Abstract
Background:
Achilles injury risk in women’s collegiate gymnastics is 10-fold higher than in all other collegiate sports. This study aims to identify risk factors for Achilles tendon ruptures in collegiate female gymnasts.
Hypothesis:
Gymnasts with Achilles tendon ruptures will be more likely to report early gymnastics specialization, elite-level training before college, and performance of high-difficulty skills on floor and vault.
Study Design:
Cross-sectional study.
Level of Evidence:
Level 4.
Methods:
Anonymous surveys were distributed to current and former collegiate female gymnasts, aged 18 to 30 years, via coaches, athletic trainers, Twitter, and ResearchMatch. Information about Achilles tendon ruptures, gymnastics-related injuries, sport specialization, event/skills participation, and medication use were collected.
Results:
A total of 581 gymnasts were included. One hundred gymnasts (17.2%; 95% CI: 14.1%-20.3%) reported Achilles tendon ruptures during collegiate training or competition. Most ruptures (91%) occurred on floor exercise; 85.7% of these occurred during back tumbling-take-off. Compared with gymnasts without ruptures during college, a greater percentage of gymnasts with ruptures competed at a Division I program, trained elite, competed difficult vaults and floor passes before and during college, competed in all 4 events during college, identified as Black/African American, and used retinoid medications.
Conclusion:
Achilles tendon ruptures are more common in women’s collegiate gymnastics than other sports. Competing at the elite level, performing difficult floor and vault skills, and competing in all 4 events may increase the risk for an Achilles tendon rupture. Potential nontraining risk factors include retinoid exposure and Black/African American race. Future studies regarding the mechanisms of Achilles tendon ruptures in female collegiate gymnasts are warranted.
Clinical Relevance:
Collegiate gymnasts who compete at the elite level, perform high levels of difficulty on floor and vault, and compete in all 4 events may be at increased risk for Achilles tendon ruptures.
Keywords: gymnastics, female athlete, Achilles tendon
Achilles tendon injuries place a significant burden on an athlete’s physical, mental, and emotional well-being and often result in significant time away from sport.7,38,46 Because of successive plyometric jumping and landing demands, significant eccentric loading on the Achilles tendon occurs particularly in gymnasts.7,46 Chan et al 7 utilized the National Collegiate Athletic Association (NCAA) Injury Surveillance Program (ISP) database to examine the rates of Achilles tendon injuries across 16 NCAA sports from 2005 to 2016. The NCAA-ISP prospectively tracks and collects de-identified injury information for collegiate athletes and undergoes verification and quality assurance processes. 7 They found that women’s gymnastics had a significantly higher Achilles injury rate (16.73/100,000 athlete exposures) compared with all other men’s and women’s NCAA sports; the second-highest rate was 4.26/100,000 athlete exposures in men’s basketball. 7 Furthermore, the rate of severe Achilles injury was the highest in women’s gymnastics (7.87/100,000 athlete exposures), compared with all other NCAA sports. Severe Achilles injury was defined as any injury that was season- or career-ending, had a play time loss greater than 30 days, or required surgery. 7 Similarly, rates of all or severe Achilles injuries due to contact or noncontact mechanisms were the highest in women’s gymnastics. 7 These findings, in addition to the observation that 20 Achilles ruptures were reported in just the first 3 months of the 2020 women’s NCAA gymnastics season, 8 suggest a disproportional risk of Achilles tendon ruptures in collegiate women’s gymnastics that warrant further investigation in this population.
Biomechanical data have revealed that gymnastics skills, such as tumbling on the spring floor, can produce forces at the foot and ankle up to 15 times the body weight,5,41 with bone-on-bone forces at the tibiotalar joint reported as high as 23 times the body weight. 5 Young gymnasts aspiring to compete at the collegiate level are often exposed to these high forces before and during puberty and must perform very high repetitions for skill acquisition. 17 Additionally, gymnastics encourages early specialization and year-round training, 31 which are known risk factors for increased overall injury risk and reduced long-term career performance.2,30,31,39
Proposed risk factors for Achilles tendon rupture in women’s gymnastics and other aerial sports include landing surfaces, training period, desire for high scores, technique, lack of footwear, age, and lack of concentration 46 ; however, no studies have examined risk factors for Achilles tendon rupture in NCAA women’s gymnastics programs. Furthermore, the influence of early specialization, training load, prior Achilles tendon injury, performance of high-level floor tumbling passes, and event specialization on the risk of spontaneous Achilles tendon rupture has not been delineated in this population. Therefore, the primary aim of this study was to assess the prevalence of spontaneous Achilles tendon ruptures during NCAA competition or training in collegiate female gymnasts. The secondary aim was to determine factors that may increase the risk of Achilles tendon rupture. We hypothesized that (1) gymnasts who had an age of specialization <10 years or competed at the elite level would be at greater risk of Achilles tendon rupture; (2) gymnasts who perform difficult skills on floor or compete in all 4 events in college would be at greater risk of Achilles tendon rupture, compared with gymnasts who perform lower-level tumbling skills or specialize in 1 or 2 events; and (3) gymnasts with prior Achilles tendon injury would be at greater risk of Achilles tendon rupture than those without prior Achilles injury or pain.
Methods
Ethics and Study Design
This cross-sectional survey study was approved by the institutional review board. The survey was anonymous and taken only 1 time by each participant. An alternative method of consent was obtained via an informed consent screening question.
Recruitment
Current and former gymnasts who competed at the NCAA level (Division I, II, III) and were aged 18 to 30 years were eligible to take the survey. Male gymnasts, female gymnasts who did not compete at the collegiate level, and those outside the age range were excluded.
The survey was distributed to potential participants via coaches, athletic trainers, social media, and ResearchMatch. If they wished to participate, a description of the study and instructions for participation were provided, with a link to the survey. Clicking on the survey link brought up an informed consent screen containing information on the study, confidentiality language, and contact information for the study. Participation was entirely voluntary. Those who agreed to participate answered a research authorization question, followed by a screening question to ensure they were 18 to 30 years old. All responses were anonymous, and no identifying information was collected.
