Heart rate variability: Obtaining the stress score from SDNN values

Abstract

BACKGROUND:

Heart Rate Variability (HRV) is a non-invasive method of assessing the autonomic nervous system response during exercise and fatigue.

OBJECTIVE:

The aim of the study was to analyze the validity and feasibility of the stress score index (SS) calculated from SDNN values during exercise.

METHODS:

18 Men performed 2 running tests: 1) incremental exercise test; 2) 10-minute constant load test. Subjects underwent HRV analysis during the constant load test, before both tests, and afterward in a seated position at 3 intervals (0’–5’, 5’–10’, 10’–15’). The relationship between SDNN and SD2 was analyzed before, during, and after the test. SS was calculated as 1/SD2*1000. The Bland-Altman test analyzed the reliability of ESS. The bias, limits of agreement (LoA), standard deviation of difference, intraclass correlation (ICC), and person coefficient were calculated.

RESULTS:

The bias was 0.15 $\pm$ 2.54 (UperLOA: 2.54; LowerLOA: $-$ 2.23). In all conditions, SD2 and SDNN showed a positive linear relationship ( $r$ 2 $=$ 0.986); SS and ESS were correlated; and SS and ESS described a positive linear relationship ( $r$ 2 $=$ 0.993).

CONCLUSIONS:

The SS index calculation from SDNN is a reliable alternative during exercise.

Keywords

HRV fatigue stress index Poincaré plot SD2

1. Introduction

Heart Rate Variability (HRV) is a non-invasive method of assessing the autonomic nervous system and cardiac function [1, 2]. It consists of analysing the fluctuations in the time intervals between adjacent heartbeats [2]. Its utility and feasibility have been widely studied, with its application in health and sports being particularly important [3, 4, 5, 6, 7, 8, 9]. As HRV analysis only requires the heartbeat signal to obtain the results, its use has been extended to wearable devices (heart belts, watches, smartphones) [5, 10, 11, 12]. The algorithms that these devices typically incorporate are based on the HRV guidelines [1, 2]. The calculations of HRV are mostly based on linear parameters, such as the root-mean-square of the consecutive differences between normal heartbeats (RMSSD) and the standard deviation of normal sinus beats (SDNN) [10, 11, 13]. These parameters have been studied as markers of parasympathetic activity [2], but there are more variables that can be used as physiological stress indicators. Specifically, the Poincaré Plot is a geometric and non-linear method that plots each R-R interval as a function of the preceding R-R interval, with the values of each pair of consecutive R-R intervals defining a point on the graph. Two different indices can be derived from this diagram, SD1, which indicates parasympathetic activity, and SD2, which is the inverse of sympathetic activity [1, 14]. However, while SD1 is a direct indicator of parasympathetic activity, understanding SD2 as an indicator of sympathetic activation could lead to confusion, as it is inversely related. Therefore, there are some studies that highlight another indicator of sympathetic activity, the Stress Score (SS), which is obtained from the Poincaré Plot SD2 index [8, 9, 15, 16, 17, 18, 19, 20, 21]. SS has been shown to be a valid index to control and monitoring training load in many sports [18, 19, 20]. For example, SS has been useful tool for tracking individual assimilation of weekly workload, including the match, in professional football players [8]. Similarly, SS appears to be the best HRV predictor of internal load during recovery in hot and humid conditions [15]. SS shows a good relationship with an internal physiological variable as it is the creatine kinase (CK) before and after a handball competition [22]. Although SS can be a reliable tool, it requires the SD2 index for its calculation, which current devices (smartwatch, mobile phone etc.) hardly provide. These devices provide HRV data based on SDNN or RMSSD calculation [5, 23, 24]. SDNN and SD2 described a close linear relationship under resting conditions [9]. In this line, if these two variables had the same behavior during exercise, we could infer SD2 from SDNN measurement, and in consequence, the SS index. However, to our knowledge, the relationship between SD2 and SDNN during exercise has never been studied. Moreover, there is no wearable device or smartphone app that could provide the SS index SS. We hypothesized that it would be possible to estimate SS from SDNN data during exercise. Therefore, the aim of the present study was to establish a valid equation for estimating SS from SDNN measurements. Secondly, the study analyzed the validity and feasibility of the SS index calculated from SDNN values during exercise.

