Journal Of Physical Activity and Health, 2009, 6, 638-643 2009 Human Kinetics, Inc.

Early-Onset Arthritis in Retired Nationai Footbaii League Piayers

Yvonne M. Golightly, Stephen W. Marshall, Leigh F. Callahan, and Kevin Guskiewicz

Background: Injury has been identified as a potential risk factor for osteoarthritis. However, no previous study has addressed playingcareer injuries and subsequent osteoarthritis in a large sample of former athletes. The purpose of this study was to describe the prevalence and determinants of arthritis and osteoarthritis in retired professional football players. A/cAods: Self-reported arthritis prevalence and retrospectively-recalled injury history were examined in a cross-sectional survey of 2,538 retired football players. Results: Football players reported a high incidence of injury from their professional playing days (52.8% reported knee injuries, 74.1% reported ligament/tendon injuries, and 14.2% reported anterior cruciate ligament tears). For those under 60 years, 40.6% of retired NFL players reported arthritis, compared with 11.7% of U.S. males (prevalence ratio =3.5, 95%CI: 3.3 to 3.7). Within the retired NFL player cohort, osteoarthritis was more prevalent in those with a history of knee injury (prevalence ratio = 1.7,95%CI: 1.5 to 1.9) and ligament/tendon injury (prevalence ratio = 1.6. 95%CI: 1.4 to 1.9). Conclusions: In males under the age of 60, arthritis is over 3 times more prevalent in retired NFL players than in the general U.S. population. This excess of early-onset arthritis may be due to the high incidence of injury in football. Keywords: osteoarthritis, sports injuries, knee injury Golightly and Marshall are with the Dept of Epidemiology. University of North Carolina at Chape! Hill. Callahan is with the Thurston Arthritis Research Center, University of North Carolina at Chapel Hill. Guskiewicz is with the Dept of Exercise and Sport Science, University of North Carolina at Chapel Hill.

Arthritis is one of the most prevalent chronic health problems in the United States.'' Arthritis comprises more than 100 medical conditions that affect the joints of 46 million American adults (21% of U.S. adult population).'* Osteoarthritis (OA) is the most common form of arthritis in Americans (approximately 27 tnillion). followed by fibromyalgia (5.0 million) and rheumatoid arthritis (1.3 million).'- As a leading cause of disability among adults, arthritis has a considerable impact on mobility, basic daily activities, employment, and quality of life.^"'' In addition, the burden of poorer physical and mental health appears to be associated with arthritis.^ Identifying populations at risk for arthritis is important for developing and providing optimal treatment and prevention strategies. Injury, particularly to the joints, has been identified as a potential risk factor for OA,"^ Annually, sports contribute to 7 million injuries in Americans,'" and previous reports have suggested that the development of OA may be more prevalent and occur prematurely in individuals who regularly participate in sports." The incidence of injury is higher in football than in any other sport of comparable popularity.'^" Previous literature examining the association between injuries and OA in football players is very sparse and based primarily on observations of small groups.'"*'^ The purpose of this study was to examine the prevalence of self-reported physician-diagnosed arthritis, particularly OA, in retired NFL players and to propose potential explanations (ie, primary playing position, years of playing professional football, and other health conditions) where an increased prevalence of OA is apparent. The prevalence of arthritis in the NFL retirees was compared with the prevalence in the general U.S. male population (using the National Health Interview Survey) and the relationship between injury and OA within the retired NFL players cohort was examined.

638

Arthritis in Football Players

639

Materials and Methods

study Participants
AU living retired professional football players registered with the National Football League Retired Player's Association were eligible to participate in this study. Eligible retired players included pre-World War II professional football players through to recent retirees. Participants completed a general heath survey. By completing and submitting the survey, participants agreed to participate in this study. The study protocol was approved by the Institutional Review Board at The University of North Carolina at Chapel Hill.

ciated with OA: age at time of survey, body mass index (BMl) at time of survey, change in BMI from professional football to time of survey, years of playing professional football, primary playing position, current comorbid conditions (hypertension, cardiovascular disease, diabetes, and depression). Playing position was grouped into 3 categories based on typical height, weight, and type of activities performed during sport participation: 1) lineman (offensive line, defensive line, and tight end), 2) linebacker and running back, and 3) wide receiver, comerback/safety, quarterback, and special teams. We also compared SF-36 scores on the physical and mental health components in those with and without OA.

