Concussion in the National Football League: An Overview for Neurologists

The authors’ studies have yielded a great deal of data regarding the biomechanics of head injury and the clinical picture of mild traumatic brain injury (MTBI) in the National Football League (NFL). The research has demonstrated the link between the effects of biomechanical forces on the brain and the clinical symptomatology of the concussed players. New insights into the mechanisms of injury are leading to new ways of protecting football players from the effects of MTBI. The clinical data validate the effectiveness of the current NFL physician approach to the evaluation and treatment of the player who sustains MTBI. There are still many more questions to answer and much more knowledge to be gained from continuing research in this area.

Clinical neurologists often encounter a patient with concussion—mild traumatic brain injury (MTBI)—in the office weeks or months after the injury. In this setting, the physician’s evaluation is frequently limited by a lack of information regarding the actual mechanisms of injury and the objectively determined clinical situation immediately following the injury. This makes it difficult to know the evolution of the clinical picture in the first few days to weeks after the traumatic event. The physician-scientist can use the athletic playing fields as a “laboratory” for the study of MTBI. The authors have been fortunate to have the opportunity to study concussions in such a setting—the National Football League (NFL).

The authors review the results of these studies here in the hope that this will help the clinical neurologist better understand the nature of MTBI, and thus improve the evaluation and treatment of all patients who have this condition. In the early 1990s, Al Toon and Merrill Hoge, two prominent NFL players, retired because of prolonged postconcussion syndrome. Commissioner Paul Tagliabue sanctioned the creation of a committee consisting of team physicians, team athletic trainers, and outside experts to objectively study and research MTBI.

The committee’s first several meetings were dedicated to developing a definition of MTBI that would be used by club medical staffs when diagnosing and reporting an injury. The definition was purposely broad, in hopes of allowing the committee to capture as much data as possible. The commissioner mandated that all MTBIs be reported, blinded, and deposited in a single repository for future analysis. Medical conferences were held with team physicians and athletic trainers.

After much discussion, the committee decided that research and data collection should focus on biomechanical and clinical analysis of the injury. Biomechanical analysis was critical because the ability of football helmets to prevent MTBI was poorly understood, and helmet manufacturer’s biomechanical research was underfunded. Understanding of the biomechanical forces that result in MTBI would precipitate objective helmet improvements, higher design standards, and greater injury prevention. Collection and study of prospective data analysis would increase sensitivity to the injury and allow clinicians to make informed decisions on a potentially unique patient population, based upon objective data. The results of the research would be published in peer-reviewed medical literature.

Biomechanics

The biomechanical information was generated from studies performed under the auspices of the NFL MTBI Committee, with the assistance of Biokinetics and Associates (Ottawa, Canada). The data were collected using video analysis of a number of NFL head impacts, some resulting in an MTBI and some not resulting in MTBI. After video analysis, the impacts were reconstructed in the laboratory using Hybrid III test dummies (General Motors, Detroit, Michigan).

Regarding impacts that resulted in clinical MTBI, the authors’ studies yielded the following findings. With concussion, the average impact speed was 9.3 ± 1.9 m/s (20.8 ± 4.2 mph). No concussions or clinical MTBIs occurred in striking players. Our analysis showed that the striking player lines up his head, neck, and torso, and strikes the other player obliquely on the face mask or where the face mask attaches to the helmet, usually below the center of gravity of the head or on the side of the helmet above the center of gravity. The striking player has a higher effective mass than the struck player because the striking player uses his whole body in the impact, whereas the struck player’s head only is involved in the impact. As a result, more momentum is transferred to the struck player, causing a rapid change in head velocity, delta V (ΔV). The peak head acceleration for concussed players averaged 98 ± 28 g, with duration of 15 msec. The peak head acceleration for the uninjured struck players averaged 60 ± 24 g. The difference in these peak head accelerations is statistically significant.

The authors’ studies revealed that clinical concussion was primarily related to translational head acceleration. Clinical concussion was strongly correlated with Severity Index (SI), Head Injury Criterion (HIC), peak translational acceleration, and head ΔV. The incidence of clinical concussion was correlated to a lesser degree with peak rotational acceleration. As noted above, the average duration of the impact was 15 msec. ΔV increases with increased translational acceleration. Our findings indicate that impact speeds of more than 7 m/s (15.6 mph) and 90 g offer a line of delineation between the occurrence of clinical concussion and the absence of clinical concussion.

