Research Article

*Correspondence to Leena Sharma, Feinberg School of Medicine, Northwestern University, 303 East Chicago Avenue, 3-715, Chicago, IL 60611

Funded by:
  NIH (National Institute of Arthritis and Musculoskeletal and Skin Diseases); Grant Number: P60-AR-48098, R01-AR-48748
  NIH (National Center for Research Resources); Grant Number: RR-00048

Received: 20 August 2003; Accepted: 16 March 2004

10.1002/art.20348  About DOI

Knee osteoarthritis (OA) is responsible for a substantial portion of the disability attributable to OA as a whole. Its prevalence increases dramatically with age. Knee OA is a leading cause of disability in older individuals ([1]); its impact is likely to increase, given the doubling of the population of individuals ages 65 years or older that is expected to occur in the next 30 years.

Patellofemoral (PF) involvement in knee OA is relatively frequent and is associated with function limitation at least comparable with and perhaps exceeding the limitation associated with tibiofemoral OA ([2]). Despite this, natural history studies continue to emphasize the tibiofemoral compartments. Neither the natural history of PF OA nor risk factors for PF OA radiographic progression have been characterized. Knowledge of these factors will allow identification of the individuals in whom OA is most likely to worsen, and will support development of interventions to delay progression.

Patellofemoral pathology (e.g., dislocation, lateral pressure syndrome) is more common on the lateral side of the PF joint, due to a lateral reaction force vector ([3]). The relatively lateral position of the tibial tubercle in full-knee extension produces the Q-angle (i.e., the angle formed by the intersection of the line of application of the quadriceps force with the center line of the patellar tendon) ([3]), as shown in Figure 1. The Q-angle adds a strong, laterally directed component to the contact force ([3][4]). With each contraction, the anatomic origin (proximal femur) and insertion (tibial tubercle) points of the quadriceps muscle onto the bone move closer together. Because this action occurs through the patella, which is more medially placed than are these 2 points, a bowstring effect draws the patella laterally ([3]).

In keeping with the presence of this lateral vector and with the finding that other PF pathologies are more common on the lateral side of the joint, lateral PF OA is more common than is medial PF OA. Evidence for this comes from cross-sectional studies ([5]); the relative predilection of the PF compartments to OA progression has not been examined in a longitudinal study.

Risk factors for PF OA progression have not been characterized. Varus-valgus alignment influences joint forces at both the tibiofemoral and PF compartments. In a varus knee, the load-bearing axis passes medial to the knee, creating a moment arm that increases forces across the medial tibiofemoral compartment ([6]). In valgus, the lateral position of the load-bearing axis results in increased force across the lateral tibiofemoral compartment ([6]). Consistent with these biomechanical effects, varus alignment and valgus alignment have been shown to increase the risk of medial and lateral tibiofemoral OA progression, respectively ([7]). At the PF compartment, varus increases the Q-angle (Figure 1), increasing medial PF forces, and valgus decreases the Q-angle, increasing lateral PF forces ([8]). Whether alignment influences PF OA progression has not been examined.

In the present study, we tested the following hypotheses: that lateral PF OA progression is more common than is medial PF OA progression, that varus alignment increases the likelihood of medial PF OA progression, and that valgus alignment increases the likelihood of lateral PF OA progression.