Heaith Survey of Retired NFL Players

The general health survey included detailed questions about musculoskeletal injuries that the retired player experienced during his football career and his postretirement health conditions, including physician-diagnosed arthritic conditions. Questions on demographics, playing history, current diet, and current health status [Short Form-36 (SF-36) Health Survey] were included. The SF-36 is a 36-question quality of life survey that consists of physical function and mental health scales."" A higher score on a scale indicates a better health status. A total of 3,647 individuals were initially maiied the questionnaire in May 2001. The survey was maiied again to nonrespondents in August 2001 and February 2iX)2. After these mailings, nonrespondents were contacted by telephone, and consenting players completed the survey over the telephone. The questionnaire was readministered to 25 of the original respondents 18 to 24 months after the initial questionnaire to determine reliability of this instrument.

2001 National Health Interview Survey (NHIS)

The logical comparison group for the NFL retirees is the U.S. male population, matched on age. We obtained data on this population from the NHIS, a continuing, nationwide survey of approximately 40,000 households that provides health and demographic data on approximately 100,000 persons in the civilian noninstitutionalized population of the United States.'^ For our analyses, arthritis was defined as a "yes" response to the question, "Have you ever been told by a doctor or other health professional that you had arthritis?" Unfortunately, the NHIS questions do not distinguish between OA and degenerative arthritis.

Statistical Anaiysis
T-tests and chi-square tests were used to compare age, BML change in BMI from time of play to time of survey, years of professional football, primary professional playing position, cardiovascular disease, hypertension, diabetes, depression, and SF-36 physical function and mental health scores between groups with and without OA. Descriptive analyses were conducted to determine the prevalence of arthritis (OA, rheumatoid arthritis, and fibromyalgia) and OA, stratified by age. The age-specific prevalence of self-reported arthritis in the general U.S. male population was determined from the 2001 NHIS data, using sampling weights and SAS 9.1 (SAS Institute, Cary, NC) to account for the sampling design. Associations between presence of OA and types of injury were determined using binary regression models to estimate prevalence ratios, adjusting for age and BMI at the time of the survey.

Arthritis and Injuries

For analysis purposes, arthritis was defined on the Health Survey of Retired NFL Players as a "yes" response to 4 questions asking if the retiree has ever been told by a physician or health professional that he has or had OA, degenerative arthritis, rheumatoid arthritis, or fibromyalgia. Since OA is a degenerative disease of joint cartilage, it was defined as a "yes" response to either the question about OA or degenerative arthritis. Retired NFL players also provided a detailed history of injuries during their professional football career through an 18-item checklist of the most common and/or clinically significant football injuries. For analyses, the injury history data were grouped into the following categories: knee, arm. ankle, ligament, tendon, joint dislocations, fractures, and disc ruptures/hem i ati ons.

Results
The overall response rate was 70% (2,538 players), with complete data for arthritis and OA from 2,529 and 2,523 respondents, respectively. The retired players had a mean age of 53.8 years (range 24 to 95 years), a mean BMI of 30.5 (range 14.8 to 56.2), played an average of 6.6 years (range 1 to 26) of professional football.

Participant Characteristics
We examined the following participant characteristics as covariates in our analyses because they may be asso-

