The accurate measurement of core body temperature (Tcore) is essential to protect individuals from heat injury during exposure to high levels of thermal stress (2). High levels of thermal stress are not necessarily commensurate with high ambient temperature and/or relative humidity, as many protective clothing ensembles worn by both athletic (e.g., American football) and occupational (e.g., emergency services personnel wearing conventional, chemical, biological, radiological, and nuclear protective clothing) populations impair metabolic heat dissipation (7). When these protective clothing ensembles are worn, even moderate-intensity exercise can cause uncompensable heat stress situations where Tcore continues to rise unless the exercise intensity is reduced and/or the protective clothing is removed (24,34).
In the occupational setting, by controlling the working time limit under known environmental conditions, it is possible to carefully manage an employee's exposure to heat stress (15,16,31). Whereas application of these methods of management result in low productivity by definition, most employees (usually 95%) could be safely exposed for longer work times (2). For first responders to emergency situations, the ability to extend their working time close to their own safe physiological limit is especially important, and it is the most efficient and effective way to maximize operational effectiveness during the initial stages of operations (2).
Recent advances in technology have made a small telemetric pill, which is swallowed and passed through the gastrointestinal (GI) tract, a commercially available and accurate method for measuring Tcore during typical locomotor activities (9,11,23). For emergency service personnel, accurate measurement of Tcore could minimize the risk of heat illness while maximizing operational effectiveness, especially during chemical, biological, radiological, and nuclear incidents that involve wearing additional personal protective equipment. However, to be a cost-effective and practical solution for measuring Tcore in these situations, the GI pill would have to be ingested shortly before operational deployment in response to an emergency call. In addition, standard operating procedures usually require the regular consumption of fluids when possible to minimize the adverse effects of dehydration and prolong exercise tolerance (25,26,35).
Despite the knowledge that elapsed time following GI pill ingestion affects the pill's location in the GI tract (12), and the temperature of a pill located in the stomach or upper GI tract can be affected by food and drink (5,10,19,38), there is a lack of empirical evidence detailing the extent and duration of the influence of ingested fluids on GI pill temperature measurement. Therefore, the aim of this study was to quantify the effect of the timing of GI pill ingestion and subsequent cool water ingestion on GI pill temperature.
METHODS
Ten (four female) operational UK firefighters volunteered to take part in this study. Their mean (± SD) age, height, and mass were 29 ± 5 yr, 1.76 ± 0.04 m, and 78 ± 10 kg, respectively. Ethical approval for the study was obtained from the University of Birmingham's sport and exercise science ethics committee. All firefighters gave written informed consent after being fully advised of the demands and possible risks associated with participation in the study.
Three simultaneous measures of core temperature were used during the study. Rectal temperature (Trec) was used to determine whether GI pill temperature (Tint), as measured by a temperature pill ingested 11.5 h before the start of the measurement period (T1int), was affected by the cool water ingestion. A second temperature pill (T2int) was used to quantify the effect of the cool water ingestion on a pill ingested at the start of the measurement period.
Prior to the start of the study, the rectal and intestinal temperature sensors were checked for accuracy at 39.0°C, using a precision water bath (GD100-P12, Grant Instruments, Cambridge, UK) and United Kingdom Accreditation Service calibrated precision thermometer (T600i, Digitron Ltd, Torquay, UK). All thermistors were within ± 0.1°C of the water bath temperature after 5 min of submersion.
The study was undertaken during a 2-d period at the Fire Service College in Moreton-in-Marsh, UK. At 2200 h on day 1, all 10 firefighters ingested T1int (CorTemp, HQ Inc, Palmetto, FL) before overnight sleep. The following morning, by choice, only five firefighters inserted a soft insertion thermistor (Grant Instruments, Cambridge, UK) to a depth of 10 cm beyond the anal sphincter, as indicated by a bead attached to the cable. Rectal temperature was recorded every 20 s by connecting the thermistor lead to a portable data logger (SQ800, Grant Instruments, Cambridge, UK), which was carried throughout the monitoring period in a small backpack. Rectal temperature was monitored for approximately 4 h (until lunch) to minimize unnecessary discomfort for the firefighters and to allow unobstructed running during the multistage fitness test performed in the afternoon. Ambient temperature and humidity were monitored using a combined temperature and humidity sensor (Hygroclip S, Rotronic, Crawley, UK) connected to the SQ800 data logger.
