Monitoring Skeletal Muscle Oxygen Saturation Kinetics during Graded Exercise Testing in NCAA Division I Female Rowers

Drake Allen Eserhaut ^1,2* Joseph M Deleo ¹ Jessica A Provost ¹ Kathryn E Ackerman ^3,4,5 Andrew C. Fry ^1,2

• ¹ Jayhawk Athletic Performance Laboratory - Wu Tsai Human Performance Alliance, Lawrence, KS, United States
• ² Department of Health, Sport, and Exercise Science, University of Kansas, Lawrence, KS, United States
• ³ Wu Tsai Female Athlete Program - Department of Orthopedics, Boston Children's Hospital, Boston, MA, United States
• ⁴ Neuroendocrine Division, Massachusetts General Hospital, Boston, MA, United States
• ⁵ Harvard Medical School, Boston, MA, United States

Purpose: The purpose of this study was to analyze how skeletal muscle oxygen saturation (SmO2) kinetics change during exercise in female rowers both acutely and longitudinally in relation to blood lactate (BLa). We also aimed to determine the agreement and statistical equivalence between physiological thresholds derived from SmO2 and BLa kinetics. Methods: Twenty-three female NCAA Division I rowers were tested throughout the 2023-2024 academic year. Of these, 11 athletes completed at least two near-infrared spectroscopy (NIRS)-equipped GXTs, with physiological data analyzed for longitudinal changes. A 7x4-minute discontinuous GXT protocol was performed by all athletes. First and second SmO2 breakpoints (SmO2BP1 and SmO2BP2) were estimated via piecewise linear regression modeling, and BLa thresholds (LT1 and LT2) were calculated using ADAPT software. Paired-samples t-tests assessed differences, and equivalence was tested using two one-sided tests (TOST). Agreement was determined using Bland-Altman analysis yielding mean differences (MD) and 95% limits of agreement (LoA). Intraclass correlation coefficients (ICC2,1) were also calculated. Results: No difference was found between SmO2BP2 and LT2 (MD = -5.76W [95% LoA = -38.52 to 22.25W], p = 0.134), moderate-to-good levels of agreement (ICC2,1=0.67 [95% CI: 0.36-0.85], p<0.001), and no statistical equivalence (p=0.117). This was not the case for SmO2BP1 and LT1, with NIRS significantly underestimating LT1 (MD = -8.14W [95% LoA = -38.90 to 27.37W], p = 0.026), poor-to-moderate agreement (ICC2,1=0.24 [95% CI: -0.13-0.58], p=0.10), and no statistical equivalence (p=0.487). Additionally, SmO2 recovery kinetics (SmO2resat) during 1-minute rest intervals increased in response to graded increases in exercise intensity (p<0.001, η2p=0.71), with higher intensities appearing to blunt this effect (step 6 – step 7: MD=-0.16%•s-1, p=0.69). No statistically significant changes were observed in LT’s or SmO2BP’s throughout the 2023-2024 season. Conclusions: In female collegiate rowers, NIRS may be a tool that compliments BLa testing when determining the second lactate threshold (i.e., LT2). However, significant inter-individual variablility exists between SmO2BP2 and LT2 paired with a lack of statistical equivalence suggest the two are not inter-changeable. While not a standalone replacement, if used in combination with traditional BLa testing methods NIRS may be a complimentary tool that helps inform individual athlete training zone prescription.