sign in sign up
<<
Home , Anabolic steroid

Authors: Scott Horn, DO Patricia Gregory, MD Musculoskeletal Kevin M. Guskiewicz, PhD, ATC

Affiliations: From the Department of Physical Medicine and Rehabilitation (SH, ORIGINAL RESEARCH ARTICLE PG), The University of North Carolina at Chapel Hill; Department of Exercise and Science (KMG), The University of North Carolina at Chapel Hill, Chapel Hill, North Self-Reported Anabolic-Androgenic Carolina. Use and Musculoskeletal Correspondence: All correspondence and requests for reprints should be addressed to Scott Findings from the Center for the Study of Retired Horn, DO, 20 Porters Glen Place, Durham, NC 27713. Health Survey of Retired NFL Players

Disclosures: This work was part of a research ABSTRACT presentation for the University of Horn S, Gregory P, Guskiewicz KM: Self-reported anabolic-androgenic steroids use North Carolina Physical Medicine and Rehabilitation Residency Program. and musculoskeletal injuries: Findings from the center for the study of retired athletes Some of these results were published health survey of retired NFL players. Am J Phys Med Rehabil 2009;88:192–200. in an abstract form and poster presentation at the 2008 Association Objective: The relationship between musculoskeletal injuries and an- of Academic Physiatrists Annual abolic-androgenic steroids is not well understood. The purpose of our Meeting, Anaheim, CA, February 20, 2008. study was to investigate the association between self-reported anabolic- androgenic steroids use and the prevalence of musculoskeletal injuries in 0894-9115/09/8803-0192/0 a unique group of retired professional players. American Journal of Physical Medicine & Rehabilitation Design: A general health questionnaire was completed by 2552 Copyright © 2009 by Lippincott retired professional football players. Survey data were collected between Williams & Wilkins May 2001 and April 2003. Results of self-reported musculoskeletal DOI: 10.1097/PHM.0b013e318198b622 injuries were compared with the use of anabolic-androgenic steroids using frequency distributions and ␹2 analyses. Results: Of the retired players, 9.1% reported using anabolic-androgenic steroids during their professional career. A total of 16.3% of all offensive line and 14.8% of all defensive line players reported using anabolic-androgenic steroids. Self-reported anabolic-androgenic steroids use was significantly associated (P Ͻ 0.05) with the following self-reported, medically diagnosed, joint and cartilaginous injuries in comparison with the nonanabolic-androgenic steroids users: disc herniations, knee ligamentous/meniscal , elbow injuries, neck stinger/burner, spine injury, and foot/toe/ankle injuries. There was no association between anabolic-androgenic steroids use and reported muscle/tendon injuries. Conclusions: Our findings demonstrate that an association may exist between anabolic-androgenic steroids use and the prevalence of reported musculoskeletal injury sustained during a professional football career, partic- ularly ligamentous/joint-related injuries. There may also be an associated predis- position to selected types of injuries in anabolic-androgenic steroids users.

