Sage Journals: Discover world-class research

Abstract

Energy deficiency (ED) and psychological stress affect athlete health. Given the emerging capabilities of wearable technology, the purpose of this study was to explore associations of wearable technology metrics in relation to lab-based measures of ED and psychological stress. We investigated the associations between (a) wearable-derived heart rate variability (HRV), resting heart rate (RHR), exercise strain and recovery, and (b) laboratory-derived measures of metabolism (resting metabolic rate (RMR), total triiodothyronine (TT₃)) and a validated stress assessment (Recovery-Stress Questionnaire for Athletes (RESTQ)-52 items) in NCAA swimmers (n = 23, 10 male, 13 female) during heavy training. Swimmers were grouped by degree of metabolic adaption to ED using the ratio of actual-to-predicted RMR (utilizing the validated cutoff of <0.94) and by sex. Metabolically suppressed swimmers had lower HRV (81 ± 27 ms vs.110 ± 35 ms, p = 0.04). HRV correlated positively with RMR (kcal·kg LBM⁻¹·day⁻¹; where LBM stands for lean body mass) (r = 0.45; p = 0.03). HRV was negatively correlated with sport-specific (r = −0.46; p = 0.03) and total stress (r = −0.46; p = 0.03). In males, HRV correlated negatively with general stress (r = −0.72; p = 0.02) and total stress (r = −0.74, p = 0.01) (RESTQ). Additionally, in males only, the strain was correlated positively with RESTQ recovery–stress balance (r = 0.69; p = 0.03), and negatively correlated with general stress (r = −0.81, p = 0.01), and sport stress (r = −0.89, p < 0.01). No correlations between HRV, RHR, strain or recovery, and stress variables were observed in females. Associations between wearable technology measures of HRV, RHR, strain, and recovery with validated measures of ED and psychological stress should continue to be explored with a focus on underlying mechanisms and moderating influences of biological sex.

Keywords

Accelerometer body mass heart rate variability photoplethysmography recovery resting metabolic rate

Introduction

Inadequate energy intake relative to energy expenditure, that is, energy deficiency (ED), negatively impacts health and sport performance in soccer players¹ and junior elite swimmers,² and can lead to serious health consequences such as the Female and Male Athlete Triad (Triad).^3–5 Therefore, the prevention of ED is key to preventing and managing Triad-related outcomes.³ Measuring ED and metabolic status is challenging due to inaccuracies and under-reporting of dietary intake³ and measures of metabolic suppression such as resting metabolic rate (RMR) and serum total triiodothyronine (TT₃)⁶ are expensive, time-consuming, and often unattainable by practitioners. Similarly, athletes are highly susceptible to additional stressors, including training and competitions that cause physiological, metabolic, and psychological stress coupled with other lifestyle stressors, such as academic and social stress.⁷ Recent research has attempted to relate psychological stress with eating behavior. In athletes, stress has been related to mental health comorbidities,⁸ and greater reported psychological stress has been associated with poor energy status in rowers undertaking high-volume training without supplemental increase in calories.⁹ Similarly, in active individuals, exercising women with functional hypothalamic amenorrhea presented with both chronic adaptations to poor energy status (low RMR ratios) and greater indicators of psychological stress compared to eumenorrheic women.¹⁰ Psychological stress (measured using a validated questionnaire such as the Recovery-Stress Questionnaire for Athletes (RESTQ))¹¹ can exacerbate physical stress and lead to overtraining syndrome.^12,13 Such measures have been demonstrated to be useful in determining training and recovery status in athletes.¹⁴

The increased use of wearable technology by sports teams warrants more research to test whether wearable metrics are valid and reliable indicators of athlete health and performance parameters. A starting point for this work is to determine how wearable metrics compare to other validated measures of interest. The WHOOP performance optimization system (WHOOP Inc., Boston, MA) is a popular wrist-worn device commonly used by athletes to obtain biofeedback about their stress and recovery through wearable measurement of heart rate variability (HRV), resting heart rate (RHR), strain (a measure of “cardiovascular load” or physical stress), and recovery under various training loads, allowing for monitoring adaptation to sport training.¹⁵ Coaches are using both individual and aggregate data from wearables to monitor internal and external load measures and prescribe recovery in order to minimize the risk of overtraining and/or injury.¹⁵ Unique to the WHOOP, the device measures HRV during sleep, a reflection of the fluctuation in time intervals between consecutive heartbeats as influenced by an autonomic nervous system.¹⁶ Prior research has identified that HRV is associated with physical training adaptation in athletes,¹⁷ that is, exercise capacity (as measured by VO₂ max, exercise post-oxygen consumption,¹⁸ and blood lactate¹⁹), exercise tolerance,¹⁷ psychological stress¹⁶ (as measured by pre-competitive anxiety),²⁰ suppressed metabolism (as measured by RMR, relative RMR, and free triiodothyronine)^21,22 and the onset of overtraining syndrome.^13,23 Specifically, in elite cyclists subjected to a progressive increase in training load, both HRV and RMR were simultaneously reduced, and measures of HRV were related to two indices of metabolism, that is, relative RMR and free triiodothyronine, whereby a higher HRV corresponded with a higher RMR and greater hormonal concentrations.²² Reductions in caloric intake have demonstrated effects on reductions in both HRV^21,24,25 and RMR.^26,27 As well, a high and consistent training load coupled with ED has been theorized to exacerbate overtraining effects^13,28 on autonomic nervous system function in endurance athletes¹³ that may present as low HRV. Therefore, exploration of wearable-derived HRV and its association with metabolic compensation and psychological stress during heavy training is warranted, as wearable technology may hold promise as a biofeedback tool for monitoring adaptation to training, thereby improving the prevention of the Athlete Triad and overtraining syndrome.

