Summary of evidence of risk factors for structural knee OA (semiquantitative synthesis)
| Risk factor | Follow-up (years) and participants (n) | Phase of investigation (number of studies) | ROB | Inconsistency | Indirectness | Imprecision | Publication bias | Upgrading factors | Modified GRADE | Direction and magnitude* with structural OA |
| Age at injury38 | 14 (134) | Phase II (1) | X | Unclear | X | ✓ | ✓ | X | ++ | Association of injury age and OA odds is unclear |
| BMI 2 years post-ACL sx38 | 14 (134) | Phase II (1) | X | Unclear | X | ✓ | ✓ | ✓ | +++ | ↑ BMI 2 years post ACLRc may ↑ MTF and PFJ OA odds up to 3.5 times |
| Pre-injury activity level81 | 2 (421) | Phase II (1) | X | Unclear | X | ✓ | ✓ | X | ++ | Association of pre-injury activity level and OA odds is unclear |
| Manual occupation38 | 14 (134) | Phase II (1) | X | Unclear | X | X | ✓ | X | + | Association of pre-injury occupation and OA odds is unclear |
| Follow-up length84 | 8 (249) | Phase I (1) | XX | Unclear | X | ✓ | ✓ | X | + | Association of longer time since injury and OA odds is unclear |
| Injury to sx38 54 75 89 | 10–16 (585) | Phase I (2) Phase II (2) | X | ✓ | X | ✓ | ✓ | X | + | Length of time from injury to knee sx may not change OA odds |
| Bilateral knee injury72 | 7 (146) | Phase II (1) | ✓ | Unclear | ✓ | Unclear | Unclear | X | + | Association of bilateral knee injury and OA odds is unclear |
| ACL+concomitant injury 74 89 113 | 15–37 (504) |
Phase I (2) Phase II (1) | X | ✓ | ✓ | ✓ | ✓ | ✓ | +++ | ACL+concomitant injury (vs ACL only) ↑ OA odds up to 7.5 times |
| Second ACL71 | 5 (84) | Phase I (1) | X | Unclear | ✓ | Unclear | ✓ | X | + | Association of second ACL tear and OA odds is unclear |
| Meniscal injury72 | 7 (146) | Phase II (1) | ✓ | Unclear | ✓ | Unclear | Unclear | X | + | Association of meniscal injury and OA odds is unclear |
| Bilateral knee sx72 | 7 (146) | Phase II (1) | ✓ | Unclear | ✓ | Unclear | Unclear | X | + | Association of bilateral knee sx and OA odds is unclear |
| Index knee sx72 | 7 (146) | Phase II (1) | ✓ | Unclear | ✓ | X | X | ✓ | ++ | Association between index knee sx (vs no injury) and OA odds is unclear |
| ACLnoSx50 85 98 | 12–15 (318) | Phase I (2) Phase II (1) | XX | ✓ | X | X | ✓ | X | + | ACLnoSx (vs ACLR Sx) may not change OA odds |
| Early ACLRc44 61 | 5–37 (266) | Phase I (1) Phase II (1) | X | X | ✓ | X | ✓ | X | + | Early ACLRc (vs delayed or rehab only) may not change OA odds |
| Delayed ACLRc44 50 61 64 | 5–37 (438) | Phase I (3) Phase II (1) | X | X | X | X | ✓ | X | + | Association of delayed ACLRc (vs early or no sx) and OA odds is unclear |
| ACL+meniscus sx74 | 32–37 (127) | Phase I (1) | X | Unclear | X | Unclear | ✓ | X | + | Association between ACL+meniscal sx (vs ACL only) and OA odds is unclear |
| ACLRc+meniscal sx (future)85 | 12 (103) | Phase II (1) | X | Unclear | X | ✓ | ✓ | ✓ | ++ | Association between ACLRc+meniscal sx (future) and OA odds is unclear |
| ACLRp42 48 50 55 | 2–30 (304) | Phase I (1) Phase II (3) | XX | Unclear | ✓ | Unclear | ✓ | X | + | ACLRp (vs ACLRc) may not change OA odds |
| ACLRc approach47 | 12 (53) | Phase II (1) | ✓ | Unclear | ✓ | X | X | X | ++ | Association of ACLRc open (vs closed) and OA odds is unclear |
| ACLRc graft bundle number35 49 58 | 2–5 (624) | Phase II (3) | X | X | ✓ | X | ✓ | X | + | Double bundle allograft (vs single) may ↓ OA odds by 0.56 times |
| ACLRc graft type57 81 86 | 2–10 (587) | Phase I (1) Phase II (2) | XX | X | ✓ | X | ✓ | X | + | Association of autograft (vs allo) and OA odds are unclear |
