PREDICTION OF MAXIMAL OXYGEN CONSUMPTION FROM SUBMAXIMAL GAS EXCHANGE DATA IN PATIENTS WITH HEART FAILURE

Abstract

Maximal oxygen consumption (VO2) is a strong predictor of mortality in patients with heart failure and is used to stratify patients for cardiac transplantation. However, many patients with heart failure do not achieve maximal exercise, and the endpoint of exercise frequently varies among those who achieve maximal exercise. PURPOSE: The purpose of this study was to determine if submaximal exercise gas exchange data can predict maximal VO2. METHODS: Fifty consecutive patients' exercise tests were reviewed which had respiratory exchange ratio (RER) greater than 1.10. Maximal exercise data were averaged with the 30 second time interval and graphed against the RER. Logarithmic, linear, exponential, and power regression lines were fitted to the maximal data. Peak VO2 was predicted with submaximal gas exchange by omitting all data where RER > 0.98. Regression equations were recalculated and predicted VO2 was determined by inserting RER = 1.10 into the new equations. For reference purposes, maximal data were compared to a rolling time average and predicted values were compared to VO2 obtained by regression of the patient's achieved peak RER. Additionally, the effect of variable peak exercise endpoints upon maximal VO2 was evaluated by inserting values of RER = 1.00, 1.10, 1.20, and 1.30 into the equations. RESULTS: Using only submaximal exercise data, the logarithmic regression equation most accurately predicted peak VO2. The reference VO2 was 16.99 ± 3.99 ml/kg/min and the VO2 extrapolated to the actual RER achieved was: 16.97 ± 3.76, 17.78 ± 4.06, 18.92 ± 4.79, 20.46 ± 5.70 ml/kg/min for logarithmic, linear, power, and exponential respectively. Logarithmic regression analysis predicted maximal VO2 the best. The absolute bias between reference VO2 and logarithmic predicted VO2 was 1.59 ± 1.36 ml/kg/min. When the endpoint of peak exercise was unknown, the logarithmic predicted VO2 at a values of RER = 1.0, 1.1, 1.2, and 1.3 were: 14.24 ± 2.97, 16.22 ± 3.50, 18.14 ± 4.06, and 19.81 ± 4.59 ml/kg/min respectively. The mean increment in VO2 between 1.00 and 1.10 was 1.98 ± 0.73, between 1.10 and 1.20 was 1.80 ± 0.67, and 1.20 and 1.30 was 1.65 ± 0.61 ml/kg/min. CONCLUSION: The use of logarithmic regression analysis can be used to predict maximal VO2 in patients with congestive heart failure. Peak VO2 was typically predicted from limited submaximal data to within 1.5 ml/kg/min, while an exercise test that is considered maximal typically had more than twice as much error depending upon the endpoint RER.