Pulmonary arterial hypertension with signs of venous/capillary involvement in connective tissue diseases: paradigms and paradoxes

Abstract

Pulmonary arterial hypertension (PAH) with features of venous and/or capillary involvement, formerly pulmonary veno-occlusive disease (PVOD), represents a rare and severe subset of pulmonary hypertension. While PVOD is typically idiopathic, heritable or drug/toxin-induced, “PVOD-like” features have long been recognised in PAH associated with connective tissue diseases (CTDs), especially systemic sclerosis (SSc). This review synthesises the available evidence on this phenotype, integrating published data with an unreported cohort of 25 patients. Venular remodelling is a frequent histological finding in SSc-PAH lung explants, but high-resolution computed tomography (HRCT) signs of PVOD are inconsistently observed at PAH presentation, reflecting a progression of the lung vasculopathy and/or an unmasking effect of pulmonary vasodilators. Patients often exhibit major functional limitation, profound impairment in gas transfer and severe haemodynamic compromise. Assessing venular involvement in SSc-PAH poses unique challenges, due to possible lung fibrosis on HRCT and differential diagnosis with an occult post-capillary component. Despite the lack of robust evidence, PAH-approved therapies are commonly used, although with caution due to a high risk of pulmonary oedema with pulmonary vasodilators. As these patients also display significant arteriolar involvement, which can benefit from these drugs, dedicated treatment strategies warrant further investigation. Prognosis remains dismal, with a 5-year survival around 45–50%. Similar PVOD-like patterns have also anecdotally been reported in systemic lupus erythematosus, rheumatoid arthritis, mixed connective tissue disease, Sjögren syndrome and inflammatory myopathies, with comparable features. Overall, PVOD-like disease in CTD-PAH constitutes a high-risk phenotype with unresolved pathogenic and management challenges, requiring refined patient stratification and improved therapeutic strategies.

Shareable abstract

PVOD-like disease in CTD-PAH is a high-risk phenotype marked by progressive venular remodelling, diagnostic challenges and therapeutic uncertainty. Improved stratification and dedicated treatment strategies are urgently needed. https://bit.ly/4wdg43T

Introduction

Pulmonary arterial hypertension (PAH) with features of venous/capillary involvement, formerly referred to as pulmonary veno-occlusive disease (PVOD) and/or pulmonary capillary haemangiomatosis (PCH), is a rare and severe subset of pulmonary hypertension (PH). Compared to other forms of PAH, it is characterised by a preferential remodelling of small pulmonary veins and capillaries, resulting in increased vascular resistance, elevated pulmonary pressures and right-heart failure [1, 2]. Despite considerable progress in its pathogenic characterisation, its prognosis remains dismal due to a poorer response to drugs approved for PAH, the frequent occurrence of pulmonary oedema induced by pulmonary vasodilators and the absence of a validated therapeutic alternative [3].

PVOD/PCH may occur sporadically, often in association with exposure to organic solvents or chemotherapeutic agents, or as a heritable condition with autosomal recessive transmission caused by biallelic mutations in the Eukaryotic Translation Initiation Factor 2 Alpha Kinase 4 (EIF2AK4) gene [2–4]. Aside from these prototypic situations, features of venous/capillary involvement have also been reported in PAH patients with connective tissue diseases (CTDs), and especially with systemic sclerosis (SSc) [5]. This scenario is often referred to as “PVOD-like”, reflecting potential differences in terms of pathogenesis, clinical presentation and disease course compared to idiopathic/heritable PVOD [6–8].

The term “CTD” refers to a heterogeneous group of immune-mediated diseases characterised by inflammation, tissue damage and fibrosis, and sharing common clinical and serological features. It encompasses several conditions, sometimes overlapping in the same patient, namely SSc, systemic lupus erythematosus (SLE), mixed connective tissue disease (MCTD), Sjögren syndrome (SjS) and inflammatory myopathies (such as antisynthetase syndrome). Antinuclear antibodies are typically positive, although this is not always mandatory nor sufficient to make a CTD diagnosis. Due to pathogenic and phenotypic similarities, rheumatoid arthritis (RA) is often included within the CTD spectrum, although not associated with anti-nuclear antibody (ANA) positivity. As per their systemic nature, CTDs can induce PH by several mechanisms, which are sometimes intricated, including pulmonary vasculopathy (group 1), left-heart disease (LHD) (group 2), interstitial lung disease (ILD) (group 3) and/or chronic thromboembolic PH, particularly in the context of antiphospholipid syndrome (group 4) [9].

Despite its clinical relevance, the occurrence of PVOD-like features in CTDs has mostly been reported through case reports and series; a comprehensive overview of the available evidence on this critical topic is currently lacking. In this review, we try to fill this gap by synthesising the available data on the clinical and therapeutic characteristics of CTD-PAH patients with venous/capillary involvement, distinguishing SSc cases from other, potentially less documented, CTDs. To further our contribution to the field, we also included a previously unreported cohort of 25 patients followed in our centre (supplementary methods) and presented the compiled data from both sources alongside those of a large idiopathic/heritable PVOD cohort [10] for reference. In doing so, we hope to clarify current knowledge, identify areas of consistency and discrepancy, and explicitly highlight the major gaps that remain unresolved to guide future research efforts.

Venous/capillary involvement is a classical occurrence in SSc-PAH but with possible differences from idiopathic/heritable PVOD

The occurrence of venular/capillary involvement during the course of CTD-PAH is best documented in SSc, with 20 published references representing a total of 87 patients [8, 11–28]. The most common condition associated with PAH in Western countries [29], SSc is characterised by widespread vasculopathy, abnormal immunologic activation and extensive fibrosis, making it one of the most severe systemic autoimmune diseases [30]. We present hereafter the compiled data of our literature review (table 1, supplementary table), alongside our own cohort of 25 patients and a population of idiopathic/heritable PVOD [10] for reference.

TABLE 1.

