‘Functionally, it is obvious that the pulmonary and circulatory apparatus are one unit’1

Cardiopulmonary interactions in chronic obstructive pulmonary disease (COPD) have fascinated physiologists and physicians for decades.1-3 Impaired cardiac function might result from the combined effects of (a) mechanical and (b) gas exchange abnormalities caused by COPD. Chiefly amongst them are: (a) gas trapping due to expiratory flow limitation and lung hyperinflation and (b) an increased physiological dead space (VDphys) due to a preponderance of alveolar units with reduced capillary perfusion relative to preserved alveolar ventilation and hypoxia.3 A central tenet of cardiopulmonary interactions in COPD is that they are more likely to influence the right (rather than the left) chambers as the former are more susceptible to the consequences of increased lung volumes and intrathoracic pressure excursions.4 This is the case because the right ventricle (RV) and pulmonary circulation are inextricably integrated. Moreover, due to RV's intrinsic properties as a high-volume, low-pressure pump with thinner and more compliant walls, it is more susceptible to acute mechanical perturbations of the respiratory system (e.g., exacerbations) compared with the left ventricle (LV). Thus, RV adaptation to added mechanical loading is strongly influenced by the severity of pressure overload, volume overload and alterations in intrinsic contractility, alone or in variable combinations.5

In this context, it has been postulated that RV end-diastolic volume indexed by body surface area (EDVI) may decrease in tandem with high mean intrathoracic pressure swings reducing venous return from the inferior vena cava and increasing right atrial pressure, with both reducing RV preload.6 It may also be negatively influenced by any external compression of pulmonary vessels caused by overinflated lungs and high intra-abdominal pressures on expiration.7 Conversely, RVEDVI may increase if high intrinsic positive end-expiratory pressure and pulmonary vasoconstriction (due to hypoxia or acidosis) increase afterload leading to RV overdistension.8, 9 These abnormalities might be potentiated by: (a) a low RV ejection fraction, (b) hypervolaemia and (c) any decrease in the functioning pulmonary capillary network either by destructive emphysema or compression by regional lung hyperinflation.3 However, if these circulatory derangements are severe enough to reduce RV systolic volume (SV) and subsequently LVSV via series interaction10 and/or cause a leftward septal displacement via mechanical interaction,11 cardiac output may decrease, reducing venous return and, consequently, RVEDVI. Thus, the prevailing effects of worsening lung hyperinflation, greater neuromechanical uncoupling of the respiratory system and higher pulmonary vascular resistance on RV dimensions will ultimately depend on the balance of forces acting to decrease RV preload (↓ RVEDVI) versus those acting to increase its afterload without reducing cardiac output (↑ RVEDVI) (Figure 1).

Schematic representation of the potential haemodynamic effects of gas trapping and lung hyperinflation secondary to worsening expiratory flow limitation during acute exacerbations of chronic obstructive pulmonary disease. For simplification purposes, only the ventricles are depicted. Right ventricle (RV) preload is negatively influenced by a high mean intrathoracic pressure (ITP) which compresses the inferior vena cava and increases right atrium pressure thereby decreasing venous return and, consequently, RV end-diastolic volume (RVEDV). Lung hyperinflation and a high intrinsic positive end-expiratory pressure (PEEPi) may compress the large and small pulmonary arterial vessels and the heart chambers: These effects, in association with alveolar (alv) hypoxia and acidosis, tend to increase RV afterload and, consequently, RVEDV. However, if there is a severe increase in RV afterload, underfilling of the left ventricle (LV) and a leftward shift of the interventricular septum may hamper the ‘ideal’ LV length–tension relationship for the prevailing filling pressure (Frank–Starling mechanism). Low dynamic lung compliance and high PEEPi require a more negative intra-pleural pressure at the start of inspiration (PPL) thereby increasing LV transmural pressure (PTM) and LV afterload. The latter may also increase secondary to high systemic vascular resistance (SVR) induced by sympathetic over-excitation. Those abnormalities may lead to low right and left stroke volume, decreasing venous return in a vicious circle. Thus, RV dimensions may depend on the dynamic balance of forces acting to decrease RV preload (reducing RVEDV) versus those acting to increase its afterload without compromising cardiac output (increasing RVEDV). Reproduced from Oliveira et al.,12 with permission