Key Words: Anabolic Agents, Football, Ligaments, Musculoskeletal System

192 Am. J. Phys. Med. Rehabil. ● Vol. 88, No. 3 animal studies. It has been suggested that the Anabolic use in the United States dates mechanism of these injuries is based on an alter- back to the 1950s when Bob Hoffman, inventor of ation of collagen structure.9 Methandienone in- the barbell and famous York Barbell Company and jected into exercised mice produced collagen ab- Gym, as well as coach of the United States Olympic normalities, which has been suggested as a change weight lifting team, learned of anabolic-androgenic that may increase the risk of tendon rupture.11 steroids (AAS) at the 1954 World Championships in Other studies conducted on rats also indicated ste- Vienna. Subsequently, the appointed physician of roid-induced alterations in tendon elasticity, which the Olympic weight lifting team, John Ziegler, gave may place the tendon at an increased risk of fail- the steroid to the entire United States Olympic ure.12,13 The pathophysiologic theory, up to this weight lifting team in the 1960 Rome games, but point, is that there may be a rapid strength adap- still lost to the Soviets as they had already devel- tation produced by AAS in , which is oped the next generation of anabolic performance not equally matched by less vascular and slower enhancers.1–3 adapting tendons. This may place the tendon at a The prevalence and the documentation of higher risk for injury because it is the weakest link AAS in both noncompetitive athletes as well as in the chain.14 competitive or elite athletes vary widely. Accord- ing to the Monitoring the Future Study by the There have been limited reports of the preva- National Institute on Abuse, 2.7% of 12th lence of MSI in NFL players. Most of the published graders admitted to using steroids in their life- reports on NFL players have focused on the epide- 15–17 time.4 The National Collegiate Athletic Associa- miology of concussion. These studies describe tion released a 5-yr drug report, which showed the mechanism of injury as a translational accel- that during year-round testing, 1% of student- eration force resulting from the impact to the 15 athletes randomly tested, tested positive for ste- facemask, side, or back of the helmet. These roids in 1999–2000. Incidence seems to vary injuries are most commonly seen in quarterbacks, greatly by with football ranking second receivers, or defensive secondary players.16 These highest (3%) to water polo.5,6 repeated concussive injuries were also associated There is very little, if any, established data with a higher prevalence of mild cognitive impair- regarding the prevalence of AAS in professional ment and memory problems in later life.17 Another football players. It is generally believed that AAS study evaluated the health status of retired NFL was introduced to professional football around players who played in Super Bowl III.18 This study 1963, when Alvin Roy, a York weight lifter, became showed that although there was a high prevalence strength coach of The American Football League’s of arthritis, health status was similar to age- San Diego Chargers. Over the next few years, play- matched controls. One study that evaluated the ers and coaches moved teams and AAS use spread rate of shoulder injuries among players on a North 7 through professional and college football players. American Professional Football team19 showed that Most of the assumptions of prevalence during the the most common magnetic resonance imaging 1970s and 1980s are based on testimonies of finding among these players was rotator cuff con- former players. For instance, Pat Donovan, a Dallas tusions related to a direct blow. Cowboys offensive lineman for 9 yrs who retired in MSI have not been widely evaluated in NFL 1983, commented that AAS use ran as high as players, but there are a few studies that evaluated 60%–70% on the Cowboys’ offensive and defensive National Collegiate Athletics Association play- lines.8 Reports on the suspected usage by profes- ers.20,21 One study evaluated the rates of injury sional football players during the same era varied, among Big Ten Football conference at both spring with players suggesting that anywhere from 40% to and fall practices.20 These rates were noted to be 90% of players had used anabolic steroids.8 In an unpublished survey to current National Football League higher in the spring compared with the fall. An- other study that evaluated the level of injury (NFL) players in 1990, with a respondent rate of only 21 7.5% (n ϭ 120), 28% overall, and 67% of offensive among draft prospects to the NFL showed prev- linemen reported previous use.9 alent rates of ankle, knee, neck, and hand injuries. Although the effects of AAS on many different The impact of AAS use on the muscle has been organ systems may not be well understood, many previously documented, but the impact on the of the effects on the muscular system in terms of musculoskeletal system and risk of injury has not increased strength and hypertrophy are well docu- been previously evaluated. Therefore, the purpose mented.10 However, the effects of AAS on muscu- of our study was to evaluate the associations be- loskeletal injuries (MSI) are not well documented tween self-reported AAS and MSI. To our knowl- or understood. The link between MSI and AAS is edge, this is the first human study conducted to based on a few published case reports and selected evaluate such associations.

March 2009 Steroid Use and Musculoskeletal Injuries 193 METHODS ical and bodily pain, and general health, as well as a Participants mental health component score, which include scores of vitality, social functioning role, emotional, We studied a diverse group of retired profes- and mental health. These scores were compared with sional football players, including recent retirees age- and gender-specific population-based norms es- and those who played professional football before tablished by previous researchers. World War II. All participants played a minimum of The Center for the Study of Retired Athletes at two seasons of professional football. We studied the University of North Carolina at Chapel Hill this group using a self-reported general health initially mailed the general health questionnaire in questionnaire approved by The University of North May 2001, followed by remailings to nonrespon- Carolina Institutional Review Board. It was ex- dents in August 2001 and February 2002. The Cen- plained at the beginning of the survey that partic- ter for the Study of Retired Athletes at the Univer- ipants would not be identified and that research sity of North Carolina at Chapel Hill then began records would be kept confidential. By completing telephoning nonrespondents at different times of and submitting the survey, participants were ac- the day and completed the questionnaire over the knowledging that they agreed to take part in this telephone. The Center for the Study of Retired research study. Athletes at the University of North Carolina at Chapel Hill then conducted a reliability check of General Health Questionnaire the general health questionnaire by readminister- The general health questionnaire was first sent ing the instrument to 83 of the original respon- to all living members of the NFL Retired Player’s dents 24–36 mos later to establish a high level of Association (n ϭ 3683) through the Center for the agreement between selected responses. Study of Retired Athletes at the University of North Carolina at Chapel Hill. The questionnaire involved Data Analysis a variety of items about musculoskeletal, cardio- Descriptive frequency analyses were conducted vascular, and neurologic conditions that the re- to first provide an overview of the sample relative tired player may have experienced during and after to selected variables such as race, marital status, his football career. The musculoskeletal section, education, position played, and calculated playing which was the focus of this study, included ques- body mass index (BMI) (based on reported height tions about the number of different , ligament, and playing weight). ␹2 analyses were then con- and muscle injuries that were sustained and med- ducted to identify associations and trends between ically diagnosed by a healthcare professional or anabolic steroid use and types of MSI, playing BMI, physician at any point during their high school, position played, era played, diagnosis of , college, armed forces, and professional careers. The and use. Level of significance for all anal- questionnaire also asked retired players about the yses was set a priori at P Ͻ 0.05. All statistical total number of injuries, , time withheld analyses were performed using SPSS statistical from play because of injury, and whether they are software (SPSS Inc., Chicago, IL).23 still affected by these injuries. Solicitation of infor- mation focused on tendonous, ligamentous, and RESULTS cartilage injuries in the surrounding joints in the Participants upper and lower limbs. Additional information was Of the original 3683 general health surveys captured regarding subsequent injuries sustained sent to retired players, 2552 (69.3%) were com- after retirement. The retired athletes also com- pleted. The age of the respondents averaged 53.8 pleted a section on , which in- (Ϯ13.4) yrs, with an average professional football- cluded questions about alcohol, , and ste- playing career of 6.6 (Ϯ3.6) yrs. Respondents re- roid use both during and after completion of their ported having played organized football (junior professional careers. Of particular interest to our high school, high school, college, armed service, study was the question on steroid use. The ques- and professional) for an average of 15.1 (Ϯ4.3) yrs. tion was phrased as follows: “During the time in When considering the prevalence of AAS use, 233 which it was acceptable to use performance-en- (9.1%) of the retired players reported they used hancing steroids, did you use steroids?” AAS during their professional career. The highest In addition, the questionnaire included the reported prevalence (20.3%) was observed in those SF-3622 measurement model for functional assess- who played primarily during the 1980s (␹2 ϭ ment of health and well-being, which addresses 122.88; df ϭ 5; P ϭϽ0.005). Prevalence of self- how well the retired functions with activi- reported steroid use by race also revealed a signifi- ties of daily living. From the Short Form-36, we cant association, because whites (11.9%) were more calculated a physical health composite score, which likely to have used AAS compared with nonwhites includes scores of physical functioning, role of phys- (3.5%) (Table 1).