Characteristics of systemic sclerosis (SSc)–pulmonary arterial hypertension (PAH) patients with features of venous/capillary involvement (“SSc-pulmonary veno-occlusive disease (PVOD)-like”), with data from an original cohort and previously published cases, and comparison with idiopathic/heritable (i/h) PVOD patients

	N	“SSc-PVOD-like” Original cohort (N=25)	N	“SSc-PVOD-like” Published cases [8, 11–28] (N=87)	i/h PVOD [10] (N=327)
Demographics and past medical history
Female, n (%)	25	22 (88)	72	58 (81)	100 (31)
Age at PVOD diagnosis (years), mean±sd	25	64±8.5	56	63±13	65±14
History of cigarette smoking, n (%)	25	8 (32)	15	7 (47)	–
Coronary artery disease, n (%)	25	2 (8)	7	0 (0)	–
Systemic hypertension, n (%)	25	12 (48)	7	3 (43)	–
Left ventricular systolic dysfunction, n (%)	25	0 (0)	50	0 (0)	–
Left ventricular diastolic dysfunction, n (%)	23	1 (4)	49	1 (2)	–
COPD/emphysema, n (%)	25	11 (44)	54	1 (2)	0 (0)
SSc diagnosis
SSc subtype	25		49		–
Limited, n (%)		21 (84)		42 (86)
Diffuse, n (%)		4 (16)		7 (14)
Autoantibodies	25		20		–
ACA, n (%)		19 (76)		16 (80)
ATA, n (%)		2 (8)		1 (5)
Other, n (%)		4 (16)		3 (15)
ILD (status at PVOD diagnosis)	25		21		–
None, n (%)		15 (60)		12 (57)
Limited, n (%)		9 (36)		2 (10)
Extensive, n (%)		1 (4)		7 (33)
PVOD diagnosis
Disease duration					–
Since SSc diagnosis (years), median (range)	25	9 (1–16)	12	4 (2–11.5)
Since PAH diagnosis (years), median (range)	25	1 (0–2)	13	0 (0–1)
PVOD and PAH concomitance	25		16		–
PVOD suspected at the time of PAH diagnosis, n (%)		10 (40)		5 (31)
PVOD suspected during PAH follow-up, n (%)		15 (60)		11 (69)
Chest HRCT suspicion
Radiological findings	25		15
Centrilobular ground-glass opacities, n (%)		23 (92)		12 (80)	NA
Mediastinal lymphadenopathy, n (%)		21 (84)		8 (53)	NA
Interlobular septal thickening, n (%)		12 (48)		10 (67)	NA
Number of findings per patient	25		82
0–1 sign, n (%)		0 (0)		5 (6)	NA
≥2 signs, n (%)		25 (100)		77 (94)	NA
Histological confirmation, n (%)	25	0 (0)	42	17 (41)	NA
Baseline characteristics at PVOD diagnosis
WHO functional class	25		68
Class I/II, n (%)		11 (44)		10 (15)	49 (15)
Class III/IV, n (%)		14 (56)		58 (85)	278 (85)
6MWD	25		47
Absolute value (m), mean±sd		328.2±101		209±151	242±160
6MWD <440 m, n (%)		20 (80)		47 (100)
Cardiac biomarkers
NT-proBNP (ng·L−1), median (range)	15	1447 (259–2099)	10	1830 (390–3000)	2040 (413–4491)
BNP (ng·L−1), median (range)	14	253 (133.5–495.5)	22	60 (43–306)	273 (78–531)
Elevated NT-proBNP/BNP, n (%)	25	19 (76)	4	3 (75)	260 (80)
Arterial blood gas
PaO2 (mmHg), mean±sd	11	69±17	32	60±11	57±13
PaCO2 (mmHg), mean±sd	11	32±5	25	32±6	32±5
PaO2 <60 mmHg, n (%)	11	4 (36)	10	6 (60)
Pulmonary function tests
FVC (% pred), mean±sd	25	91.7±19	58	96±27	95±22
TLC (% pred), mean±sd	23	90.4±10	29	91±15	93±17
FEV1 (% pred), mean±sd	25	83.9±17	71	86±18	88±20
DLCO (% pred), mean±sd	22	39.7±8	77	36.4±15	33±13
DLCO/VA (% pred), mean±sd	19	45.4±10	28	43.7±15	42±16
DLCO <60%, n (%)	22	21 (95)	72	71 (99)	NA
DLCO <40%, n (%)	22	9 (41)	11	6 (55)	NA
Right-heart catheterisation
mPAP (mmHg), mean±sd	25	40.7±11	81	42±9	45±11
mPAP >35 mmHg, n (%)	25	17 (68)	47	46 (98)	NA
PAWP (mmHg), mean±sd	25	8.2±3	76	9±3	9±4
RAP (mmHg), mean±sd	25	6.8±4	51	7±4	8±5
PVR (WU), mean (± sd)	25	7±3.1	75	9±5	9±4
PVR >5 WU, n (%)	25	16 (64)	47	46 (98)	NA
CO (L·min−1), mean±sd	25	4.6±0.9	59	4.0±1.4	4.4±1.3
CI (L·min−1·m−2), mean±sd	25	2.7±0.6	69	2.4±0.8	2.4±0.7
CI <2.5 L·min−1·m−2, n (%)	25	12 (48)	9	7 (78)	NA
SVI (mL·m−2), mean±sd	25	33±9	1	28.6	33±11
SvO2 (%), mean±sd	19	65.4±7	39	64.7±11	60±9
Treatment strategy
PAH therapies
Already ongoing before PVOD diagnosis	25		26		–
None, n (%)		0 (0)		6 (23)
Single therapy, n (%)		10 (40)		11 (42)
Dual therapy, n (%)		13 (52)		8 (31)
Triple therapy, n (%)		2 (8)		1 (4)
Initial strategy following PVOD diagnosis	25		30
None, n (%)		0 (0)		12 (40)	49 (15)
Single therapy, n (%)		13 (52)		11 (37)	221 (68)
Dual therapy, n (%)		10 (40)		6 (20)	57 (17)
Triple therapy, n (%)		2 (8)		1 (3)	0 (0)
Treatments stopped at PVOD diagnosis or during follow-up	25		10
Endothelin receptor antagonist, n (%)		1 (4)		6 (60)	40 (12)
Phosphodiesterase 5 inhibitor, n (%)		3 (12)		6 (60)	13 (4)
Prostacyclin, n (%)		0 (0)		3 (30)	3 (1)
Maximal tolerated strategy during follow-up	25		15
None, n (%)		0 (0)		7 (47)	57 (17)
Single therapy, n (%)		6 (24)		5 (33)	136 (42)
Dual therapy, n (%)		9 (36)		3 (20)	134 (41)
Triple therapy, n (%)		10 (40)		0 (0)	0 (0)
Corticosteroids/immunosuppressants	25		6		–
Already ongoing before PVOD diagnosis, n (%)		7 (28)		3 (50)
Escalation following PVOD diagnosis or during follow-up, n (%)		1 (4)		4 (67)
Diuretics
Already ongoing before PVOD diagnosis, n (%)	25	16 (64)	7	3 (43)	NA
Escalation following PVOD diagnosis or during follow-up, n (%)	25	17 (68)	36	21 (58)	NA
PE
Episodes of PE during follow-up
Patients with ≥1 PE episode, n (%)	25	16 (64)	64	29 (45)	NA
Time between PAH diagnosis and first PE episode (months), median (range)	25	22 (5.2–42)	31	1.5 (0.8–4)	NA
Number of PE episodes, median (range)	25	1 (0–2)	17	1 (0–1)	NA
Treatment de-escalation due to PE, n (%)	25	4 (16)	43	17 (40)	NA
Status at last follow-up
Follow-up duration (months), median (range)	25	47 (20–70)	8	1 (0.5–3.5)	21 (8–38)
Lung transplantation, n (%)	25	0 (0)	35	3 (8.6)	41 (13)
Death, n (%)	25	16 (64)	35	20 (57)	167 (51)