Data Collection
Data were collected on gymnasts’ age, race/ethnicity, and geographic location of college attended. Race/ethnicity was self-reported; the race and ethnic categories were based on definitions provided by the National Institutes of Health and included “American Indian/Alaska Native,” “Asian,” “Black or African American,” “Hispanic or Latino,” “Native Hawaiian or Other Pacific Islander,” and “White.” 33 Next, the age when gymnasts started gymnastics, highest level reached in club gymnastics, average hours trained per week during their highest level before college, and the age when gymnasts started competing at their highest level before college were obtained. The following precollege data were collected: vaulting difficulty (Yurchenko Layout 1½, Yurchenko Layout 2/1, Tsukahara tuck/pike 1/1, Tsukahara Layout 1/1, Tsukahara Layout 1½, Tsukahara Layout 2/1 were considered difficult), floor difficulty (E-level twisting and D- or E-level flipping skills: back triple full or more, back double tuck, back double pike, back full in/full out/half in half out, back double layout, Arabian double front were considered difficult), and history of injuries that required the gymnasts to miss or modify gymnastics participation for ≥2 weeks.
Collegiate training and competition data were collected, including years competed in college, division of school attended, events trained during college, difficult vaults competed (same as above), and difficult floor passes competed (same as above). Gymnasts reporting an Achilles tendon rupture were asked to provide information on how, when, and on which event the rupture occurred. They were also asked if they had experienced Achilles tendon pain, calf pain, heel pain, or swelling in the 4 weeks before the rupture, and whether any injections were performed on or around their Achilles tendon. Last, a series of questions regarding female athlete triad characteristics (bone health, menstrual health, and nutrition) 34 and medication exposure (retinoids, oral steroids, fluoroquinolone, or statins) during college were asked. The medication exposure question did not specify retinoid type (eg, systemic or topical), but retinol-A was given as an example.
Statistical Analysis
Continuous data are summarized as means and standard deviations (SDs), and discrete data are summarized as frequencies and percentages. The prevalence rate of Achilles tendon rupture is reported with 95% CIs. Comparisons between gymnasts with and without Achilles tendon ruptures were analyzed using the 2-sample t test for continuous variables and the Pearson’s chi-square test for discrete variables. The 95% CIs of differences are reported for these comparisons. Crude and adjusted effects of race, elite training, difficult vaults/floor skills competed before and during college, and retinoid use were done using log-binomial regression. Parameter estimates are reported as crude and adjusted risk ratios (aRRs) and 95% CIs.P values <0.05 were considered statistically significant. All analyses were performed with Stata, Version 14.2 (StataCorp).
Results
Participant Flow
A total of 695 potential participants accessed the survey from July 29 to November 23, 2020. Of these, 9 declined participation, and 73 were outside the age range. An additional 32 participants were withdrawn owing to incomplete responses to the Achilles tendon rupture question. The number of gymnasts who were included in the study was 581.
Gymnast Characteristics
Gymnast characteristics are shown in Table 1. The mean current age was 23.0 ± 3.4 years, and 505 (87.2%) respondents identified as White. Most responses came from those who competed at Division I schools (72.8%). More gymnasts attended college in the Midwest (35.1%) compared with other regions. Of the 581 gymnasts, 86 (14.8%) reached the elite level before college.
Table 1.
Gymnast characteristics
| Characteristic | n a | Count (%) |
|---|---|---|
| Current age, y, mean ± SD | 581 | 23.0 ± 3.4 |
| Current or past NCAA gymnast | 581 | Current: 346 (59.6) Past: 235 (40.4) |
| Race | 579 | American Indian/Alaska Native = 5 (0.9) Asian = 41 (7.0) Black or African American = 46 (7.9) Hispanic or Latino = 23 (4.0) Native Hawaiian or Other Pacific Islander = 7 (1.2) White = 505 (87.2) Other = 6 (1.0) |
| NCAA division | 577 | I = 423 (72.8) II = 42 (7.2) III = 112 (19.3) |
| Geographic region | 578 | Northeast = 117 (20.1) South = 149 (25.6) Midwest = 204 (35.1) West = 108 (18.6) |
| Competed elite? | 581 | Yes = 86 (14.8) No = 495 (85.2) |
| Started training level 10/elite level at age 13 y or younger | 536 | Yes = 176 (32.8) No = 360 (67.2) |
NCAA, National Collegiate Athletic Association.
Excludes missing data.
Prevalence of Achilles Tendon Ruptures
One hundred gymnasts reported Achilles tendon ruptures during NCAA training or competition, for a prevalence rate of 17.2% (95% CI: 14.1%-20.3%); of these, 6 gymnasts reported 2 or more ruptures. The mean age of participants when an Achilles tendon rupture occurred was 19.9 ± 1.4 years. There were no significant differences in rupture prevalence across the 4 years of college. Most ruptures (91%) occurred on floor exercise, and the remainder (9%) occurred on vault. The majority of ruptures reported on floor occurred during the back tumbling-take-off (85.7%) (Figure 1). Of the 9 ruptures reported on vault, 7 occurred during the spring board impact (77.8%) (Figure 1). Of those who sustained an Achilles tendon rupture, 70.9% reported Achilles tendon pain in the preceding 4 weeks (Table 2). Sixteen (16.0%) gymnasts reporting receiving injections in or around their Achilles tendon during college; 8 received the injections before rupture, and 8 received them after rupture.
Figure 1.
Mechanism of injuries reported on floor and vault. The numbers of gymnasts who sustained Achilles tendon rupture during back tumbling back handspring take off, back tumbling round off impact, back tumbling landing, front tumbling landing, and other skills on floor are shown on the left. The numbers of gymnasts who sustained Achilles tendon rupture during board impact and landing on vault are shown on the right.
Table 2.
Details of Achilles tendon rupture during college
| Rupture Details | n a | Count (%) |
|---|---|---|
| Year in college when rupture occurred | 96 | Freshman = 27 (28) Sophomore = 26 (27) Junior = 22 (23.0) Senior = 24 (25.0) Fifth year = 1 (1.0) |
| Events training before rupture | 100 | Vault = 87 (87) Bars = 77 (77) Beam = 82 (82) Floor = 96 (96) |
| Symptoms during 4 weeks before rupture | 86 | Achilles tendon pain = 61 (70.9) Calf pain = 30 (34.9) Heel pain = 20 (23.3) Swelling = 17 (19.8) Other = 22 (25.6) |
Excludes missing data.
