Siatta B. Dunbar and Margot Putukian
Concussion is a common and important injury for a variety of sports participants at every level of play. Whether there are gender differences in the incidence of concussion, the mechanisms of injury, or in the clinical presentation of concussion, as well as the recovery from injury, remain areas of interest. This chapter evaluates the literature that exists exploring the sex and gender differences in sports-related concussion
The Centers for Disease Control and Prevention estimates that between 2001 and 2009, 2.6 million children under the age of 19 were treated for sports-related injuries and 6.5%, or 173,285 of those injuries, were traumatic brain injuries (TBI) (1). Concussion, the most common type of TBI (2), is also the most common head injury occurring in sports, accounting for 8.9% and 5.8% of injuries at the high school and college levels, respectively (3). Certain factors appear to increase the risk for concussion or to delay recovery when a concussion injury occurs, such as a prior history of concussion, migraine or headache disorder, learning disabilities/attention deficit hyperactivity disorder (ADHD), and mood disorders. There are data that suggest that females may have an increased incidence of concussion, and there are also limited data to suggest that there may be differences in the mechanism of injury, the symptoms reported at baseline as well as after injury, and the cognitive and overall recovery from concussion between men and women. This chapter explores the limited research in this area.
INJURY EPIDEMIOLOGY IN WOMEN VERSUS MEN
It can be challenging to compare concussion injury epidemiology due to differences in the definition of concussion, differences in what constitutes an injury (e.g., playing time loss versus injury evaluated by medical staff), the quality of the data, specifically the quality of the denominator in terms of exposures, and finally the reporting system (e.g., whether reported by medical personnel such as an athletic trainer or team physician or self-reported). Though there are data at the college and high school levels, there is a paucity of data at the youth level as well as at the professional level. The Institute of Medicine has recently reviewed the available sport-related concussion data (4).
In 2007, the data from 1998 to 1999 through 2003 to 2004 from the National Collegiate Athletic Association (NCAA) were published and demonstrated that the incidence of concussion was 5.3 compared to 3.9 per 1,000 athlete exposures (AEs) in women compared to men’s soccer, respectively, and 4.7 compared to 3.2 per 1,000 AEs in women’s compared to men’s basketball (5). This same study reported an incidence of concussion in women’s softball of 4.3 compared to 2.5 per AEs for baseball. At approximately the same time, another study evaluated 100 high schools and 180 colleges and again demonstrated an increased incidence of concussion in female soccer and basketball players compared to their male counterparts (3). Other studies have also confirmed that for sports using the same rules (e.g., soccer, basketball, and baseball/softball), females report a higher incidence of concussion compared to males (7–12). The most recent data from the NCAA, reviewing the incidence of injury in both game and practices, continue to demonstrate a higher incidence of concussion in sports with the same rules for women compared to men (6). In the same NCAA review, the highest overall incidence of concussion occurred in wrestling, followed by men’s football, men’s ice hockey, women’s field hockey, women’s soccer, women’s lacrosse, men’s lacrosse, men’s soccer, women’s basketball, men’s basketball, softball, women’s volleyball, and finally baseball. At the high school level, recent data between 2010 and 2012 reported the incidence of game and practice concussions in girls compared to boys soccer and girls compared to boys basketball was 3.4 and 1.9 and 2.1 and 1.6 per 10,000 AEs, respectively, and 1.6 compared to 0.5 per 10,000 AEs in softball compared to baseball (12). When evaluating the percentage of total injuries in practice and games that are concussions, the recent NCAA data (6) show that for football, 35% of injuries are concussions, followed by women’s soccer (12%), women’s basketball (10%), men’s basketball (8%), men’s soccer (6%), wrestling and softball (both 5%), with all other sports less than 5%.
In terms of mechanism of injury, some data exist suggests that in sports with similar rules there may be subtle differences in the mechanism of injury between sexes, with player contact to surface or equipment being more common in females compared to males (7). Dick (7) found that for basketball and soccer, the mechanism of concussion in males was more likely due to impact from other players (81% in males versus 66% in females), whereas the mechanism of concussion in females was more likely due to contact with the ball (8% in males versus 18% in females) or a surface (8% in males versus 13% in females). It is important to understand that there are very few prospective data to evaluate mechanism of concussive injury, and therefore conclusions must be considered preliminary.