640

Golightly et al

Compared with players without OA, players with OA were older, had higher BMIs, and had poorer health status on the physical and mental health components of the SF-36 (Tables I and 2). Compared with linebacker/ running back playing positions, linemen were associated with a higher prevalence of OA. Wide receiver, comerback/safety, quarterback, and special teams positions were less likely to report OA (Table 1). Players with OA were more likely to report cardiovascular disease, hypertension, diabetes, and depression than those without OA (Table 2). Overall, 43.4% of retirees players reported a physician diagnosis of any type of arthritis and 37.9% refwrted a physician diagnosis of OA (Table 3), Arthritis was far more common in the retirees compared with the general U.S males (43.4% vs. 19.5%; Figure 1). This disparity was greatest in males under the age of 60 years: 40.6% of retirees and 11.7% U.S. males reported arthritis in this age group. On the other hand, in males 60 years and older, 49.2% of retirees and 41.9% U.S. males reported arthritis. The arthritis prevalence ratio was 2.2 (95%CI: 2,1, 2.3) across all ages groups and 2.5 (95%CI: 3.3, 3.7) in those under 60 years of age. The prevalence ratio was even higher in the younger age groups: 4.5 (95%CI: 4.1, 4.9) in those under 50 years and 6.5 (95%CI: 5.7, 7.5) in those under 40 years. The retired NFL players reported a high prevalence of injury during their professional playing days. The types of injuries reported included: any knee injuries (52.8%), anterior cruciate ligament tears (ACL; 14.2%), meniscus tears (18.4%), any arm injuries (36.1%), any ankle injuries (24.3%), any ligament/tendon injuries (74,1%), any ligament injuries (35.6%), any tendon injuries (32.8%), joint dislocations (35.7%), fracture (0.8%), and disc ruptures/herniations ( 10.9%). In models adjusting for age and BMl at time of survey, individuals with history of injury (specifically, ACL and meniscus tears, shoulder, elbow, or wrist injury, any ligament injury, any joint dislocation, any fracture, and any disc rupture/hem iation) all had an elevated prevalence of OA (Table 4). The results of unadjusted analyses (not shown) were very similar to the adjusted models.

Discussion
In this study, the prevalence of arthritis was higher in retired NFL players than in the general US male population. However, the excess of arthritis was concentrated in those under the age of 60 years, and was far less apparent in older age categories. This suggests that the onset of arthritis is accelerated, but not increased over the whole lifespan, in the NFL retiree population. Longevity of play did not appear to have a large influence on whether OA was developed, as shown in Table I. The high incidence and severity of joint injury in football likely accelerates development of degenerative arthritis or OA, creating an excess of early-onset arthritis in NFL

retirees. This hypothesis is supported by the fact that we observed a strong association between injury history and OA, regardless of type or location of injury. Tbe existing literature on sports-related injury and OA supports this potential association. A small study of American football players examined 20 years after high school competition demonstrated an increased prevalence of knee OA in players who had sustained a knee injury.'^ In a study of young Swedish female soccer players who sustained anterior cruciate ligament injury, 51% had radiographie knee OA after 12 years.'^ Furthermore, il has been argued that repair or reconstruction of cruciate ligaments does little to delay the onset of OA following knee injury.'** In addition, medial meniscectomy may be associated with knee OA.^" Future OA research of football players should attempt to examine both injury and surgery history. The transition from an active playing career to retirement is challenging for any professional athlete, and perhaps even more so in a high-profile sport like football. Physical activity was not measured longitudinally in our data, and thus, the relationship between decreased physical activity and accelerated development of OA and comorbidities in this cohort is unknown. Potentially, decreased physical activity during retirement may hasten the development of OA and comorbidities. Alternatively, the development of physical limitations due to OA may vastly complicate an already difficult transition and may initiate a cascade of" health problems. The frequency of comorbidities has been suggested to be higher among individuals with rheumatoid arthritis and OA than those without arthritis.^ Our data also show that OA is associated with a range of comorbidities that are associated with lack of physical activity, including overweight/obesity, hypertension, cardiovascular disease, and diabetes. Tbe onset of OA may result in physical activity limitations that lead to the development of these comorbidities. The forced inactivity combined with tbe pain associated witb the OA may explain the increased risk for depression in this group. The pattern of a medical condition leading to the development of other associated conditions may be described as a "snowball effect", which is difficult to stop once it begins. Thus, injuries during the playing career of NFL players shape not only their playing career, but also their health in retirement. Injury history, in a global rather than joint-specific sense, during a playing career is strongly associated with OA in retirement. When stratifying by playing position, offensive and defensive linemen reported the greatest prevalence of OA. in addition, these players are more likely to have the highest BMI (during playing career and retirement) and highest risk for cardiovascular disease. The significance of the finding that injury leads to early-onset of arthritis is not limited to professional athletes. Although NFL players are Subject to particularly high levels of musculoskeletal stress, we expect that a similar association between sports injuries and OA holds true in tbe general population. The NFL cohort is a research population in which