PATIENTS AND METHODS

Patient sample

The Mechanical Factors in Arthritis of the Knee (MAK) study is a natural history study of knee OA, conducted at Northwestern University. MAK participants were recruited from the community through periodicals targeting seniors, 67 neighborhood organizations, the Aging Research Registry of the Buehler Center on Aging of Northwestern University, and referrals from medical centers. Inclusion and exclusion criteria were based on those proposed at a National Institute of Arthritis and Musculoskeletal and Skin Diseases/National Institute on Aging-sponsored workshop for OA progression studies ([9]). Inclusion criteria were definite osteophyte presence (i.e., Kellgren/Lawrence radiographic grade  2 [[10]]) in 1 or both knees and at least a little difficulty (Likert category) with more than 2 items on the Western Ontario and McMaster University Osteoarthritis Index physical function scale ([11]). Exclusion criteria were as follows: intraarticular corticosteroid injection into either knee within the previous 3 months; history of avascular necrosis, rheumatoid arthritis, or other inflammatory arthritis; periarticular fracture; Paget's disease; villonodular synovitis; joint infection; ochronosis; neuropathic arthropathy; acromegaly; hemochromatosis; Wilson's disease; osteochondromatosis; gout; pseudogout; osteopetrosis; and bilateral total knee replacement or a planned replacement within the next year. Patients who had previously undergone unilateral knee replacement were eligible if they had OA in the nonreplaced knee. Institutional review board approval was obtained, and all participants gave informed consent.

Measurement of alignment and body mass index (BMI)

To assess alignment, a single, anteroposterior radiograph of both lower extremities was obtained using a 51 × 14-inch graduated grid cassette, adhering to a protocol we previously described ([7]). Alignment was measured on the full-limb radiograph as the angle formed by the intersection of the line connecting the centers of the femoral head and intercondylar notch with the line connecting the centers of the ankle talus and tibial spines. A knee was described as being in varus when alignment was >0° in the varus direction, and valgus when alignment was >0° in the valgus direction. A single experienced reader made all measurements. For this reader, reliability was high for measurements of both varus (intraclass correlation coefficient [ICC] 0.99) and valgus (ICC 0.98) alignment, using a set of radiographs obtained from patients with OA ([7]). BMI was measured as weight (kg)/height (m²).

Measurement of outcome

PF and tibiofemoral radiographs of both knees were obtained in all participants. To visualize the PF compartment, weight-bearing, 30° flexion, axial (i.e., skyline) views of the PF joint were obtained according to a protocol that specified participant positioning and technical acquisition parameters ([12]). Tibiofemoral radiographs were obtained with the patient in the standing, semiflexed position according to the Buckland-Wright protocol (including knee position, beam alignment, markers to account for magnification, and fluoroscopic confirmation of knee position) ([12]). Radiographs were obtained in a single unit by 2 trained technicians. Foot maps made at baseline were used at the 18-month evaluation.

The hypotheses required assessment of PF OA progression separately in the medial and lateral PF compartments. Unlike the lateral view, the skyline view provides a view of each PF compartment. Joint space assessment is the only approach that allows separate evaluation of medial and lateral PF OA progression. Joint space can be directly measured or graded. Whether direct measurement of the PF joint space is meaningful in longitudinal studies has been the focus of a small number of studies, with conflicting results ([13][14]). Of note, Ciccutini et al recently reported that with every increase (worsening) in grade of the PF joint space on the skyline view, patellar cartilage volume on magnetic resonance imaging was reduced ([15]). This was the case for both PF compartments.

OA progression was defined consistently in each PF and tibiofemoral compartment. The medial PF, lateral PF, medial tibiofemoral, and lateral tibiofemoral compartments were each graded separately using the OA Research Society International atlas-based scales described by Altman et al ([16]), in which grade 0 = no narrowing, grade 1 = possible narrowing, grade 2 = definite narrowing, and grade 3 = severe narrowing. Progression was defined as an increase (worsening) of the grade between baseline and 18 months. Readings were performed in a random order by a single experienced reader. For this reader, the reliability, using kappa coefficients, was good (e.g.,   = 0.80-0.86).