6MWD: 6-min walk distance; ACA: anti-centromere antibodies; ATA: anti-topoisomerase antibodies; BNP: brain natriuretic peptide; CI: cardiac index; CO: cardiac output; D_LCO: diffusing capacity of the lung for carbon monoxide; FEV₁: forced expiratory volume in 1 s; FVC: forced vital capacity; HRCT: high resolution computed tomography; ILD: interstitial lung disease; mPAP: mean pulmonary arterial pressure; NA: not applicable/available; NT-proBNP: N-terminal pro-brain natriuretic peptide; P_aO2: arterial oxygen tension; P_aCO2: arterial carbon dioxide tension; PAWP: pulmonary arterial wedge pressure; PE: pulmonary oedema; PVR: pulmonary vascular resistance; RAP: right atrial pressure; SVI: stroke volume index; S_vO2: mixed venous oxygen saturation; TLC: total lung capacity; V_A: alveolar volume; WHO: World Health Organization; WU: Wood unit.

Epidemiology

The exact prevalence of PVOD-like involvement in SSc-PAH remains uncertain. Venular and capillary remodelling are frequent histological findings in the lungs of these patients, occurring in 50–100% of post-transplant and post mortem assessments in pathological series [19, 26, 31]. Conversely, the classic high-resolution computed tomography (HRCT) triad of PVOD signs (centrilobular ground-glass opacities, mediastinal lymphadenopathy and interlobular septal thickening) has been reported at variable frequencies, ranging from 7% to 61.5% of patients with SSc-PAH [18, 21, 24, 25]. Interestingly, the lowest prevalence estimates are found in studies where radiological assessment was performed at the time of PAH diagnosis in untreated patients. This suggests that PVOD-like features in SSc may be absent at PAH presentation but can appear later during follow-up, reflecting a progression of the lung vasculopathy over time and/or the effect of therapeutic escalation that unmasks a sub-clinical venular involvement.

Histopathology

The pathological hallmark of PVOD/PCH is a significant remodelling of the lung vasculature predominantly located to the venous and capillary circulation. Typical changes include intimal fibrosis and medial hypertrophy of the septal veins and pre-septal venules, leading to their progressive narrowing and occlusion, as well as patchy proliferation and congestion of the capillaries within the alveolar wall. Arteriolar lesions reminiscent of idiopathic/heritable PAH, with the notable exception of plexiform lesions, are also frequently encountered, although usually to a lesser extent than in PAH patients [32].

Pathological descriptions of SSc-PAH lungs with venous/capillary lesions were provided in several case reports [11–14, 17, 20, 23, 27, 28] and histological series [19, 26, 31] (figure 1). Interestingly, the number of cardinal PVOD signs on chest HRCT accurately reflects the presence and severity of venous and capillary remodelling on lung histology [18]. Venular lesions appear less severe than in idiopathic PVOD [26], and PCH-like lesions less frequent [26], although cases of isolated capillary involvement have been reported [12, 13]. However, these observations are derived from a specific subgroup (SSc-PH-ILD patients undergoing lung transplantation) and should be interpreted with caution. Arteriolar lesions are almost invariably present and may be more important than in patients without PVOD-like disease [26], which suggests that venular remodelling in SSc-PAH reflects a distinctive phenotype of the SSc pulmonary vasculopathy (figure 2).

FIGURE 1.

Lung parenchyma from a 63-year-old male patient with systemic sclerosis (SSc)–pulmonary arterial hypertension (PAH) with clinical features of venular involvement. This case depicts representative histopathological findings of pulmonary veno-occlusive disease (PVOD)-like alterations in SSc-PAH. Focal venous, capillary and arterial remodelling are illustrated, reminiscent of idiopathic/heritable PVOD/pulmonary capillary haemangiomatosis (PCH) without fulfilling all their classical histopathological criteria. a) Note the rather discrete but well perceivable subpleural interstitial fibrosis with thickening of alveolar septa, but also solid fibrotic areas. Also, note the two bronchiolo-vascular bundles with important wall remodelling of the pulmonary arteries (PAs) (asterisk, one is magnified and depicted in the inset). Medial hyperplasia and massive concentric intimal fibrosis are present. Of importance, the remodelling of the PAs appears to be unrelated to the fibrotic process beneath the pleura. Haematoxylin–eosin (HE) stain, magnification ×100. b) Close-up of the PCH-like remodelling with a patchy area of alveolar septal thickening (left), as compared with normal septa (right). Note the two small arterioles which show rather concentric muscularisation (centre), just at the border of the PCH-like area, this association is typically seen in PVOD. HE stain, magnification ×200. c) PA remodelling (asterisk), patchy thickening of the alveolar septa with capillary congestion and PCH-like appearance (C), and muscular thickening and fibrosis of two branches of a septal vein (V) are all present in this photo. HE stain, magnification ×100. d) Close-up of the PCH-like area and muscular thickening and fibrosis of a septal vein.