The high impact speeds, ΔVs, head accelerations, and duration of 15 milliseconds are exceptionally high velocities, accelerations, and long durations compared with other types of head impacts. The average peak force to the head in NFL impacts is in the range for cranial fractures with short duration impacts for the unprotected head. In view of these results, it appears clear that the helmet shell and padding were functioning well in distributing the load and lowering the risk for more serious brain injuries and cranial fractures from these impacts. These biomechanical data on NFL impacts have given the scientific community new insights on head tolerances for impacts of 15 msec duration. In view of the authors’ findings, it appears that translational acceleration should be the primary measure for the assessment of helmet protection performance. Because our studies clearly showed that translational acceleration and rotational acceleration are related, efforts to decrease the peak translational acceleration, SI, or HIC will also result in proportional decreases in rotational acceleration.

As part of the authors’ studies, we investigated the location and the direction of helmet impacts resulting in MTBI in the NFL. Twenty-nine percent of the helmet impacts involved loading of the face mask (67% of these occurred between the angles of 0° and 45°). Seventy-one percent of the helmet impacts involved impacts to the helmet shell: 22% involved impacts with the ground, 20% involved impacts with another player’s shoulder pad or arm, 7% involved impacts with the leg of another player, and 50% involved impacts with another player’s helmet. Overall, 61% of all NFL MTBI impacts involved collision with the other player’s helmet. Seventy-six percent of face mask impacts occurred below the center of gravity of the head. More than 80% of all the impacts to the helmet shell occurred above the head center of gravity. Impacts to the face mask occurring between 0° and 45° below the head center of gravity exhibited the highest average impact velocity of all NFL impacts, and the lowest ΔVs of all NFL MTBI impacts.

Falls in which the back of the helmet struck the ground resulted in the lowest average impact velocities and the highest ΔVs. This occurs because when the back of the helmet hits the ground, the closing velocity only is coming from the one player, but the ΔV is highest because it is increased by the rebound off of the ground. The lowest translational accelerations with MTBI occur with face mask impacts; the highest translational accelerations occur with falls to the back of the head. Falls to the ground in which the back of the helmet strikes the ground exhibit the lowest rotational accelerations and velocities of all NFL impacts.

For the striking player, impact velocities were just as high as the impact velocities for the struck players, but the ΔV and peak translational accelerations were lower than for the struck players who sustained concussion. The reason for this is that the effective mass of the striking player is much greater than that of the struck player, because the effective mass of the striking player includes a greater mass from the torso. The striking player hits the other player using the top of the helmet because he lowers his head and aligns his head, neck, and torso to impact the head of the struck player. This procedure results directly in more effective mass and energy transfer to the struck player. The rotational velocities and accelerations of the striking player, in contrast, are very similar to those of the concussed struck player. The authors’ data also indicated that with face mask impacts, there was a horizontal direction of loading. Impacts to the face mask from an oblique angle twist the head (head rotation with eyes right or left) while accelerating it, resulting in an increased lateral component of acceleration.

Further analysis of the biomechanics of the striking player indicates that the key to the concussive blow is the head-down position, which increases the mass of the striking player by 67% as a result of the coupling of his torso into the collision. This results in the transfer of more momentum to the struck player. The biomechanical analysis of the striking player suggested ways to possibly diminish the risk of concussion in helmet to helmet collisions. Stricter enforcement of rules against head-down tackling techniques can lower the risk of concussion.

Players by rule are supposed to tackle with their heads up. Tackling in the head-up position results in diminished torso inertia in the striking player, which imparts diminished impact force to the struck player, which should lower the risk of concussion. Diminishing the stiffness of the top crown region of the helmet should result in lower impact force from the striking player, leading to diminished head acceleration for the struck player and increased duration of impact when the top of the helmet is used in making the tackle. This should also lower the risk of concussion. Finally, diminishing the mass of the helmet should result in lower inertia of the striking player. If the helmet weight were to be decreased by 20%, this would result in a 6% reduction in the mass of the striking player’s head and a 6% decrease in the collision impact resulting from head-down impact.

