Body Anthropometric Characteristics and Rectal Temperature Cooling Rates in Women With Hyperthermia

In this cross-sectional laboratory study, the independent variable was the BSA:LBM ratio group (high group = 4.161 ± 0.232 m²/kg·10², low group = 3.759 ± 0.214 m²/kg·10²). The dependent variables were T_rec, CWI duration, BSA, LBM, body mass index (BMI), percentage of body fat, and maximal oxygen consumption (o₂max). We calculated T_rec cooling rates using T_rec and CWI durations. We also measured exercise intensity, environmental chamber temperature and relative humidity, and hydration status to ensure consistent testing conditions between groups.

We estimated sample size a priori using the following assumptions: an α level of .05, 80% power, a difference in T_rec cooling rates of 0.11°C/min,¹²  and an SD of 0.08°C/min.⁹  Based on these assumptions, we needed 8 people in each group to detect a difference. Therefore, we purposefully recruited women with high or low BSA:LBM ratios; our goal was an average difference in BSA:LBM ratio of at least 0.40 m²/kg·10² between groups based on previous research¹²  in men that demonstrated statistical significance in cooling rates with this difference. A convenience sample of 20 healthy, physically active (performed moderate to vigorous activity for at least 30 minutes, 3 times per week) women volunteered for this study. Four women chose to discontinue testing and were excluded from the final data analysis: 3 discontinued because of the difficulty of the exercise protocol on day 2, whereas the fourth woman had an elevated T_rec before testing.

A total of 16 physically active, healthy women with various BSA:LBM ratios completed this study. We rank ordered these participants by BSA:LBM ratio and evenly divided them into 1 of 2 groups post hoc (8 in each group; Table). Volunteers were excluded if they self-reported any of the following: (1) an injury that impaired their ability to run; (2) any self-reported neurologic, metabolic, gastrointestinal, respiratory, or cardiovascular disease diagnosed by a physician; (3) taking any medication (eg, diuretics) that affected fluid balance or temperature regulation; (4) a sedentary lifestyle (ie, active for <30 minutes, 3 times per week)²¹; (5) a history of heat-related illness (eg, heat exhaustion) in the 6 months preceding data collection; (6) current illness or fever; (7) being a current smoker; (8) allergy to cold; or (9) current pregnancy. Three participants completed the study while using a low-dose levonorgestrel (<35 mg, <1 nmg/d release) intrauterine birth control device and did not experience a menstrual cycle. All other participants denied taking birth control medication and were considered eumenorrheic. All participants provided written informed consent before testing, and all procedures were approved by the Central Michigan University Institutional Review Board.

Participants reported for testing on 2 days separated by 48 hours. Data collection for both days occurred between 9 am and 4 pm to reduce fluctuations in body temperature due to circadian rhythms.²²  Participants completed both days of testing during their follicular phase of menstruation (days 1–14 after start of menses). They were asked to arrive each day properly hydrated (having consumed ≥500 mL of water the night before testing) and fasting for 2 hours.

On the first testing day, participants answered a health history questionnaire. We measured their height, and they emptied their bladders so we could assess hydration using a refractometer (model SUR-Ne refractometer; Atago USA Inc). If urine specific gravity was >1.020,²³  they were considered too hypohydrated for testing. We reminded them to consume more fluids and asked them to report ≥24 hours later.

Euhydrated participants were weighed nude to the nearest hundredth of a kilogram (model Defender 5000; Ohaus Corp). They dressed in workout apparel and underwent body composition measurements using dual-energy x-ray absorptiometry (DEXA; Lunar Prodigy; General Electrics Co) so we could obtain LBM, fat mass, and body fat percentage. These data were used to calculate BSA via the following formula: Body mass (kg)^0.425 × Height (cm)^0.725 × 0.007184.²⁴  Afterward, we measured the o₂max (model VMax Console System; CareFusion) during exercise. This was necessary so we could ensure that exercise intensity was consistent among participants on the second testing day. Briefly, this test consisted of participants running on a treadmill at a speed of 8 km/h (5 mph). After 2 minutes, the speed increased to 8.9 km/h (5.5 mph). After another minute, the speed increased to 9.6 km/h (6 mph). The treadmill grade was then increased by 1% per minute until the participant's oxygen consumption plateaued, the participant quit, or a rating of perceived exertion of 20 was reported.²⁵  The rating of perceived exertion was reported every 2 minutes. After maximal exercise testing, they were excused.

