BACKGROUND: Heart rate recovery (HRR) during exercise testing is an independent predictor of prognosis. The relative predictive power of computational analysis of HRR as a function of resting and maximum heart rate (HR) compared with direct measurement of the drop in HR has not been determined. HYPOTHESIS: We aimed to improve on the prognostic value of HRR by the use of mathematical representations of HRR kinetics. METHODS: In all, 2,193 patients who underwent exercise testing, coronary angiography, and clinical evaluation were followed up for 10.2 +/- 3.6 years. Mathematical functions were used to model HRR as a function of resting (HR(Rest)), maximum HR (HR(Peak)) and time (t): (a) HRR= HR(Rest) + (HR(peak) - HR(Rest)) X e(-kt) and (b) HRR= HR(Rest) + (HR(peak) - HR(Rest)) e(-kt2) RESULTS: Equation (b) provided the best fit of the recovery HR curve. An abnormal HRR at 2 min was a better predictor of mortality than HRR at 1, 3, or 5 min. At 2 min, HRR also predicted mortality better than computational models of HRR, relating HRR as a function of maximum and resting HRs. After adjusting for univariately significant predictors of mortality, HRR, age, exercise capacity, and maximum HR were chosen in order as the best predictors of mortality. CONCLUSION: Even though the computational models of HRR and the determination of HRR at different time intervals were significant predictors of mortality, the simple discrete measure of HRR at 2 min was the best predictor of mortality. At 2 min, HRR outperformed age, METs, and maximum exercise HR in predicting all-cause mortality.
BACKGROUND: Heart rate recovery (HRR) during exercise testing is an independent predictor of prognosis. The relative predictive power of computational analysis of HRR as a function of resting and maximum heart rate (HR) compared with direct measurement of the drop in HR has not been determined. HYPOTHESIS: We aimed to improve on the prognostic value of HRR by the use of mathematical representations of HRR kinetics. METHODS: In all, 2,193 patients who underwent exercise testing, coronary angiography, and clinical evaluation were followed up for 10.2 +/- 3.6 years. Mathematical functions were used to model HRR as a function of resting (HR(Rest)), maximum HR (HR(Peak)) and time (t): (a) HRR= HR(Rest) + (HR(peak) - HR(Rest)) X e(-kt) and (b) HRR= HR(Rest) + (HR(peak) - HR(Rest)) e(-kt2) RESULTS: Equation (b) provided the best fit of the recovery HR curve. An abnormal HRR at 2 min was a better predictor of mortality than HRR at 1, 3, or 5 min. At 2 min, HRR also predicted mortality better than computational models of HRR, relating HRR as a function of maximum and resting HRs. After adjusting for univariately significant predictors of mortality, HRR, age, exercise capacity, and maximum HR were chosen in order as the best predictors of mortality. CONCLUSION: Even though the computational models of HRR and the determination of HRR at different time intervals were significant predictors of mortality, the simple discrete measure of HRR at 2 min was the best predictor of mortality. At 2 min, HRR outperformed age, METs, and maximum exercise HR in predicting all-cause mortality.
Authors: Luiz Eduardo Ritt; Ricardo Brandão Oliveira; Jonathan Myers; Ross Arena; Mary Ann Peberdy; Daniel Bensimhon; Paul Chase; Daniel Forman; Marco Guazzi Journal: J Cardiopulm Rehabil Prev Date: 2012 May-Jun Impact factor: 2.081
Authors: Olguín C Henríquez; San Martín E Báez; A Von Oetinger; Jamett R Cañas; Campillo R Ramírez Journal: Biol Sport Date: 2013-04-11 Impact factor: 2.806