Prevalence of total hip and knee arthroplasty in former National Football League players: comparison with the general US population and other populations of professional athletes

Abstract

Objectives: To evaluate the prevalence of total hip arthroplasty (THA) and total knee arthroplasty (TKA) in a population of former National Football League (NFL) players. Methods: Participants were 3913 former NFL players (participation in years 1960–2019) who completed either an online or mailed survey that included self-reported TKA and THA, year(s) of surgery and date of birth. The prevalence of TKA and THA was reported by age category and compared to published cohorts of athlete populations and general population of non-athletes in the USA. Results: 12.3% and 8.1% of sample reported TKA and THA, respectively. The prevalence of both TKA and THA was higher in former NFL players compared to US non-athletes across all ages. Prevalence of TKA was not statistically higher than in other former athlete cohorts but performed at younger ages. The prevalence of TKA and THA was higher than in other cohorts of former NFL players. Conclusion: Former NFL players had higher prevalence of arthroplasty than the general population, suggesting prior participation in American-style football may contribute to elevated risk for arthroplasty at younger ages. Understanding risk factors in style of play, lifestyle and other contributors is important to improve joint health of this population.

INTRODUCTION

There has been a rise in the prevalence of total hip arthroplasty (THA) and total knee arthroplasty (TKA) in the USA over the past 30 years, particularly in younger populations.¹ Studies have examined the prevalence of THA and TKA in former professional athletes including soccer players,² rugby players³ and other athletes in impact sports.⁴ To date, two studies have reported on THA and TKA in former National Football League (NFL) players.^{5 6} In one study of players who participated in Super Bowl III in 1969 (age 62±3y, n=36), two-thirds reported arthritis, seven had TKA and one had THA.⁶ Another study examining data from former players in years 1929–2001 (n=2432) reported age, joint injuries, weight and playing position as risk factors associated with increased risk of TKA and THA, with reported prevalence of 7.7% and 4.6%, respectively.⁵ While both studies provide a historical understanding of the risk of TKA and THA in this population, neither provides recent prevalence in former players, nor compares prevalence rates across age to the general population and other professional athletes. Further studies suggest that high body mass index (BMI),^{7 8} weight gain⁹ and intensive physical activity at work⁷ are risk factors for TKA and THA. Given that former NFL players have some of the highest BMIs among athletes, experience weight gain throughout their careers¹⁰ and regularly engage in intensive physical activity, we hypothesised this population may be at increased risk for total joint arthroplasties, higher than previously reported. To test this hypothesis, we evaluated the prevalence of arthroplasty in former NFL players by age and compared the rate of arthroplasty in former NFL players to that in the general US population and other athlete cohorts.

METHODS

The Football Players Health Study (FPHS) at Harvard University is an ongoing study evaluating health outcomes in former American-style football (ASF) players. The methods of recruitment for this cohort and verification of eligibility to participate have been noted previously.¹⁰ Living former ASF players were identified using the NFL Player Association and online public source Pro Football Reference who participated during years 1960 to present. Former players with verified contact information were each sent a 76-question survey including items regarding their health status. This included report of prior TKA and THA queried as follows: ‘Since leaving active professional football have you had any of the following surgical procedures: (knee joint replacement) or (hip joint replacement)?’ Participants who answered ‘Yes’ also reported approximate year(s) of surgery. Responses were collected and tabulated using REDCap (Vanderbilt, Nashville, Tenn)^{11 12} for online surveys or Scantron (Scantron Corporation, Tustin, CA, USA) for paper surveys. For participants with history of multiple joint arthroplasties (including the same joint or bilateral joint replacement), the date of initial surgery was used as the index measure to determine the age at the time of arthroplasty. Participants reporting history of both knee and hip arthroplasty were also reported using the date for first surgery as index case for each arthroplasty and were reported in prevalence of hip and knee arthroplasty in the total cohort. Statistical analyses were conducted with R (R Core Team, 2018).

χ ² tests were used to identify differences in arthroplasty prevalence between the FPHS cohort and the general population and other athlete populations, using published values in other cohorts. Date of birth and year of arthroplasty were used to determine approximate age at the time of arthroplasty. The study was approved by the institutional review board (IRB) at Harvard T.H. Chan School of Public Health (Protocol:18-1365) and Beth Israel Deaconess Medical Center (2014P000308). Waiver of documentation of consent was obtained from the reviewing IRBs. All participants received a copy of the consent and participation was optional.