Two hours following breakfast and 11.5 h after ingesting T1int (i.e., 0930 h), all firefighters ingested T2int with their first 250-mL bolus of chilled (5-8°C) water to start the 8.5-h monitoring period. The temperatures of the GI pills were monitored every 20 s using two portable ambulatory data recorders (CorTemp, HQ Inc, Palmetto, FL) carried at the bottom of the backpack. Each pill and data recorder worked on a different frequency band (T1int on 262 kHz and T2int on 300 kHz) to avoid interference.
During the 8.5-h monitoring period, the firefighters performed repeated cycles of intermittent activity consisting of 30 min of indoor seated rest, followed by 30 min of outdoor general firefighting activities. All but the last 30-min activity period involved the firefighters walking, jogging, lifting, carrying, or using typical firefighting equipment (e.g., hoses, pumps) while performing general duties or training drills. However, the final work period of the day (1600 h) involved all firefighters performing the multistage fitness test (33) to volitional exhaustion. Standard personal protective equipment (trousers, T-shirt plus boots, overtrousers, and bunker jacket) was worn throughout the day, except during the multistage fitness test, when shorts, T-shirt, and running shoes were worn. The purpose of the multistage fitness test was to raise Tcore in a safe, controlled manner to temperatures typically encountered during emergency response training scenarios.
During the first 2 min of each 30-min rest period, each firefighter ingested a further 250-mL bolus of chilled water. A nonstandardized lunch was consumed at 1350 h, and this was the only time the firefighters were allowed to eat and drink outside the prescribed 250 mL of chilled water consumed at the start of each 30-min rest period.
All results are reported as means ± SD. All statistics analyses were carried out using the SPSS package version 12 (SPSS Inc, Chicago, IL). A Student's t-test was used to compare the accuracy of the thermistors at 39°C. A Pearson product-moment correlation was used to assess the relationship between variables. The level of agreement between two methods of measuring Tcore was assessed using the 95% limits of agreement (LoA) method proposed by Bland and Altman (3,4). The LoA are reported as bias ± (1.96 SD of the differences between paired measurements). The level of significance was set at P < 0.05. A systematic bias of less than ± 0.1°C and 95% LoA of less than ± 0.3°C were delimited as being required for accepting that two methods of measuring Tcore agree (6,11,30).
RESULTS
The mean water bath temperature was 39.00 ± 0.01°C while the accuracy of the thermistors was ascertained. The mean temperature for the rectal probes (N = 12) was 39.02 ± 0.03°C (P = 0.03), with a range of 0.10°C. The mean temperature for the intestinal pills (N = 20) was 39.06 ± 0.02°C (P < 0.01), with a range of 0.08°C. The mean intestinal pill temperature was higher than the mean rectal probe temperature (P = 0.01). However, all thermistors were accurate to within the manufacturers' specifications of ± 0.1°C at 39°C. There was also interdata logger variability with the intestinal pill system. When the temperature of one pill was recorded by 10 different loggers (every 20 s, averaged for the same 5-min measurement period), the mean pill temperature was 39.06 ± 0.02°C, with a range of 0.05°C.
The mean ambient temperature and humidity during the 30-min indoor rest periods were 17.2 ± 1.3°C and 35 ± 2% and 7.0 ± 3.2°C and 63 ± 7% during the 30-min outdoor activity periods, respectively.
Figure 1 shows the mean Trec and T1int during the monitoring period, in which Trec ranged from 37.07 ± 0.15°C to 37.77 ± 0.40°C (first 4 h only), and T1int ranged from 37.23 ± 0.22°C to 38.37 ± 0.11°C (8.5 h). The mean bias between Tcore measurements (Trec - T1int) was −0.15°C (P < 0.01), with 95% LoA of ± 0.22°C. The correlation between the paired temperature differences and the average temperature (r = 0.00, P > 0.05) and between the absolute residual paired temperature difference (from the mean bias) and the average temperature (r = −0.04, P > 0.05) suggest no evidence for proportional bias or heteroscedasticity in the data (3). The absence of proportional bias or heteroscedasticity is demonstrated in Figure 2, which shows the difference between T1int and Trec during the first 4 h of the monitoring period for each individual (N = 5) and the group.
