Share this article on:

Sweating Is Greater in NCAA Football Linemen Independently of Heat Production


Medicine & Science in Sports & Exercise: February 2012 - Volume 44 - Issue 2 - p 244–252
doi: 10.1249/MSS.0b013e31822cb4e2
Basic Sciences

Purpose: The study’s purpose was to investigate whether differences in local sweat rates on the upper body between American football linemen (L) and backs (B) exist independently of differences in metabolic heat production.

Methods: Twelve NCAA Division I American football players (6 linemen (mass = 141.6 ± 6.5 kg, body surface area (BSA) = 2.67 ± 0.08 m2) and 6 backs (mass = 88.1 ± 13.4 kg, BSA = 2.11 ± 0.19 m2)) cycled at a fixed metabolic heat production per unit BSA of 350 W·m−2 for 60 min in a climatic chamber (tdb [dry bulb temperature] = 32.4°C ± 1.0°C, twb [wet bulb temperature] = 26.3°C ± 0.6°C, v [air velocity] = 0.9 ± 0.1 m·s−1). Local sweat rates on the head, arm, shoulder, lower back, and chest were measured after 10, 30, and 50 min of exercise. Core temperature, mean skin temperature, and HR were measured throughout exercise.

Results: Because metabolic heat production per unit surface area was fixed between participants, the rate of evaporation required for heat balance was similar (L = 261 ± 35 W·m−2, B = 294 ± 30 W·m−2, P = 0.11). However, local sweat rates on the head, arm, shoulder, and chest were all significantly greater (P < 0.05) in linemen at all time points, and end-exercise core temperature was significantly greater (P = 0.033) in linemen (38.5°C ± 0.4°C) relative to backs (38.0°C ± 0.2°C) despite a ∼25% lower heat production per unit mass. The change in mean skin temperature from rest was greater in linemen (P < 0.001) after 15, 30, 45, and 60 min, and HR was greater in linemen for the last 30 min of exercise.

Conclusions: Football linemen sweat significantly more on the torso and head than football backs independently of any differences in metabolic heat production per unit BSA and therefore the evaporative requirements for heat balance. Despite greater sweating, linemen demonstrated significantly greater elevations in core temperature suggesting that sweating efficiency (i.e., the proportion of sweat that evaporates) was much lower in linemen.

1Thermal Ergonomics Laboratory, School of Human Kinetics, University of Ottawa, Ottawa, Ontario, CANADA; and 2Division of Sports Medicine, Department of Family Medicine, University of South Florida, College of Medicine, Tampa, FL

Address for correspondence: Ollie Jay, Ph.D., Thermal Ergonomics Laboratory, School of Human Kinetics, University of Ottawa, Ottawa, Ontario, Canada K1N 6N5; E-mail:

Submitted for publication April 2011.

Accepted for publication July 2011.

Heat stress has become a primary concern among sports organizations during the past couple of decades. Athletes participating in American football training camps during the summer months are at a particular risk of physiological heat strain due to high ambient temperatures and humidity, the metabolic demands of the sport (22), and the thermal resistance of the equipment worn (1,27). Since 1995, 39 athletes have died of heat-related illnesses while participating in the sport at both the amateur and the professional levels of competition (9).

Athletes participating in American football typically represent a wide range of morphological variation when compared with other popular summer sports (e.g., running, cycling, or soccer) (14). Vast differences in physical characteristics such as total body mass, body surface area (BSA), percentage of body fat, and surface area-to-mass ratio may potentially predispose some groups of players to a greater risk of critically high core temperatures (Tcore) (18,20,21,35). Indeed, a recent study by Godek et al. (12) demonstrated that larger “linemen” such as offensive linemen, defensive tackles, and defensive ends had significantly larger elevations in Tcore during a National Football League (NFL) training camp relative to smaller “backs” such as defensive backs, wide receivers, and linebackers. However, the factors responsible for higher Tcore in linemen remain unclear. The primary focus of research has been on the potential role of hydration, and although marked dehydration can clearly exacerbate heat stress (40), it does not seem to be the principal reason for position-related differences in Tcore elevation (12).

Evaporative heat loss through sweating is the primary avenue of heat dissipation during exercise, particularly in hot environments where ambient air temperatures approach skin temperature (31). The regional distribution of sweating has never been measured in football players. Moreover, Godek et al. (13,15) are the only group to investigate positional differences in whole-body sweating. They showed that linemen (total body mass = 133 ± 15 kg, BSA = 2.62 ± 0.16 m2) produced ∼1 L·h−1 more sweat than physically smaller backs (total body mass = 89 ± 5 kg, BSA = 2.13 ± 0.10 m2) (13). However, these were on-field measurements during an actual NFL training camp, and as such, exercise intensity was likely greater in the heavier linemen in comparison with the lighter backs. Both whole-body and local sweat rate (LSR) in a fixed environment are directly influenced by the rate of evaporation required for heat balance (Ereq), which, in turn, is primarily determined by the rate of metabolic heat production per unit surface area (2,17,26,36). It follows that regional sweating responses between linemen and backs cannot be independently compared unless both groups are exercising at the same rate of metabolic heat production per unit BSA under the same environmental conditions (10).

