Demographic Factors and Instantaneous Lower Extremity Injury Occurrence in a National Collegiate Athletic Association Division I Population

National Collegiate Athletic Association (NCAA) athletes experience approximately 13.8 injuries per 1000 athlete-exposures during games and 4 injuries per 1000 athlete-exposures during practices, with >50% of all injuries involving the lower extremities (LEs).¹ As many as 76% of female NCAA Division I athletes are injured in a given season,² whereas only 43% of all athletes are reported to be injured.³ Over a 4-year collegiate career, at least 1 sport-related injury was reported by 90% of student-athletes.⁴ Furthermore, knee, lower leg, ankle, and foot injuries account for more than half of all severe injuries, defined as restricting sport participation for >3 weeks, as reported in NCAA programs.⁵ Given that injuries have long-lasting repercussions on physical function⁶ and result in increased financial burden,⁷ we need reliable identification of injury risk factors to provide clinicians with the means to mitigate future injury occurrences. Understanding the combined influences and interactions among various demographic factors and the LE injury occurrence over time is a necessary step toward this aim.

The inclusion of multiple risk factors in the prediction of sport-related injury is important because of the interactions among potential risk factors such as sex, sport, and body mass index (BMI). Relatively few researchers^8–10 have assessed the combined effect of such demographic factors on the risk of LE injury at the collegiate level. Accurate prediction of LE injury occurrence in athletes provides clinicians the opportunity to implement injury-prevention programs to target athletes most at risk. In a systematic review of the literature, Montalvo et al¹¹ showed that females participating in contact sports were approximately 3 to 5 times more likely to sustain a noncontact anterior cruciate ligament (ACL) injury than males, although evidence is mixed regarding whether this risk extends to other LE injuries.^11,12 For instance, Beynnon et al⁸ observed no differences in the incidence rates of first-time ankle sprains between male and female athletes; however, in basketball, female athletes had a greater risk of a first-time ankle sprain than male athletes. These results suggest a possible interaction between sex and sport in terms of injury rates, such that injury risks are not the same for each sex, and they demonstrate a need to include such factors as determinants of all LE injuries.

BMI has been documented to play a role in the LE injury risk, but most relevant investigations have been conducted in military or youth populations. In a prospective study including college-aged cadets, researchers¹³ revealed that BMI was a predictor of ACL injury in females. Higher BMI has been reported to contribute to increased risk of first-time musculoskeletal injury¹⁴ and ankle injury¹⁵ in soldiers. However, the effect of BMI on injury risk in adolescents has been inconclusive, with some authors¹⁶ documenting a greater risk of injury in individuals with high BMI (incidence rate ratio = 2.07; 95% CI = 1.00, 6.94) and others¹⁷ finding protective effects of high BMI (odds ratio = 0.64; 95% CI = 0.51, 0.80). The effect of BMI on injury risk might be clearer when accounting for sex, sport, or both. Compared with normal-weight athletes in the same sport, obese softball and girls' basketball players sustained more knee injuries, whereas obese wrestling, volleyball, and football athletes sustained more ankle and foot injuries.¹⁸ These results further argue for examinations of interactions among sex, sport, and BMI.

Additionally, previous injury is a well-documented risk factor for subsequent injury.^19,20 Researchers^21–24 have indicated that athletes with a history of LE injuries, including hamstrings strains, ACL ruptures, Achilles tendon ruptures, and ankle sprains, were at a higher risk of subsequent LE injury in high school, collegiate, and professional athletics. To this end, investigators have proposed that previous injury results in neuromuscular deficits, muscle imbalances, and changes in LE biomechanics.^21,25 Coupled with inadequate rehabilitation, subsequent injuries often ensue.