;)
FIGURE 1: Mean rectal temperature (T rec) and mean intestinal (T1int) temperature during the 30-min work:rest intervals during the study. Note that T rec was discontinued after 4 h. SD bars are reported every 0.5 h for clarity. The peak at 7 h corresponds to the firefighters performing the multistage fitness test.
FIGURE 2: Difference between intestinal temperature (T1int) and rectal temperature (T rec) during the first 4 h of the monitoring period for each individual and the group (N = 5). Vertical dashed lines represent the ingestion of 250 mL of chilled water.
Figure 3 shows the difference between T2int and T1int during the whole monitoring period for each individual (N = 10) and the group. The large fall in T2int in comparison with T1int is clearly visible in several individuals during the first 5 h of the monitoring period and again following the 7-h and 8-h points in two individuals (Fig. 3). However, water ingestion had a highly variable effect both within and between individuals in transiently reducing T2int in comparison with T1int (Table 1 and Fig. 3). In general, the transient reduction in T2int became progressively smaller as time following ingestion increased. During the first 5 h of the study, the mean reduction in T2int following water ingestion lasted 32 ± 10 min before recovering to within 0.3°C of the mean T1int (Fig. 3). While many firefighters showed a negligible maximum difference (± 0.1°C) between T1int and T2int after 3-5 h, two firefighters still had large transient differences (peaking at 2.0°C and 6.3°C) between pill temperatures when 250 mL of water was ingested some 8 h following the ingestion of T2int (Fig. 3). It is interesting to note that in these two firefighters, there was no evidence of water ingestion cooling T2int at 6 h (Fig. 3).
FIGURE 3: Difference between intestinal temperatures measured by a pill ingested the night before (T1int) or at the start of the monitoring period (T2int) for each individual and the group (N = 10). Vertical dashed lines represent the ingestion of 250 mL of chilled water (black) or lunch (gray).
TABLE 1: Median, maximum, and minimum temperature difference (T1int-T2int) within each hourly cycle of rest/activity after ingestion of T2int.
DISCUSSION
The main finding of this study is that ingestion of a 250-mL bolus of cool water (5-8°C) can transiently decrease GI pill temperature (more than 2°C) in some individuals, even when the water is ingested 8 h after pill ingestion (Table 1). This negates the use of the GI pill as a valid measure of Trec for approximately 30-60 min after the ingestion of cool fluids (Fig. 3). Hence, Tint is unsuitable for the measurement of Tcore in emergency service personnel, who would need to ingest the pill with fluids immediately prior to deployment, and who would continue to drink fluids when possible during operations to maximize operational effectiveness (25,26,35). In the sport and exercise medicine setting, careful planning should allow the pill to be ingested the night before performing activities when Tint needs to be monitored.
The initial transient decline of T2int in response to cool water ingestion is attributable to the pill's direct contact with the water in the mouth, esophagus, and stomach. It has been suggested that a GI pill will pass into the small intestine within 1 h (27), and this process will be aided by consuming a small meal (18,21). As the pill passes along the small intestine, the influence of the cool water on T2int will diminish as the water will have more time to equilibrate to Tcore before any possible contact with the pill. This general effect can be seen in the mean response shown in Figure 3. Both the magnitude of the decrease in T2int and the time before T2int approaches T1int decreases as the time following pill ingestion increases.
It is unlikely that water ingestion transiently decreases T2int only as a result of direct contact with the pill in the stomach and along the upper regions of the small intestine. Two subjects experienced significant (2-6°C) decreases in T2int at 7-8 h after pill ingestion (Fig. 3). In addition, these subjects demonstrated no noticeable decreases in T2int in the preceding 2 h before this period (i.e., 5-6 h). One possible explanation for these findings is that the ingested water may cause localized cooling of areas of the small and large intestines in close proximity to the stomach and duodenum. The transverse colon, for example, lies directly underneath the stomach, attached by the greater omentum.
In this study, assuming that the temperature of the 250 mL of cool water increased from 6°C to 37.5°C, heat lost through conduction from the tissues surrounding the stomach region to the ingested volume of water would be approximately 33 kJ (25), enough to lower the surrounding tissue temperature initially by several degrees centigrade. Therefore, the temperature registered by a GI pill will clearly be influenced by its position in the GI tract and its proximity to the stomach when cool fluids are ingested. As pill transit times have been reported to range between 8 h (21) and 136 h (23), it is unrealistic to presume that increasing the time between pill ingestion and Tint measurement will necessarily decrease the possible influence of cold water ingestion on Tint, unless the pill resides close to the rectum (22). In addition, increasing transit time increases the likelihood of the pill passing out of the GI tract. Approximately 20-30% of active or well-trained individuals will pass the pill in a bowel movement within 12 h of ingestion after overnight sleep (20,32). The only practical way to be sure that the pill is located near the rectum and free from the modifying effects of fluid and food is to use the pill as a rectal suppository (17).