194 Horn et al. Am. J. Phys. Med. Rehabil. ● Vol. 88, No. 3 TABLE 1 Demographics 233 ؍ Steroids—Yes, n 2179 ؍ Steroids—No, n (90.3%) (9.1%) ␹2

Era played (n ϭ 2411) 1940s 96 (100) 0 (0) 122.33; df ϭ 5(P Ͻ 0.005) 1950s 296 (99) 3 (1) 1960s 544 (95.4) 26 (4.6) 1970s 568 (90.3) 61 (9.7) 1980s 413 (79.7) 105 (20.3) 1990s or later 261 (87.3) 38 (12.7) Race (n ϭ 2355) White 1516 (88.1) 205 (11.9) 37.907; df ϭ 1(P Ͻ 0.005) Nonwhite 612 (96.5) 22 (3.5) Education (n ϭ 2400) HS diploma 10 (71.4) 4 (28.6) 8.436; df ϭ 4(P ϭ 0.08) Some college 303 (89.1) 37 (10.9) AA degree 34 (87.2) 5 (12.8) BS degree 1266 (90.2) 137 (9.8) Postgraduate degree 555 (91.9) 49 (8.1) Playing BMI (n ϭ 2404) Ͻ18.50 3 (100) 0 (0) 74.191; df ϭ 3(P Ͻ 0.005) 18.51–24.99 156 (97.5) 4 (2.5) 25.00–29.99 1176 (94.2) 72 (5.8) Ͼ30.00 836 (84.2) 157 (15.8) Primary position (n ϭ 2409) Quaterback 127 (97.7) 3 (2.3) 85.063; df ϭ 8(P Ͻ 0.005) Running back 290 (97.0) 9 (3.0) Wide receiver 217 (97.3) 6 (2.7) Offensive line 489 (83.7) 95 (16.3) Tight end 113 (89.7) 13 (10.3) Defensive line 242 (85.2) 42 (14.8) Linebacker 274 (87.5) 39 (12.5) Defensive back 312 (95.1) 16 (4.9) Special teams 112 (91.8) 10 (8.2)

Subject Characteristics well as shoulder dislocation/shoulder injury did not In terms of BMI, 157 (15.8%), of the 933 (157 ϩ present significant associations with AAS use (Tables 836) respondents reporting a BMI of Ͼ30, also 2 and 3). reported using AAS. This represented a significant Although there were several lower-limb inju- association (␹2 ϭ 74.19; df ϭ 3; P ϭϽ0.005), ries that were not associated with AAS use, we whereby those with increasing BMIs were more found that AAS users had a higher prevalence (or likely to have used AAS during their playing ca- association) of injuries to the MCL, ACL, meniscus, reers. Stratified by playing position, the highest foot and toe, ankle, and knee (Tables 4 and 5). percentage of self-reported AAS lies in the positions Interestingly, musculotendinous injuries (up- typically requiring the larger size and strength per and lower limbs) were not more prevalent in including 95(16.3%) of 584 offensive line players, the AAS group. 42 (14.8%) of 284 defensive line players, 39 (12.5%) of 313 linebackers, and 13 (10.3%) of 126 Other Medical Conditions tight ends (␹2 ϭ 85.06; df ϭ 8; P ϭϽ0.005) (Table We observed significant associations between 1). We did not observe significant associations for depression diagnosis and steroid use (␹2 ϭ 24.14; education level and the likelihood of having used df ϭ 1; P ϭϽ0.005), and the likelihood of in- AAS (␹2 ϭ 8.44; df ϭ 4; P ϭ 0.08). creased alcohol consumption on a weekly basis (␹2 ϭ 41.44; df ϭ 3; P ϭϽ0.005) (Table 6). Injury Type We observed a significant association between In considering MSI of the upper limb and steroid use and a lower prevalence of diabetes and spine, we observed, an increased likelihood for , but higher prevalence of back pain, osteo- self-reported AAS users to have sustained inju- arthritis, depression, and attention deficit disorder ries to the elbow, wrist or hand, neck, and spine (Table 6). There was not a significant association (Tables 2 and 3). Biceps and triceps ruptures, as between all other medical conditions in the ques-

