SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. SUBJECTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. ROLE OF THE STUDY SPONSOR
  9. REFERENCES

Objective

To determine if self-reported current and early adult life knee malalignment is associated with knee chondrocalcinosis (CC).

Methods

A case–control study embedded in the Genetics of Osteoarthritis and Lifestyle (GOAL) study was performed. A validated self-reported line-drawing instrument was mailed to 3,022 participants of GOAL, inquiring about knee malalignment (straight, valgus, varus) in their third decade (20s) and currently. Self-reported weight in 20s and present weight and height measured at the initial visit were used to calculate body mass index (BMI) in 20s and currently. Occupational risk was present if participants self-reported mechanically demanding activities in their longest held job. Cases were participants with radiographic CC at any tibiofemoral (TF) joints. Osteoarthritis was defined as a Kellgren/Lawrence score of ≥3 at TF joints or patellofemoral joints. Odds ratio (OR), adjusted OR, and 95% confidence interval (95% CI) were calculated for associations with knee CC.

Results

A total of 2,167 participants responded to the questionnaire, of which 7.5% respondents had knee CC. Knee CC was associated with self-reported varus (adjusted OR 1.77 [95% CI 1.05, 2.98]) and any knee malalignment in 20s (adjusted OR 1.64 [95% CI 1.02, 2.64]). This association was not restricted to the mechanically loaded TF compartment. There was no association between current knee malalignment and knee CC. Age (adjusted OR 1.73 [95% CI 1.36, 2.19]) and knee OA (adjusted OR 2.88 [95% CI 1.89, 4.38]) were associated with knee CC. There was no association between BMI in 20s (adjusted OR 1.14 [95% CI 0.92, 1.42]), occupational risk (adjusted OR 1.32 [95% CI 0.94, 1.85]), and knee CC.

Conclusion

Early life knee malalignment, predominantly varus, is an independent risk factor for knee CC. This association is not restricted to the mechanically loaded compartment, implying a generalized predisposition throughout the knee.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. SUBJECTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. ROLE OF THE STUDY SPONSOR
  9. REFERENCES

Chondrocalcinosis (CC), the calcification of fibro- and hyaline articular cartilage, most commonly results from calcium pyrophosphate (CCP) crystal deposition (1). CCP crystal deposition may be an asymptomatic incidental finding on imaging (CC), or may be associated with acute CCP crystal arthritis, osteoarthritis (OA), and/or chronic CCP crystal inflammatory arthritis (2).

CCP crystals form extracellularly, and raised extracellular pyrophosphate (PPi) is the key metabolic precursor to their formation (3). This appears to be a local metabolic abnormality specific to synovial fluid and presumably to the joint cartilage, as the PPi concentration is normal in both plasma and urine in people with CC (3). Both fibro- and hyaline articular chondrocytes secrete PPi (4) and are possible contributors to intraarticular PPi levels. Mechanical loading of a joint could elevate intraarticular PPi levels. This is as mechanically loaded chondrocytes release more ATP to the extracellular compartment (5), which may then be metabolized to PPi by NTPPPH ectoenzyme plasma cell glycoprotein I (6).

Mechanical loading at the knee, the most common site of CC, is influenced by body weight, frontal plane knee malalignment, and joint use (7–9). Although body mass index (BMI) does not appear to associate with CC (10, 11), the association between current frontal plane knee malalignment and CC has been examined in 2 small cross-sectional studies only, and with conflicting results (12, 13). The association between knee malalignment in early adult life, or joint use, and knee CC has not been examined before. However, both of these are risk factors for knee OA (14, 15).

In order to clarify if mechanical loading at the knee is associated with knee CC, we undertook a case–control study with retrospective questioning concerning self-reported knee malalignment and BMI, both current and in the third decade of life (20s), and exposure to mechanically demanding occupational activities, and examined their association with knee CC. We hypothesized that self-reported knee malalignment, higher BMI, and presence of occupational risk may predispose to knee CC. Therefore, the objectives of this study were to examine 1) whether self-reported current knee malalignment and knee malalignment in 20s is associated with the risk of knee CC, and 2) the possible interactions of knee malalignment with BMI, occupational risk, and knee OA on predisposition to CC.

