Evolution of Obstructive Lung Function in Advanced Pulmonary Arterial Hypertension

To the Editor:

Pulmonary arterial hypertension (PAH) is characterized by increased pulmonary arterial impedance leading to exercise intolerance and right heart failure. With improvement in therapies targeting right heart failure, there is an increasing interest in novel pathways by which PAH diminishes exercise capacity, a key therapeutic target. Although restrictive lung disease is well recognized in PAH, peripheral airway obstruction (1, 2) and dynamic hyperinflation (3, 4) have also been described in prior cross-sectional PAH studies. We sought to characterize the evolution of obstructive physiology in advanced PAH and to investigate potential mechanisms using a combination of imaging and histopathology.

Patients with PAH and subsequent lung transplant at the University of California, Los Angeles with available explant pathology, high-resolution computed tomography (HRCT) chest imaging, and sequential pulmonary function were included (Institutional Review Board #11-003042). The studies closest to the PAH diagnosis and transplant dates were chosen for analysis. No patients were on PAH-specific therapy at diagnosis, and all were receiving triple combination PAH-specific therapy (including parenteral prostacyclin) at transplant. Measurements of arterial and venous total blood volume normalized for total lung volume (TBV/TLV) (5) and percent emphysema (%LAA−950 [% of parenchymal voxels below −950 Hounsfield units (HU) on inspiratory phase]) were obtained using the Chest Imaging Platform (www.chestimagingplatform.org) (6). Measures of air trapping (%LAA−856 [% of parenchymal voxels below −856 HU on expiratory phase]) (7) were obtained using proprietary software (IBIS) (8). Airway short axis diameters were measured at two airway trees (RB1 and RB10; generations 3, 4, and 5) in patients with available imaging at PAH diagnosis and transplant (9). Obstructive lung disease (OLD) was defined as FEV₁/FVC z-score (FEV₁/FVCz) (10) within the fifth percentile. Data are presented as medians and interquartile ranges for continuous variables. Differences were assessed using the Wilcoxon signed rank test (paired samples) with two-sided P values. Correlative analyses were performed using Spearman correlation. Statistical analyses were performed in R 4.0.1 (R Foundation for Statistical Computing). P values <0.05 were considered significant.

Demographics, pulmonary function, chest imaging parameters, and hemodynamic data (at diagnosis and transplant) for this PAH cohort are shown in Table 1. Patients were generally young, never-smoker females with a diagnosis of PAH from 1993 and 2014 and who underwent transplant from 2004 to 2020. The final PAH diagnosis was based on clinical assessment and review of radiology and pathology and included idiopathic PAH (n = 10), pulmonary capillary hemangiomatosis (n = 2), amphetamine or anorexigen use (n = 2), and corrected congenital (n = 1). The median time between diagnostic right heart catheterization and transplant was 8.7 years. There were no significant changes in pulmonary hemodynamics between diagnosis and transplant (Table 1).

Table 1.

Demographic, HRCT Chest Imaging, and Hemodynamic Data of a PAH Cohort (N = 15, Unless Otherwise Stated) from Diagnosis to Lung Transplantation

Demographics		CT-based Parenchyma
Age at diagnosis, yr	35 (29 to 44)	%LAA−950	1.15 (0.51 to 2.77)
Age at transplant, yr	49 (37 to 53)	>1% LAA−950	7/15
Disease duration, yr	8.7 (6.4 to 15.9)	>2% LAA−950	4/15
Sex, F/M	13/2	>5% LAA−950	1/15
Smoking >5 pack-years	1/15	%Air trapping (n = 5)	0 (0 to 1.5)

Change in Pulmonary Function
	n	Diagnosis	Transplant	P Value*	Difference/Year
FEV1, L	13	2.36 (2.13 to 2.77)	1.52 (1.28 to 1.95)	0.0002	−0.14 (−0.18 to −0.11)
FEV1% predicted	13	82.2 (62.0 to 86.3)	56.0 (45.7 to 61.0)	0.0002	−4.3 (−5.5 to −2.5)
FVC, L	13	2.89 (2.56 to 3.34)	2.25 (1.99 to 2.64)	0.0002	−0.16 (−0.21 to −0.10)
FVC, % predicted	13	80.4 (69.7 to 90.8)	62.2 (58.9 to 70.6)	0.0005	−3.6 (−5.2 to −2.9)
FEV1/FVC	13	76.0 (70.0 to 82.0)	71.0 (66.5 to 76.5)	0.008	−1.1 (−1.3 to −0.5)
FEV1/FVCz	13	−0.8 (−1.7 to 0.2)	−1.5 (−2.3 to −0.90)	0.006	−0.10 (−0.14 to −0.04)
FEF25–75, L/s	11	2.54 (1.84 to 2.88)	1.08 (0.70 to 1.53)	0.002	−0.10 (−0.14 to −0.04)
DlCO, ml/min/mm Hg	9	16.4 (13.5 to 18.6)	14.6 (13.3 to 15.1)	0.02	−2.7 (−5.3 to −1.9)
DlCO % predicted	9	74.8 (54.2 to 83.1)	62.9 (46.7 to 74.5)	0.02	−3.4 (−5.5 to −0.8)
DlCO/VA	6	3.9 (2.5 to 4.2)	3.5 (2.0 to 3.9)	0.18	−0.05 (−0.11 to 0.0)
DlCO/VA % predicted	6	87.8 (58.9 to 98.4)	83.0 (48.4 to 88.7)	0.44	−1.0 (−2.3 to 0.36)

