Letter by Naeije and Mélot Regarding Article, “Exercise-Induced Pulmonary Arterial Hypertension”
To the Editor:
In a recent issue of Circulation, Tolle et al reported on exercise-induced pulmonary arterial hypertension (PAH) as a well-circumscribed syndrome with clinical cardiopulmonary exercise testing and hemodynamic criteria.1 The authors are to be commended for this article, which is original and clinically relevant. However, we were intrigued by the presentation of their results as log-log plots of mean pulmonary arterial pressure (PAP) as a function of predicted oxygen consumption (V̇o2).
According to the authors, the logPAP-logV̇o2 relationship normally shows an inflection point with increased slope above the anaerobic threshold (“takeoff pattern”), in contrast to a “plateau pattern” identified in exercise-induced PAH and in established PAH. Because V̇o2 and cardiac output are tightly correlated, accounting for “highly concordant” PAP versus cardiac output (Q) and PAP versus V̇o2 log-log plots, the authors explain these patterns by exercise-induced pulmonary vasoconstriction in normal subjects but not in patients with pulmonary vascular disease.
Until now, pulmonary vascular pressure-flow relationships have always been reported to be best described by a linear approximation.2,3 However, when flow was increased by exercise, negative extrapolated pressure intercepts were observed in some studies in patients with heart failure and/or pulmonary vascular diseases.4,5 A disproportionate increase in pulmonary artery pressure at exercise may be explained by an increased pulmonary vascular resistance or an increased filling pressure of the left heart. In the article by Tolle et al, there was no increase in either pulmonary vascular resistance or pulmonary capillary wedge pressure at the highest level of exercise in either normal controls or PAH patients.
Therefore, we do not see any explanation for “takeoff” and “plateau” patterns of PAP versus V̇o2 log-log plots. Could it be an artifact related to an excessive manipulation of data in the presentation of the results? Although this is not clearly stated in the Methods, we assume that the reason to present results as log-log plots was that the distributions of the variables of interest tested negative for normality. The reason for V̇o2 presented as a percentage predicted or V̇o2 instead of Q is more difficult to understand.
It would be interesting to see the results shown as PAP versus Q, which is probably the most appropriate way to describe the functional state of the pulmonary circulation. On these real source data, one would be curious to see some statistical comparison between the slopes of the double linear regression, with power regression analysis to confirm the results obtained by double linear regression (with a slope coefficient <1 for the plateau pattern and >1 for the takeoff pattern).
In our opinion, the article by Tolle et al is convincing about the notion of exercise-induced PAH as a clinically relevant syndrome but less so about the nonlinearity of pulmonary vascular pressure-flow relationships at exercise.
Tolle JJ, Waxman AB, Van Horn TL, Pappagianopoulos PP, Systrom DM. Exercise-induced pulmonary arterial hypertension. Circulation. 2008; 118: 2183–2189.
Reeves JT, Dempsey JA, Grover RF. Pulmonary circulation during exercise. In: Weir EK, Reeves JT, eds. Pulmonary Vascular Physiology and Physiopathology. New York, NY: Dekker; 1989: 107–133.
Naeije R. Pulmonary vascular function. In: Peacock AJ, Rubin LJ, eds. Pulmonary Circulation. Diseases and Their Treatment. 2nd ed. London, UK: Arnold; 2004: 3–13.
Janicki JS, Weber KT, Likoff MJ, Fishman AP. The pressure-flow response of the pulmonary circulation in patients with heart failure and pulmonary vascular disease. Circulation. 1985; 72: 1270–1278.