Clinical Features of Patellofemoral Joint Osteoarthritis
Clinical Features of Patellofemoral Joint Osteoarthritis
Between August 2002 and September 2003, 819 people attended the research clinic, of whom 745 were eligible for the current analysis (mean (SD) age 65.2 (8.6) years; 55% female; mean (SD) body mass index 29.6 (5.2) kg/m). Reasons for ineligibility were: participants declining radiography (n = 2), incomplete radiographic data (total knee replacement in index knee (n = 15), unlabelled PA view (n = 2), absent patella (n = 2), uninterpretable skyline view (n = 5)), existing diagnosis of inflammatory arthritis verified by medical record review (n = 16), no knee pain in the last six months (n = 32).
When applying the lower threshold definition of radiographic OA, the numbers of participants classed as no radiographic OA, isolated patellofemoral joint OA, isolated tibiofemoral OA, and combined patellofemoral/tibiofemoral joint OA were 236 (32%), 178 (24%), 30 (4%) and 301 (40%), respectively. Due to the small number with isolated tibiofemoral joint OA, modelling was limited to comparing the clinical features of no radiographic OA, isolated patellofemoral joint OA and combined patellofemoral/tibiofemoral joint OA.
In addition to age, gender and body mass index, a total of 21 risk factors, clinical signs and symptoms were significantly different on at least one pair-wise comparison (Additional file 1), suggesting them as diagnosis-relevant indicators. Due to small numbers we were unable to include previous menisectomy or total knee replacement in the contralateral knee.
The regression function for isolated patellofemoral joint OA compared to no radiographic OA had the lowest AUC and greatest overlap in predicted probabilities. Difficulty descending stairs (adjusted OR 1.83; 95% CI 1.13, 2.96) and the presence of coarse crepitus (definite crepitus: aOR 2.46; 1.32, 4.60) were marginally informative when added to age, sex and body mass index but added little discriminative power (AUC 0.71 (95% CI 0.66, 0.76) vs 0.69 (0.64, 0.74); Χ = 1.23; P = 0.264).
Combined patellofemoral/tibiofemoral joint OA was distinguished from no radiographic OA by older age, higher body mass index, patient-reported onset following injury (aOR 2.18; 1.07, 4.44), stiffness on waking (1.92; 1.10, 3.34), difficulty descending stairs (2.53; 1.40, 4.57), palpable effusion (for example, mild effusion: 3.08; 1.75, 5.42), fixed flexion deformity (7.58; 2.08, 27.58), coarse crepitus (for example, definite crepitus: 3.38; 1.75, 6.55) and lower knee flexion range of motion (0.96; 0.94, 0.99). Female gender and the patient-reported whole leg pain (0.28; 0.13, 0.61) tended to indicate no radiographic OA.
Compared with isolated patellofemoral joint OA, individuals with combined patellofemoral/tibiofemoral joint OA were more likely to be older, female, obese and have varus deformity (2.11; 1.18, 3.75), palpable effusion (for example, mild effusion: 2.82; 1.70, 4.69), bony enlargement (for example, definite bony enlargement: 3.01; 1.56, 5.81), fixed flexion deformity (2.11; 1.04, 4.28) and lower knee flexion range of motion on examination (0.96; 0.94, 0.99).
In the final multinomial model, with the isolated patellofemoral group as the reference, the probability of subtypes of any knee OA was a joint function of age, gender, body mass index, patient-reported whole leg pain and difficulty descending stairs and, on physical examination, intercondylar gap, palpable effusion, fixed flexion deformity, bony enlargement, coarse crepitus and knee flexion range of motion ( Table 3 ). Classification based on the 'balance of probabilities' was correct in 392 (67%) instances. A confident (≥80% probability) correct diagnosis of isolated patellofemoral joint OA and combined patellofemoral/tibiofemoral joint OA was possible in 0 and 79 (28%) cases, respectively.
When applying the more stringent cut-off for radiographic OA ('moderate to severe OA'), the numbers of participants classed as no/mild radiographic OA, isolated patellofemoral joint OA, isolated tibiofemoral joint OA and combined patellofemoral/tibiofemoral joint OA were 453 (61%), 99 (13%), 123 (17%) and 70 (9%), respectively.