Comparisons Between Gymnasts With and Without Achilles Tendon Ruptures During College
Demographics, geographic location, and NCAA division were compared between gymnasts with and without Achilles tendon ruptures during NCAA training or competition. Significantly more gymnasts with ruptures were of Black/African American race (15.0% vs 6.5% without ruptures; 95% CI of differences: 1.2% to 15.9%; P = 0.004). Additionally, more gymnasts with ruptures attended college in the West (29.0% vs 16.5%; 95% CI of differences: 3.0%-22.0%; P = 0.004) and competed at a Division I institution (83.5% vs 71.3%; 95% CI of differences: 3.8%-20.7%;P = 0.01) (Table 3), compared with those without ruptures.
Table 3.
Race, geographic location, and NCAA division
| Demographics | No Achilles Rupture (n = 481), n (%) | Achilles Rupture (n = 100), n (%) | 95% CI of Differences | P |
|---|---|---|---|---|
| Race | ||||
| Native American/Alaska Native | 5 (1.0) | 0 (0.0) | 0.1%, 2.0% | 0.30 |
| Asian | 30 (6.3) | 11 (11.0) | –11.0%, 1.8% | 0.09 |
| Black/African American | 31 (6.5) | 15 (15.0) | –15.9%, –1.2% | 0.004 |
| Hispanic/Latino | 21 (4.4) | 2 (2.0) | 0.0%, 5.7% | 0.27 |
| Native Hawaiian/Pacific Islander | 6 (1.3) | 1 (1.0) | 0.0%, 2.4% | 0.83 |
| White | 426 (88.9) | 79 (79.0) | 1.5%, 18.4% | 0.007 |
| Unknown | 5 (1.0) | 1 (1.0) | 0.0%, 2.2% | 0.97 |
| Geographic region | ||||
| Northeast | 97 (20.3) | 20 (20.0) | –8.3%, 8.9% | 0.95 |
| South | 130 (27.2) | 19 (19.0) | –0.5%, 16.9% | 0.09 |
| Midwest | 172 (36.0) | 32 (32.0) | –6.1%, 14.1% | 0.45 |
| West | 79 (16.5) | 29 (29.0) | –22.0%, –3.0% | 0.004 |
| NCAA division | ||||
| I | 342 (71.3) | 97 (83.5) | –20.7%, –3.8% | 0.01 |
| II | 36 (7.5) | 6 (6.2) | –4.0%, 6.7% | 0.65 |
| III | 102 (21.3) | 10 (10.3) | 3.9%, 18.0% | 0.01 |
Boldface indicates those with P values <0.05. NCAA, National Collegiate Athletic Association.
Table 4 shows comparisons regarding precollege training habits, skills competed, and injuries sustained before and during college. Compared with gymnasts without ruptures, a higher percentage of gymnasts with ruptures trained at the elite level before college (22.0% vs 13.3% without ruptures; 95% CI of differences: 0.0%-17.4%; P = 0.03), competed difficult floor skills before college (92.9% vs 85.4%; 95% CI of differences: 1.54%-13.4%; P = 0.05), and competed difficult vaults before college (28.3% vs 16.6%; 95% CI of differences: 2.2%-21.1%; P = 0.007). A higher percentage of gymnasts with ruptures were also found among those who competed difficult floor skills during college (92.1% vs 82.9%; 95% CI of differences: 2.5%-15.9%; P = 0.03) and competed difficult vaults during college (21.6% vs 10.6%; 95% CI of differences: 1.8%-20.1%; P = 0.006). Compared with gymnasts who specialized in 1 to 3 events during college, a significantly greater percentage of gymnasts in the rupture group competed in all 4 events during college (59.0% vs 47.0%; 95% CI of differences: 1.4% to 22.6%; P = 0.03). Furthermore, there were no significant differences in vault and/or floor specialization during college between gymnasts with and without Achilles tendon ruptures (Table 4). There were higher percentages of precollege ankle and hip injuries in the no-rupture group compared with the rupture group (ankle—71.7% vs 60.0%; 95% CI of differences: 1.3%-0.2%; P = 0.02) (hip: 12.0% vs 5.0%; 95% CI of differences: 2.0%-13.0%; P = 0.03). There were no significant differences in the percentages of gymnasts with and without Achilles tendon ruptures who reported calcaneal apophysitis before or Achilles tendinopathy before or during college. Similar percentages of gymnasts with and without Achilles tendon ruptures specialized ≤age 10 years and trained >20 hours per week (Table 4).
Table 4.
Training, event skills, and injuries
| Characteristic | No Achilles Rupture (n = 481) | Achilles Rupture (n = 100) | 95% CI of Differences | P |
|---|---|---|---|---|
| Age started gymnastics, mean ± SD | 4.1 ± 2.0 | 4.3 ± 2.0 | –0.64, 0.21 | 0.33 |
| Specialized at ≤10 y | 376 (78.3) | 84 (84.8) | –14.5%, 1.5% | 0.14 |
| Elite or level 10 training at <13 y old a | 140 (31.5) | 36 (39.6) | –19.0%, 2.8% | 0.13 |
| >20 h trained per week before college | 421 (87.5) | 88 (88.0) | –7.5%, 6.6% | 0.90 |
| Elite training before college | 64 (13.3) | 22 (22.0) | –17.4%, 0.0% | 0.03 |
| Difficult vaults competed before college | 80 (16.6) | 28 (28.3) | –21.1%, –2.2% | 0.007 |
| Difficult floor passes competed before college | 411 (85.4) | 92 (92.9) | –13.4%, –1.5% | 0.05 |
| Difficult vaults competed during college | 39 (10.6) | 19 (21.6) | –20.1%, –1.8% | 0.006 |
| Difficult floor passes competed during college | 321 (82.9) | 82 (92.1) | –15.9%, –2.5% | 0.03 |
| Calcaneal apophysitis or Achilles tendinopathy before college | 99 (20.6) | 24 (24.0) | –12.5%, 7.0% | 0.45 |
| Achilles tendinopathy during college | 34 (7.1) | 4 (4.0) | –1.4%, 7.5% | 0.26 |
| All-arounder during college | 226 (47.0) | 59 (59.0) | –22.6%, –1.4% | 0.03 |
| Specialists on vault, not floor | 33 (12.9) | 8 (19.5) | –19.4%, 6.2% | 0.26 |
| Specialists on floor, not vault | 68 (26.7) | 11 (26.8) | –14.8%, 14.4% | 0.98 |
| Specialists on floor and vault | 108 (42.4) | 21 (51.2) | –25.3%, 7.6% | 0.29 |
Boldface indicates those with P values <0.05.
