Integrating Hemodynamics Identifies an Extreme Pulmonary Hypertension Phenotype

Pulmonary hypertension (PH) is a highly morbid disease defined foremost by elevated mean pulmonary artery pressure (mPAP) measured during right heart catheterization (RHC) at rest. Patients are classified further into one of three PH hemodynamic subgroups based on specific pulmonary artery wedge pressure (PAWP) and pulmonary vascular resistance (PVR) thresholds: pre-capillary PH (PAWP ≤15 mmHg + PVR ≥3.0 WU), isolated post-capillary PH (PAWP >15 mmHg + PVR <3.0 WU), and combined pre-/post-capillary PH (PAWP >15 mmHg + PVR ≥3.0 WU).¹

Achieving an accurate hemodynamic diagnosis is essential to ensure the appropriate clinical management of PH patients.² However, a binary approach to hemodynamic interpretation may over-generalize risk assessment for individual patients.³ Prior studies have focused largely on differences in individual hemodynamic parameters for prognostication,⁴ even though point-of-care assessment in PH hinges on integrating data collected during RHC.⁵ Further, little is established on the association between PAWP and mortality despite the importance of PH to cardiovascular morbidity in left heart disease.⁶

To address these dilemmas, we analyzed RHC and outcome data from the Veterans Affairs Clinical Assessment, Reporting, and Tracking Program (VA-CART) (2008–2016). The methods for assembling the study cohort have been reported recently.⁷ We focused our analysis on patients with mPAP ≥19 mmHg, which is independently associated with increased mortality in this study population,⁴ with ≥1 year follow-up (median 1153 [570–1971] days). A Cox proportional hazards model incorporating age, sex, race, and 19 other covariates was used to assess the association between PAWP and all-cause mortality.⁷ The Colorado Multiple Institutional Review Board approved this study with a waiver of informed consent.

From a study population of 32,725 patients (66 [61–74] yr; 97% male, 30.1 [26.1–35.0] kg/m²), a history of left heart disease or congestive heart failure (CHF) was observed in 26,255 (80.2%) patients, and 29,352 (89.7%) patients had systemic hypertension.⁷ We observed a bimodal (i.e., U-shaped) distribution in PAWP relative to all-cause mortality (Figure). Among all PAWP 1 mmHg bins, the absolute mortality rate and adjusted hazard for mortality were lowest for patients with PAWP=12 mmHg, which was selected as the reference point for further analyses involving PAWP. From this approach, PAWP=15 mmHg was the minimal PAWP bin above 12 mmHg at which the mortality hazard was ≥1.0. Similarly, PAWP=11 mmHg was the minimal PAWP bin below 12 mmHg at which the mortality hazard was ≥1.0.

Figure. The association between pulmonary artery wedge pressure (PAWP) and all-cause mortality in patients with pulmonary hypertension.

Upper panel illustrates hazard ratio for mortality as a function of PAWP adjusted for demographic variables and medical co-morbidities as described previously⁴ (green line with 95% confidence interval band, left axis) and absolute mortality (gray bars, right axis) in 1 mmHg bins. Vertical lines separate PAWP into groups described further in the text (PAWP <12, 12–15, >15 mmHg). The number of patients in each bin is indicated at the base of the bars and the percentage with co-morbid lung disease (COPD + interstitial lung disease) is enclosed by parentheses. The lower panel illustrates the median right atrial pressure (RAP), mean pulmonary artery pressure (mPAP), pulmonary vascular resistance (PVR), and cardiac output (CO) for each PAWP bin.

Indeed, the association between PAWP and adjusted mortality spanned three general ranges: <12, 12–15, and >15 mmHg, representing 14.6% (N=4,793), 22.2% (N=7,280), and 63.1% (N=20,652) of the study population, respectively. There was no meaningful difference in age across these patient subgroups (66.7 [61.3–66.9] vs. 66.9 [61.7–74.0] vs. 66.5 [51.2–73.5] yr, P=0.009]; however, a step-wise increase was observed in the prevalence of CHF (46.4 vs. 50.4 vs. 71.5%, P<0.001), atrial flutter (19.1 vs. 24.5, vs. 36.6%, P<0.001), chronic kidney disease (24.9 vs. 27.9 vs. 37.4%, P<0.001) and obstructive sleep apnea (11.4 vs. 13.1 vs. 15.5%, P<0.001). By contrast, connective tissue disease (4.8 vs. 3.6 vs. 2.6%, P<0.001) and chronic obstructive pulmonary disease (COPD) (41.15 vs. 34.6 vs. 35.0%, P<0.001) were more common in the PAWP <12 mmHg subgroup compared to the 12–15 mmHg and >15 mmHg subgroups.