FIGURE 2.

Vascular lesions and pathological features of various phenotypes of pulmonary arterial hypertension (PAH) (adapted from Montani et al. [2]). Typical lesions affecting the three compartments of the pulmonary circulation are illustrated. Arterial lesions include intimal fibrosis and medial hypertrophy, with complex plexiform lesions occurring only in idiopathic/heritable (i/h) PAH. Capillary lesions are characterised by exuberant proliferation and congestion of endothelial cells within the alveolar wall (capillary haemangiomatosis). Venular lesions include intimal fibrosis and medial hypertrophy of the septal veins and pre-septal venules. Their relative distribution on the pulmonary vasculature according to disease phenotype is also depicted. In i/h PAH, vascular remodelling is almost exclusively restricted to the arterial compartment. Although venular lesions have been reported in specific subgroups, such as BMPR2 (bone morphogenetic protein receptor type 2) mutation carriers [61], the prevalence and clinical significance of such venular involvement remains uncertain due to the lack of systematic assessment. In i/h pulmonary veno-occlusive disease (PVOD), lesions typically affect all three vascular compartments, with a preferential involvement of the pulmonary venous system. In systemic sclerosis (SSc)-PAH, arteriolar lesions are constantly found, but the degree of venous/capillary involvement can vary between patients and over time within the same patient.

Importantly, it should be noted that these pathological descriptions almost exclusively stem from lung explants and are therefore inherently biased by the lack of systematic assessment performed in every patient. As lung transplantation in the setting of SSc is rare and limited to a highly selected population (typically young patients with few comorbidities and an inadequate response to PAH-approved therapies), it cannot be excluded that these patients represent a subgroup characterised by predominant venular involvement, particularly given their poor therapeutic response. Moreover, the frequent coexistence of ILD represents a potential confounder, as the observed vascular lesions may be a secondary consequence of adjacent fibrotic remodelling, thereby complicating the interpretation of PVOD-like lesions. In addition, available histopathological data are subject to significant sampling variability, potentially limiting their generalisability.

Few pathobiological studies are available to provide a functional insight into the mechanisms underlying the venous vasculopathy in SSc-PAH. Lower expression of interferon-α receptor 1 in total-lung homogenates is noted in patients with a predominantly venous disease, suggesting that different pathways are involved depending on the histological phenotype [33]. Expression patterns of platelet-derived growth factor receptor-β (PDGFR-β) were studied in the pulmonary vascular bed of SSc-PAH, idiopathic PAH and idiopathic/heritable PVOD patients; while differences were observed in the idiopathic PAH group, the immunoreactivity patterns of PDGFR-β in the SSc-PAH patients were similar to those of the idiopathic/heritable PVOD patients [34].

Clinical presentation

The demographic and SSc characteristics of SSc-PAH patients do not significantly differ based on the presence of a venous involvement [18, 21]; most patients are middle-aged females with a limited cutaneous subset and anti-centromere antibodies in both cases. However, the female predominance seems opposite to the usual sex ratio observed in sporadic PVOD. SSc is usually diagnosed before PAH (median 9 (range 1–16) years in our cohort); and venous involvement is often suspected during PAH follow-up (60% in our cohort, 69% in published cases) rather than at its presentation. Most patients did not display significant parenchymal involvement (96% in our cohort, 67% in published cases), although this may simply reflect the challenges of detecting signs of PVOD in the presence of extensive ILD on HRCT.

Functional capacities are altered for the majority of SSc-PAH patients with venous involvement; more than half are classified in World Health Organization (WHO) functional class III or IV (56% in our cohort, 85% in published cases) and most of them walk a total distance below 440 m during the 6-min walk test (80% in our cohort, 100% in published cases). This functional impairment is not statistically different from that observed in SSc patients with isolated PAH in previous studies [18, 21], but appears better in our cohort than in published cases and in idiopathic/heritable PVOD. This may reflect a publication bias towards more severe cases and/or earlier PAH diagnosis in SSc due to the recent implementation of systematic screening programmes.

Altered alveolar–capillary membrane diffusion is almost invariably present, with a large proportion of patients presenting with a markedly low diffusing capacity for carbon monoxide (D_LCO) and hypoxaemia in both our cohort and published cases, in ranges that are significantly different than in SSc patients without venous involvement [18, 21, 24, 25]. Diffusion parameters on pulmonary function tests are significantly associated with the number and nature of PVOD signs on HRCT [24] and the severity of venular remodelling on lung histology [35]. Interestingly, D_LCO values appear slightly higher in our cohort than in idiopathic/heritable PVOD patients, despite a higher proportion of ILD.

Chest imaging

Chest imaging, essentially through HRCT, remains a cornerstone in the assessment of venular involvement in SSc-PAH patients. The classic triad of interlobular septal thickening, centrilobular ground-glass opacities and mediastinal lymphadenopathy reflects venular congestion, interstitial oedema and lymphatic dilation secondary to post-capillary obstruction. When present, this constellation of findings can raise suspicion for venular disease and help differentiate it from purely arteriolar forms of PAH [7].

In SSc-PAH, however, the identification of PVOD-like changes on HRCT faces distinctive challenges (figures 3 and 4). First, parenchymal changes due to ILD, a frequent co-occurrence in these patients, can obscure or mimic subtle signs of venular involvement [18, 27] (figure 3). Second, PVOD-like features on HRCT are undistinguishable from pulmonary oedema due to left-heart conditions [7], which are common comorbidities in SSc [36]. Third, these radiographic signs are sensitive to intravascular volume status and may be absent in hypovolaemic patients (figure 4). As such, a high clinical suspicion of venular involvement in SSc-PAH patients without evocative signs on HRCT should prompt retesting at a later time-point, especially under treatment. Finally, pulmonary arteriolar remodelling in pre-capillary PH (including idiopathic and heritable PAH) may be associated with subtle parenchymal changes on CT, such as centrilobular micronodules, peripheral neovascularisation or lobular areas of ground-glass attenuation [37], which probably reflect increased pressure and flow within the arteriolar circulation. Although usually rare and mild, these changes may be over-attributed to venular involvement in the context of SSc-PAH.