Significance & Innovations

  • This is one of the few studies to systematically examine the risk factors for knee chondrocalcinosis (CC).

  • It is the largest study to examine the association between knee malalignment and knee CC.

  • It demonstrates that knee malalignment in 20s and not current knee malalignment is associated with knee CC, and that this association is independent of confounding factors, including age, sex, body mass index, knee osteoarthritis, and occupational knee loading.

SUBJECTS AND METHODS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. SUBJECTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. ROLE OF THE STUDY SPONSOR
  9. REFERENCES

Study design.

A case–control study embedded in the Genetics of Osteoarthritis and Lifestyle (GOAL) study was performed. Recruitment for the GOAL study occurred between 2002 and 2006. GOAL comprises 3,170 individuals: 1,007 with clinically severe hip OA, 1,042 with clinically severe knee OA, and 1,121 without knee or hip OA (15). A followup questionnaire with a validated self-reported instrument for knee malalignment was mailed to patients who were alive and available in 2008 (16). The details of GOAL and the followup questionnaire are available elsewhere (15).

Radiographs.

All participants had posteroanterior weight-bearing knee radiographs using the SynaFlexer positioning frame (Synarc) and skyline views of the patellofemoral (PF) joints. All knee radiographs were graded for the presence or absence of CC (linear calcification in the fibro- or hyaline articular cartilage) at the medial and lateral compartments of each tibiofemoral (TF) joint, and for changes of OA at the TF joints and PF joints. All radiographs were assessed for CC and OA changes by a senior research metrologist (SD). Frontal-plane knee malalignment (anatomic axis) in degrees corrected to the second decimal place was measured on knee radiographs of all GOAL participants using a validated method (17), as part of ongoing work on GOAL radiographs. The intraclass coefficient for intrarater reliability was 0.99 (95% confidence interval [95% CI] 0.97, 1.00).

Participants.

Cases were participants with radiographic CC in either knee, and controls were participants without radiographic CC in both knees. The case definition was independent of the presence of knee OA, or that of CC at other joints (hands, hips, symphysis pubis).

Exposures.

Self-reported frontal plane knee alignment.

Information about current and early adult life (20s) self-reported frontal plane knee malalignment was collected using a validated line diagram instrument in the followup GOAL questionnaire (15). This line diagram instrument has good patient and intraobserver reproducibility (κ = 0.73 and 0.89, respectively) (16). Using this instrument, participants separately self-reported their current and early adult life knee malalignment as severe varus, mild varus, straight legs, mild valgus, or severe valgus. Participants with severe or mild varus were considered to have varus knee malalignment. Similarly, those with severe or mild valgus were considered to have valgus knee malalignment.

In the GOAL study, participants self-reported a single grade of knee malalignment that was derived from and applied to both of their knees. In order to establish that the single self-reported knee malalignment grade derived from both knees is a valid measure of knee malalignment at each knee, we examined the symmetry of radiographic knee malalignment in participants in the GOAL control group with no hip or knee OA. For this, we selected all GOAL control participants with a Kellgren/Lawrence (K/L) score of 0 at both the TF joints and hips and without a history of knee surgery or injury. We assumed that in the absence of confounding effects of OA, surgery, injury, and Paget's disease (an exclusion criterion for recruitment into the GOAL study), the controls with normal lower extremities would demonstrate their constitutional knee alignment and that this would be symmetric, as would be expected in people in their 20s.

Other covariates.

Information about age and sex was collected at the initial visit. Current height and weight were measured in centimeters and kilograms, respectively. All participants self-reported their body weight in their 20s in kilograms. Self-reported body weight in 20s and current height were used to calculate BMI in 20s (kg/m2). Current height and weight were used to calculate current BMI.

