Article Text
PDF
Abstract
Introduction The present study determined the (1) day-to-day reliability of basal heart rate (HR) and HR variability (HRV) measured by the Equivital eq02+ LifeMonitor and (2) agreement of ultra short-term HRV compared with short-term HRV.
Methods Twenty-three active-duty US Army Soldiers (5 females, 18 males) completed two experimental visits separated by >48 hours with restrictions consistent with basal monitoring (eg, exercise, dietary), with measurements after supine rest at minutes 20–21 (ultra short-term) and minutes 20–25 (short-term). HRV was assessed as the SD of R–R intervals (SDNN) and the square root of the mean squared differences between consecutive R–R intervals (RMSSD).
Results The day-to-day reliability (intraclass correlation coefficient (ICC)) using linear-mixed model approach was good for HR (0.849, 95% CI: 0.689 to 0.933) and RMSSD (ICC: 0.823, 95% CI: 0.623 to 0.920). SDNN had moderate day-to-day reliability with greater variation (ICC: 0.689, 95% CI: 0.428 to 0.858). The reliability of RMSSD was slightly improved when considering the effect of respiration (ICC: 0.821, 95% CI: 0.672 to 0.944). There was no bias for HR measured for 1 min versus 5 min (p=0.511). For 1 min measurements versus 5 min, there was a very modest mean bias of −4 ms for SDNN and −1 ms for RMSSD (p≤0.023).
Conclusion When preceded by a 20 min stabilisation period using restrictions consistent with basal monitoring and measuring respiration, military personnel can rely on the eq02+ for basal HR and RMSSD monitoring but should be more cautious using SDNN. These data also support using ultra short-term measurements when following these procedures.
- sports medicine
- physiology
- cardiology
Data availability statement
Data are available upon reasonable request.
This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.
Statistics from Altmetric.com
WHAT IS ALREADY KNOWN ON THIS TOPIC
The cardiovascular system is a frequently studied target for physiological monitoring.
WHAT THIS STUDY ADDS
Following restrictions consistent with basal monitoring and a 20 min stabilisation period, basal heart rate and RMSSD (an index of heart rate variability) exhibited good day-to-day reliability when measured by the Equivital eq02+ LifeMonitor over 5 min.
Adjusting the model to include the influence of respiratory measurements collected with a metabolic cart slightly improved the day-to-day reliability of RMSSD measurement.
Ultra short-term measurement (1 min) of RMSSD has a very modest negative bias compared with short-term measurements taken over 5 min.
HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY
This study supports that military personnel can rely on the eq02+ for basal HR and RMSSD monitoring and supports the use of 1 min measurements of basal HR and RMSSD provided that users follow the same restrictions and 20 min stabilisation period used in the present study.
Introduction
The cardiovascular system is frequently studied for physiological monitoring due to the relative ease of employing non-invasive wearable technology.1 2 Heart rate variability (HRV), which describes the beat-to-beat fluctuations of the heart (ie, R–R interval), is often used to evaluate overtraining and under-recovery.3 Assessment of HRV can be performed using 24 hours, short-term (5 min) and ultra short-term (<5 min) time periods.3 4 These measurements reflect different underlying physiology and are not interchangeable. For example, HRV measured over 24 hours is the clinical gold standard assessment and is influenced by circadian rhythm, body temperature, metabolism, sleep and the renin–angiotensin–aldosterone system.5 Short-term HRV is the result of four interdependent processes: (1) the net effect of the interaction between sympathetic and parasympathetic neural outflow; (2) respiratory sinus arrythmia, which is described by the vagally mediated alterations in heart rate (HR) due to respiration; (3) baroreflex-mediated changes in cardiac control; and (4) adjustments in vascular tone.6 7 The current literature suggests that 60 s is the minimum duration needed for ultra short-term time-domain based analyses as a proxy for short-term HRV, whereas more than 60 s is required for frequency-domain and non-linear analyses.6
Periodic HRV assessment is likely insufficient for practical use in certain fields,8 9 so it is likely that near daily assessment is needed to characterise ongoing HRV status.10 If the basal status of military personnel is to be characterised by HRV assessment, then it is critical that the devices and methodology used by the military provide reliable data. Understanding the day-to-day reliability of obtaining HRV in controlled settings (ie, reducing potential confounders), where HRV values would likely be more consistent, with devices currently used by military personnel is a needed first step for this purpose.