The authors also performed biomechanical analyses of the struck player in MTBI collisions. In the impact phase of these collisions, the peak head acceleration of the struck player is 94 ± 28 g and ΔV (velocity change) is 7.2 ± 1.8 m/s. During the impact phase, the movement of the struck player’s head is only 20.2 ± 6.8 mm and 6.9° ± 2.5° near the end of the impact (time 10 msec). After the impact phase, there occurs a phase of rapid head displacement in the struck player. During this phase, there is a fourfold increase in head displacement. The head displacement increases to 87.6 ± 21.2 mm and 29.9° ± 9.5°, and results in neck tension and bending. This phase of rapid head displacement occurs at approximately 20 msec following initial impact. Impacts to the front of the helmet result in rotation around the superior-inferior axis of the struck player’s head because the impact is forward of the neck centerline. Rotation about the superior-inferior axis averages 17.6° ± 12.7°, with a twist moment of 17.7 ± 3.3 Nm and neck tension of 1,704 ± 432 N at 20 msec.

As noted earlier, the authors’ studies have shown that HIC correlates best with NFL concussion risk. Mathematical analysis reveals that HIC is proportional to ΔV to the fourth power divided by head displacement to the 1.5th power (ΔV ⁴/d ^1.5). Therefore, relatively small decreases in head ΔV will have a large effect on HIC and concussion risk because the change in ΔV affects HIC by a factor to the fourth power. This suggests that players who have stronger necks will have decreased concussion risk. Stronger necks result in diminished head accelerations in the struck player, decreased ΔV in the struck player, and decreased displacement of the head and neck in the struck player. This has a significant effect on lowering HIC and concussion because of the mathematical equation noted above. As a result of this mathematical relationship, a 10% reduction in head ΔV will result in a 34% reduction in HIC if the head displacement (d) is constant. A 10% increase in head displacement (such as would occur with increased padding and thickness of the helmet) would result in a 15% decrease in HIC, but because the ΔV is often increased by the change in head displacement, the actual decrease in HIC is often less than would be expected.

During the late response (approximately 20 msec after impact), head displacement and rotation loads the neck, resulting in increased neck tension. The largest neck moments are caused by increased rotational moments, and these occur at 40 msec after impact. An increase in neck stiffness will result in decreased peak head acceleration and head ΔV, leading to a large decrease in HIC and concussion risk. Differing neck strengths may help explain the increased susceptibility to concussion seen in younger athletes, because of their relatively weaker neck muscles. This may also play a role in the increased susceptibility to concussion in some female athletes. Differing neck strengths and stiffness may also be an individual difference between NFL players, which can help to explain why some players sustain a clinical concussion and others do not, even though they experience similar impact speeds and accelerations. This suggests that strength training programs aimed at increasing the strength of the neck muscles that resist head rotation and lateral bending might help lower HIC via the mechanism of lowering head ΔV.

Impacts to the face mask result in larger head rotations in the struck player than do impacts directly to the helmet shell. Impacts to the face mask result in a greater head rotation than those to the shell because the face mask is further from the head center of gravity and SI axis of the neck than the shell itself.

The authors further analyzed the biomechanical data using a finite element computer model that simulates the fine anatomic detail and tissue level characteristics of the human head and brain to study head impacts. This model is well-suited for the investigation of strain and strain-rates (tissue deformation) in the brain following head impacts. Our finite element studies of NFL impacts reveal some interesting results. High mid-to-late time-frame strain and strain rate in the midbrain and fornix regions correlated strongly with the occurrence of clinical MTBI. Our computer analysis revealed that the areas of high strain and strain-rates move to the midbrain later in the brain response than the first 10 msec during the impact phase. The average mid-to-late time frame strain was higher in athletes who sustained clinical MTBI than in those who did not sustain clinical MTBI. These findings indicate that the occurrence of clinical MTBI (concussion) is related to brain deformation that occurs after the primary head impact and momentum transfer. This mid-to-late phase of the brain response involves rapid displacement and rotation of the head after the ΔV and rotational velocity changes have occurred.