In addition to age, gender and body mass index, a total of 26 risk factors, clinical signs and symptoms were significantly different on at least one pair-wise comparison (Additional file 3), suggesting them as diagnosis-relevant indicators. This list of 26 potential indicators included all but two (patient-reported whole leg pain and incident pain) of those found to be associated in the 'any OA' models and seven additional indicators (patient-reported locking and significant interference with activities and, on physical examination, intermalleolar gap, pain on patellofemoral joint glide/compression, quadriceps strength, multiple local tender points, and timed single-leg standing balance).
All groups with moderate to severe knee OA were older and more obese than those with no-mild radiographic osteoarthritis (ROA). However, neither age nor body mass index appeared to differ between subsets with 'moderate to severe OA' after adjustment for covariates. Patient-perceived onset following injury, intercondylar gap > 0 cm (a crude measure of varus malalignment), palpable effusion, bony enlargement, fixed flexion deformity and lower knee flexion range of motion tended to be associated with tibiofemoral disease. By contrast, a recalled episode of dramatic swelling in the past, intermalleolar gap > 0 cm (valgus malalignment), markedly reduced knee extensor strength, and pain on PFJ compression appeared to indicate patellofemoral joint disease.
In the final multinomial model, with the isolated patellofemoral group as the reference, the probability of subsets of 'moderate to severe OA' was a joint function of age, sex, body mass index, patient-perceived time since the onset and onset following injury, patient-recalled dramatic swelling, self-reported difficulty descending stairs and, on physical examination, varus malalignment, valgus malalignment, pain on patellofemoral joint glide/compression, palpable effusion, fixed flexion deformity, bony enlargement, mediolateral instability, coarse crepitus, quadriceps strength and knee flexion range of motion (Table 4). Classification based on balance of probabilities was correct in 467 (68%) instances. Correct confident diagnosis of isolated patellofemoral joint OA, isolated tibiofemoral joint OA and combined patellofemoral/tibiofemoral joint OA was not possible in any cases.
Results
Study Participants
Between August 2002 and September 2003, 819 people attended the research clinic, of whom 745 were eligible for the current analysis (mean (SD) age 65.2 (8.6) years; 55% female; mean (SD) body mass index 29.6 (5.2) kg/m). Reasons for ineligibility were: participants declining radiography (n = 2), incomplete radiographic data (total knee replacement in index knee (n = 15), unlabelled PA view (n = 2), absent patella (n = 2), uninterpretable skyline view (n = 5)), existing diagnosis of inflammatory arthritis verified by medical record review (n = 16), no knee pain in the last six months (n = 32).
Comparative Clinical Features: 'Any OA'
When applying the lower threshold definition of radiographic OA, the numbers of participants classed as no radiographic OA, isolated patellofemoral joint OA, isolated tibiofemoral OA, and combined patellofemoral/tibiofemoral joint OA were 236 (32%), 178 (24%), 30 (4%) and 301 (40%), respectively. Due to the small number with isolated tibiofemoral joint OA, modelling was limited to comparing the clinical features of no radiographic OA, isolated patellofemoral joint OA and combined patellofemoral/tibiofemoral joint OA.
In addition to age, gender and body mass index, a total of 21 risk factors, clinical signs and symptoms were significantly different on at least one pair-wise comparison (Additional file 1), suggesting them as diagnosis-relevant indicators. Due to small numbers we were unable to include previous menisectomy or total knee replacement in the contralateral knee.
Isolated Patellofemoral Joint OA vs No Radiographic OA
The regression function for isolated patellofemoral joint OA compared to no radiographic OA had the lowest AUC and greatest overlap in predicted probabilities. Difficulty descending stairs (adjusted OR 1.83; 95% CI 1.13, 2.96) and the presence of coarse crepitus (definite crepitus: aOR 2.46; 1.32, 4.60) were marginally informative when added to age, sex and body mass index but added little discriminative power (AUC 0.71 (95% CI 0.66, 0.76) vs 0.69 (0.64, 0.74); Χ = 1.23; P = 0.264).