Occupational risk factors for the job held for the longest time were scored from 0–5 for each of the following: squatting for ≥1 hour/week, kneeling for ≥1 hour/week, performing heavy work while standing for ≥1 hour/week, lifting 25 kg ≥10 times/week, and lifting 50–100 kg ≥1 time/week. Participants were divided into those with a score of ≥1 and those without any occupational risks.

TF joints and PF joints were allocated a K/L score for each knee. Radiographic knee OA was defined if the K/L score was ≥3 at either TF joint or PF joint on any side (18).

Statistical analysis.

A paired t-test was used to examine the symmetry of the anatomic axis between the right and left knees in the GOAL control group. Cases with CC were compared with controls without CC. An independent-sample t-test was used to compare the mean and the chi-square test was used to compare frequencies between groups. Odds ratio (OR) and 95% CI were calculated for all associations. Binary logistic regression was used to adjust for confounding factors such as age, sex, and current BMI or BMI in 20s, as appropriate (model 1). Further adjustment was made for knee OA and occupational risk if there was an association between risk factors and knee CC in model 1 (model 2). Ordinal logistic regression was used to examine the association between number of knees (0–2) with CC (as a dependent variable) and self-reported knee malalignment, adjusted for other covariates as in model 1, and model 2.

The GOAL study is a case–control study originally designed to investigate risk factors for large joint OA. Two-thirds of participants have knee or hip OA, which may cause confounding due to oversampling of participants with OA. In order to reduce this confounding, we examined positive associations between knee malalignment and knee CC in the GOAL population stratified by their knee OA status (knee OA present: K/L score ≥3 in any TF joint or PF joint; knee OA absent: K/L score <3 at TF joints and PF joints on both sides).

We also examined any associations between knee malalignment and knee CC in a subset of the GOAL population standardized with the age and sex prevalence of TF joint OA (K/L grade of ≥2) from the Zoetermeer survey (19). The Zoetermeer survey was a population-based radiographic survey of more than 6,500 participants. All participants ages >44 years (n = 2,969, 52.4% women, age range 45 to >80 years) had weight-bearing anteroposterior knee radiographs. The prevalence of K/L grade ≥2 TF joint OA was 18.5% and 18.1% on the right and left sides, respectively (19). For this, GOAL participants with TF joint OA (K/L grade ≥2) were classified according to their sex and age in 5-year intervals (age 45–49 years, etc.). The prevalence of TF joint OA for each age and sex stratum was obtained from the Zoetermeer survey, the requisite number of cases with TF joint OA were randomly selected from each age and sex stratum using SPSS, and their data were merged with those without TF joint OA to yield a subset of the GOAL population standardized to the Zoetermeer study for prevalence of knee OA (19).

Analyses for interactions were only carried out if there was an association between self-reported current or early adult life knee malalignment and knee CC. Interaction was examined using stratified analysis and logistic regression models. Self-reported current or early adult life knee malalignment was entered in pairs with either 1) BMI at the same age (less than, or greater than or equal to the mean BMI), 2) occupational risk (present, absent), or 3) knee OA (present, absent) in the logistic regression models. The adjusted OR for interaction (ORint) and 95% CI were estimated after adjusting for other risk factors (models 1 and 2). Due to a small number of participants with valgus knee malalignment, analysis of interaction was only carried out for any knee malalignment versus straight knees. The level of statistical significance was set at P values less than or equal to 0.05. All statistical analysis was carried out using SPSS, version 14.

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. SUBJECTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. ROLE OF THE STUDY SPONSOR
  9. REFERENCES

Demographics.

Of the 3,070 participants (250 [8.1%] with knee CC), 2,167 were included in the analysis (Figure 1). The mean ± SD age and BMI were 66.55 ± 7.61 years and 23.10 ± 3.53 kg/m2, respectively. There were 1,057 women (48.8%). Radiographic knee OA was present in 1,104 participants (50.8%). Self-reported varus, valgus, and straight knee alignment in 20s was present in 147 (7.4%), 46 (2.3%), and 1,794 (90.3%) participants, respectively. Similarly, self-reported current varus, valgus, and straight knee alignment was present in 172 (8.0%), 145 (6.8%), and 1,825 (85.2%) participants, respectively. Knee CC was present in 162 cases and absent in 2,005 controls. Knee CC was bilateral in 53.7% of cases.

thumbnail image

Figure 1. Number of participants at each stage of the study. GOAL = Genetics of Osteoarthritis and Lifestyle; CC = chondrocalcinosis; +ve = positive; -ve = negative.

