The risk of knee pain and knee osteoarthritis in professional footballers

Parekh, Sanjay M. (2017) The risk of knee pain and knee osteoarthritis in professional footballers. PhD thesis, University of Nottingham.

[img] PDF (Final Thesis Version) (Thesis - as examined) - Repository staff only until 14 July 2019. Subsequently available to Repository staff only - Requires a PDF viewer such as GSview, Xpdf or Adobe Acrobat Reader
Download (10MB)



Knee osteoarthritis (KOA) is a common complex disorder. Although previously believed to be degenerative, KOA is in fact a regenerative condition, compensating against insults sustained at the joint. However, a failure of this compensatory repair process, especially in the presence of constitutional and local joint factors, increases the risk of KOA and inevitably leads to joint-failure (Dieppe and Lohmander, 2005, Arden and Nevitt, 2006, Sandell, 2012).

Diagnosis of KOA may be made via clinical presentation, imaging, or using clinical algorithms, which may be a combination both in addition to biochemical diagnostic tests (Brandt et al., 2003). Knee pain (KP) is the most common symptom, and in the general population its prevalence is 25% (Peat et al., 2001a). Patients may also experience early morning stiffness of the joint and reduced function. Physician-observed signs include crepitus, restricted movement, and bony and soft tissue swellings (Abhishek and Doherty, 2013).

Although considered the gold standard to diagnose KOA, plain film radiography is not without its limitations (Wick et al., 2012). Clinicians, however, favour radiography because it can easily discern two key features of the condition: joint space narrowing (JSN), a surrogate of cartilage loss, and the formation of osteophytes on the joint margin (Roemer et al., 2014). Assessment of radiographs is most commonly undertaken using Kellgren-Lawrence (KL) grade verbal descriptors (Altman et al., 1986a). The prevalence of radiographic KOA (RKOA) may be higher than KP, but there is a discordance between people reporting symptoms and those with structural change (RKOA) (Peat et al., 2001a, Bedson and Croft, 2008).

A plethora of constitutional risk factors and joint-specific biomechanical factors increase the risk of KOA, including joint injury and occupation (Suri et al., 2012, Silverwood et al., 2015). One such occupation, which has a greater risk of injury are professional footballers (Drawer and Fuller, 2002) and knee injuries account for 17% of all footballing injuries (Ekstrand et al., 2011). Football is one of the most common team sports worldwide, with over 265 million people worldwide play the game (FIFA, 2007a), and of these, 110,000 are male professional footballers (FIFA, 2007b). Although perceived that that footballers are at great risk of long-term consequences such as KOA, due to their high risk of injury, the current evidence supporting this is limited (Kuijt et al., 2012, Tran et al., 2016).

The previous studies observing KOA in footballers are difficult to generalise to the wider football population. This is for a number of reasons, including recruitment of inadequate sample, absence of inappropriate control groups, and differing case definitions, all resulting in a large variation in prevalence of KOA. Thus, there exists a need for a comprehensive study to determine the true prevalence and risk of KOA in retired professional footballers compared to the general population.


(1) To determine the prevalence and risk of KOA (measured as KP, RKOA and requirement for total knee replacement (TKR)) in retired professional footballers compared to the general population;

(2) To determine the specific factors (constitutional, biomechanical and football-specific) that are associated with an increased risk of each of these outcomes (KP, RKOA and TKR) within footballers.


The Nottingham University Hospitals NHS Trust and the Nottingham Research Ethics Committee (Refs 14/EM/0045; 14/EM/0015) approved this study, which was registered on the portal (NCT02098044; NCT02098070).

This study design involved carrying out two cross-sectional studies. The Football Study involved distributing 4775 postal questionnaire surveys to retired professional footballers via multiple sources, including football clubs, their former players’ associations and the Professional Footballers Association (PFA). The Knee Pain and Related Health in the Community Study (KPIC) involved distributing 40,500 postal questionnaires, via 12 general practice surgeries, to both men and women in the East Midlands general population. However, only men formed the control group for this study. The inclusion criteria for both the footballers and control participants was the same: men aged 40 and older.

