604 FOOTBALL HELMETS Lewis et al. FOOTBALL HELMETS AND BLUNT HEAD INJURIES Do Football Helmets Reduce Acceleration of Impact in Blunt Head Injuries? LAWRENCE M. LEWIS, MD, ROSANNE NAUNHEIM, MD, JOHN STANDEVEN, PHD, CARL LAURYSSEN, MD, CHRIS RICHTER, MD, BRIAN JEFFORDS, MD Abstract. Several recent studies suggest that acceleration of the head at impact during sporting activities may have a detrimental effect on cognitive function. Reducing acceleration of impact in these sports could reduce neurologic sequelae. Objective: To measure the effectiveness of a regulation football helmet to reduce acceleration of impact for both low- and moderate-force impacts. Methods: An experimental paired study design was used. Male volunteers between 16 and 30 years of age headed soccer balls traveling approximately 35 miles per hour bareheaded and with a helmet. An intraoral accelerometer worn inside a plastic mouthpiece measured acceleration of the head. The helmet also had an accelerometer placed inside the padding. For more forceful impacts, cadaver heads, both with and without helmets, were instrumented with intraoral (IO) and intracranial (IC) accelerometers and struck with a pendulum device. Simultaneous IO and IC accelerations were measured and compared between helmeted and unhelmeted cadaver heads. The main outcome was mean peak acceleration of the head and/or brain associated with low- and moderate-force impacts with and without protective headgear. Results: Mean peak Gs, measured by the mouthpiece accelerometer, were significantly reduced when the participants heading soccer balls were wearing a helmet (7.7 Gs with vs 19.2 Gs without, p = 0.01). Wearing a helmet also significantly lowered the peak Gs measured intraorally and intracranially in cadavers subjected to moderate-force pendulum impacts: 28.7 Gs with vs 62.6 Gs without, p < 0.001; and 56.4 Gs with vs 81.6 Gs without, p < 0.001, respectively. Conclusions: A regulation football helmet substantially reduced the peak Gs associated with heading a soccer ball traveling at moderately high velocities. A helmet was also effective in reducing the peak acceleration both intraorally and intracranially for impacts significantly more forceful than heading a soccer ball. Key words: football helmets; sports injuries; head injuries; acceleration. ACADEMIC EMERGENCY MEDICINE 2001; 8:604 609 HELMETS have been shown to reduce the risk of significant intracranial injury in motorcycle and bicycle riders. 1 3 Furthermore, motorcycle and bicycle helmets along with football and hockey helmets reduce the acceleration experienced by metallic headforms undergoing impact testing. 4 What has not been demonstrated, to the best of our knowledge, is whether these helmets reduce the peak acceleration of the head in humans participating in various sporting activities and, if so, by what factor. This study was undertaken to measure and quantify the impact-attenuating properties of regulation football helmets From the Emergency Medicine Division, Washington University School of Medicine (LML, RN, CR, BJ), Human Performance Laboratory, Barnes-Jewish Hospital (JS), and Department of Neurosurgery, Washington University School of Medicine (CL), St. Louis, MO. Received September 1, 2000; revisions received December 11, 2000, and December 22, 2000; accepted January 8, 2001 Address for correspondence and reprints: Lawrence M. Lewis, MD, Division of Emergency Medicine, Washington University Physicians, Washington University Medical Center, Campus Box 8072, 660 South Euclid Avenue, St. Louis, MO 63110-11093. Fax: 314-362-0478; e-mail: lewisl@msnotes.wustl.edu when worn by human volunteers or cadavers undergoing head accelerations of a magnitude that may be experienced during sporting activities. METHODS Study Design. We performed an experimental paired study design measuring the peak acceleration of impact experienced by volunteers when heading a soccer ball, both without any head protection and with a regulation football helmet. All volunteers gave informed consent to participate in this study. Parental consent was also obtained for all minors. The study was approved by the Human Studies Committee at Washington University. Study Setting and Population. Three male volunteers between the ages of 16 and 30 years with high-school level soccer experience who were familiar with the correct technique of heading a soccer ball were used for this study. Balls were kicked by one participant from mid-field approximately 30 yards from the participant who was expected to head the ball.
