==== Front J Funct Morphol Kinesiol J Funct Morphol Kinesiol jfmk Journal of Functional Morphology and Kinesiology 2411-5142 MDPI 10.3390/jfmk5030046 jfmk-05-00046 Article Test and Evaluation of Heart Rate Derived Core Temperature Algorithms for Use in NCAA Division I Football Athletes Hagen Joshua 1 Himmler Aaron 2 Clark Joseph 2 Ramadan Jad 1 Stone Jason 1 Divine Jon 2 Mangine Robert 2* 1 Rockefeller Neuroscience Institute, West Virginia University, Morgantown, WV 26505, USA; joshua.hagen@hsc.wvu.edu (J.H.); jramadan@hsc.wvu.edu (J.R.); jason.stone1@hsc.wvu.edu (J.S.) 2 Department of Athletics, University of Cincinnati, Cincinnati, OH 45221, USA; kuehnhav@ucmail.uc.edu (A.H.); joseph.clark@uc.edu (J.C.); divinej@ucmail.uc.edu (J.D.) * Correspondence: manginre@ucmail.uc.edu 06 7 2020 9 2020 5 3 4631 5 2020 01 7 2020 © 2020 by the authors.2020Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).The purpose of this study was to assess the validity of utilizing heart rate to derive an estimate of core body temperature in American Football athletes. This was evaluated by combining commercially available Zephyr Bioharness devices, which includes an embedded estimated core temperature (ECT) algorithm, and an ingestible radio frequency core temperature pill during the highest heat injury risk timepoint of the season, summer training camp. Results showed a concordance of 0.643 and 78% of all data points fell within +/−1.0 °F. When the athletes were split into Upper (>/=6.0%) and Lower (<6.0%) body composition groups, there was a statistical improvement in accuracy with the Upper Body Fat% reaching 0.834 concordance and 93% of all values falling within +/−1.0 °F of the Gold Standard. Results suggest that heart rate derived core temperature assessments are a viable tool for heat stress monitoring in American football, but more work is required to improve on accuracy based on body composition. exertional heat illnessathlete monitoringcore temperature ==== Body 1. Introduction A growing abundance of applied sport science initiatives include strategies for mitigating the deleterious effects of heat stress, which are preventable through the implementation of proper training and effective heat stress monitoring (HSM) protocols. The ramifications of heat stress, such as heat-induced severe cramping, edema, rhabdomyolysis, or heat strokes to name a few, are often characterized as exertional heat illnesses (EHI) [1,2,3,4], and arise when homeostasis cannot maintain a core body temperature beneath 40.5 °C (considered as excessive elevations in core body temperature during physical exertion). This process is primarily achieved through constant heat exchange between the human body and the ambient environment via evaporation, radiation, convection, and conduction [2]. Athletes, whom inherently endure bouts of intense and voluminous physical exertion throughout their training cycles, rely on these thermoregulatory processes to counterbalance inevitable elevations in core body temperature as a result of metabolic heat production, a necessary by-product of exercise. However, if an athlete’s physiological systems, namely their cardiovascular and nervous systems as well as their skin, are unable to prevent excess elevations in core body temperature then they are at a severe risk for injury, or in the most severe cases, fatality [2,5]. A position statement on exertional heat illnesses in collegiate athletes released in 2015 from the National Athletic Trainers Association related core temperatures exceeding 40.5 °C with exertional heat stroke [2], a catastrophic event for athletes that must be avoided at all costs. While not all heat illnesses are temperature-mediated, it is worth noting that exertional heat strokes are among the leading causes of sudden death in collegiate athletes [6]. Fortunately, heat stress morbidity and mortality are largely avoidable with proper HSM protocols. In general, HSM strategies are designed to ensure the health and well-being of athletes, especially during training periods comprising bouts of high volume or intensity (often times training is a combination of both) conducted under extreme environmental conditions, such as high humidity and/or ambient temperatures. For example, American