The explanation(s) for a potential difference in concussion incidence as well as mechanism of injury is unclear and may be multifactorial. Some of the differences in incidence could be related to a reporting bias, which postulates that females are more likely than males to report symptoms of a concussion to a health provider. Additionally, there may exist a treatment bias where athletic trainers or other health care providers may manage female athletes, especially those of younger age, more conservatively, therefore possibly skewing the data. Other potential explanations for incidence and mechanism of injury include biomechanical differences, such as head-to-neck ratio and strength, and the neuroprotective effect of sex hormones, most notably progesterone. There may also be sex differences in symptom scores and cognitive function at baseline as well as post-concussive injury. There is limited literature exploring sex and gender differences in concussion and more research is needed to answer critical questions regarding concussion, most importantly how injury and complications can be prevented.
COGNITIVE FUNCTION AND SYMPTOMS AT BASELINE VERSUS POST-INJURY
Concussion is characterized as a “subset of traumatic brain injury” and is defined as “a complex pathophysiological process affecting the brain, induced by biomechanics forces” (13). The diagnosis of a concussion can be challenging, without a clear biomarker or diagnostic test, relying instead on a constellation of subjective and objective data. Following a concussion there are typically alterations in function that cover several domains—presence of athlete reported symptoms, physical signs, and behavioral, postural, and cognitive changes. Symptoms and cognitive function are two important components of the multimodal assessment of concussion. Several guidelines for managing concussion have been published (13–19,25), and many of these endorse the concept of “baseline” testing as part of the preseason to assess symptoms and findings that can then be repeated after an injury occurs. The NCAA recently released guidelines stating that every athlete, no matter how minimal the risk for concussion, should have a baseline assessment for concussion (18). There are standardized tools such as the Sideline Assessment of Concussion (SAC) as well as the more recent Sideline Concussion Assessment Tool-3 (SCAT-3), which can be used to assess concussion (20,21). The SCAT-3, which incorporates the SAC, has been shown to be both sensitive and specific to concussion (22). When baseline assessments are available and compared with post-injury assessments, a drop in score of 3.5 points on the SCAT-2 demonstrated a sensitivity and specificity of the SCAT-2 of 96% and 81%, respectively. When a baseline assessment is not available, sensitivity and specificity were 91% and 83% when a cutoff value is used (22,41). Both paper and pencil as well as computerized neuropsychological (NP) testing have been utilized to evaluate concussive injury, and testing can be used for both baseline and post-injury assessments of cognitive function. The utility of NP testing in assessing and tracking recovery after concussion has been reviewed (23,24,26,27). NP testing evaluates brain-behavior relationships and provides objective information regarding speed of information processing, memory recall, attention and concentration, reaction time, scanning and tracking ability, and problem solving. Similar to differences in incidence of injury, research has shown that there are differences between males and females in their baseline reporting of concussion symptoms and measures of cognitive function using NP testing both at baseline as well as after concussive injury (17,28,29,37). Specifically as it relates to NP testing at baseline, Covassin et al. (28,38,54) found that females scored better on verbal memory with no statistically significant differences in other testing domains. Interestingly, Colvin et al. (29), in their baseline evaluation of soccer players with a history of concussion, found that females had significantly slower reaction times than males. When evaluated after sustaining a concussion, results have shown that females have scored lower on visual memory as compared to males (31,38,53).
The Post Concussion Symptom Score (PCSS) is a subjective measure showing the presence and type of symptom as well as the athlete’s perceived severity of that given symptom. There are a total of 22 symptoms, graded 0–6 and covering four domains: physical, cognitive, emotional, and sleep. Severity totals can range from 0 (no symptoms) to 132 (presence of all symptoms rated as most severe). Covassin et al. (28) studied the presence and degree (mild 1–2, moderate 3–4, and severe 5–6) of baseline symptoms, males compared to females. Of the total 1,209 athletes studied, “68% of males and 76% of females” endorsed baseline symptoms with the difference between the sexes reaching statistical significance for “headache, nausea, fatigue, sleep disturbance, drowsiness, sensitivity to light and noise, irritability, sadness, nervousness, feeling more emotional, and difficulty concentrating.” For both males and females, fatigue was most prominent at 26% and 29%, respectively. Overwhelmingly, both males and females reported their symptoms as being “mild (1–2 on a 0–6 scale), at 66% and 73%, respectively.” These data underscore the importance of an individualized approach to concussion management and the importance of evaluating post-concussion NP testing and PCSS in comparison to baseline values.
Covassin et al. (37) explored how depression affects baseline NP testing and baseline concussion symptoms score. A total of 1,616 collegiate and high school athletes completed Immediate Post-Concussion Assessment and Cognitive Testing (ImPACT), PCSS, and the Beck Depression Inventory assessments. Results showed that those with severe depression “scored worse on visual memory” and endorsed presence of symptoms in the “somatic (headache, dizziness), cognitive (memory problems, fogginess), emotional (anxiety, depression), and sleep (more/less sleep, trouble sleeping) domains of the PCSS.” At baseline, females overall had more symptoms in the “cognitive, emotional and sleep clusters,” but there was no correlation between females and the presence or severity of depression.
