The Risk of Knee Osteoarthritis in Professional Soccer Players—a Systematic Review With Meta-Analyses

Abstract

Background: We address the question whether professional soccer players with and without macroinjury of the knee joint are at an elevated risk for knee osteoarthritis.

Methods: A systematic review with meta-analyses was conducted. The study protocol was prospectively registered (registration number CRD42019137139). The MEDLINE, EMBASE, and Web of Science databases were searched for relevant publications; in addition, forward searching was performed, and the listed references were considered. All steps of the process were undertaken independently by two reviewers, and any discordances were resolved by consensus. For all publications whose full text was included, the methods used were critically evaluated. The quality of the evidence was judged using the GRADE criteria.

Results: The pooled odds ratio for objectively ascertained osteoarthrosis of the knee was 2.25 (95% confidence interval [1.41-3.61], I² = 71%). When only radiologically ascertained knee osteoarthrosis was considered, the odds ratio was 3.98 [1.34; 11.83], I² = 58%). The pooled risk estimator in studies in which knee joint macroinjury was excluded was 2.81 ([1.25; 6.32], I² = 71%).

Conclusion: A marked association was found between soccer playing and knee osteoarthritis in male professional soccer players. For female professional soccer players, the risk of knee osteoarthritis could not be assessed because of the lack of data. Knee injuries seem to play an important role in the development of knee osteoarthritis in professional soccer players.

Figure 1

Meta-analysis of objectively confirmed knee osteoarthritis*¹ in professional soccer players *¹ The meta-analysis included only studies in which the diagnosis of knee osteoarthritis was obtained by objective means, i.e., radiographically, status post knee replacement, or registry data showing hospital admission due to knee osteoarthritis. *² In all studies the comparison group was made up of population-based probands. 95% CI, 95% Confidence interval; SE, standard error

Figure 2

Sensitivity analysis 1: Meta-analysis of radiographically confirmed knee osteoarthritis*¹ in professional soccer players *¹ The meta-analysis included only studies in which the diagnosis of knee osteoarthritis was confirmed radiographically. *² The comparison groups were made up of population-based probands (Iosifidis et al., 2015; Roos et al., 1994) and professional sports shooters (Kujala et al., 1995). 95% CI, 95% Confidence interval; SE, standard error

Figure 3

Meta-analysis 2: Meta-analysis of the risk of knee osteoarthritis in professional soccer players, excluding those with macrotrauma*¹ *¹ The meta-analysis included only professional soccer players without macroinjuries. *² In all studies the comparison group was made up of population-based probands. 95% CI, 95% Confidence interval; SE, standard error

Figure 4

Meta-analysis 3: Meta-analysis of the risk of knee osteoarthritis in professional soccer players with and without macrotrauma*¹ *¹ The meta-analysis included professional soccer players with and without macroinjuries. *² In all studies the comparison group was made up of population-based probands. 95% CI, 95% Confidence interval; SE, standard error

eFigure 1

PRISMA flow chart * Studies with risk estimators for professional soccer, permitting the two research questions to be answered

eFigure 2

Sensitivity analysis 2: Meta-analysis on the risk of knee osteoarthritis in professional and amateur soccer players*¹ *¹ The meta-analysis additionally included studies that did not differentiate between professional and amateur soccer. *² In all studies the comparison group was made up of population-based probands. 95% CI, 95% Confidence interval; SE, standard error

eFigure 3

Sensitivity analysis 3: Meta-analysis on the risk of knee osteoarthritis in professional soccer players in studies with a high overall risk of bias*¹ *¹ The meta-analysis included only studies with a high overall risk of bias. *² In all studies the comparison group was made up of population-based probands. 95% CI, 95% Confidence interval; SE, standard error

eFigure 4

Sensitivity analysis 4: Meta-analysis on objectively confirmed knee osteoarthritis in professional soccer players, excluding studies in which the outcome was determined by self-reporting*¹ *¹ The meta-analysis did not include the two studies that recorded the outcome “status post knee joint replacement” as reported by participants (Fernandes et al., 2018; Tveit et al. 2012). *² In all studies the comparison group was made up of population-based probands. 95% CI, 95% Confidence interval; SE, standard error

eFigure 5

Meta-analysis 4: Meta-analysis of the risk of knee osteoarthritis in professional soccer players, without adjustment for knee joint injuries*¹ *¹ The meta-analysis included only risk estimators for which there was no adjustment for knee joint injuries. *² The comparison groups were made up of population-based probands (Fernandes et al., 2018; Roos et al., 1994; Tveit et al., 2012) and professional sports shooters (Kujala et al., 1995). 95% CI, 95% Confidence interval; SE, standard error

eFigure 6

Meta-analysis 5: Meta-analysis of the risk of knee osteoarthritis in professional soccer players, with adjustment for knee joint injuries*¹ *¹ The meta-analysis included only risk estimators for which there was adjustment for knee joint injuries. *² The comparison groups were made up of population-based probands (Fernandes et al., 2018; Roos et al., 1994; Tveit et al., 2012) and professional sports shooters (Kujala et al., 1995). 95% CI, 95% Confidence interval; SE, standard error

eFigure 7

Exposure–risk gradient for soccer playing and knee osteoarthritis (second-degree polynomial), calculated according to Vrezas et al. (2010) and Seidler et al. (2008)