football is one (but not the only) sport that comprises high magnitudes of physical workload, often times performed in harsh environments [7,8,9]. Prior to each competitive season, football athletes participate in summer training camp, which imposes tremendous amounts of fatigue on the athletes and likely occurs during some of the hottest days within a calendar year [9,10,11]. In fact, the National Collegiate Athletic Association (NCAA) recently made drastic alterations to the practice specifications (e.g., number of practices per day, amount of collision contact) permitted during preseason training camp periods. The NCAA’s deliberate attempt to mitigate EHI is a result of the increased awareness surrounding EHI in athletes [1,2,6,12], which ultimately stem from tragic events that saturated mainstream media including but not limited to the deaths of pro-bowl offensive tackle, Korey Stringer, in August of 2001 and University of Maryland football athlete, Jordan McNair, who sustained an EHI in May 2018, and succumbed in June 2018. Both athletes suffered exertional heat strokes during the summer months. Moreover, the Coronavirus 2 (COVID-19) pandemic that halted the entire global sports domain in March 2020 and prevented training and monitoring protocols to progress normally poses a new and unique challenge to the sports performance realm as it is likely that athlete training statuses were negatively impacted [13]. Indeed, athletes are at increased risk for EHI when first returning to practice following a relatively long “break”, such as the initial summer months following the spring semester for collegiate fall athletes [3,4], which further exposes the importance of expeditiously incorporating HSM strategies into athlete training and recovery. As previously alluded to, EHIs are largely preventable with careful considerations applied to HSM protocols that are implemented and refined throughout the various training periods (e.g., preseason, offseason, competitive season). Yet, typical procedures garnering research quality data that measure core body temperature comprise invasive and uncomfortable instruments, such as rectal or esophageal temperature probes, that are impractical for routine use in athletics. Thus, efforts to develop more practical approaches for measuring core temperatures ensued [14,15,16]. However, concerns regarding the validity and reliability for many of these instruments, such as axillary, oral, or tympanic sensors are questioned throughout the extant literature [16,17,18,19,20]. One alternative strategy utilizes telemetric pills that are swallowed by athletes and transported through their digestive system within 24–48 h [18,19,21]. Telemetric pills were previously validated as an acceptable instrument for quantifying core body temperature in humans [22,23], although limitations for daily or routine use beyond unit costs do exist. Unfortunately for many athletes, when telemetric pills are utilized, data collection typically requires the donning of a data logging device, which is impractical for contact sports such as American football, rugby, or mixed martial arts. Alternatively, remote transmissions of data are sometimes possible, which likely still necessitates that at least one human data recorder (presumably need much more for large groups of athletes such as American football teams) meticulously and manually records data throughout any given practice or training session. Furthermore, previous research reported that the accuracy of telemetric pills is sensitive to fluid ingestion [18] and ingestion timing [19], posing unique challenges for athletic groups that are constantly hydrating to prevent EHIs and other harmful effects of dehydration. Interestingly, recent investigation of the utilization of serial raw heart rate signals to estimate core body temperature during ambulation reported an impressive overall bias of −0.03 ± 0.32 °C when compared to an ingested telemetric pill [14], which was later replicated in a similar model conducted in first responders [24]. The estimated core temperature (ECT) algorithm is based on a Kalman filter model using a series of heart rate (HR) measurements as the leading indicator; equations and MATLAB code available in