Change in Hemodynamics
	n	Diagnosis	Transplant	P Value*
mPAP, mm Hg	13	53 (48 to 55)	55.5 (51.0 to 68.8)	0.31
PAWP, mm Hg	12	10.5 (8.8 to 12.5)	14.0 (12.5 to 15.0)	0.17
CI, L/min/m2	12	2.4 (2.0 to 3.1)	2.4 (1.8 to 3.0)	1.0
PVR, Wood units	12	11.2 (7.5 to 12.2)	11.2 (9.6 to 14.6)	0.62

Correlates with FEV1/FVCz at Transplant				Correlates with Yearly Change in FEV1/FVCz
	Transplant	R	P Value†		R	P Value†
Age at transplant, yr	49 (37 to 53)	0.16	0.56	Age at diagnosis	0	1.0
mPAP, mm Hg (n = 14)	56 (51 to 69)	−0.57	0.03	mPAP, %Δ/yr (n = 13)	−0.07	0.83
PVR, mm Hg (n = 13)	11.2 (7.5 to 12.2)	−0.04	0.91	PVR, %Δ/yr (n = 12)	−0.66	0.02
PA diameter, mm	42 (38 to 48)	−0.66	0.01
Arterial TBV/TLV, ml vessel/L lung	31 (29 to 34)	−0.83	<0.0001
Venous TBV/TLV, ml vessel/L lung	17 (15 to 18)	−0.44	0.09

Definition of abbreviations: CI = cardiac index; CT = computed tomography; FEV₁/FVCz = FEV₁/FVC z-score; FEF_25–75 = forced expiratory flow, midexpiratory phase; HRCT = high-resolution computed tomography; %LAA−950 = percentage of low attenuation areas below −950 Hounsfield units; mPAP = mean pulmonary arterial pressure; PA = (main) pulmonary artery; PAH = pulmonary arterial hypertension; PAWP = pulmonary arterial wedge pressure; PVR = pulmonary vascular resistance; TBV = total blood volume; TLV = total lung volume.

Data are shown as median (interquartile range) unless otherwise stated.

^*P values from Wilcoxon signed rank test.

^† P values are from Spearman correlation.

At the time of transplant, the majority of patients demonstrated absolute FEV₁ (14/15) and FVC (13/15) values within the fifth percentile, and 6/15 patients satisfied OLD criteria (Figure 1). Between diagnosis and transplant, there were significant yearly declines in absolute FEV₁ (liters, median) −0.14 (interquartile range [IQR], −0.18 to −0.11) and FVC −0.16 (IQR, −0.21 to −0.10). The FEV₁/FVC ratio significantly declined from 76.0 (IQR, 70.0–82.0) to 71.0 (IQR, 66.5–76.5) between diagnosis and transplant, representing a yearly decline of 1.1 (IQR, 0.5 to 1.3). Similarly, the FEV₁/FVCz significantly declined from −0.8 (IQR, −1.7 to 0.2) to −1.5 (IQR, −2.3 to −0.90), representing a yearly decline of −0.10 (IQR, −0.14 to −0.04). As such, the majority of patients advanced to lower (more negative) FEV₁/FVCz percentile classifications between diagnosis and transplant (Figure 1), consistent with worsening obstructive lung function. Based on imaging at transplant, there was none to minimal evidence for emphysema (N = 15; 1/15 with %LAA−950 > 5%) or air trapping (n = 5; ⩽1.5% by %LAA–856 HU) (Table 1) and no findings to support any other parenchymal or airway disorder.

Figure 1.

The left upper panels illustrate the changes in absolute FEV₁ and FVC (in liters; center lines represent medians; boxes encompass middle two quartiles) illustrating a loss of both FEV₁ and FVC between the diagnosis of pulmonary arterial hypertension and lung transplantation. The left lower panel illustrates the evolution of FEV₁/FVC z-score (FEV₁/FVCz) for each subject between those same time points illustrating the decline in FEV₁/FVCz (y-axis: FEV₁/FVCz; z-axis: FEV₁/FVCz percentile). The left lower panel inlet accounts for the number of subjects within each FEV₁/FVCz percentile comparing two time points, pulmonary arterial hypertension diagnosis and lung transplantation. The top right panel displays two representative vascular reconstructions (subjects #14 and #12) with an accompanying color legend depicting the spectrum of pulmonary artery sizes ranging from 0.8 mm² (red) to 100 mm² (blue), based on cross-sectional area. Importantly, these two subjects demonstrate the association between generally larger pulmonary arteries and an expanded thoracic blood volume with a more negative FEV₁/FVCz. Similarly, the axial sections exhibit the relatively severe main pulmonary artery dilatation in the subject with lower FEV₁/FVCz compared with the subjects with modest main pulmonary artery dilatation and higher FEV₁/FVCz (right bottom panels). PA = pulmonary artery.

