A retrospective analysis of American football hyperthermia deaths in the United States
- First Online:
- Cite this article as:
- Grundstein, A.J., Ramseyer, C., Zhao, F. et al. Int J Biometeorol (2012) 56: 11. doi:10.1007/s00484-010-0391-4
- 326 Views
Over the period 1980–2009, there were 58 documented hyperthermia deaths of American-style football players in the United States. This study examines the geography, timing, and meteorological conditions present during the onset of hyperthermia, using the most complete dataset available. Deaths are concentrated in the eastern quadrant of the United States and are most common during August. Over half the deaths occurred during morning practices when high humidity levels were common. The athletes were typically large (79% with a body mass index >30) and mostly (86%) played linemen positions. Meteorological conditions were atypically hot and humid by local standards on most days with fatalities. Further, all deaths occurred under conditions defined as high or extreme by the American College of Sports Medicine using the wet bulb globe temperature (WBGT), but under lower threat levels using the heat index (HI). Football-specific thresholds based on clothing (full football uniform, practice uniform, or shorts) were also examined. The thresholds matched well with data from athletes wearing practice uniforms but poorly for those in shorts only. Too few cases of athletes in full pads were available to draw any broad conclusions. We recommend that coaches carefully monitor players, particularly large linemen, early in the pre-season on days with wet bulb globe temperatures that are categorized as high or extreme. Also, as most of the deaths were among young athletes, longer acclimatization periods may be needed.
KeywordsAmerican footballHyperthermiaClimateUnited States
Over the last 50 years (1960–2009), there have been 123 documented cases of American-style football players in the United States dying from heat-related illnesses (Mueller and Colgate 2010), including 58 well-documented cases from 1980–2009. Football players are particularly susceptible to heat-related illnesses in part because of the timing of pre-season practice, often in the late summer when conditions are hot and humid. A climate study of optimal practice times in Alabama found that there were no suitable times for outdoor practices in full uniform in August (Francis et al. 1991). Additionally, the level of clothing football players wear may increase metabolic heat production and inhibit cooling (e.g., Fox et al. 1966; Mathews et al. 1969; Kulka and Kenney 2002; McCullough and Kenney 2003; Brothers et al. 2004; Armstrong et al. 2010). A full football uniform, for example, may impede evaporative heat loss by 60–70% (Mathews et al. 1969). Finally, football players tend to have physical characteristics that may increase heat storage, such as greater muscle mass, lower body surface area to body mass ratios, and higher body fat percentages (Godek et al. 2004).
A variety of safety guidelines for practicing under different environmental conditions have been established by organizations such as Sports Medicine Australia, The American College of Sports Medicine, and the American Academy of Pediatrics (Armstrong et al. 1996; Binkley et al. 2002; American Academy of Pediatrics, Committee on Sports Medicine and Fitness 1991, 2000; Sports Medicine Australia 2010). These guidelines rely on a derived variable called the wet bulb globe temperature (WBGT; Yaglou and Minard 1957) that integrates the influence of air temperature, humidity, and solar radiation. Depending on the threshold, recommendations are made about the level of risk and whether changes in how the practice is conducted are needed. Cooper et al. (2006), however, noted that these standards are generic and do not account for the particular sport, the equipment worn, or fitness of the athlete.
Football-specific thresholds based on air temperature and humidity were established for different levels of dress including full pads, light pads, and shorts only by Kulka and Kenney (2002). Using subjects that walked on a treadmill wearing different types of clothing and under different meteorological conditions, they identified levels of uncompensable heat stress in which the body becomes unable to maintain a thermal balance. A limitation of this method is that it makes assumptions about the level of activity and does not account for other meteorological factors such as solar radiation or wind that may affect exertional heat stress (Kulka and Kenney 2002). Coyle (2003) expressed thresholds for uncompensable heat stress for the different levels of dress in terms of WBGT using empirical relationships between the black globe temperature and dry bulb temperature.
Several studies have examined environmental conditions and health outcomes in football players. Cooper et al. (2006) investigated the use of the WBGT and safety thresholds established by the American College of Sports Medicine (ACSM) at five universities in the southeastern United States. They noted that the highest risk of heat-related illness was in the first few weeks of practice in August. Most illnesses (97%) occurred under high-risk or extreme-risk categories. The clothing-specific standards of Kulka and Kenney (2002) were studied during football training sessions. Godek et al. (2004) noted that in many cases players in partial or full football uniforms were able to train at levels that exceeded thresholds without any signs of heat-related illnesses because periods of inactivity allowed the athletes to cool and maintain a thermal balance.
In particular, the meteorological conditions during football hyperthermia deaths have rarely been examined. Only two studies have performed retrospective analyses relating weather conditions to hyperthermia deaths in football players. Fox et al. (1966) studied nine deaths from 1959 to 1962. They noted that all were interior linemen in full football uniforms and most died during the first 2 days of practice. Deaths occurred under a range of conditions from higher temperatures and lower relative humidities to lower temperatures and higher relative humidities. Kulka and Kenney (2002) added data on three deaths in 2001 to Fox et al.’s study and placed these values in the context of risk thresholds that accounted for level of clothing. They found that all cases fell at or above the level of uncompensable heat stress for exercise in a full football uniform.