Statistical analysis

Knees with the most severe grade of PF joint space narrowing in either the medial or lateral PF compartment at baseline were excluded from analyses of PF OA progression. The odds of progression were analyzed from logistic regression, using generalized estimating equations (GEE) to include data from 1 or both knees of each participant. The use of GEE provides estimates of the odds ratios (ORs) that correctly account for potentially correlated observations in the same individual. Alignment ORs were calculated for medial and lateral PF OA progression (both unadjusted and controlled for age, sex, and BMI). The associated 95% confidence intervals (95% CIs) were calculated. Both  isolated  medial or lateral PF OA progression (i.e., OA progression occurring in 1 PF compartment but not the other) and  any  medial or lateral PF OA progression (e.g., any medial PF OA progression was defined as PF OA progression in the medial compartment only as well as bicompartmental PF OA progression) were examined as outcomes. If malalignment does not play an important role in bicompartmental PF OA progression, then a smaller OR would be anticipated when the  any  definition was applied.

Next, the unadjusted and adjusted ORs for the outcome of PF OA progression, using baseline varus alignment as a continuous variable (in degrees; varus as a positive value, neutral as 0, and valgus as a negative value), were analyzed by logistic regression using GEE. Similarly, the relationship between baseline valgus alignment as a continuous variable and lateral PF OA progression was examined.

RESULTS

In this study, 237 participants, of whom 7 (3%) did not return at 18 months (5 had died, and 2 could not be located), were followed up longitudinally. Of the 230 remaining subjects, 13 had the most severe grade of PF joint space narrowing at baseline in both knees and were therefore excluded, leaving 217 participants with at least 1 knee without advanced PF OA. Among these 217 participants, 155 (71%) were women, the mean ± SD age was 68.4 ± 10.8 years, and the mean ± SD BMI was 30.6 ± 6.1. Thirty-seven of their knees had advanced PF OA at baseline; exclusion of these knees left 397 knees, which were the sample in this study. These 397 knees included 211 varus knees (53%) with a mean ± SD severity of varus malalignment of 5.0 ± 3.7°, 158 valgus knees (40%) with a mean ± SD severity of valgus malalignment of 3.9 ± 2.9°, and 28 neutral knees (7%) (0° alignment).

Lateral PF OA progression was more common than medial PF OA progression. Of the 397 knees, 60 (15%) (95% CI 11-19) had isolated medial PF OA progression (i.e., medial PF OA progression and no lateral PF OA progression), 120 (30%) (95% CI 25-35) had isolated lateral PF OA progression, and 41 (10%) (95% CI 7-14) had progression in both the medial and lateral PF compartments.

Figure 2 shows the baseline alignment status of knees with isolated medial PF OA progression and isolated lateral PF OA progression. Thirty-nine knees (65%) with isolated medial PF OA progression were varus at baseline, 18 (30%) were valgus, and 3 (5%) were neutral. Of knees with isolated lateral PF OA progression, 58 (48%) were valgus, 54 (45%) were varus, and 8 (7%) were neutral. The distribution of alignment in knees with any medial PF OA progression and in knees with any lateral PF OA progression was similar to that observed in knees with isolated progression. Of the 41 knees progressing in both medial and lateral PF compartments, 26 were varus, 12 valgus, and 3 neutral.

To test the hypotheses regarding the compartment-specific effect of alignment on PF OA progression, we calculated the OR for PF OA progression associated with a given baseline alignment. Table 1 shows the OR for medial PF OA progression associated with baseline varus alignment. Varus (versus nonvarus) alignment significantly increased the odds of medial PF OA progression, after adjusting for age, sex, and BMI, whether the outcome was isolated medial PF OA progression (adjusted OR 1.85) or any medial PF OA progression (adjusted OR 1.98). Table 2 shows the OR for lateral PF OA progression associated with valgus alignment. After adjustment, valgus (versus nonvalgus) alignment was associated with an increased likelihood of isolated lateral PF OA progression only.

We also examined the relationship between severity of malalignment as a continuous variable and medial and lateral PF OA progression, adjusting for age, sex, and BMI. Severity of varus was significantly associated with both isolated (adjusted OR [per 3° increment in varus] 1.22, 95% CI 1.03-1.45) and any (adjusted OR [per 3° increment in varus] 1.28, 95% CI 1.08-1.51) medial PF OA progression. Severity of valgus was not associated with an increase in the adjusted OR for any lateral PF OA progression. The relationship between valgus severity and isolated lateral PF OA progression approached significance (P = 0.10).