In a study of 280 collegiate athletes, Putukian evaluated the SCAT-2 as well as baseline screening assessments of both depression (Patient Health Questionnaire-9 [PHQ-9]) and anxiety (Generalized Anxiety Disorder-7 [GAD-7]), history of loss of consciousness (LOC), migraine, and prior concussion (22,41). The SCAT-2 includes subsets of symptoms, cognitive function (the SAC assessment), and postural stability (modified Balance Error Scoring System). In evaluating for baseline demographics that might explain a difference in SCAT-2 scores, no significant gender differences emerged. In addition, no effect was seen when history of migraine, concussion, or loss of consciousness (LOC) was evaluated. Elevations in both the PHQ-9 and GAD-7 did correlate with SCAT-2 performance, and when this was evaluated further it was found that these scores correlated with increases in symptom score and not the subsets of cognitive function or balance. This study was limited in terms of evaluating the effect of sex or gender on incidence of and outcome from concussion given the small number of concussions during the study period.
After sustaining a concussion, there may be sex differences in the reporting of concussion symptoms. Colvin et al. (29) studied 234 soccer players with a history of previous concussion, which included 93 males and 141 females, ranging from 8 to 24 years old. Results showed that female players reported significantly higher total symptom scores, 25.6 compared to 14.0 reported by males. Conversely, Broshek et al. (30) demonstrated that males and females had a similar mean number of reported symptoms post-concussion, 5.07 compared to 6.66, respectively. However, within these reported symptoms, females self-reported “concentration problems, fatigue, lightheadedness, and seeing flyspecks” more significantly than males. Of note, the most reported symptom by both males and females was headache at approximately 82% and 86%, respectively. Although baseline NP testing reportedly had no differences in the Broshek study, there was no mention or comparison of post-concussion symptoms to baseline symptoms and if a gender difference exists. Several years later Covassin and colleagues (31) studied 79 college athletes, at preseason and at 2 and 8 days post-injury, and noted no sex or gender differences with baseline symptom scores, but they did conclude that males had “significantly higher symptom scores for sadness and vomiting” post-injury. When post-concussion symptoms were evaluated in athletes who played a similar sport, in this case soccer, and controlled for modifiers (ADHD, learning disability, and migraines), Covassin et al. (53) demonstrated that females “reported a greater number of total concussion symptoms at eight days” post-concussion. Specifically the total reported scores for females was 11.9 (±15.70) as compared to 5.3 (±7.40) for males and there were statistically significant gender differences in the “sleep” and “migraine-cognitive-fatigue” clusters. Of note, in the Covassin et al. study, there was no comparison to baseline symptoms, and although the study was controlled for modifiers, it may not be possible to fully conclude that the increased post-concussive symptoms (PCSs) were purely due to sex. Therefore, it appears that females as compared to males will report more symptoms at baseline. However, given that PCSs have not been compared to baseline, we are unable to conclude from these studies whether symptoms post-injury are due to persistent deficits after concussion versus symptoms potentially due to other conditions. Why sex differences might occur in other realms of assessment (e.g., cognitive function and balance) is unclear.
BALANCE/POSTURAL ASSESSMENT AT BASELINE VERSUS POST-INJURY
After sustaining a concussion, there have been reported balance and postural deficits (32). These deficits represent the complex interdependent relationship of the sensory and motor systems that make up our vestibular system that can be tested by vestibulo-ocular reflex (VOR), the vestibulospinal reflex (VSR) and the vestibulocollic reflex (VCR) techniques (32). The VOR helps to maintain vision during head movement, VSR maintains posture and balance, and VCR stabilizes the head. Guskiewicz et al. (33) demonstrated that the Balance Error Scoring System (BESS) is a “practical, valid, and cost-effective method of objectively assessing postural stability” with deficits returning to baseline by “day 3 post-injury.” Subsequently, Murray et al. (34) performed a literature search to determine the reliability of postural testing after sustaining a concussion and demonstrated that the BESS has a “moderate test-retest reliability.” This was more clearly defined by Finnoff et al. (35) who showed that a change in the total BESS score of greater than 7.3, with the same rater, and greater than 9.3 points, with a different rater, is required to conclude that the difference from baseline is related to changes in postural sway rather the individual completing the test.