Download figure to PowerPoint

Symmetry of knee malalignment.

In the 600 GOAL control participants with no structural radiographic changes at either knee or hip, the mean ± SD anatomic axis was 182.80° ± 2.01° at the right knee and 182.84° ± 2.00° at the left knee, respectively. The mean difference in anatomic axis angle between the right and left knees was −0.03 (95% CI −0.14°, 0.09°; P = 0.61).

Risk factors for CC.

Cases with knee CC were older and were more likely to be men, have knee OA, report varus knee alignment in their 20s, have high BMI in their 20s, and have an occupational risk (Table 1). Both varus and any knee malalignment in 20s was associated with knee CC after adjusting for other risk factors (Table 2). There was no association between self-reported current knee malalignment, current BMI, and knee CC. Higher BMI in 20s was associated with knee CC after adjusting for age and sex. However, after adjusting for knee OA, any knee malalignment in 20s, and occupational risk, there was no convincing association between BMI in 20s and knee CC (adjusted OR 1.14 [95% CI 0.92, 1.42]; model 2). The presence of occupational risk was not associated with knee CC after adjusting for age, sex, and BMI.

Table 1. Demographic and mechanical risk factors for knee chondrocalcinosis (CC)*
 Case (CC positive), n = 162Control (CC negative), n = 2,005P
  • *

    Means and percentages calculated using the number of participants for whom data were available for that variable.

Age, mean ± SD years70.59 ± 5.6366.22 ± 7.65< 0.001
Women, no. (%)66 (40.70)991 (49.40)0.033
Body mass index in 20s, mean ± SD kg/m223.60 ± 3.5522.99 ± 3.530.036
Current body mass index, mean ± SD kg/m228.62 ± 4.2629.26 ± 5.210.130
Knee osteoarthritis, no. (%)122 (75.30)979 (48.80)< 0.001
Any knee malalignment in 20s, no. (%)26 (17.70)166 (9.00)0.001
Varus knee in 20s, no. (%)22 (15.40)124 (6.9)< 0.001
Valgus knee in 20s, no. (%)4 (3.20)42 (2.50)0.607
Any knee malalignment currently, no. (%)26 (16.70)289 (14.60)0.481
Varus knee currently, no. (%)18 (12.20)153 (8.3)0.106
Valgus knee currently, no. (%)8 (5.80)136 (7.40)0.475
Occupational risk, no. (%)79 (48.80)819 (40.80)0.049
Table 2. Multivariate analysis of risk factors for knee chondrocalcinosis*
 Case, no.Control, no.Crude OR (95% CI)Adjusted
Adjusted OR (95% CI)Adjusted OR (95% CI)
  • *

    OR = odds ratio; 95% CI = 95% confidence interval; BMI = body mass index.

  • Adjusted for age, sex, BMI in 20s, and current BMI for current knee malalignment.

  • Full adjustment undertaken for all variables listed where any malalignment was used for knee malalignment. Age and BMI were converted to tertiles: sex was 1 = female, 0 = male, and knee OA; varus, valgus, and any knee alignment in 20s; and occupational risk were 1 = present, 0 = absent.

  • §

    Model 1 nonsignificant.