We explored, in separate analyses of the varus knees and the valgus knees, the frequency with which the same sides of the PF and tibiofemoral compartments progressed together (Table 3). Synchronous medial progression was not common. Of 211 varus knees at risk for progression in both compartments, 65 progressed in the medial PF compartment and 50 progressed in the medial tibiofemoral compartment, but only 8 of these progressed in both compartments. In fact, in varus knees, synchronous medial progression appeared to be less common than medial tibiofemoral/lateral PF OA progression (Table 3). Of 158 valgus knees at risk, 70 progressed in the lateral PF compartment, 24 progressed in the lateral tibiofemoral compartment, but only 5 progressed in both. Synchronous lateral progression was no more common than was lateral PF/medial tibiofemoral progression. In all of the knees considered together, lateral PF progression was more common than medial PF progression, and medial tibiofemoral progression was more common than lateral tibiofemoral progression.

DISCUSSION

Our results showed that lateral PF OA progression was more common than medial PF OA progression. Varus alignment at baseline was associated with a 2-fold increase in the likelihood of medial PF OA progression, whether the outcome was isolated medial PF OA progression or any medial PF OA progression. Valgus alignment was associated with an increased likelihood of isolated lateral PF OA progression but was not associated with any lateral PF OA progression.

The finding that lateral PF OA progression was twice as common as medial PF OA progression possibly reflects the presence of mechanical factors that inherently stress the lateral PF compartment even in healthy knees, and is consistent with prior cross-sectional studies suggesting that lateral PF OA is more common than medial PF OA ([5][17]). The patella lengthens the lever arm of the quadriceps muscle in achieving extension or resisting flexion, allows a wider distribution of compressive stress on the femur, acts as a guide for the quadriceps tendon to centralize the input for the 4 aspects of the quadriceps muscle, contributes to the capsular tension of the knee, and serves as a bony shield to protect the trochlear and condylar cartilage ([18][19]). However, with each quadriceps contraction, the patella is drawn laterally.

Varus alignment (as either a dichotomous or a continuous variable) increased the likelihood of medial PF OA progression, theoretically through an increase in the Q-angle, which increases stress to the medial PF compartment. This change in the Q-angle (in vitro) resulted in 2 types of contact-pattern change: a medial shift of the contact area, with unloading of the lateral facet, or a separation of contact areas into distinct lateral and medial regions ([4]). Of note, in our study, knees with isolated medial PF OA progression were more often varus at baseline than valgus. Varus alignment was linked to PF OA progression even when a bicompartmental pattern was included. In theory, bicompartmental PF OA progression in the setting of varus may reflect the varus stress on the medial PF, coupled with inherent stresses on the lateral PF not specifically related to alignment.

In contrast, knees with lateral PF OA appeared to be varus at baseline as often as they appeared to be valgus. Valgus alignment did increase the likelihood of isolated lateral PF OA but did not increase the likelihood of any lateral PF OA. The weaker association between valgus alignment and lateral PF OA progression than between varus and medial PF OA progression may relate to less severe valgus (i.e., mean ± SD 3.9 ± 2.9°) than varus (5.0 ± 3.7°) in this sample. Alternatively, lateral PF OA progression may hinge upon factors other than valgus alignment that stress the lateral PF compartment, such as patellar laxity, subluxation, or dislocation. By predisposing to recurrent patellar dislocation, impairments such as medial PF ligament injury, patella alta, and lateral femoral condyle dysplasia may contribute to lateral PF OA development and/or progression. These factors need to be examined further in epidemiologic and natural history studies focusing on PF OA.