RESULTS

Of 15 643 eligible former players, 3913 agreed to participate (average age 52.4 years). TKA and THA were reported in 12.3% and 8.1% of participants, respectively (tables 1 and 2). The prevalence of arthroplasty increased across age and was higher for TKA compared to THA (figure 1). The prevalence of TKA was higher than THA for all age groups above 60 years.

Table 1.

Prevalence of knee and hip joint replacement by age at the time of the survey in FPHS cohort

	Knee joint replacement (89 responses missing)		Hip joint replacement (114 responses missing)
Age category (years)	Yes	No	Yes	No
<50	24 (1.48%)	1596 (98.52%)	27 (1.68%)	1584 (98.32%)
50–59	69 (7.60%)	839 (92.40%)	68 (7.53%)	835 (92.47%)
60–69	180 (23.14%)	598 (76.96%)	104 (13.54%)	664 (83.85%)
70–79	170 (36.96%)	290 (63.04%)	90 (20.85%)	367 (79.15%)
80+	28 (46.67%)	32 (53.33%)	18 (29.51%)	43 (70.49%)
Total	471 (12.31%)	3355 (87.71%)	307 (8.08%)	3492 (91.92%)

FPHS, Football Players Health Study.

Table 2.

Prevalence of knee and hip joint replacement by reported age at the time of surgery in FPHS cohort

Age category (years)	TKA	THA
<40	24 (5.10%)	19 (6.19%)
40–49	55 (11.68%)	42 (13.68%)
50–59	135 (28.66%)	90 (29.32%)
60–69	147 (31.21%)	90 (29.32%)
70–79	56 (11.89%)	27 (8.79%)
80+	1 (0.21%)	3 (0.98%)
Total	418 (53, or 11.25%, missing year of TKA)	271 (36, or 11.73%, missing year of THA)

% Reflects the percentage of joint replacement that occurred in that age category.

FPHS, Football Players Health Study; THA, total hip arthroplasty; TKA, total knee arthroplasty.

Figure 1.

Prevalence of total knee arthroplasty and total hip arthroplasty was reported by age category of former NFL Players with the Football Players Health Study (FPHS) cohort and general US population. Knee arthroplasty was more common that hip arthroplasty for both populations. Arthroplasty was more common for FPHS cohort than general US population across all ages. NFL, National Football League.

When compared to the prevalence of arthroplasty in the general population,¹ the prevalence in the FPHS cohort was significantly higher in each age category (table 3). The prevalence of THA reported in FPHS cohort age 50–59 years exceeded the prevalence for nearly every age category in the general population, including the 90+ category.¹

Table 3.

Prevalence of knee and hip joint replacement in age-comparable males in the general population and in the FPHS cohort

	Kremers, 2015 (general population)		FPHS cohort N=3,913
	n	%	n	%	P value
Knee replacement
<50
Yes	85 195	0.08	24	1.48	<0.005
Total	105 709 894		1620
50–59
Yes	245 935	1.20	69	7.60	<0.005
Total	20 512 532		908
60–69
Yes	510 244	3.64	180	23.14	<0.005
Total	14 029 131		778
70–79
Yes	541 038	7.26	170	36.96	<0.005
Total	7 447 964		460
	Kremers, 2015 (general population)		FPHS cohort N=3,913
	n	%	n	%	P value
Hip replacement
<50
Yes	122 401	0.08	27	1.68	<0.005
Total	105 709 894		1611
50–59
Yes	205 995	1.00	68	7.52	<0.005
Total	20 512 532		904
60–69
Yes	295 869	2.11	104	13.54	<0.005
Total	14 029 131		768
70–79
Yes	280 383	3.76	90	19.69	<0.005
Total	7 447 964		457

FPHS, Football Players Health Study.

Former players in the FPHS cohort were, on average, younger than other professional athlete populations in which joint arthroplasty has been studied, including another cohort of former NFL players (all p<0.05, figure 2). Despite being younger, the overall prevalence of TKA for former NFL players (12.3%) was non-statistically higher than 9% reported in rugby players³ and 11.1% reported in soccer players² (p=0.1235 and 0.1794, respectively). The FPHS cohort also had a higher prevalence of both reported knee (12.3% vs 7.7%, p<0.001) and hip replacement as compared to the Davies cohort of former ASF players (8.1% vs 4.6%, p<0.001).⁵

Figure 2.