The instantaneous injury risk is a metric largely unexplored in the current sports medicine literature and may offer valuable information for clinicians seeking to mitigate the potential for injury. Often quantified via time-to-event analyses, instantaneous injury risk captures both if and when an athlete sustains an injury. In this way, it expands on a simple dichotomous answer of yes or no and infers the rate at which athletes become injured. A Cox regression is the most common form of time-to-event analysis, as it yields both hazard rate ratios (ie, if an injury occurs) and survival curves (ie, when an injury occurs).^{14,22,24,26–29} A hazard ratio (HR) of 1.20 indicates that a particular group has a 20% greater chance of being injured at any given time than a reference group, whereas the reference group will take 1.20 times longer to be injured. Survival curves provide a visual representation of when an individual can be expected to sustain injury. A sharp decline seen in a curve indicates a drop in the “survival” of the group (ie, more of the group experiences an injury at this time). This information would be valuable for resource-limited clinicians as they seek to mitigate future injuries via targeted interventions.

The purpose of our study was to characterize the instantaneous LE injury risk (ie, if and when an injury occurs) in a sample of NCAA Division I athletes using the following demographic factors: sex, sport, BMI, and previous LE injury. Our operational definition of instantaneous LE injury risk was injury risk at any given point in time after baseline measurement. We hypothesized that the instantaneous LE injury risk would be greater (ie, higher incidence at an earlier time) in females, athletes with a BMI of >25 kg/m²,³⁰ and athletes with a previous LE injury.

METHODS

Procedures

Lower extremity injury was defined as any injury occurring to the lower limb, hip, or lumbar spine that required medical attention from an athletic trainer or medical doctor and resulted in restriction of participation for ≥1 day beyond the day of injury. All injuries meeting our definition of LE injury were captured, regardless of when or where they happened; injuries were not limited to organized team practices or competitions. Lumbar spine injuries were included because lower back injuries are known to create lower limb motor control deficits.³¹ Because evidence^27,32 has indicated that concussions may influence the potential for subsequent LE injury, we also captured and reported concussions. A history of LE injury was obtained via a review of athletes' medical records maintained by the university's sports medicine and athletic training department. Injuries that occurred during athletes' time at the university were recorded by staff athletic trainers. All injuries that occurred before arrival at the university were self-reported by athletes in questionnaires completed upon arrival at the university as freshmen or transfer students. After each academic year, data on injuries sustained during the study (prospective injuries) were exported from the athletic training department's injury-management software and compiled. Data extracted from each prospective injury record are detailed in Table 1. Athletes were tracked until (1) their first LE injury, (2) the end of their collegiate career, or (3) September 1, 2016, whichever occurred first. If a single athlete completed multiple data-collection sessions (eg, 2013 and 2014), only the earliest session data were used. Athletes were ultimately tracked for 237 ± 235 consecutive days (range = 1–856 days).

Table 1 Descriptive Frequency Data for All Predictor Variables

View Fullscreen

RESULTS

A total of 278 athletes (119 males, 159 females; age = 19.07 ± 1.21 years, height = 175.48 ± 11.06 cm, mass = 72.24 ± 12.87 kg) were recruited from the varsity basketball, cross-country, soccer, softball, tennis, and volleyball teams. Fifty-eight athletes (21%) had no previous LE injury or prospectively identified injuries. Eighty-three athletes (30%) had previous injuries only. A total of 25 athletes (9%) sustained prospective injuries only. A total of 112 athletes (40%) sustained both previous and prospective injuries. Predictor variables are detailed in Table 1. Descriptive information for all prospective injuries is provided in Table 2.

Table 2 Descriptive Data of Prospective Injuries Gathered (N = 137)^a

View Fullscreen

In the all-athletes univariate model, female and male athletes had similar instantaneous LE injury risks (HR = 1.29; 95% CI = 0.91, 1.83; P = .16; Figure 1). Compared with athletes with no previous LE injury, athletes with previous injuries were not more likely to sustain a prospective LE injury (HR = 1.09; 95% CI = 0.76, 1.54; P = .64; Figure 2). Although the difference was not statistically significant, we found a trend toward a higher (ie, shorter time to first injury) risk of instantaneous LE injury in overweight athletes (BMI >25 kg/m²) than that in athletes with a BMI of <25 kg/m² (HR = 1.23; 95% CI = 0.84, 1.82; P = .29). Basketball, soccer, and softball players had a higher risk of LE injury than cross-country athletes (HR range = 2.78–3.49; each P = .002; Figure 3). The full univariate results are presented in Table 3.