FIGURE 3.

Representative chest computed tomography (CT) images of systemic sclerosis (SSc)–pulmonary arterial hypertension (PAH) patients with venular involvement, acquired at ultra-high resolution using photon-counting detector CT technology. a–c) Chest CT images of a 59-year-old male systemic sclerosis (SSc)-PAH patient showing subpleural parenchymal opacities, predominant in the lower lobes (green arrowheads in a), possibly compatible with SSc–interstitial lung disease (ILD) (nonspecific interstitial pneumonia-like pattern), with no evidence of venular involvement. Of note, lung perfusion assessed by spectral imaging was normal (c). d–f) Chest CT images of the same patient 2 months later, after initiation of PAH vasodilators, showing disseminated septal lines (yellow arrows in d), mediastinal lymphadenopathies (yellow asterisk in e) and bilateral pleural effusions (red asterisks in d), compatible with venular involvement associated with SSc-PAH. Additional ground-glass opacities (yellow arrowhead in d) appeared on top of those previously documented (green arrowheads in d) and were attributed to venular overload rather than to the underlying ILD. Lung perfusion imaging revealed patchy perfusion defects disseminated throughout both lungs (blue arrowheads in f).

FIGURE 4.

Representative chest computed tomography (CT) images of systemic sclerosis (SSc)–pulmonary arterial hypertension (PAH) patients with venular involvement, acquired at ultra-high resolution using photon-counting detector–CT technology. a–c) Chest CT images of a 60-year-old female SSc-PAH patient showing typical features of venular involvement: diffuse ground-glass opacities (yellow arrowheads in a), disseminated septal lines (yellow arrows in b) and mediastinal lymphadenopathies (yellow asterisk in a). d–f) Chest CT images of the same patient 2 years later, showing a complete resolution of the venular anomalies after optimal diuretic treatment. This showcases the influence of intravascular volume status on the detection of pulmonary veno-occlusive disease signs on chest CT.

Novel imaging technologies hold promises for improving sensitivity and specificity in the detection of venular involvement in SSc-PAH [38]. Among them, dual-energy computed tomography (DECT) enables the assessment of regional lung perfusion through the creation of iodine maps. When evaluated with DECT, SSc-PAH patients with venous involvement almost always display patchy perfusion defects, occasionally associated with pulmonary embolism-type perfusion defects, disseminated throughout the lungs [39]. Interestingly, perfusion anomalies seem less frequently encountered in SSc-PAH in the absence of PVOD-like features [39]. Other techniques, such as photon-counting detector computed tomography, which allows the study of lung microvasculopathy at ultra-high resolution and the assessment of lung perfusion through spectral acquisition, can further revolutionise the field of SSc-PAH imaging [38] (figure 3).

Pulmonary haemodynamics

Although pulmonary vascular remodelling extends to the post-capillary venules, SSc-PAH with venous involvement displays a typical pre-capillary PH profile on right heart catheterisation, with normal pulmonary arterial wedge pressure (PAWP) [6]. Patients usually present with severe haemodynamics, with mean pulmonary arterial pressure (mPAP) values frequently above 35 mmHg (68% in our cohort, 98% in published cases), pulmonary vascular resistance (PVR) values above 5 Wood units (WU) (64% in our cohort, 98% in published cases) and cardiac index (CI) values below 2.5 L·min⁻¹·m⁻² (48% in our cohort, 78% in published cases). The number of PVOD signs on HRCT significantly correlates with mPAP values [24]. Except for one notable exception (that included patients with extensive ILD and group 3 PH) [26], studies have consistently observed significantly higher mPAP and PVR values compared to SSc-PAH without venous involvement; while differences in cardiac output and index did not always reach statistical significance [18, 21, 24, 25]. Haemodynamic alterations also appear milder than in idiopathic/heritable PVOD patients.

Importantly, as both can experience pulmonary oedema under pulmonary vasodilators, SSc-PAH with venular involvement can be confused with combined pre- and post-capillary SSc-PH, especially in case of normal PAWP values. Indeed, due to the systemic nature of the disease and its occurrence late in life, SSc patients are at risk for LHD and heart failure with preserved ejection fraction (HFpEF), either through common cardiovascular comorbidities or SSc-specific heart involvement [36, 40]. A thorough work-up should be performed to identify signs of structural LHD, such as left atrial dilation and abnormal mitral flow on echocardiography, or myocardial fibrosis on cardiac magnetic resonance imaging. Patients with borderline PAWP values (13–15 mmHg) and/or large v-waves on PAWP tracing should undergo provocative testing (fluid challenge, exercise haemodynamics) to uncover occult post-capillary PH [41] (figure 5).

FIGURE 5.

Pulmonary arterial wedge pressure (PAWP) tracings a) before and b) after fluid challenge in a 77-year-old female with systemic sclerosis–pulmonary arterial hypertension (SSc-PAH) and heart failure with preserved ejection fraction (HFpEF). a) PAWP tracing at basal state. PAWP was measured at 13 mmHg in a normovolaemic patient (right atrial pressure 7 mmHg), with mean pulmonary arterial pressure at 29 mmHg, cardiac index at 2.61 L·min⁻¹·m⁻² and pulmonary vascular resistance at 3.41 Wood units. Note the large v-waves, suggestive of a left-heart disease despite a PAWP value below the 15-mmHg threshold. b) PAWP tracing after fluid challenge (500 mL of normal saline over 5 min). PAWP rose to 26 mmHg, confirming combined pre- and post-capillary pulmonary hypertension in this patient with SSc-PAH and HFpEF. Note that, due to technical limitations and signal artefacts, the positioning of a-waves may appear suboptimal (especially at baseline). This does not affect the validity of the PAWP measurement (which was carefully determined at end-diastole after manual correction to appropriately identify the a-wave and averaged over multiple cardiac cycles), nor the identification of v-waves.