Table 1

Selected Characteristics of NFL Retirees and Association with OA Total sample (N = 2538) With 0A (n = 960) 55.0 13.0 31.1 4.4 4.7 12.9 6.8 3.5 Without OA" (n = 1578) 53.1 13.6 30.1 4.1 4.0 10.9 6.5 3.6 Prevalence ratio 1.03 1.15 1.05 1.06 95% CI 1.01-1.05 1.09-1.22 0.88-1.25 0.99-1.14 P-value <0.0I <0.01 0.61 0.07

Age at time of survey m years (mean SD) BMI at time of survey (mean kg/ m^SD) Change in BMl'' (%) Years of professional football (mean

53.8 13.4 30.5 4.3 4.2 11.7 6.6 3.6

SD))
Primary professional playing position Lineman Linebacker/running back Wide receiver, cornerback/safety, quarterback, and special teams

41.0% 25.5% 28.5%

47.7% 24.7% 23.7%

36.9% 26.0% 31.5%

1.10 1.00 0.86

1.I0-I.I7 (reference) 0.74-0.99

<0,01 0.04

Report of physician-diagnosed OA or degenerative anhritis. (BMl at time of survey - playing BMI) / playing BMI.

Table 2 Comorbid Conditions Associated with OA, and Gnerai Heaith Status, in NFL Retirees
Totai sample (N = 2538) Cardiovascular disease (%) Hypertension (%) Diabetes (%) Depression (%) SF-36: Physical function (mean SDl SF-36: Mental health (mean SD) 15.2 29.7 6.2 11.4 With OA (n = 960) 19.4 35.0 Without OA" (n = 1578) 12.6 26.5 5.1 8.4 53.0 8.5 49.4 9.1 P-value <0.01 <0.01 <0.01 <0.0]
<0.01 <O.OI

8.0
16.4 50.0 9.9 4O.8I1.I

51.99.2
46.1 10.7

Report of physician-diagnosed OA or degenerative anhritis.

Table 3 Age-Specific Prevalence of Seif-Reported Arthritis in NFL Retirees and Maies in National Heaith interview Survey (NHIS)
Prevalence of arthritis in US males (NHIS 2001) Age category (years) 24-34 35-39 40-44 45^9 50-54 55-59 60-64 65-69 70-74 75-H Total n/N % (unweighted) (weighted)
110/2893 128/1555 182/1604 222/1388 308/1291 287/1020 274/497 268/657 268/619 472/1013 2533/13110

Prevalence of arthritis in retired NFL players (2001)

Arthritis prevalence ratio Prevalence ratio 6.3 3.8 3.1 2.4 1.9 1.6 1.3 1.3 1.1 1.1 2.2

Prevalence of OA in retired NFL players (2001)

%
29.5 34.2 38.1 40.9 47.0 46.1 47.7 50.5 49.7 49.5 43.4

n/N
51/173 91/266 102/268 113/276 157/334 172/373 133/279 100/198 90/181 90/182 1099/2530

95% CI 5.0-7.9 3.2^.5 2.7-3.7 2.1-2.8 1.7-2.2 1.4-1.8 1.2-1.5 1.1-1.4 1.0-1.3 0.9-1.2 2.1-2.3

n/N
47/173 83/265 99/268 98/275 140/334 148/371 109/279 85/197 77/181 70/181 956/2524

%
27.2 31.3 37.0 35.6 41.9 39.9 39.1 43.1 42.5 38.7 37.9

4.7 8.9 12.1 16.9 24.1

29.4 36.0 40.1 43.8 46.9 19.5

641

642

Golightly et al

60%
50% -

s
Q.

g 40% H

5 30% H

U.S. Maies

<24

24-34

35-39

40-44

45-49

50-54

55-59

60-64

65-69

70-74

75+

Age (in years)