Having shown the BESS to be a valid tool to assess postural stability, Zimmer et al. (36) used it with 437 Division II athletes to determine if there is a difference in baseline BESS based on sport and then subsequently based on sex. They looked at a total of seven sports, and when evaluated as a whole, soccer athletes had the least amount of errors at baseline (16.43 errors) and basketball had the most (22.02 errors). When reordered by sex, scores for women’s soccer (15.35 errors) were statistically significant as compared to men’s basketball (25.04 errors), men’s lacrosse (22.52 errors), and football (22.22 errors). It seems logical that soccer players would have better BESS results at baseline given the need to perform a majority of the sport-specific skills on one leg. Therefore, this further underscores the importance of baseline testing or at a minimum having sport-specific normative data that can be used in the evaluation and management of athletes after sustaining a concussion.
In another study using the BESS, athletes who had sustained a concussion were evaluated to see if there were any correlations with sex or age (38). A total of 222 athletes were tested: 157 male and 65 female, of which 150 were from high school and 72 from college. The athletes were analyzed at days 1, 2, and 3 post-concussion. Although not statistically significant, the results showed that college females had a mean of 21 errors compared to 13 for college males after sustaining a concussion. However, the results were opposite at the high school level, with females having 17 errors and males having 19 errors. Of note, the statistically significant finding from this study was the gradual improvement in total BESS of all participants from day 1 to 3.
A recent retrospective chart review of 206 patients, age 6 and over, with a concussion diagnosis (80 that were sport-related, ages 10–65) was completed to determine if there was any significant correlation among age, gender, BESS, SAC, and King-Devic (K-D) testing. The results showed no sex effect on SAC, BESS, and K-D; however, there was a significant correlation between female sex, older age, and a higher symptom score. Similarly, there was a correlation, as expected, between older age and higher BESS errors. Limitations of this study include the inclusion of concussions that were not sport-related as well as a better stratification of BESS performance by age (39). In reviewing the current literature available on sex differences and BESS testing, what is apparently missing are prospective longitudinal data. Zimmer et al. (36) showed a sex difference in baseline BESS, but there was no correlation to post-concussion data in the subjects that were studied. Similarly, Covassin et al. (38) demonstrated a sex difference in post-concussion BESS with college females scoring the worst, but there was no baseline BESS for these subjects. Therefore, more research is needed to evaluate the changes from baseline to post-concussion across the currently established measures used (i.e., BESS and NP testing) to evaluate for any differences based on sex.
LENGTH OF RECOVERY
The decision to begin a return to play (RTP) progression should be done on an individual basis and in a stepwise progression. Several position statements agree with this concept (13–19,25) and recommend a RTP protocol that begins with light aerobic activity (to maintain a heart rate greater than 70% of maximum) and, if tolerated, progresses over a 24-hour period to sport-specific exercises, noncontact drills, and finally sport-specific activity and full contact. Of note, although widely accepted, this stepwise progression is not an evidence-based approach. Therefore, it is important to understand that each concussion is different and that individualized management is necessary. While the overwhelming literature supports the concept that the large majority of concussions will resolve in 7 to 10 days (13,40), more recent published data demonstrates a more conservative approach with RTP taking 14 to 21 days (22). Additionally, one must consider “modifiers” that may prolong recovery, such as female sex, younger age, history of previous concussions, or history of ADHD, learning disability, depression, or migraines (13,16). A recent meta-analysis has found that symptoms persisted for 15 days in high school athletes compared to 6 days in college athletes, though NP testing deficits were similar in both groups (52).
Bazarian et al. (42) evaluated 1,425 individuals that sustained a mild traumatic brain injury (MTBI) to determine correlation between sex, PCS, and return to baseline (determined as the length of time to return to normal activities and the amount of work days missed). Close to half of the study subjects were female, and the results showed that females were more likely to have PCS scores greater than 16 and to miss more than 7 days of normal activities. Of the three parameters, the only one that was statistically significant was the correlation between female sex and PCS score. At 3 months post-injury, the mean PCS score for males was 4.0 compared to 7.0 for females and was most apparent between ages 14 and 56. Similarly, King (43) found a significant correlation between female sex and prolonged PCS, defined as symptoms being present 12 to 18 months post-injury, with a mean age of 35.9 years. In the adolescent population (ages 11 to 18), 67.7% of females took 1 month or more to return to baseline and 41.2% took more than 2 months (44). Though more research is needed, there appears to be evidence that female gender increases the risk for having prolonged PCS.
WHAT EXPLAINS THE DIFFERENCE?
Although research has shown an increased incidence of concussion in females compared to males participating in sports with the same rules, and there appear to be differences in both baseline and post-injury assessments of symptoms, cognitive function, and postural stability, there remains an unclear understanding as to the etiology of these differences. Possible explanations include reporting bias, treatment bias, the effect of fluctuating sex hormones across the menstrual cycle, and biomechanical differences.