Treatment modalities

The best treatment modalities of SSc-PAH patients with venous/capillary involvement remain elusive. Indeed, PAH clinical trials usually consider significant venous involvement as an exclusion criterion and do not report on the PVOD-like status of their SSc-PAH subgroup [42]. As such, it is unclear whether their results can be safely generalised to patients with an established venous component.

Although no robust data currently exist to support this strategy, pulmonary vasodilators are commonly used in SSc-PAH with venous involvement. Most patients are actually already treated at the time of PVOD-like diagnosis, either on single (40% in our cohort, 42% in published cases), double (52% in our cohort, 31% in published cases) or triple (8% in our cohort, 4% in published cases) therapy. The confirmation of venous involvement usually leads to treatment de-escalation, as attested by the increase in patients on single therapy in our cohort (52% versus 40%, respectively) or left untreated in published cases (40% versus 23%, respectively). The maximal tolerated combination achieved during follow-up varies according to sources; a majority of patients (76%) reached double or triple therapy in our cohort while this remained rare (20%) in published cases, which may again reflect a publication bias towards more severe cases. Interestingly, idiopathic/heritable PVOD patients also seldom reached dual (41%) or triple (0%) therapy, which might suggest a better tolerance of these drugs in case of SSc. As such, a sequential treatment, consisting of a first-line single therapy with consecutive escalation to dual therapy in case of adequate tolerance, may be an interesting approach in this setting [42].

The relevance of anti-inflammatory or immunosuppressive drugs in these patients is also insufficiently documented. Previous studies have generally found corticosteroids, conventional immunosuppressants and biologics to have no major benefit on SSc-PAH [42], although whether a differential effect could exist between patients with and without venous involvement has never been investigated. We could only identify four cases [12, 13, 17, 27] in which the diagnosis of PVOD-like involvement led to a change in immunosuppressive strategy; corticosteroid doses were increased in all of them and cyclophosphamide added in one patient [12]. This approach always failed to induce any significant clinical improvement [12, 13, 17, 27].

As expected, pulmonary oedema is a frequent occurrence, both in our cohort (64%) and previously reported cases (45%). Registry data estimate a prevalence ranging from 29% to 50% in SSc-PAH patients with radiological suspicion of venous involvement and virtually null in patients without [18, 21]. Pulmonary oedema usually developed after starting or increasing PAH-targeted therapy, and limited therapeutic escalation in 16% of cases in our cohort and 40% cases from the literature. This discrepancy may reflect our local practice, which favours continuing specific PAH drugs as long as signs of volume overload can be controlled by diuretics, rather than systematically de-escalating the vasodilator therapy. Consistently, diuretic escalation is frequently needed at PVOD-like diagnosis and follow-up (68% in our cohort, 58% in published cases), but vasoactive drugs are rarely required.

Outcome

The prognosis of SSc-PAH patients with venous involvement remains unfavourable, with 5-year survival estimated at 48% in our cohort and 43% in published cases. The presence of classical signs of PVOD on chest HRCT has consistently been associated with increased mortality, with a quantitative effect of the number of signs present [24]. Survival is significantly worse than in SSc-PAH patients without venous involvement [18, 21, 24, 25], although not independently from other usual prognostic factors (age, PVR, D_LCO, 6-min walk distance, N-terminal prohormone of brain natriuretic peptide, WHO functional class) [21, 25].

Signs of venous/capillary involvement have also been described in PAH associated with other CTDs

Features of venous and/or capillary involvement have also been described in PAH associated with CTDs other than SSc, although these situations appear less frequent. Overall, we identified 15 case reports (table 2) representing a total of 16 patients, with four cases of SLE [19, 43–45], four cases of RA [19, 46–48], two cases of MCTD [49, 50], two cases of SjS [51, 52], one case of ASS [53] and three cases of undifferentiated CTD (UCTD) [54–56]. Interestingly, some patients had alternative explanations to their venous involvement: features of SSc were present in one case of MCTD (with anti-centromere antibodies) [50] and one case of UCTD (with Raynaud's phenomenon and telangiectasias) [55]; heritable PVOD was suspected in one patient with SjS and a biallelic mutation of the EIF2AK4 gene [51]; and three patients with RA, a disease which is no longer considered to drive significant pulmonary vascular remodelling [57], were ANA-positive (including one with anti-centromere antibodies) [46–48], suggesting an undiagnosed overlap with another CTD.

TABLE 2.

Characteristics of nonsystemic sclerosis connective tissue disease (CTD)–pulmonary arterial hypertension (PAH) patients with features of venous/capillary involvement: results from a literature review

