Author(s): RIESCO, J., LÓPEZ, L., JUAN, L., PROTZEN, G., GONZÁLEZ, A., BOULLOSA, D., Institution: UNIVERSITY OF LEON, Country: SPAIN, Abstract-ID: 1685

INTRODUCTION:
The lactate threshold (LT) is the intensity at which the lactate production and clearance are balanced, beyond which a continuous increase in lactate becomes inevitable [1]. The LT is widely used for training prescription and performance prediction. Traditionally, the estimation of LT has relied on invasive and costly methods. Non-invasive and more applicable alternatives have gained increasing attention, including Dynamic Detrended Fluctuation Analysis (DDFA), which is a fractal analysis technique of heart rate (HR) in non-stationary conditions. DDFA captures changes in cardiac autonomic regulation during dynamic conditions, and it has revealed meaningful associations and good agreement between LT and DDFA-derived LT (DT) during incremental testing [2]. More recently, DDFA-based algorithms have been incorporated into commercial smartwatches (i.e., Suunto), allowing low-cost and real-time LT estimation in field-based settings.
METHODS:
Twelve well-trained runners (26 ± 6 years; 10 km performance: 3:33 ± 0:18 min/km) volunteered for participation. After warming up, the test started at 8 km/h and was increased by 1 km/h every 4 minutes until exhaustion. Blood lactate was assessed at the last 30 seconds of each stage, and LT was determined by the Log-Poly-Mod DMax method. During the test, participants wore a running smartwatch (Suunto Race S, Finland) with a HR chest strap (Polar H10, Finland) connected. The smartwatch automatically estimated the LT by using a DT method (ZoneSense, Suunto App, Finland). Intraclass Correlation Coefficient (ICC) and Bland-Altman (BA) analyses were used to assess agreement and systematic bias between the speed and HR associated with DT and LT.
RESULTS:
The ICC for speed between DT (15.5 ± 1.0 km/h) and LT (15.8 ± 1.4 km/h) showed good agreement (ICC = 0.77, p < 0.001), with a bias of −0.25 km/h observed (LoA: −1.95 to 1.45 km/h). For HR, the ICC between DT (169.7 ± 10.1 bpm) and LT (167.1 ± 12.5 bpm) also showed good agreement (ICC = 0.82, p < 0.001), with a mean bias of −2.8 bpm, and substantial individual variability (LoA: −13.0 to 7.4 bpm).
CONCLUSION:
The DDFA-based algorithm provided by the Suunto wearable device allows a valid, non-invasive, low-cost, and real-time estimation of LT speed and HR in well-trained runners. However, the substantial individual variability in HR highlights the need for individual-level interpretation. We suggest that DT can be used as an important data source to support fine-tuning decisions in day-to-day training monitoring.

REFERENCES:
[1] Faude et al., 2009; [2] Kanniainen et al., 2023.