Best evidence synthesis (MRI features arranged from top to bottom according to the number of studies included)
| Association (sizes) | Number of studies: positive/total (%) | |||||
|---|---|---|---|---|---|---|
| Studies | Study design | Crude | Adjusted | Adjusted confounders | All | High quality |
| Cartilage defects (level of evidence: conflicting) | ||||||
| Scored using semiquantitative scores | ||||||
| Pelletier et al31 | CS | r=0.09, p=0.38 | – | NA | 6/12 (50%) | 3 (1 C, 2 CS)/6 (2 C, 3 CS) (50%) |
| Phan et al32 | CS | r Not mentioned, NS | – | NA | ||
| Torres et al19 | CS | β=1.03 (95% CI 0.6 to 1.5) | 0.53 (0.08 to 0.98) | Age and BMI | ||
| Hayes et al26 | CS | Positive, p=0.001 | – | NA | ||
| Kornaat et al29 | CS | – | OR 1.12 (99% CI 0.4 to 3.2) | Age, sex, BMI, intrafamily effects | ||
| Link et al30 | CS | positive, p<0.05 | – | NA | ||
| Scored quantitatively | ||||||
| Raynauld et al14 | C | r=−0.25, NS (WOMAC), r=0.12, NS (VAS) | – | NA | ||
| Wluka et al16 | C | r=0.28, positive, p=0.002 | – | NA | ||
| Fernandez-Madrid et al25 | CS | NS | – | NA | ||
| Sowers et al34 | CS | Positive, p<0.0001 | – | NA | ||
| Dunn et al23 | CS | Positive, p<0.05 | – | NA | ||
| Scored using other methods (ie, quantitatively after giving contrast agent) | ||||||
| Anandacoomarasamy et al35 | CS | r=−0.21, p=0.07 | – | NA | ||
| Bone marrow lesion (level of evidence: moderate) | ||||||
| Kornaat et al20 | C | – | Mean difference (increasing BML)=2 (95% CI −8 to 11) | Age, sex, BMI, intrafamily effects | 5/8 (63%) | 4 (CS)/5 (1 C, 4 CS) (80%) |
| Hayes et al26 | CS | Positive, p=0.001 | – | NA | ||
| Felson et al24 | CS | – | OR=3.31 (95% CI 1.5 to 7.4) | Age, sex, radiological and effusion score | ||
| Link et al30 | CS | p>0.05 | – | NA | ||
| Lo et al36 | CS | Positive, RR BML scores versus no BML: | Positive: | Effusion and synovitis | ||
| 1:1.3 | 1:1.2 | |||||
| 2:2.1 | 2:1.9 | |||||
| 3:2.3 | 3:2.0 | |||||
| p For trend=0.0009 | p For trend 0.006 | |||||
| Phan et al32 | CS | r Not mentioned, NS | – | NA | ||
| Sowers et al34 | CS+ | OR=5.0 (95% CI 2.4 to 10.5) | – | NA | ||
| Torres et al19 | CS+ | β=5.0 (95% CI 3.0 to 7.0) | 3.72 (1.8 to 5.7) | Age and BMI | ||
| Osteophytes (level of evidence: limited) | ||||||
| Presence | ||||||
| Fernandez-Madrid et al25 | CS | NS | – | NA | 2/6 (33%) | 0/2 (CS) (0%) |
| Hayes et al26 | CS | Positive, p<0.001 | – | NA | ||
| Kornaat et al29 | CS | – | OR=1.05 (99% CI 0.4 to 2.9) | Age, sex, BMI, intrafamily effects | ||
| Link et al30 | CS | p>0.05 | – | NA | ||
| Torres et al19) | CS | β=1.2 (95% CI 0.6 to 1.7) | B=0.5 (0.07 to 0.94) | Age and BMI | ||
| Signal strength | ||||||
| Sengupta et al33 | CS | – | PR=0.94 (0.8 to 1.1) | Age, sex, BMI | ||
| Meniscal lesion (level of evidence: conflicting) | ||||||
| Bhattacharyya et al18 | CS | – | p=0.7 | Age | 2/6 (33%) | 1/3 (CS) (33%) |
| Fernandez-Madrid et al25 | CS | NS | – | NA | ||
| Hayes et al26 | CS | Positive, p=0.001 | – | NA | ||
| Kornaat et al29 | CS | – | Tears: OR=1.26 (99% CI 0.6 to 2.7), subluxation: OR=1.03 (99% CI 0.5 to 2.2) | Age, sex, BMI, intrafamily effects | ||
| Link et al30 | CS | p>0.05 | – | NA | ||
| Torres et al19 | CS | Tears: β=3.3 (95% CI 0.9 to 5.8) | Tears: 2.0 (0.6 to 3.4) | Age and BMI | ||
| Subluxation: β=15.0 (95% CI −0.3 to 30.3) | Subluxation: 2.22 (−6.9 to 11.3) | |||||
| Effusion and synovitis (level of evidence: moderate) | ||||||
| Hill et al12 | C | – | Effusion: OR=1.2 (95% CI −8.1 to 10.5), synovitis: OR=3.2 (95% CI 1.04 to 5.3) | Age, sex, BMI, cartilage score at baseline, effusion score, BML score, change in effusion and BML score. | 6/8 (80 %) | 4 (1 C, 3 CS)/4 (1 C, 3 CS)(100%) |