The Equivital eq02+ LifeMonitor (Cambridge, UK) is a commercially available physiological monitoring system that includes a 2-lead ECG sampling at 256 Hz for measurement of R–R interval and was recently validated for HR measurement during intermittent exercise.11 The previous system, the eq02, provided HRV measurements that were highly correlated with gold standard continuous ECG measurement during periods of low activity and low artefact.12 The day-to-day reliability of basal HRV measurements taken with the modern system, the eq02+, has not been studied. The agreement between ultra short-term and short-term HRV as measured by the eq02+ is also not known. Therefore, the purpose of the present study was to (1) determine the day-to-day reliability of basal HR and HRV assessed by the Equivital eq02+ LifeMonitor in US Army soldiers and (2) determine the agreement of ultra short-term HRV (ie, 1 min average) compared with the traditional short-term duration (ie, 5 min average) using this system. The data presented herein were pooled from two larger studies, one previously published13 and the other unpublished, that used the same procedures for a comprehensive baseline assessment performed on separate days.
Materials and methods
Participants
Twenty-three healthy, active-duty US Army soldiers (sex assigned at birth: 5 females and 18 males) participated in this study (table 1). Participants were included if their age was 18–44 years and were at least recreationally active (≥ 2 days per week of aerobic or resistance exercise for ≥30 min). Due to procedures in the larger studies, participants were excluded if they had any medical condition or injury that compromised their ability to exercise, body mass >129.3 kg (due to treadmill weight limit specifications), pregnant or any history of gastrointestinal diseases, conditions or surgeries (eg, diverticulitis, appendectomy).
Participant characteristics
Experimental protocol
The experimental protocol was designed to meet recommended criteria for measuring basal metabolic rate.14 Participants were familiarised with all equipments. Participants completed two experimental trials separated by >48 hours and arrived at the lab after abstaining from high-intensity exercise for >48 hours, alcohol and vigorous exercise for >24 hours and caffeine, nicotine and food intake for >10 hours. Participants were instructed to drink ≥500 mL of water both the night before and the morning of each visit. All participants arrived at the lab at the same time for a given individual (06:00–09:00 hours). These procedures were necessary to ensure similarity in conditions during preceding days prior to the lab visits.
On arrival, participants confirmed adherence to the protocol restrictions. Participants provided ratings of subjective sleep quality from the previous night and the past 7 days via a 11-point scale where 0=terrible, 1–3=poor, 4–6=fair, 7–9=good and 10=excellent. Participants provided a urine sample to assess hydration status via urine-specific gravity by refractometry (Goldberg TS Meter, Reichert, Buffalo, New York, USA). On visit 1, only height was measured using a stadiometer (Model 213, Seca, Hamburg, Germany). On both visits, nude body mass was measured with standard scale (Model 876, Seca, Hamburg, Germany). On visit 2, only body composition was assessed by dual-energy X-ray absorptiometry (DPX-IQ, Lunar Corporation, Madison, Wisconsin, USA).
After baseline assessment, participants were instrumented with the eq02+ LifeMonitor (Equivital, Cambridge, UK) and additional HR monitor (H10, Polar Electro, Kempele, Finland). Proper fitting of the eq02+ was ensured per manufacturer’s instructions. Technical limitations precluded the ability to validate the eq02+ against the H10 because data from the H10 device were not collected on a beat-to-beat basis (collected as 5 s averages with the metabolic cart). Participants were outfitted with a mask to collect expired gases for continuous respiratory measurements with a calibrated metabolic cart (TrueOne 2400, ParvoMedics, Sandy, Utah).