ACADEMIC EMERGENCY MEDICINE June 2001, Volume 8, Number 6 605 Study Protocol. The velocity of the kicked soccer ball was measured using a hand-held radar gun by a member of our local law enforcement agency, at a point approximately 10 feet from impact. Participants who were heading the ball were instrumented with a PCB triaxial accelerometer (PCB Piezotronics, Inc., Depew, NY), which was inserted into a standardized plastic mouthpiece. For helmeted impacts, the participants had a second triaxial accelerometer (IS-100 Techmark, Lansing, MI) placed in the padding of the helmet, near the crown of the head, just touching the participant s hair. Peak accelerations during impact were recorded and measured by both accelerometers in three orthogonal directions: the x-axis (anteroposterior); the y-axis, perpendicular to the x-axis and lying in a plane that bisects the head through the ears; and the z-axis, which lies in the plane running from the feet through the crown of the head. The vector sum of all three accelerations was calculated and displayed for both accelerometers (Fig. 1). In order to better determine the intrinsic differences in the acceleration seen by an accelerometer in helmet padding from that in a mouthpiece, one helmeted volunteer and two instrumented cadaver heads underwent blunt impacts with a pendulum device. The pendulum consisted of a 110- centimeter (cm) long cord with a bowling ball weighing 6.8 kilograms tethered at the end. For the volunteer, the pendulum was allowed to swing freely under the influence of gravity from a position 45 degrees from neutral, a chord distance (from helmet surface to bowling ball surface) of 65 cm. For cadaver heads, the cord length from the ceiling had to be 115 cm. The resulting impacts were from 40 degrees (68 cm) and from 51 degrees (89 cm). Unhelmeted cadaver heads had accelerations measured with an intraoral accelerometer. Finally, cadaver heads were also instrumented with an intracranial accelerometer attached to the sphenoid wing. The peak Gs recorded from this accelerometer were used to compare simultaneous intracranial acceleration with that measured by the mouthpiece accelerometer. Data Analysis. Because the data were not normally distributed, peak accelerations of impact as measured by the intraoral accelerometer were compared between the helmeted and the unhelmeted participants using the Mann-Whitney rank sum test. An alpha value less than 0.05 was considered significant. For the cadaver study, pre-impact velocities were calculated using standard physical equations. These impacts resulted in accelerations that could be reasonably reproduced and calculated. The peak accelerations that resulted from these impacts, as measured by the accelerometer in the helmet, were compared with those measured by the accelerometer in the mouthpiece using a paired t-test. Differences in the measured mean peak acceleration Figure 1. The vector summation of accelerations experienced in all three axes when heading a soccer ball.
606 FOOTBALL HELMETS Lewis et al. FOOTBALL HELMETS AND BLUNT HEAD INJURIES TABLE 1. Mean Peak Gs (Measured Intraorally and in the Helmet Padding) Experienced When Heading a Soccer Ball and with Pendulum Impacts Both with and without a Helmet Helmeted Soccer Player Unhelmeted Soccer Player Helmeted Pendulum Impact Peak Gs (intraoral) 7.7 19.2 15.0 Peak Gs (helmet padding) 49.3 N/A 58.1 between helmeted and nonhelmeted cadaver heads were determined using a two-way analysis of variance. Differences in the mean peak accelerations measured by the helmet accelerometer and the mouthpiece accelerometer were used to infer quantitative differences between the actual accelerations experienced in these two locations. Comparisons were performed using a paired t test. All comparisons were considered significant at the 0.05 level. RESULTS Pre-impact velocity measurements of the soccer ball were very consistent, averaging 39.3 ( 1.8) miles per hour. The mean peak acceleration measured in the helmet padding of helmeted volunteers heading a soccer ball traveling at the above velocity was 49.3 ( 1.2) Gs. Concurrent peak Gs as measured in the oral cavity were 7.7 ( 1.0) Gs (p < 0.001). The mean peak acceleration, as measured intraorally, was significantly lower in helmeted participants than in those without a helmet: 7.7 Gs ( 1.0) vs 19.2 ( 2.2) Gs, respectively (p = 0.01) (Table 1). The data from the pendulum impacts are as follows: Mean peak acceleration as measured in the helmet padding in the human volunteer trials was 58.1 ( 2.0) Gs vs 15.0 ( 0.1) Gs measured intraorally by the mouthpiece accelerometer (p < 0.04). Simultaneous measurement of intracranial and intraoral accelerations