On day 2, participants again emptied their bladders completely, had their hydration assessed, and were weighed nude. They self-inserted a rectal thermistor (model YSI 4600 Precision Thermometer with 401 probe; Advanced Industrial Systems, Inc) 15 cm past the anal sphincter²⁶  and dressed in athletic clothing (ie, socks, shoes, undergarments, running shorts, sports bra, and T-shirt) and a disposable rain poncho to expedite the increase in body temperature during exercise. They entered an environmental chamber (temperature = ∼40°C, relative humidity = 40%) and stood on a treadmill (model 1850; Proform Performance) for 10 minutes to adjust to the hot conditions, and T_rec was recorded.

Participants walked at 4.8 km/h (3 mph) for 3 minutes. Next, they ran at a treadmill speed (0% incline) corresponding to 80% of their predetermined o₂max for 2 minutes. We calculated 80% of o₂max via a linear regression model using data from the incremental o₂max testing performed on day 1. This sequence was repeated until T_rec reached 39.5°C. We recorded T_rec every 5 minutes during the exercise portion of the experiment. After T_rec reached 39.5°C, participants stopped the treadmill, removed the poncho, and immersed themselves up to the neck in an 1135.6-L capacity, noncirculating water tub (model 4247; Rubbermaid) kept at a temperature of ∼10°C; this temperature falls within expert recommendations²  for treating EHS. We circulated the water every 2 minutes, and participants remained immersed until T_rec reached 38°C. We recorded T_rec every 30 seconds during the CWI portion of the experiment. The CWI duration was recorded when T_rec reached 38°C. Participants exited the water bath and sat in the environmental chamber for 15 minutes. They exited the environmental chamber, dried themselves completely, and removed the rectal thermistor. They were weighed nude a second time, dressed in dry clothes, and were excused. No fluids were given to participants at any time during testing.

We assessed the data for normality and calculated means and SDs. We used independent-samples t tests to determine if differences between groups existed for T_rec cooling rates during CWI, afterdrop cooling rates, nadir values postcooling, exercise durations, BSA:LBM ratios, preimmersion water bath temperature, BSA, LBM, BMI, body fat percentages, and o₂max. Pearson product moment correlation coefficients were calculated to determine if any physical characteristic was correlated with T_rec cooling rates.

Given that exercise and CWI durations differed among participants, we only statistically compared T_rec at times common to all participants. Separate repeated-measures analyses of variance were used to determine if differences in T_rec existed between groups during exercise and CWI. Sphericity was assessed using the Mauchly test. Geisser-Greenhouse adjustments to P values and degrees of freedom were made when sphericity was violated. Tukey-Kramer post hoc tests were then used if interactions or main effects were observed. We did not examine the simple main effects of time because this did not address our research questions. Significance was accepted when P < .05 (version 2007; Number Cruncher Statistical Software).

By design, our groups differed by BSA:LBM ratio: the high group had a 0.402 m²/kg·10² greater BSA:LBM ratio than did the low group (t₁₄ = 3.6, P = .001; Table). The LBM (t₁₄ = 2.2, P = .02), body fat percentage (t₁₄ = 2.8, P = .007), and o₂max (t₁₄ = 1.9, P = .04) also differed between groups. In contrast, BSA (t₁₄ = 0.3, P = .40) and BMI (t₁₄ = 0.9, P = .17) did not differ between groups.

Both groups had similar T_rec during exercise (F_2,20 = 2.8, P = 1.0) and exercised for similar durations (t₁₄ = 1.3, P = .11; Figure) and in similar environmental temperatures and relative humidities (Table). Preimmersion water bath temperature was also similar between groups (t₁₄ = 0.4, P = .72; Table). Rectal temperatures during CWI (F_2,16 = 0.11, P = .77) and CWI durations were comparable in each group (Figure). The groups had similar CWI cooling rates (t₁₄ = 0.3, P = .39), afterdrop cooling rates (t₁₄ = 0.49, P = .69; Table), and nadir values (t₁₄ = 0.3, P = .61). The BSA:LBM ratio (r = 0.14, P = .59), LBM (r = −0.10, P = .70), BSA (r = −0.01, P = .97), BMI (r = 0.37, P = .16), and body fat percentage (r = 0.29, P = .28) were not correlated with T_rec cooling rates.

None of the physical characteristics that we measured (ie, BSA:LBM ratio BSA, LBM, body fat percentage) were important in determining how quickly women with hyperthermia cooled during CWI. This finding is in stark contrast to the results of others¹²  who observed that their high BSA:LBM ratio group cooled ∼1.7 times faster than did the low BSA:LBM ratio group (3.15 ± 0.079 m²/kg·10² versus 2.75 ± 0.085 m²/kg·10²). Similarly, Fowkes-Godek et al²⁰  reported that American football lineman with lower BSA:LBM ratios cooled 0.10°C/min slower than did cross-country runners (0.16 ± 0.06°C/min versus 0.26 ± 0.05°C/min).