Participants laid in supine rest on a cot in a dimly lit room, while remaining awake and while continuous expired gas measurements were taken. Due to the influence of body posture on integrated cardiovascular control, a 20 min period of supine rest before measurements was taken to allow stabilisation of blood pressure and body fluid compartmental shifts.15 16 After 20 min supine rest, HRV was assessed during spontaneous breathing (ie, participants continued to breathe normally) at minutes 20–25 (short-term) and minutes 20–21 (ultra short-term). Due to the relation between cardiorespiratory fitness and HRV,17 18 data from the maximal oxygen uptake (V̇O2max) testing from the larger studies, which occurred after supine measurements,13 19 are included in the present manuscript to provide context for the cardiorespiratory fitness of the participants.
Data analysis
The interbeat interval from minutes 20–25 of supine rest was exported from the eq02+ sensor electronics module with the Equivital Manager software (V.2.11.1.272, Hidalgo Limited, Cambridge, UK). Data were processed in R (V.4.3.1, Vienna, Austria) to time align these data with the metabolic cart data. We were unable to confidently identify ectopic beats because we did not have continuous ECG recordings. Therefore, data were neither cleaned nor smoothed, and data were not removed or filtered (ie, all data are included in the results). Short-term HRV data were binned as a 5 min average over this period and ultra short-term data were taken at minute 20. Time domain analyses were used to determine the square root of the mean squared differences between consecutive R–R intervals (RMSSD) and the SD of R–R intervals (SDNN). SDNN provides an estimate of the overall HRV and RMSSD provides insight into cardiac parasympathetic activation.4 Time domain analyses were used to quantify HRV because it is the only recommended HRV analysis over a 60 s duration.6 Additionally, there is some evidence supporting that RMSSD is less affected by respiration and may be a better proxy of HRV during spontaneous breathing.20–22
Statistical analysis
Data were analysed using R. Data were pooled from two existing datasets and therefore, an a priori power analysis for sample size determination was not conducted. After confirming normality of data with Shapiro-Wilk test, two-tailed paired t-tests were used to assess differences between visits 1 and 2 in ambient temperature, relative humidity, sleep quality, urine-specific gravity, respiratory rate and minute ventilation, and HRV metrics taken over 5 min periods. The primary analysis for the day-to-day reliability of HR and HRV metrics between visits was performed with intraclass correlation coefficients (ICCs) by calculating the proportion of total variation among individuals with the unadjusted linear mixed-effects model approach with random effects of participants on intercepts23 (Model 1). Additionally, linear mixed-effects models with fixed effects of respiratory rate (Model 2), minute ventilation (Model 3) and both respiratory rate and minute ventilation (Model 4) were used to calculate ICC to assess the influence of respiration om HRV. We also included subjective ratings of overnight sleep quality or weekly sleep quality into the model to investigate the influence of perceived sleep quality on the reliability of basal HR and HRV parameters. ICCs were interpreted by the method of Koo and Li,24 where reliability was considered poor if <0.5, moderate if between 0.5 and 0.75, good if between 0.75 and 0.9 and excellent if >0.9. Bland-Altman plots were used to assess the agreement between HRV metrics taken over 5 min and 1 min periods by including both visits in the analysis.25 To account for repeated measures with this approach, a mixed-effects model was used to calculate bias and 95% limits of agreement (LoA).26 27 As a subanalysis, data were also disaggregated by visit using the method of Bland-Altman. Statistical significance was set at α=0.05. Data are presented as mean with 95% CIs with the exception of table 1, which is presented as mean (SD).