Denominators are 444 (no tear) and 91 (tear).
Comparisons involving medication use and menstrual cycle characteristics during college are shown in Table 5. Retinoid medications were more commonly used in gymnasts who had an Achilles tendon rupture, compared with those who had no rupture (12.1% vs 4.7%; 95% CI of differences: 0.7%-14.1%;P = 0.005). No differences in oral steroid or hormonal contraceptive use were observed between groups. Additionally, no significant differences in menstrual cycle characteristics were observed between groups (Table 5).
Table 5.
Menstrual cycle characteristics, hormonal contraceptive use, and medication use during college
| Characteristic | No Achilles Rupture (n = 481), n (%) | Achilles Rupture (n = 100), n (%) | 95% CI of Differences | P |
|---|---|---|---|---|
| Age of menarche, y, mean ± SD | 14.5 ± 1.0 | 14.4 ± 0.2 | –0.3, 0.5 | 0.69 |
| Primary amenorrhea | 222 (46.7) | 45 (47.0) | –11.1%, 10.8% | 0.98 |
| Oligomenorrhea a | 36 (21.3) | 5 (19.0) | –13.1%, 18.7% | 0.74 |
| Secondary amenorrhea a | 8 (4.7) | 0 (0.0) | 1.5%, 7.9% | 0.25 |
| Hormonal contraceptive use | 298 (62.3) | 70 (70.0) | –17.6%, 2.3% | 0.15 |
| Oral contraceptive use | 252 (52.7) | 59 (59.0) | –17.2%, 4.0% | 0.23 |
| Retinoid use | 22 (4.7) | 12 (12.1) | –14.1%, –0.7% | 0.005 |
| Oral steroid use | 43 (9.2) | 11 (11.1) | –8.7%, 4.8% | 0.55 |
| Fluoroquinolone antibiotic use | 21 (4.5) | 5 (5.1) | –5.3%, 4.1% | 0.80 |
Boldface indicates those with P values <0.05.
Oligomenorrhea and secondary amenorrhea were determined in nonhormonal contraceptive users only.
Screening questions regarding the female athlete triad, depression and anxiety were also compared between groups (Table 6). Compared with those without ruptures, significantly lower percentages of gymnasts with ruptures reported worries about body weight and composition (76.0% vs 85.4%; 95% CI of differences: −18.3% to −0.4%; P = 0.02), limiting or controlling food intake (63.0% vs 78.0%; 95% CI of differences: −25.2% to −4.8%; P = 0.002), and eating in secret (12.0% vs 21.7%; 95% CI of differences: −17.1% to −2.4%; P = 0.03). There were no differences in past stress fractures, low bone density, or reports of anxiety, depression, or obsessive/compulsive thoughts or behaviors between groups. Body mass index (BMI) was also compared between groups; however, no significant differences were found (Table 6).
Table 6.
Female athlete triad and mental health during college
| Characteristic | No Achilles Rupture (n = 481), n (%) | Achilles Rupture (n = 100), n (%) | 95% CI of Differences | P |
|---|---|---|---|---|
| Worries about weight/body composition | 409 (85.4) | 76 (76.0) | 0.4%, 18.3% | 0.02 |
| Limits/controls food intake | 372 (78.0) | 63 (63.0) | 4.8%, 25.2% | 0.002 |
| Tries to lose weight to meet weight/image requirements in sport | 314 (65.6) | 61 (61.0) | –5.9%, 15.0% | 0.39 |
| Weight affects way she feels about herself | 382 (79.7) | 74 (74.0) | –3.6%, 15.1% | 0.20 |
| Worries about losing control over eating amount | 207 (43.2) | 38 (38.0) | –5.3%, 15.7% | 0.34 |
| Vomits or uses diuretics/laxatives after eating | 31 (6.5) | 11 (11.0) | –11.0%, 2.0% | 0.11 |
| Current or past eating disorder | 101 (21.2) | 19 (19.2) | –6.6, 10.6% | 0.66 |
| Eats in secret | 104 (21.7) | 12 (12.0) | 2.4%, 17.1% | 0.03 |
| Past stress fracture | 116 (24.3) | 27 (27.3) | –12.6%, 6.6% | 0.53 |
| Low bone density | 25 (5.3) | 5 (5.0) | –4.4%, 5.0% | 0.91 |
| Body mass index, kg/m2, mean ± SD | 22.8 (2.3) | 22.5 (2.3) | –0.2, 0.8 | 0.20 |
| Anxiety, depression, or obsessive/compulsive thoughts or behaviors | 309 (64.8) | 63 (63.0) | –8.6%, 12.2% | 0.74 |
Boldface indicates those with P values <0.05.
Factors Associated With Achilles Tendon Ruptures
Log-binomial regression analyses were conducted to determine if elite training, retinoid use, race, all-around status, difficult vaults competed, or difficult floor skills competed were associated with an increased risk of Achilles tendon ruptures during NCAA competition or training. In the crude model, Black or African American race, elite training, difficult floor skills/vaults competed precollege, difficult floor skills/vaults competed during college, all-around status, and retinoid medication exposure during college were all significantly associated with a greater risk of Achilles tendon ruptures (Table 7; P < 0.05). Results from the adjusted model demonstrate that Black or African American race was independently associated with a greater risk of Achilles tendon ruptures (aRR: 1.9; 95% CI: 1.2-2.9; P = 0.008). Retinoid medication use was also independently associated with a greater risk of Achilles tendon ruptures (aRR: 1.8; 95% CI: 1.1-2.9; P = 0.01) (Table 7).
Table 7.