We propose 3 reasons for the different observations.^12,20  First, women tend to have greater overall adiposity, less LBM, and a higher BSA:LBM ratio than do men. Lemire et al¹⁰  noted that, even when matched for BSA:LBM ratio, women cooled ∼1.7 times faster than did men (0.22 ± 0.07°C/min versus 0.12 ± 0.03°C/min). They concluded that the greater LBM in men accounted for the difference in cooling rates because they retained heat longer. Given that our groups had lower LBM, it is not surprising that their cooling rates were higher, lending further support to the observations of Lemire et al.¹⁰  Second, both of our groups had considerably higher BSA:LBM ratios than did the male participants in the study of Friesen et al.¹²  In fact, the BSA:LBM ratios of our low group was >1.00 m²/kg·10² greater than that of their low BSA:LBM ratio group composed of men (ie, 2.75 ± 0.086 m²/kg·10²). There may be a value at which the BSA:LBM ratio ceases to clinically affect T_rec cooling rates when CWI is used to treat hyperthermia. Third, Friesen et al¹²  used considerably colder water (temperature = 2°C) to treat their participants than we did, yet our cooling rates surpassed theirs. This may suggest that the ability to dissipate heat through a greater BSA:LBM ratio is more effective than colder water baths. However, this possibility requires further research because differences between sexes exist for behavioral or pacing strategies, running economy, and heat-generating capacity.^10,27 

Another main observation in our study was that the T_rec cooling rates of both groups were excellent and well exceeded expert recommendations (ie, >0.10°C/min).⁶  Our T_rec cooling rates were consistent with those of other scientists^4,10,28  who demonstrated female T_rec cooling rates ranging from 0.22 ± 0.07°C/min to 0.27°C/min. Physical characteristics did not affect female T_rec cooling, so clinicians do not need to account for these variables when treating patients with hyperthermia and, perhaps, EHS. Moreover, predetermined cooling-duration estimates for women with hyperthermia cannot, and should not, be used to guide the treatments of patients with EHS. In a retrospective analysis of 6 cooling studies, Poirier et al¹³  stated that physical characteristics could not predict CWI durations. Whereas LBM was the best predictor of cooling times, it still explained only 14% of the variability in cooling times, but the large residual errors would have resulted in the removal of 65% of their participants with hyperthermia from CWI too soon.¹³  Because EHS mortality and morbidity is related to the amount of time spent in hyperthermia,^2,29  we strongly advise clinicians to insert an indwelling rectal thermistor and monitor T_rec continuously during and after CWI instead of using physical characteristics to estimate cooling times. This strategy has 2 distinct advantages. First, it will help prevent dangerous hypothermic afterdrop in women because they have less LBM to generate metabolic heat post-CWI. Second, it will prevent the removal of the patient from the lifesaving CWI before T_rec is reduced sufficiently.

We acknowledge the limitations of our study. First, while all testing took place within a 7-hour window (9 am to 4 pm), minor discrepancies among participants' T_rec may have occurred. Coyne et al²²  tracked body temperature continuously for 11 cycles and reported an average follicular-phase body temperature of 36.6°C ± 0.16°C, indicating that minimal variations (<0.2°C) in body temperature occurred over the course of a day during the follicular phase. Eighty-eight percent (14/16) of our participants started testing within the same 4-hour block of time (between 9 am and 1 pm). Moreover, the T_rec before exercise, rate of temperature increase during exercise, and final postexercise T_rec in each group were almost identical, making it unlikely that circadian rhythms affected our data. Second, for safety reasons, none of our women experienced EHS. Third, 3 of our participants were using birth control at the time of testing. However, multiple researchers have suggested that the reason for temperature increases during the luteal phase is that progesterone increases while estrogen decreases. Without the decline in estrogen in conjunction with the minute dose of progesterone, any temperature change from this medication in these 3 participants was unlikely.^30,31  Finally, we did not measure skin temperature during exercise or CWI.

Several body anthropometric characteristics did not affect how quickly women with hyperthermia cooled when completing CWI. Consequently, health care professionals do not need to account for body composition when treating women with exertional hyperthermia. Moreover, given the lack of correlation between physical characteristics and T_rec cooling, we could not identify predetermined CWI duration estimates to ensure safe T_rec. Instead, women suspected of having EHS should be rapidly diagnosed using T_rec, followed by immediate CWI.²  Future investigators should examine the influence of physical characteristics on cooling in women with hyperthermia and, more specifically, in women who experience EHS because of the disparity in findings between controlled laboratory studies and field observations.¹³  Overall, CWI was highly effective in cooling women with hyperthermia who had various physical characteristics and should be used for women with EHS because it saves lives.⁴ 

We thank the Central Michigan University Office of Research and Graduate Studies and College of Health Professions for funding this study. We also thank Megan Keen, Mike Szymanski, and Courtney Zickmund for their assistance during data collection.