Given that the number of total sweat glands across the body is determined by the age of 2 yr (25), mass changes that occur beyond this point will not be coupled with a further increase in the total number of sweat glands. Therefore, in adults, the density of sweat glands decreases linearly with increasing BSA (3). Although the diameter and capacity of each sweat gland may be much larger in individuals with a greater BSA (34), a lower evaporative efficiency (i.e., the percentage of sweat produced that evaporates) may occur in larger players by virtue of a lower overall coverage of skin with sweat, resulting in a greater required sweat production to achieve a given amount of evaporation and possibly greater increases in Tcore.

The aim of the present study was to investigate whether differences in regional upper body sweat rates between football linemen and backs exist independently of differences in metabolic heat production. It was hypothesized that at the same metabolic heat production per unit BSA, significantly greater sweating and Tcore would still be observed in larger linemen relative to smaller backs indicating a lower evaporative efficiency in linemen.

Back to Top | Article Outline



After approval of the experimental protocol by the University of Ottawa Research Ethics Committee and University of South Florida (USF) Biomedical Institutional Review Board and after receiving informed consent, 12 NCAA Division I football players volunteered to participate in the study. For a comparison between the largest and the smallest players, six linemen (two offensive guards, two offensive tackles, and two defensive tackles) and six backs (two wide receivers, two linebackers, one safety, and one cornerback) were recruited on the basis of position and body mass. Total body mass, height, and body fat percentage were taken before all experimental testing. Body fat percentage was determined using air displacement plethysmography (BOD POD®; Life Measurements, Inc., Concord, CA). BSA was estimated according to the equation provided by DuBois and DuBois (7). Mean participant characteristics are presented in Table 1.

Back to Top | Article Outline


Core Temperature

Tcore was measured using a telemetric pill sensor (HQ, Inc., Palmetto, FL). Pills were ingested the night before testing just before going to bed at ∼11:00 p.m. for trials scheduled for earlier than noon the following day and at ∼7:00 a.m. the same day for trials scheduled for later than noon. This allowed appropriate transit time for the pill to be located within the intestine when the participant arrived for testing (5,11,33). Pills were administered to players by the USF athletic training staff to ensure the proper procedure was followed. Tcore data were sampled every 15 s using a telemetric receiver (Data Recorder w/HR HT130042; HQ, Inc.) and were subsequently uploaded to a laptop computer (ThinkPad T60; Lenovo, North York, Ontario, Canada) at the end of the experimental trial.

Back to Top | Article Outline

Mean skin temperature

Skin temperature was measured at four points (chest, shoulder, thigh, and calf) using a wireless adhesive temperature sensor (iButton DS1922L; Embedded Data Systems, Lawrenceburg, KY) secured to the skin surface using surgical tape (e.g., Blenderm; 3M, London, Ontario, Canada). Each location was first shaved to remove any excess body hair and then wiped down with an alcohol swab before the placement of the iButton sensor. Mean skin temperature (Tsk) was estimated using the following weighted regional proportions: shoulder (30%), chest (30%), quadriceps (20%), and back calf (20%) (32). The iButton system is a self-contained unit, and therefore, no real-time data were displayed during the trial. Skin temperatures were sampled every 15 s, and data were uploaded immediately after the trial. Although often used in the food industry and geological studies, iButtons have been recently validated for use in physiological applications (37,39).

Back to Top | Article Outline

Local Sweat Rate

LSR was measured using a modified version of the technical absorbent method previously reported by Havenith et al. (19) and Smith and Havenith (38). Before each experimental trial, a set of absorbent material patches (number 2164, Laminated Air-laid; Technical Absorbents, Ltd., Grimsby, UK) was cut to size, individually placed in airtight ziplock freezer bags, marked, and then weighed to the nearest 0.1 mg using a precision scale (Mettler Toledo AE240; Mississauga, Ontario, Canada). A total of five distinct sweat zones were measured: forehead (LSRhead) measured above the eyebrows and below the hairline at the midline of the nose, forearm (LSRarm) measured at the midpoint between the medial epicondyle of the humerus and the styloid process of the ulna on the supinated forearm, chest (LSRchest) measured 5 cm below the clavicle at the center of left pectoral muscle, shoulder (LSRshoulder) measured 8 cm above the inferior angle of the left scapula on the infraspinous fossa, and lower back (LSRlowback) measured 6 cm above the iliac crest and 5 cm to the left of the lumbar spinous processes. The surface area of all patches was 64 cm2 (8.0 cm × 8.0 cm) with the exception of that used for the forehead, which was 36 cm2 (6.0 cm × 6.0 cm) to accommodate for the smaller sampling site. Surrounding each sweat patch was a 1-cm-wide frame to prevent contamination of the sweat sample caused by the dripping of sweat from adjacent areas on the body.