	Patient ID and reference
	#1 [46]	#2 [54]	#3 [47]	#4 [51]	#5 [52]	#6 [49]	#7 [43]	#8 [19]	#9 [19]	#10 [50]	#11 [44]	#12 [45]	#13 [53]	#14 [56]	#15 [55]	#16 [48]
Demographics
Gender	M	F	F	F	F	F	F	F	F	F	F	F	M	F	F	F
Age at PVOD diagnosis (years)	59	46	36	28	21	47	18	46	58	64	28	20	55	41	67	69
CTD diagnosis
CTD type	RA	UCTD	RA	SjS	SjS	MCTD	SLE	SLE	RA	MCTD	SLE	SLE	ASS	UCTD	UCTD	RA
ANA	Pos	Pos	Pos	Pos		Pos	Pos			Pos	Neg	Pos	Pos	Pos	Pos	Pos
Titre	1:160	1:80	1:640				1:160			1:1280			1:160	1:640	1:64 1:4096	1:5120
Pattern			Homogeneous				Diffuse					Diffuse		Speckled	Homogeneous speckled	Centromere
Anti-DNA Abs	Neg				Pos		Pos			Pos		Pos
Anti-ENA Abs	Neg			SSA+	RNP+ SSA+ SSB+ Sm−	U1− RNP+ SSA− SSB− Sm−	RNP+			U1-RNP+ SSA− Scl-70− Cm+			Jo1+
APL				Neg	LAC− ACL−		LAC+ ACL+				LAC+			LAC− ACL−
RF		Neg	Pos	Neg		Pos						Pos		Pos	Neg	Pos
Anti-CCP Abs	Pos					Neg
Complement levels							Norm				Low	Low		Norm		Low
Interstitial lung disease (status at PVOD diagnosis)	Yes	Yes	No	No	No	No	No			No	No		Yes	No
PVOD diagnosis
Disease duration
Since CTD diagnosis (years)		0	16	0	8	20	2			0			0.25		0	25
Since PAH diagnosis (years)	6	0	0	0	0	0	0			0	0	0	0.25	2	0	0
Chest HRCT suspicion
Centrilobular ground-glass opacities	No			Yes	Yes	No	Yes			Yes			Yes
Mediastinal lymphadenopathy	No			No	Yes	No	No			Yes			Yes
Interlobular septal thickening	No			No	Yes	No	Yes			Yes			Yes
Histological confirmation	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	No	Yes	Yes	Yes	Yes	Yes	Yes
Characteristics at PVOD presentation
CTD activity (aside from PAH)	Nonactive	Active	Uncertain	Uncertain	Active	Uncertain	Active			Active	Active	Uncertain	Active	Uncertain	Uncertain	Uncertain
WHO functional class	III	III	II	IV	III	III	IV			II	III	IV	III	III	II	IV
6-min walk distance (m)					443					310
Cardiac biomarkers
NT-proBNP (ng·L−1)				2482
BNP (ng·L−1)	843.3									861
Arterial blood gas
PaO2 (mmHg)	61.9	52	58		65.6			57.7	60.7		54	24	53		39	72
PaCO2 (mmHg)	33.7	32	18		38.4						27	24			37	43
Pulmonary function tests
FVC (% pred)	60	79	30.5													75
TLC (% pred)		58	44.4					60	79
FEV1 (% pred)	68	45	31					102	77							58
DLCO (% pred)	8.8	35					32			18		54				38
Right-heart catheterisation
mPAP (mmHg)	52		45	62	41		33	49	55	40			65		33	51
PAWP (mmHg)	2		11	10	4		11			11			8		3	10
RAP (mmHg)	5				5			15	12				10		2
PVR (WU)	12.1			25	16.8					6.17			21		15.3	21.5
CO (L·min−1)	4.14			2.05	2.20					4.7			2.7		2.8	1.9
CI (L·min−1·m−2)	2.46				1.61			2.65	1.94				1.6			1.0
SvO2 (%)					59								55
Treatment strategy
Conventional PAH therapies
Before PVOD diagnosis	None	None	None	None	None	None	None			None	None		ERA+PDE5i+inh then i.v. PC	None	None	None
After PVOD diagnosis	ERA+PDE5i	None	None	ERA+PDE5i	None	None	None	i.v. PC	i.v. PC	ERA+PDE5i	i.v. PC		ERA+PDE5i	None	None	None
Corticosteroids and immunosuppressants
Before PVOD diagnosis	None	CS	None	None	None	None	None			None		None		None	None	None
After PVOD diagnosis	None	CS+AZA	CS+CYC	HCQ	CS+AZA	CS	CS			CS+CYC		CS		CS+HCQ	CS	CS
PE
Episodes of PE during follow-up	No	No	Yes	No	No	No	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Time between PAH diagnosis and first PE episode (months)			0				5			1	0.25		0.25		0.5	0.25
Treatment de-escalation due to PE	No	No	No	No	No	No	No			Yes	No		Yes	No	No	No
Status at last follow-up
Follow-up duration (months)	96	24	5	36	7	0.25	5			4	1	72	0.25	36	1
Lung transplantation	No	No	No	No	No	No	Yes	Yes	Yes	No	No	No	No	No	No	No
Death	Yes	No	Yes	No	No	Yes	No	No	No	No	Yes	Yes	Yes	Yes	No	No

ACL: anti-cardiolipin; ANA: anti-nuclear antibodies; anti-CCP Abs: anti-cyclic citrullinated peptide antibodies; anti-DNA Abs: anti-deoxyribonucleic acid antibodies; anti-ENA Abs: anti-extractable nuclear antigen antibodies; APL: anti-phospholipid; ASS: anti-synthetase syndrome; AZA: azathioprine; BNP: brain natriuretic peptide; CI: cardiac index; Cm: centromere; CO: cardiac output; CS: corticosteroids; CYC: cyclophosphamide; D_LCO: diffusing capacity for carbon monoxide; ERA: endothelin receptor antagonist; F: female; FEV₁: forced expiratory volume in 1 s; FVC: forced vital capacity; HCQ: hydroxychloroquine; HRCT: high resolution computed tomography; inh: inhaled; LAC: lupus anticoagulant; M: male; MCTD: mixed connective tissue disease; mPAP: mean pulmonary arterial pressure; Norm: normal; Neg: negative; NT-proBNP: N-terminal prohormone of brain natriuretic peptide; P_aCO2: arterial carbon dioxide tension; P_aO2: arterial oxygen tension; PC: prostacyclin analogues; PAWP: pulmonary arterial wedge pressure; PDE5i: phosphodiesterase 5 inhibitor; PE: pulmonary oedema; Pos: positive; PVOD: pulmonary veno-occlusive disease; PVR: pulmonary vascular resistance; RA: rheumatoid arthritis; RAP: right atrial pressure; RF: rheumatoid factor; RNP: ribonucleoprotein; SjS: Sjögren syndrome; SLE: systemic lupus erythematosus; Sm: Smith; SSA: Sjögren syndrome related antigen A; SSB: Sjögren syndrome related antigen B; S_vO2: mixed venous oxygen saturation; TLC: total lung capacity; U1-RNP: U1-ribonucleoprotein; UCTD: undifferentiated connective tissue disease; WHO: World Health Organization.

CTD was diagnosed before PAH presentation in all patients. Signs of venous/capillary involvement were often present at PAH diagnosis (11/14 cases), but could also appear during follow-up (3/14 cases), especially after starting pulmonary vasodilators [43–56]. The diagnosis of PVOD-like involvement was frequently suspected based on clinical presentation (D_LCO <60% in 6/6 cases and/or hypoxaemia in 7/11 cases) and/or compatible chest HRCT pattern (≥2 classical signs noted in 3/7 cases). Less commonly, pulmonary venous/capillary remodelling was an unsuspected pathological finding on a lung biopsy or autopsy [48, 50]. Overall, histological confirmation was obtained in 15/16 cases.