Results
There were no differences in environmental conditions between visits (ambient temperature: visit 1, 22.2°C (95% CI: 21.5 to 22.8); visit 2, 22.4°C (95% CI: 21.9 to 22.9), p=0.194; relative humidity: visit 1, 41.5% (95% CI: 35.4 to 48.7); visit 2, 44.5% (95% CI: 39.8 to 49.1), p=0.075). There were no differences between visits in subjective rating of overnight sleep quality (visit 1: 7 a.u. (95% CI: 6 to 8); visit 2: 7 a.u. (95% CI: 6 to 8), p=0.583) or in sleep quality in the past 7 days (visit 1: 7 a.u. (95% CI: 6 to 8); visit 2: 7 a.u. (95% CI: 6 to 8), p=0.866). Hydration status did not differ between visits (urine-specific gravity: visit 1, 1.018 a.u. (95% CI: 1.014 to 1.022); visit 2, 1.017 a.u. (95% CI: 1.012 to 1.021), p=0.307). There were no differences between visits in respiratory rate (visit 1: 15 breaths/min (95% CI: 13 to 16); visit 2: 14 breaths/min (95% CI: 13 to 15), p=0.172) or minute ventilation (visit 1: 6.6 L/min (95% CI: 6.0 to 7.2); visit 2: 6.9 L/min (95% CI: 6.2 to 7.5), p=0.359).
Between-visit reliability of 5 min measurements
Basal HR was lower by two bpm (95% CI: −4 to –1) on visit 2 compared with visit 1 (figure 1A). SDNN did not differ between visits (figure 1B), but RMSSD was higher by 14 ms (95% CI: 3 to 25) on visit 2 compared with visit 1 (figure 1C). HR and RMSSD had good day-to-day reliability, whereas the reliability of SDNN was moderate (table 2). SDNN and RMSSD had less certainty in the precision of the mean for ICC compared with HR given the wide range of the 95% CIs (table 2). When considering the influence of respiration, there was no change in the reliability of basal HR, but the reliability of SDNN was further reduced (table 2). In contrast, the reliability of RMSSD slightly improved when considering respiration, which was largely due to minute ventilation as a fixed effect (table 2). Overall, the reliability of the day-to-day variation in basal HR was slightly improved when including a fixed effect of overnight sleep quality, whereas the reliability of SDNN and RMSSD was slightly worsened (Models 1–4, table 2). Including rating of weekly sleep quality did not meaningfully alter the ICC of basal HR but led to slight reductions in the reliability for SDNN and RMSSD as compared with overnight sleep quality (Models 1–4, table 2).
Between-visit comparisons of heart rate variability metrics taken over a 5 min period following 20 min supine rest. Basal heart rate (A), the SD of R–R intervals (SDNN, B) and the square root of the mean squared differences between consecutive R–R intervals (RMSSD, C) were analysed by two-tailed paired t-test. Mean data with 95% CIs (large symbols with error bars) are presented with individual data points. n=23 (5 females and 18 males).
Between-visit ICCs)
Agreement between 1 min and 5 min measurements
When visits were combined and repeated measures were accounted for statistically, there was no bias between basal HR measured for 1 min and 5 min (0 bpm; 95% CI: −1 to 1; 95% LoA: −4 to 5; p=0.511, figure 2A). SDNN measured for 1 min had a bias of −4 ms (95% CI: −5 to –2; 95% LoA: −15 to 7) compared with 5 min (p<0.0001, figure 2B) and RMSSD measured for 1 min had a bias of −1 ms (95% CI: −2 to 0; 95% LoA: −7 to 4) compared with 5 min (p=0.023, figure 2C). When data were disaggregated by visit, there were similar trends compared with the pooled analysis and between visits for data on visit 1 (HR—bias: 1; 95% CI: 0 to 2; 95% LoA: −4 to 6; SDNN—bias: −4; 95% CI: −6 to –1; 95% LoA: −14 to 7; RMSSD—bias: −1; 95% CI: −2 to 0; 95% LoA: −6 to 4) and visit 2 (HR—bias: 0; 95% CI: −1 to 1; 95% LoA: −5 to 4; SDNN—bias: −4; 95% CI: −7 to –1; 95% LoA: −16 to 8; RMSSD—bias: −1; 95% CI: −2 to 0; 95% LoA: −7 to 5).
