Table 3 Patient characteristics studied as determinants for clinical knee OA progression in the included studies
Determinant | Author [ref.] year | Instrument of measurement | Definition of knee OA progression | OR/RR/HR/β (95 % CI) | Association with OA prognosis* |
|---|---|---|---|---|---|
Age | Muraki [25] 2012 | Per 5 years increase | Incident knee pain at follow-up (baseline K/L ≥2) | OR 1.01 (0.95–1.07) | o |
(N = 10,043) | Holla [20] 2010 | Continuous (years) | Progressing/remaining in poor WOMAC-PF quintiles | OR 0.97 (0.94–1.00) | – |
|  | Holla [21] 2014 | Continuous (years) | Poor vs good outcome WOMAC-PF trajectory | OR 0.94 (0.88–1.00) | – |
|  | Collins [15] 2014 | Continuous (years) | Severe vs no pain WOMAC pain trajectory | OR 0.92 (0.89–0.96) | – |
|  | Riddle [29] 2009 | Continuous (years) | Knee joint surgery | OR 1.07 (1.02–1.11) | + |
|  | Riddle [30] 2012 | Continuous (years) | Knee joint surgery | RR 1.04 (1.01–1.23) | + |
Female sex | Collins [15] 2014 | Female vs male | Severe vs no pain WOMAC pain trajectory | OR 3.0 (1.5–6.2) | + |
(N = 3,066) | Muraki [25] 2012 | Female vs male | Incident knee pain at follow-up (baseline K/L ≥2) | OR 1.32 (0.94–1.84) | o |
Ethnicity | Collins [15] 2014 | Non-white vs white | Severe vs no pain WOMAC pain trajectory | OR 3.3 (1.7–6.6) | + |
(N = 2,585) | Holla [20] 2010 | Non-western vs western | Progressing/remaining in poor WOMAC-PF quintiles | OR 4.03 (1.06–15.4) | + |
Education level | Collins [15] 2014 | <College ≥ college | Severe vs no pain WOMAC pain trajectory | OR 5.1 (2.3–11.2) | + |
(N = 2,531) | Riddle [29] 2009 | ≤ High school graduate | Knee joint surgery | OR 2.40 (1.09–5.28) | + |
Body mass index (N = 4,857) | Holla [20] 2010 | Continuous | Progressing/remaining in poor WOMAC-PF quintiles | OR 1.06 (1.02–1.11) | + |
Sharma [34] 2003 | Per 5 unit increase | Progressing/remaining in poor WOMAC-PF quintiles | OR 1.15 (0.89–1.46) | o | |
Collins [15] 2014 | Obese vs non-obese | Severe vs no pain WOMAC pain trajectory | OR 2.3 (1.2–4.4) | + | |
Holla [21] 2014 | Continuous | Moderate vs good outcome WOMAC-PF trajectory | OR 1.12 (1.07–1.18) | + | |
Muraki [25] 2012 | Per 5 units increase | Incident knee pain at follow-up (baseline K/L ≥2) | OR 1.54 (1.90–1.82) | + | |
Riddle [29] 2009 | ≤30 vs >30 kg/m2 | Knee joint surgery | OR 2.66 (1.20–5.92) | + | |
Bodyweight change | Riddle [31] 2013 | ≥–10 % vs –4.9 to +4.9 % |  | β 4.07 (1.49–6.65) | – |
(N = 1,410) |  | –9.9 to –5 % vs –4.9 to +4.9 % | Increase self-reported limitations (WOMAC-PF) | β 0.01 (–1.87 to 1.89) | o |
|  |  | +5 to +9.9 % vs –4.9 to +4.9 % | Increase in pain (WOMAC); similar results, not tabulated | β 1.08 (–0.91 to 3.07) | o |
|  |  | ≥ + 10 % vs –4.9 to +4.9 % |  | β –5.36 (–8.74 to –2.00) | + |
Knee compression force | Henriksen [18] 2013 | Change in peak knee joint compressive force | Change in KOOS-4 | LSMD –2.4 (–6.8 to 1.9) | o |
(N = 157) |  | Change in walking speed, m/s | LSMD –0.01 (–0.05 to 0.03) | o | |
Co-morbidity count | Holla [20] 2010 | ≥3 vs none | Progressing/remaining in poor WOMAC-PF quintiles | OR 1.53 (0.93–2.53) | o |
(N = 3,672) | Holla [21] 2014 | ≥3 vs < 3 | Poor vs good outcome WOMAC-PF trajectory | OR 3.28 (1.62–6.64) | + |
|  | Collins [15] 2014 | ≥1 vs 0 | Severe vs no pain WOMAC pain trajectory | OR 2.0 (1.0–3.9) | + |
