Monitoring core body temperature is one of the best methods to reduce the risk of heat injury in athletic, occupational, and military settings. The manner in which core body temperature is measured represents a balance between accuracy, reliability, and logistical practicality. The "gold standard" of measurement in experimental research settings is a thermistor or a thermocouple probe inserted into the esophagus (3,20) and, in field settings, the rectum (19,25). However, these approaches are largely impractical for field use when continuous core body temperature measurements are needed. The advent of ingestible temperature sensors (ITS), which transmit intestinal temperature (T int), has allowed for the continuous monitoring and recording of core body temperature without the logistical limitations imposed by laboratory techniques designed principally for constrained data collection.
The validity and the reliability of using T int via ITS as a surrogate for core body temperature have been demonstrated under controlled conditions (4,5,7-9,18,21,31). Gant et al. (9) found good agreement between T int and T rec during intermittent exercise 10 h after ITS ingestion, and they concluded that T int measures were reliable between repeated trials when allowing 10 h between ingestion and measurement. O'Brien et al. (21) compared T int to both rectal (T rec) and esophageal (T eso) temperatures during rest and exercise as well as during warm and cold water immersion. They concluded that 12 h after ingestion, T int measures via an ITS were valid measures of core body temperature relative to T rec and T eso but suggested that ITS measures could be influenced by temperature variations along the GI tract. Although ITS measures of T int have been shown to be valid and reliable, mitigating factors such as movement of the ITS within the GI tract may limit the use of ITS in measuring core body temperature for comparison between repeated trials.
Gastrointestinal motility is unpredictable and can play a large role in transit time of an ITS device out of the stomach and within the GI tract itself. The rate of GI motility is determined acutely by numerous factors such as dietary content including the use of caffeine, alcohol, and medication (23,24,26), exercise (13,16), time of day (26), emotional state (26), and dehydration (30), and by additional factors such as age (11,17,26), sex (11,26), training and fitness (14,22), or nicotine inhalation (17). As these factors are difficult to control, the transit time and the location of an ITS device within the GI tract could be highly variable and may alter T int measures. Using standardized procedures to reduce the influence of some of these factors may decrease the variability of GI motility. McKenzie and Osgood (18) observed ITS transit times of 12.5 to 134.5 h (∼0.5-5.5 d) and suggest that volunteers ingest an ITS at the same time every day to ensure no data loss when measuring T int over extended periods.
The timing of ITS ingestion can alter T int measures such that they are not valid compared with conventional measures. Manufacturers of ITS devices recommend ingestion 3-5 h before exercise (HQinc, Palmetto, FL, and Minimitter Inc., Bend, OR). Inside this 3- to 5-h window, the amount of time the ITS resides within or near the stomach may affect T int values due to food and fluid consumption (15,27,31). It is also possible that T int values will differ along the GI tract during rest and exercise when the elapsed time between ITS ingestion and T int measurements is extended beyond 5 h. Discrepancies between T int measured via ITS and other measures of core body temperature (T rec, T eso) have been reported and range from 0.2°C to 2.2°C when an ITS is ingested 2-9 h before T int measurement (5,15,27). Although the authors speculated as to the reason for the range of temperature differences observed, they did not specifically address that the discrepancies may have been due to differences in the amount of time elapsed between ingestion and measurement of T int. More recently, Wilkinson et al. (31) reported that 5 h is sufficient time to eliminate effects of cold fluid ingestion for the majority of volunteers but suggest that an ITS be ingested 10 h before exercise to completely eliminate these effects. Gant et al. (9) supports the 10-h ingestion timing before activity to allow more time for the progression of the ITS along the GI tract where motility and differences between T int and T rec may be decreased. However, 10 h may result in the loss of an ITS in some volunteers if the measurement period extends beyond 12 h (18,21), thus requiring ingestion of a second ITS.
To date, the only study we are aware of where volunteers ingested two ITS many hours apart is reported by Wilkinson et al. (31). Volunteers ingested a pair of ITS 11.5 h apart for the purpose of identifying and observing transient temperature differences (>2.0°C) due to cold water ingestion over time. This report demonstrated that the majority of ITS are unaffected by cold fluid ingestion 5 h after ingestion; however, it does not provide data regarding the level of agreement between paired ITS after 5 h. The purpose of the current study was to quantify the agreement between a pair of ITS ingested 24 h apart by determining the amount of time the temperature differences would be greater than typical diurnal variations, and precision of the ITS device itself.