The authors also found a number of significant correlations between the finite element brain responses and the signs, symptoms, and outcome of clinical MTBI. High strains and strain rates in the mid-to-late time period in the fornix and the corpus callosum correlated with players not returning to play on the day of the injury. High strains and strain rates in the mid-to-late time period in the fornix and the midbrain also correlated with the occurrence of memory and cognition problems. High strain rates occurring in the mid-to-late time period in the midbrain correlated with the occurrence of loss of consciousness. Dizziness correlated with high strain rates in the early time period in the temporal lobes and the orbital frontal cortex. These findings provide strong evidence of a link between the biomechanical effects of NFL impacts to the head and the clinical picture of MTBI.

The biomechanics of the finite element brain responses revealed further insights into the mechanism of MTBI in NFL impacts. The results indicated that there is a delayed response of the brain resulting from impact acceleration of the cranium. Immediately after the impact there is a low strain response in brain regions near the impact site (coup) areas. Because NFL impacts are primarily oblique or lateral, the early regions of these strains are often in the temporal lobe. A short time later, in what is termed the “mid-time phase response,” the areas of increased strain move to the opposite side of the brain from the area of initial impact loading (contrecoup areas). In NFL impacts, this is usually the opposite temporal lobe. These finite element model findings offer insight into a possible mechanism to explain coup-contrecoup injuries.

During the late time phase (approximately 20 msec after impact) the areas of high strains and strain rates move to the midbrain and other midline regions. This migration of areas of high strain and strain rates is not caused by wave propagation, which occurs over much shorter time durations than those seen in this model. The migration of the areas of increased strain and strain rate results from the motion of the head secondary to the ΔV and the resultant rapid free-motion displacement and rotation of the cranium. This also offers some insights into further means of preventing clinical MTBI. It suggests that attention should be paid to reducing the mid-to-late strains and strain rates in the midbrain as a new area of prevention. Perhaps this could be done by focusing on the role of the neck musculature. Taken in conjunction with the authors’ findings in the biomechanical analysis of the struck player, this supports the notion that strength training to increase neck strength and stiffness may be a means of lowering the risk of MTBI.

Over the past 10 years, helmet performance has been improved by the manufacturers installing thicker padding and providing fuller coverage over larger regions of the helmet shell. Pendulum testing of different helmets demonstrates that the newer helmet designs and padding reduced the risk of concussion in 7.4 m/s (16.6 mph) and 9.3 m/s (20.8 mph) impacts oblique on the face mask or on the helmet shell. Testing also showed that at the highest impacts of 11.2 m/s (25.1 mph), the helmet padding in the newer helmets bottomed out and the head responses increased dramatically, resulting in no improvement in concussion risk. Studies on the newer football helmets have demonstrated that they reduce concussion risk by 10% to 20% in collisions representative of the NFL player experience. By using thicker, more energy-absorbent padding and more padding lower on the side and back of the helmets and around the ears, the helmet manufacturers have improved the safety of NFL players. Their changes have resulted in increased energy absorption by the helmet, decreased head accelerations, and therefore diminished risk of concussion.

Epidemiology

The authors performed a 6-year clinical study of MTBI in the NFL. Data were collected prospectively. We collected complete injury data, initial clinical evaluation data, and follow-up clinical evaluation data for 787 MTBIs that occurred in preseason, regular, or playoff games for 6 seasons between 1996 and 2001. One hundred additional MTBIs from practice sessions were included in the database as well. Data were collected from forms filled out by team physicians and athletic trainers at the time of the initial clinical examination and at the time of follow-up examinations. Clinical symptoms and signs and treatments were recorded.

The definition introduced by the MTBI Committee in 1996 and used for the remainder of the study is as follows. A reportable concussion was defined as a traumatically induced alteration in brain function, which is manifested by:

•

Alteration of awareness or consciousness, including but not limited to being dinged, dazed, stunned, woozy, foggy, amnesic, or unconscious.
•

Signs and symptoms commonly associated with postconcussion syndrome, including persistent headaches, vertigo, light-headedness, loss of balance, unsteadiness, syncope, near syncope, cognitive dysfunction, memory disturbance, hearing loss, tinnitus, blurred vision, diplopia, visual loss, personality change, drowsiness, lethargy, fatigue, and inability to perform usual daily activities.