March 2009 Steroid Use and Musculoskeletal Injuries 195 TABLE 2 Upper-limb and back injuries No. Injuries 0 (%) 1 (%) 2؉ (%) ␹2

Shoulder injury Steroid use (n ϭ 229) 58.8 21.8 20.1 0.85; df ϭ 2(P ϭ 0.66) No steroid use (n ϭ 2154) 61.2 20.1 18.8 Shoulder dislocation Steroid use (n ϭ 231) 76.3 15.9 7.8 2.15; df ϭ 2(P ϭ 0.342) No steroid use (n ϭ 2162) 76.1 16.1 10.8 Elbow dislocation/sep Steroid use (n ϭ 232) 85.3 12.5 2.2 16.61; df ϭ 2(P Ͻ 0.005) No steroid use (n ϭ 2175) 92.1 5.7 2.2 Elbow injury Steroid use (n ϭ 229) 82.7 17.9 12.7 24.29; df ϭ 2(P Ͻ 0.005) No steroid use (n ϭ 2159) 69.4 10.4 6.9 Wrist/hand injury Steroid use (n ϭ 229) 53.3 19.7 27.1 12.92; df ϭ 2(P Ͻ 0.005) No steroid use (n ϭ 2150) 65.2 15.4 19.4 Neck injury Steroid use (n ϭ 228) 62.3 17.1 20.6 12.92; df ϭ 2(P Ͻ 0.005) No steroid use (n ϭ 2148) 75 12.4 12.7 Back injury Steroid use (n ϭ 227) 60.8 25.1 14.1 26.70; df ϭ 2(P Ͻ 0.005) No Steroid use (n ϭ 2151) 75.7 13.6 10.7 Neck stinger/burner Steroid use (n ϭ 227) 33.5 8.8 57.7 71.50; df ϭ 2(P Ͻ 0.005) No Steroid use (n ϭ 2139) 62.0 6.5 31.5 Disc herniation Steroid use (n ϭ 231) 79.2 12.1 8.7 29.14; df ϭ 2(P Ͻ 0.005) No steroid use (n ϭ 2162) 90.4 6.3 3.3 Bicep/triceps rupture Steroid use (n ϭ 232) 94.8 5.2 0 3.091; df ϭ 2(P Ͻ 0.213) No steroid use (n ϭ 2176) 94.3 4.5 1.2

tionnaire, which included, but was not limited to, findings, based on self-report of a large sample of disease, , , , retired professional players, are somewhat contrary to peripheral vascular disease, renal disease, , the anecdotal observations reported in a group of rheumatoid arthritis, Alzheimer’s disease, or Par- single patient case reports.24–31 Our questionnaire kinson’s disease. return rate of nearly 70% was unusually high, pro- viding a strong argument supporting the validity of DISCUSSION the findings. Nine percent (n ϭ 233) of the respon- Our study findings of retired professional football dents reported using AAS at some point during their players suggest a potential link between AAS use and football career. This finding is less than the number the likelihood of sustaining more ligamentous and previously reported in an unpublished survey done in cartilaginous injuries during a playing career. Our 1991.9

TABLE 3 Upper-limb injuries with odds ratios Steroids—Yes (%) Steroids—No(%) ␹2 Odds Ratio/95% CI

Neck injury 37.7 25.0 17.09; df ϭ 1(P Ͻ 0.005) 1.81 (1.36–2.41) Neck stinger/burner 66.5 38.0 69.35; df ϭ 1(P Ͻ 0.005) 3.25 (2.43–4.33) Back injury 39.2 24.3 23.83; df ϭ 1(P Ͻ 0.005) 2.01 (1.51–2.67) Disc herniations 20.8 9.6 27.52; df ϭ 1(P Ͻ 0.005) 2.48 (1.75–3.51) Shoulder injury 41.9 38.8 0.841; df ϭ 1(P ϭ 0.359) 1.14 (0.86–1.50) Shoulder dislocation 23.7 26.9 1.11; df ϭ 1(P ϭ 0.293) 0.84 (0.61–1.16) Biceps/triceps rupture 5.2 5.7 0.13; df ϭ 1(P ϭ 0.721 0.90 (0.49–1.65) Elbow injury 30.6 17.3 24.21; df ϭ 1(P Ͻ 0.005) 2.11 (1.56–2.85) Elbow dislocation 14.7 7.9 12.42; df ϭ 1(P Ͻ 0.005) 2.01 (1.35–2.99) Wrist/hand injury 46.7 36.8 12.70; df ϭ 1(P Ͻ 0.005) 1.64 (1.24–2.16)