All volunteers were provided informational briefings and gave voluntary and informed written consent to participate. Investigators adhered to policies for protection of human subjects as prescribed in Army Regulation 70-25 and US Army Medical Research and Materiel Command Regulation 70-25. The research was conducted in adherence with the provisions of 32 CFR Part 219. The study protocol was approved in advance by the Human Use Review Committee at the US Army Research Institute of Environmental Medicine and the Human Subjects Research Review Board at the US Army Medical Research and Materiel Command (USAMRMC). Eight volunteers (two female) were included in the analysis. An additional five volunteers participated but were not included in the analysis. The volunteers ranged in age from 18 to 32 yr. All volunteers were members of the United States Army, were of a moderate to high fitness level, and took part in physical training on a regular basis.
The investigation took place over three consecutive days in the Mojave Desert. Because this study was conducted in the field without laboratory access, no attempt was made to calibrate ITS as has been suggested by numerous publications (4,9,21,27,31). Therefore, we administered the ITS per the manufacturer specifications. Figure 1 presents the time line of an individual day of the study. Each morning, after breakfast, volunteers orally ingested an ITS (VitalSense Jonah Ingestible Capsule; Minimitter Inc.). From 1300 to 1700 h, volunteers performed structured, intermittent activities that included light, moderate, and high intensity exercise while carrying a load of approximately 15 kg. This structured activity occurred in ambient temperatures of 38-46°C and was designed to elicit elevations in T int, which were recorded approximately every min on a portable data recorder (VitalSense Monitor, Minimitter Inc.) for each volunteer. On the second day, volunteers ingested a second ITS (24 h after the ingestion of the previous ITS), and T int measures were recorded simultaneously from both ITS using dual channels of the same VitalSense Monitor. On the third day, three male volunteers retained both of the previously ingested ITS and performed the same procedures as the previous 2 d for observational purposes. No dietary, lifestyle, or nutritional restrictions were imposed, except to abstain from alcohol. Throughout all experimental testing, the eight volunteers included in the analysis were provided with cold water and permitted to drink ad libitum but exhibited no transient decreases of T int often associated with cold fluid consumption as described by Wilkinson et al. (31). The additional five volunteers, who participated in the structured activities, were not included in the analysis because ITS-5 exhibited transient decrease in T int from cold fluid ingestion (n = 1) (8) or they expelled an ITS less than 29 h after ingestion (n = 4). Upon completion of all testing, investigators verified that the ITS telemetry signal was no longer present before volunteers were allowed to remove the "MRI incompatible" safety wristband.
At the beginning of the 4-h structured activity on the second day, the ITS ingested the previous day was in the GI tract for 29 h (ITS-29), and the ITS ingested that morning was in the GI tract for 5 h (ITS-5). During the 4-h structured activity period, comparisons were made between simultaneous T int recordings for each pair of ITS (ITS-29 vs ITS-5). It has been established that the rate of motility decreases along the GI tract such that an object, such as a bolus of food, or ITS, located in the lower small intestine would undergo less motility compared with one in the upper small intestine and greater motility compared with one in the colon (26). Therefore, this analysis was grounded on the premise that the ITS-29, in the GI tract for 29 h, would be more established within the lower GI tract whereas ITS-5 would reside somewhere in the upper GI tract (9). As an additional observation, the three male volunteers who had not expelled the previously ingested ITS by the third day allowed for a comparison of T int from an ITS ingested 53 h (ITS-53) and 29 h (ITS-29) before the 4-h period of structured activity.
Criteria for T int comparisons
Criteria for T int comparisons between the ITS ingested 24 h apart were made using a meaningful threshold of acceptance of ±0.25°C. This threshold was determined to be meaningful by the typical day-to-day variability in rectal temperature when controlling for time of day as well as the precision of the ITS device itself. Consolozio et al. (6) reported ±0.25°C as the typical standard deviation of normal resting rectal temperatures in a large group of volunteers (n > 80). The precision of the ITS used in this study was ±0.10°C (Minimitter Inc.). For the purpose of this study, the range of acceptance was chosen such that it was larger than the precision of the instrumentation and approximately equal to normal core body temperature variability. This same difference is small enough to allow detection of differences commonly considered to have physiological and psychological consequences (10) as well as differences commonly associated with circadian and ovulatory core body temperature rhythms (7,28,29).