2. Material and methods

2.1 Participants

Eighteen male athletes participated in the study (25.17 $\pm$ 5.84 years; 178.37 $\pm$ 8.77 cm; 70.2 $\pm$ 12.73 kg). Inclusion criteria were that they were physically active, drug- and smoke-free, did not exercise for 24 hours before the test, and did not consume stimulant drinks or foods before each test. Each participant completed the Physical Active Readiness Questionnaire (PAR -Q) before being enrolled in the study. Before the start of the study, all subjects signed a written informed consent form in accordance with the Declaration of Helsinki. The Ethics Committee of the University Study Centre-Cardenal Spínola CEU approved the present study.

2.2 Design

All participants completed two different exercise tests 7 days apart. The first was an incremental exercise test to exhaustion with was gas analysis. Respiratory gas exchange was recorded during the incremental test. The first ventilatory threshold, second ventilatory threshold and maximal aerobic speed (MAS) were calculated [25]. The second test was a 10-min constant exercise test at 70% of MAS obtained from the previous incremental exercise test one week earlier.

All subjects underwent a seated HRV analysis in pre-test (5 minutes) and post-test (3 different conditions): from 0’ to 5’, from 5’ to 10’ and from 10’ to15’ of the recovery. The study also measured HRV during the 10-minute constant load test. The total number of HRV registers was 162. 19 registers had to be eliminated due to technical complications during the tests. The total number of HRV data analysed was 143.

2.3 Measurements

The R-R interval monitor Firstbeat Bodyguard 2 (Firstbeat Technologies Ltd, Jyväskylä, Finland) collected the data for the HRV analysis, performed with the algorithms of Kubios HRV v2.0 (Biosignal Analysis and Medical Imaging Group, Department of Physics, University of Kuopio, Kuopio, Finland). The philters provided by the Kubios software were applied when necessary (no more than 10% error) after a double visual inspection to detect artefacts. The parameters obtained from Kubios were SDNN and SD2.

2.4 Stress score and estimated stress score

Table 1
Descriptive data from SD2 and SD2 estimated (ESD2) in all conditions

Test	SD2	ESD2	ICC	Pearson
Rest	73.32 $\pm$ 16.81	73.25 $\pm$ 16.27	0.972	0.972
GXT	–	–	–	–
0–5 min_GXT	19.93 $\pm$ 10.22	20.27 $\pm$ 9.57	0.989	0.991
5–10 min_GXT	27.83 $\pm$ 11.13	26.83 $\pm$ 10.38	0.996	0.998
10–15 min_GXT	32.84 $\pm$ 14.29	31.58 $\pm$ 13.24	0.996	0.999
Rest	66.89 $\pm$ 19.07	67.05 $\pm$ 17.37	0.959	0.963
70% MAS	29.41 $\pm$ 14.87	34.4 $\pm$ 18.85	0.942	0.969
0–5 min_70%	31.58 $\pm$ 19.62	30.97 $\pm$ 18.11	0.995	0.998
5–10 min_70%	39 $\pm$ 22.16	37.73 $\pm$ 20.43	0.993	0.996
10–15 min_70%	46.72 $\pm$ 27.07	46.93 $\pm$ 28.36	0.992	0.993

Abbreviations: Poincaré plot SD2 index (SD2); estimated SD2 (ESD2); intra-class correlation (ICC); graded exercise test (GXT); maximum aerobic speed (MAS).

Figure 1.

Relationship between SD2 and SDNN with all the data measured ( $n=$ 143).

In the study, the stress score was calculated using the following calculation (1/SD2*1000) [9]. To obtain the estimated stress score (ESS), this work applied the equation obtained from the relationship between SD2 and SDNN measured in all conditions ( $y=$ 1.232 $x$ $+$ 1.915) (Fig. 1); this gave an estimated SD2 (ESD2) from SDNN. The ESS was calculated from the obtained ESD2 (1/ESD2*1000).

2.5 Statistical analysis

The study analysed the statistics using the software IBM SPSS statistics 27 (International Business Machines Corporation, New York, USA) and presented all data as mean $\pm$ standard deviation for each variable. The Shapiro-Wilk test analysed the normal distribution of the data. The linear regression model described the relationship between the variables and provided the equations that defined each relationship. The Pearson coefficient for correlations was also used for all variables. To determine the agreement and consistency of the measurements, the intraclass correlation coefficient (ICC) was calculated. Reliability between measured and estimated variables was assessed using Bland-Altman analysis. Bias, limits of agreement (LoA) and standard deviation of difference (SD) were calculated [26]. The significance value was $p<$ 0.05 for each statistical analysis.