Age, years     
 Tertile 1: ≤6318688111
 Tertile 2: 64–70606533.51 (2.05–6.01)3.43 (2.00–5.88)2.46 (1.41–4.30)
 Tertile 3: ≥71846644.84 (2.87–8.13)4.61 (2.74–7.78)3.65 (2.11–6.31)
Sex     
 Male961,014111
 Female669910.70 (0.51–0.97)0.78 (0.56–1.09)§
Straight knee in 20s1211,669111
 Any malalignment in 20s261662.16 (1.37–3.40)2.10 (1.32–3.35)1.64 (1.02–2.64)
 Varus knee in 20s221242.45 (1.50–3.99)2.33 (1.40–3.88)1.77 (1.05–2.98)
 Valgus knee in 20s4421.31 (0.46–3.72)1.29 (0.44–3.74)§
Straight knee currently1301,692111
 Any malalignment currently262891.17 (0.76–1.82)1.17 (0.75–1.82)§
 Varus knee currently181531.53 (0.91–2.58)1.37 (0.81–2.34)§
 Valgus knee currently81360.77 (0.37–1.60)0.85 (0.40–1.80)§
BMI in 20s, kg/m2     
 Tertile 1: ≤21.4535676111
 Tertile 2: 21.46–23.78636461.89 (1.23–2.89)1.80 (1.16–2.78)1.70 (1.07–2.69)
 Tertile 3: ≥23.80626481.85 (1.21–2.84)1.65 (1.06–2.57)1.34 (0.84–2.16)
BMI currently, kg/m2     
 Tertile 1: ≤26.5059665111
 Tertile 2: 25.51–30.82546820.89 (0.61–1.31)0.84 (0.57–1.24)§
 Tertile 3: ≥30.83496580.84 (0.57–1.24)0.91 (0.61–1.37)§
Knee osteoarthritis     
 Absent401,026111
 Present1229793.20 (2.21–4.62)2.65 (1.81–3.88)2.88 (1.89–4.38)
Occupational risk     
 Absent831,186111
 Present798191.38 (1.00–1.90)1.32 (0.94–1.85)§

As there was no association between current knee malalignment and knee CC, further analyses were restricted to self-reported knee malalignment in 20s. Varus or any knee malalignment in 20s was associated with increasing number of knees with CC. The adjusted OR (model 2) for having 1 more knee with CC was 1.19 (95% CI 0.41, 3.41) for valgus, 1.87 (95% CI 1.12, 3.12) for varus, and 1.73 (95% CI 1.08, 2.77) for any knee malalignment in 20s. Self-reported any knee malalignment in 20s was associated with bilateral knee CC compared to unilateral knee CC (adjusted OR 2.66 [95% CI 1.02, 6.97]; model 2). There was a similar association between varus knee malalignment in 20s and bilateral knee CC (adjusted OR 2.50 [95% CI 0.87, 7.19]; model 2), although it did not reach statistical significance.

The association between knee malalignment in 20s and knee CC was examined in the GOAL population stratified according to their knee OA status. The adjusted OR for associations between varus and any knee malalignment in 20s and knee CC were comparable for both strata. The associations were statistically significant only in the strata with knee OA (Table 3).

Table 3. Association between knee malalignment in 20s and knee chondrocalcinosis stratified by the knee OA status*
 Crude OR (95% CI)Adjusted
Adjusted OR (95% CI)Adjusted OR (95% CI)
  • *

    Knee osteoarthritis (OA) is regarded as present if the Kellgren/Lawrence score was ≥3 in any tibiofemoral or patellofemoral joint, or is otherwise absent. OR = odds ratio; 95% CI = 95% confidence interval.

  • Adjusted for age, sex, and body mass index (BMI) in 20s.

  • Full adjustment undertaken for all variables.

  • §

    No cases with chondrocalcinosis. Age and BMI were converted to tertiles: sex was 1 = female, 0 = male, and knee OA; varus, valgus, and any knee alignment in 20s; and occupational risk were 1 = present, 0 = absent.