Given the importance of the need to properly match fueling with exercise energy expenditure (EEE) for the prevention of ED on athlete health and performance, and as well, given prior observations that wearable-derived HRV may be associated with some indicators of metabolic adaptation, our interest was to explore whether physiological indications of ED can be detected by wearable technology. As well, prior observations indicate that HRV may relate to psychological stress,²⁰ we also explored whether WHOOP measures (HRV, RHR, strain, and recovery) may be significantly associated with a validated, sports-specific questionnaire commonly used to capture sports-specific and general stress, as well as recovery and the balance between stress and recovery, that is, the RESTQ.^14,29 Therefore, the emerging reports of the use of wearable technology to assess physiological and potentially psychological factors that may impact health and performance, the purpose of this study was to (i) examine the associations among WHOOP measures (HRV, RHR, strain, and recovery) and laboratory measures of metabolism (RMR and TT₃), and (ii) examine associations among WHOOP measures and validated assessments of stress and recovery (RESTQ) in male and female Division 1 swimmers during the heaviest phase of a training season. Lastly, since the susceptibility to ED may vary based on sex,^4,5 we explored whether the relationships between WHOOP-derived variables and psychological stress and ED were sex-dependent.

Materials and methods

Participants

Twenty-seven members of a NCAA Division 1 Swim team participated in this study (11 males, 16 females). These athletes comprised NCAA competitors, Olympic Trials qualifiers, College Swimming & Diving Coaches Association of America All-Americans, and national and international team members and competitors from varying countries. All participants were in good health, between the ages of 18–22 years, and not presently injured or requiring any major training modifications.

Design

This cross-sectional study examined the association of WHOOP measures with metabolism and psychological stress in 23 elite male (n = 10) and female (n = 13) collegiate swimmers during the heaviest training phase of the NCAA swim season (Figure 1). The study was approved by the University Institutional Review Board. All participants provided consent prior to study participation. Data collection occurred over six weeks, lasting the duration of the training phase prior to training load reduction, preparing athletes for championship competition. Training data and physiological data from the WHOOP wearable were collected daily and averaged over two weeks to obtain an aggregate view of physiological responses. A validated survey assessing stress and recovery, that is, RESTQ,¹¹, and laboratory-based physiological testing (RMR, body composition, and blood sampling) were collected at one time during the measurement period.

Figure 1.

NCAA Division 1 swimming season training plan. This figure depicts the training plan of the swim team for the NCAA sport season. Bold represents the measurement time period utilized for data collection.

Methodology

Training

Figure 1 presents the season training plan and the measurement time period. Both sexes followed this plan over the course of the season, with the same training focus during each segment. Weekly training routines consisted of three resistance (weightlifting) training sessions, six in-water training sessions, and one in-water recovery training session. Training load was quantified as yards/day, yards/week, and min/week of training. Subjective training load was measured as the weekly average of ratings of perceived exertion (RPE) using the Borg CR10 scale³⁰ during training sessions. Energy expended during training was measured directly using the WHOOP heart rate and three-dimensional-accelerometry analysis (WHOOP Inc.) and is described as EEE (kcal/day).

Anthropometrics and body composition

Total body mass was measured using a physician's scale (Seca Model 770; Seca, Hamburg, Germany), and height was measured by using a stadiometer to the nearest 0.5 cm. Body mass index was calculated as weight divided by height squared (kg/m²). Body composition was determined using dual x-ray absorptiometry (DXA) (Hologic Horizon-W, Model 201331) performed by an International Society of Clinical Densitometry-certified technician. Total percent body fat, fat mass, fat-free mass, and lean body mass were quantified.

Resting metabolic rate (RMR)

RMR was measured by indirect calorimetry using a ventilated hood and metabolic cart (Vmax Encore Metabolic Cart; Vyaire, Chicago, IL) following a 12-hour fast. Participants abstained from caffeine and alcohol for 24 hours, and testing occurred following participants’ day off of training or following their recovery day. Calculation of RMR was performed as previously published.⁶

A validated biomarker of ED is the ratio of measured-to-predicted RMR ratio,⁶ where a low ratio represents metabolic compensation to ED. Low RMR ratios below a specific threshold, dependent upon the prediction equation utilized, reflect a reduction of RMR on a per kilogram tissue basis, therefore relative RMR (kcal per kg lean body mass (LBM)) is often also reported. This ratio significantly predicted low TT₃ in a population of exercising women.⁶ DXA-derived predicted RMR equation was used in the ratio of measured-to-predicted RMR calculations as previously described.^6,31,32 RMR ratios were categorized into suppressed (<0.94) or non-suppressed (≥0.94) groups for analysis, based on a validated cutoff for RMR DXA-derived ratio utilized in a previous investigation.⁶

Serum TT₃

Blood was collected from an antecubital vein immediately following the RMR test. Samples were allowed to clot at room temperature. Serum was then assayed for TT₃ using an Immulite chemiluminescent analyzer (Immulite; Siemens Healthcare, Erlangen, Germany) through competitive immunoassay. Analytical sensitivity for the TT₃ assay was 35 ng/dL (0.54 nmol/L). The intraassay and interassay coefficients of variation were 13.2% and 15.6%, respectively.