| Sx post-ACLRc62 | 15 (109) | Phase I (1) | XX | Unclear | X | Unclear | X | X | + | Association of sx post-ACLRc and OA odds is unclear |
| ACLRc revision79 | 8 (108) | Phase I (1) | XX | Unclear | ✓ | Unclear | X | X | + | Association of ACLRc revision and OA odds is unclear |
| Meniscal repair28 | 14 (62) | Phase I (1) | XX | Unclear | X | Unclear | ✓ | X | + | Association of meniscal repair and OA odds is unclear |
| Cartilage sx45 | 10 (57) | Phase II (1) | X | Unclear | ✓ | X | ✓ | X | ++ | Association of cartilage sx with OA odds is unclear |
| Anterior tibial motion74 76 78 82 91 94 | 7–37 (1193) | Phase I (6) | XX | Unclear | X | Unclear | X | X | + | Association of anterior tibial motion post-ACLRc and OA odds is unclear |
| Accessory TF motion38 82 | 7–14 (197) | Phase I (1) Phase II (1) | XX | Unclear | ✓ | Unclear | Unclear | X | + | Association of accessory TF motion post-ACLRc and OA odds is unclear |
| AKP post-ACLRc77 | 20 (142) | Phase II (1) | ✓ | Unclear | X | X | ✓ | X | ++ | Association of AKP post-ACLRc with OA risk is unclear |
| Return to pivoting sport73 78 89 | 5–22 (473) | Phase I (2) Phase II (1) | XX | ✓ | X | Unclear | ✓ | X | + | Association of return to pivoting sport with OA odds is unclear |
| TF ROM post-ACL or ACLRc29 69 74 91 94 | 7–37 (1035) | Phase I (4) Phase II (1) | XX | ✓ | X | ✓ | ✓ | ✓ | + | Association of TF ROM post-ACL or ACLRc with OA odds is unclear |
| Hop post-ACL or ACLRc59 71 91 | 5–37 (425) | Phase I (4) | XX | Unclear | X | Unclear | X | ✓ | + | Association of hop function post-ACL or ACLRc and TF OA odds is unclear Asymmetry of hop function 4 years post-ACL may ↑ PF OA odds 5.1 times |
| Strength 4 years post-ACL tear74 | 32–37 (127) | Phase I (1) | X | Unclear | X | ✓ | ✓ | ✓ | + | Association of knee extension and flexion strength and OA odds is unclear |
| SR function post-ACL or ACLRc71 74 89 | 5–37 (421) | Phase I (2) Phase II (1) | X | ✓ | X | Unclear | ✓ | X | + | Association of SR knee function post-ACL or ACLRc and OA odds is unclear |
| SR physical activity 4 years post-ACL74 | 32–37 (127) | Phase I (1) | X | Unclear | X | Unclear | ✓ | X | + | Association of physical activity post-ACL or ACLRc and OA odds is unclear |
| Proximal medial tibial angle75 | 10 (117) | Phase I (1) | X | Unclear | X | X | ✓ | X | + | Association of proximal medial tibial angle and OA odds is unclear |
| Anatomic axis angle75 | 10 (117) | Phase I (1) | X | Unclear | X | X | ✓ | X | + | Association of anatomic axis angle and OA odds is unclear |
| Inflammatory biomarkers63 | 16 (69) | Phase II (1) | X | Unclear | X | ✓ | ✓ | X | ++ | Association of assessed biomarkers and OA odds is unclear |
Bold: Factor(s) with moderate to high quality evidence and OR, or RR consistently positive (more than one analysis) with the largest estimate ≥2.0.
*Unclear (estimates significant in inconsistent directions), may (estimates mostly significant in the same direction or mostly not significant), increase or decrease (estimates consistently significant in the same direction).
++, low-certainty evidence; ✓, no limitations; ++++, high-certainty evidence; +, very low-certainty evidence; +++, moderate-certainty evidence; ACL, ACL tear; ACLnoSx, ACL tear with no sx; ACLRc, ACL reconstruction; ACLRp, ACL repair; AKP, anterior knee pain; CKS, Cincinnati Knee Score; GRADE, Grading of Recommendations Assessment, Development and Evaluation; MTF, medial tibiofemoral; n, number; OA, osteoarthritis; PFJ, patellofemoral joint; ROB, risk of bias; SR, self-reported; Sx, surgery; TF, tibiofemoral; x, time; X, serious limitations; XX, very serious limitations.