Combined Patellofemoral/Tibiofemoral Joint OA vs No Radiographic OA
Combined patellofemoral/tibiofemoral joint OA was distinguished from no radiographic OA by older age, higher body mass index, patient-reported onset following injury (aOR 2.18; 1.07, 4.44), stiffness on waking (1.92; 1.10, 3.34), difficulty descending stairs (2.53; 1.40, 4.57), palpable effusion (for example, mild effusion: 3.08; 1.75, 5.42), fixed flexion deformity (7.58; 2.08, 27.58), coarse crepitus (for example, definite crepitus: 3.38; 1.75, 6.55) and lower knee flexion range of motion (0.96; 0.94, 0.99). Female gender and the patient-reported whole leg pain (0.28; 0.13, 0.61) tended to indicate no radiographic OA.
Combined Patellofemoral/Tibiofemoral Joint OA vs Isolated Patellofemoral Joint OA
Compared with isolated patellofemoral joint OA, individuals with combined patellofemoral/tibiofemoral joint OA were more likely to be older, female, obese and have varus deformity (2.11; 1.18, 3.75), palpable effusion (for example, mild effusion: 2.82; 1.70, 4.69), bony enlargement (for example, definite bony enlargement: 3.01; 1.56, 5.81), fixed flexion deformity (2.11; 1.04, 4.28) and lower knee flexion range of motion on examination (0.96; 0.94, 0.99).
In the final multinomial model, with the isolated patellofemoral group as the reference, the probability of subtypes of any knee OA was a joint function of age, gender, body mass index, patient-reported whole leg pain and difficulty descending stairs and, on physical examination, intercondylar gap, palpable effusion, fixed flexion deformity, bony enlargement, coarse crepitus and knee flexion range of motion ( Table 3 ). Classification based on the 'balance of probabilities' was correct in 392 (67%) instances. A confident (≥80% probability) correct diagnosis of isolated patellofemoral joint OA and combined patellofemoral/tibiofemoral joint OA was possible in 0 and 79 (28%) cases, respectively.
Comparative Clinical Features: 'Moderate to Severe OA'
When applying the more stringent cut-off for radiographic OA ('moderate to severe OA'), the numbers of participants classed as no/mild radiographic OA, isolated patellofemoral joint OA, isolated tibiofemoral joint OA and combined patellofemoral/tibiofemoral joint OA were 453 (61%), 99 (13%), 123 (17%) and 70 (9%), respectively.
In addition to age, gender and body mass index, a total of 26 risk factors, clinical signs and symptoms were significantly different on at least one pair-wise comparison (Additional file 3), suggesting them as diagnosis-relevant indicators. This list of 26 potential indicators included all but two (patient-reported whole leg pain and incident pain) of those found to be associated in the 'any OA' models and seven additional indicators (patient-reported locking and significant interference with activities and, on physical examination, intermalleolar gap, pain on patellofemoral joint glide/compression, quadriceps strength, multiple local tender points, and timed single-leg standing balance).
All groups with moderate to severe knee OA were older and more obese than those with no-mild radiographic osteoarthritis (ROA). However, neither age nor body mass index appeared to differ between subsets with 'moderate to severe OA' after adjustment for covariates. Patient-perceived onset following injury, intercondylar gap > 0 cm (a crude measure of varus malalignment), palpable effusion, bony enlargement, fixed flexion deformity and lower knee flexion range of motion tended to be associated with tibiofemoral disease. By contrast, a recalled episode of dramatic swelling in the past, intermalleolar gap > 0 cm (valgus malalignment), markedly reduced knee extensor strength, and pain on PFJ compression appeared to indicate patellofemoral joint disease.
In the final multinomial model, with the isolated patellofemoral group as the reference, the probability of subsets of 'moderate to severe OA' was a joint function of age, sex, body mass index, patient-perceived time since the onset and onset following injury, patient-recalled dramatic swelling, self-reported difficulty descending stairs and, on physical examination, varus malalignment, valgus malalignment, pain on patellofemoral joint glide/compression, palpable effusion, fixed flexion deformity, bony enlargement, mediolateral instability, coarse crepitus, quadriceps strength and knee flexion range of motion (Table 4). Classification based on balance of probabilities was correct in 467 (68%) instances. Correct confident diagnosis of isolated patellofemoral joint OA, isolated tibiofemoral joint OA and combined patellofemoral/tibiofemoral joint OA was not possible in any cases.
Source...