Author(s): WILL, J., UEBERSCHÄR, O., DUNST, A.K., Institution: HOCHSCHULE MAGDEBURG-STENDAL, Country: GERMANY, Abstract-ID: 2675

INTRODUCTION:
Mader's metabolic model is increasingly applied to individualise endurance training by integrating VO2max, efficiency and vLamax into a mechanistic description of energy metabolism. Model validity, however, critically depends on accurate parameter determination. In elite rowing, vLamax is typically derived from short maximal sprints, yet practical application often yields inconsistent metabolic profiles and implausible threshold predictions. This study evaluated sprint-derived vLamax for metabolic modelling in rowing and investigated an indirect estimation approach based on blood lactate concentration (BLC) thresholds.
METHODS:
Twenty-nine elite rowers (4 females, 25 males) completed a 15-s maximal sprint test, an incremental ramp test to determine relative VO2max (66.60 +/- 7.76 mL/min/kg), and an incremental step test to establish the VO2-power relationship for calibration of rowing economy and y-intercept within Mader's metabolic model. BLC diagnostics during incremental testing determined LT1 via the lactate minimum concept and LT2 (representing the MLSS) using a modified maximal deviation method (Dmax). Within model simulations, LT1 was defined as the point of maximal pyruvate deficit, corresponding to maximal fat oxidation (FatMax). vLamax was obtained:
(i) from the sprint test using a fixed alactic time of 3 s, and
(ii) indirectly by fitting vLamax within the model to reproduce the measured LT2 response.
Internal model consistency was evaluated by predicting LT1 using the LT2-anchored vLamax and comparing against measured LT1.
RESULTS:
Sprint-derived vLamax (0.34 +/- 0.08 mmol/L/s) resulted in systematic overestimation of both LT1 (predicted: 250.1 +/- 45.9 W vs measured: 211.7 +/- 37.2 W) and LT2 (349.6 +/- 51.0 W vs 323.9 +/- 41.5 W, p < 0.001). This yielded metabolic profiles inconsistent with empirical data due to a systematic upward bias in fractional utilisation at LT2 (85.5 +/- 2.9% VO2max vs 80.9 +/- 2.4% VO2max) and elevated simulated BLCs (3.9 +/- 1.7 mmol/L vs 2.7 +/- 0.8 mmol/L). Indirect vLamax estimation anchored to measured LT2 improved reconstruction of LT1 (223.9 +/- 32.3 W), reducing mean absolute prediction error from 43.2 +/- 31.2 W to 16.0 +/- 13.2 W (p < 0.001). Model predictions indicated higher internal consistency, as the LT2-anchored vLamax showed an improved alignment with measured LT1 (mean error: 12.2 +/- 16.9 W).
CONCLUSION:
Sprint-derived vLamax appears less suitable for metabolic modelling in elite rowing. Threshold-anchored indirect estimation provides a more coherent parameter set, improves the simultaneous reconstruction of LT1 and LT2, and enhances internal model consistency. This approach may provide a superior basis for individualised training prescription in athletes with discipline-specific constraints on maximal sprint performance.