| Fernandez-Madrid et al25 | CS | Effusion: positive, p<0.001 | – | NA | ||
| Synovitis: NS | – | |||||
| Hayes et al26 | CS | Effusion: positive, p<0.001 | – | NA | ||
| Synovitis: positive, p<0.001 | – | |||||
| Kornaat et al29 | CS | – | Effusion: OR=10.0 (99% CI 1.1 to 149) | Age, sex, BMI, intrafamily effects | ||
| Link et al27 | CS | Effusion: p>0.05 | – | NA | ||
| Lo et al*36 | CS+ | Effusion: RR BML scores versus no BML: | Synovitis and BML | |||
| 1:1.8 | 1:1.7 | |||||
| 2:2.4 | 2:2.0 | |||||
| 3:3.1 | 3:2.6 | |||||
| p For trend<0.0001 | p For trend=0.0004 | |||||
| Synovitis: | Synovitis: | |||||
| 1:1.9 | 1:1.4 | |||||
| 2:1.9 | 2:1.5 | |||||
| 3:2.3 | 3:1.9 | |||||
| p For trend 0.20 | p For trend=0.22 | |||||
| Torres et al19 | CS | β=15.0 (95% CI −8.2 to 38.2) | 9.8 (0.4 to 19.3) | Age and BMI | ||
| Pelletier et al21 | C | Effusion: r=0.07, positive, p=0.71 (WOMAC); r=0.01,positive, p=0.93 (VAS) | – | NA | ||
| Knee ligament abnormalities (level of evidence: limited) | ||||||
| Amin et al22 | CS | – | ACL: positive, p<0.05 | Age, sex, BMI and cartilage scores | 2/5 (40%) | 2/2 (CS) (100%) |
| Hill et al28 | CS | ACL: positive, p=0.0004 | – | NA | ||
| Link et al30 | CS | ACL: p>0.05, MCL and LCL: p>0.05 | – | NA | ||
| Torres et al19 | CS | β (95% CI) | Age and BMI | |||
| ACL: 5.0 (−13.0 to 23.0) | ACL: 6.8 (−5.4 to 19.0) | |||||
| MCL: 0 (−11.9 to 11.9) | MCL: −6.10 (−14.0 to 1.7) | |||||
| LCL:15.0 (95% CI −8.2 to 38.2) | LCL: 29.5 (17.8 to 41.1) | |||||
| Hayes et al26 | CS | ACL and PCL: p=0.23, MCL and LCL, p=0.86 | – | NA | ||
| Subchondral cysts (level of evidence: limited) | ||||||
| Hayes et al26 | CS | Positive, p<0.001 | – | NA | 1/5 (20%) | 0/2 (CS) (0%) |
| Kornaat et al29 | CS | – | OR=1.71 (99% CI 0.8 to 3.6) | Age, sex, BMI, intrafamily effects | ||
| Link et al27 | CS | p>0.05 | – | NA | ||
| Fernandez-Madrid et al22 | CS | NS | – | NA | ||
| Torres et al19 | CS | β=2.50 (95% CI −0.4 to 5.4) | 0.82 (−0.5 to 2.1) | Age and BMI | ||
| Bone attrition (level of evidence: conflicting) | ||||||
| Hernández-Molina et al27 | CS | OR=2.1 (95% CI 1.4 to 3.4) | 1.2 (0.7 to 2.0) | Age, sex, BMI, K&L grade, presence of BML and effusion | 1/2 (50%) | 1/2 (CS) (50%) |
| Torres et al19 | CS | β=3.3 (95% CI 1.8 to 4.9) | 1.9 (0.7 to 3.1) | Age and BMI | ||
Author's name in bold indicates high-quality studies; ‘positive’ in front of p values indicates significant positive association sizes.
r: (Spearman's or Pearson's) correlation coefficient between MR feature of interest and pain in continuous scale (WOMAC pain subscale or VAS); in a cohort study the correlation coefficient showed the association between changes of the MRI features with the changes in pain during the follow-up. OR, odds of having pain (in cross-sectional studies) or increasing pain (in cohort studies) when a MRI feature is present or increasing comparing to the odds when MRI feature is absent. β is regression coefficient representing the increase in knee pain severity associated with increase in lesion score, PR, prevalence (odds) ratio.
ACL, anterior cruciate ligament; BMI, body mass index; BML, bone marrow lesion; C, cohort, CS, cross-sectional studies; K&L, Kellgren and Lawrence; LCL, lateral cruciate ligament; MCL, medial cruciate ligament; NA, not applicable; NS, not significant; PCL, posterior cruciate ligament; VAS, Visual analogue scale; WOMAC, Western Ontario and McMaster Scoring.