Abstract
Cardiopulmonary exercise tests provide additional data for the noninvasive diagnosis of pulmonary veno-occlusive diseasehttps://bit.ly/3p6Dzc5
To the Editor:
Pulmonary veno-occlusive disease (PVOD) is a rare form of pulmonary hypertension that shares some clinical and haemodynamic features with idiopathic pulmonary arterial hypertension (PAH). However, suspicion of PVOD is crucial, considering that PAH-specific treatment may precipitate life-threatening pulmonary oedema and lung transplant should be considered from diagnosis [1, 2].
The absence of pathogenic variants in the EIF2AK4 gene and the prohibitive risk of performing a lung biopsy in these patients often prevents a definitive diagnosis of PVOD [3–5]. Therefore, PVOD diagnosis frequently relies on the identification of other indicators with a high associated likelihood of PVOD, namely: decreased diffusing capacity of the lung for carbon monoxide (DLCO) and typical high-resolution computed tomography (HRCT) features [5–7]. However, their sensitivity and specificity are far from perfect and additional diagnostic tools are missing. We hypothesise that cardiopulmonary exercise testing (CPET) might reveal characteristic patterns of exercise performance in PVOD patients, strengthening its suspicion and diagnosis.
We studied 23 patients diagnosed with PVOD, referred to a national referral centre for pulmonary hypertension. Among them, 16 patients carried pathogenic biallelic variants in EIF2AK4; two presented the three typical HRCT features; two developed pulmonary oedema on PAH-specific treatment; and, in the remaining three, PVOD diagnosis was only possible by histological examination of lung specimens after transplantation. The control group consisted of 52 consecutive PAH patients on regular follow-up (24 idiopathic PAH and 28 heritable PAH associated with BMPR2).
Patients underwent symptom-limited incremental CPET on a cycle ergometer. Exercise variables were measured at rest, ventilatory threshold and peak exercise. Oxygen saturation was monitored by pulse oximeter. Clinical and haemodynamic data closest to CPET were analysed.
The mean±sd age was 39.8±11.9 years, without differences between PVOD and PAH groups (37.7±11.7 versus 40.7±12; p=0.28). PVOD patients did not show female sex predominance (43.5% versus 76.9%, p=0.08). They presented worse functional class (FC) (FC I: none PVOD versus 13 PAH; FC II: 11 PVOD versus 31 PAH; and FC III: 12 PVOD versus eight PAH; p<0.01) and lower DLCO levels (32.8±7.8 versus 78.7±13.2%; p<0.01) than PAH patients. Most CPET (85.3%) was performed in patients receiving PAH-specific treatment, including 32% on systemic prostacyclins. 11 patients were not receiving any PAH-specific medication: nine (seven PAH and two PVOD) who underwent CPET at diagnosis and two PVOD who did not tolerate it.
PAH patients achieved higher work rates than PVOD patients. There was a significant reduction in predicted peak oxygen uptake (VO2), oxygen pulse and VO2 levels at ventilatory threshold in PVOD patients compared to PAH (table 1).
The main exponents of ventilatory efficiency were increased to a greater extent in PVOD when compared to PAH patients, both minute ventilation (VE)/carbon dioxide production (VCO2) at ventilatory threshold (EqCO2VT) and VE/VCO2 slope. Moreover, minimum oxygen saturation was lower in PVOD patients. End-tidal carbon dioxide pressure (PETCO2) at rest was reduced in both groups, although more profoundly in PVOD (table 1). After stratifying by FC, these differences observed between PVOD and PAH were maintained in patients in FC II, while a non-significant trend was observed in the group of patients in FC III.
Interestingly, PVOD patients had lower pulmonary vascular resistance (PVR), although similar cardiac output (table 1). In fact, the three PVOD patients with mildly elevated PVR (between 3 and 4 Wood units) exhibited profound alterations of CPET parameters: a predicted VO2 below 60%, an EqCO2VT above 51 and a VE/VCO2 slope exceeding 45.
Most relevant CPET variables were strongly associated with the definitive diagnosis of PVOD, especially when adjusted for PVR. These associations persisted after adjusting for FC. Receiver operating characteristic analyses and goodness-of-fit tests were performed to estimate how accurately CPET variables, individually or in combination, identify PVOD patients. As a result, predicted VO2, VE/VCO2 slope, EqCO2VT and PETCO2 showed the highest areas under the curve (AUC). Their combination with PVR increased AUC for all CPET variables. After selecting those models with adequate goodness-of-fit, predicted VO2 showed the highest AUC (0.89, 95% CI 0.81–0.98), closely followed by VE/VCO2 slope (0.85, 95% CI 0.76–0.94). The combination of predicted VO2, VE/VCO2 slope and PVR achieved the greatest discriminative power (AUC 0.974, 95% CI 0.946–1) compared with individual models (p<0.001), constituting the final PVOD diagnostic model.