Log-binomial regression analysis of factors associated with Achilles tendon ruptures
| Characteristic | Crude RR | 95% CI | P | Adjusted RR | 95% CI | P |
|---|---|---|---|---|---|---|
| Black or African American race | 2.0 | 1.3, 3.2 | 0.002 | 1.9 | 1.2, 2.9 | 0.008 |
| Elite training | 1.6 | 1.1, 2.5 | 0.02 | 1.4 | 0.9, 2.1 | 0.15 |
| Precollege difficult floor skills/vaults | 2.0 | 1.0, 4.2 | 0.06 | 1.3 | 0.6, 2.9 | 0.55 |
| NCAA difficult floor skills/vaults | 1.9 | 1.2, 3.1 | 0.007 | 1.6 | 0.9, 2.8 | 0.09 |
| All-around gymnast in college | 1.5 | 1.0, 2.2 | 0.03 | 1.1 | 0.7, 1.6 | 0.76 |
| Retinoid exposure | 2.2 | 1.3, 3.6 | 0.002 | 1.8 | 1.1, 2.9 | 0.01 |
Boldface indicates those with P values <0.05. NCAA, National Collegiate Athletic Association; RR, risk ratio.
Discussion
In this study, the prevalence of Achilles tendon ruptures during NCAA training or competition in 581 surveyed gymnasts was 17.2% (95% CI: 14.1%-20.3%). Seventy-eight percent of the ruptures occurred on the back-tumbling back handspring take-off on floor. A significantly higher percentage of gymnasts with Achilles tendon ruptures competed at the elite level before college, performed difficult floor skills and vaults before and during college, competed in all 4 events (all-around), competed at a Division I school, used retinoid medications in college, and identified as Black or African American. Adjusted regression analyses show that Black or African American race and retinoid use were independent risk factors for Achilles tendon ruptures.
A healthy Achilles tendon is necessary for rapid transfer of energy during locomotion, and athletes have an increased risk of injury because of the large cyclic loads experienced during running and jumping. 1 Current trends suggest that the prevalence of Achilles tendon injuries within the general population is increasing.13,23 Past epidemiological data showed that ruptures occur most commonly among men aged 20 to 39 years, with women aged 40 to 59 years having the second-highest risk for injury. 23 Among the athletic population, most athletes who suffer from spontaneous Achilles tendon ruptures participate in recreational levels, although 8% to 20% of ruptures occur at the competitive level. 24 Among the professional cohort, it has been shown that approximately 25% of professional athletes are unable to return to play after an Achilles tendon rupture. 18 In terms of gender comparison, although studies have shown that men are at increased risk, 23 a recent study by Chan et al 7 discovered that women’s collegiate gymnastics had the highest rates of Achilles tendon injuries when compared with 16 other men’s and women’s NCAA sports. Specifically, the rate of all Achilles injuries (ie, complete and partial Achilles tendon tears/ruptures) in women’s gymnastics was 4-fold greater than in men’s basketball, which held the second-highest reported injury rate per 100,000 athlete exposures. 7 For severe Achilles injuries, including spontaneous ruptures, the rate in women’s gymnastics was found to be 16 times higher than in women’s track, 13 times higher than in men’s basketball, and 60 times higher than in women’s soccer. 7
By the time a gymnast begins his/her collegiate career, he/she has been involved in gymnastics training for an average of approximately 13 years. 26 Prior studies have demonstrated that past training variables, such as repetitive overuse and intensity of sports training and competition, are risk factors for Achilles tendon ruptures. 46 We hypothesized that these variables would influence risk of rupture in collegiate gymnasts. As early sport specialization has been shown to increase the risk of injury,2,30,31,39 we hypothesized that specialization in gymnastics before age 10 years would be a risk factor for Achilles tendon rupture. However, we found no differences in age of specialization between the rupture and no-rupture groups when assessing at cut-offs below ages 8 through 12 years. Overall, 63.0% of the gymnasts specialized at age 8 and less, 69.4% at age 9 and less, 79.4% at age 10 and less, 83.8% at age 11 and less, and 90% at age 12 and less. The lack of difference in gymnasts’ age at specialization suggests that training duration might not influence risk of Achilles tendon rupture. However, training volume might, as we found a significantly greater percentage of ruptures among those who competed at the elite level before college. This is understandable given the context of elite gymnastics, as the skill requirements involve greater difficulty than level 10 or lower. Reports have shown elite training schedules to range from 30 to 42 hours per week, compared with 24 to 36 hours per week among levels 9 and 10. 40 In this study, 70% of the gymnasts that reached the elite level reported training 30+ hours per week, compared with only 12% of level 9 or 10 gymnasts. Prior work by Emerson et al 11 has demonstrated that both male and female elite gymnasts have greater Achilles tendon thickness compared with nonelite gymnasts, and that gymnasts, regardless of level of expertise, have greater Achilles tendon thickness, hyperechoic areas, and paratenon thickening than nongymnast controls. These structural changes seen in gymnasts’ Achilles tendons suggest that gymnastics training volume may play a role in future tendinopathy or tears. In addition to increased volume of training, greater difficulty of elements performed, especially on floor and vault, may lead to increased loading of the Achilles tendon and potentially escalate injury risk. Our data showed a greater percentage of Achilles tendon ruptures in gymnasts who trained both difficult floor passes and vaults before college. There were no significant differences in the number of gymnasts who trained >20 hours per week before college in the rupture versus nonrupture groups. We did not collect any data on training hours during college; however, NCAA rules limit mandatory practices to no more than 20 hours per week. 32
In addition to precollege risk factors for Achilles tendon rupture, we examined how training and other factors during college might influence risk. We found that a greater percentage of gymnasts who performed difficult floor passes and vaults during college sustained Achilles ruptures. The take-off and landing of floor tumbling skills have been shown to reach up to 15 times the gymnast’s bodyweight, and more difficult skills likely result in loads at the upper end of that range compared with less difficult skills requiring less height and rotations.5,41 Unlike in precollege training, where all gymnasts participate in all 4 events, collegiate gymnasts may specialize in any number of events, while only some compete in all 4 events, also known as all-around competitors. Prior work has shown that event participation influences injury risk, with floor participation holding the highest risk of injury. 6 Our study did find that the majority of Achilles ruptures occurred on floor. Additionally, a significantly higher percentage of gymnasts with ruptures competed in the all-around. This suggests that overall loading with training multiple events might be a factor in Achilles tendon rupture. A growing body of evidence suggests that repetitive microtrauma to the Achilles tendon may cause degenerative changes in the mechanical properties that reduce its capacity to handle rapid high force eccentric loading.20,48 These include increased type III collagen levels with a disorganized collagen matrix, compared with type I collagen with a more organized collagen matrix, markers of chronic inflammation, and an ingrowth of neovascularization.9,43 Gymnasts who compete more difficult skills and all 4 events are exposed to greater Achilles tendon repetitive loading, which may predispose them to tendon structural changes. Finally, we found that a significantly greater percentage of gymnasts with ruptures attended a college located in the western United States. Many western states (California, Utah, Oregon, Washington, Arizona) belong to historically competitive conferences in women’s collegiate gymnastics. A possible explanation for the increase in ruptures reported among gymnasts competing in this region could be attributed to the training intensity and greater difficulty of skills performed before and during college. Future work should examine the relationship between tendon pathology and training load in terms of events trained/competed and skills performed in a prospective manner.