Total knee arthroplasty was performed at younger ages within the Football Players Health Study (FPHS) compared to studies in other athlete populations and a separate report in former NFL players. Overall prevalence data were non-statistically different between groups. NFL, National Football League.

DISCUSSION

Our results demonstrate that former NFL players have an overall prevalence of 12.3% and 8.1% for TKA and THA, respectively. This study adds to the limited reports of arthroplasty in former NFL players and is the largest cohort report to date. Furthermore, this study expands upon findings of Davies et al ⁵ with rates of TKA and THA higher in the current study population than previously observed in former NFL players participating in a more distant era from 1929 to 2001.⁵

Differences in TKA between sports

The prevalence of TKA was similar to other non-ASF athlete populations but performed at a younger age. Compared to the general population, former NFL players had higher rates of TKA and THA across all ages. The mechanism for higher rate of arthroplasty in former NFL players is unknown; we hypothesise that the combination of traumatic joint injuries, higher playing weight and other factors may contribute. Previous studies suggest that the combined effects of physical activity at work,⁷ BMI,^{7 8} joint injuries associated with osteoarthritis⁵ and weight gain are risk factors for arthroplasty.⁹

Future directions

The current report expands on earlier findings from an ongoing study evaluating the largest cohort of former NFL players, the FPHS at Harvard University. An earlier report identified early-life weight gain during the time of high school to college was independently associated with the development of cardiometabolic disease, sleep apnoea and chronic pain.¹⁰ When separated by field position, linemen gained greater weight than non-linemen and were also more likely to be classified as having chronic pain.¹⁰ The current findings from this report on high rate of arthroplasty, along with prevalence of potential risk factors from other populations, illustrate the need for translational strategies to address the high and premature rate of arthritis and joint disease. Treatment should address modifiable risk factors in this cohort and include lifestyle factors of diet and exercise to lower BMI. However, the interconnectivity of other disease states suggests a model of care addressing obstructive sleep apnoea,¹³ metabolic syndrome,¹⁴ along with mood and neurocognitive issues, which may influence pain experience and thresholds. Addressing inflammation burden may be important given the known association to disease states including metabolic syndrome,¹⁴ obstructive sleep apnoea¹⁵ and neuroinflammation that has been postulated to contribute to widespread pain.¹⁶ In addition, new treatments that address the underlying inflammatory joint catabolic pathways may be targets for early intervention with new injectable medications and orthobiological agents.^17–19

Limitations

Limitations include the possibility of selection bias. It is not possible to know whether players who underwent arthroplasty were more or less likely to enrol in this study. The self-reported data in our study may overestimate the true prevalence of TKA and THA. However, there appears to be good agreement between self-reported data and hospital data in TKA and THA.⁹ Our cross-sectional design and comparison to former athlete populations precludes determination of explanatory mechanisms specific to sport, including forces sustained during sport, differences in pain threshold and documented weight gains through sports participation. The study design limits accounting precisely for determinants of health that influence the outcome of having an arthroplasty, both in the general US population and within our former NFL population studied. Despite these limitations, this study represents the largest study to date to report prevalence rates for TKA and THA in former NFL players.

Conclusion

Our study identified higher rates of THA and TKA within former NFL players than previously reported in a contemporary cohort. Although the rate of TKA was not statistically different compared to former rugby and soccer players, the younger age of our cohort suggests an earlier burden of advanced joint disease and symptomatic joint pain in former NFL players. The high rate of arthroplasty in this population requires further investigation to identify methods for injury prevention and to optimise treatment of arthritis.

What are the new findings.

• The prevalence of total knee arthroplasty and total hip arthroplasty within the cohort of former National League Football (NFL) players was 12.3% and 8.1%, respectively.

• Findings suggest former NFL players have total knee arthroplasty at younger average age but at similar rates to other sport populations of soccer and rugby.

• The elevated rate of total hip arthroplasty and total knee arthroplasty of former NFL players compared to the general US population across ages suggests a sport-specific style of play, athlete characteristics, health co-morbidities and other lifestyle factors need to be further explored.

• The young age of this cohort suggests further translational research is needed to understand methods to optimise non-surgical management of knee and hip injuries.