The definition of concussion used by the MTBI Committee is a natural extension of a much earlier one from the Committee of the Congress of Neurological Surgeons in 1966 that defined concussion as “a clinical syndrome characterized by immediate transient impairment of neural function such as alteration of consciousness, disturbance of vision, equilibrium, etc, due to mechanical forces.”

In the 1913 games (two teams per game) over the 6 years there were 787 MTBIs. The average annual incidence of MTBIs was 131.2 ± 26.8 MTBIs per year. The average rate of occurrence was 0.41 MTBIs per game.

Positions played had a significant impact upon the incidence of MTBIs. The offensive team players had more MTBIs than defensive team players. More concussions occurred among defensive secondary (18.2%), kicking unit players (16.6%), and wide receivers (11.9%) than other position players. The authors determined the injury rates per 100 game positions to determine the relative risks of MTBI according to position. The risk was highest for quarterbacks (1.62 concussions per 100 game positions). It was somewhat lower for wide receivers (1.23), tight ends (0.94), and defensive secondary players (0.93). Punters, return unit players, kickers, and holders had relatively low risks. All backs had three times the relative risk of MTBI compared with all linemen.

The highest frequency of injury occurred in passing plays (35.8%), followed by rushing plays (31.3%), kickoffs (15.9%), and punts (9.5%). When injuries rates per 1000 plays were considered, the relative risk of MTBI on kickoff plays (9.29 per 1000 plays) was four times the risk in rushing plays and passing plays, and 2.5 times that in punting plays; 60.5% of MTBIs were associated with tackling, and 29.5% were associated with blocking. In the majority of the cases (67.7%), MTBI was related to striking of the helmet by the helmet of another player.

On initial clinical evaluation following MTBI, the three most common symptoms that occurred were headaches (55.0%), dizziness (41.8%), and blurred vision (16.3%). At least one symptom of cognitive or memory dysfunction was experienced by 45.9% of players. 16.1% of players returned to the game immediately, 35.6% returned to play later in the same game, 44% did not return to the same game but were not hospitalized, and 2.4% were hospitalized.

Players who did not return to play on the day of the injury statistically had an increased number of symptoms compared with the players who did return to play on the day of injury. The occurrence of any memory or cognitive problem was statistically associated with the player not returning to play on the day of the injury. Loss of consciousness occurred in 9.3% of the cases. The occurrence of loss of consciousness was related to the player missing a longer period of time before returning to play. Players who sustained loss of consciousness average 5.0 ± 7.5 lost days following injury, which was 2.6 times longer than those who did not have loss of consciousness (1.9 ± 5.3 days).

This 6-year study was unique in that it was prospective. It cast a wide net, using a broad definition of MTBI to be very inclusive. A standardized reporting form for all teams was used. All of the players were evaluated by team physicians and trainers and had forms filled out by team physicians and athletic trainers, therefore increasing confidence in the validity and reliability of the medical information. All of the reporting forms from team doctors and athletic trainers were completed at the time of the evaluation of the player, and therefore were not subject to the vagaries of recall at a later time. None of the subjects was lost to follow-up.

Results of the authors’ clinical studies help to validate the results of our biomechanical analyses. The players who had the highest risk of MTBI in our study are the same players who make up most of the cases in the video analysis (open-field, high-velocity impacts). There are clear correlations between the biomechanical features and the epidemiologic findings of our study. Quarterbacks have the highest risk of MTBI because they are immobile or slowly moving, and are struck at high velocities by other players while the quarterback is often unaware of the impending impact. As a result, the high velocity of the striking player is transferred to the head of the quarterback, resulting in large ΔV and large accelerations of the head of the quarterback. Defensive backs and wide receivers have a higher relative risk of MTBI as well because they are moving at higher speeds. They are more often struck in high-velocity, high-acceleration impacts. They are often struck in midair, resulting in a fall backward and hitting the back of the helmet on the ground. These high head accelerations often exceed tolerance levels. As noted, linemen have a lower risk of MTBI. This is because they move at slower velocities over shorter distances, resulting in lower head ΔVs and accelerations.