Many young gymnasts are diagnosed with calcaneal apophysitis, ankle sprains, and anterior talocrural impact injuries that negatively influence foot and ankle integrity.6,15,45 We hypothesized that those with a history of these injuries might be at a greater risk of spontaneous Achilles tendon rupture. Contrary to expectation, there were no significant differences in the percentages of gymnasts with and without Achilles tendon ruptures who had calcaneal apophysitis or Achilles tendinopathy before or during college. Interestingly, gymnasts who reported a spontaneous Achilles tendon rupture were less likely to have had prior ankle or hip injuries. One possible explanation is that these gymnasts with prior ankle and hip injuries may have been limited in their ability to train more difficult skills or as many events later on in their careers because of setbacks in training. Additionally, prior lower limb injuries may have resulted in reduced skill repetition during training, decreasing the overall loading on the Achilles tendon and further protecting from future rupture. Despite these findings, 70.9% of those who experienced an Achilles tendon rupture during college reported pain 4 weeks before the injury. Past studies exploring the presence of prior symptoms have been mixed, with reports of Achilles tendon pain before ruptures ranging from 4% to 32% of cases reported.4,22,37 Given the range identified in past studies pertaining to prerupture pain, as well as the high percentage of prior symptoms reported in the current study, monitoring for prerupture symptoms is warranted.
One unexpected finding from the study was that Black or African American race was independently associated with a greater risk of Achilles tendon rupture in gymnasts. The association between race and tendon injuries has been shown in epidemiological studies of Achilles tendon injuries.10,23,36 Davis et al 10 explored the rate of Achilles tendon repair from 1994 to 1996 and found those who identified as African American to be at greater risk of undergoing surgical repair compared with those who identified as White. Similarly, a 2007 study explored racial differences in tendon rupture incidence over a 4-year period using a military database and found a significant increase in lower-extremity major tendon rupture in Black US service members compared with White service members. 36 Furthermore, White et al 47 found the rate of major tendon ruptures (Achilles, patellar, pectoralis major, and quadriceps tendons) to be 13 times greater for Black men compared with White men. Past literature has also explored whether racial differences in tendon properties may contribute to the disproportionate rates of ruptures seen in African Americans. Fukashiro et al 12 investigated the racial differences in viscoelastic characteristics of the triceps surae muscle group between Black and White collegiate sprinting athletes and found that muscle stiffness among Black athletes was significantly greater compared with White athletes, and greater tendon stiffness has been proposed to increase the risk of rupture. Racial differences in ABO blood type have also been proposed to explain the increased prevalence of ruptures among certain groups. Garratty et al 14 explored ABO and Rh(D) phenotype frequencies and found that a greater percentage of Black or African Americans in the United States have type O blood than White race. In an early study, Józsa et al 19 found that those with ABO blood type O have an increased risk of tendon rupture. 19 Two years after that study, Kujala et al 21 discovered a similar result, finding a lower A:O blood type ratio among those who reported an Achilles tendon rupture than those without a rupture (1.0 vs 1.42). However, more recent studies done outside the United States did not find any association between ABO blood type and Achilles tendon rupture.25,29 More research is needed to determine the relationship between race and systemic factors that influence tendon health and whether these differences directly contribute to greater rupture risk.
Medications such as corticosteroid injections or oral steroids are known risk factors for Achilles tendon rupture, as they can interfere with tendons’ healing processes and may mask pain associated with tendon damage, putting an athlete at a greater risk of injury. 28 Other medications such as fluoroquinolone antibiotics and oral contraceptive pills can also contribute to tendon pathology.16,28 In our study, there were no differences in the use of oral steroids, fluoroquinolone antibiotics, or oral contraceptive pills between gymnasts with or without Achilles tendon ruptures during NCAA competition/training. However, a statistically greater percentage of gymnasts with Achilles tendon rupture took retinoids than those without rupture. Retinoids are frequently used for the treatment of acne in adolescents and have been shown to contribute to advanced bone age, premature epiphyseal fusion, and Achilles tendinopathy due to the mechanism of a retinoic acid-metabolizing enzyme.3,27,35 Beytemür et al 3 explored the histopathological and biomechanical effects of isotretinoin on Achilles tendons in rats and discovered tendon properties, including elastic modulus, yield force, and ultimate tensile force, to be significantly lower in isotretinoin-exposed rats. A possible clinical implication of our results suggests that reducing retinoid medication use during collegiate gymnastics may lower the risk of Achilles tendon ruptures. However, future work should clarify oral versus topical retinoids, and more clinical studies are needed to determine the effects of systemic and topical retinoid medications on tendon pathology.
As past studies have shown that the female athlete triad increases the risk of sustaining musculoskeletal injuries,42,44 we hypothesized that gymnasts who reported characteristics of the triad (disordered eating patterns) would be more likely to have had an Achilles tendon rupture. However, data from the triad screening portion of the survey revealed the opposite: worries about body weight composition, limited food intake, and eating in secret were significantly less common in gymnasts who had an Achilles tendon rupture. There were no differences in reported menstrual cycle irregularity between groups. There were also no significant associations between BMI and ruptures, which eliminates the suspicion that elevated BMI could have contributed to increased loading and ultimately a greater risk of an Achilles tendon rupture. Further research investigating the relationship between nutrition, disordered eating, and Achilles tendon ruptures is warranted, particularly performed in a prospective manner with a more in-depth exploration of triad variables than is possible with a survey.
