A comparison of lactate indices during ramp exercise using modelling techniques and conventional methods.

The aim of this study was to compare the lactate indices provided by single- and double-breakpoint models with lactate thresholds obtained with conventional methods. Arterial samples for the determination of lactate concentrations were drawn from eight participants at rest and every minute during a ramp test (15 W x min(-1)) on a cycle ergometer. Lactate thresholds were determined from a blood lactate concentration equal to 4 mM (LT(4)), from an increase of 1 mM above the resting level (Delta1 mM), and from indirect methods using ventilatory parameters. Other indices were computed from the modelling of the lactate curve using an exponential function (LSI), a polynomial function (Dmax), a semi-log model (SLog), a parabola plus delay model (Mod P), and a two-breakpoint model (Mod M). Mod P and Mod M showed poor agreement with the other methods. LT(4), Dmax, LSI, and respiratory exchange ratio equal to 1 were correlated with each other (0.81 <or= R <or= 0.92) and their mean differences ranged from 2.8 to 15 W, with limits of agreement within the range +/- 24.6 to +/- 42.4 W. These results question the interest in breakpoints models to detect lactate thresholds, knowing that LT(4), LSI, Dmax, and respiratory exchange ratio equal to 1 provide indices that occur at similar power outputs.