View Figure

• Download as Powerpoint
• Download Figure

View Figure

• Download as Powerpoint
• Download Figure

View Figure

• Download as Powerpoint
• Download Figure

Table 3 All-Athletes Univariate Cox Regression Model Predicting the Prospective Injury Risk^a

View Fullscreen

When stratified by sex, males with a BMI of >25 kg/m² were at an increased risk of LE injury compared with males who had a BMI of <25 kg/m² (HR = 2.21; 95% CI = 1.20, 4.07; P = .009; Figure 4A). Females with a BMI of <25 kg/m² exhibited similar LE injury risk rates as females with a BMI of >25 kg/m² (HR = 0.80; 95% CI = 0.48, 1.36; P = .41; Figure 4B). Males with a previous injury were not at higher risk of future injury than males without previous injuries (HR = 1.74; 95% CI = 0.95, 3.19; P = .07). The sex-stratified univariate results are shown in Table 4.

View Figure

• Download as Powerpoint
• Download Figure

Table 4 Sex-Stratified Univariate Cox Regression Model Predicting the Prospective Injury Risk^a

View Fullscreen

Accordingly, we considered the predictors of sex, sport, BMI, and a sex-by-BMI interaction in the multivariate analysis. Previous LE injury was excluded. The multivariate model indicated that basketball (HR = 2.91; 95% CI = 1.37, 6.18; P = .005), soccer (HR = 2.76; 95% CI = 1.44, 5.32; P = .002), and softball (HR = 3.30; 95% CI = 1.47, 7.42; P = .003) athletes had a higher LE injury risk than cross-country athletes. The instantaneous LE injury risk was greater in female than in male athletes (HR = 1.55; 95% CI = 1.00, 2.39; P = .05) and greater in male athletes with a BMI of >25 kg/m² than that in all other athletes (HR = 0.44; 95% CI = 0.19, 1.00; P = .05); these findings were of borderline significance. Full multivariate results are supplied in Table 5.

Table 5 Multivariate Cox Regression Model Predicting Prospective Injury Risk^a

View Fullscreen

DISCUSSION

Our study is one of the first to examine the association between demographic factors and the instantaneous LE injury risk in Division I collegiate athletes. We hypothesized that the instantaneous LE injury risk would be greater in females, athletes with a BMI of >25 kg/m², and athletes with previous LE injury. The results partially supported our hypothesis. Specifically, in the multivariate model, females did exhibit a tendency toward a greater instantaneous LE injury risk than males. Additionally, males with a BMI of >25 kg/m² showed an increased risk of LE injury compared with males with a BMI of <25 kg/m², but this difference was not evident in females or when males and females were analyzed together. Contrary to our hypothesis, and surprisingly given earlier research,^19,20 previous LE injury was not a predictor of prospective LE injury occurrence. A strength of our work is the use of instantaneous injury risk as the outcome of interest. Instantaneous injury risk is operationalized as an HR, which is similar to a relative risk ratio but inferred at each point in time. In other words, an HR of 1.20 indicates that, compared with a reference group, individuals have a 20% greater chance of sustaining an injury on any given day. Alternatively, individuals in the lower-risk group will take 1.20 times longer to become injured.³³

The LE injury rates we observed were comparable with those reported earlier. In an NCAA Division I athlete population, 43% of student-athletes experienced an LE injury over a single season.³ Although our study was conducted over several years, our data demonstrated that approximately 40% of the Division I athletes sustained at least 1 LE injury within 150 days after baseline (Figure 2).

Researchers^11,12 have determined that females are at increased risk of LE injury compared with males. Our results indicated that, overall, females displayed a nonsignificant tendency toward a greater instantaneous LE injury risk than in males; however, in the sex-stratified analysis, this sex effect was not statistically significant when isolating athletes with no injury history (Table 4). Although sex was not a predictor of LE injury in our univariate analysis, it was significant in the multivariate model. Combined with BMI, the effect of sex on LE injury risk was even stronger. Therefore, if sex is to be used as a predictor of injury, BMI must be included because it affects each sex differently. The BMI appears to affect the risk for LE injury in males much more than in females, as seen in Figure 4.