When one compares the clinical symptoms of MTBI in the NFL and MTBI in non-athletes and other athletes, one concludes that the symptoms and signs are generally similar. As noted, 55% of NFL players complained of headache on initial evaluation. Literature review revealed that 30% to 90% of non-athletes complain of headache and 40% to 85% of other athletes complain of headaches.

Dizziness occurred in 45% of NFL players after MTBI. Other studies have shown that 53% of non-athletes complained of dizziness on initial evaluation. Blurred vision occurred in 16% of NFL players and was found in 14% of non-athletes in other studies. Photophobia occurred in 4.1% of NFL athletes and was found in 7.2% of non-athletes. Cognitive and memory problems occurred in 45.9% of NFL players. Impairment of immediate recall was much more frequent than disorientation. This has important clinical implications for the sideline physician. It is not enough to ask the player only the year, month, and date; the physician must specifically test for immediate recall when evaluating MTBI patients on the sideline. The 39.5% incidence of memory problems in NFL players on initial evaluation is consistent with the results in non-athletes examined 4 weeks after injury: 19% had memory problems, 21% had concentration problems. This is also consistent with the extensive literature on neuropsychological testing in high school and college players.

Somatic complaints such as fatigue, anxiety, personality change, irritability, and sleep disturbance were found in 20.1% of NFL players following MTBI. Similar complaints were found in 50% to 84% of non-athletes, and fatigue specifically was found in 29% of non-athletes 4 weeks after injury. Only one player experienced a seizure following MTBI. This was a generalized tonic-clonic seizure that occurred in the locker room about 30 minutes following the injury. That player made a full recovery. The clinical course of MTBI in the NFL is somewhat different than in the general population. For the great majority of NFL players there is no prolonged disability or prolonged absence from play. Only 2.9% of NFL players missed more than 9 days before returning to play. Most of the MTBIs were self-limiting, and the players made a full spontaneous and rapid recovery. It is also important to note that only 9.3% of the authors’ MTBI cohort sustained loss of consciousness. It is thus clear that loss of consciousness is not a common occurrence in NFL MTBIs. If the physicians and athletic trainers only diagnose MTBI when loss of consciousness is present, over 90% of MTBIs will be missed.

Repeat Concussions

The authors performed a more detailed analysis of players who had repeat MTBIs. Of the 887 total MTBIs occurring in 6 years during games or practices, we found that 650 players were involved. One hundred and sixty players (24.6%) had repeat concussions during the 6-year period, 51 players (7.9%) had three or more concussions, and 1 player had seven concussions. The median duration between the first and second MTBI was 374.5 days. Of the 38 repeat MTBIs that occurred within a 90-day window, the median time interval was 31.5 days. Thirty-six of these 38 injuries occurred during regular games in the same season. Only six repeat MTBIs occurred within 2 weeks of the initial injury. The risk of repeat concussion is statistically increased in quarterbacks only. The risk is decreased in offensive linemen.

The authors’ analysis revealed that the signs and symptoms were essentially the same for players experiencing only one MTBI and those whose first MTBI later turned out to be the first of repeat MTBIs. The repeat MTBI was different from the first only by the presence of increased somatic complaints, but was otherwise similar to the initial MTBI. There was no difference in the frequency of loss of consciousness for more than 1 minute in single or repeat concussions. There was no difference in the total number of signs or symptoms between single or repeat concussions. There was no difference in the total number of signs or symptoms in single or repeat concussions related to the time interval between the initial and repeat events. There was no difference in the time interval to return to play between single, initial, or repeat concussions. There was no significant difference in the incidence of postconcussion syndrome (missing more than 7 days) between single injury, initial injury and repeat injury.

The authors also evaluated players who had three or more concussions over the 6-year period, and compared them with players who had two or one concussions during the 6-year period. The average duration between concussions was 1 year. There was no difference seen compared with those who had two or one concussions in any clinical characteristic. Our studies thus reveal that the incidence of repeat concussions in the NFL is relatively low: 160 of the 3228 players in the NFL during this 6-year period sustained multiple MTBIs (5%). There were no cases of second impact syndrome seen in our study, and in fact the second impact syndrome has never been seen in the NFL.