3. Results

Table 2
Descriptive data from SS and SS estimated (ESS) in all conditions ( $n=$ 143)

Test	SS	ESS	ICC	Pearson
Rest	14.46 $\pm$ 3.91	14.44 $\pm$ 3.89	0.971	0.971
GXT	–	–	–	–
0–5 min_GXT	64.55 $\pm$ 33.27	61.08 $\pm$ 28.69	0.984	0.994
5–10 min_GXT	43.04 $\pm$ 21.66	43.72 $\pm$ 19.9	0.995	0.998
10–15 min_GXT	37.78 $\pm$ 20.76	38.43 $\pm$ 19.41	0.996	0.998
Rest	16.71 $\pm$ 4.84	19.97 $\pm$ 4.58	0.967	0.968
70% MAS	48.86 $\pm$ 35.71	43.86 $\pm$ 33.42	0.995	0.997
0–5 min_70%	42.57 $\pm$ 22.14	41.99 $\pm$ 20.45	0.994	0.997
5–10 min_70%	32.27 $\pm$ 14.23	32.91 $\pm$ 14.26	0.997	0.997
10–15 min_70%	28.55 $\pm$ 14.69	28.93 $\pm$ 15	0.997	0.997

Abbreviations: Stress score (SS); estimated stress score (ESS); intra-class correlation (ICC); maximum aerobic speed (MAS).

Figure 2.

Relationship between SD2 and estimated SD2 (ESD2) with data from all conditions ( $n=$ 143).

Table 1 describes the measured SD2 and ESD2 (mean $\pm$ SD) in each condition. It also shows the agreement and relationship between the variables. Table 2 summarises the SS and ESS (mean $\pm$ SD) in each condition and the agreement and relationship between the variables. Figure 1 shows a plot of the linear relationship between SD2 and SDNN with all data measured ( $r$ 2 $=$ 0.986). Figure 2 describes a positive linear relationship between SD2 and ESD2 with all data measured in all conditions ( $r$ 2 $=$ 0.986). SS and ESS show a positive linear relationship with data measured in all conditions ( $r$ 2 $=$ 0.993) (Fig. 3). Figure 4 shows the agreement between SS and ESS using a Bland-Altman plot analysis. Bias was 0.15 $\pm$ 2.54 (UperLOA: 2.54; LowerLOA: $-$ 2.23).

Figure 3.

Relationship between Stress Score (SS) and estimated Stress Score (SSE) with data from all conditions ( $n=$ 143). Identity line is represented in a dotted line.

Figure 4.

Bland-Altman plot ( $n=$ 143). The limits of agreement are represented in dotted lines ( $-$ 1.96s to $+$ 1.96s).

4. Discussion

The main finding of this study is that SS can be derived from SDNN values during exercise. The data show that there is a linear relationship between SD2 and SDNN. Thus, these results are on the same line as the work published by [9] under resting conditions. This relationship suggests that SDNN is not only an indicator of parasympathetic activity, but also represents the reversal of sympathetic activity.

This finding provides an opportunity to estimate SD2 from SDNN. That is, SS, which requires the SD2 index for its calculation, can be estimated from the estimate of SD2 by SDNN. This assumption is confirmed by comparing SS and ESS because the agreement between the two variables is high, with a mean difference of 0.15 $\pm$ 2.54 SS points (Fig. 4).

SS has been demonstrated a good physiological internal load marker. The SS value is a reliable indicator of exercise load regulation [8, 9, 15, 16, 18, 19, 20, 21]. However, nowadays it is difficult to measure it in a practical way from the most usual apps. Our results enable the measurement of SS using SDNN. This new approach offers athletes and coaches the opportunity to incorporate SS into their daily routine. These results could have an important practical application, as most devices and apps do not include a non-linear HRV index to monitor exercise load. In contrast, linear parameters such as SDNN or RMSSD are usually available [5, 23, 24]. Moreover, this study proved that estimation of SS is possible under three different conditions: rest, exercise, and post-exercise recovery. The correlations were statistically significant under all conditions, so the estimation of SS by SDNN measurement is valid regardless of training intensity and time of determination.