To our knowledge, this is the largest cohort of CPET in confirmed PVOD patients. CPET data in PVOD exhibited characteristic features distinguishable from those in PAH: 1) a significantly greater ventilatory inefficiency demonstrated by higher VE/VCO2 slope and EqCO2VT values, consistent with data previously published [8]; 2) a more severe functional impairment revealed by lower peak VO2 and earlier ventilatory threshold; 3) CPET parameters may be profoundly altered in PVOD patients with mildly elevated PVR; and 4) predicted VO2 and VE/VCO2 slope in combination with PVR showed predictive power for PVOD diagnosis.
The histopathological mechanisms involved in the exercise response of PAH patients have been widely described. The vascular obliteration and the increased PVR in PAH do not allow the alveolar physiological recruitment; hence, ventilation–perfusion mismatch is further aggravated [9]. These changes lead to tissue hypoxaemia, early ventilatory threshold and reduced peak VO2 [9].
Despite the worse exercise capacity and prognosis associated with PVOD, various studies did not observe haemodynamic differences between PVOD and PAH [6, 10]. Moreover, our research actually shows significantly lower pulmonary pressures in the POVD population. Importantly, previous studies observed an association between CPET parameters and haemodynamic severity in PAH [9, 11]. However, in our study, the decrease in VO2 and the degree of ventilatory inefficiency seemed both disproportionate to the haemodynamic severity in PVOD patients, supporting the assumption that additional factors beyond PVR play a critical role in its pathophysiology.
We expose some possible mechanisms. 1) The vascular remodelling occurring in PVOD, which includes venular intimal fibrosis, venular muscular hyperplasia and capillary proliferation, leads to a lower pulmonary capillary blood volume and alveolar membrane diffusion. These alterations are independent of haemodynamic severity and reflected in reduced DLCO values [3, 12]. 2) The capillary congestion and interstitial oedema decrease gas exchange, further impairing ventilatory efficiency [13]. 3) The significant hypoxaemia throughout exercise could exacerbate myocardial ischaemia and right ventricular failure, and may justify the flattening behaviour of oxygen pulse in PVOD, despite a normal cardiac output at rest [14]. 4) The hypoxaemia and low cardiac output reduce peripheral oxygen delivery and induce an early onset of lactic acidosis, which results in earlier ventilatory threshold and lower VO2 [14].
Currently, HRCT is the key noninvasive test when PVOD is suspected. Two thirds of PVOD patients have at least two HRCT characteristic signs [6, 15]; however, their absence does not exclude PVOD. Interestingly, HRCT signs are less common at the initial stages of the disease, while we found that CPET exhibits alterations even at the early phases. Predicted VO2 and VE/VCO2 slope showed high discriminative power which further improved in combination with PVR. In the current scenario, where the identification of PVOD patients remains challenging, we encourage the incorporation of a CPET model that combines predicted VO2, VE/VCO2 slope and PVR with the currently available noninvasive diagnostic tools in suspected PVOD patients. However, larger studies would be needed to validate this model and establish appropriate cut-off points.
In conclusion, PVOD and PAH show different exercise patterns, where disproportionate changes between haemodynamics and both functional and ventilatory impairment are of particular interest. The combination of predicted VO2, VE/VCO2 slope and PVR showed the highest ability to accurately identify PVOD, positioning CPET as a promising additional tool for the noninvasive diagnosis of PVOD.
Shareable PDF
Supplementary Material
This one-page PDF can be shared freely online.
Shareable PDF ERJ-00115-2021.Shareable
Footnotes
Conflict of Interest: C. Pérez-Olivares has nothing to disclose.
Conflict of Interest: T. Segura de la Cal has nothing to disclose.
Conflict of Interest: Á. Flox Camacho has nothing to disclose.
Conflict of Interest: J. Nuche has nothing to disclose.
Conflict of Interest: J. Tenorio has nothing to disclose.
Conflict of Interest: A. Martínez Meñaca has nothing to disclose.
Conflict of Interest: A. Cruz Utrilla has nothing to disclose.
Conflict of Interest: J. de la Cruz-Bertolo has nothing to disclose.
Conflict of Interest: M. Pérez Nuñez has nothing to disclose.
Conflict of Interest: F. Arribas-Ynsaurriaga has nothing to disclose.
Conflict of Interest: P. Escribano Subías has nothing to disclose.
Support statement: This work was supported through project “Bases Genetico Moleculares de la Medicina de Precisión en la Hipertensión Arterial Pulmonar”, funded by Instituto Carlos III; Ministerio de Economıa y Competitividad (award number: PI 18/01233) to P E-S (the funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript). A.C.U. holds a research training contract “Rio Hortega” (CM20/00164) from the Spanish Ministry of Science and Innovation (Instituto de Salud Carlos III). J.N. is recipient of a predoctoral grant (Jordi Soler Soler) through CIBERCV. Funding information for this article has been deposited with the Crossref Funder Registry.
- Received January 14, 2021.
- Accepted February 2, 2021.
- Copyright ©The authors 2021. For reproduction rights and permissions contact permissions{at}ersnet.org