The authors’ clinical examinations did not reveal an increased frequency of cognitive or memory impairments in players who had multiple concussions compared with players who had single concussions. We also found that cognitive or memory impairments occurred at the same frequency in the group who had three more concussions compared with the group who had two or fewer concussions.

Players Out 7 or More Days

The authors performed an analysis of the 72 cases of a player missing more than 7 days following an MTBI. These account for 8.1% of all NFL MTBIs. Quarterbacks had the highest odds ratio of such prolonged postconcussion syndrome, but this was not statistically significant. Players who were out for more than 7 days had more signs and symptoms on initial evaluation than those who ultimately were out less than 7 days. Certain symptoms are statistically correlated with a player ultimately being out for more than 7 days. These symptoms include disorientation to time, loss of consciousness for greater than 1 minute, the presence of retrograde amnesia, fatigue, overall cognition problems, problems with immediate recall, and general memory problems. Players who were out for more than 7 days also had more total number of symptoms on initial evaluation compared with players who were out for less than 7 days. The players who were out for more than 7 days averaged 4.64 symptoms or signs, versus those who were out for less than 7 days, who averaged 2.58 symptoms or signs on initial evaluation. Of the players who ultimately were out for more than 7 days, 72.2% had been removed from play and not returned to play on the day of the injury. An additional 12.5% had been hospitalized following the injury. Only 6.9%of these players had returned to play immediately.

The clinical course of players who ultimately were out for more than 7 days was evaluated by the authors. The initial examination usually occurred on the sideline or in the locker room, and the first follow-up usually occurred the next day (median 19 hours later). Most of the symptoms started to decrease between the initial and first follow-up, except for memory problems, fatigue, irritability, and sleep disturbance. By the fourth follow-up (median 4.7 days after injury) all memory and cognition problems had resolved. At that time, 17.7% of the group who ultimately would be out for more than 7 days still had headaches, 16.7% had dizziness, and 25% had photophobia. By the seventh follow-up examination all of the symptoms except for headaches had resolved.

Video analysis of some of the cases who ultimately were out for 7 days revealed that most of these were caused by open-field, high-speed impacts. The impacts of two of the players who were ultimately out for more than 7 days had been reconstructed in the laboratory as part of our biomechanical studies. These two cases had very high impact speeds, well above the median.

The authors’ analysis revealed that postconcussion symptoms resolve much more quickly in NFL players than in non-athletes. A large percentage of NFL players recover fully from MTBI within minutes to 1 hour. A smaller percentage has symptoms for up to 2 days before recovery. In the authors’ opinion, these players fall into the recovery phase of concussion, not postconcussion syndrome. Our analysis indicates that postconcussion syndrome occurs in only 8.1% of NFL MTBIs. We define this as the group having persistent symptoms or signs for 3 to 5 days or longer, who are thus out for more than 7 days. The majority of these players (80.5%) become fully asymptomatic in 7 to 11 days, and therefore are out for less than 14 days. A very small number of players (19.5% of the more than 7day-out group) ultimately missed more than two games.

There are thus four groups of NFL MTBIs that can be characterized by clinical course of recovery: (1) the immediate recovery group (56.0%) become asymptomatic in less than 1 hour, and return to play on the day of the injury; (2) the early recovery group (35.9%) have symptoms and signs for 1 hour to 2 days, and then become asymptomatic and return to play in less than 7 days; (3) the short-duration postconcussion syndrome group (6.5%) have symptoms for 3 to 10 days, and are held out of play for more than 7 and less than 14 days—they miss one game; and (4) the prolonged postconcussion syndrome group (1.6%) have symptoms for more than 10 days, and miss two or more games.

The authors’ studies found that there were certain early predictors on initial evaluation that a player might fall into the group that ultimately would be out for more than 7 days. Loss of consciousness for more than 1 minute was associated with increased risk of being out for 7 days. Memory problems at the onset, disorientation to time or general cognitive difficulties on initial evaluation all were associated with increased risk of being out for more than 7 days. The total number of signs and symptoms on initial evaluation was also a good predictor of ultimate 7 days plus out. The group that ultimately would be out for 7 or more days had an average of 4.64 symptoms or signs at the time of initial evaluation. The group that ultimately would be out for less than 7 days had an average of 2.58 symptoms and signs at the time of initial evaluation. We also found that players who were hospitalized after the injury had an increased chance of ultimately being out for more than 7 days.

