eTable 5. Study characteristics.

Study	Study region	Time of recruitment/follow-up	Exposure group	Comparison group	Clinical endpoint (outcome)
Retrospective cohort study
Kujala et al., 1994	Finland	n.d.	Male previous professional soccer players (n = 313) who had represented Finland at least once at the Olympic Games, World Cup, European Championship, or at national level, recruited from a partial sample of the study by Sarna et al. (1993) (e88) Response rate (questionnaire): 84% Average age: 45.4 years (range 23–76)	Male probands (n = 1712) from the population who had been classed as fully healthy at the time of their military training, matched for age and place of residence, recruited via public archives Response rate (questionnaire): 77% Average age: 44.1 years (range 24–86)	Hospital admission due to knee osteoarthritis (according to the national register of hospital discharges, 1970–1990) (Excluded: infection of lower extremity, rheumatoid arthritis)
Studie	Study region	Time of recruitment/follow-up	Cases	Controls	Exposure
Case–control studies
Sandmark & Vingard, 1999	Sweden	n.d.	Male (n = 325) and female (n = 300) cases, born between 1921 and 1938, with knee joint replacement due to primary tibiofemoral knee osteoarthritis in the period 1991–1993 according to the Swedish knee replacement registry Response rate: Men: 88%, women: 79% (Excluded: trauma or surgery of the knee joint or surrounding structures, rheumatoid arthritis, systemic disease affecting the joints [poliomyelitis, rachitis, etc.], musculoskeletal malalignment)	Male (n = 264) and female (n = 284) population controls, born between 1921 and 1938, recruited randomly from the Swedish population registry Response rate: Men: 80%, women: 77% (Excluded: trauma or surgery of the knee joint or surrounding structures, rheumatoid arthritis, systemic disease affecting the joints [poliomyelitis, rachitis, etc.], musculoskeletal malpositions)	Professional or amateur soccer during the ages of 15 and 50 years
Thelin et al., 2006	Sweden	n.d.	Male (n = 338) and female (n = 440) cases (age: ≤ 70 years) with advanced severe or moderate tibiofemoral knee osteoarthritis according to the radiographic registries of six southern Swedish hospitals or following knee joint surgery (osteotomy or joint replacement) or with moderate cartilage degeneration or a joint space of ≤ 3 mm Response rate: 94.3% (non-responder analysis performed) Average age (cases and controls): 62.6 years (range 51–70) (Excluded: knee osteoarthritis, minor cartilage degeneration, joint space ≥ 4 mm, knee osteoarthritis of Ahlbäck grade 1 oder 2, patellofemoral knee osteoarthritis, chronic inflammatory joint disease)	Male (n = 293) and female (n = 402) population controls (age: ≤ 70 years), matched for age, sex, and commune, recruited from the Swedish population registry Response rate: 84.2 % (non-responder analysis performed) Average age (cases and controls): 62.6 years (range 51–70)	Professional or amateur soccer for at least 1 year after the age of 15 years
Vrezas et al., 2010	Germany	n.d.	Male cases (n = 295) with radiographically diagnosed knee osteoarthritis (Kellgren–Lawrence score ≥ 2) with chronic symptoms (age: 25–70 years), recruited via orthopedic departments and orthopedists in the community Response rate: 61% (non-responder analysis performed)	Male population controls (n = 327; age: 25–70 years), recruited randomly from the residents’ registries of the cities Frankfurt and Offenbach, Germany Response rate: 55% (non-responder analysis performed)	Hours spent playing soccer (whole lifetime; unclear whether professional or amateur soccer)
Studie	Study region	Time of recruitment/follow-up	Exposure group	Comparison group	Clinical endpoint (outcome)
Cross-sectional studies
Fernandes et al., 2018	UK	12/2014–07/2015	Male former professional soccer players (n = 1207) aged > 40 years who had played in the top four divisions in England, recruited via 22 professional football organizations (with a partial sample who had had radiographic examination of the knee joint: n = 470) Response rate: questionnaire 25.3%, radiographs n.d. Average age: 59.0 years (± 11.7) Body mass index (current): 27.3 kg/m 2 (± 3.2)	Male probands (n = 4085) from the population who were not incurably ill, were capable of giving written informed consent, and whose primary care physicians discerned no other grounds for exclusion, recruited from12 primary care practices (with a partial sample who had had radiographic examination of the knee joint: n = 491) Response rate*1: questionnaire 23.8%, radiographs 1.2% Average age: 62.9 years (± 10.4) Body mass index (current): 27.5 kg/m 2 (± 4.7)	Diagnosis of knee osteoarthritis by a physician (self-reported) Knee joint replacement (self-reported) (Given the inclusion and exclusion criteria, no data extraction of the radiographic examination results was performed, because no response data on the radiographic examinations in the professional soccer players were published and the response rate among the controls was 1.2%.)