For each sample period, the contact site was first wiped with a dry towel to ensure that no moisture remained on the skin in the sampling area. The sweat patch (with frame) was then removed from the airtight bag and placed on the appropriate site and subsequently held in place with a light compression garment. A headband (Skull Wrap, product 8000072; Under Armour, Inc., Baltimore, MD), forearm sleeve (Forearm Shiver, product 8000033; Under Armour, Inc.), and T-shirt (Compression Full T, product 1000039; Under Armour, Inc.) were used for the forehead, forearm, and torso, respectively, and ensured an even distribution of pressure. After exactly 5 min, the absorbent patch was removed from the surface of the skin, the frame was discarded, and the patch was placed in its own individualized prelabeled ziplock freezer bag. After the conclusion of the trial, each absorbent patch was weighed inside its ziplock bag to ensure that no moisture escaped, and LSR for each site was calculated using the difference between pre- and postapplication weight (patch with no frame + bag) divided by the surface area of the patch (forehead = 36 cm2, forearm = 64 cm2) and the duration of application (5 min), giving values in milligrams per square centimeter per minute (mg·cm−2·min−1).

Back to Top | Article Outline

Metabolic data

Metabolic data were collected using a portable indirect calorimetry unit (model K4b2; COSMED, Chicago, IL) (8,30) and displayed in real time on a laptop (ThinkPad T60; Lenovo). Subjects were equipped with a face mask that covered both the mouth and the nose and were instructed to breathe normally. Metabolic measurements were collected continuously during the 60-min exercise protocol with the exception of the three LSR sample periods at 10, 30, and 50 min of exercise.

Back to Top | Article Outline


HR was monitored continuously throughout the trial using a Polar RS400X coded transmitter (Polar Electro Canada Inc., Lachine, Canada) and stored in the aforementioned indirect calorimetry unit.

Back to Top | Article Outline


Participants were transported in an air-conditioned vehicle to the laboratory situated at the School of Public Health at USF, Tampa, FL, in a well-hydrated state and having consumed a small meal. The number of linemen and backs being tested in the morning and afternoon sessions was balanced to ensure that no order effect for time of day was present. When participants first arrived, they were asked to change into the appropriate clothing (shorts, socks, and running shoes (i.e., 0.1 clo)), and a researcher then verified that the telemetric pill was present within their body by placing the telemetric receiver near the lower back. Once the pill was located, participants were then instrumented with the four skin temperature sensors, and body weight was recorded. Next, participants entered the thermal chamber (tdb [dry bulb temperature] = 32.4°C ± 1.0°C, twb [wet bulb temperature] = 26.3°C ± 0.6°C, v [air velocity] = 0.9 ± 0.1 m·s−1) and were equipped with the portable indirect calorimetry system while sitting on an upright bike ergometer. Participants rested for 5-min to obtain baseline values for both metabolic and temperature data after which they began cycling for 60 min at a fixed cadence of 80 rpm at an external workload sufficient to elicit a fixed heat production of 350 W·m−2.

There were three separate sweat sampling periods during each trial. Each individual LSR measurement lasted exactly 5 min. The measurements on the forehead and forearm commenced after 10, 30, and 50 min of exercise, whereas the measurements on the chest and upper and lower back commenced 90 s after those on the forearm and forehead, i.e., after 16.5, 31.5, and 51.5 min of exercise (Fig. 1). The participant continued to cycle throughout each sweat sampling period. After the completion of the torso sweat sample measurements, the metabolic unit was placed back on the participant until the next sweat sample period.

Back to Top | Article Outline


Metabolic energy expenditure

Metabolic energy expenditure (M) was obtained from minute-average values for oxygen consumption (V˙O2) in liters per minute and the RER using equation 1 (29):

where ec is the caloric equivalent per liter of oxygen for the oxidation of CHO (21.13 kJ) and ef is the caloric equivalent per liter of oxygen for the oxidation of fat (19.62 kJ).

Back to Top | Article Outline

External work

External work (W) was regulated and measured directly using an upright cycle ergometer (Ergomedic 828E; Monark Exercise, Langley, WA).