The findings of our study should be considered with the following limitations. As with any survey, recall bias when reporting past training and injury data is an important consideration, especially in gymnasts who completed their collegiate training several years before taking the survey. Additionally, gymnasts who sustained Achilles tendon ruptures may have been more inclined to participate in the survey than those who did not. Furthermore, there may be a lack of generalization because of unequal distributions among geographic regions and NCAA divisions, primarily because of the anonymous nature of the survey. In addition, the validity and reliability of this survey have not been established in prior studies. Only a small percentage of survey responses contained missing data, but it is possible this limited our analysis. Our study did not collect any data on loading factors such as exact numbers of skill repetitions, take-off angle measurements, technique analysis, or landing surfaces used, which likely play a role in Achilles tendon injury. Last, because of the lack of exposure information and training hours from the gymnast participants, we could not directly compare our findings to previous literature that expressed injury rate per athlete exposures.
Conclusion
Female collegiate gymnasts are at the greatest risk of Achilles tendon ruptures compared with all other college sports. The prevalence of Achilles tendon ruptures during NCAA training or competition among female gymnasts surveyed was 17.2% (95% CI: 14.1%-20.3%). Potential risk factors identified by this survey included high-level training before college, performance of difficult floor skills and vaults before and during college, competing as an all-around gymnast during college, use of retinoid medications, and identification as Black or African American. This study serves as a first step to unearth the risk factors associated with the unparalleled prevalence of Achilles tendon ruptures in women’s collegiate gymnastics. Future studies investigating the mechanisms for Achilles tendon rupture in female gymnasts, including acute and chronic training variables and systemic factors, will hopefully lead to identification of at-risk athletes and effective Achilles tendon injury prevention strategies.
Footnotes
The authors report no potential conflicts of interest in the development and publication of this article.
This work was supported by grant number UL1 TR 002384 from the National Center for Advancing Translational Sciences (NCATS) of the National Institutes of Health (NIH) for REDCap use.
References
- 1. Arampatzis A, Karamanidis K, Morey-Klapsing G, De Monte G, Stafilidis S. Mechanical properties of the triceps surae tendon and aponeurosis in relation to intensity of sport activity. J Biomech. 2007;40:1946-1952. [DOI] [PubMed] [Google Scholar]
- 2. Bell DR, Post EG, Biese K, Bay C, Valovich McLeod T. Sport specialization and risk of overuse injuries: a systematic review with meta-analysis. Pediatrics. 2018;142:e20180657. [DOI] [PubMed] [Google Scholar]
- 3. Beytemür O, Yüksel S, Tetikkurt ÜS, Genç E, Olcay E, Güleç A. Isotretinoin induced achilles tendinopathy: histopathological and biomechanical evaluation on rats. Acta Orthop Traumatol Turc. 2018;52:387-391. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Böhm E, Thiel A, Czieske S. Achilles tendon rupture. Anamnestic and morphologic studies and considerations on the etiology [in German]. Sportverletz Sportschaden. 1990;4:22-28. [DOI] [PubMed] [Google Scholar]
- 5. Bruggemann G. Mechanical load in artistic gymnastics and its relation to apparatus and performance. In: Leglise M, ed. Symposium Medico-Technique. International Gymnastics Federation; 1999:17-27. [Google Scholar]
- 6. Campbell RA, Bradshaw EJ, Ball NB, Pease DL, Spratford W. Injury epidemiology and risk factors in competitive artistic gymnasts: a systematic review. Br J Sports Med. 2019;53:1056-1069. [DOI] [PubMed] [Google Scholar]
- 7. Chan JJ, Chen KK, Sarker S, et al. Epidemiology of Achilles tendon injuries in collegiate level athletes in the United States. Int Orthop. 2020;44:585-594. [DOI] [PubMed] [Google Scholar]
- 8. College Gym News. Injury updates. Accessed February 10, 2021. https://collegegymnews.com/injury-updates
- 9. Dakin SG, Newton J, Martinez FO, et al. Chronic inflammation is a feature of Achilles tendinopathy and rupture. Br J Sports Med. 2018;52:359-367. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Davis JJ, Mason KT, Clark DA. Achilles tendon ruptures stratified by age, race, and cause of injury among active duty U.S. Military members. Mil Med. 1999;164:872-873. [PubMed] [Google Scholar]
- 11. Emerson C, Morrissey D, Perry M, Jalan R. Ultrasonographically detected changes in Achilles tendons and self reported symptoms in elite gymnasts compared with controls—an observational study. Man Ther. 2010;15:37-42. [DOI] [PubMed] [Google Scholar]
- 12. Fukashiro S, Abe T, Shibayama A, Brechue WF. Comparison of viscoelastic characteristics in triceps surae between Black and White athletes. Acta Physiol Scand. 2002;175:183-187. [DOI] [PubMed] [Google Scholar]
- 13. Ganestam A, Kallemose T, Troelsen A, Barfod KW. Increasing incidence of acute Achilles tendon rupture and a noticeable decline in surgical treatment from 1994 to 2013. A nationwide registry study of 33,160 patients. Knee Surg Sports Traumatol Arthrosc. 2016;24:3730-3737. [DOI] [PubMed] [Google Scholar]
- 14. Garratty G, Glynn SA, McEntire R. ABO and Rh(D) phenotype frequencies of different racial/ethnic groups in the United States. Transfusion. 2004;44:703-706. [DOI] [PubMed] [Google Scholar]
- 15. Hart E, Meehan WP, 3rd, Bae DS, d’Hemecourt P, Stracciolini A. The young injured gymnast: a literature review and discussion. Curr Sports Med Rep. 2018;17:366-375. [DOI] [PubMed] [Google Scholar]
- 16. Hicks KM, Onambélé-Pearson G, Winwood K, Morse CI. Oral contraceptive pill use and the susceptibility to markers of exercise-induced muscle damage. Eur J Appl Physiol. 2017;117:1393-1402. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Jemni M. The Science of Gymnastics: Advanced Concepts. Routledge; 2017. [Google Scholar]