The questionnaires, developed based on previously literature, were similar to capture detailed information about KP, undergoing a TKR and putative risk factors for KOA, including knee injuries, surgery and alignment. The questionnaires also gathered information regarding demographics, medical and occupational history, general health and current medication.

Following this, footballers and controls who consented had radiographic assessments of both their knees, including weight-bearing semi-flexed posterior-anterior (PA) view using the Rosen template (Rosenberg et al., 1988) and a seated 30° flexion skyline view. A single observer (GSF) scored all the radiographs as a single mixed batch using HIPAX Dicom software. In addition to the KL grades, the Nottingham Line Drawing Atlas (NLDA) was used (Nagaosa et al., 2000) (Wilkinson et al., 2005), which scored composite joint space narrowing (JSN), composite osteophyte, and a combined global score for each knee.

Primary outcomes observed were current KP, RKOA (measured using the NLDA) and TKR. Secondary outcomes observed were ever having KP (chronic), physician-diagnosed KOA, RKOA (measured using KL grades) and radiographic CC. Power calculations determined the sample size for the questionnaire survey and the radiographic survey. Categorical variables presented as frequency and percent and compared using a chi-squared test. Continuous variables presented as mean and standard deviation and compared using a t-test.

The risk of KOA (measured for each outcome independently) in footballers compared to the controls was determined using a generalised linear model (GLM) with a Poisson distribution, and adjusted for known risk factors (including age, body mass index (BMI) and previous knee injury). The specific risk factors within footballers associated with outcomes of KOA (namely KP, RKOA and TKR) were determined using multivariate logistic regression.


1207 footballers (response rate of 25.3%) and 4085 control men responded to the Football and KPIC studies respectively, which was far lower than studies previously conducted in both populations. Following this, 470 footballers and 500 men consented to undergoing radiographic assessment of their knees. For participants who returned the questionnaire (footballers and controls), characteristics were compared between those who underwent a knee radiograph and those who did not. Age and sustaining a knee injury were the main factors significantly difference in both.

Footballers were significantly older (3.9 years) than the controls, but were gender-matched (males-only) and had a similar BMI. Footballers had a significantly greater number of injuries (64.5% v. 23.3%) compared to the controls. They also had significantly more body pain (74.7% v. 69.8%) and therefore took more pain-relief medication (61.9% v. 28.5%). However, footballers suffered from far fewer comorbidities compared to the controls (29.4% v. 45.7%).

Footballers had a far greater prevalence of both primary and secondary outcomes. The prevalence of KP was almost twice as great in footballers (52.2%) compared to the controls (26.9%) and this increased prevalence was regardless of age. The peak prevalence of KP also occurred at least ten years earlier in footballers compared to the controls. Although the prevalence of physician-diagnosed KOA was much lower than the prevalence of KP in footballers (28.3%), it was more than double that of the controls (12.2%). Additionally, footballers (11.1%) had almost three times greater prevalence of TKR compared to the controls.

In terms of radiographic measures, the prevalence of NLDA-defined RKOA was much greater in footballers (64.0%) compared to the controls (35.2%) and this remained the case for KL grade-defined RKOA (28.5% v. 14.1%). A stricter KL grade (KL>3) accounts for the lower prevalence compared to NLDA-defined RKOA. Radiographic CC was also just less than three times more prevalent in footballers (24.3%) compared to the controls (8.8%). Within footballers, RKOA (NLDA) was greater in the right (52.6%) compared to the left (45.3%) knee.

The greater prevalence conferred an increased risk of KOA for all outcomes. Footballers were almost twice as likely to report KP [aRR 1.92, 95% CI 1.78-2.07], more than twice as likely to report RKOA (NLDA-defined) [aRR 2.14, 95% CI 1.87-2.45] and almost three times more likely to report TKR [aRR 2.79, 95% CI 2.42-3.23] compared to the controls (following adjustment for age and BMI). The risk of having a physician diagnosis of KOA was 2.6 times greater [aRR 2.62, 95% CI 2.32-2.96] and the risk of radiographic CC was 3.4 times greater [aRR 3.36, 95% CI 2.41-4.69] in footballers compared to the controls. RKOA measured using KL grades was also slightly greater compared with the NLDA definition [aRR 2.44, 95% CI 1.88-3.15].