The hemodynamic profile (median [IQR]) for patients in the PAWP <12, 12–15, and >15 mmHg ranges was: right atrial pressure (RAP) 6 [4–8] vs. 8 [6–10] vs. 12 [9–16] mmHg, P<0.001; mPAP 22 [20–27] vs. 24 [21–28] vs. 34.0 [28–41] mmHg, P<0.001; PVR 2.6 [2.0–3.8] vs. 2.0 [1.5–2.9] vs. 2.1 [1.3–3.2] WU, P<0.001; cardiac output (CO) 5.2 [4.3–6.2] vs. 5.3 [4.4–6.3] vs. 4.9 [4.0–6.1] L/min, P<0.001; transpulmonary gradient 14 [11–18] vs. 10 [8–14] vs. 10 [7–15] mmHg, P<0.001; and RA/PAWP ratio 0.64 [0.45–0.82] vs. 0.58 [0.43–0.75] vs. 0.54 [0.41–.68], P<0.001.

Compared to patients with PAWP 12–15 mmHg, the adjusted hazard ratio for mortality was increased significantly in patients with PAWP <12 mmHg (HR 1.17, 95% CI: 1.11–1.25; X ²=27.0, P<0.001) and PAWP >15 mmHg (HR 1.19, 95% CI: 1.14–1.24; X²=54.4, P<0.001). Compared to PAWP=12 mmHg, patients with PAWP=15 mmHg had higher mPAP (23 [21–27] vs. 25 [22–29] mmHg, P<0.001) and lower PVR (2.1 [1.6–2.9] vs. 1.9 [1.3–2.8] WU, P<0.001), but similar CO (5.3 [4.4–6.3] vs. 5.3 [4.4–6.3] L/min, P=0.828).

We identify an extreme PH hemodynamic profile among patients with PAWP <12 mmHg that is characterized by elevated PVR, transpulmonary gradient, and RAP/PAWP ratio. These data suggest that there is opportunity to prognosticate PH patients with low PAWP even in the absence of cor pulmonale, as the median CO (5.2 L/min) was above the range independently associated with mortality in this population and the follow-up period in this study is beyond longevity expectations when frank end-stage right heart failure is diagnosed at the time of RHC.^7–9 This subgroup was enriched with lung disease-PH, but may also include some patients with combined pre-/post-capillary PH on diuretic therapy and a small number of pulmonary arterial hypertension patients. Our data also provide the first comprehensive information on the association between elevated PAWP and mortality hazard in PH. In the PAWP >15 mmHg subgroup, RAP was also elevated but the RAP/PAWP ratio was decreased relative PAWP <12 and 12–15 mmHg patients, indicative of biventricular heart failure. These findings reinforce the importance of elevated left heart filling pressure to prognosis in post-capillary PH.

The adjusted mortality risk was unchanged between PAWP 12–15 mmHg. We have shown previously that in this study population, there is a continuous relationship between mPAP and mortality hazard,⁴ suggesting that mPAP may be particularly useful for risk stratification of patients in this hemodynamic zone between pre- and post-capillary PH. Importantly, reference PAWP levels alternative to 12 mmHg as well as other statistical approaches could have been used to define the elevated vs. low PAWP subgroups. The retrospective design and inaccessibility to comprehensive patient-level data are important methodological limitations that prevent this study from identifying an optimal hemodynamic profile among patients with elevated mPAP, which is likely to vary by clinical and comorbidity profile. Rather, these data lay the framework for an evidence-based approach to delineating PAWP ranges that separate pre- and post-capillary PH subgroups. The study population was largely male, which may limit the generalizability of our findings to women or other sex-balanced cohorts. Prospective studies are therefore needed to validate our data, as over-representation of males and referral bias are important considerations when interpreting our findings.

To the best of our knowledge, this study provides the first comprehensive analysis involving multiple hemodynamic parameters modeled simultaneously and linked to clinical outcome in patients with PH. Applying this approach to a large RHC referral population inclusive of patients with elevated mPAP and highly enriched with cardiopulmonary co-morbidities identified a bimodal distribution in mortality. This included a particularly high-risk hemodynamic profile defined by low PAWP, elevated PVR and widened transpulmonary gradient. Although elevated PAWP is known to associate with adverse clinical events in left heart failure,¹⁰ the current data expand on this concept by illustrating an important risk continuum associated with PAWP in post-capillary PH. Overall, our findings support integrating RHC variables as key step toward improving PH classification, and suggest that PAWP <12 mmHg and >15 mmHg may be reasonable thresholds delineating pre- and post-capillary PH. Additional data are needed to clarify the relevance of mPAP and other variables for prognosticating patients across a hemodynamic transition zone between pre- and post-capillary PH that emerged in this study.