Buller et al. [14]. The model contains two relationships, a time update and observational component. The ECT updates in 1-min increments by using updated HR data (time update) to modulate the previous value (observational). As such, the possibility for real time core temperature monitoring in athletes (applicable to military populations as well) through existing live athlete monitoring platforms that already obtain heart rate data garnered intrigue from researchers and practitioners alike [25,26,27]. Still, integrating the predictive algorithm developed through these previous efforts requires additional validations to ensure heart rate data obtained from the respective commercial devices are adequate enough for the algorithm to retain acceptable levels of accuracy. Further, while the published model for ECT [14,24] was rigorously developed and validated, this was done on non-athlete populations and not during sport-specific activities. Specifically, the nature of American Football in practice and game settings requires brief periods of high intensity effort followed by brief periods of rests (e.g., in-between plays) [7,28]. Therefore, the purpose of this study was to evaluate the accuracy of heart rate derived core body temperature estimations, which were obtained from a commercial electrocardiogram (ECG) chest strap using the published ECT algorithm. More specifically, data were collected on American football athletes during the highest risk time period of the season, summer training camp [3,4]. Temperature estimations were compared to a telemetric pill that was ingested by the athletes prior to several team practices. A secondary purpose was to examine potential influences that athlete body composition may impose on estimation accuracy. We hypothesized that ECT calculations would perform within 2 °F with respect to the telemetric pill. The +/−2 °F (1.1 °C) enables a range between 102.9 and 106.9 °F (39.4–41.6 °C, with a midpoint at 40.5 °C) thus allowing the monitoring of temperature ranges indicative of EHI risk [2]. Due to the novelty of this data collection, previous data were unavailable for use in determining a required sample size a priori. However, a post-hoc power analysis was conducted using Bland Altman statistics calculated from the data. Post-hoc power analysis will aid in justifying results presented herein and will be considered for future applications of ECT in sport and research. These data will reveal critical information relevant to athletes and sport practitioners interested in HSM as a strategy for mitigating EHI. 2. Materials and Methods 2.1. Participants A total of 13 male Division I football athletes participated in this study. Demographic information is listed in Table 1. All subjects gave their informed consent for inclusion before they participated in the study. The study was conducted in accordance with the Declaration of Helsinki. This research was covered under the approved IRB Protocol at the University of Cincinnati (#2017-3008, Approval #00003152). The athletes were additionally grouped into “Upper BF%” (>6.0 Body Fat) and “Lower BF%” ( 2; HA2: U ≤ 2. To power this, two values were used [34]: standardized difference limit (the mean of differences divided by the standard deviation of the differences), and the standardized agreement limit (2 divided the standard deviation of the differences). Here, the standardized difference limit is −0.192/0.822 = −0.23, and the standardized agreement limit is 2/0.822 = 2.43. With these two values, a sample size of 402 measurement pairs would be required to achieve 80% power at the 0.05 level for the hypothesis test above. However, when this same post hoc power analysis was run using only the Upper BF% group data, only 14 measurement pairs would be required to achieve 80% power at the 0.05 level. 