Table 1.

Prospective cohort studies examining knee pain as a predictor of knee structural outcomes

Author, year	Participants	Exposure	Outcome measure	Main results
Outcome: change in cartilage volume or thickness
Raynauld et al., 2004 [10]	40 patients with symptomatic knee OA 2-year follow-up (N = 32)	WOMAC pain	Cartilage volume loss; slow (< 2% global cartilage loss) and rapid (> 15% global cartilage loss) progressors	Trend for higher baseline knee pain in rapid progressors than in slow progressors (49.9 ± 6.1 vs. 34.0 ± 4.4, p = 0.05)
Wluka et al., 2004 [11]	81 healthy post-menopausal women 2.5-year follow-up (N = 57)	WOMAC pain	Change in tibial cartilage volume	Baseline knee pain not associated with change in tibial cartilage volume (data not shown)
Wluka et al., 2004 [1]	132 people with symptomatic knee OA 2-year follow-up (N = 117)	WOMAC pain	Change in tibial cartilage volume	The severity of baseline knee pain did not predict subsequent tibial cartilage volume loss (r = 0.13, p = 0.14) There was a weak association between worsening of knee pain and increased tibial cartilage volume loss (r = 0.28, p = 0.002)
Raynauld et al., 2006 [12]	110 patients with symptomatic knee OA 2-year follow-up (N = 107)	WOMAC pain	Change in cartilage volume for the entire knee (global) and medial and lateral knee compartments	Medial compartment cartilage volume loss was associated with simultaneous knee pain change at 2 years (β coefficient − 0.45, p = 0.03)
Pelletier et al., 2007 [13]	110 patients with symptomatic knee OA 2-year follow-up (N = 107)	WOMAC pain	Change in knee cartilage volume from subregions	Baseline WOMAC pain not associated cartilage volume loss in medial central femoral condyle or medial central tibial plateau An increase in WOMAC pain score associated with cartilage volume loss in medial central tibial plateau (β coefficient − 0.26, p = 0.007)
Eckstein et al., 2011 [5]	718 participants with radiographic knee OA (K-L grades 2–4) 12-month follow-up	Questionnaire categorized as no, infrequent or frequent pain	Change in cartilage thickness in the central subregion of medial weight-bearing femoral condyle	Change in cartilage thickness − 12 μm in knees without pain vs − 54 μm in those with frequent pain at baseline, p = 0.01 The percentage of “progressors” (knees with cartilage thinning) was greater in knees with frequent pain than in those without pain for total joint (29% vs 16%, p = 0.004) and medial compartment (23% vs 13%, p = 0.015), but not for lateral compartment (20% vs 16%, p = 0.29)
Saunders et al., 2012 [14]	912 randomly selected individuals, 53% had radiographic OA in the medial compartment, 24% had radiographic OA in the lateral compartment 2.9-year follow-up (N = 399)	WOMAC pain	Change in tibial cartilage volume	Pain independently predicted lateral tibial cartilage volume loss. WOMAC knee pain: B − 0.14 (95% CI, − 0.22, − 0.05); knee pain yes/no: B − 0.96 (95% CI, − 1.91, − 0.00). No significant associations for medial tibial cartilage volume loss
Outcome: radiographic incidence or progression of OA
Spector et al., 1992 [15]	169 patients with OA of the hands or knees 11-year follow-up (N = 63)	VAS pain	Radiographic progression defined by > 10% reduction in joint space width or a K-L grade increase ≥ 1	9 of 15 people with knee pain at baseline had radiographic progression, compared with 6 of 16 without knee pain at baseline (p = 0.20)
Cooper et al., 2000 [16]	583 people from a population cohort 5.1-year follow-up (N = 354)	Have you had pain in or around your knee on most days for at least 1 month, at some time during the last year?	Incident and progressive OA defined using thresholds of both K-L grade 2 and K-L grade 1	Knee pain associated with incident OA when defined by K-L ≥ 1 (OR 2.9, 95% CI 1.2–6.7) but not K-L ≥ 2 (OR 1.3, 95% CI 0.6–2.7) No association between knee pain and progressive OA defined by either K-L ≥ 1 (OR 0.8, 95% CI 0.4–1.7) or K-L ≥ 2 (OR 2.4, 95% CI 0.7–8.0)
Wolfe et al., 2002 [17]	1507 patients with symptomatic knee OA	VAS pain	Joint space narrowing score = 3	Knee pain not associated with progression to maximum joint space narrowing (data not shown)
Miyazaki et al., 2002 [6]	106 patients with medial compartment knee OA 6-year follow-up (N = 74)	Knee rating system of the Hospital for Special Surgery	Radiographic progression defined as > 1 grade increase in narrowing of joint space width of the medial compartment	Baseline knee pain not associated with radiographic progression (OR 0.93, 95% CI 0.78–1.11, p = 0.43)
Mazzuca et al., 2005 [46]	174 obese women with unilateral knee OA 16- and 30-month follow-up	WOMAC pain	Joint space narrowing > 0.50 mm	Baseline WOMAC pain > 11 as a predictor of joint space narrowing > 0.50 mm: Month 16 (n = 73): sensitivity 77%, positive predictive value 36%, specificity 59%, negative predictive value 89% Month 30 (n = 70): sensitivity 65%, positive predictive value 45%, specificity 62%, negative predictive value 78%
Muraki et al., 2012 [18]	3040 people from a population-based cohort 3.3-year follow-up (N = 2262)	Have you experienced right/left knee pain on most days in the past month, in addition to now?	Incident radiographic knee OA; progressive radiographic knee OA	Knee pain at baseline associated with incident K-L grade > 3 knee OA (OR 2.53, 95% CI 1.59–4.00) and progressive radiographic knee OA (OR 2.63, 95% CI 1.81–3.81), but not incident K-L grade > 2 knee OA
Driban et al., 2016 [19]	1930 participants with no radiographic knee OA in either knee 4-year follow-up	WOMAC pain	Incident accelerated knee OA: at least one knee progressed to end-stage knee OA (K-L grade > 3)	Individuals with accelerated knee OA had greater WOMAC pain (OR 2.00, 95% CI 1.33–3.00)

Abbreviations: K-L Kellgren Lawrence, OA Osteoarthritis, VAS Visual analogue scale, WOMAC Western Ontario and McMaster Universities Osteoarthritis Index