This study illustrates that the prevalence of PF involvement in a knee OA cohort can be substantial. Participants were recruited from a variety of community sources, applying the established and widely used definition of knee OA requiring definite osteophyte presence in a tibiofemoral compartment. Of the 230 participants who completed the 2 visits, 50 had severe PF OA in at least 1 knee.

There is a paucity of longitudinal information on PF OA. Knowledge of PF OA has been based on several cross-sectional studies, which provided evidence that several factors are associated with the presence of this condition, including age ([20]), sex ([20]), knee injury ([20][21]), Heberden's nodes ([21]), BMI ([20][22]), elite weight lifting ([23]), and elite running or tennis ([24]). In a clinic-based study, 28% of individuals with isolated PF OA, but none with combined PF/tibiofemoral OA, had a history of patellar dislocation or subluxation ([5]). There was no evidence of joint hypermobility syndrome in patients with PF OA ([5]).

We previously observed (in this cohort) that varus alignment increased the odds of medial tibiofemoral OA progression (age-, sex-, and BMI-adjusted OR 4.09, 95% CI 2.20-7.62), and that valgus alignment increased the odds of lateral tibiofemoral OA progression (age-, sex-, and BMI-adjusted OR 4.89, 95% CI 2.13-11.20) ([7]). These results, in conjunction with the current results, suggest that alignment influences OA progression in both tibiofemoral and PF compartments. The mechanical effect of alignment is different in the 2 major compartments, which perhaps explains the difference in the magnitude of the ORs for tibiofemoral versus PF OA progression; the tibiofemoral compartments appear to be more vulnerable to the effect of alignment. It is not known whether the standing skyline view is as sensitive in detecting PF OA progression as is the semiflexed fluoroscopically confirmed anteroposterior knee radiograph in detecting tibiofemoral OA progression; if the skyline view were less sensitive, alignment may appear to have a more potent effect on tibiofemoral OA progression. To date, in longitudinal studies, varus-valgus alignment is the only factor identified that influences the risk of progression in both of the major knee compartments.

During the 18-month study period, OA progression in both the medial PF and medial tibiofemoral compartments (or in both the lateral PF and lateral tibiofemoral compartments) in the same knee was not common. These results are consistent with the past report of no correlation between the severity of tibiofemoral and PF OA within the same knee ([5]) and the recent finding of no correlation between patellar and tibial cartilage volume change in the same knee over 2 years ([25]). The current results support the concept that the mechanism of the alignment effect on OA progression is different in the PF compartment than in the tibiofemoral compartment. Among all of the knees, the lateral PF and the medial tibiofemoral compartments were most vulnerable to progression, perhaps reflecting the inherently greater stress under which these compartments are required to function, even under healthy conditions.

Limitations of this study include the inability to more formally examine the frequency of isolated and combined compartment progression among all varus knees and all valgus knees, due to a relatively small sample size. It would have been of interest to look at the Q-angle as well as other factors that increase stress on the lateral PF compartment; the relationship between such factors and the risk of incident and progressive PF OA should be examined in future studies.

In theory, intervention to lessen the force imbalances created by varus-valgus malalignment may delay OA progression in both the tibiofemoral and PF compartments. Interventions have been developed to target malalignment-imposed stresses in knee OA, including wedge insole orthoses. Both biomechanical and clinical studies of these interventions have focused on the tibiofemoral compartment. Their effect at the PF compartment should be examined. The results of the current study also support the need to develop new interventions that specifically target malalignment-imposed stresses on the osteoarthritic PF compartment.

In conclusion, we found evidence that lateral PF OA progression was more common than medial PF OA progression. Varus-valgus alignment influenced the likelihood of PF OA progression in a compartment-specific manner. Varus alignment at baseline was associated with a 2-fold increase in the odds of medial PF OA progression, whether the outcome was isolated medial PF OA progression or any medial PF OA progression. Valgus alignment was associated with a 1.6-fold increase in the odds of isolated lateral PF OA progression.