Back to Top | Article Outline

Sensible heat loss

The combined rate of convective (C) and radiative (R) heat exchange at the skin was calculated using equations 2 and 3, respectively (31):

where fcl is clothing area factor (because participants were very lightly clothed, this was assumed to be equal to BSA). tcl is mean temperature of the clothed body in degrees Celsius (assumed to be equal to Tsk). Thermal resistance of clothing was assumed to be negligible because participants were seminude. hc is the convective heat transfer coefficient in watts per square meter per kelvin and was estimated to be 8.3v0.6 (31). hr is the linear radiative heat transfer coefficient in watts per square meter per kelvin. hr was calculated using equation 4 (31):

where ε is the area-weighted emissivity of the clothing body surface (assumed to be 1.0); σ is the Stefan–Boltzmann constant, 5.67 × 10−8 W·m−2·K−4; Ar/BSA is the effective radiative area of the body in square meters (assumed to be 0.70 because the participants were seated); and tr is mean radiant temperature in degrees Celsius and was equal to tdb.

Back to Top | Article Outline

Respiratory heat loss by convection and evaporation

Respiratory heat loss by convection (Cres) and evaporation (Eres) was calculated using equation 5 (31):

where M is metabolic heat production in watts per square meter, Pa is partial pressure of water vapor in the ambient air in kilopascal, and tdb is dry bulb temperature in degrees Celsius.

Back to Top | Article Outline

Maximum evaporative capacity of the ambient environment

Maximum theoretical rate of evaporation possible in the ambient environment (Emax) was calculated using equation 6 (31):

where w is skin wettedness (during maximum evaporation, skin is assumed to be completely wet; therefore, w is equal to 1.0), Pa is the water vapor pressure in the ambient air in kilopascals, Psk,s is the partial water vapor pressure at the skin in kilopascals (assumed to be equal to saturated water vapor pressure at skin temperature), and he is the evaporative heat transfer coefficient in watts per square meter per kilopascal calculated using the Lewis relation (he = 16.5hc) (31). Evaporative heat transfer resistance of clothing was considered to be negligible because the participants were seminude.

Back to Top | Article Outline

Rate of Evaporation Required for Heat Balance

Ereq was calculated using equation 7:

Back to Top | Article Outline

Statistical Analyses

The data were first evaluated using a two-way ANOVA using the fixed factors of positional group (two levels, i.e., linemen (L) and backs (B)) and time (three or four levels, i.e., rest and 10, 30, and 50 min). Dependent variables were Tcore, Tsk, ΔTsk, HR, LSRhead, LSRarm, LSRchest, LSRshoulder, LSRlowback, and the ratio of each LSR relative to the overall average upper body sweat rate. Any significant main effects or interactions between main effects were subjected to post hoc analyses using independent t-tests. Furthermore, metabolic heat production was compared between linemen and backs using an independent t-test. The significance level was set at an α of 0.05 for all comparisons.

Back to Top | Article Outline


Back to Top | Article Outline

Heat balance parameters

Mean heat balance parameters for both linemen and backs throughout 60 min of exercise are given in Table 2. As intended, similar rates of metabolic heat production per unit surface area (M − W) were observed between the L and B groups. Because sensible heat loss (C + R) and respiratory heat loss (Cres + Eres) were similar between the L and B groups, so was the rate of Ereq. Moreover, the maximum theoretical rate of evaporation possible in the ambient environment (Emax) was not different between the L and B groups, assuming both groups reached the same maximum skin wettedness. This resulted in a similar Ereq/Emax ratio.

Back to Top | Article Outline


LSR was significantly greater in the L group relative to the B group for measurements taken on the forearm, shoulder, chest, and forehead (all P < 0.05). However, no significant differences were observed between the L and B groups for LSR measurement on the lower back (P = 0.704). No significant interactions were found between time and player position for any of the LSR measurements; therefore, the differences in LSR between the L and B groups were sustained throughout exercise for arm, shoulder, chest, and forehead measurements, and LSR on the lower back remained similar between the L and B groups throughout exercise. The LSR measured in both the L and B groups after 10, 30, and 50 min of exercise are illustrated in Figure 2.

The relative contribution of each LSR site to the overall mean upper body sweat rate throughout exercise separated according to position group is illustrated in Figure 3. For both the L and the B groups, the LSR on the forehead was almost double the average upper body sweat rate, whereas the LSR on the chest, forearm, and shoulder was ∼0.7 to 0.8 times that of the average upper body sweat rate. The relative contributions of forehead, forearm, shoulder, and chest sweat rates to the average upper body sweat rate were the same in the L and B groups; however, the relative contribution of the lower back to average upper body sweat rate was significantly lower in the L group relative to the B group (P = 0.006).