Assessment of self-reported stress and recovery

On one occasion prior to a training session, participants completed the RESTQ¹¹ for assessment of mood and feelings of recovery and stress. The RESTQ was developed for use in athletes to assess subjective feelings of stress and recovery. Stress and recovery subscales are divided separately into (a) general, (b) sport-specific, and (c) total (combined general and sport-specific categories).¹¹ The Total Recovery–Stress Balance was determined using the stress and recovery difference (total recovery–total stress), with higher total recovery-stress scores indicating greater recovery, and lower scores reflecting greater stress.¹¹

WHOOP measures

The WHOOP Performance Optimization System (WHOOP Inc.) is a wrist-worn multi-sensor (tri-axial accelerometer, optical sensor, capacitive touch sensor, and ambient temperature sensor) device. Athletes wore the WHOOP device on their non-dominant wrist continuously for the duration of the study. Data were relayed to a cloud-based analytics platform and retrieved for analysis. This study included an analysis of the WHOOP measures of HRV, RHR, and training data encompassing strain and recovery measures along with purposeful EEE (kcal). Although WHOOP assesses sleep metrics, these measures were not included in the present analysis.

The device uses photoplethysmography for resting HRV, RHR, and EEE measurements.³³ HRV is measured during the last 5 minutes of slow-wave sleep using the root mean square of the successive differences in the interbeat intervals, quantified in milliseconds. To determine caloric expenditure, the device estimates the users’ metabolic rate using a proprietary formula based on the users’ entered height, weight, and gender, then as heart rate rises during activity, WHOOP uses their proprietary formula: “Calories Burned = BMR + function (Heart Rate)” (WHOOP Inc.). The WHOOP also provides a proprietary measure of cardiovascular load experienced during a 24-hour period referred to as “strain” scaled from 0 to 21. Strain is determined daily as an aggregate of cardiovascular load accumulated for training sessions and throughout the day using measured heart rate (photoplethysmography) and the duration of time spent in heart rate zones (zones are 0%–49%, 50%–59%, 60%–69%, 70%–79%, 80%–89%, and 90%–100% derived from a user's age-based predicted maximum heart rate) (WHOOP Inc.). WHOOP recovery is an indication of a daily measure of how prepared your body is to perform or take on strain from activity. Recovery is determined and “scored” by the device utilizing measures of HRV, RHR, and sleep inputted into a proprietary algorithm for determination. The WHOOP has been validated for use in determining sleep measures, heart rate, and HRV in healthy young individuals (both males and females) at rest.^34,35

Statistical analysis

Normality was tested using the Shapiro–Wilk statistic. Outliers were located and removed, and Levene's test was utilized to determine the homogeneity of variance. Participants were analyzed as a whole group, by sex, and by whether they were metabolically suppressed or non-suppressed based on the ratio of measured-to-DXA-predicted RMR utilizing the proposed ratio cutoff of <0.94 indicating metabolic suppression as previously described.⁶ WHOOP variables of interest for comparison to metabolism and indications of stress consisted of HRV, RHR, strain, and recovery. Correlations were calculated using Pearson's correlation analysis. Independent t-tests were conducted to determine group differences between sexes and between suppressed and non-suppressed participant groups. Data were analyzed using SPSS Statistical Software (version 26, Chicago, IL) and reported as mean ± SD. A p-value of <0.05 was considered significant.

Results

Participants

Twenty-seven participants enrolled in the study and data from 23 participants were included in analyses. Two participants with outlying data points were excluded from the analysis, as was one who prematurely reduced training volume prior to data collection completion, and one who engaged in excessive supplementation and dieting practices. No participants dropped out. Participants were primarily Caucasian with two participants identifying as African American/Caribbean, and Latin American, respectively. Descriptive information about the participants is displayed in Table 1. As expected, male swimmers had significantly greater height, weight, lean and fat-free mass, and lower fat mass and body fat percentage (p < 0.05; t-test) compared to female swimmers.

Table 1.

Participant characteristics.