196 Horn et al. Am. J. Phys. Med. Rehabil. ● Vol. 88, No. 3 TABLE 4 Lower-limb injuries No. Injuries 0, % 1, % 2؉,% ␹2

Medial collateral ligament Steroid use (n ϭ 231) 67.1 22.5 10.4 15.91; df ϭ 2(P Ͻ 0.005) No steroid use (n ϭ 2166) 78.6 14.8 6.6 Anterior collateral ligament Steroid use (n ϭ 232) 69.8 21.1 9.1 20.74; df ϭ 2(P Ͻ 0.005) No steroid use (n ϭ 2161) 81.9 13.7 0.9 PCL Steroid use (n ϭ 231) 88.7 8.7 2.6 5.07; df ϭ 2a (P ϭ 0.079) No steroid use (n ϭ 2163) 92.8 5.4 1.8 12.2; df ϭ 1(P ϭ 0.034) LCL Steroid use (n ϭ 231) 90.5 8.2 1.3 3.32; df ϭ 2(P ϭ 0.190) No steroid use (n ϭ 2164) 91.5 5.9 2.6 Ankle ligament tear Steroid use (n ϭ 231) 73.6 16.9 9.5 13.11; df ϭ 2(P Ͻ 0.005) No steroid use (n ϭ 2168) 83.2 10.9 6.0 Meniscus Steroid use (n ϭ 230) 57.8 19.1 23.0 17.05; df ϭ 2(P Ͻ 0.005) No steroid use (n ϭ 2164) 68.0 18.7 13.3 Foot and toe injuries Steroid use (n ϭ 229) 64.6 19.2 16.2 11.22; df ϭ 2(P Ͻ 0.005) No steroid use (n ϭ 2152) 74.8 14.3 10.9 Ankle injuries Steroid use (n ϭ 227) 54.2 23.3 22.5 7.61; df ϭ 2(P ϭ 0.022) No steroid use (n ϭ 2144) 63.0 17.3 19.7 Hip dislocation/fracture Steroid use (n ϭ 232) 97.4 2.6 0 1.39; df ϭ 2(P ϭ 0.501) No steroid use (n ϭ 2175) 97.2 2.3 0.6 HS/quad tear Steroid use (n ϭ 232) 72.4 16.8 10.8 2.44; df ϭ 2(P ϭ 0.295) No steroid use (n ϭ 2169) 73.6 13.5 12.9 Hip/thigh injuries Steroid use (n ϭ 227) 76.2 14.1 9.7 1.26; df ϭ 2(P ϭ 0.533) No steroid use (n ϭ 2155) 75.8 12.4 11.8 Knee injuries Steroid use (n ϭ 230) 29.1 27.4 43.5 17.63; df ϭ 2(PϽ 0.005) No steroid use (n ϭ 2160) 42.1 26.3 31.6 Knee/patella dislocations Steroid use (n ϭ 231) 90.0 6.5 3.5 1.96; df ϭ 2(P ϭ 3.75) No steroid use (n ϭ 2174) 92.6 5.0 0.3 Calf/Achilles tendon tear Steroid use (n ϭ 232) 93.1 6.9 0 5.73; df ϭ 2(P ϭ 0.57) No steroid use (n ϭ 2173) 93.5 4.8 1.7 Ankle/foot fracture Steroid use (n ϭ 230) 81.7 13.9 4.3 2.52; df ϭ 2(P ϭ 2.83) No steroid use (n ϭ 2170) 85.5 10.6 3.8 aFailed to meet expected value of 5; recategorized to 0 and 1ϩ injuries; ϩ reanalysis.

The notion that anabolic steroid usage is asso- mechanism of what places the cartilaginous struc- ciated with musculotendinous pathologies as op- tures at a higher risk of injury is unclear. Perhaps posed to damage to more inert structures such as the cartilaginous structures do not adapt as fast (or ligaments and cartilage was unsupported in our much at all) to the increased study and is inconsistent with previous theory. and force output. This may put them at greater risk Likewise, the absence of a significant association of injury, as these structures (not the tendon) are between steroid use and muscle/tendon injury was the weakest link in the chain. In addition, we surprising, as previous literature has suggested an speculate that this association could be due in part increased risk of tendon ruptures among steroid to the increased mass and stresses placed on the users. This theoretical increased risk of tendon ligaments and cartilage in players who typically rupture is based on two animal studies and a group take AAS. of case reports, which was theorized to be caused A factor to consider is that some of these by alterations in tendon elasticity.11,12 The exact athletes may have used AAS after their injury with

March 2009 Steroid Use and Musculoskeletal Injuries 197 TABLE 5 Lower-limb injury with odds ratios Steroids—Yes (%) Steroids—No (%) ␹2 Odds Ratio