4. Discussion The primary motivation behind this study was to assess the accuracy of using heart rate data to derive an estimated core body temperature in Division I NCAA football athletes during the highest EHI risk time of the season. When assessed for every participant across all practices, the result was a CCC of 0.643. Statistically with CCC, this falls in the “poor” range, but this translates to 78% of the values falling between +/−1.0 °F. The decision to utilize this technology and method during training comes down to the practitioners, and if they are willing to accept that most data points are within +/− 1.0 °F accuracy to help augment their HSM surveillance process. More interestingly, when the participants were split by body composition into an Upper BF% group and Lower BF% group, the results were very different. The Upper BF% group, defined as above 6% BF with a range between 6% and 18.8%, showed a substantially and statistically significant improvement in CCC with a value of 0.834, compared to 0.515 for the Lower BF% group. This is also reflected in an improvement in percentage of values within +/−1.0 °F accuracy up to 93%, as compared to 62% for the Lower BF% group. The ECT algorithm has been well published, cited, and used in military populations, but not formally investigated for validity in athletics. Currently, this algorithm is only specific to heart rate data, and the model was trained and validated on military personnel. One explanation for the increased accuracy for the Upper BF% group is that the ECT algorithm was trained on participants that were in a higher body composition range, of 13–18% [14], which closer matches the Upper BF% group compared to the Lower BF% group. Physiologically, there are at least two other factors to consider due to body composition differences, surface area and metabolic demand of tissue. Potential differences in athletes’ body mass-to-body surface area ratio (BM/BSA) might suggest that there may be a greater surface area available for evaporative cooling based on upon an single athlete’s anthropometric/body composition makeup [35]. Therefore, future developments on ECT algorithms should take this into account. A second factor to consider comprises the differing metabolic demands of skeletal muscle tissue compared to adipose tissue. Muscle tissue possesses inherently greater metabolic demands, both at rest and during exertion, when compared to fat thus impacting the overall energy expenditure, which should be investigated as an additional factor for thermal regulation applications. Future studies powered to and designed to incorporate body composition and BSA are being considered. Due to the applied research aspect of this study, several limitations are evident. (1) Number of subjects: due to the cost associated with the one time use core temperature pills, a certain number were able to be procured with the budget provided. This number was distributed across 13 athletes to enable multiple training sessions with the same subjects. A follow-on study should include a larger number of athletes and use the data collected in this study to provide power analysis to include proper distribution among different body compositions for investigating updated algorithms. (2) Number of data points: the pills selected for this study were not able to log data, and required a reader to be placed on the back of the subject for data transfer. This significantly limited the total number of data points, with trainers only able to collect data between series during training. Future research should utilize a pill that enables continuous logging at 30–60 s increments that allows for post-training download. This will enable a single subject to collect between 120 and 240 unique data points during a 2-h practice, and will be well above the full group post hoc power analysis within a handful of subjects. (3) Potential interference with cold water ingestion: due to the early morning practice times, a 3-h pre-training ingestion time was logistically feasible, whereas a 6–8 h window that likely could further mitigate interference would be more ideal. (4) Based on the assessment of the core pills, a majority of the core temperatures in the subjects did not exceed 102.5 °F. Additional data should be collected to investigate data points exceeding 105 °F where clinical significance becomes more evident. By expanding the number of subjects and data points suggested previously, the potential for expanding the range of temperatures is likely to increase. This, along with the data presented in this study suggest that modifications can be investigated in the ECT algorithm to improve accuracy for