To determine the percent of time that ITS measures were outside of the threshold of acceptance, we calculated the difference between the recorded T int for each pair of ITS approximately every minute. T int differences were plotted against time for each volunteer. The time intervals when T int differences were >0.25°C and ≤0.25°C were compared with the entire time and expressed as a percent. This analysis is very similar to the simple and intuitive nonparametric Bland-Altman (1) approach to assess agreement between two measures. Data gaps for each subject >5 min were not included in the analysis. Regression analysis was used to compare agreement between ITS-29 and ITS-5 (r 2) and to determine the magnitude (SEE) and the uniformity (residuals vs predicted) (12) of the differences across the range of core body temperatures observed. The observational comparison of three male subjects on the third day was examined in similar fashion, but with a descriptive aim only, due to the small number of subjects.
Recorded T int values from all ITS ranged from 36.94°C to 39.24°C. Linear regression analysis revealed 67% explained variance between ITS-29 and ITS-5. The typical magnitude of the differences (SEE) was 0.24°C, and these differences were uniform across the entire range of observed temperatures as determined using both runs test and visual inspection (12). However, the maximum T int difference between these paired ITS was as high as 0.83°C with a minimum difference of 0.00°C. T int differences between paired ITS for each volunteer are plotted against time in Figure 2. T int differences between ITS-29 and ITS-5 (Fig. 2) for all eight volunteers were within the ±0.25°C threshold of acceptance 56.2% of the time. The remaining 43.8% of the time, T int differences were outside of the threshold of acceptance with a mean difference of 0.40°C. The individual volunteer ITS pair percentages outside of the threshold of acceptance for ITS-29 versus ITS-5 ranged from 0.0% to 81.8% with a mean percentage of 45.0% (SD = 28.6%). We also observed that the T int differences between ITS-53 and ITS-29 for three volunteers were within the threshold of acceptance 81.5% of the time. The individual ITS pair percentages outside of the threshold of acceptance for these ITS pairs ranged from 8.0% to 29.8% with a mean percentage of 18.8% (SD = 10.9%).
The purpose of this study was to determine whether ingestion timing of an ITS would alter T int measures beyond typical diurnal variations and the precision of the ITS device itself. To determine whether a meaningful difference of ±0.25°C existed between two ITS devices ingested 24 h apart, T int measures were compared during 4 h of activity of varying intensity. Five hours after ingestion of a second ITS, T int differences were outside of the threshold of acceptance for 43.8% of the time with a mean difference of 0.40°C.
The differences observed between ITS-29 and ITS-5 may be due to differences in location along the GI tract. As contents move farther along the GI tract, motility slows (26) and ITS are likely subject to less temperature variability. Although the literature supports this notion, comparing both ITS with another measure of core body temperature (T rec and T eso) is needed to confirm this phenomenon. Although only observed in three individuals, ITS-53 versus ITS-29 were within the threshold of acceptance, a larger percentage of time (82.5%) of the 4-h observation, further supporting our contention. Most investigators who have allowed ≥10 h between ITS ingestion and measurements (9,21) report temperature differences between T int and T rec similar (0.2-0.3°C) to the threshold of acceptance in the current investigation. Conversely, reports of differences as large as ∼2.2°C between T int and T rec are possibly explained by shorter time between ingestion and activity, coupled with cold fluid ingestion (5,27).
Core temperature differences outside of the established threshold of acceptance may impact volunteer safety, athlete performance, and conclusions drawn from experimental research. The ITS difference greater than 0.25°C chosen for the threshold of acceptance of the current study is similar to several publications that identify differences between T rec and T int greater than 0.27°C (5,9,15,31) as meaningful. The consequence of temperature differences between paired ITS has the same implications as differences between a single ITS and T rec. Although the average absolute difference in temperature between ITS-29 and ITS-5 appears acceptable (r 2, SEE, uniform residuals), a closer inspection shows that 5 h after ingestion of an ITS, T int falls outside the threshold of acceptance 43.8% of observed time. Measuring and recording temperatures outside of the threshold of acceptance may place volunteers/athletes at increased risk for heat injury (10), result in degraded cognitive and physical performance (10), limit effect size of research protocols, or fail to identify documented diurnal and menstrual variations (7,28,29).