Hip/thigh injury 23.8 24.2 0.17; df ϭ 1(P ϭ 0.897) 0.98 Hip dislocation/fracture 2.6 2.8 0.37; df ϭ 1(P ϭ 0.848) 0.92 MCL injury 32.9 21.4 15.91; df ϭ 1(P Ͻ 0.005) 1.80 ACL injury 30.2 18.1 19.48; df ϭ 1(P Ͻ 0.005) 1.95 PCL injury 11.3 7.2 5.00; df ϭ 1(P ϭ 0.025) 1.64 LCL injury 9.5 8.5 0.251; df ϭ 1(P ϭ 0.616) 1.13 Meniscus injury 42.2 32.0 9.69; df ϭ 1(P ϭ 0.002) 1.55 Knee injury 70.9 57.9 14.44; df ϭ 1(P Ͻ 0.005) 1.77 Knee/patella dislocations 10.0 7.4 1.38; df ϭ 1(P ϭ 0.165) 1.38 Ankle ligament tear 26.4 16.8 13.10; df ϭ 1(P Ͻ 0.005) 1.77 Ankle injury 45.8 37.0 6.80; df ϭ 1(P ϭ 0.009) 1.44 Ankle/foot fracture 18.3 14.5 2.37; df ϭ 1(P ϭ 0.124) 1.32 Calf/achilles tendon tear 6.9 6.5 0.45; df ϭ 1(P ϭ 0.833) 1.06 Foot/toe injury 35.4 25.2 11.00; df ϭ 1(P ϭ 0.001) 1.62

the intention of promoting a faster recovery, al- ries, and the presence of a posterior collateral lig- though the question in the survey was written in ament injury. such a way that reported the use of “perfor- The anabolic affects of steroids, as a general class, mance-enhancing steroids.” Although there are in clinical short-term studies produce a body weight clinical data that suggest AAS may protect from gain of ϳ3–5 kg34,35 and increases in strength repre- muscle in those with cachectic HIV- sented by the one repetition maximum in the bench infected or cancer as well as patients with trau- press.35,36 This well-documented increase in weight ma,32,33 it is unclear how much benefit this and strength can be very beneficial for professional would be to the athlete. Anecdotally this is a football players in such a physical sport, particularly major issue to consider, because many athletes for certain positions. Anecdotal reports that the will reportedly use AAS after an injury for fear of athletes that play positions requiring the most size losing their job, if they cannot perform at the and strength (defensive/offensive line, linebacker, previous level once the injury has healed. Al- and tight end) had higher prevalence of AAS were though it is a possibility that alternative use of supported by our findings. These positional players AAS contributed to our findings, given the way also typically have a greater BMI and our observed the question was phrased in the survey, we ex- significant association between AAS usage and size pect that this was a negligible number. was not surprising. It should be noted that there was a significant Our findings also demonstrated an association association between most, but not all, of the lower- between lifetime prevalence of depression and AAS limb ligament/joint structures and AAS use. There use. Clinical studies have shown a lifetime preva- was a significant association between AAS and the lence of ϳ13% in the general population and have frequency of MCL, ACL, meniscus, and ankle inju- found depressive symptoms related to withdrawal

TABLE 6 Other medical conditions Steroid—Yes (%) Steroids—No (%) ␹2

Diabetes 2.6 (n ϭ 6) 6.3 (n ϭ 138) 5.31; df ϭ 1; (P ϭ 0.010) Cancer 6.0 (n ϭ 14) 9.8 (n ϭ 213) 3.52; df ϭ 1; (P ϭ 0.034) Back pain 73.4 (n ϭ 171) 54.5 (n ϭ 1186) 30.44; df ϭ 1; (P Ͻ 0.005) Osteoarthritis 44.8 (n ϭ 103) 37.0 (n ϭ 805) 5.30; df ϭ 1; (P ϭ 0.013) Depression 21.0 (n ϭ 49) 10.3 (n ϭ 224) 24.14; df ϭ 1; (P Ͻ 0.005) Attention deficit disorder 6.5 (n ϭ 15) 2.8 (n ϭ 61) 9.21; df ϭ 1; (P ϭ 0.005) Days of week had EtOH during pro career None 4.8 (n ϭ 11) 20.8 (n ϭ 452) 41.443; df ϭ 3; (P Ͻ 0.005) 1–2 46.3 (n ϭ 107) 43 (n ϭ 934) 3–4 29 (n ϭ 67) 24.5 (n ϭ 532) 5–7 19.9 (n ϭ 46) 11.7 (n ϭ 254)