In conclusion, the present study shows that the relationship between SDNN and SD2 is constant not only at rest, but also during and after exercise. This finding provides a new approach for the use of SS by athletes and coaches during their training program. Thanks to this new approach, the SS index could be easily measured using actual smart watch and apps.

Author contributions

CONCEPTION: Salazar-Martínez E., Naranjo Orellana J., Sarabia-Cachadiña E.

PERFORMANCE OF WORK: Salazar-Martínez E., Naranjo Orellana J., Sarabia-Cachadiña E.

INTERPRETATION OR ANALYSIS OF DATA: Salazar-Martínez E.

PREPARATION OF THE MANUSCRIPT: Salazar-Martínez E., Sarabia-Cachadiña E.

REVISION FOR IMPORTANT INTELLECTUAL CONTENT: Naranjo Orellana J.

SUPERVISION: Salazar-Martínez E., Naranjo Orellana J., Sarabia-Cachadiña E.

Ethical approval

The Ethics Committee of the University Study Centre-Cardenal Spínola CEU approved the present study in 4-5-2021. Procedures were carried out in concordance with the Declaration of Helsinki. All subjects were informed before the study and signed a written informed consent form in accordance with the Declaration of Helsinki.

Footnotes

Acknowledgments

The investigators would like to thank the sport students whose cooperation made this study possible and Carlos Muriel for his assistance.

Conflict of interest

The authors have no conflicts of interest to report.

References

Malik

. Heart rate variability. Standards of measurement, physiological interpretation, and clinical use. Eur Heart J. 1996; 17(3): 354-81.

Shaffer

Ginsberg

. An overview of heart rate variability metrics and norms. Front Public Health. 2017; 5: 258.

Coyne

Coutts

Newton

Haff

. Training load, heart rate variability, direct current potential and elite long jump performance prior and during the 2016 olympic games. J Sports Sci Med. 2021; 20(3): 482-91.

Flatt

Howells

. Effects of varying training load on heart rate variability and running performance among an Olympic rugby sevens team. J Sci Med Sport. 2019; 22(2): 222-6.

Hernando

Roca

Sancho

Alesanco

Bailón

. Validation of the apple watch for heart rate variability measurements during relax and mental stress in healthy subjects. Sensors (Basel). 2018; 18(8).

Javaloyes

Sarabia

Lamberts

Moya-Ramon

. Training prescription guided by heart rate variability in cycling. Int J Sports Physiol Perform. 2018; 1-28.

Nakamura

Antunes

Nunes

Costa

Esco

Travassos

. Heart rate variability changes from traditional vs. Ultra-Short-Term Recordings in Relation to Preseason Training Load and Performance in Futsal Players. J Strength Cond Res. 2020; 34(10): 2974-81.

Naranjo Orellana

De la Cruz

Sarabia

De Hoyo

Domínguez-Cobo

. Heart rate variability: A follow-up in elite soccer players throughout the season. Int J Sports Med. 2015; 36(11): 881-6.

Naranjo Orellana

de la Cruz Torres

Sarabia Cachadiña

de Hoyo

Domínguez Cobo

. Two new indexes for the assessment of autonomic balance in elite soccer players. Int J Sports Physiol Perform. 2015; 10(4): 452-7.

10.

Piatrikova

Willsmer

Altini

Jovanović

Mitchell

LJG

Gonzalez

, et al. Monitoring the heart rate variability responses to training loads in competitive swimmers using a smartphone application and the banister impulse-response model. Int J Sports Physiol Perform. 2021; 16(6): 787-95.

11.

Natarajan

Pantelopoulos

Emir-Farinas

Natarajan

. Heart rate variability with photoplethysmography in 8 million individuals: a cross-sectional study. Lancet Digit Health. 2020; 2(12): e650-e7.

12.

Georgiou

Larentzakis

Khamis

Alsuhaibani

Alaska

Giallafos

. Can wearable devices accurately measure heart rate variability? A Systematic Review. Folia Med (Plovdiv). 2018; 60(1): 7-20.

13.