The analysis used in the current study used linear regression (12) and a method akin to the simplified nonparametric Bland-Altman approach (1) to compare agreement between ITS-29 and ITS-5. This combination affords both a conventional (12) and a simplified (1) but highly interpretable means for comparison. By plotting the error between ITS measures against time (Fig. 2), differences relative to the threshold of acceptance could be determined at any given moment over the 4-h period. Two recent publications (5,9) used conventional Bland-Altman analysis (2) to conclude that T int is a valid measure of core body temperature, using T rec as the criterion standard. Both authors report correlation values (r) >0.85 and similar mean biases that are well within the respective precision described therein, which suggests that these methods of analyses may provide useful clinical information. Despite similar conclusions, these two studies report drastically different 95% limits of agreement. Gant et al. (9) reported a 95% limits of agreement of ±0.22°C and Casa et al. (5) reports ±0.99°C. Although these 95% limits of agreement signify the relatively narrow and wide spread of the data, respectively, little information is conveyed relative to a meaningful difference. By applying a nonparametric analysis to the Bland-Altman plots of these two studies (1), all data points can be compared with the same meaningful value, which allows for an equivalent comparison of these analyses to the current study. By using the same threshold of acceptance as that designated herein (±0.25°C), we find that approximately 41% of the temperature differences reported by Casa et al. (5) are within the threshold of acceptance, whereas approximately 85% of the temperature differences reported by Gant et al. (9) are within the threshold of acceptance. Methodological differences between those two studies and the current study, such as elapsed time after ingestion and drinking cold fluids, may help to explain the differences in percentages of agreement.
The timing of ITS ingestion will affect temperature variability between ITS and other measures of core body temperature, including a second ITS. At minimum, ideal timing would allow for an ITS to travel enough distance into the intestines to avoid effects of fluid ingestion while simultaneously ensuring that the ITS is not expelled. Five hours is sufficient to eliminate the effects of fluid ingestion for the majority of volunteers (31), although 10 h of ingestion timing before activity appears to be the consensus of current ITS research to eliminate effects of fluid ingestion (9,31) and to ensure that the ITS is not passed (18,21). By ingesting a second ITS 5 h before activity, the current study demonstrates that the difference between ITS-5 and ITS-29 is within the threshold of acceptance for a much smaller percentage of time (56.2%) than the temperature differences between T rec and ITS ingested 10 h before activity (approximately 85%), as reported by Gant et al. (9). These results reinforce the notion that ingesting an ITS 10 h before activity would provide a higher level of agreement compared with 5 h. Interestingly, ingesting an ITS 29 h before activity does not appear to be more advantageous than ingesting an ITS 10 h before activity. The current study does not provide a comprehensive analysis of the advantages or limitations of using ITS for research nor clinical purposes.
The timing for ingestion of an ITS is critical for accurate and reliable core body temperature monitoring. This study shows that T int measured 5 h after ingestion may still differ significantly (>0.25°C) from T int measured 29 h after ITS ingestion. Although gastrointestinal motility is highly variable and strongly influenced by numerous, difficult to control factors, it appears that in many cases T int agreement with other core body temperature measures is improved when more than 5 h are allowed to elapse between ITS ingestion and measurement. Coaches, athletic trainers, and researchers must balance their need for accuracy and safety with the limitations of ingestion of an ITS.
The authors would like to thank SSG Jorge Diaz for technical assistance during data collection and Ms. Brett Ely for her technical assistance. The authors would also like to thank the volunteers who participated in this study.
The views, opinions, and/or findings in this report are those of the authors and should not be construed as official Department of the Army position, policy, or decision unless so designated by other official designation. The results of the present study do not constitute endorsement by ACSM. All experiments were carried out in accordance to state and federal guidelines.
1. Bland JM, Altman DG. Measuring agreement in method comparison studies. Stat Methods Med Res
2. Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet
3. Brengelmann GL. Dilemma of body temperature measurement. In: Shiraki K, Yousef MK, editors. Man in Stressful Environments
. Springfield: Thomas; 1987. p. 5-22.
4. Byrne C, Lim CL. The ingestible telemetric body core temperature sensor: a review of validity and exercise applications. Br J Sports Med
5. Casa DJ, Becker SM, Ganio MS, et al. Validity of devices that assess body temperature during outdoor exercise in the heat. J Athl Train
6. Consolazio CF, Johnson RE, Pecora LJ. Physiological variability in young men. In: Consolazio CF, Johnson RE, Pecora LJ, editors. Physiological Measurements of Metabolic Functions in Man
. New York: McGraw Hill Book Co.; 1963. p. 453-80.