|  | Pisters [27] 2012 | Per unit CIRS increase | Increase in self-reported limitations (WOMAC-PF) | β 3.69 (1.66–8.23) | + |
|  | Van Dijk [37] 2011 | Per unit CIRS increase | Increase in self-reported limitations (WOMAC-PF) | β –0.147 (not provided) | + |
|  |  |  | Increase in performance-based limitations (TWT) | β 0.150 (not provided) | + |
Mental health | Collins [15] 2014 | CES-D ≥16 vs CES-D < 16 | Severe vs no pain WOMAC pain trajectory | OR 8.8 (3.1–25.2) | + |
(N = 6,659) | Sharma [34] 2003 | Per 5 points score | Progressing/remaining in poor WOMAC-PF quintiles | OR 0.58 (0.39–0.86) | – |
|  | Riddle [30] 2012 | Per unit SF-12 MCS score | Knee joint surgery | RR 1.07 (1.04–1.10) | + |
Vitality | Holla [21] 2014 | Per unit SF-36 HS score | Poor vs good outcome WOMAC-PF trajectory | OR 0.96 (0.94–0.98) | – |
(N = 871) | Van Dijk [37] 2011 | Per unit SF-36 MOS score | Increase in self-reported limitations (WOMAC-PF) | β 0.157 (not provided) | – |
|  |  |  | Increase in performance-based limitations (TWT) | β –0.229 (not provided) | – |
Pain coping | Alschuler [9] 2013 | CSQ subscale praying or hoping | ≥20 % change in combined NRS and WOMAC-PF score | Not provided | + |
(N = 1,048) |  | CSQ subscale catastrophizing | ≥20 % change in combined NRS or WOMAC-PF score | Not provided | + |
Holla [20] 2010 | PCI subscale distraction | Progressing/remaining in poor WOMAC-PF quintiles | OR 1.26 (0.98–1.62) | o | |
|  | PCI subscale worrying | OR 0.63 (0.66–0.73) | – | ||
Holla [21] 2014 | PCI subscale resting | Poor vs good outcome WOMAC-PF trajectory | OR 1.16 (1.02–1.31) | + | |
Pisters [27] 2012 | PCI subscale resting | Increase in self-reported limitations (WOMAC-PF) | β 23.3 (1.93–280.7) | + | |
|  |  |  | Increase in performance-based limitations (TWT) | β 3.13 (1.95–5.03) | + |
Morning stiffness | Holla [20] 2010 | <30 minutes, yes vs no | Progressing/remaining in poor WOMAC-PF quintiles | OR 1.37 (0.99–1.88) | o |
(N = 832) |  |  |  |  |  |
Knee injury | Muraki [25] 2012 | Previous knee injury | Incident knee pain at follow-up (baseline K/L ≥2) | OR 2.91 (1.26–6.82) | + |
(N = 1,313) |  |  |  |  |  |
Knee surgery | Riddle [30] 2012 | History of knee surgery | Knee joint surgery | RR 2.04 (1.33–3.13) | + |
(N = 4,670) |  |  |  |  |  |
Pain medication use | Riddle [30] 2012 | For pain, aching or stiffness | Knee joint surgery | RR 1.64 (0.87–3.12) | o |
(N = 6,516) | Lapane [22] 2015 | NSAID usage vs not | MICD of WOMAC pain | β –0.88 (–2.22 to 0.46) | o |
|  |  |  | MICD of WOMAC-PF | β –4.27 (–8.84 to 0.31) | o |
|  |  |  | MICD of WOMAC stiffness | β –0.72 (–1.56 to 0.12) | o |
Bisphosphonate use | Laslett [24] 2014 | Yes vs no | Decrease in WOMAC pain | β 0.69 (–0.54 to 1.92) | o |
(N = 323) |  |  | Decrease in WOMAC function | β 0.05 (–3.85 to 3.95) | o |
|  |  |  | Decrease in WOMAC stiffness | β –0.24 (–0.75 to 0.27) | o |
|  |  |  | Decrease in NRS after 3 and 4 years, but not 5 years | β –1.15 (–1.94 to –0.36) | + |
Glucosamine/chondroitin use | Yang [38] 2015 | Yes vs no | MICD of WOMAC pain | β 0.68 (–0.16 to 1.53) | o |
|  |  | MICD of WOMAC-PF | β 1.28 (–1.23 to 3.79) | o | |
(N = 1,625) |  |  | MICD of WOMAC stiffness | β 0.41 (0–0.82) | o |
History of HRS | Riddle [30] 2012 | Yes vs no | Knee joint surgery | RR 2.73 (0.93–8.07) | o |
(N = 4,670) |  |  |  |  |  |