Irrespective of sex, we noted a clear sport effect regarding the instantaneous LE injury risk. Cross-country and tennis displayed the lowest LE injury rates, whereas basketball and soccer showed the highest rates among sports that included both male and female athletes. This finding was expected because of the nature of each sport (eg, noncontact versus contact sports, anticipated versus unanticipated movements). Approximately 30% of cross-country athletes will experience an LE injury over the course of a year, whereas approximately 70% of basketball athletes will experience an LE injury in that same time (Figure 3). Athletes playing softball or volleyball, which are exclusively female sports, sustained many LE injuries early in the tracking period, as depicted by the initial sharp decline of the survival curve; however, the values for both plateaued after approximately 200 days.

Similar patterns were observed in other survival curves. For instance, the survival probability dropped drastically in females with no previous injury and males with a BMI of >25 kg/m² (Figure 4A) throughout the first 200 days. This result suggested that the LE injury occurrence was highest in the first 200 days of activity after a preparticipation examination and plateaued after that point. We acknowledge that tracking by consecutive days is a limitation in not accounting for an athlete's exposure to sport; still, our data collection took place in both the fall and spring and was based on individual athlete availability, with all sports being represented at each data collection. Thus, although we were unable to reference a specific point in the season when injury was more likely to occur, the cyclical nature of our data collection mitigated any seasonal effects. The reader should be mindful that only the time to first injury was considered. Therefore, it is possible that some individuals were reinjured after returning to play and that these subsequent injuries were not captured. Nevertheless, clinicians can use these survival curves to extrapolate the initial LE injury occurrence for a specific population.

Overweight athletes (BMI >25 kg/m²) were at a 23% higher risk of LE injury than athletes with a BMI of <25 kg/m². This univariate finding pertaining to BMI was not statistically significant, possibly because we included only the initial measure of BMI as a predictor. Any changes in BMI that an athlete experienced over the season or in subsequent years were not taken into account. Researchers examining both high school and military populations have identified higher injury incidence rates in individuals who were overweight or obese.^14–16,18 Exploration of BMI as an injury risk factor in collegiate populations is an area with limited literature; most data pertain to military or youth and high school populations. Using BMI as a predictor of injury risk may pose greater challenges in a collegiate than a youth or a high school population. Division I collegiate athletes are more elite than youth and high school populations, and BMI measures often overestimate adiposity in elite athletic populations due to the increased relative muscle mass typically seen in high-level athletes.³⁴ Because they are more elite, collegiate athletes are a more homogeneous population than youth or high school cohorts and will likely display less BMI variability. The BMIs in our study were 23.1 ± 2.3 kg/m² for males and 23.7 ± 9.2 kg/m² for females. These data are comparable with previous data from a large military cohort in which the respective male and female mean BMIs were 23.4 ± 2.7 kg/m² and 22.0 ± 2.0 kg/m², with the ACL injury risk associated with BMIs of >1 SD above the mean.¹³

Our multivariate findings revealed an interaction that was of borderline significance between sex and BMI—females with a BMI of <25 kg/m² had a 56% decreased risk of instantaneous LE injury compared with males who had a BMI of >25 kg/m². The BMI value may be more useful as a risk factor in a sport for which a larger or smaller body size is preferred. For example, in 1 study,¹⁸ more than half of the high school football players who were injured were overweight or obese. Athletes playing certain positions in football—offensive line and defensive line, for instance—are often encouraged to have a larger body size to meet the demands of those positions. Interestingly, in our investigation, males with a BMI of >25 kg/m² were at an increased risk of LE injury compared with males who had a BMI of <25 kg/m², even though we did not include football data. In a systematic review of female athletes, Collings et al³⁵ showed that a higher BMI predicted LE injury; however, the mean difference was small (0.5 kg/m²). This finding, combined with our multivariate results, suggested an interaction between BMI and sex. Because BMI can represent different metrics between sexes (eg, higher BMI in males often reflects increased relative muscle mass, whereas higher BMI in females often reflects increased relative fat mass),³⁴ future authors should use more accurate measures of body composition, such as fat-free mass or fat mass, to analyze the influence of body mass on injury risk.