These findings of the present study also demonstrate that the practice of giving subjects a cold drink before commencing measurement of Tint, to confirm that the pill is far enough down the GI tract to be unaffected by fluid intake (10,19,36), is unreliable. While the pill may be in a position in the GI tract to be unaffected by fluid intake during this preliminary test, it may subsequently move to a position in close proximity to the stomach or duodenum, where localized cooling of the surrounding tissues may influence Tint once more (Fig. 3).
Ingesting the pill before overnight sleep and allowing at least 10 h before measurement of Tint appears to offer the best possibility of the pill being unaffected by subsequent fluid ingestion. In this study, there were no apparent reductions in T1int as a result of the water ingestion 11.5 h after pill ingestion (Fig. 2). A similar finding was reported after 10 h following repeated ingestion of 4°C water (11). Shorter ingestion periods (2-3 h) have clearly shown that fluid consumption affects Tint (19,38). Although ingesting fluids close to body temperature should minimize this effect (8), it would be difficult to accurately quantify, especially when core body temperature is changing during exercise. In addition, warm fluid ingestion (40°C) has a detrimental impact on voluntary hydration (14,39) and would negate any potential thermoregulatory benefit that may result from an increased heat storage capacity for fluids cooler than Tcore (25,35).
There was good agreement between T1int and Trec, with T1int consistently measuring 0.15°C higher than Trec. If the difference in the accuracy of temperature measurement between the temperature pills and rectal probes at 39°C is accounted for (0.04°C), the true bias is 0.11°C. This bias in favor of slightly higher pill temperatures compared with Trec is consistent with the findings of several recent investigations that have reported similar biases of 0.07°C (6,21), 0.20°C (9), and 0.15°C (11). This bias should be accounted for when interpreting an individual's thermal status (11). However, failure to correct for this small bias when using Tint as a critical exercise termination temperature should not be a major concern, as it will effectively lower the equivalent Trec critical temperature limit by approximately 0.1-0.2°C. The cause of this temperature difference between Tint and Trec remains to be elucidated, but it may result from the existence of a temperature gradient along the GI tract (6).
Early studies have reported problems with manufacturer temperature calibration of the GI pills (19), with corrections of up to 0.8°C being required in some pills (21,27,37). The temperature pills transmit a low-frequency radio wave, which is susceptible to electromagnetic interference (6,21). Therefore, any calibration system (e.g., heated water bath) must ensure that the pills are shielded from any electromagnetic heating device (27,38). In addition, the low-frequency radio waves may rebound off some metal water bath tanks and cause distortion to the transmitted signal, so plastic water bath tanks are recommended for pill calibration. In this study, all 20 temperature pills were within the manufacturer's calibration limit of ± 0.1°C, with the largest difference between any two pills of 0.08°C and the greatest difference from the mean temperature (39.06°C) of ± 0.04°C. However, it is still recommended that all GI pills are individually calibrated to ensure accuracy before use (6).
The specific logger used to process the radio wave received from the pill will add additional variability to the recorded pill temperature. When 10 different data loggers that had just been factory calibrated were used to measure the temperature of one pill at 39°C, there was a maximum interlogger difference of 0.05°C. This small but significant variability is inherent with the current sensitivity of the electrical components within the data logger, but it highlights the need to calibrate each pill with the specific logger used during measurement to maximize measurement accuracy (6). Even so, with some variability inherent between pills, data loggers, and the gold standard calibration standards used to determine temperature accuracy between the United Kingdom and United States (United Kingdom Accreditation Service and National Institute of Standards and Technology; both certify to within ± 0.05°C), an accuracy to within ± 0.1°C should be expected for commercially available temperature pill systems.