	All (N = 23)		Male (n = 10)		Female (n = 13)
	Mean ± SD	Range	Mean ± SD	Range	Mean ± SD	Range	p value
Demographics
Age (years)	19.6 ± 1.0	18.0–21.0	20.1 ± 0.9	18.0–21.0	19.2 ± 0.9	18.0–21.0	0.046
Height (cm)	179 ± 8	163–194	186 ± 5	179–194	173 ± 5	163–182	0.001
Weight (kg)	74.8 ± 10.7	60.3–99.4	83.8 ± 8.6	70.6–99.4	67.9 ± 6.2	60.3–79.0	0.001
BMI (kg/m²)	23.0 ± 1.9	19.6–26.4	24.1 ± 1.9	20.4–26.4	22.6 ± 1.8	19.6–24.8	0.063
Body fat (%)	22.0 ± 5.2	14.2–29.5	16.6 ± 1.9	14.2–19.5	26.7 ± 1.6	24.1–29.5	0.001
Fat mass (kg)	15.9 ± 2.9	11.2–21.1	13.7 ± 2.1	11.2–18.3	17.8 ± 1.9	14.5–21.1	0.001
LBM (kg)	54.7 ± 11.0	39.7–74.7	65.6 ± 6.4	55.8–74.7	46.4 ± 4.3	39.7–54.6	0.001
FFM (kg)	57.4 ± 11.4	42.0–77.8	68.5 ± 6.7	58.1–77.8	48.8 ± 4.6	42–57.7	0.001
Training
Week volume (yards/week)	38,850 ± 3720	34,650- 45,250	39,175 ± 4782	34,650- 45,250	38,575 ± 2835	34,650- 45,250	0.730
Day volume (yards/day)	6700 ± 720	5775–7775	6550 ± 797	5775–7775	6825 ± 655	5775–7775	0.316
Duration (min/week)	1100 ± 42	968–1135	1080 ± 58	968–1135	1113 ± 15	968–1135	0.099
sRPE (1–10)	6 ± 0.8	5–7.9	7 ± 0.8	5–7.9	6 ± 1.3	5–7.9	0.367
EEE (kcal/day)	717 ± 272	366–1284	901 ± 279	474–1284	578 ± 177	366–874	0.003
Metabolism
mRMR (kcal/day)	1570 ± 274	1131–2041	1831 ± 178	1513–2041	1370 ± 118	1132–1562	0.001
pRMR (kcal/day)	1696 ± 308	1281–2284	2005 ± 164	1761–2284	1457 ± 112	1281–1670	0.001
RMR (kcal/kg LBM/day)	29 ± 2	25–32	28 ± 2	25–30	30 ± 2	27–32	0.005
RMR ratio	0.93 ± 0.05	0.84–1.00	0.92 ± 0.05	0.84–1.00	0.93 ± 0.04	0.85–0.99	0.247
TT₃ (ng/dL)	90.4 ± 16.1	64.7–133.5	87.8 ± 11.6	67.8–131.5	92.5 ± 19.4	64.7–133.5	0.050
Stress
General stress	15 ± 6	7–33	16 ± 8	7–33	14 ± 4	9–22	0.332
Sport specific stress	7 ± 3	2–13	8 ± 8	3–13	6 ± 2	2–10	0.215
Total stress	22 ± 8	10–46	25 ± 10	10–46	20 ± 6	11–28	0.222
Recovery
General recovery	17 ± 4	10–24	16 ± 4	10–23	18 ± 3	14–24	0.106
Sport specific recovery	14 ± 3	8–20	13 ± 4	8–20	14 ± 2	11–19	0.338
Total recovery	31 ± 6	18–42	28 ± 8	18–42	32 ± 5	27–42	0.159
Total recovery–stress balance
Final score (recovery-stress)	8.5 ± 12.5	−28–28	4.3 ± 14.7	−28–20	12.3 ± 10	−1–28	0.130
WHOOP measures
HRV (ms)	95 ± 34	40–167	85 ± 36	40–167	102 ± 31	56–155	0.237
RHR (bpm)	52 ± 5	44–63	52 ± 6	44–61	52 ± 5	44–63	0.698
Strain (AU: 0–21)	14.3 ± 1.6	10.5–17.3	14.5 ± 1.2	12.3–16.8	14.2 ± 1.8	10.5–17.3	0.621
Recovery score (0–100%)	59 ± 5	49–73	59 ± 6	51–73	59 ± 5	44–66	0.883

BMI: body mass index; LBM: lean body mass; FFM: fat-free mass; sRPE: session rating of perceived exertion; EEE: exercise energy expenditure; RMR: resting metabolic rate; mRMR: measured RMR; pRMR: predicted RMR; TT₃: total triiodothyronine; HRV: heart rate variability; RHR: resting heart rate; SD: standard deviation.

Training

Weekly and daily training volume, duration, RPE, and EEE are depicted in Table 1. The training group breakdown consisted of 8 sprints (5 male; 3 female), 10 middle-distance (2 male; 8 female), and 5 distance swimmers (3 male; 2 female). Participants averaged 38,850 ± 3720 yards/week, 6700 ± 720 yards/training day, and trained 1100 ± 42 minutes/(approximately 18 hours)/week (Table 1). No significant sex differences in training parameters existed, except EEE (kcal/day) was greater in males than females (p = 0.003).

WHOOP variables and metabolism

Descriptive variables for metabolism and from the WHOOP are presented in Table 1. Male swimmers had significantly greater measured RMR and predicted RMR, but relative RMR was greater in female swimmers, while there was no significant difference in RMR ratio and TT₃ between sexes. Correlations between WHOOP variables and metabolic variables are presented in Table 2. HRV was positively correlated with relative RMR (kcal/LBM) in all participants (r = 0.448; p = 0.032). However, no correlation was evident between HRV and either the RMR ratio or serum TT₃ concentrations in all participants. There were no significant relationships evident between other WHOOP variables (RHR, strain, or recovery) and indications of metabolism (RMR ratio, relative RMR, or TT₃) when analyzing all participants together. Results illustrating the relationship between HRV and metabolic status are depicted in Figure 2. The metabolically suppressed participants (n = 12; 7 men, 5 women, RMR ratio <0.94) exhibited a significantly lower mean HRV than non-suppressed participants (n = 11; 4 men, 7 women) (p = 0.039). There were no differences between suppressed and non-suppressed in body weight, body fat, fat-free mass, or TT₃.

Figure 2.

