Double gas transfer factors (DLCO-DLNO) at rest in patients with congenital heart diseases correlates with their ventilatory response during maximal exercise

Abstract

Aim: Exercise capacity is altered in congenital heart diseases (CHD) with potentially impaired pulmonary perfusion adaptation during exercise, such as in single ventricle or in significant pulmonary regurgitation. This study aimed to evaluate the value of double gas transfer factor analysis, at rest in conjunction with postural manoeuvres, to explore the various components of pulmonary gas transfer and its association with exercise capacity.

Methods: A total of 40 subjects (24 CHD, 16 controls) underwent a combined measurement of lung diffusing capacity for carbon monoxide and nitric oxide (DLCO-DLNO) to determine pulmonary membrane diffusion (Dm) and Vcap, in sitting then supine position. CHD patients performed a maximal cardiopulmonary exercise test.

Results: Compared to normal controls, the CHD group's DLNO, DLCO, Vcap, and alveolar volume (AV) at rest, in the sitting position were depressed, whereas the DLCO/AV and DLNO/AV were similar. The magnitude of Dm and Vcap adaptation induced by postural change was similar in both groups, indicating a preserved pulmonary capillary recruitment capacity in CHD. In the CHD group, at rest, for each ml of postural-induced increase in Vcap we observed during exercise a VE/VCO2 slope decrease of 0.46 (95% CI[0.83; 0.098]), indicating a better ventilatory response to exercise.

Conclusion: CHD patients with impaired pulmonary circulation have a reduced Dm and Vcap mainly due to decreased pulmonary volume but maintain a normal capacity to adapt these parameters through a simple recruitment manoeuver. Vcap adaptation evaluated at rest predicts the level of ventilatory efficiency during exercise, which represents a main limiting factor in these CHD patients.

Keywords: Combined measurement of lung diffusion; Congenital heart disease; Exercise capacity; Pulmonary circulation; Pulmonary regurgitation; Single ventricle.

Fig. 1

Independent effects of CHD and supine position on the sub-components of gas transfer factors for NO and CO. Legend: Graph summarising the independent effects of CHD and supine posture (from the centre) on the transfer factors (DLNO, DLCO) and their sub-components (DmCO, θCO, DmNO and Vcap), based on the Roughton-Forster's equation and alveolar volume (AV), represented in separated blocks. Red arrows indicate the significant positive effect (= increase), while blue arrows indicate the significant negative effect (= decrease) on the target variable. The numbers indicate the mean effect size expressed in % of basal value (i.e. x for CHD, y for supine) and could be interpreted as: the target variable would decrease by x % in patients with CHD or would increase by y % when switching from sitting to supine posture. When no significant effect could be established, the connection link was drawn in grey colour and dotted line. Abbreviation: AV, alveolar volume; DLCO, total diffusing capacity of the lung for carbon monoxide; DLNO, total diffusing capacity of the lung for nitric oxide; KCO, ratio of total diffusing capacity of the lung for carbon monoxide over alveolar volume; KNO, ratio of total diffusing capacity of the lung for nitric oxide over alveolar volume; DmCO, diffusing capacity of the membrane for carbon monoxide; DmNO, Diffusing capacity of the membrane for nitric oxide; Vcap, total volume of blood in the lung capillaries exposed to alveolar air.

Fig. 2

Relationship between VO2peak, VE/VCO2 slope and position related change in gas transfer factors. Legend: the graph represents the result of a Bayesian linear correlation analysis, including the marginal effect of 4 linear regression models (left side) to estimate the values of VE/VCO2 slope and VO2peak (Y-axis) as a linear function of the change in DLNO, DLCO and Vcap (X-axis). On the left column, the lines indicate the mean predicted values, while the curvilinear bands represent the confidence interval of this prediction. The right column shows the estimated posterior distribution of the Pearson's r coefficient measuring the linear correlation between the studied variables. The lighter area in each Kernel Density Estimate (KDE) plots represents the highest density interval (95% of Highest Density Interval, HDI) while the 3 labelled vertical lines indicate the 5th, 50th and 95th centiles of the distribution. The red vertical lines and red boxes determine the thresholds for null hypothesis (H0: r = 0, with tolerance interval of ±0.1).