Only 6.9%of players who were eventually out for more than 7 days had been allowed to return to play on the day of the injury. This indicates that team physicians and athletic trainers in the NFL are effective in screening out the most severely injured players on the sidelines shortly after injury. The authors also found that photophobia at initial evaluation was associated with increased risk of being out for more than 7 days. Results also indicate that the symptom of fatigue at the first follow-up evaluation is associated with an increased risk of being out for more than 7 days. Other findings on follow-up evaluations that were associated with increased risk of being out for more than 7 days were memory or cognitive impairments. Our results indicate that persistent headaches are the most common reason for extended delays in return to play in NFL players.

Return to Play on the Day of Injury

The authors’ results indicated that 49.5% of players return to play on the day of injury. No players were considered for return to play until and unless they were asymptomatic and had completely normal clinical neurological examinations, including mental status. The final decision regarding medical clearance to return to play was made by the team physician using clinical expertise and judgment. The median time interval between injury and return to play was 5 minutes in the group that immediately returned, and 17 minutes in the group that was removed from the game for a period of time, rested, and then returned later to the game. Forty-one percent of the players returning to the same game or practice were out from play for more than 15 minutes.

The authors found no relationship between repeat concussions and the timing of the player’s return to play. Of all the players returning to play in the same game, only 12% had a second concussion the same season involving more than 7 days out. This number is similar to the 10% of cases of 7 plus days out among those who did not return to play that day. Forty-five total players had a repeat concussion in the same season; 20 of these had been removed from play on the day of the injury or hospitalized, and 25 had returned to play on the same day.

Players who returned to play immediately had the lowest mean total number of signs and symptoms (1.52) followed by those who rested and returned (2.07 signs and symptoms). There was a significantly lower incidence of cognitive or memory problems in those who returned to play on the same day compared with players who did not return to play on the day of the injury. None of the players who returned to play on the day of the injury had experienced loss of consciousness for more than 1 minute.

There were nine cases (in 8 players) in which the player experienced loss of consciousness for less than 1 minute and subsequently returned to play on the day of the injury. All but 2 of these had injuries near the end of the game. Six of these 8 players had injuries in later seasons, but these numbers are too small for a statistical analysis. Of the 439 injured players who returned to play on the day of the injury, only 10 subsequently were out for more than 7 days. This group had more signs and symptoms on average than the 429 other players who had returned to play on the day of the injury.

In the group of players who returned to play on the day of the injury, there were no cases of subdural hematoma, epidural hematoma, or other intracranial hematoma or cerebral edema. There were no cases of second impact syndrome. Compared to the players who did not return to play on the day of the injury, there was no increased risk of the players ultimately being out for more than 7 days (prolonged postconcussion syndrome) or of having a repeat concussion. There were a small number of players in this group who ultimately were out for more than 7 days, which emphasizes the importance of follow-up medical evaluation in the days after MTBI, even in those players who had returned to play on the day of the injury.

Based on the results of the authors’ 6-year study and the experience of NFL team physicians and certified athletic trainers, the League has recently reaffirmed current NFL medical practice regarding the evaluation and treatment of players who sustain MTBI ( Box 1 ). As noted previously, loss of consciousness is infrequent in NFL MTBIs. Over the course of our study, only nine players who sustained brief loss of consciousness returned to play on the day of the injury. Only one of those nine players returned to play on the day of the injury in each of the 2000 and 2001 seasons. None of those nine players returned to play on the day of the injury during the 1999 season. Although there is no evidence that any of these players experienced any adverse consequences as a result of returning to play on the day of the injury, NFL team physicians have adopted a conservative approach to these cases. The League has recently reaffirmed that those players who are determined by team physicians to be unconscious as the result of head impact should not return to play on the day of the injury (see Box 1 ). In order to further the NFL players’ understanding of concussion, the League has also recently developed and disseminated to all its players an information sheet on the subject ( Box 2 ).

Box 1

Only gold members can continue reading. Log In or Register to continue