How much of the incremental protocol does a lactate threshold need? Agreement between three-stage and five-stage testing across seven lactate-derived indices in swimmers.

BACKGROUND

Incremental lactate testing is widely used to characterize the lactate-velocity relationship in swimmers, but the highest-intensity stages increase testing burden and may be difficult to execute consistently.

OBJECTIVE

To determine whether a five-stage protocol can be reduced to three stages without materially changing seven lactate-derived velocity indices, and whether model complexity affects reconstruction.

METHODS

Ninety highly trained swimmers completed a 5 × 200 m freestyle incremental test at 80-100% of 200 m personal-best performance. Stage-4 and stage-5 lactate values were reconstructed from the first three stages using participant-specific linear extrapolation and a long short-term memory (LSTM) model. Seven lactate-derived indices were recalculated and compared with full-protocol values using limits of agreement, typical error, ICC(A,1) and equivalence testing.

RESULTS

Lactate reconstruction was inaccurate for both models, yet v@Exp-Dmax showed close agreement with the full protocol (bias -0.008 m·s-1; limits of agreement -0.023 to +0.007 m·s-1; ICC 0.988) and met the pre-specified equivalence criterion. The other six indices did not meet equivalence. LSTM provided no advantage over linear extrapolation, and reconstructed v@Exp-Dmax preserved the performance-association pattern observed with the full protocol.

CONCLUSIONS

A three-stage protocol may support lower-burden longitudinal monitoring of v@Exp-Dmax, but protocol simplification is index-specific and does not establish physiological threshold validity.