We then examined correlations between FEV₁/FVCz and imaging/hemodynamic measures. Increased mean pulmonary arterial pressure, main pulmonary artery (PA) diameter, and TBV/TLV (arterial, not venous) were associated with lower FEV₁/FVCz (Table 1 and Figure 1) at transplant. Between diagnosis and transplant, increasing pulmonary vascular resistance (%Δ/year) was associated with decreasing FEV₁/FVCz (Δ/year). In six patients with available HRCT chest imaging at diagnosis and transplant, the three patients with the most rapid decline in FEV₁/FVCz (Δ/year) (−0.11, −0.09, and −0.04 vs. −0.02, 0.05, and 0.19) demonstrated a greater enlargement in main PA diameter (%Δ/year) (12.8%, 3.6%, and 4.8% vs. 2.4%, 3.1%, and 0%) compared with the remaining three patients, respectively. Similarly, patients with a greater enlargement in main PA diameter revealed a larger decrease (%Δ/year) in aggregate airway (−13.9%, −4.5%, and 1.0% vs. −5.3%, 3.3%, and 0.9%), respectively. In four of these six patients with imaging amenable to vascular reconstruction at diagnosis and transplant, the two patients with the largest decline in FEV₁/FVCz (−0.11 and −0.4 vs. 0.05 and 0.19) had greater increases in arterial TBV/TLV (18.2% and 24.7% vs. 9.7% and 8.3%) compared with the two remaining patients, respectively.

The explanted lungs did not demonstrate any histopathology that would otherwise explain the OLD (n = 6) or progressively obstructive lung function (n = 9) in our cohort. Pathologic examination of all explanted specimens did not reveal evidence of emphysema, interstitial fibrosis, or small/large airway disease. Notably, there was no evidence for acute or chronic airway disease as indicated by the absence of airway inflammation, fibrosis, narrowing or collapse, or submucosal or mucous gland hyperplasia.

Based on sequential pulmonary function, this study revealed progression of obstructive lung function during a median of 8.7 years between the diagnosis of PAH and subsequent transplant. Development of airflow obstruction correlated best with chest imaging variables indicative of an expanded thoracic blood volume, despite unchanged pulmonary hemodynamics. Air trapping, a common chest imaging correlate for airway obstruction, was not seen in the subjects with available inspiratory and expiratory images in this cohort. Congruent with these imaging findings, explant pathology confirmed the absence of an alternate explanation for obstructive lung function. To our knowledge, this is the first study to evaluate sequential pulmonary function and to demonstrate the evolution of obstructive physiology in PAH.

Prior authors have postulated that airway compression related to PA dilatation (11, 12) may be a potential mechanism to explain peripheral airway obstruction (1, 2) in PAH. Our study provides an additional framework to support this hypothesis by relating the progression of obstructive physiology (based on sequential pulmonary function) to the engorgement or enlargement of the pulmonary artery tree (based on HRCT chest imaging). In support of obstructive physiology, dynamic hyperinflation and related dyspnea during exercise have also been reported in PAH, despite normal FEV₁/FVC ratios (3). Nevertheless, the definitive mechanism for obstructive physiology in PAH remains unanswered.

Our study is limited by its retrospective nature and small sample size. Importantly, the cohort was intentionally restricted to include PAH with explant pathology, which allowed definitive exclusion of alternate etiologies of obstruction. HRCT imaging methods of measuring emphysema and air trapping rely on threshold methods that are not validated in PAH, particularly in the context of alveolar or interstitial abnormalities, which may be present in advanced PAH (13). Only five patients had inspiratory and expiratory images available for analysis. The inherent effort dependence of pulmonary function testing and the initiation of PAH-specific therapy after diagnosis are additional factors that potentially contributed to loss of lung function. As this patient cohort focused on advanced PAH requiring transplant, generalization to less severe PAH may not be justified.

In conclusion, our analysis suggests that patients with advanced PAH may display progressive obstructive physiology associated with chest imaging parameters indicative of increased thoracic blood volume, but without air trapping. Explant pathology verified the absence of another explanation for obstructive physiology. The findings support the hypothesis that enlarged or engorged pulmonary arteries directly or indirectly modulate dynamic airway function. In addition, we propose that PAH be added to the existing assembly of respiratory diseases known to manifest both restrictive and obstructive physiology. Whether this obstructive physiology contributes to worsening $\dot{\text{V}}$/$\dot{\text{Q}}$ mismatch and gas exchange in PAH or if it responds to bronchodilator therapy remains to be determined. The consideration of sequential lung function as an endpoint in future PAH clinical trials may be warranted. Future research is required to confirm the evolution of obstructive physiology and pinpoint the underlying mechanism in PAH.