in the cadaver model can be seen in Table 2. There were statistically significant differences seen between intraoral and intracranial measurements of mean peak Gs, with intraoral measurements demonstrating consistently lower readings. Mean peak intracranial acceleration was measured to be 35.3 ( 2.0) Gs vs 20.5 ( 0.7) Gs intraorally in helmeted cadaver heads at an impact distance of 68 cm (p < 0.001). The preimpact velocity at this distance was calculated to be 2.3 m/sec. The mean peak acceleration for the 89-cm impact distance was 56.4 ( 1.0) Gs intracranially vs 28.7 ( 0.4) Gs intraorally, p < 0.001. The pre-impact velocity at this distance was calculated to be 2.9 m/sec. Mean peak accelerations were significantly higher in the unhelmeted cadaver heads regardless of the site of measurement (Table 2). The mean peak acceleration measured intraorally had a high correlation with that measured by an intracranial accelerometer (r = 0.94 for helmeted impacts and r = 0.84 for non-helmeted impacts). DISCUSSION The use of helmets is recommended and even mandated in many sports where there is a reasonable chance of experiencing a significant blunt impact to the head. 5 8 Epidemiologic studies in cyclists suggest that the use of helmets significantly decreases the incidence of serious head injury. 3,9,10 Using mortality data from 1984 1988, Sacks et al. calculated that as many as 500 deaths from bicycle injuries per year could be prevented by the universal use of bicycle helmets. 11 Football helmets, like other protective helmets (i.e., motorcycle and bicycle helmets), were originally intended to decrease the incidence of serious intracranial injury and appear to have been effective in this regard. The incidence of football-related head injuries resulting in intracranial hemorrhage has decreased with the use of more protective helmets. 12 What about the effect helmets have on less serious but more common brain injuries such as concussion? Epidemiologic studies regarding the incidence of concussion in high school, collegiate, and professional football and hockey players are mixed, although most data suggest a slight decrease at least in high school and collegiate football. 13 15 However, the use of helmets in professional football and hockey have not prevented the continued occurrence of concussions in those sports. Our data suggest that helmets decrease the peak acceleration of the head (as mea- TABLE 2. Mean Peak Gs as Measured Intracranially and Intraorally in a Pendulum Impact Model Using Helmeted and Non-helmeted Cadaver Heads Mean Peak Gs Unhelmeted Cadaver Head Helmeted Cadaver Head p-value Intracranial (68 cm chord distance) velocity 2.3 m/sec 52.1 1.4 35.3 2.0 <0.001 Intraoral (68 cm chord distance) velocity 2.3 m/sec 40.4 2.4 20.5 0.7 <0.001 Intracranial (89 cm chord distance) velocity 2.9 m/sec 81.6 2.2 56.4 1.0 <0.001 Intraoral (89 cm chord distance) velocity 2.9 m/sec 62.6 4.1 28.7 0.4 <0.001
ACADEMIC EMERGENCY MEDICINE June 2001, Volume 8, Number 6 607 sured intracranially) by about one third when heads are subjected to a significant blunt force. Intraoral measurements of peak Gs demonstrated a reduction in acceleration of about 50% under the same conditions. Even accelerations in the relatively low range of 10 to 20 Gs (as measured intraorally) were decreased more than 50% by using a helmet. There are implications from this study regarding future research as well as public safety. Regarding the research implications, any studies to be performed in players to try to determine the actual accelerations associated with various sports are limited to the potential locations for placement of the accelerometer. Placement in helmet padding appears to significantly overestimate the actual accelerational forces experienced. Values measured in helmet padding ranged from two and one-half to three and one-half times the peak Gs measured intraorally. However, intraoral placement using an instrumented mouthpiece underestimates the actual G forces that are transmitted intracranially. These differences result from two factors: first, the accelerometer in the helmet padding does not reflect much impact attenuation (i.e., it is not in a very protected position), whereas the accelerometer in the mouthpiece sustains the full impact attenuation of the helmet, and second, the difference in the positions of the accelerometers relative to the center of gravity of the head and the resulting contribution of angular acceleration to the initial linear acceleration of impact (Fig. 2). Nonetheless, intraoral acceleration correlated very well with intracranial acceleration in our study. This correlation is evidence of the relatively small contribution