Where do football hyperthermia deaths occur?
When do deaths occur (annual, intra-seasonally, time of day, time of practice season)?
Which athletes (age, position, weight, level of dress) are most vulnerable?
Under what meteorological conditions and safety thresholds do the deaths occur?
Materials and methods
Data on heat-related deaths of football players were obtained from the National Center for Catastrophic Injury Research (NCCIR; Mueller and Colgate 2010). The dataset includes information on the death date, age of deceased, height, weight, school, city, state, and the sport being played that caused the death. The first death recorded in the database occurred in 1955 and the last in 2009. We used 1980–2009 as a study period because deaths before 1980 lacked information in the majority of the data fields.
In order to compile a more complete record of the data, extensive research was conducted to fill in missing data in the NCCIR dataset and to add additional information. It is important to note, however, that no additional deaths were added to the original NCCIR database. Media reports were the principal source of information but lawsuit publications and obituaries were also used. Most of the news articles were obtained using the Lexis–Nexis academic database. Local newspapers were contacted for several cases where no information was found in the Lexis–Nexis database.
Additional data fields were added to the existing database that would allow better analysis of the conditions under which the death occurred. These data fields included the exposure date, time of day the athlete was practicing or participating in football-related outdoor exercise, the point in the football season the incident occurred (first day of practice, etc.), position, pre-existing medical conditions, type of activity (practice, conditioning workout, etc.), clothing worn, and latitude/longitude coordinates of exposure location. The exposure date is a critical field for assessing the meteorological conditions that were present during the onset of hyperthermia symptoms. In many cases, the exposure date and death date were different. Clothing ensembles were categorized according to Kulka and Kenney (2002) as full football uniform (helmet, undershirt, shoulder pads, jersey, and game pants with thigh, knee, and hip pads), practice uniform (helmet, undershirt, shoulder pads, jersey, and shorts), and shorts only.
Meteorological data were used to assess the environmental conditions under which the fatalities occurred. All hourly observations of temperature, dew-point temperature, relative humidity, wind speed, and cloud cover were from the Integrated Surface Hourly (ISH) dataset (Lott et al. 2001). However, we retrieved data for 1991–2005 from the National Solar Radiation Database (NSRDB) because of ease of use of the dataset (NREL 2007). Meteorological observations were obtained from the nearest station to the location where the exposure occurred. Fatalities were only examined, however, if an observing station was within 50 km and recorded data during the approximate exposure time. The average distance between the station and the heat-related death was 19.2 km, with a standard deviation of 13.8 km.
We reconstructed meteorological conditions during the approximate exposure period. While the exact exposure time was not available for the majority of the cases, whether it occurred at a morning or afternoon practice was often indicated. Thus, two time periods were defined: a morning session was assumed to occur somewhere between 8:00 am and 12:00 pm local daylight time (LDT), and an afternoon session was assumed to occur somewhere between 2:00 and 6:00 pm LDT. For these time periods, the maximum temperature, relative humidity and dew-point temperature at the time of maximum temperature, the average temperature, average dew-point temperature, average relative humidity, and average cloud cover were calculated. Two derived meteorological variables, the WBGT and the Heat Index (HI), were computed using the average morning or afternoon data. As noted earlier, WBGT is a widely used measure of environmental conditions assessed for safety standards in athletics and industry (Cooper et al. 2006). We used the software package Heat Stress Advisor (Zunis Foundation 2010) which uses algorithms developed by Coyle (2000) to convert standard meteorological variables (e.g., air temperature, relative humidity, cloud cover) to the wet bulb globe temperature (WBGT). The Heat Index is based on the apparent temperature by Steadman (1979) and modified for operational purposes by the U.S. National Weather Service (NWS). The NWS issues heat-health warnings based on the HI and it is therefore familiar to people in the United States. An empirical equation by Schoen (2005), requiring air and dew-point temperatures, is used to compute the HI.
Spatial and temporal patterns
Temporal patterns of football hyperthermia fatalities
Time of day
Time of season
Morning (8:00 am–12:00 pm LDT)
Afternoon (2:00–6:00 pm LDT)
1st practice with pads
Characteristics of football hyperthermia fatalities. Percent of High School (HS) linemen is rounded to the nearest percent and based on data in Laurson and Eisenmann (2007)
All linemen n
Linemen ≤18 years n (%/%HS linemen)
Obese Class I (30–34)
Obese Class II (35–39)
Obese Class III (≥40)
For further analysis, the positions were aggregated as linemen (offensive and defensive), backs (two fullbacks, one running back, one defensive back), and other (one wide receiver). Among American-style football players, offensive and defensive linemen tend to be the largest players on the field with higher levels of body fat percentage and greater absolute strength (Pincivero and Bompa 1997). Offensive and defensive linemen tend to mirror each other by position, with offensive linemen attempting to protect the passer and clear the way for runners to advance the ball while defensive linemen try to disrupt the passer and tackle any running backs before they can gain yardage. Offensive and defensive backs along with wide receivers tend to have lower body fat levels and lower absolute strength, but are faster and have greater aerobic capacity (Pincivero and Bompa 1997).