Knee OA negative   
 Straight knee in 20s111
 Any malalignment in 20s1.19 (0.28–5.11)1.56 (0.35–6.86)1.57 (0.35–6.92)
 Varus knee in 20s1.82 (0.42–7.93)2.21 (0.49–9.98)2.21 (0.49–9.99)
 Valgus knee in 20s§§§
Knee OA positive   
 Straight knee in 20s111
 Any malalignment in 20s1.75 (1.07–2.86)1.68 (1.01–2.77)1.67 (1.00–2.76)
 Varus knee in 20s1.85 (1.09–3.15)1.74 (1.00–3.02)1.73 (0.99–3.00)
 Valgus knee in 20s1.36 (0.46–4.01)1.30 (0.44–3.88)1.30 (0.44–3.87)

The association between knee malalignment in 20s and knee CC was examined in the participants (n = 1,573) with an age- and sex-standardized prevalence of TF joint OA according to the Zoetermeer survey. There was a statistically significant association between self-reported varus knee malalignment in 20s and knee CC (OR 2.81 [95% CI 1.43, 5.34] and adjusted OR 2.92 [95% CI 1.44, 5.91], model 1; adjusted OR 2.12 [95% CI 1.03, 4.40], model 2) and between any knee malalignment in 20s and knee CC (OR 2.57 [95% CI 1.40, 4.73] and adjusted OR 2.77 [95% CI 1.48, 5.20], model 1; adjusted OR 2.11 [95% CI 1.10, 4.03], model 2). However, the association between self-reported valgus knee malalignment in 20s and knee CC was not statistically significant (OR 1.96 [95% CI 0.58, 6.61] and adjusted OR 2.22 [95% CI 0.63, 7.87], model 1; adjusted OR 1.96 [95% CI 0.54, 7.10], model 2). This may be due to a small number of cases with valgus knee malalignment in 20s.

There was no interaction between knee malalignment in 20s and either occupational risk (ORint 1.41 [95% CI 0.56, 3.55] and adjusted ORint 1.41 [95% CI 0.55, 3.63], model 1; adjusted ORint 1.32 [95% CI 0.51, 3.41], model 2), BMI in 20s (ORint 0.93 [95% CI 0.38, 2.31] and adjusted ORint 0.89 [95% CI 0.36, 2.25], model 1; adjusted ORint 0.91 [95% CI 0.38, 3.45], model 2), or knee OA (ORint 1.47 [95% CI 0.32, 6.86] and adjusted ORint 1.18 [95% CI 0.25, 5.58], model 1; adjusted ORint 1.17 [95% CI 0.25, 5.52], model 2) as risk factors for knee CC.

Compartment-specific association with risk factors.

CC was present in the medial and lateral TF joint compartment in 99 (75.0%) and 122 (92.2%) participants at the right knee, respectively, and in 78 (69.0%) and 102 (90.3%) participants at the left knee, respectively (P < 0.001 for each knee). On crude analysis, varus knee malalignment in 20s was associated with CC in both medial and lateral tibiofemoral compartments (Table 4). However, this association was not significant for the medial tibiofemoral compartment of the left knee (OR 1.95 [95% CI 0.94, 4.01]). Also, on adjusting for all risk factors (model 2), this association was lost for the medial compartment of the right knee (adjusted OR 1.74 [95% CI 0.91, 3.33). This analysis was restricted to varus versus straight knees, as a small number of participants self-reported valgus knee malalignment in their 20s.

Table 4. Knee malalignment in 20s and risk of CC in medial and lateral tibiofemoral joint compartment*
Knee alignmentCCCrude OR (95% CI)Adjusted
PresentAbsentAdjusted OR (95% CI)Adjusted OR (95% CI)
  • *

    CC = chondrocalcinosis; OR = odds ratio; 95% CI = 95% confidence interval.

  • Adjusted for age, sex, and body mass index (BMI) in 20s.

  • Full adjustment undertaken for all variables. Age and BMI were converted to tertiles: sex was 1 = female, 0 = male, and knee osteoarthritis; varus, valgus, and any knee alignment in 20s; and occupational risk were 1 = present, 0 = absent.