In a recent publication in Respirology, Leong et al.13 report some thought-provoking results showing a negative independent prognostic value of ↑ RVEDVI, measured by dynamic computed tomography, in patients suffering an exacerbation of moderate-to-severe COPD. Exacerbations are characteristically associated with those aforementioned factors that jointly conspire to decrease RVEDVI, particularly acute-on-chronic hyperinflation.14 Although (total) supine lung volumes on imaging did not differ between patients with or without RVEDVI, it is noteworthy that the ‘static’ and dynamic operating lung volumes were not measured by body plethysmography or derived from serial inspiratory capacity manoeuvres, respectively. Patients showing ↑ RVEDVI had higher carbon dioxide venous tensions (PvCO2), a finding usually associated with greater mechanical constraints on tidal volume expansion in the setting of high physiological dead space.15 Thus, it could be speculated that ↑ RVEDVI reflected the dominant upstream consequences of increased RV afterload with high pulmonary arterial pressures leading to altered RV distension.5 Although lack of pressure–volume metrics, including pulmonary vascular resistance, precludes definitive mechanistic assertions, these are the abnormalities expected in an RV under acute strain.8 In fact, lower RV ejection fraction in the↑ RVEDVI group suggests some degree of incipient right systolic dysfunction.

There is, however, another critical piece of information that should be taken into consideration to interpret ↑ RVEDVI in patients with poorer prognosis: they had higher cardiac output due to larger stroke volume, compared with those with preserved RVEDVI. As there is no a priori reason to believe that patients with ↑ RVEDVI have lower pulmonary vascular resistance, increased cardiac output likely further increased RV afterload. The hyperdynamic circulatory state could be related to hypervolaemia. Although a low estimated glomerular filtration rate was an exclusion criterion, it is unknown whether patients with ↑ RVEDVI had a higher combined burden of cardiocirculatory-metabolic co-morbidities and/or a more relevant pre-exacerbation history of fluid retention due to some degree of cardio-renal axis impairment or other factors, compared with their counterparts. As mentioned, they were more hypercapnic (but not more hypoxaemic): the systemic responses to high PaCO2 (e.g., increased activity of the renin–angiotensin–aldosterone system, atrial natriuretic peptide and vasopressin) reduce renal blood flow, increasing water and sodium retention.16 Alternatively, or in addition, increased cardiac output may have been a physiological response to higher peripheral O2 requirements due to increased work of breathing in patients facing the worst exacerbations as shown by higher DECAF (Dyspnea, Eosinopenia, Consolidation, Acidemia and Atrial Fibrillation) scores.17, 18 In other words, the hyperdynamic status may simply reflect the increased blood flow requirements of the active respiratory muscles in patients facing more severe mechanical ventilatory constraints. In this scenario, the increased O2 cost of ventilation19 reflects greater neuromechanical dissociation in more hyperinflated patients and/or higher ventilatory demands in those with more extensive ‘wasted’ ventilation. Moreover, higher CO2 production secondary to increased metabolic demands may have contributed to higher PvCO2 in the ↑ RVEDVI group.

A clearer elucidation of the mechanisms of ↑ RVEDVI may have relevant clinical implications. There is growing evidence that cardiac dysfunction does occur during exacerbations and, as seen in the Leong et al.'s study, portends poor prognosis.20 Based on the mechanistic considerations outlined above: (a) avoiding volume overload, (b) controlling hypercapnia, (c) reducing pulmonary vascular resistance and, in particular, (d) decreasing work and O2 cost of breathing might collectively improve cardiopulmonary function in those with ↑ RVEDVI and, hopefully, improve prognosis. Amongst the adjunct strategies to address (b) and (d), non-invasive ventilation, particularly bi-level positive airway pressure,21 stands out. Of note, Leong et al.13 report that patients with ↑ RVEDVI did not receive non-invasive ventilation more frequently than their counterparts. Future prospective studies in patients with COPD during exacerbations, which examine the effects of carefully optimized, non-invasive mechanical ventilation on cardiopulmonary interactions, respiratory symptoms and prognosis, will undoubtedly allow us to further refine individualized management strategies.