In our dataset, linemen had an average weight (BMI) of 120.07 kg (34.75) compared with 105.91 kg (31.85) for backs. For reference, we compared the BMI of 23 linemen ≤18 years of age with an extensive dataset of 3,683 high school linemen compiled by Laurson and Eisenmann (2007) (Table 2). The linemen who died from heat-related illnesses were large, with approximately 87% classified as obese (BMI >30) compared with 30% in the Laurson and Eisenmann (2007) dataset. Also, linemen represent a large and disproportionate percentage of deaths (86%) relative to their numbers on the field (45%). Finally, we considered the level of dress of the athletes. Of the 20 cases with data, 8 were in in shorts only, 8 were in practice gear, and 4 were in full football uniforms.
Average meteorological conditions on days with hyperthermia deaths
Air temperature (°C)
Relative humidity (%)
Dew-point temperature (°C)
Heat index (°C)
Afternoon (14 cases)
Morning (19 cases)
Statistics on football hyperthermia fatalities by risk category
American College of Sports Medicine and Sports Medicine Australia
American Academy of Pediatrics
NWS Heat Index
Uncompensable heat stress
No limits (<24°C)
No warning (<26.7°C)
Full uniform (24.7–28.4°C)
Longer rest periods (24–25.9°C)
Practice uniform (28–29.6°C)
Stop activity if not acclimatized (26–29°C)
Extreme caution (32.3–40.6°C)
Shorts only (31.6–33.1°C)
Extreme/Cancel event (>28°C)
Cancel event (>29°C)
Extreme danger (>54.4°C)
Integrated exposure conditions over the morning or afternoon for clothing specific standards were investigated using WBGT thresholds adapted by Coyle (2003) from the temperature and relative humidity thresholds of Kulka and Kenney (2002). The WBGT categories involve a range of values because of the interaction between temperature and relative humidity. Nevertheless, the results using the lower and upper WBGT thresholds for each clothing category were identical (Table 4). Here, one of the two athletes in a full uniform practiced in conditions that exceeded the level of uncompensable heat stress, five of the eight athletes in practice gear were exposed to conditions that exceeded the threshold, and only two athletes in shorts exceeded the threshold.
Discussion and conclusions
From 1980 to 2009, 58 football players died from heat-related illnesses. Our results create a temporal, demographic, and meteorological profile of hazardous conditions for football hyperthermia deaths. Consistent with findings by Cooper et al. (2006) and Fox et al. (1966), heat-related injuries are most prominent in the first half of August, when unconditioned and unacclimatized players begin pre-season practice under hot and humid conditions. Over half the deaths occur during morning practices which, although cooler, have higher humidity levels that increase heat stress. The level of dress has been a concern in studies of football players and heat stress as helmets and pads can inhibit cooling mechanisms. Many of the deaths, however, were among players wearing no pads and minimal clothing. Finally, consistent with Fox et al. (1966), large linemen, especially those with BMIs exceeding 30, are disproportionately represented among the deaths.
Meteorologically, the fatalities occurred on days that tended to be unusually hot and humid by local standards. More importantly, the conditions were categorized as extremely dangerous (i.e., high or extreme risk) for heat-related illnesses by the American College of Sports Medicine using the WBGT. The HI, however, tended to underestimate the degree of hazard as most deaths occurred under conditions that were assigned lower risk levels. Football-specific thresholds for uncompensable heat stress were a useful metric for athletes in practice uniforms but less so for those dressed in shorts only. Indeed, most of the deaths for athletes in shorts occurred below the level of uncompensable heat stress. A possible explanation for this discrepancy is that the deaths for athletes wearing only shorts all occurred early in the pre-season (all on or before 8 August), suggesting that a lack of acclimatization may have made the players more vulnerable to heat-related illnesses. Too few data on athletes wearing full pads were available to draw any definitive conclusions.
Coaches should carefully monitor players, particularly big linemen, early in the practice season. The complete suite of meteorological conditions, not merely temperature, should be accounted for when deciding whether to practice and the intensity of the practice. In particular, coaches should be aware that simply holding practice in the morning does not alleviate hyperthermia threats due to higher humidity levels. Derived measures of heat stress such as the WBGT, along with established safety thresholds from organizations like the American College of Sports Medicine, are useful guides for assessing levels of hazard and in dictating activity levels. Finally, the great number of deaths early in the pre-season practice schedule suggests that proper acclimatization is important. Most of the deaths were among youth athletes who may have less physical preparation or opportunity for acclimatization than college-level or professional athletes (Bergeron et al. 2005). Thus, longer acclimatization periods may be an important strategy for reducing the number of hyperthermia-related deaths. Bergeron et al. (2005) offer suggested acclimatization plans for both high school and youth football players. At high school level, for instance, they suggest a 14-day acclimatization plan where coaches carefully adjust the level of dress, and length and intensity of practice sessions as the players gain fitness and adapt to the ambient meteorological conditions.
We would like to thank Dr. Fred Mueller of the National Center for Catastrophic Injury Research for kindly providing data on heat-related deaths of football players.