lower body composition athletes. Due to the nature of American Football, additional investigation for football specific exertion should be studied. This is a sport with substantial equipment worn, and repeated high intensity bouts with rest in-between. This can be 10–20 s between plays, and 5–10 min between series during practices and games. This differs greatly from sports like soccer, where long durations of aerobic and anaerobic work is endured. Even more different are tactical populations like the military which can endure hours upon hours of steady state movements with substantial equipment. It is possible that enhanced algorithms can be created for these cases. 5. Conclusions HSM is an important tool that can be used in exertional training settings to help reduce the risk of EHIs. However, for this to become standard practice, scientifically validated tools must be used, but just as important, must be logistically feasible to implement on a daily basis by the training staff. Data must be trusted, reliable, and instantly actionable. The data presented in this study shows that deriving ECT off of accurate heart rate data shows promise in HSM applications. Additionally, the ECT algorithm was implemented in a commercial monitoring system and can be added to other systems with minimal software modifications, which is important for daily use applications and not research. It is ultimately up to the practitioners what level of error they are comfortable with in using this as a surveillance tool, but this data shows that it is feasible and logistically possible with commercial systems. Future work should be done investigating modified ECT algorithms off of heart rate that are specific to body composition and sport/position for athletics applications. For tactical and military populations, additional algorithm work could include age, body composition, gender, specialty. Additionally, as the wearable technology sector continues to evolve, additional sensors should be investigated for inclusion in algorithm work as well, such as accelerometers, skin temperature, ambient temperature, and potentially sweat rate. Acknowledgments The team would like to acknowledge the University of Cincinnati Training staff for support in data collection. Author Contributions Conceptualization, J.H., A.H., J.C., J.D. and R.M.; Data curation, J.H., A.H. and J.R.; Formal analysis, J.H., J.R. and J.S.; Funding acquisition, R.M.; Investigation, J.H., A.H., J.C. and R.M.; Methodology, A.H., J.C., J.D. and R.M.; Resources, R.M.; Supervision, J.D. and R.M.; Validation, A.H. and J.C.; Visualization, J.H., J.R. and J.S.; Writing—original draft, J.H., J.R. and J.S.; Writing—review and editing, J.H., A.H., J.C., J.R., J.S., J.D. and R.M. All authors have read and agreed to the published version of the manuscript. Funding This research received no external funding. Conflicts of Interest The authors declare no conflict of interest. Dedication The authors would like to dedicate this research in the loving memory of Matthew Robert Mangine Jr. Figure 1 Bland–Altman plot: Zephyr estimated core temperature (ECT) versus Gold Standard. Figure 2 Concordance plot: ECT versus Gold Standard. Figure 3 Histogram of ECT versus Gold Standard: number of observations (Y-axis) within given ranges of error (X-axis). Figure 4 Side-by-side Bland–Altman plots for Upper BF% (left) group vs. Lower BF% (right) group. Figure 5 Concordance plots of Upper BF% group (left) and Lower BF% group (right). Figure 6 Histogram plots for Upper BF% (left) and Lower BF% (right) groups, where the percent of total observations (Y-axis) is plotted against given ranges of error (X-axis). jfmk-05-00046-t001_Table 1Table 1 Characteristics of participants. Characteristic All Participants “Upper BF%” “Lower BF%” Positions included OT, LB, TE, RB, S S, CB, WR, QB, LB Number of participants 13 6 7 Height range (in) 70–78” 70–78” 70–75” Height average (in) 73.2” 74.2” 72.4” Weight range (lbs) 178–315 212–315 178–215 Weight average (lbs) 217 240 196 Body fat range (%) 3–18.8 6.7–18.8 3–6.0 Body fat average (%) 7.3 10.1 4.8 jfmk-05-00046-t002_Table 2Table 2 Practice conditions. Practice Number Practice Start Time Practice Dry Bulb Temperature (°F) (Start/Finish) Practice Wet