Bland-Altman plots to compare the difference between 1 min and 5 min measurements of basal heart rate (A), the SD of R–R intervals (SDNN, B) and the square root of the mean squared differences between consecutive R–R intervals (RMSSD, C). Bias (solid bold red line) with 95% CIs (light red shading) was calculated using a mixed-effects model to account for repeated measures between visit 1 and visit 2. The 95% limits of agreement (solid black line) with 95% CIs (grey shading) are presented. n=23 (5 females and 18 males).
Discussion
There were several novel findings of the present study. First, when preceded by a 20 min stabilisation period using restrictions consistent with basal monitoring, basal HR and RMSSD measured for 5 min with the eq02+ exhibited good day-to-day reliability, while the reliability of SDNN of basal HRV was, at best, moderate given the imprecise 95% CIs. Second, the reliability of basal RMSSD was only slightly improved when considering respiration in the model. Third, when comparing 1 min versus 5 min measurements within the same visit, the eq02+ had no bias in basal HR, but there was a very modest, yet statistically significant, bias between 1 and 5 min measurements of SDNN (~−4 ms) and RMSSD (~−1 ms). Collectively, these findings indicate that the basal HR and RMSSD measurements are reliable with the eq02+ when following basal monitoring procedures, but the reliability of quantifying SDNN is less certain.
In the present study, we employed a well-controlled experimental design and found that the eq02+ provides a reliable day-to-day measure of basal HR in US Army soldiers when using the controls employed in the experimental protocol. The day-to-day reliability of the HR measurements in the present study was higher compared with previous studies assessing resting HR,28 which could be due to a number of factors, including population studied and differences in the experimental approach (eg, previsit restrictions, timed supine rest). We also found that basal HR was ~2 bpm lower on visit 2 compared with visit 1. Despite the extensive controls used in the present study and that participants were familiarised with all study procedures and equipment, additional familiarisation in future studies with primary outcomes related to basal HR may be warranted.
We found that the reliability of the basal HRV was moderate-to-good depending on the outcome variable, but the substantial variation noted in the 95% CIs for SDNN reduced the reliability. The SDNN and RMSSD values derived from the eq02+ in the present study were similar to values we have obtained using 3-lead ECG sampling at 1000 Hz during spontaneous and paced breathing protocols.29 The reliability of basal RMSSD in the present study was slightly improved when considering respiration (Model 4), which was largely driven by ventilation (Model 3) compared with respiratory rate (Model 2). In contrast, the reliability of basal HR appears to be independent of respiration, whereas the reliability of basal SDNN is reduced when considering respiration in the model. These data add to a large body of literature that respiration is an important consideration when measuring basal short-term HRV. However, the marginal improvement in the reliability of basal RMSSD when including respiration suggests that military personnel can rely on the eq02+ as having good reliability whether adjustments for respiration are made. Paced breathing protocols, such as to a metronome at 0.25 Hz,29 add further complexity, although relatively minor, to protocols assessing basal HRV that already include many restrictions for participants and may reduce practicality. Given that the reliability of RMSSD was only slightly improved with the inclusion of respiration, our interpretation is that the reliability of RMSSD as a basal short-term HRV measurement is good both with and without consideration of respiration, but caution may be warranted because of the relatively wide-ranging 95% CIs in both Models 1 and 4.
We also found a very modest bias for lower SDNN and RMSSD when measured for 1 min compared with 5 min, suggesting that 1 min measurements provided sufficient resolution to capture basal HRV. These data are supported by previous findings in military trainees where 1 min time domain-based analyses had good agreement with a 5 min period.10 In dynamic settings, HRV is influenced by other factors (eg, ambulation) that were mitigated by using a controlled basal approach and thus, the current data should not be interpreted as support for the use of short-term or ultra short-term HRV assessment in active transient states. Translating the findings of the present study into practice will require further research, such as with assessing the recovery from physical and mental stress in basic training, which may be a vulnerable state with heightened risk for suboptimal performance or a window of opportunity that can be leveraged to enhance training adaptations.30 Further research should determine whether these restrictions or the length of the stabilisation period can be reduced.