The relationship between HRV and RMR and stress in all swimmers. (A) The positive correlation between HRV and relative resting metabolic rate (RMR per kg LBM) in all participants (R = 0.45; p = 0.03). (B) Comparison of HRV between RMR ratio groups. HRV for all, male and female participants during the two-week measurement time period. Participants categorized as “Suppressed” (below the RMR ratio cutoff value (<0.94) of measured to DXA predicted RMR) are displayed using black bars, and “Non-Suppressed” participants (equal to or above the RMR ratio cutoff value) are displayed using grey bars. Statistically significant differences between metabolism groups for “ALL” noted with *p < 0.05. (C) The negative correlation between HRV and sport-specific stress (RESTQ) in all swimmers (R = −0.46; p = 0.03).

Table 2.

Correlations between WHOOP variables, metabolism, stress, and recovery.

WHOOP variables:	HRV	RHR	Strain	Recovery
All swimmers
Metabolism
RMR ratio	0.33	−0.32	0.33	0.22
Relative RMR	0.45^a	−0.37	0.37	−0.03
TT₃	0.03	0.42	−0.21	−0.02
Stress and recovery
General stress	−0.37	0.32	−0.32	0.06
Sport-specific stress	−0.46^a	0.19	−0.36	−0.05
Total stress	−0.46^a	0.32	−0.38	0.03
General recovery	0.12	−0.18	−0.01	0.02
Sport-specific recovery	−0.24	−0.09	−0.05	−0.06
Total recovery	−0.05	−0.18	−0.03	−0.01
Recovery–stress balance	0.26	−0.28	0.23	−0.03
Male swimmers
Metabolism
RMR ratio	0.49	−0.32	0.33	0.65^a
Relative RMR	0.53	−0.47	0.89^a	0.29
TT₃	−0.35	0.42	0.12	0.03
Stress and recovery
General stress	−0.72^a	0.75^a	−0.81^a	−0.02
Sport-specific stress	−0.60	0.48	−0.89^a	−0.17
Total stress	−0.74^a	0.72^a	−0.89^a	−0.07
General recovery	0.26	−0.52	0.17	0.11
Sport-specific recovery	−0.20	−0.02	0.09	−0.07
Total recovery	0.05	−0.30	0.14	0.03
Recovery–stress balance	0.54	−0.66^a	0.69^a	0.06
Female swimmers
Metabolism
RMR ratio	0.06	−0.30	0.25	−0.29
Relative RMR	0.25	−0.55	0.35	−0.32
TT₃	0.19	0.43	−0.30	−0.03
Stress and recovery
General stress	0.31	−0.30	−0.03	0.20
Sport-specific stress	−0.27	−0.07	−0.11	0.07
Total stress	0.11	−0.28	−0.07	0.20
General recovery	−0.26	0.17	−0.07	−0.08
Sport-specific recovery	−0.48	−0.23	−0.10	−0.02
Total recovery	−0.40	−0.02	−0.10	−0.06
Recovery–stress balance	−0.29	0.13	−0.01	−0.15

HRV: heart rate variability; RHR: resting heart rate; RMR: resting metabolic rate; TT₃: total triiodothyronine. Correlations are described with R values in table. 'a' denotes statistical significance (p<0.05) between variables.

There was no association between HRV and relative RMR (kcal/kg LBM) or RMR ratio when males and females are considered separately. Similarly, no correlation was evident between HRV and serum TT₃ concentrations in participants when grouped by sex. WHOOP strain was positively correlated with relative RMR (r = 0.890; p = 0.001), and WHOOP recovery score was positively correlated to RMR ratio in male swimmers (r = 0.653; p = 0.041), but neither relationship was evident in females or when analyzing both sexes together. There were no relationships evident between RHR and metabolic variables in any analysis.

WHOOP measures and stress

Descriptive variables from the WHOOP and RESTQ stress and recovery are presented in Table 1. There were no significant differences in any WHOOP measures or any RESTQ stress or recovery measures between the sexes. Correlations are presented in Table 2. We found a negative correlation between HRV and sport-specific stress (r = −0.462; p = 0.026) and total stress (r = −0.459; p = 0.028) in all participants. No other correlations between HRV, RHR, strain, or recovery and general stress, any recovery subscales (general, sport-specific, and total), or recovery–stress balance variables were found in all participants.

Sex-specific analysis revealed a negative correlation in males between HRV and general stress (r = −0.724, p = 0.018) and total stress (r = −0.741, p = 0.014), indicating that a higher HRV was correlated with lower total and general stress. In male swimmers only, RHR was positively correlated to general (r = 0.745; p = 0.013), and total stress (r = 0.722; p = 0.018), and negatively correlated with recovery–stress balance (r = −0.657; p = 0.039), indicating a higher RHR was associated with higher reported stress and a worse/lower recovery–stress balance. Male swimmers demonstrated a negative correlation between strain and general stress (r = −0.811, p = 0.004), sport-specific stress (r = −0.889, p = 0.001), and total stress (r = −0.891, p = 0.001). Additionally, in males only, there was a positive correlation between strain and total recovery–stress balance (r = 0.688; p = 0.028), while no correlations were observed in females. Finally, no correlations were evident between WHOOP recovery and RESTQ stress or recovery variables in either sex.