A number of novel, noninvasive techniques that aim to accurately measure Tcore are currently under development. These include the measurement of insulated skin temperature (40), brain tunnel temperature (13), and microwave emission radiometry (28). However, these methods remain to be incorporated into a simple, friendly, and comfortable device with proven validity against a recognized gold standard measure of Tcore during both rest and strenuous exercise (29). At present, rectal temperature is still considered the gold standard for Tcore measurement for thermal strain in both occupational and athletic populations (1).
CONCLUSION
In conclusion, this study has shown that a GI temperature pill ingested immediately prior to physical activity cannot be used to measure core body temperature accurately in all individuals for 30-60 min after the ingestion of cool fluids during the following 8-h period. Therefore, Tint is not a suitable measurement site for use by emergency service personnel to assess their individual levels of heat strain during emergency deployments. In contrast, if a temperature pill is ingested before overnight sleep, there is good agreement between Tint and Trec (≥ 10 h after ingestion), with the accuracy of Tint being independent of fluid ingestion after this time.
This work was supported by the Home Office's CBRN Science and Technology Programme.
REFERENCES
1. American College of Sports Medicine. Position stand: exertional heat illness during training and competition.
Med Sci Sports Exerc. 2007;39(3):556-72.
2. Bernard TE, Kenney WL. Rationale for a personal monitor for heat strain.
Am Ind Hyg Assoc J. 1994;55:505-14.
3. Bland MJ, Altman DG. Measuring agreement in method comparison studies.
Stat Methods Med Res. 1999;8:135-60.
4. Bland MJ, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement.
Lancet. 1986;i:307-10.
5. Brake DJ, Bates GP. Deep body core temperatures in industrial workers under thermal stress.
J Occup Environ Med. 2002;44:125-35.
6. Byrne C, Lim CL. The ingestible telemetric body core temperature sensor: a review of validity and exercise applications.
Br J Sports Med. 2007;41:126-33.
7. Cheung SS, McLellan TM, Tenaglia S. The thermophysiology of uncompensable heat stress. Physiological manipulations and individual characteristics.
Sports Med. 2000;29:329-59.
8. Ducharme MB, McLellan TM, Moroz D, Buguet A, Radomski MW. A 36-hour comparison of core temperature at rest and during exercise using rectal probe and pill telemetry. In:
Proceedings of the Australian Physiological and Pharmacological Society International Thermal Physiology Symposium. Wollongong (Australia): Australian Physiological and Pharmacological Society; 2001. p. 28P.
9. Edwards B, Waterhouse J, Reilly T, Atkinson G. A comparison of the suitabilities of rectal, gut, and insulated axilla temperature for measurement of the circadian rhythm of core temperature in field studies.
Chronobiol Int. 2002;19:579-97.
10. Fox RH, Goldsmith R, Wolff HS. The use of a radio pill to measure deep body temperature.
J Physiol. 1962;160(Suppl.):22-3.
11. Gant N, Atkinson G, Williams C. The validity and reliability of intestinal temperature during intermittent running.
Med Sci Sports Exerc. 2006;38(11):1926-31.
12. Gibson TM, Redman PJ, Belyavin AJ. Prediction of oesophageal temperature from core temperatures measured at other sites in man.
Clin Phys Physiol Meas. 1981;2:247-56.
13. Haddadin AS, Abreu MM, Silverman DG, Luther M, Hines RL. Noninvasive assessment of intracranial temperature via the medical canthal-brain temperature tunnel.
Anesthesiology. 2006;105:A70.
14. Hubbard RW, Sandick BL, Matthew WT, et al. Voluntary dehydration and alliesthesia for water.
J Appl Physiol. 1984;57:868-73.
15. International Standards Organization.
Ergonomics of the Thermal Environment-Analytical Determination and Interpretation of Heat Stress Using Calculation of the Predicted Heat Strain. Geneva (Switzerland): International Standards Organization; 2004.
16. International Standards Organization.
Hot Environments-Estimation of the Heat Stress on Working Man, Based on the WBGT-Index (Wet Bulb Globe Temperature). Geneva (Switzerland): International Standards Organization; 1989.
17. Keatinge WR, Nield PJ. Use of rectal radiopill to monitor human body core temperature during four mile swim across Beagle Channel, Tierra del Fuego.
J Physiol. (Lond.) 1990;430:93.
18. Kolka MA, Levine L, Stephenson LA. Use of an ingestible telemetry sensor to measure core temperature under chemical protective clothing.
J Therm Biol. 1997;22:343-9.