- 18. Johns W, Walley KC, Seedat R, Thordarson DB, Jackson B, Gonzalez T. Career outlook and performance of professional athletes after Achilles tendon rupture: a systematic review. Foot Ankle Int. 2021;42:495-509. [DOI] [PubMed] [Google Scholar]
- 19. Jozsa L, Balint JB, Kannus P, Reffy A, Barzo M. Distribution of blood groups in patients with tendon rupture. An analysis of 832 cases. J Bone Joint Surg Br. 1989;71:272-274. [DOI] [PubMed] [Google Scholar]
- 20. Klatte-Schulz F, Minkwitz S, Schmock A, et al. Different Achilles tendon pathologies show distinct histological and molecular characteristics. Int J Mol Sci. 2018;19:404. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Kujala UM, Järvinen M, Natri A, et al. ABO blood groups and musculoskeletal injuries. Injury. 1992;23:131-133. [DOI] [PubMed] [Google Scholar]
- 22. Lea RB, Smith L. Non-surgical treatment of tendo achillis rupture. J Bone Joint Surg Am. 1972;54:1398-1407. [PubMed] [Google Scholar]
- 23. Lemme NJ, Li NY, DeFroda SF, Kleiner J, Owens BD. Epidemiology of Achilles tendon ruptures in the United States: athletic and nonathletic injuries from 2012 to 2016. Orthop J Sports Med. 2018;6:2325967118808238. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Leppilahti J, Orava S. Total Achilles tendon rupture. A review. Sports Med. 1998;25:79-100. [DOI] [PubMed] [Google Scholar]
- 25. Leppilahti J, Puranen J, Orava S. ABO blood group and Achilles tendon rupture. Ann Chir Gynaecol. 1996;85:369-371. [PubMed] [Google Scholar]
- 26. Ling D, Sleeper M, Casey E. Identification of risk factors for injury in women’s collegiate gymnastics with the gymnastics functional measurement tool. PM R. 2020;12:43-48. [DOI] [PubMed] [Google Scholar]
- 27. Luthi F, Eggel Y, Theumann N. Premature epiphyseal closure in an adolescent treated by retinoids for acne: an unusual cause of anterior knee pain. Joint Bone Spine. 2012;79:314-316. [DOI] [PubMed] [Google Scholar]
- 28. Maffulli N. Current concepts in the management of subcutaneous tears of the Achilles tendon. Bull Hosp Jt Dis. 1998;57:152-158. [PubMed] [Google Scholar]
- 29. Maffulli N, Reaper JA, Waterston SW, Ahya T. ABO blood groups and Achilles tendon rupture in the Grampian Region of Scotland. Clin J Sport Med. 2000;10:269-271. [DOI] [PubMed] [Google Scholar]
- 30. McGowan J, Whatman C, Walters S. The associations of early specialisation and sport volume with musculoskeletal injury in New Zealand children. J Sci Med Sport. 2020;23:139-144. [DOI] [PubMed] [Google Scholar]
- 31. Myer GD, Jayanthi N, Difiori JP, et al. Sport specialization, part I: does early sports specialization increase negative outcomes and reduce the opportunity for success in young athletes? Sports Health. 2015;7:437-442. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32. National Collegiate Athletic Association. Defining countable athletically related activities. Accessed June 17, 2021. https://www.ncaa.org/sites/default/files/Charts.pdf
- 33. National Institutes of Health. Racial and ethnic categories and definitions for NIH diversity programs and for other reporting purposes. Accessed March 1, 2021. https://grants.nih.gov/grants/guide/notice-files/NOT-OD-15-089.html
- 34. Nattiv A, Loucks AB, Manore MM, et al. American College of Sports Medicine position stand. The female athlete triad. Med Sci Sports Exerc. 2007;39:1867-1882. [DOI] [PubMed] [Google Scholar]
- 35. Nilsson O, Isoherranen N, Guo MH, et al. Accelerated skeletal maturation in disorders of retinoic acid metabolism: a case report and focused review of the literature. Horm Metab Res. 2016;48:737-744. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36. Owens B, Mountcastle S, White D. Racial differences in tendon rupture incidence. Int J Sports Med. 2007;28:617-620. [DOI] [PubMed] [Google Scholar]
- 37. Park YH, Kim TJ, Choi GW, Kim HJ. Achilles tendinosis does not always precede Achilles tendon rupture. Knee Surg Sports Traumatol Arthrosc. 2019;27:3297-3303. [DOI] [PubMed] [Google Scholar]
- 38. Peterson JG, Tjong VK, Mehta MP, Goyette BN, Patel M, Kadakia AR. A qualitative assessment of return to sport following Achilles tendon repair.J Orthop. 2021;23:46-51. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39. Post EG, Trigsted SM, Riekena JW, et al. The association of sport specialization and training volume with injury history in youth athletes. Am J Sports Med. 2017;45:1405-1412. [DOI] [PubMed] [Google Scholar]
- 40. Ritchie L, Ronshaugen N, Sygo J. Growth and development in gymnastics. In: Sweeney E, ed. Gymnastics Medicine. Springer International; 2020:55-73. [Google Scholar]
- 41. Sands B, Caine DJ, Borms J. Scientific Aspects of Women’s Gymnastics. Medicine and Sports Science. Vol. 45. Karger; 2003. [Google Scholar]
- 42. Scheid JL, Stefanik ME. Drive for thinness predicts musculoskeletal injuries in Division II NCAA female athletes. J Funct Morphol Kinesiol. 2019;4:52. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43. Splittgerber LE, Ihm JM. Significance of asymptomatic tendon pathology in athletes. Curr Sports Med Rep. 2019;18:192-200. [DOI] [PubMed] [Google Scholar]
- 44. Sweeney E, Howell DR, Seehusen CN, Tilley D, Casey E. Health outcomes among former female collegiate gymnasts: the influence of sport specialization, concussion, and disordered eating. Phys Sportsmed. Published online November 23, 2020. doi: 10.1080/00913847.2020.1850150. [DOI] [PubMed] [Google Scholar]
- 45. Thomas RE, Thomas BC. A systematic review of injuries in gymnastics. Phys Sportsmed. 2019;47:96-121. [DOI] [PubMed] [Google Scholar]
- 46. Wertz J, Galli M, Borchers JR. Achilles tendon rupture: risk assessment for aerial and ground athletes. Sports Health. 2013;5:407-409. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 47. White DW, Wenke JC, Mosely DS, Mountcastle SB, Basamania CJ. Incidence of major tendon ruptures and anterior cruciate ligament tears in US Army soldiers. Am J Sports Med. 2007;35:1308-1314. [DOI] [PubMed] [Google Scholar]
- 48. Winnicki K, Ochała-Kłos A, Rutowicz B, Pękala PA, Tomaszewski KA. Functional anatomy, histology and biomechanics of the human Achilles tendon—a comprehensive review. Ann Anat. 2020;229:151461. [DOI] [PubMed] [Google Scholar]