angular acceleration had to overall acceleration in these impact experiments. Further studies evaluating actual game conditions should probably be performed using accelerometers placed in the mouthpiece and not in the helmet padding. The public safety implications result from the fact that even well-constructed helmets appear to reduce the intracranial acceleration by only around 30%. From the work of Gurdjian et al., 16 Ommaya et al., 17 and others, it would seem that decreasing the acceleration of impact should help protect against concussion. However, Ommaya and colleagues have shown that concussion and traumatic unconsciousness may also occur from high accelerational impulse, a mechanism that may not be amenable to protection by helmet design, but may be amenable to protective gear that reduces the amount of head movement in relation to the neck. 17,18 Another factor related to the incidence of head injury in sport has to do with the psychological effect protective gear may have on athletes. The use of helmets may embolden players to sustain more Figure 2. Linear and angular acceleration components, as seen by accelerometers at different locations, relative to the center of gravity of the head. violent impacts and may foster a gradual change in the style of play that makes injuries more likely. 19 Rugby, a sport played without helmets, has a rate of reported concussion similar to that of football. 20 This study supports the idea that regulation football helmets significantly reduce the acceleration of impact over a fairly large range (from 10 to more than 60 Gs). It also supports the fact that acceleration measured intraorally correlates well with measurements of intracranial acceleration. Recent neuropsychological testing of soccer and football players suggests that they may have chronic neurologic dysfunction and subtle, but measurable intellectual deterioration. 21,22 Whether heading the soccer ball contributes to this is presently unclear. 23,24 Further work is needed to determine the threshold of acceleration required to cause cumulative neurologic dysfunction and to find methods to help protect athletes from potentially developing cognitive impairment. LIMITATIONS AND FUTURE QUESTIONS It is important to recognize that our experimental
608 FOOTBALL HELMETS Lewis et al. FOOTBALL HELMETS AND BLUNT HEAD INJURIES design measured linear acceleration and not rotational acceleration. It would require nine accelerometers (or three triaxial accelerometers, one in each plane of motion) to accurately measure rotational acceleration. Rotational acceleration would be expected to result from any eccentric blow to the head or from central blows sufficiently forceful to cause the head to move beyond the limited range of translational displacement. Rotational acceleration has been shown to be more likely to result in neuronal injury than linear acceleration because it results in shearing strains within the brain. However, even pure linear acceleration of the head, due to anatomic and vibrational properties of the brain, will result in acceleration of various parts of the brain at different rates. For any given impact, acceleration of the cerebral cortex differs from that of relatively fixed brainstem or falx. Previous work by our group showed that simultaneous recordings from accelerometers in different parts of the brain showed significant differences in accelerations for the same impact. 25 Accelerometers simply placed within the brain substance often moved in relationship to the brain and subsequently gave results difficult to reproduce. The reason we chose the sphenoid bone for attachment of the accelerometer was to keep the position fixed. As long as the position of the accelerometer remains fixed in its relationship to the brain (or brainstem), then values from various impacts can be precisely compared. This is the assumption made for all helmet testing. Although our study measured changes only in linear acceleration, and not rotational or angular acceleration, there is a relationship between the two. The linear acceleration of the head is directly related to the magnitude of the force transmitted to the head from impact. Rotational acceleration of the head is also driven by this force (Fig. 2), with a second component, limited by the initial impacting force, derived from frictional forces between the impacting surface and the surface of the head. Thus, the maximum possible angular acceleration of the head drops as the linear acceleration decreases. Future work should extend the range of force of impact and measure acceleration in several areas within the cranium to determine the extent of differential acceleration and shearing