Back to Top | Article Outline


No significant difference in Tcore was observed between the L and B groups during preexercise rest or during the initial 45 min of exercise. However, a significant interaction between player position and time (P = 0.012) was observed with Tcore significantly higher in the L group relative to the B group during the last 15 min of exercise (P = 0.033). A comparison between the time-dependent changes in Tcore throughout 60 min of exercise of linemen and backs is illustrated in Figure 4A.

Back to Top | Article Outline


No significant main effect of position was found for absolute Tsk; however, analysis did reveal a significant interaction between time and player position (P ≤ 0.001). Although absolute Tsk was not significantly different between the L and B groups throughout 60 min of exercise, the L group demonstrated much greater changes in Tsk from preexercise rest relative to the B group after 10 min (L = 1.20°C ± 0.16°C, B = 0.41°C ± 0.18°C, P < 0.001), 30 min (L = 1.58°C ± 0.22°C, B = 0.74°C ± 0.17°C, P < 0.001), and 50 min (L = 1.52°C ± 0.24°C, B = 0.77°C ± 0.26°C, P < 0.001) of exercise. A comparison between the time-dependent changes in Tsk throughout the 60-min exercise of L and B is illustrated in Figure 4B.

Back to Top | Article Outline


Mean HR responses throughout 60 min of exercise in both the L and B groups are illustrated in Figure 4C. The HR responses were similar between the L and B groups throughout the initial 30 min of exercise; however, a significantly greater HR was observed in the L group relative to the B group after 45 (P = 0.007) and 60 min (P = 0.002) of exercise.

Back to Top | Article Outline


The present study reveals, for the first time, that when metabolic heat production (MW) per unit BSA and therefore the Ereq per unit BSA are fixed between football linemen and backs during exercise in a hot environment (Table 2), linemen demonstrate significantly greater sweat rates on the forehead, chest, shoulder, and forearm relative to backs. Furthermore, despite heat production per unit BSA being the same and heat production per unit mass actually being ∼25% lower in linemen (6.0 ± 0.5 W·kg−1) in comparison with backs (8.2 ± 0.8 W·kg−1) by virtue of their significantly lower surface area-to-mass ratio, Tcore was in fact ∼0.5°C higher in linemen at the end of 60 min of exercise. Although previous research with both NFL and NCAA football programs have demonstrated greater whole-body sweating and larger elevations in Tcore in linemen (12,13,15,16), the experimental conditions were such that previous studies could not identify whether different levels of heat stress were simply due to differences in metabolic heat production or also differences in the potential for heat dissipation. Our data suggest that football linemen have a compromised potential for heat dissipation independently of any difference in metabolic heat production.

Greater end-exercise elevations in Tcore suggest a greater heat storage per unit mass in linemen; therefore, their lower heat production per unit mass must have been paralleled by a reduced heat dissipation despite significantly greater upper body sweating. As such, the proportion of sweat produced that evaporated (sweating efficiency) must have been much less in the L group compared with the B group. Early research shows that by the age of 2 yr, the total amount of sweat glands covering the human body has already been established (25). Although there is interindividual variation in the absolute number of sweat glands, it has been shown that as BSA and adipose tissue increase (typical in football linemen), distension of the skin reduces the number of sweat glands per unit of surface area effectively decreasing sweat gland density (SGD) in larger individuals (3). With a lower SGD, the fraction of the total skin surface that could potentially be covered by secreted sweat (maximum skin wettedness) may be compromised because the spaces between neighboring sweat glands are larger. Sweat gland diameter and subsequently the capacity for sweat production per gland are greater in individuals with a greater BSA (34), and although this could compensate somewhat for a lower SGD in larger individuals in terms of sweat coverage over the skin, it likely elevates the proportion of secreted sweat that does not evaporate, thus contributing further to a reduction in their sweating efficiency. In a given environment, a lower maximum skin wettedness in linemen would directly reduce their maximum theoretical rate of evaporation possible (Emax—equation 6) relative to backs and subsequently increase their Ereq/Emax ratio despite fixing environmental conditions and ensuring a similar Ereq between positional groups (by prescribing the same metabolic heat production per unit BSA). Bain et al. (2) demonstrated that as the actual Ereq/Emax ratio increases, LSR measured on the forehead and forearm increases in parallel, presumably because of a reduction in sweating efficiency (6).

linemen demonstrated a much greater body fat percentage (Table 1), and this could have partially contributed to the observed differences in end-exercise Tcore between position groups. The lower specific heat capacity of fat tissue would have yielded a lower average specific heat of the body in linemen (3.44 kJ·kg−1·°C−1) relative to backs (3.53 kJ·kg−1·°C−1), and therefore, a given amount of heat stored in the body would have theoretically produced a slightly greater change in Tcore in linemen. However, although specific heat was <3% lower in linemen, heat production per unit mass was >25% lower; thus, a greater heat storage per unit mass almost certainly occurred in the linemen group even when accounting for differences in body composition. It could be argued that greater adipose tissue may have contributed to the apparent reduction in the potential for heat dissipation in linemen. However, although a greater subcutaneous fat layer may have played a role in the slightly lower resting skin temperatures (35) observed in linemen before the start of exercise, because of a lower thermal conductivity (23), it is likely that any insulative fat layers were bypassed by blood flow after the onset of exercise (21)—a notion supported by similar absolute skin temperatures between position groups during exercise.

