar to that of PAWPM. Finally, PVR derived from momentary PAWP, in contrast to conventional PVR, failed to provide prognostic information.

A precise wedge‐derived EDP assessment requires to account for the transmission time from the LAP as well as for the electromechanical time delay. Consequently, the EDP is expected to occur 130–200 ms following the QRS onset, , in line with the current findings of a median 180 ms time delay for the Z‐point. Braunwald et al. demonstrated that the Z‐point pressure comprised a more accurate EDP estimate compared with PAWPM ; however, its utility is limited owing to its rather infrequent demarcation on the PAWP waveform. Recently, the mid‐A‐wave pressure has been proposed as a preferable EDP estimate. Indeed, we found that PAWPmid‐A reliably represents PAWPZ and thus EDP. Nevertheless, as currently demonstrated, PAWPmid‐A might occasionally overestimate the PAWPZ particularly in cases of steep A‐wave pressure decay, which can occur with increased LV end‐diastolic stiffness.

The absence of coordinated atrial contraction might accentuate the disparity between PAWPM and EDP, and it has been shown that in AF, PAWPM significantly overestimates EDP. Accordingly, in AF, the latest consensus statement recommends instantaneous wedge pressure measurements at 130 to 160 ms following the QRS onset. The lack of LV pressure recordings in our study does not allow for a direct comparison between PAWP at the aforementioned recommended time intervals and the corresponding LVEDP. Nonetheless, our results indicate that measurements at the earliest part (i.e. 130 ms) of the suggested time‐period are more justifiable considering that PAPD befell roughly 100 ms after the QRS onset. Furthermore, case‐by‐case analysis of the instantaneous PAWP occurring concurrently with the PAPD revealed good agreement with the corresponding PAWP130. Of note, ECG‐gated PAWP measurements provided similar values to the corresponding simultaneous pressures as well as to PAWP130. Considering the superior feasibility of the ECG‐gated approach in clinical practice, our results support the use of PAWPQRS as the measurement of choice in AF.

The preferential employment of instantaneous end‐diastolic PAWP measurements is currently supported even in the context of differentiation between PH‐LHD and PAH. It is however debatable whether this new approach could provide superior results in this setting. Indeed, it is commonplace for any experienced invasive cardiologist that patients with PAH do not exhibit substantial pressure pulsatility on PAWP recordings, an observation most probably ascribed to the dampening effect of significantly elevated PVR. In contrast, significant pressure pulsatility (the presence of large V‐waves) is suggestive of PH‐LHD. Accordingly, evasion of the pulsatile component of the PAWP, by employing instantaneous measurements, would preferentially impact on PH‐LHD cases and by yielding lower PAWP values might perplex the differentiation between PH‐LHD and PAH as corroborated by our results showing that end‐diastolic PAWP measurements provided lower specificity than PAWPM in this regard. PAWPM > 15 mmHg has historically been employed for differentiation between PAH and PH‐LHD. Of note, in our cohort, lower values (for PAWPM and PAWPmid‐A/130 an optimal diagnostic cut‐off of 13 mmHg, and for PAWPQRS of 12 mmHg) have been identified, in line with the ongoing discussion for lower differentiating PAWP levels.

Along with the studies refuting the utility of DPG , , , , , , arose the concept that the prognostic inconsistencies associated with this measurement originated, at least partially from an inappropriate approach to DPG derivation, not taking into account the temporal occurrence of the applied pressures. , , Wright and colleagues first attempted to standardize DPG calculation proposing instantaneous PAWP at QRS onset, thereby limiting the distorting effects of pressure pulsatility on DPG. Temporal adjustment of PAWP assessment is also suggested in the current consensus statement, which furthermore proposes that even PVR used for PH classification should also be based on instantaneous PAWP. Although the suggested instantaneous PAWP do indeed evade some physiological incongruencies of PAWPM, these adjustments did not render improved prognostic or diagnostic information in our study. This might not be surprising, considering what exactly the PAWP is expected to represent in the specific setting. For DPG calculation, approximation of the end‐diastolic LV pressure is important; thus, end‐diastolic PAWP is likely the method of choice. However, for PVR, accounting for both the pulsatile and the static pressures as provided by PAWPM might better represent the haemodynamic load imposed on the right ventricle. In light of the current findings, before introducing momentary PAWP in place of PAWPM, cautious consideration and most probably more data are required to clarify whether the use of instantaneous PAWP measurements should really be generalized.

Limitations

Our investigation has limitations. The study cohort was relatively small; however, it is still among the largest of its kind considering the detailed analysis of invasive pressure waveforms. Conductance catheters would have provided better accuracy than fluid‐filled catheters; on the other hand, the current approach represents everyday clinical practice. We did not perform LV catheterization, which hampers the comparability of our data; however, that would not have been ethically defendable.

In accordance with guideline recommendations, we measured PAWP at end‐expiration. This approach may lead to an overestimation of pulmonary vascular pressures in case of hyperventilation and/or obstructed airways. Whether averaging pulmonary pressure measurements over multiple respiratory cycles would be the preferred method was not addressed in the present study.

Our study did not include a control group of healthy individuals, which confines our results to HF and PAH; however, regarding that all participants underwent invasive testing, the recruitment of such control cohort would also raise major ethical issues.

Conclusions

Instantaneous PAWP measurements evade the distorting effect of pressure pulsatility and reliably represent EDPs. Nevertheless, these adjustments failed to translate into diagnostic or prognostic superiority compared with conventional mean PAWP‐derived indices.

Conflict of interest

None.

Funding

A.I.N. was supported by the János Bolyai Scholarship (Bolyai Foundation) of the Hungarian Academy of Sciences.

Figure S1. Flowchart explaining patient composition of the study cohort. RHC, right heart catheterisation; HF, heart failure; ARVC, arrhythmogenic right ventricular cardiomyopathy; PH, pulmonary hypertension; PAPM, mean pulmonary artery pressure; PAWPM, mean pulmonary artery wedge pressure.

Click here for additional data file.