19. Kolka MA, Quigley MD, Blanchard LA, Toyota DA, Stephenson LA. Validation of a temperature telemetry system during moderate and strenuous exercise.
J Therm Biol. 1993;18:203-10.
20. Laursen PB, Suriano R, Quad MJ, et al. Core temperature and hydration status during an ironman triathlon.
Br J Sports Med. 2006;40:320-5.
21. Lee SM, Williams WJ, Schneider SM. Core temperature measurement during supine exercise: esophageal, rectal and intestinal temperatures.
Aviat Space Environ Med. 2000;71:939-45.
22. Livingstone SD, Grayson J, Frim J, Allen CL, Limmer RE. Effect of cold exposure on various sites of core temperature measurements.
J Appl Physiol. 1983;54:1025-31.
23. McKenzie JE, Osgood DW. Validation of a new telemetric core temperature monitor.
J Therm Biol. 2004;29:605-11.
24. McLellan TM. Heat strain while wearing the current Canadian or new hot-weather French NBC protective clothing ensemble.
Aviat Space Environ Med. 1996;67:1057-62.
25. McLellan TM, Cheung SS. Impact of fluid replacement on heat storage while wearing protective clothing.
Ergonomics. 2000;43:2020-30.
26. McLellan TM, Cheung SS, Latzka WA, et al. Effects of dehydration, hypohydration and hyperhydration on tolerance during uncompensable heat stress.
Can J Appl Physiol. 1999;24:349-61.
27. Mittal BB, Sathiaseelan V, Rademaker AW, Pierce MC, Johnson PM, Brand WN. Evaluation of an ingestible telemetric temperature sensor for deep hyperthermia applications.
Int J Radiat Oncol Biol Phys. 1991;21:1353-61.
28. Moran DS, Eliyahu U, Heled Y, Rabinovitz S, Hoffman J, Margaliot M. Core temperature measurement by microwave radiometry.
J Therm Biol. 2004;29:539-42.
29. Moran DS, Mendal L. Core temperature measurement. Methods and current insights.
Sports Med. 2002;32:879-85.
30. Muir IH, Bishop PA, Lomax RG, Green JM. Prediction of rectal temperature from ear canal temperature.
Ergonomics. 2001;44:962-72.
31. National Institute for Occupational Safety and Health.
Criteria for a Recommended Standard. Occupational Exposure to Hot Environments. Revised Criteria 1986. Washington (DC): United States Department of Health and Human Services, National Institute for Occupational Safety and Health; 1986. Publication no. 86-113.
32. O'Brien C, Hoyt RW, Buller MJ, Castellani JW, Young AJ. Telemetry pill measurements of core temperature in humans during active heating and cooling.
Med Sci Sports Exerc. 1998;30(3):468-72.
33. Ramsbottom R, Brewer J, Williams C. A progressive shuttle run test to estimate maximum oxygen uptake.
Br J Sports Med. 1988;22:141-4.
34. Selkirk GA, McLellan TM. Physical work limits for Toronto firefighters in warm environments.
J Occup Environ Hyg. 2004;1:199-212.
35. Selkirk GA, McLellan TM, Wong J. The impact of various dehydration volumes for firefighters wearing protective clothing in warm environments.
Ergonomics. 2006;49:418-33.
36. Sleivert GG. Using microtechnology to monitor thermal strain and enhance performance in the field.
Int J Sports Physiol Perform. 2007;2:98-102.
37. Sparling PB, Snow TK, Millard-Stafford ML. Monitoring core temperature during exercise: ingestible sensor vs. rectal thermistor.
Aviat Space Environ Med. 1993;64:760-3.
38. Stephenson A, Quigley MD, Blanchard LA, Toyota DA, Kolka MA.
Validation of Two Temperature Pill Telemetry Systems in Humans during Moderate and Strenuous Exercise. Natick (MA): U.S. Army Research Institute of Environmental Medicine; 1992. Technical report T10/92.
39. Szlyk PC, Sils IV, Francesconi RP, Hubbard RW, Armstrong LE. Effects of water temperature and flavoring on voluntary dehydration in men.
Physiol Behav. 1989;45:639-47.
40. Taylor NAS, Wilsmore BR, Amos D, Takken T, Komen T.
Insulated Skin Temperature. Indirect Indices of Human Body-Core Temperature. Melbourne (Australia): Defence Science Technology Organisation; 1998. DSTO-TR-0752.