Notwithstanding the apparently lower sweating efficiency in linemen, the potential differences in maximum skin wettedness between positional groups do not explain the physiological origin of the differences in sweating, particularly during the early stages of exercise. It is well established that the magnitude of elevation in skin temperature above baseline values promotes thermoregulatory sweating in the heat at equivalent Tcore (28). In the present study, linemen demonstrated significantly greater increases in Tsk from baseline throughout exercise (Fig. 4B) compared with B. Because Tcore was only statistically different between position groups during the final 15 min of exercise (Fig. 4A), a greater activation in peripheral thermosensors may be the primary cause for the greater LSR observed in linemen. When sweat rate at each location is expressed as relative values compared with the average upper body sweat rate (i.e., mean of all five LSR measurements), linemen do not seem to exhibit a disproportionately greater elevation in sweat rate in any one particular region (Fig. 4). Furthermore, the distribution of relative sweat rates followed the same pattern in both position groups at four of five sites with the LSR on the forehead being almost double (L = 184% ± 26%, B = 174% ± 24%) the average upper body sweat rate, with the arm, chest, and shoulder all exhibiting values that were ∼70%–80% of the average upper body sweat rate. All relative values are not dissimilar from those previously reported in the literature (38) with the exception of the lower back showing lower absolute (L = 1.20 ± 0.36 mg·min−1·cm−2, B = 1.26 ± 0.48 mg·min−1·cm−2) and relative (L = 87% ± 23%, B = 57% ± 12% ) sweat rates in comparison with previous studies. This discrepancy is likely due to our measurement site being situated more toward the lateral aspect of the torso as opposed to directly over the lumbar spine to ensure complete contact between the technical absorbent patch and the skin surface.

Aerobic fitness was likely lower in linemen relative to backs because HR was greater in linemen during the final 30 min of exercise (Fig. 3C) despite oxygen consumption per unit mass being lower in linemen (L = 20.5 ± 1.3 mL O2·kg−1·min−1, B = 28.3 ± 2.6 mL O2·kg−1·min−1). A greater aerobic fitness no doubt facilitates a greater maximum sweat rate (4); however, greater sweating was observed in the position group that was probably less fit. Furthermore, submaximal whole-body (17,26,36) and local (2) sweat rates have been clearly demonstrated to be primarily determined by the metabolic heat production (W·m−2) irrespective of relative exercise intensity (i.e., percentage of V˙O2max) and Tcore, hence the design of the present experiment fixed heat production per unit BSA between position groups.

Although the approach adopted in the present study allows us to identify potential differences in the physiological capacity to dissipate heat, any observed on-field differences in heat stress between players are also likely due to differences in metabolic heat production. No research has thus far been conducted that has accurately measured on-field energy expenditure, and future research is needed to determine the relative contribution of metabolic heat production to the exacerbation of heat stress during live training camp activities. SGD at each sweat sample site was not measured in the present study. Although a greater SGD was assumed in linemen by virtue of their greater BSA using evidence in the literature (3,13,24,25,34), a quantitative measurement of SGD and average sweat production per gland in both linemen and backs would support the proposed mechanism for the difference in sweating efficiency between groups. Future research studies should also include the evaluation of LSR on the lower extremities in both linemen and backs to provide a greater insight into individual differences in heat dissipation potential. The possible influence of the technical absorbent material on both local skin temperature and skin saturation levels is unclear. An increase in skin saturation seems highly unlikely as the single highest observed sweat rate (1213 mg on the forehead in 5 min) equated to only 2.1% of the capacity of the absorbent patch (57,175 mg for a 36-cm2 patch). A greater local skin temperature at the sample site could have occurred, which, in turn, could have influenced LSR; however, this potential effect was minimized because of short sample periods (5 min) and the use of small patch sizes. The fit of the compression T-shirt used to keep the absorbent patches in place and the subsequent pressure exerted on the sweat sample sites may have been slightly different between participants. However, it is unlikely that this fit effect is responsible for any differences observed in LSR between the linemen and backs because compression shirts of different sizes were selected depending upon the size of the participant, and greater LSR were also observed on the forearm and forehead, which were not covered by the compression T-shirt.

