Non-invasive estimation of stroke volume during exercise from oxygen in heart failure patients

In heart failure, oxygen uptake and cardiac output measurements at peak and during exercise are important in defining heart failure severity and prognosis. Several cardiopulmonary exercise test-derived parameters have been proposed to estimate stroke volume during exercise, including the oxygen pulse (oxygen uptake/heart rate). Data comparing measured stroke volume and the oxygen pulse or stroke volume estimates from the oxygen pulse at different stages of exercise in a sizeable population of healthy individuals and heart failure patients are lacking.

We analysed 1007 subjects, including 500 healthy and 507 heart failure patients, who underwent cardiopulmonary exercise testing with stroke volume determination by the inert gas rebreathing technique. Stroke volume measurements were made at rest, submaximal (∼50% of exercise) and peak exercise. At each stage of exercise, stroke volume estimates were obtained considering measured haemoglobin at rest, predicted exercise-induced haemoconcentration and peripheral oxygen extraction according to heart failure severity.

A strong relationship between oxygen pulse and measured stroke volume was observed in healthy and heart failure subjects at submaximal (R2 = 0.6437 and R2 = 0.6723, respectively), and peak exercise (R2 = 0.6614 and R2 = 0.5662) but not at rest. In healthy and heart failure subjects, agreement between estimated and measured stroke volume was observed at submaximal (-3 ± 37 and -11 ± 72 ml, respectively) and peak exercise (1 ± 31 and 6 ± 29 ml, respectively) but not at rest.

In heart failure patients, stroke volume estimation and oxygen pulse during exercise represent stroke volume, albeit with a relevant individual data dispersion so that both can be used for population studies but cannot be reliably applied to a single subject. Accordingly, whenever needed stroke volume must be measured directly.