Background: Given the recent development of newer and less-invasive treatments for pulmonary hypertension, and the long wait for lung transplantation, early and correct diagnosis of this condition is increasingly important. The purpose of this study was to determine and improve the accuracy of a non-invasive, cardiopulmonary exercise-testing algorithm for detecting a pulmonary vascular limit to exercise.
Methods: We performed 130 consecutive, incremental cycling-exercise tests for exertional symptoms with pulmonary and radial artery catheters in place. Pulmonary vascular limit was defined as pulmonary vascular resistance at maximum exercise >120 dynes. sec/cm(5) and a peak-exercise systemic oxygen delivery <80% predicted, without a pulmonary mechanical limit or poor effort. We applied a previously reported non-invasive exercise-test-interpretation algorithm to each patient and sequentially manipulated branch point threshold values to maximize accuracy.
Results: The sensitivity of the original non-invasive algorithm for pulmonary vascular limit was 79%, specificity was 75%, and accuracy was 76%. Sensitivity did not change with systematic alteration of branch-point threshold values, but specificity and accuracy improved to 88% and 85%, respectively. Accuracy improved most by modifying the threshold values for percent predicted maximum oxygen uptake and carbon dioxide output ventilatory equivalents at lactate threshold.
Conclusion: Non-invasive cardiopulmonary exercise testing is a useful tool for detecting and excluding a pulmonary vascular limit and for determining whether abnormal pulmonary hemodynamics limit aerobic capacity.