Iosifidis et al., 2015	Greece	n.d.	Male former professional soccer players (n = 121) aged ≥ 40 years who had played at national level and participated in at least one European Championship or World Cup and had been active up to the age of 25 years, recruited via sports organizations (with a partial sample who had had radiographic examination of the knee joint: n = 91) Response rate: 84.4% Average age (all professional athletes included): 50.1 years (± 8.5) Body mass index (all professional athletes included, at age 20 years): 22.3 kg/m 2 (±1.6)	Male probands (n = 181) from the population aged ≥ 40 years, resident in the same region, who had been classed as fully healthy at the time of their military training (with a partial sample who had had radiographic examination of the knee joint: n = 163) Response rate: 86.6% Average age: 50.7 years (± 10.0) Body mass index (at age 20 years): 21.5 kg/m 2 (± 1.6)	Clinical knee osteoarthritis (pronounced pain or functional impairment of knee joint in recent years and also at least one of the following clinical signs: limited mobility, pain, crepitus, malposition) Radiographically diagnosed knee osteoarthritis (Kellgren–Lawrence score ≥ 2, no information on whether or not a tangential radiograph of the patella was obtained) (Excluded: lower extremity surgery, bone or soft-tissue trauma, inflammatory joint disease)
Kujala et al., 1995	Finland	1985	Male former professional soccer players (n = 31) who had represented Finland at least once at the Olympic Games, World Cup, European Championship, or at national level, recruited from a partial sample of the study by Sarna et al. (1993) (e88) Response rate: 83.8% Average age: 56.5 years (± 5.7) Body mass index (at age 20 years): 22.9 kg/m 2 (±1.4)	Male professional sports shooters (n = 29) Response rate: 82.9% Average age: 61.0 years (± 4.3) Body mass index (at age 20 years): 22.49 kg/m 2 (± 2.3)	Radiographically diagnosed knee osteoarthritis (Kellgren–Lawrence score ≥ 2)
Roos et al., 1994	Sweden	1988	Male former professional soccer players (n = 71; top two divisions in Sweden) and amateur soccer players (n = 215; third or lower divisions in Sweden), all aged ≥ 40 years,), who had been active up to the age of 25 years,recruited from all 10 football clubs in the city of Malmö Response rate: 100% Average age: Professional soccer players: 62.7 years, amateur soccer players: 53.2 years	Male probands (n = 572) from the population, matched for age, recruited randomly via the Swedish population registry Response rate: 100% Average age: 55.5 years	Radiographically diagnosed knee osteoarthritis (The radiographic archives of a hospital and two private practices were searched for the knee radiographs of all participating soccer players and controls. Radiographs of 253 soccer players and probands were found. Radiographic knee osteoarthritis was diagnosed according to the criteria of Ahlback (1968) (e35): joint space in p.a. view of tibiofemoral joint only half as great as in the other compartment of the same knee joint or the same compartment of the contralateral knee joint or joint space height of > 3 mm in one compartment. Other changes, e.g., osteophytes were ignored.)
Tveit et al., 2012	Sweden	n.d.	Male former professional soccer players (n = 397) who had been active at national or international level, recruited from an archived book on former professional athletes, via archives of the Swedish Olympic Committee, and from the study by Nilsson & Westlin (1971) (e89) Response rate: 74% Median age (of all investigated professional athletes): 70 years (range 50–93)	Male probands (n = 1368) from the population, matched for age, recruited via the Swedish population registry Response rate: 64% Median age: 70 years (range 51–93)	Diagnosis of knee osteoarthritis by a physician (self-reported) Knee joint replacement (self-reported) (Excluded: earlier knee joint fractures)

*¹ The response rate relates to the Knee Pain and Related Health in the Community Study, which covered men and women.n.d., No data