Bulb Temperature (°F) (Start/Finish) Practice Relative Humidity (%) (Start/Finish) 1 7:00 pm 81/69 68/66 51/87 2 11:00 am 81/85 71/70 61/46 3 9:00 am 80/85 72/74 79/61 4 9:00 am 81/88 76/79 79/66 5 12:00 pm 76/78 62/63 43/43 6 1:00 pm 79/81 68/68 54/49 jfmk-05-00046-t003_Table 3Table 3 Bland–Altman statistics for Zephyr ECT vs. Gold Standard. Bland-Altman Statistic Value Lower 95% CI Upper 95% CI Bias (Δ°F) −0.192 −0.333 −0.05 Lower LOA −1.803 −2.046 −1.56 Upper LOA 1.42 1.178 1.662 jfmk-05-00046-t004_Table 4Table 4 Bland–Altman statistics for Zephyr ECT vs. Gold Standard. Statistic Value Lower 95% CI Upper 95% CI CCC 0.643 0.534 0.731 jfmk-05-00046-t005_Table 5Table 5 Subgroup Bland–Altman statistics: Upper BF% and Lower BF%. Group Bland–Altman Statistic Value Lower 95% CI Upper 95% CI Upper BF% Bias (Δ°F) −0.247 −0.388 −0.106 Lower BF% Bias (Δ°F) −0.153 −0.374 0.068 Upper BF% Lower LOA −1.26 −1.502 −1.017 Lower BF% Lower LOA −2.076 −2.456 −1.697 Upper BF% Upper LOA 0.766 0.524 1.009 Lower BF% Upper LOA 1.77 1.39 2.15 jfmk-05-00046-t006_Table 6Table 6 Subgroup concordance correlation coefficients (with 95% CI). Group Statistic Value Lower 95% CI Upper 95% CI All Participants CCC 0.643 0.534 0.731 Upper BF% CCC 0.834 0.734 0.898 Lower BF% CCC 0.515 0.336 0.659 jfmk-05-00046-t007_Table 7Table 7 Histogram values for Figure 6. Group Values within +/−0.5 °F Values within +/−1.0 °F All 54% 78% Upper BF% 57% 93% Lower BF% 46% 62% ==== Refs References 1. Gamage P. Fortington L. Finch C. Epidemiology of exertional heat illnesses in organised sports: A systematic review J. Sci. Med. Sport 2019 22 S83 S84 10.1016/j.jsams.2019.08.092 2. Casa D.J. DeMartini J.K. Bergeron M.F. Csillan D. Eichner E.R. Lopez R. Ferrara M.S. Miller K.C. O’Connor F. Sawka M.N. National Athletic Trainers’ Association Position Statement: Exertional Heat Illnesses J. Athl. Train. 2015 50 986 1000 10.4085/1062-6050-50.9.07 26381473 3. Cooper E.R. Ferrara M.S. Casa D.J. Powell J.W. Broglio S.P. Resch J. Courson R.W. Exertional Heat Illness in American Football Players: When Is the Risk Greatest? J. Athl. Train. 2016 51 593 600 10.4085/1062-6050-51.8.08 27505271 4. Cooper E.R. Ferrara M.S. Broglio S.P. Exertional Heat Illness and Environmental Conditions during a Single Football Season in the Southeast J. Athl. Train. 2006 41 332 336 17043703 5. Rav-Acha M. Hadad E. Heled Y. Moran D.S. Epstein Y. Fatal exertional heat stroke: A case series Am. J. Med. Sci. 2004 328 84 87 10.1097/00000441-200408000-00003 15311166 6. Harmon K.G. Asif I.M. Maleszewski J.J. Owens D. Prutkin J.M. Salerno J.C. Zigman M.L. Ellenbogen R. Rao A.L. Ackerman M.J. Incidence, Cause, and Comparative Frequency of Sudden Cardiac Death in National Collegiate Athletic Association Athletes: A Decade in Review Circulation 2015 132 10 19 10.1161/CIRCULATIONAHA.115.015431 25977310 7. Fullagar H.H.K. McCunn R. Murray A. Updated Review of the Applied Physiology of American College Football: Physical Demands, Strength and Conditioning, Nutrition, and Injury Characteristics of America’s Favorite Game Int. J. Sports Physiol. Perform. 2017 12 1396 1403 10.1123/ijspp.2016-0783 28338375 8. Hoffman J.R. The Applied Physiology of American Football Int. J. Sports Physiol. Perform. 2008 3 387 392 10.1123/ijspp.3.3.387 19211949 9. Hoffman J.R. Physiological demands of American football Sports Science Exchange 2015 28 1 6 10. Hoffman J.R. Kang J. Ratamess N.A. Faigenbaum A.D. Biochemical and Hormonal Responses during an Intercollegiate Football Season Med. Sci. Sports Exerc. 2005 37 1237 1241 10.1249/01.mss.0000170068.97498.26 16015144 11. Stone J.D. Kreutzer A. Mata J.D. Nystrom M.G. Jagim A.R. Jones M.T. Oliver J.M. Changes in Creatine Kinase and Hormones Over the Course of an American Football Season J. Strength Cond. Res. 2019 33 2481 2487 10.1519/JSC.0000000000001920 28394834 12. Carter R. Exertional Heat Illness and Hyponatremia Curr. Sports Med. Rep. 2008 7 S20 S27 10.1249/JSR.0b013e31817f38ff 13. Jukic I. Calleja-González J. Cos F. Cuzzolin F. Olmo J. Terrados N. Njaradi N. Sassi R. Requena B. Milanovic L. Strategies and Solutions for Team Sports Athletes in Isolation due to COVID-19 Sports 2020 8 56 10.3390/sports8040056 32344657 14. Buller M.J. Tharion W.J. Cheuvront S.N. Montain S.J. Kenefick R.W. Castellani J. Latzka W.A. Roberts W.S. Richter M. Jenkins O.C. Estimation of human core temperature from sequential heart rate observations Physiol. Meas. 2013 34 781 798 10.1088/0967-3334/34/7/781 23780514 15. Niedermann R. Wyss E. Annaheim S. Psikuta A. Davey S. Rossi R.M. Prediction of human core body temperature using non-invasive measurement methods Int. J. Biometeorol. 