198 Horn et al. Am. J. Phys. Med. Rehabil. ● Vol. 88, No. 3 of AAS as high as 41%.37,38 In our study (Table 6), Another limitation lies in the variability of 21% of those who used steroids reported depres- steroid use within each player that reported using sion vs. 10% in the nonsteroid user group (␹2 ϭ AAS. The question, as was phrased in the question- 24.14; df ϭ 1; P Ͻ 0.005). The cause of association naire, simply asked “if” the player had ever used in our data is unclear. Depression may be directly and did not take into account the dosage, type, related to steroid withdrawal or depression may frequency, or time frame of AAS use. There is a place the player at a higher risk of using AAS. We large spectrum of different that exist and are speculate that it is probably caused by a combina- often used in conjunction with each other as well tion of both factors. as consecutively. The effect that this has on the Although not the main focus of the study, we data is unclear. did observe for other medical conditions that are often thought to be associated with anabolic ste- CONCLUSION roid use (Table 6). No clear association of AAS and Our findings suggest an association between AAS diabetes is available in the medical literature. use and the incidence of MSI sustained during a There is one case study of a bodybuilder with a professional football career. Of particular interest was 15-yr history of using AAS that developed acute the association between ligamentous/joint-related in- renal failure, , and diabetes.39 Interest- juries and AAS use, whereas musculotendinous inju- ingly, our findings demonstrated a lower preva- ries were surprisingly unrelated in most cases. Given lence of diabetes among the AAS users (2.6%) vs. the limitations of our study, more data are needed to nonusers (6.3%). These values are both lower than further examine the relationship between AAS use general population prevalence (8.1%).40 Because and MSI. To our knowledge, this is the first study in most of those who used AAS had a high BMI, one humans to evaluate such an association. would think that a greater percentage of these subjects would be at higher risk for diabetes, which ACKNOWLEDGMENTS calls into question some of our findings. However, We thank Amy Matthews and the Carolina if we consider that the highest prevalence of type II Center for the Study of Retired Athletes for gath- diabetes is in those of 60 yrs of age or older, we may ering and organizing the database. We also thank not have sufficient number of steroid users from Dr. Michael Lee for his continuous support, Dr. the mid-1970s to mid-1980s who have reached this Vicki Kowlowitz for her invaluable comments on age. This supposition may also help explain some of the manuscript, and Dr. Charles Yesalis for all the our other findings, such as a lower prevalence of work he has done on the subject and assisting on cancer among steroid users and no significant as- ideas when the study was in its infancy. sociation seen between a number of other medical conditions. The data may not have captured many REFERENCES of the medical conditions (such as heart disease, 1. Hartgens F, Kuipers H: Effects of androgenic-ana- bolic steroids in athletes. Sports Med 2004;34: stroke, liver disease, peripheral vascular disease, 513–54 and renal disease) that occur later in life. In addition to the aforementioned limitation, 2. Millmen RB, Ross EJ: Steroids and nutritional sup- plement use in athletes. Am J Addict 2003;12:48–54 our study is also limited by the use of retrospective self-report data. The study is limited by the uncer- 3. Mcduff DR, Baron David: Substance use in athletics: tainty of how well the retired players recalled their A perspective. Clin Sports Med 2005;24:885–97 diagnosed injuries during their careers, as well as their willingness to report steroid use. However, 4. Monitoring the Future Study: Trends in Prevalence given that we were asking for “diagnosed injuries of Various Drugs for 8th-Graders, 10th-Graders, and 12th-Graders. Bethesda, MD, National Institute on by a physician or health professional,” many of Drug Abuse, 2006 which required medical attention including surgi- 5. NCAA Drug Testing 2004–2005: NCAA Committee cal or therapeutic intervention or both, we believe on Competitive Safeguards and Medical Aspects of that the findings are valid. In an attempt to help Sports. December 2006. Available at: http://www.ncaa. validate the self-report data, we asked a sample of org/wps/ncaa?ContentIDϭ485 83 retired players to complete the identical general 6. NCAA study of substance abuse habits of college health questionnaire a second time between 36 and student-athletes. June 2001. Available at: http:// 48 mos after completing the original question- www1.ncaa.org/membership/ed_outreach/health- naire. Intraclass correlation analysis of select vari- safety/drug_ed_progs/2005/DrugStudy2005_ ables revealed a relatively high consistency (r ϭ ExecutiveSummary.pdf 0.84–0.92) when responses were compared between 7. Fair JD: Muscletown USA: Bob Hoffman and the the two questionnaires. This lead us to believe that Manly Culture of York Barbell. University Park, PA, the self-reported items were likely accurate. Pennsylvania State University Press, 1999