The underlying CTD could be considered as clinically active in a large proportion of patients (6/14 cases). CTD-associated ILD was noted in 3/11 cases, with significant parenchymal extension noted in two of them [46, 54]. Importantly, PH was always considered as disproportionate in these patients and could not be solely explained by the lung infiltration. Similarly, acute pulmonary embolism was documented in one patient with progressing PH, but with no sign of chronic thromboembolic pulmonary disease [49].

Although this may reflect a publication bias, most patients were considered as severe at time of PVOD-like diagnosis, with WHO functional class III/IV in 11/14 cases, mPAP >35 mmHg in 9/11 cases, PVR >5 WU in 7/7 cases and CI <2.5 L·min⁻¹·m⁻² in 7/8 cases. Profound hypoxaemia requiring oxygen supplementation occurred in at least three cases [46, 48, 53], including one on a background of extensive RA-ILD [46].

Treatment strategies differed considerably between patients and mostly consisted in various combinations of PAH drugs and/or immunosuppressants, with very variable outcomes. At the time of PVOD-like presentation, most patients were treatment-naïve, both in terms of corticosteroids/immunosuppressants (10/12 cases) and of PAH therapies (14/15 cases). The identification of venous/capillary involvement led to a change in the immunosuppressive strategy in 11/12 cases, with a positive outcome noted only in one patient [52]. Drugs approved for PAH were started or modified in about half the patients (8/15 cases) with the remaining cases left untreated.

Importantly, the majority of patients (11/16 cases) developed pulmonary oedema during the course of their disease, most commonly following the initiation of PAH-specific therapy. Precise timing of oedema onset was often unavailable due to incomplete data in the reports. After a median follow-up of 5 (1–36) months, lung transplantation occurred in 3/16 patients, and death in 7/16 cases.

Conclusion

Pulmonary venular involvement has long been recognised in CTD-PAH and mentioned in pathological reports as early as the 1890s [58]. However, its clinical significance and therapeutic implications remain incompletely understood, leaving several important questions unresolved.

A prevailing paradigm is that PVOD-like changes are frequent in CTD-PAH, especially SSc, as supported by historical histopathological series [18]. Yet, this contrasts with recent paradoxical findings of low prevalence on HRCT screening [25], suggesting that venular lesions may be a dynamic process that progresses during the course of the disease. This discordance underscores the need for systematic HRCT assessment repeated over time, to properly determine its prevalence at CTD-PAH diagnosis, its incidence during follow-up and its prognostic value. Modern chest imaging technologies will also probably facilitate earlier diagnosis [38]. Additionally, suspicion of venular involvement should not rely solely on imaging findings and integrate broader clinical findings (such as marked diffusion impairment, rest or exercise-induced hypoxaemia, pulmonary oedema under PAH vasodilators) [2], while acknowledging the potential interference of comorbidities (ILD, HFpEF) in their occurrence.

The reclassification of PVOD from a distinct group (1') to a subset of group 1 PAH (1.5) and the removal of CTDs as recognised causes of PVOD reflect another paradigm shift [1, 59]. A conceptual dichotomy has emerged between “true” PVOD (either heritable, toxic or idiopathic) and “PVOD-like” disease observed during CTDs [6–8], relying on putative clinical and pathophysiological differences. However, to our knowledge, a direct comparison of these two PVOD phenotypes is currently lacking, with the exception of one retrospective study limited by insufficient patient characterisation [8]. In the absence of dedicated pathobiological data, it remains unclear how CTD-associated PVOD-like disease overlaps with, or segregates from, both CTD-PAH without venular involvement and idiopathic/heritable PVOD.

Importantly, this new framework does not yet translate in terms of therapeutic strategies. In idiopathic/heritable PVOD, PAH therapies have no demonstrated effect on survival and are frequently poorly tolerated [10], with a significant risk of pulmonary oedema. By analogy, current practice tends to apply this paradigm of caution to PVOD-like disease associated with CTDs [42]. However, unlike idiopathic/heritable PVOD, “CTD-PVOD” is almost invariably associated with a significant arteriolar component, which paradoxically is likely to benefit from standard therapies. This highlights the crucial need to investigate specific treatment strategies dedicated to these patients. Novel PAH therapies that act on alternative noncanonical pathways, such as the activin signalling inhibitor sotatercept, can improve vascular remodelling without inducing pulmonary vasodilation and as such could be a promising therapeutic option in these patients [60].

Overall, the field of venous involvement in CTD-PAH remains filled with evolving paradigms, unresolved paradoxes and, most importantly, unanswered clinical questions. Addressing these critical gaps is mandatory to refine patient stratification, define adapted treatment strategies and improve its currently dismal prognosis.

Points for clinical practice

• In CTD-PAH, especially SSc-PAH, venous/capillary involvement should be suspected in patients with markedly reduced D_LCO, hypoxaemia, suggestive signs on HRCT and/or occurrence of pulmonary oedema under PAH vasodilators.

• HRCT assessment can be challenging in these patients as SSc-ILD may mimic or obscure PVOD-like features, and these may also be seen in pulmonary oedema due to LHD.

• Due to frequent cardiovascular comorbidities, differential diagnosis with occult HFpEF should always be considered, especially in patients with borderline PAWP values.

• PAH-approved therapies may be considered but should be introduced and escalated cautiously because of a high risk of pulmonary oedema.

Questions for future research

• What is the actual prevalence and incidence of venous/capillary involvement in CTD-PAH?

• What are the pathogenic similarities and differences between CTD-PAH with and without venular involvement and idiopathic/heritable PVOD?

• Can these patients safely benefit from conventional PAH vasodilators? Is there a place for newer nonvasodilating therapies such as activin-signalling inhibitors?

Supplementary material

Please note: supplementary material is not edited by the Editorial Office, and is uploaded as it has been supplied by the author.

Supplementary methods

ERR-0159-2026.SUPPLEMENT

Supplementary table

ERR-0159-2026.SUPPLEMENT