Data on U.S. Males is from the National Health Interview Survey, United States, 2001

Figure 1 Age-specific prevalence of arthritis (x-axis not to scale)

Table 4 Associations Between OA and Type of Injury in NFL Retirees^

OA No OA 44.9% 11.0% 15.8% 33.4% 23.3% 69.3% 30.0% 32.2% 32.3% 0.4% 8.7%

= 949)
Any knee injury Anterior cruciate ligament tear Meniscus tear Any arm injury Any ankle injury Any ligament or tendon injury Any ligament injury Any joint dislocation Any tendon injury Any fracture Disc rupture/hemiation
' Report history of ai least one occurrence of specified injury. ^ Adjusting for age and BMI at the time of the survey.

Prevalence ratio'' 1.72 1.45 1.29 1.21 1.08 1.61 1.47 1.28 1.06 1.75 1.39

95% CI 1.541.92 1.30-1.62 1.151.44 1.101.34 0.97-1.21 1.40-1.85 1,33-1.62 1.161.41 0.951.18 1.28-2.40 1.231.57

P-value <0.01 <0.0I <0.01 <0.01 0.17 <0.01 <0.01 <0.01 0.28 <0.01 <0.0I

66.0%
19.5%

22.6% 40.2% 26.1% 82.2% 44.9% 41.4%

33.4%

1.4% 14.3%

to Study the effect of those stresses in the general population. Injury prevention and good rehabilitation after injury may be key elements to maximizing health status over one's lifetime, not just in the NFL, but in the general population, too. This study includes a large, representative sample of retired football players with a wide range of ages and professional football experiences. The response rate to the mailed survey was high at 70%. Several limitations are present in this study. The NHIS does not include specific questions about OA or degenerative arthritis, but had a general question about physician-diagnosed

arthritis. However, OA is the most common form of arthritis. In addition, the Health Survey of Retired NFL Players did not include joint-specific questions about arthritis. Thus, analyses of the association between OA and injury were limited because we were unable to link the site of injury to the site of OA. Compared with the 2001 survey, the 2002 NHIS survey includes a tnore comprehensive self-report arthritis question ("some form of arthritis, rheumatoid arthritis, gout, lupus, or fibromyalgia"), which may be more comparable to question used in the NFL questionnaire. A similar selfreport arthritis question was used in the Behavioral Risk

Arthritis in Football Players

643

Factor Surveillance System with moderate validity compared with rheumatologist assessment of arthritis.^' Despite differences in the phrasing ofthe arthritis questions, the prevalence of self-report arthritis among adult males is very similar between the 2001 and 2002 NHIS data. Thus, we chose to use the 2001 NHIS data since it was the same year that the retired NFL player data were collected. Another limitation of this study is the measure of OA. Self-report of physician-diagnosed OA may not be as valid as radiographie evidence of joint disease. Selfreport of knee or hip joint pain or swelling as a predictor of OA diagnosis has been found to have a sensitivity > 40% and a specificity > 80% in a sample of Australian older twins." Although we did not obtain radiographie or medical record data for this study, we suspect that the retired NFL players may be able to more accurately self-report OA, specifically physician-diagnosed OA, than the general population because they are better aware of their physical condition. In a readministration of our instrument on a subsample of 83 subjects, 3 to 5 years after the original administration of the instrument, we found a kappa of 0.49 (95% CI 0.30 to 0.68) for OA and 0.52 (95% CI 0.34 to 0.70) for overall arthritis. Similar findings have been reported elsewhere in the literature.^^

Conclusion
Arthritis is more prevalent, especially in younger age categories, in retired NFL players than in U.S. males. The type, frequency, and severity of injury likely account for differences in arthritis prevalence between retired football players and the general U.S. male population. Further research of the retired NFL players should examine the relationship of injury type with the specific joint site of OA and whether specific joint OA is associated with the comorbidities of cardiovascular disease, hypertension, and depression.