7. Coyne MD, Kesick CM, Doherty TJ, Kolka MA, Stephenson LA. Circadian rhythm changes in core temperature over the menstrual cycle: method for noninvasive monitoring. Am J Physiol Regul Integr Comp Physiol
8. Easton C, Fudge BW, Pitsiladis YP. Rectal, telemetry pill and tympanic membrane thermometry during exercise heat stress. J Therm Biol
9. Gant N, Atkinson G, Williams C. The validity and reliability of intestinal temperature during intermittent running. Med Sci Sports Exerc
10. Goldman RF. Introduction to heat-related problems in military operations. In: Pandolf KB, Burr RE, Wenger CB, editors. Medical Aspects of Harsh Environments
. Washington (DC): Office of The Surgeon General at TMM Publications; 2001. p. 3-49.
11. Graff J, Brinch K, Madsen JL. Gastrointestinal mean transit times in young and middle-aged healthy subjects. Clin Physiol
12. Hopkins WG. Bias in Bland-Altman but not regression validity analyses. Sportscience
13. Keeling WF, Martin BJ. Gastrointestinal transit during mild exercise. J Appl Physiol
14. Koffler KH, Menkes A, Redmond RA, Whitehead WE, Pratley RE, Hurley BF. Strength training accelerates gastrointestinal transit in middle-aged and older men. Med Sci Sports Exerc
15. Kolka MA, Quigley MD, Blanchard LA, Toyota DA, Stephenson LA. Validation of a temperature telemetry system during moderate and strenuous exercise. J Therm Biol
16. Leiper JB, Nicholas CW, Ali A, Williams C, Maughan RJ. The effect of intermittent high-intensity running on gastric emptying of fluids in man. Med Sci Sports Exerc
17. Madsen JL, Graff J. Effects of ageing on gastrointestinal motor function. Age Ageing
18. McKenzie JE, Osgood DW. Validation of a new telemetric core temperature monitor. J Therm Biol
19. Moran DS, Mendal L. Core temperature measurement: methods and current insights. Sports Med
20. Nielsen B, Nielsen M. Body temperature during work at different environmental temperatures. Acta Physiol Scand
21. O'Brien C, Hoyt RW, Buller MJ, Castellani JW, Young AJ. Telemetry pill measurement of core temperature in humans during active heating and cooling. Med Sci Sports Exerc
22. Oettle GJ. Effect of moderate exercise on bowel habit. Gut
23. Pfeiffer A, Hogl B, Kaess H. Effect of ethanol and commonly ingested alcoholic beverages on gastric emptying and gastrointestinal transit. Clin Investig
24. Rao SS, Welcher K, Zimmerman B, Stumbo P. Is coffee a colonic stimulant? Eur J Gastroenterol Hepatol
25. Ronneberg K, Roberts WO, McBean AD, Center BA. Temporal artery temperature measurements do not detect hyperthermic marathon runners. Med Sci Sports Exerc
26. Schuster MM, Crowell MD, Koch KL. Schuster Atlas of Gastrointestinal Motility
. 2nd ed. Hamilton (Ontario): BC Decker Inc; 2002. pp. 2-18, 149, 178, 203-88, 363-74.
27. Sparling PB, Snow TK, Millard-Stafford ML. Monitoring core temperature during exercise: ingestible sensor vs. rectal thermistor. Aviat Space Environ Med
28. Stephenson LA, Kolka MA. Thermoregulation in women. In: Holloszy JO, editor. Exercise and Sport Sciences Reviews
. Baltimore (MD): Williams and Wilkins; 1993. pp. 231-62.
29. Stephenson LA, Wenger CB, O'Donovan BH, Nadel ER. Circadian rhythm in sweating and cutaneous blood flow. Am J Physiol Regul Integr Comp Physiol
30. van Nieuwenhoven MA, Vriens BE, Brummer RJ, Brouns F. Effect of dehydration on gastrointestinal function at rest and during exercise in humans. Eur J Appl Physiol
31. Wilkinson DM, Carter JM, Richmond VL, Blacker SD, Rayson MP. The effect of cool water ingestion on gastrointestinal pill temperature. Med Sci Sports Exerc