2013 58 7 15 10.1007/s00484-013-0687-2 23760405 16. Ganio M.S. Brown C.M. Casa D.J. Becker S.M. Yeargin S.W. McDermott B.P. Boots L.M. Boyd P.W. Armstrong L.E. Maresh C.M. Validity and Reliability of Devices That Assess Body Temperature during Indoor Exercise in the Heat J. Athl. Train. 2009 44 124 135 10.4085/1062-6050-44.2.124 19295956 17. Huggins R. Glaviano N. Negishi N. Casa D.J. Hertel J. Comparison of Rectal and Aural Core Body Temperature Thermometry in Hyperthermic, Exercising Individuals: A Meta-Analysis J. Athl. Train. 2012 47 329 338 10.4085/1062-6050-47.3.09 22892415 18. Wilkinson D.M. Carter J.M. Richmond V.L. Blacker S.D. Rayson M.P. The Effect of Cool Water Ingestion on Gastrointestinal Pill Temperature Med. Sci. Sports Exerc. 2008 40 523 528 10.1249/MSS.0b013e31815cc43e 18379216 19. Goodman D.A. Kenefick R.W. Cadarette B.S. Cheuvront S.N. Influence of Sensor Ingestion Timing on Consistency of Temperature Measures Med. Sci. Sports Exerc. 2009 41 597 602 10.1249/MSS.0b013e31818a0eef 19204591 20. Lim C.L. Byrne C. Lee J.K. Human thermoregulation and measurement of body temperature in exercise and clinical settings Ann. Acad. Med. Singap. 2008 37 347 18461221 21. Bergeron M.F. McLeod K.S. Coyle J.F. Core body temperature during competition in the heat: National boys’ 14s junior tennis championships Br. J. Sports Med. 2007 41 779 783 10.1136/bjsm.2007.036905 17562747 22. Byrne C. Lim C.L. The ingestible telemetric body core temperature sensor: A review of validity and exercise applications Br. J. Sports Med. 2006 41 126 133 10.1136/bjsm.2006.026344 17178778 23. Kolka M.A. Levine L. Stephenson L.A. Use of an ingestible telemetry sensor to measure core temperature under chemical protective clothing J. Therm. Boil. 1997 22 343 349 10.1016/S0306-4565(97)00032-6 24. Buller M.J. Tharion W.J. Duhamel C.M. Yokota M. Real-time core body temperature estimation from heart rate for first responders wearing different levels of personal protective equipment Ergonomics 2015 58 1830 1841 10.1080/00140139.2015.1036792 25967760 25. Buller M.J. Welles A.P. Friedl K. Wearable physiological monitoring for human thermal-work strain optimization J. Appl. Physiol. 2018 124 432 441 10.1152/japplphysiol.00353.2017 28798200 26. Friedl K. Military applications of soldier physiological monitoring J. Sci. Med. Sport 2018 21 1147 1153 10.1016/j.jsams.2018.06.004 29960798 27. Notley S.R. Flouris A.D. Kenny G.P. On the use of wearable physiological monitors to assess heat strain during occupational heat stress Appl. Physiol. Nutr. Metab. 2018 43 869 881 10.1139/apnm-2018-0173 29726698 28. Adams W.M. Belval L.N. Hosokawa Y. Grundstein A.J. Casa U.J. Heat Stress During American Football Heat Stress in Sport and Exercise Springer Science and Business Media LLC Berlin/Heidelberg, Germany 2019 203 218 29. HQInc Available online: https://hqinc.net (accessed on 1 June 2020) 30. Medtronic Available online: www.zephyranywhere.com (accessed on 1 June 2020) 31. Lin L.I.-K. A Concordance Correlation Coefficient to Evaluate Reproducibility Biometrics 1989 45 255 10.2307/2532051 2720055 32. Bland J.M. Altman D. Statistical methods for assessing agreement between two methods of clinical measurement Lancet 1986 327 307 310 10.1016/S0140-6736(86)90837-8 33. National Oceanic and Atmospheric Administration (NOAA) U.S. Local Climatological Data (LCD) Available online: https://data.nodc.noaa.gov/cgi-bin/iso?id=gov.noaa.ncdc:C00684 (accessed on 24 June 2020) 34. Lu M.-J. Zhong W.-H. Liu Y.-X. Miao H.-Z. Li Y.-C. Ji M.-H. Sample Size for Assessing Agreement between Two Methods of Measurement by Bland−Altman Method Int. J. Biostat. 2016 12 12 10.1515/ijb-2015-0039 35. Kurbel S. Zucić D. Vrbanec D. Plestina S. Comparison of BMI and the body mass/body surface ratio: Is BMI a biased tool? Coll. Antropol. 2008 32 299 301 18494217