References
1. Lawrence RC, Felson DT, Helmick CG, et al. Estimates of the prevalence of arthritis and other rheumatic conditions in the United States: Part II. Arthritis Rheum. 2008;58(i):26-35. 2. Helmick CG, Felson DT, Lawrence RC, et al. Estimates of the prevalence of arthritis and other rheumatic conditions in the United States: Part I. Arthritis Rheum. 2OO8;58(l):l5-25. 3. Centers for Disease Control and Prevention. Public health and aging: Projected prevalence of self-reported arthritis or chronic joint symptoms among persons aged >65 years-United States, 2005-2030. MMWR Morh Mortal Wkly Rep. 2003:52:489^91. 4. Lethbridge-Ceiku M, Schiller JS, Bernadel L. Summary health statistics for U.S. adults: National Health Interview Survey, 2002. Vital Health Stat 0. 2()O4Mi222)A-\5l.

5. Centers for Disease Control and Prevention. Health-related quality oflife among adults with arthritis-Behavioral Risk Factor Surveillance System, U states, 1996-1998. MMWR Morb Mortal Wkly Rep. 2000;49:366-369. 6. deBock GH, Kaptein AA, Touw-Otten F, Mulder JD. Health-related quality of life In patients with osteoarthritis in a family practice setting. Arthritis Care Res. 1995;8:88-93. 7. Dominick KL, Ahem FM, Gold CH. Heller DA. Healthrelated quality of life among older adults with arthritis. Health Qual Life Outcomes. 2004;2:5. 8. Gabriel SE, Crowson CS, O'Fallon WM. Comorbidity in aiXhti.JRheuniatol. 1999;26f 11):2475-2479. 9. Davis MA. Epidemiology of osteoarthritis. Clin Geriatr Med. 1988:4:241-255. 10. Conn JM, Annest JL, Gilchrist J. Sports and recreation related injury episodes in the US population. 1997-99. Inj Prev. 2OO3;9:I17-123. 11. Lequesne MG, Dang N, Lane NE. Sport practice and osteoarthritis of the limbs. Osteoarthritis Cartilage. 1997;5:75-86. 12. Centers for Disease Control and Prevention. Sports-related injuries among high school athletesUnited States, 2005-06 school year. MMWR. 2006:55:1037-1040. 13. Hootman JM, Dick R, Agel J. Epidemiology of collegiate injuries for 15 sports: summary and recommendations for injury prevention initiatives. J Athl Train. 2007;42(2):311-319. 14. Adams ID. Osteoarthritis and sport. Clin Rheum Dis. 1976:2:523-541. 15. Moretz JA, Harlan SD, Goodrich J, Walters R. Long-term followup of knee injuries in high school football players. Am J Sports Med. 1984:12:298-300. 16. Ware JE, Sherboume CD. The MOS 36-items short-form health survey (SF-36). I. Conceptual framework and item selection. Med Care. 1992;30:473^83. 17. National Health Interview Survey (NHIS). Available at: http ://ww w.cdc. gov/nchs/about/major/nh i s/q uest_data_ related_l997__forward.htmAccessed February 2, 2007. 18. Lohmander LS, Ostenberg A. Englund M, Roos H. High prevalence of knee osteoarthritis, pain, and functional limitations in female soccer players twelve years after anterior cruciate ligament injury. Arthritis Rheum. 2OO4;50:3145-3 ! 52. 19. Lohmander LS, Roos H. Knee ligament injury, surgery, and osteoarthritis. Truth or consequences? Acta Orthop Scand. 1994;65:605-609. 20. AitSiSelmi T. Fithian D, Nevret P The evolution of osteoarthritis in 103 patients with ACL reconstruction at 17 years follow up. Knee. 2006; 13:353-358. 21. Bombard JM, Powell KE, Martin LM, Helmick CG, Wilson WH. Validity and reliability of self-reported arthritis. Am J Prev Med. 2OO5:28(3):251-258. 22. Kirk KM. Bellamy N, O'Gorman LE. et al. The validity and heritability of self-report osteoarthritis in an Australian older twin sample. Twin Res. 2002:5{2):98-106. 23. Skinner KM, Miller DR, Lincoln E. et al. Concordance between respondent self-reports and medical records for chronic conditions: experience for the Veteran Health Study. JAmbul Care Manage. 20O5;28(2): 102-110.