Prognostic value of the VE/VCO2 slope calculated from different time intervals in patients with suspected heart failure

Background Maximal exercises testing, whether involving cycling- or walking-based protocols, are often not well tolerated in patients with chronic heart failure (CHF). The peak oxygen consumption and the slope of the relation between ventilation (VE) and carbon dioxide production (VCO2) are independent predictors of outcome and help risk stratification. The prognostic usefulness of submaximal exercise testing is not clear. The aim of the present study was to assess the prognostic value of the VE/VCO2 slope when derived from data acquired from submaximal exercise. Methods 394 patients referred with breathlessness and suspected heart failure (74% male) (mean ± S.D.) age 60 ± 12 years; BMI 27 ± 5 performed a CPET to determine peak VO2 and the VE/VCO2 slope. The VE/VCO2 slope was calculated using least squares regression from data acquired from the first 25% of exercise (mean VE/VCO2 slope ± SD; 30.6 ± 5.7), 50% (29.6 ± 6.9), below the ventilatory compensation point (sub-VCP) (29.9 ± 6.8), and all data points (full slope) (32.1 ± 7.8). For each measure, patients were divided into quartiles and Kaplan–Meier curves were constructed to determine probability of death after 24 months. The prognostic value of the different classifications was assessed using the χ2 statistic from the Mantel–Cox log-rank test. Results During a mean follow-up period of 41 ± 19 months, 48 patients died. For the VE/VCO2 slope, the log-rank statistic was greatest for the full slope (χ2 = 53.7; P = 0.0001), followed by the sub-VCP (χ2 = 45.5; P = 0.0001), 50% (χ2 = 41.9; P = 0.0001), and 25% (χ2 = 26.0; P = 0.01). The pair-wise log-rank statistic between the fourth and third quartiles was also greatest using the full slope (χ2 = 25.4; P = 0.001) followed in order by the sub-VCP (χ2 = 20.1; P = 0.001), 50% (χ2 = 19.7; P = 0.001), and 25% (χ2 = 14.2; P = 0.05). Using the stratified slope measurements entered into a Cox regression analysis using a forward LR stepwise elimination procedure; only the full slope remained significant (P = 0.0001). Conclusions The VE/VCO2 slope should be calculated from all data points to optimise prognostic sensitivity. Data acquired from the first 50% of exercise and below the VCP provide adequate prognostic surrogates in patients who may not be able to perform maximal exercise testing (i.e. in patients with a respiratory exchange ratio < 1.10).