Discussion

This is the first study to explore the associations between WHOOP-derived variables and measures of metabolic compensation and stress in elite athletes. We observed a significantly lower HRV in swimmers categorized as metabolically suppressed and a positive correlation between HRV and relative RMR, suggesting a relationship between HRV and indicators of metabolic compensation in collegiate swimmers utilizing the WHOOP Performance Optimization System (WHOOP Inc.). Additionally, HRV was negatively related to both sport-specific and total stress in all swimmers. In male swimmers, we observed a positive relationship between strain and relative RMR. As well, we found a positive relation between WHOOP recovery and RMR ratio, indicating males with a higher recovery score had a higher RMR ratio. We also observed that in males, HRV was negatively correlated with stress (RESTQ general stress and total stress), indicating that higher HRV was associated with lower psychological stress. RHR was positively correlated with general and total stress, while negatively correlated with recovery–stress balance in males, indicating a higher RHR was related to higher stress and a lower/worse recovery–stress balance. The strain was positively correlated with RESTQ total recovery–stress balance, and negatively correlated to general, sport-specific, and total stress in males. While these findings promote the potential utility of the WHOOP, further research could examine whether WHOOP HRV, RHR, and strain could be used to monitor adaptation to training load and psychological stress in female and male swimmers. Our findings are also important because they offer scientific findings related to the use of the WHOOP, a device that is being utilized with increasing frequency among athletes and coaches for training monitoring.

Previously, measures of metabolism and HRV have been interpreted by some researchers as indicators of adaptations to training, where consistently low measured values for both variables serve as precursors or indicators of the onset of overtraining syndrome.^12,13 The present study is the first to show significant relationships between HRV and laboratory-based measures of metabolism (relative RMR and RMR ratio), indicating that wearable HRV may have some association with metabolic processes in the body. When energy demands are not met chronically, a reduction of RMR occurs on a per-tissue basis in order to reallocate energy toward vital life processes (metabolic compensation).²⁷ Positive training adaptations have been demonstrated to correspond to a higher HRV while negative training outcomes such as excess fatigue, worsened performance, and changes in mood correspond to a lower HRV³⁶ and negative alterations in circulating metabolic markers.^12,13 Although the underlying mechanism between HRV and its relation to indications of metabolic compensation (i.e. reduced RMR) is unclear, a diminished sensitivity to the sympathetic nervous system, that is, 'adrenal exhaustion' that may occur during chronic ED^24,37 coupled with heavy training^36,38 may manifest as a lower HRV and lower RMR, potentially explaining the low HRV observed in participants with lower RMR ratios. Supportive of this theory, Woods et al.²² found a significant positive relationship between HRV and free T3, as well as a uniform reduction in both RMR and HRV when subjecting elite male cyclists to intensified heavy training.²² Interestingly, the present study did not find an association between measures of TT₃ and HRV. However, heavy training has been shown to improve thyroid hormone production, metabolism, and clearance in elite athletes,³⁹ potentially attenuating a typical thyroid response to heavy training associated with chronic ED. Finally, we found that WHOOP strain was positively related to relative RMR, and WHOOP recovery was positively related to the RMR ratio in male swimmers alone. As such, WHOOP strain and recovery measures may indicate that athletes who are more recovered and are prepared to undertake a greater physical strain present with a higher RMR.

Stress, recovery, and the balance between them can also influence athlete health and performance. The present study found a significant relation between HRV and RESTQ sport-specific and total stress in all swimmers. In agreement with our findings, lower measured HRV has been correlated with higher reported stress and fatigue in swimmers.⁴⁰ Therefore, lower HRV in the present cohort of swimmers could potentially indicate greater feelings of stress. There was no relationship between HRV and total recovery–stress balance, or RESTQ-derived general stress or recovery subscales in all swimmers. However, HRV in male swimmers was negatively correlated to reported general stress and total stress. Regarding the sex-specific results in our male swimmers, sex-specific stress responses have been identified where males experience a greater “fight or flight” response reflective of a greater sympathetic activation compared to females’ “tend and befriend” response that is reflective of downregulated sympathetic responses to stress due to oxytocin and other sex-hormones such as estrogen, and maintained by social and cultural gender norms.⁴¹ Therefore, one could speculate that potentially males in general experience greater sympathetic activation to stress acutely, yielding greater sympathetic receptor desensitization as stress persists chronically. Logically, the positive relationship between RHR and RESTQ general stress and total stress, and the negative relationship between RHR and recovery–stress balance in the present study supports this theory, as a higher RHR associated with greater stress could indicate a greater sympathetic activation in the male swimmers. Although the mechanistic link between HRV and stress is unclear, the WHOOP HRV and/or RHR may be a potential surrogate indicator of stress and stress-related physiological and psychological changes in male athletes, however, more research is warranted.

WHOOP strain, a combined measure of cardiovascular load and physical stress demonstrated no relationship with measures of stress and recovery in all participants. In males, however, WHOOP strain was significantly correlated to RESTQ general stress, sport-specific stress, and total stress. The RESTQ has been demonstrated to be a good measure of training fatigue during both acute and chronic training,¹⁴ and as such, is a good measure of training load, not just psychological stress. The observed relationship between the RESTQ and strain may result from the ability of better-recovered athletes to consistently achieve higher cardiovascular loading during heavy training compared to their over-trained or excessively fatigued counterparts.