forces. Finally, parallel work evaluating longitudinal neuropsychiatric testing in sporting participants to detect longterm cognitive dysfunction should proceed. CONCLUSIONS Wearing a regulation football helmet significantly reduced the peak acceleration of the head experienced by participants heading a soccer ball. A similar protective effect was measured intracranially at significantly more forceful impacts (above 60 Gs). Further work is required to determine to what extent helmets continue to have a protective effect at impacts above this range. The authors thank Michael Chicoine, MD, William Shannon, PhD, Guy Genin, PhD, and Mr. Theodore Naunheim for their assistance with this project. References 1. McDermott FT, Klug GL. Differences in head injuries of pedal cyclist and motor cyclist casualties in Victoria. Med J Aust. 1982; 2:30 2. 2. Offner PJ, Rivara FP, Maier RV. The impact of motorcycle helmet use. J Trauma. 1992; 32:636 42. 3. Thompson DC, Rivara FP, Thompson RS. Effectiveness of bicycle safety helmets in preventing head injuries. A case control study. JAMA. 1996; 276:1968 73. 4. American Society for Testing and Materials. Standard Test Methods for Equipment and Procedures Used in Evaluating the Performance Characteristics of Protective Headgear. F 1446-97. 1997 Annual Book of ASTM Standards. West Conshohocken, PA: ASTM, 1997, pp 923 31. 5. 2000 Official Rules of the NHL. Section 3 Equipment; Rule 22. Protective Equipment Subset (a) and (b). Chicago: Triumph Books, 2000, pp 39 40. 6. 2000 Official Rules of Major League Baseball. William A. Murray, Chairman, Rules Committee. Objectives of the Game. Rule 1.16 (a)-(e). Chicago: Triumph Books, 1999, pp 14 5. 7. 1999 2001 Official Rules of Ice Hockey. USA Hockey, Inc. Section Three: Equipment. Rule 304: Protective Equipment subset (a)-(e). Chicago: Triumph Books, 1999, pp 31 4. 8. Official Rules of the NFL. Rule 5: Players, Substitutes, Equipment. Article 3, section (a), National Football League. Chicago: Triumph Books, 1999, p 58. 9. Thomas S, Acton C, Nixon J, et al. Effectiveness of bicycle helmets in preventing head injury in children: case control study. BMJ. 1994; 308:173 6. 10. Linn S, Smith D, Sheps S. Epidemiology of bicycle injury, head injury, and helmet use among children in British Columbia: a five year descriptive study. Canadian Hospitals Injury, Reporting and Prevention Program (CHIRPP). Inj Prev. 1998; 4:122 5. 11. Sacks JJ, Holmgreen P, Smith SM, Sosin DM. Bicycle-associated head injuries and deaths in the United States from 1984 through 1988. How many are preventable? JAMA. 1991; 266:3016 8. 12. Cantu RC, Mueller FO. Catastrophic football injuries: 1977 1998. Neurosurgery. 2000; 47:673 5. 13. Albright JP, Mcauley E, Martin RK, Crowley ET, Foster DT. Head and neck injuries in college football: an eight year analysis. Am J Sports Med. 1985; 13:147 52. 14. Buckley WE. Concussions in college football: a multivariate analysis. Am J Sports Med. 1988; 16:51 6. 15. Powell JW, Barber-Foss KD. Traumatic brain injury in high school athletes. JAMA. 1999; 282:958 63. 16. Gurdjian ES, Roberts VL, Thomas LM. Tolerance curves of acceleration and intracranial pressure and protective index in experimental head injury. J Trauma. 1966; 6:600 5. 17. Ommaya AK, Flamm ES, Mahone RH. Cerebral concussion in the monkey: an experimental model. Science. 1966; 153: 211 2. 18. Ommaya AK, Gennarelli TA. Cerebral concussion and traumatic unconsciousness: correlation of experimental and clinical observations on blunt head injuries. Brain. 1974; 97: 633 54. 19. Dietzen CJ, Topping BR. Rugby football. In: Recreational Sports Injuries. Phys Med and Rehabil Clin North Am. 1999; 10:159 75. 20. Gibbs N. Injuries in professional rugby league: a three year prospective study of the South Sydney Professional Rugby Football Club. Am J Sports Med. 1993; 21:696 700.
ACADEMIC EMERGENCY MEDICINE June 2001, Volume 8, Number 6 609 21. Matser EJT, Kessels AG, Lezak MD, Jordan BD, Troost J. Neuropsychological impairment in amateur soccer players. JAMA. 1999; 282:971 3. 22. Collins MW, Grindel SH, Lovell MR, et al. Relationship between concussion and neuropsychological performance in college football players. JAMA. 1999; 282:964 70. 23. Boden BP, Kirkendall DT, Garrett WE. Concussion incidence in elite college soccer players. Am J Sports Med. 1998; 26:238 41. 24. Barnes BC, Cooper L, Kirkendall DT, et al. Concussion history in elite male and female soccer players. Am J Sports Med. 1998; 26:433 8. 25. Lewis LM, Naunheim RS, Standeven J, et al. A comparison between magnesium headforms and instrumented human cadaver heads in measuring peak acceleration of impact [abstract]. Ann Emerg Med. 1993; 22:900.