KHC

White

Tipoe

Liu

Wong

Jesuthasan

, et al. The current state of mobile phone apps for monitoring heart rate, heart rate variability, and atrial fibrillation: Narrative review. JMIR Mhealth Uhealth. 2019; 7(2): e11606.

14.

Hsu

Tsai

Huang

Lin

Chen

, et al. Poincaré plot indexes of heart rate variability detect dynamic autonomic modulation during general anesthesia induction. Acta Anaesthesiol Taiwan. 2012; 50(1): 12-8.

15.

Abellán-Aynés

López-Plaza

Alacid

Naranjo-Orellana

Manonelles

. Recovery of heart rate variability after exercise under hot conditions: The effect of relative humidity. Wilderness Environ Med. 2019; 30(3): 260-7.

16.

García Bermejo

De La Cruz Torres

Naranjo Orellana

Albornoz Cabello

. Autonomic responses to ultrasound-guided percutaneous needle electrolysis: Effect of needle puncture or electrical current? J Altern Complement Med. 2018; 24(1): 69-75.

17.

Hoyos-Flores

Rangel-Colmenero

Alonso-Ramos

García-Dávila

Cruz-Castruita

Naranjo-Orellana

, et al. The role of cholinesterases in post-exercise hrv recovery in university volleyball players. Applied Sciences. 2021; 11(9).

18.

Muñoz-López

Naranjo-Orellana

. Individual versus team heart rate variability responsiveness analyses in a national soccer team during training camps. Sci Rep. 2020; 10(1): 11726.

19.

Muñoz-López

Nakamura

Naranjo Orellana

. Soccer matches but not training sessions disturb cardiac-autonomic regulation during national soccer team training camps. Res Q Exerc Sport. 2021; 92(1): 43-51.

20.

Nieto-Jiménez

Pardos-Mainer

Ruso-Álvarez

Naranjo-Orellana

. Training load and HRV in a female athlete: a case study. Revista Internacional de Medicina y Ciencias de la Actividad Física y el Deporte. 2020; 20(78): 321-33.

21.

Rangel-Colmenero

Hoyos-Flores

Hernández-Cruz

Miranda-Mendoza

González-Fimbres

Reynoso-Sánchez

, et al. Comportamiento de las colinesterasas tras condiciones de fatiga en corredores de fondo. Revista Internacional de Medicina y Ciencias de la Actividad Física y del Deporte. 2022; 22(85): 35-45.

22.

Miranda-Mendoza

Reynoso-Sánchez

Hoyos-Flores

Quezada-Chacón

Naranjo Orellana

Rangel-Colmenero

, et al. Stress score and Lnrmssd as internal load parameters during competition. Revista Internacional de Medicina y Ciencias de la Actividad Física y el Deporte. 2020; 20: 21-35.

23.

Flatt

Esco

. Smartphone-derived heart-rate variability and training load in a women’s soccer team. Int J Sports Physiol Perform. 2015; 10(8): 994-1000.

24.

Flatt

Esco

. Evaluating individual training adaptation with smartphone-derived heart rate variability in a collegiate female soccer team. J Strength Cond Res. 2016; 30(2): 378-85.

25.

Cerezuela-Espejo

Courel-Ibáñez

Morán-Navarro

Martínez-Cava

Pallarés

. The relationship between lactate and ventilatory thresholds in runners: Validity and reliability of exercise test performance parameters. Front Physiol. 2018; 9: 1320.

26.

Giavarina

. Understanding Bland Altman analysis. Biochem Med (Zagreb). 2015; 25(2): 141-51.

Heart rate variability: Obtaining the stress score from SDNN values

Abstract

BACKGROUND:

OBJECTIVE:

METHODS:

RESULTS:

CONCLUSIONS:

Keywords

1. Introduction

2. Material and methods

2.1 Participants

2.2 Design

2.3 Measurements

2.4 Stress score and estimated stress score

Table 1 Descriptive data from SD2 and SD2 estimated (ESD2) in all conditions

3. Results

Table 2 Descriptive data from SS and SS estimated (ESS) in all conditions ( n = 143)

Author contributions

Ethical approval

Footnotes

Acknowledgments

Conflict of interest

References

Table 1
Descriptive data from SD2 and SD2 estimated (ESD2) in all conditions

Table 2
Descriptive data from SS and SS estimated (ESS) in all conditions ( $n=$ 143)