This investigation utilized applied and laboratory measures, and while scientific rigor was of the highest priority, this study has limitations and our findings should be interpreted with caution. While we acknowledge that this study utilized a sample of convenience, a strength of our study includes a sample size larger than most utilized for athlete health research^1,2,9,22,40 and we achieved high compliance (100%) as all participants completed procedures in their entirety. We acknowledge that findings are specific to WHOOP, as the HRV measurement is unique to this device, has been shown to demonstrate some bias,^34,35 and therefore may not be applicable to other wearable devices. Additionally, while the WHOOP has been validated against gold-standard electrocardiogram for a measure of HR and HRV in healthy individuals,³⁴ we acknowledge that the WHOOP measurement of EEE has not yet been validated, and may represent an underestimation based on prior investigations.^42,43 This study was not an attempt to validate the wearable or to provide evidence of causality, but rather to explore the associations between wearable measures to determine the potential utility of the wearable as a surrogate indicator of physiological signs of ED and psychological stress that can only be assessed with validated laboratory-based and/or other established measures. We recognize the DXA RMR ratio cutoff of <0.94, identified and validated by Strock et al.,⁶ has not yet been validated in elite athletes or males. Finally, we acknowledge that results may vary depending on sport and training regimen; therefore, these findings are not generalizable to other types of athletes or an entire training season. Completing a prospective study would provide more insight into physiological and performance changes across a training season.

Conclusion

In conclusion, our results indicate that during heavy training, WHOOP-derived HRV may prove to be a valuable metric for sport scientists and may provide insights into metabolism and sport-specific stress among collegiate swimmers if further research can corroborate our findings and define underlying mechanisms. These results add to the current body of knowledge by demonstrating that wearable-derived HRV has the potential to be a surrogate indicator of chronic ED, as HRV was associated with RMR, and lower HRV values were correlated with a greater degree of metabolic compensation (lower RMR ratio). Finally, this study demonstrates that WHOOP HRV, RHR, and strain may provide insights into a male athlete's experienced stress. More research is warranted to determine whether wearable technology can reliably identify ED and general lifestyle stress in male and female swimmers and perhaps other athlete groups. As well, more work is necessary to understand the differences between sexes in the associations we observed.

Footnotes

Acknowledgements

The authors thank the members of the Women's Health and Exercise Laboratory, particularly laboratory technician Ellen Bingham for their important contributions to this research. We also appreciate the extraordinary cooperation of the swimming athletes, coaching staff, and support staff.

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iD

Emily A. Lundstrom

References

Reed

De Souza

Williams

. Changes in energy availability across the season in Division I female soccer players. J Sports Sci 2013; 31: 314–324.

Vanheest

Rodgers

Mahoney

, et al. Ovarian suppression impairs sport performance in junior elite female swimmers. Med Sci Sports Exercise 2014; 46: 156–166.

De Souza

Nattiv

Joy

, et al. 2014 female athlete triad coalition consensus statement on treatment and return to play of the female athlete triad: 1st international conference held in San Francisco, California, May 2012 and 2nd international conference held in Indianapolis, Indiana, May 2013. Br J Sports Med 2014; 48: 89.

Fredericson

Kussman

Misra

, et al. The male athlete triad—a consensus statement from the female and male athlete triad coalition part II: Diagnosis, treatment, and return-to-play. Clin J Sport Med 2021; 31: 349–366.

Nattiv

De Souza

Koltun

, et al. The male athlete triad—a consensus statement from the female and male athlete triad coalition part 1: Definition and scientific basis. Clin J Sport Med 2021; 31: 335–348.

Strock

NCA

Koltun

Southmayd

, et al. Indices of resting metabolic rate accurately reflect energy deficiency in exercising women. Int J Sport Nutr Exerc Metab 2020; 30: 14–24.

Hamlin

Wilkes

Elliot

, et al. Monitoring training loads and perceived stress in young elite university athletes. Front Physiol 2019; 10: 34.

Åkesdotter

Kenttä

Eloranta

, et al. Prevalence and comorbidity of psychiatric disorders among treatment-seeking elite athletes and high-performance coaches. BMJ Open Sport Exerc Med 2022; 8: e001264.

Woods

Garvican-Lewis

Lundy

, et al. New approaches to determine fatigue in elite athletes during intensified training: Resting metabolic rate and pacing profile. PLOS ONE 2017; 12: e0173807.

10.

Strock

NCA

De Souza

Williams

. Eating behaviours related to psychological stress are associated with functional hypothalamic amenorrhoea in exercising women. J Sports Sci 2020; 38: 2396–2406.

11.

Kellmann

Kallus

2001 Recovery-stress questionnaire for athletes. User manual. Champaign, Ill., Leeds: Human Kinetics.

12.

Cadegiani

Kater

. Body composition, metabolism, sleep, psychological and eating patterns of overtraining syndrome: results of the EROS study (EROS-PROFILE). J Sports Sci 2018; 36: 1902–1910.

13.

Meeusen

Duclos

Foster

, et al. Prevention, diagnosis and treatment of the overtraining syndrome: Joint consensus statement of the European College of Sport Science (ECSS) and the American College of Sports Medicine (ACSM). Eur J Sport Sci 2013; 13: 1–24.

14.

Saw

Main

Gastin

. Monitoring the athlete training response: Subjective self-reported measures trump commonly used objective measures: a systematic review. Br J Sports Med 2016; 50: 281–291.

15.

Bourdon

Cardinale

Murray

, et al. Monitoring athlete training loads: Consensus statement. Int J Sports Physiol Perform 2017; 12: S2-161–S2-170.

16.

McCraty

Shaffer

. Heart rate variability: New perspectives on physiological mechanisms, assessment of self-regulatory capacity, and health risk. Glob Adv Health Med 2015; 4: 46–61.

17.

Kaikkonen

Hynynen

Mann

, et al. Heart rate variability is related to training load variables in interval running exercises. Eur J Appl Physiol 2012; 112: 829–838.

18.