The present study has potentially important on-field practical applications because any intervention that would increase the evaporation of sweat without interfering with exercise intensity would be effective in mitigating heat stress risk, particularly in linemen. Because of the stationary nature of typical training camp drills for linemen, their lower self-generated wind velocity likely allows for less evaporative cooling for a given sweat production relative to their smaller counterparts (13). However, our data suggest that even at similar air velocities (self-generated or otherwise), sweating efficiency would be lower in linemen. This position group should therefore be specifically targeted using the strategic placement of large mechanical fans during drills to potentially aid in the evaporation of sweat. Furthermore, because the lower sweating efficiency in linemen is attributed to a lower maximum skin wettedness, any artificial increase in skin wettedness water misting stations (such as those currently used by some organizations) may compensate for this apparent morphologically related disadvantage in linemen and reduce their heat stress risk. However, further investigation is required to assess the efficacy of these on-field interventions.

Back to Top | Article Outline


The present study demonstrated that football linemen sweat significantly more on the torso and head than football backs independently of any differences in metabolic heat production per unit BSA and therefore Ereq. Despite greater sweating and a lower heat production per unit body mass, linemen demonstrated significantly greater elevations in Tcore suggesting that sweating efficiency (i.e., the proportion of sweat that evaporates) was much lower in linemen.

This research was supported by an NFL Charities Medical Research Grant. The authors report no conflicts of interest.

The authors thank the athletic training staff, players, and coaching staff of the USF Football Program as well as Dr. Douglas Casa of the Korey Stringer Institute at the University of Connecticut. They also thank Professor Tom Bernard of the School of Public Health at USF for generously granting them access to the climatic chamber and laboratory facilities and Dr. Candi Ashley for her assistance throughout data collection.

The results of the present study do not constitute endorsement by the American College of Sports Medicine.

