Author(s): UMEMOTO, M., HYODO, H., Institution: PHYSICAL EDUCATION, Country: JAPAN, Abstract-ID: 187

INTRODUCTION:
Blood lactate (bLA) is widely used to quantify internal load in swimming; however, repeated blood sampling limits practical implementation. Wearable sweat lactate (sLA) sensors offer a non-invasive alternative, yet evidence in a regular pool setting under intermittent protocols remains limited. This pilot study examined the feasibility of continuous sLA monitoring during swimming and its agreement with bLA-derived threshold metrics during an intermittent lactate curve test.
METHODS:
Two trained swimmers completed an intermittent lactate curve test in a standard 25-m pool (front crawl): five 200-m bouts separated by 10-min intervals with progressively higher target intensities (from submaximal to maximal). Capillary bLA was measured 3 min post-bout (and additionally 6 min after the maximal bout). Continuous sLA was recorded at 1 Hz using a wearable sensor secured at the mid-forehead with waterproof sealing; data were processed using a 13-s moving average and individual baseline correction. Sweat lactate threshold (sLT) was defined as the first clear rise above baseline on the time-series plot, and blood lactate threshold (bLT) was determined using a graph-based method. Differences across sampling timings were explored using one-way ANOVA with Bonferroni-corrected paired t-tests, acknowledging the very small sample size. Agreement between sLT-speed and bLT-speed was assessed using Bland–Altman analysis with difference defined as (bLT-speed − sLT-speed).
RESULTS:
Continuous sLA acquisition was maintained during vigorous swimming including turns, with no noticeable water ingress. bLA showed a significant main effect across sampling timings, whereas post-hoc pairwise comparisons did not detect significant differences after correction. sLA exhibited transient decreases during rest intervals, consistent with the intermittent structure of the protocol. Bland–Altman analysis showed a negative mean difference (bLT-speed − sLT-speed), indicating that sLT-speed tended to be estimated faster than bLT-speed. Limits of agreement were wide (LCL −0.56 to UCL 0.18 m·s⁻¹), suggesting substantial individual variability in agreement under the present conditions.
CONCLUSION:
Continuous sLA monitoring was feasible in a standard pool setting during an intermittent swimming lactate curve test. However, agreement between sLT-speed and bLT-speed varied and tended toward faster sLT-speed estimates, with wide limits of agreement in this pilot dataset. Further work with larger samples and repeated trials is required to stabilize agreement estimates and to compare intermittent protocols directly with continuous incremental conditions, thereby clarifying the practical utility of sLA-based threshold monitoring for coaching.