Grant

Clark

van Rensburg

DCJ

, et al. Relationship between exercise capacity and heart rate variability: supine and in response to an orthostatic stressor. Auton Neurosci 2009; 151: 186–188.

19.

Al Haddad

Laursen

Ahmaidi

, et al. Nocturnal heart rate variability following supramaximal intermittent exercise. Int J Sports Physiol Perform 2009; 4: 435–447.

20.

Morales

Garcia

García-Massó

, et al. The use of heart rate variability in assessing precompetitive stress in high-standard judo athletes. Int J Sports Med 2013; 34: 144–151.

21.

Millis

Austin

Hatcher

, et al. Metabolic energy correlates of heart rate variability spectral power associated with a 900-calorie challenge. J Nutr Metab 2011; 2011: 715361.

22.

Woods

Rice

Garvican-Lewis

, et al. The effects of intensified training on resting metabolic rate (RMR), body composition and performance in trained cyclists. PLoS One 2018; 13: e0191644.

23.

Lehmann

Foster

Dickhuth

, et al. Autonomic imbalance hypothesis and overtraining syndrome. Med Sci Sports Exercise 1998; 30: 1140–1145.

24.

Young

Landsberg

. Suppression of sympathetic nervous system during fasting. Science 1977; 196: 1473–1475.

25.

Mazurak

Gunther

Grau

, et al. Effects of a 48-h fast on heart rate variability and cortisol levels in healthy female subjects. Eur J Clin Nutr 2013; 67: 401–406.

26.

Leibel

Rosenbaum

Hirsch

. Changes in energy expenditure resulting from altered body weight. N Engl J Med 1995; 332: 621–628.

27.

Wade

Schneider

. Metabolic fuels and reproduction in female mammals. Neurosci Biobehav Rev 1992; 16: 235–272.

28.

Stellingwerff

Heikura

Meeusen

, et al. Overtraining syndrome (OTS) and relative energy deficiency in sport (RED-S): shared pathways, symptoms and complexities. Sports Med 2021; 51: 2251–2280.

29.

Kellmann

Bertollo

Bosquet

, et al. Recovery and performance in sport: Consensus statement. Int J Sports Physiol Perform 2018; 13: 240–245.

30.

Borg

. Psychophysical bases of perceived exertion. Med Sci Sports Exerc 1982; 14: 377–381.

31.

Elia

Organ and tissue contribution to metabolic rate. In: Kinney

Tucker

(eds) Energy metabolism: Tissue determinants and cellular corollaries. New York: Raven Press, 1992, pp.61–80.

32.

Hayes

Chustek

Wang

, et al. DXA: Potential for creating a metabolic map of organ-tissue resting energy expenditure components. Obes Res 2002; 10: 969–977.

33.

Georgiou

Larentzakis

Khamis

, et al. Can wearable devices accurately measure heart rate variability? A systematic review. Folia Med (Plovdiv) 2018; 60: 7–20.

34.

Bellenger

Miller

Halson

, et al. Wrist-based photoplethysmography assessment of heart rate and heart rate variability: validation of WHOOP. Sensors (Basel) 2021; 21: 3571.

35.

Berryhill

Morton

Dean

, et al. Effect of wearables on sleep in healthy individuals: A randomized crossover trial and validation study. J Clin Sleep Med 2020; 16: 775–783.

36.

Mourot

Bouhaddi

Perrey

, et al. Decrease in heart rate variability with overtraining: Assessment by the Poincare plot analysis. Clin Physiol Funct Imaging 2004; 24: 10–18.

37.

Kollai

Bonyhay

Jokkel

, et al. Cardiac vagal hyperactivity in adolescent anorexia nervosa. Eur Heart J 1994; 15: 1113–1118.

38.

Hart

Dawson

Rasmussen

, et al. β-Adrenergic receptor desensitization in man: Insight into post-exercise attenuation of cardiac function. J Physiol 2006; 577: 717–725.

39.

Rone

Dons

Reed

. The effect of endurance training on serum triiodothyronine kinetics in man: physical conditioning marked by enhanced thyroid hormone metabolism. Clin Endocrinol (Oxf) 1992; 37: 325–330.

40.

Flatt

Esco

Nakamura

. Association between subjective indicators of recovery status and heart rate variability among Divison-1 Sprint-swimmers. Sports (Basel) 2018; 6: 93.

41.

Taylor

Klein

Lewis

, et al. Biobehavioral responses to stress in females: tend-and-befriend, not fight-or-flight. Psychol Rev 2000; 107: 411–429.

42.

Bellenger

Miller

Halson

, et al. Evaluating the typical day-to-day variability of WHOOP-derived heart rate variability in Olympic water polo athletes. Sensors (Basel) 2022; 22: 6723.

43.

Støve

Hansen

ECK

. Accuracy of the Apple Watch Series 6 and the Whoop Band 3.0 for assessing heart rate during resistance exercises. J Sports Sci 2022; 40: 2639–2644.

Wearable technology metrics are associated with energy deficiency and psychological stress in elite swimmers

Abstract

Keywords

Introduction

Materials and methods

Participants

Design

Methodology

Training

Anthropometrics and body composition

Resting metabolic rate (RMR)

Serum TT3

Assessment of self-reported stress and recovery

WHOOP measures

Statistical analysis

Results

Participants

Training

WHOOP variables and metabolism

WHOOP measures and stress

Discussion

Conclusion

Footnotes

Acknowledgements

Declaration of conflicting interests

Funding

ORCID iD

References

Serum TT₃