A discussion of the disadvantage of analysing R–R intervals without continuous ECG tracings is warranted. The data exportation process did not allow for the identification or correction of ECG tracings with artefacts or to discern ectopic beats (ie, irregular beats). Therefore, it was not possible to differentiate whether artefact, if existent, was due to signal noise or potential cardiac arrhythmia. Additionally, the technical details of the signal processing are only known by the manufacturer. Thus, similar to other devices (eg, H10), the influence of sampling frequency, resolution or preprocessing algorithms used to filter background noise on R–R interval measurement is not known.31 The data exportation process used in the present study was considered more feasible for large-scale deployment. Future studies are warranted to determine the reliability of HR and HRV metrics using different software packages offered by the manufacturer for the eq02+ (eg, eqView Professional) and other devices for HRV assessment. The SDNN and RMSSD values obtained in the present study were comparable to values that our lab has observed using gold standard techniques.29 Therefore, it does not appear that there was a substantial effect on the HRV metrics in the present study, which may be due to having a relatively homogeneous sample population and rigorous experimental controls. Future studies in more diverse study populations are warranted to determine the day-to-day reliability measuring basal HR and HRV over weeks or months.
Conclusion
When preceded by a 20 min stabilisation period using restrictions consistent with basal monitoring, the eq02+ has good day-to-day reliability for basal HR and RMSSD measurements but moderate, at best, day-to-day reliability for SDNN. Including respiration as a fixed effect slightly improved the reliability of basal RMSSD but reduced the reliability of SDNN, while not altering the reliability of basal HR. Ultra short-term measurements (1 min) of SDNN and RMSSD have a very modest negative bias compared with 5 min measurements. Military personnel can rely on the eq02+ for basal HR and RMSSD monitoring when assessing basal status but should be more cautious when interpreting SDNN measurements. Additionally, using ultra short-term measurement of basal HR and RMSSD is supported by this study provided that military personnel follow the same experimental restrictions employed in this protocol.
Data availability statement
Data are available upon reasonable request.
Ethics statements
Patient consent for publication
Ethics approval
This study was approved by the Institutional Review Board (IRB protocols numbers M-10803 and M-11018) at the US Army Medical Research and Development Command (Fort Detrick, MD). Participants gave informed consent to participate in the study before taking part.
Acknowledgments
We thank the volunteers who participated in the study and Dr Nisha Charkoudian for feedback on the manuscript. Approved for public release; distribution is unlimited. The opinions or assertions contained herein are the private views of the author(s) and are not to be construed as official or reflecting the views of the Army or the Department of Defense. Any citations of commercial organisations and trade names in this report do not constitute an official Department of the Army endorsement of approval of the products or services of these organisations. This research was supported in part by an appointment to the Department of Defense (DOD) Research Participation Program administered by the Oak Ridge Institute for Science and Education (ORISE) through an interagency agreement between the US Department of Energy (DOE) and the DOD. ORISE is managed by ORAU under DOE contract number DE-SC0014664. All opinions expressed in this paper are the author’s and do not necessarily reflect the policies and views of DOD, DOE or ORAU/ORISE.
References
Footnotes
Contributors Conceptualisation: CLC, AWP, BMR, JWC, KEF, DPL. Data curation: CLC, EAS, EML, DPL. Formal analysis: CLC, DPL. Funding acquisition: DPL, JWC. Investigation: CLC, EAS, EML, BMR, DPL. Methodology: DPL, AWP, JWC. Project administration, resources, software and validation: DPL. Visualisation: CLC. Writing – original draft: CLC and DPL. Writing – review and editing: CLC, EAS, AWP, EML, BMR, JWC, KEF, DPL. Gaurantor is AWP.
Funding This research was funded by US Army Medical Research and Development Command (USAMRDC), Military Operational Medicine Research Program (MOMRP) (MO220031). This research was supported in part by appointments to the Department of Defense (DOD) Research Participation Program administered by the Oak Ridge Institute for Science and Education (ORISE).
Competing interests None declared.
Provenance and peer review Not commissioned; externally peer reviewed.