March 2009 Steroid Use and Musculoskeletal Injuries 199 8. Johnson W: Steroids: A Problem of Huge Dimen- 25. Bach BR, Warren RF, Wickiewicz TL: Triceps sions. Sports Illustrated, May 13, 1985. Available at: rupture: A case report and literature review. Am J http://vault.sportsillustrated.cnn.com/vault/article/ Sports Med 1987;15:285–89 magazine/MAG1119448/index.htm 26. David HG, Green JT, Grant AJ, et al: Simultaneous 9. Courson S: False Glory. Stamford, CT, Longmeadow bilateral quadriceps rupture: A complication of ana- Press, 1991 bolic steroid abuse. J Bone Joint Surg Am 1995;77: 10. Hoffman JR, Ratamess NA: Medical issues associated 159–60 with anabolic steroid use: Are they exaggerated? 27. Herrick RT, Herrick S: Ruptured triceps in power- J Sports Sci Med 2006;5:182–93 lifter presenting as cubital tunnel syndrome: A case 11. Lasseter JT, Russell JA: Anabolic steroid induced report. Am J Sports Med 1987;15:514–16 tendon pathology: A review of the literature. Med Sci 28. Hill JA, Suker JR, Sachs K, et al: The athletic poly- Sports Excerc 1991;23:1–3 drug abuse phenomenon: A case report. Am J Sports Med 1983;11:269–71 12. Michna H: Tendon injuries induced by exercise and anabolic steroids in experimental mice. Int Orthop 29. Kramhoft M, Solgaard S: Spontaneous rupture of 1987;11:157–62 the extensor pollicis longus tendon after anabolic steroids. J Hand Surg ͓Br͔ 1986;11:87 13. Wood TO, Cooke PH, Goodship AE: The effect of exercise and anabolic steroids on the mechanical 30. Liow RY, Tavares S: Bilateral rupture of the quadri- properties and crimp morphology of the rat tendon. ceps tendon associated with anabolic steroids. Br J Am J Sports Med 1988;16:153–8 Sports Med 1995;29:77–9 14. Miles JW, Grana WA, Egle D, et al: The effect of 31. Visuri T, Lindholm H: Bilateral distal biceps tendon anabolic steroids on the biomechanical and histolog- avulsions with use of anabolic steroids. Med Sci ical properties of rat tendons. J Bone Joint Surg Am Sports Exerc 1994;26:941–44 1992;74:411–22 32. Michelsen CV, Askanazi J, Kinney JM, et al: Effect of 15. Pellman EJ, Viano DC, Tucker AM, et al: Concussion an anabolic steroid on nitrogen balance and amino in professional football: Reconstruction of game im- acid patterns after total hip replacement. J Trauma pacts and injuries. Neurosurgery 2003;53:799–814 1982;22:410–13 16. Pellman EJ, Powell JW, Viano DC, et al: Concussion 33. Loughton SJ, Ruhling RO: Human strength and in professional football: Epidemiological features of endurance responses to anabolic steroid and train- game injuries and review of the literature—Part 3. ing. J Sports Med Phys Fitness 1977;17:285–96 Neurosurgery 2004;54:81–96 34. Friedl KE, Dettori JR, Hannan CJ, et al: Comparison of the effects of high dose and 19- 17. Guskiewicz KM, Marshall SW, Bailes J, et al: Associ- nortestoserone to a replacement dose of testosterone ation between recurrent concussion and late-life on strength and body composition in normal men. cognitive impairment in retired professional football J Steroid Biochem Mol Biol 1991;40:607–12 players. Neurosurgery 2005;57:719–26 35. O’Shea JP: The effect of anabolic steroid on dynamic 18. Nicholas SJ, Nicholas JA, Nicholas C, et al: The strength levels of weighlifters. Nutr Rep Int 1971;4: health status of retired american football players. 363–70 Super bowl III revisited. Am J Sports Med 2007;35: 1674–9 36. Hervey GR, Knibbs AV, Burkinshaw L, et al: Effects of methandienone on the performance and body 19. Cohen SB, Towers JD, Bradley JB: Rotator cuff con- composition of men undergoing athletic training. tusions of the shoulder in professional football play- Clin Sci (Lond) 1981;60:458–61 ers. Epidemiology and magnetic resonance imaging findings. Am J Sports Med 2007;35:442–7 37. Hasin DS, Goodwin RD, Stinson FS, et al: Epidemi- ology of major depressive disorder. Results from the 20. Albright JP, Powell JW, Martindale A, et al: Injury National Epidemiologic Survey on and patterns in big ten conference football. Am J Sports Related Conditions. Arch Gen Psychiatry 2005;62: Med 2004;32:1394–404 1097–106 21. Brophy RH, Barnes R, Rodeo SA, et al: Prevalence of 38. Brower KJ, Blow FC, Young JP, et al: Symptoms and musculoskeletal disorders at the NFL combine— correlates of anabolic-androgenic steroid depen- Trends from 1987–2000. Med Sci Sports Exerc 2007; dence. Br J Addict 1991;86:759–68 39:22–7 39. Young J, Anwar A: Strong diabetes. Br J Sports Med 22. Stewart AL, Hays RD, Ware JE: The MOS short-form 2007;41:335–6 general health survey: Reliability and validity in a patient population. Med Care 1988;26:724–35 40. Behavioral Risk Factor Surveillance System, 2007; analysis by the National Center for Chronic Disease 23. SPSS for Windows, Rel. 16.0. 2007. Chicago, SPSS Prevention and Health Promotion, Division of Nu- Inc trition and Physical Activity, Centers for Disease 24. Evans NA, Bowrey DJ, Newman GR: Ultrastructural Control and Prevention. Available at: http://apps. analysis of ruptured tendon from anabolic steroid nccd.cdc.gov/brfss/list.asp?catϭDB&yrϭ2007&qkeyϭ users. Injury 1998;10:769–73 1363&stateϭAll

200 Horn et al. Am. J. Phys. Med. Rehabil. ● Vol. 88, No. 3