Contrary to what has been well documented in the literature, our study revealed that, overall, athletes with previous LE injuries were not more likely to sustain a prospective LE injury than those with no history of injury. Yet when the results were stratified by sex, males with previous injuries exhibited a higher risk of future injury than males without. In recent research on military cadets, Hearn et al³⁶ demonstrated that individuals with previous injury (in a cohort that was overwhelmingly male) were at greater risk of future LE injury. Given that most athletes in our study were female, the overall effect of previous injury may have been diminished, suggesting that biological sex may moderate the relationship between previous and future injuries. Because we included only individuals available for full participation in their sport, we would have excluded athletes who had previously sustained career-ending injuries. Thus, we cannot attribute this finding to different exposure rates. One proposed explanation is that sustaining an injury only increases an individual's future injury risk for a short time after the initial injury.³⁷ The increased risk of injury is possibly only seen while athletes experience deficits in motor control, proprioception, or strength after an initial injury, a state that may exist for just a few weeks after injury. Athletes who had sustained an injury in the previous 6 weeks were excluded from our cohort; thus, we may have inadvertently excluded athletes who were truly at risk of prospective injury. Furthermore, we did not control for the time since the previous injury. If an athlete reported any previous LE injury, no matter when it occurred, it was included in our data as a previous injury. Of interest, recent authors³⁷ proposed a nonlinear relationship between factors influencing injury. Many potential factors can play a role in an athlete's risk of injury, including motor control, hormones, and the specific time in the season or game. It is possible that these factors, among others, influence the effect of previous injury on future injury. A single risk factor is unlikely to be the sole contributor to an athlete's injury. For example, in a study of Division I athletes, Hegedus et al³⁸ determined that motor control had a mediating effect between previous injury and future injury, such that greater motor control during physical performance tests exhibited a protective effect on subsequent injury. Because of this finding, researchers^22,37 have suggested that targeted rehabilitation aimed at increasing motor control can help athletes mitigate the detrimental effects of initial injury and ultimately reduce their risk of future injury.

We acknowledge that our investigation had limitations. The small sample size did not allow us to stratify the data by acute and chronic injuries, which likely have different risk factors. Also, we chose to dichotomize BMI into <25 and >25 kg/m² and to not consider underweight (BMI <18.5 kg/m²) or obese (BMI >30 kg/m²) athletes as separate categories as only 9 athletes were underweight and 9 were obese. Future authors should examine larger sample sizes to assess risk factors for acute versus chronic injuries as well as stratify the data by multiple BMI categories. Furthermore, we did not control for the time since previous injury, even if it was an injury sustained during childhood. Future researchers should consider this factor to determine if a more recent injury predicts the risk of sustaining a subsequent injury. Additionally, we captured only the time to first LE injury. Athletes may have sustained multiple injuries during the study period, but any subsequent injuries were not taken into account. Moreover, we included only BMI measurements at baseline. It is possible that BMI fluctuated over the study period. Future researchers should evaluate whether changing one's BMI can alter the injury risk, as well as devise implementation strategies to accomplish this goal. The use of BMI to measure body composition is also a limitation, as BMI does not differentiate between the lean mass typically seen in high-level athletes and adiposity. Future studies should incorporate more accurate measures of body composition, such as fat-free mass or fat mass.

CONCLUSIONS

Among our cohort of Division I athletes with no history of injury, females exhibited a greater (ie, higher incidence at an earlier time) risk of instantaneous LE injury than males. A history of LE injury was not a significant predictor of future injury. A BMI of >25 kg/m² was associated with an increased risk of instantaneous LE injury in males but not in females or the whole sample. Athletes in contact sports sustained more LE injuries than athletes in noncontact sports. Clinicians can use this information to better identify athletes most at risk for LE injuries and implement proper injury-prevention techniques to help mitigate those risks.