The risk factors for KP, RKOA (NLDA-defined) and TKR within footballers were also analysed. Following adjustment for age and BMI, a number of factors were associated with increased risk of each outcome.

The following factors were associated with increased risk of footballers having KP: being overweight [OR 1.55, 95% CI 1.16-2.06] or obese [OR 2.22, 95% CI 1.52-3.24]; having a pattern three digit ratio [OR 1.30, 95% CI 1.02-1.66]; a family history of OA (knee, hip or hand OA); and suffering from gout [OR 2.15, 95% CI 1.48-3.13]. Sustaining a knee injury [OR 4.21, 95% CI 3.27-5.42] and its consequences, surgical intervention [OR 4.37, 95% CI 3.39-5.64] and receiving an intra-articular injection [OR 2.76, 95% CI 2.18-3.50] were most the factors most significantly associated with KP.

Age [OR 1.08, 95% CI 1.05-1.10] and familial KOA or knee replacement were significantly associated with RKOA in footballers. Crystal deposition, typically in patients suffering from gout [OR 3.24, 95% CI 1.41-7.45] or with evidence of CC [OR 4.62, 95% CI 2.61-8.05] were also strongly associated with RKOA. As with KP, injury [OR 2.17, 95% CI 1.43-3.30] and its consequences, namely surgical intervention [OR 4.25, 95% CI 2.78-6.50] and receiving intra-articular injections [OR 1.60, 95% CI 1.10-2.35] were strongly associated with RKOA in footballers. A longer duration of playing football [OR 1.04, 95% CI 1.00-1.08] was also strongly associated with RKOA in footballers. For every 1000 hours trained [OR 1.13, 95% CI 1.07-1.19] and 100 matches played [OR 2.41, 95% CI 1.08-5.36], footballers had an increased risk of RKOA.

Risk factors significantly associated with footballers who had undergone a TKR, included age [OR 1.09, 95% CI 1.07-1.11], being obese [OR 1.77, 95% CI 1.00-3.12] and having gout [OR 3.11, 95% CI 1.96-4.70]. Sustaining a significant knee injury [OR 3.11, 95% CI 1.94-4.99] and receiving an intra-articular knee injection [OR 2.56, 95% CI 1.76-3.73] were also significant risk factors for footballers who underwent a TKR. However, those footballers with a longer duration of playing the game [OR 0.95, 95% CI 0.92-0.98] had a reduced risk of TKR.


These findings show footballers have a greater risk of KOA compared to the general population, reporting up to three times higher prevalence of various outcomes (KP, RKOA, physician-diagnosed KOA and TKR). The age-prevalence of all outcomes of KOA, are greater in footballers compared to the controls. The high prevalence of injuries significantly account the risk of KOA in footballers compared to the controls (even following adjustment of other risk factors).

Within footballers, knee injuries, together with subsequent investigations (specifically exploratory and interventional arthroscopy) and management (specifically intra-articular knee injections), were strongly associated with risk of KOA (KP, RKOA and TKR independently).

Football’s governing bodies need to set out and implement strategies to reduce or even prevent the risk of serious injury (thus reducing the risk of subsequent investigation). This will reduce the risk of long-term consequences, such as KOA. However, whether the Industrial Injuries Advisory Council considers the risk of KOA in footballers an industrial compensable disease remains a question.

Item Type: Thesis (University of Nottingham only) (PhD)
Supervisors: Doherty, Michael
Zhang, Weiya
Fernandes, Gwen
Keywords: English; football; footballers; soccer; knee pain; knee injuries; osteoarthritis
Subjects: W Medicine and related subjects (NLM Classification) > WE Muscoskeletal system
Faculties/Schools: UK Campuses > Faculty of Medicine and Health Sciences > School of Medicine
Item ID: 43010
Depositing User: Parekh, Sanjay
Date Deposited: 29 Aug 2017 08:59
Last Modified: 07 Sep 2017 11:50

Actions (Archive Staff Only)

Edit View Edit View