Back to Top | Article Outline


1. Armstrong LE, Johnson EC, Casa DJ, et al.. The American football uniform: uncompensable heat stress and hyperthermic exhaustion. J Athl Train. 2010; 45 (2): 117–27.
2. Bain AR, Deren TM, Jay O. Describing individual variation in local sweating during exercise in a temperate environment. Eur J Appl Physiol. 2011; 111 (8): 1599–607.
3. Bar-Or O, Magnusson LI, Buskirk ER. Distribution of heat-activated sweat glands in obese and lean men and women. Hum Biol. 1968; 40 (2): 235–48.
4. Buono MJ, Sjoholm NT. Effect of physical training on peripheral sweat production. J Appl Physiol. 1988; 65 (2): 811–4.
5. Byrne C, Lim CL. The ingestible telemetric body core temperature sensor: a review of validity and exercise applications. Br J Sports Med. 2007; 41 (3): 126–33.
6. Candas V, Libert JP, Vogt JJ. Human skin wettedness and evaporative efficiency of sweating. J Appl Physiol. 1979; 46 (3): 522–8.
7. DuBois D, DuBois EF. A formula to estimate surface area if height and weight are known. Arch Intern Med. 1916; 17: 863.
8. Duffield R, Dawson B, Pinnington HC, Wong P. Accuracy and reliability of a Cosmed K4b2 portable gas analysis system. J Sci Med Sport. 2004; 7 (1): 11–22.
9. Eichner ER. Toward ending fatal heat stroke in football players. J Athl Train. 2010; 45 (2): 105–6.
10. Gagge AP, Gonzales RR. Mechanisms of heat exchange. In: Fregly MJ, Blatteis CM, editors. Handbook of Physiology. Environmental Physiology. Bethesda (MD): American Physiological Society; 1996. p. 45–84.
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. Godek SF, Bartolozzi AR, Burkholder R, Sugarman E, Dorshimer G. Core temperature and percentage of dehydration in professional football linemen and backs during preseason practices. J Athl Train. 2006; 41 (1): 8–14; discussion 14–7.
13. Godek SF, Bartolozzi AR, Burkholder R, Sugarman E, Peduzzi C. Sweat rates and fluid turnover in professional football players: a comparison of National Football League linemen and backs. J Athl Train. 2008; 43 (2): 184–9.
14. Godek SF, Bartolozzi AR, Godek JJ. Sweat rate and fluid turnover in American football players compared with runners in a hot and humid environment. Br J Sports Med. 2005; 39 (4): 205–11.
15. Godek SF, Bartolozzi AR, Peduzzi C, et al.. Fluid consumption and sweating in National Football League and collegiate football players with different access to fluids during practice. J Athl Train. 2010; 45 (2): 128–35.
16. Godek SF, Peduzzi C, Burkholder R, Condon S, Dorshimer G, Bartolozzi AR. Sweat rates, sweat sodium concentrations, and sodium losses in 3 groups of professional football players. J Athl Train. 2010; 45 (4): 364–71.
17. Gonzalez RR, Cheuvront SN, Montain SJ, et al.. Expanded prediction equations of human sweat loss and water needs. J Appl Physiol. 2009; 107 (2): 379–88.
18. Havenith G, Coenen JM, Kistemaker L, Kenney WL. Relevance of individual characteristics for human heat stress response is dependent on exercise intensity and climate type. Eur J Appl Physiol Occup Physiol. 1998; 77 (3): 231–41.
19. Havenith G, Fogarty A, Bartlett R, Smith CJ, Ventenat V. Male and female upper body sweat distribution during running measured with technical absorbents. Eur J Appl Physiol. 2008; 104 (2): 245–55.
20. Havenith G, Luttikholt VG, Vrijkotte TG. The relative influence of body characteristics on humid heat stress response. Eur J Appl Physiol Occup Physiol. 1995; 70 (3): 270–9.
21. Havenith G, van Middendorp H. The relative influence of physical fitness, acclimatization state, anthropometric measures and gender on individual reactions to heat stress. Eur J Appl Physiol Occup Physiol. 1990; 61 (5–6): 419–27.
22. Hitchcock KM, Millard-Stafford ML, Phillips JM, Snow TK. Metabolic and thermoregulatory responses to a simulated American football practice in the heat. J Strength Cond Res. 2007; 21 (3): 710–7.
23. Jéquier E, Gygax PH, Pittet P, Vannotti A. Increased thermal body insulation: relationship to the development of obesity. J Appl Physiol. 1974; 36 (6): 674–8.
24. Kondo N, Takano S, Aoki K, Shibasaki M, Tominaga H, Inoue Y. Regional differences in the effect of exercise intensity on thermoregulatory sweating and cutaneous vasodilation. Acta Physiol Scand. 1998; 164 (1): 71–8.
25. Kuno Y. Human Perspiration. Springfield (IL): Charles C Thomas; 1956. 268 p.
26. Lustinec K. Sweat rate, its prediction and interpretation. Arch Sci Physiol (Paris). 1973; 27 (2): 127–36.
27. Mathews DK, Fox EL, Tanzi D. Physiological responses during exercise and recovery in a football uniform. J Appl Physiol. 1969; 26 (5): 611–5.
28. Nadel ER, Bullard RW, Stolwijk JA. Importance of skin temperature in the regulation of sweating. J Appl Physiol. 1971; 31 (1): 80–7.
29. Nishi Y. Measurement of thermal balance in man. In: Cena K, Clark JA, editors. Bioengineering, Thermal Physiology and Comfort. New York (NY): Elsevier; 1981. p. 29–39.
30. Parr BB, Strath SJ, Bassett DR, Howley ET. Validation of the Cosmed K4b2 portable metabolic measurement system. Med Sci Sports Exerc. 2001; 33 (5): S300.
31. Parsons KC. Human Thermal Environments. 2nd ed. London, UK: Taylor & Francis; 2003. 527 p.
32. Ramanathan NL. A new weighting system for mean surface temperature of the human body. J Appl Physiol. 1964; 19: 531–3.
33. Rav-Acha M, Heled Y, Slypher N, Moran DS. Core body temperature monitoring using the telemetric pill. Harefuah. 2003; 142 (3): 197–202, 238. [Hebrew].
34. Sato K, Sato F. Individual variations in structure and function of human eccrine sweat gland. Am J Physiol. 1983; 245 (2): R203–8.
35. Savastano DM, Gorbach AM, Eden HS, Brady SM, Reynolds JC, Yanovski JA. Adiposity and human regional body temperature. Am J Clin Nutr. 2009; 90 (5): 1124–31.
36. Shapiro Y, Pandolf KB, Goldman RF. Predicting sweat loss response to exercise, environment and clothing. Eur J Appl Physiol Occup Physiol. 1982; 48 (1): 83–96.
37. Smith AD, Crabtree DR, Bilzon JL, Walsh NP. The validity of wireless iButtons and thermistors for human skin temperature measurement. Physiol Meas. 2010; 31 (1): 95–114.
38. Smith CJ, Havenith G. Body mapping of sweating patterns in male athletes in mild exercise-induced hyperthermia. Eur J Appl Physiol. 2011; 111 (7): 1391–404.
39. van Marken Lichtenbelt WD, Daanen HA, Wouters L, et al.. Evaluation of wireless determination of skin temperature using iButtons. Physiol Behav. 2006; 88 (4–5): 489–97.
40. Yeargin SW, Casa DJ, Armstrong LE, et al.. Heat acclimatization and hydration status of American football players during initial summer workouts. J Strength Cond Res. 2006; 20 (3): 463–70.


©2012The American College of Sports Medicine