One possibility that has been eluded to is that females may be more likely to report symptoms, both at baseline as well as post-injury. This can be a challenging argument and certainly deserves further attention. The possibility of sex hormones as reason for sex disparity has also been an area of research. Animal studies have demonstrated that progesterone has “anti-apoptotic” and protective effects when administered serially to adult male rats after sustaining a “contusion of the frontal cortex” (45). The significance and reproducibility of these findings in humans has been questioned. Wunderle and colleagues (46) studied what they termed the “withdrawal hypothesis.” They postulated that the downregulation of progesterone production after sustaining an MTBI would lead to worse outcomes. They evaluated 144 patients between ages 16 and 60 and stratified them into three groups—those taking synthetic progestin (SP), those in the luteal phase (LP) of their menstrual cycle (as determined by serum progesterone concentration), and those in the follicular phase (FP) of the menstrual cycle. Those in the LP group had statistically significant lower scores on the Quality of Life (QoL) general health rating and QoL index score as compared to the SP group. The authors hypothesized that the lower scores in the LP group might be due to a drop in progesterone just after the time of injury as progesterone levels fall toward the end of the LP. The LP group had worse Riverbed Post Concussion Questionnaire (RPCQ) scores as compared to the SP and FP groups, but they were not statistically significant. The differences between the LP and FP groups, across all three measures, however, was not statistically significant. Mihalik et al. (47) evaluated if baseline neurocognitive function, postural stability, and symptoms scores were affected by being in the LP or FP of the menstrual cycle. Thirty-six females were tested using ImPACT, the Sensory Organization Test (SOT), and PCSS in both their follicular and luteal phases and there was no statistically significant difference between the menstrual phases across all three tests. Therefore baseline tests done at any phase in the menstrual cycle are valid and any deviations noted can be assumed to be secondary to the concussion. However, these results should be considered with the knowledge that cycle phase was determined with the calendar, which is far less reliable than using urinary or serum markers. Thus, additional research is needed to evaluate the role of sex hormone levels and menstrual cycle phase as an explanation for the sex differences seen in concussion.
Another area that warrants exploration is the biomechanical differences between the sexes, specifically head-to-neck ratio, which effects head stabilization and acceleration during contact. Tierney et al. (48) examined 40 physically active individuals, 20 female and 20 male, to assess gender differences in head-neck stiffness, mass, and acceleration. Males were found to have longer head-neck length, greater head-neck mass, and greater neck girth, but only the difference in mass and girth were statistically significant, with females having 43% less head-neck mass and 30% less neck girth. Although females activated their cervical stabilizing muscles faster, sternocleidomastoid (SCM) 29% faster, and trapezius 7% faster, and generated 79% more muscle activity, they had up to 70% more acceleration and 39% more displacement than males. These differences in head acceleration and displacement may increase the risk of concussion. Whether neck mass and girth can be modified and whether this translates to a decrease in head acceleration and displacement is unclear. Schmidt et al. (55) found that greater cervical stiffness and less angular displacement after perturbation, and not isometric muscle strength or size, was associated with mitigating head impact severity in high school and college football players. More research evaluating head impact biomechanics and if and how they relate to concussive injury is needed, as well as research specifically addressing gender differences in head impact biomechanics.
Mansell et al. (49) followed 36 college soccer players (17 males and 19 females) to see if an 8-week cervical resistance training program would improve dynamic stabilization when force application was known or unknown, and then evaluated for sex differences. Although females demonstrated increased extensor strength by 22.5% and increased girth by 4.5%, there was no reduction in acceleration whether the force application was known or unknown. Conversely, Eckner et al. (50) demonstrated that if the force is known, subsequent cervical muscle activation led to reductions in linear and angular velocity. These reductions held true across both sexes and in both the pediatric and adult groups. Intuitively it makes sense that if one sees the impact coming the acceleration force that the head sees is diminished. Applying Newton’s second law, where mass equals force times acceleration, if the cervical muscles are held rigid, the mass is now the body and the head (instead of just the head alone) and one would expect a decrease in the acceleration seen by the head. More research is needed in this area.
Collins et al. (51) measured neck strength in 6,662 male and female high school athletes and subsequently compared strength to concussion incidence. They found that across both sexes and all sports, with the exception of soccer, weaker neck strength consistently correlated with those athletes that had sustained a concussion. They also extrapolate that “for every one pound increase in neck strength, odds of concussion decrease by 5%.” This research supports the concept that concussion may be prevented by measures to increase strength of the neck musculature.