Right     
 Medial     
  Straight731,7041.001.001.00
  Varus131322.30 (1.24–4.26)2.09 (1.11–3.94)1.74 (0.91–3.33)
 Lateral     
  Straight891,6881.001.001.00
  Varus191262.86 (1.69–4.85)2.61 (1.52–4.51)2.11 (1.20–3.68)
Left     
 Medial     
  Straight581,7171.001.001.00
  Varus91371.95 (0.94–4.01)1.93 (0.92–4.07)1.63 (0.76–3.49)
 Lateral     
  Straight731,7021.001.001.00
  Varus171293.07 (1.76–5.37)3.08 (1.72–5.52)2.32 (1.28–4.23)

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. SUBJECTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. ROLE OF THE STUDY SPONSOR
  9. REFERENCES

This study reports that self-reported varus knee malalignment in 20s is associated with increased risk of subsequent knee CC, and that this risk is not limited to the mechanically loaded medial or lateral TF joint compartment. The association between self-reported varus knee malalignment in 20s and knee CC was independent of age, sex, BMI in 20s, knee OA, and occupational risks. Any knee malalignment in 20s associated with CC in both knees was supporting a constitutional morphologic risk since such malalignment is inevitably symmetric. We found a “dose-response” relationship between any knee malalignment in 20s and knee CC.

In addition, this study confirms that knee CC is an age-related condition with a strong “dose-response” relationship, i.e., the older the age, the greater the risk (10, 11, 20). Furthermore, this study also confirmed previous reports of association between knee CC and knee OA (10, 11, 20). However, there was no association between current knee malalignment and knee CC, suggesting that knee malalignment is less likely to be a concurrent condition or a consequence of the knee CC. There was no convincing association between BMI and knee CC, especially after adjusting for knee OA, suggesting that the association between BMI and knee CC may be confounded by knee OA (Table 2). Biomechanically demanding occupational risks were not associated with knee CC and there was no interaction between knee malalignment in 20s and occupational risk, BMI in 20s, or knee OA.

Constitutional knee alignment is observed to be symmetric in routine clinical practice. Therefore, in the GOAL study, we elicited only one self-reported grade of knee malalignment that was derived from and applied to both knees. The symmetry that we demonstrated in older GOAL control participants supports this decision to use a single assessment of alignment rather than separate self-reported assessments for each knee.

The association between early adult life knee malalignment and knee CC has not been studied before. Two small cross-sectional studies give conflicting results about the association between current knee malalignment and knee CC. In a survey of 58 ambulatory active residents of a home for the elderly (ages >70 years), varus knee malalignment (>10°) was associated with CC (predominantly at the knee) (12). There was no difference in the prevalence of clinical knee arthropathy between the groups (12). However, a subsequent case–control study of 45 CC cases and 23 OA controls did not find an association between current knee malalignment and knee CC (13). Our findings agree with the latter study.

Knee CC was significantly more common in the lateral compartment of the TF joint than in the medial compartment of the TF joint. This concurs with previous observations (21, 22) and findings from the Progetto Veneto Anziani study (11). A previous study from Nottingham, UK, also reported a higher prevalence of CC in the lateral TF joint compartment than the in the medial TF joint compartment (10). Varus knee malalignment in the third decade was associated with lateral TF joint compartment CC, and not with medial TF joint compartment CC, after adjusting for confounders. This is contrary to the pattern observed for OA where varus malalignment predisposes to medial compartment OA (15, 23). However, the lack of a compartment-specific association between early adult life knee malalignment and subsequent knee CC may be explained if both the ATP and subsequent extracellular PPi produced in increased amounts due to mechanically-induced chondrocyte “stress” can diffuse throughout the joint and not stay just at the site of production.

We did not find an association between self-reported valgus knee alignment in 20s and knee CC. This may be due to the small number of participants (n = 46) who reported valgus knee malalignment in their 20s. We also failed to show an association between knee loading due to occupational use and knee CC on multivariate analysis. This may be due to a smaller contribution of occupational usage on knee loading. There is a weaker association between occupational joint use and knee OA than between self-reported knee malalignment in 20s and knee OA (15).

There was a statistically significant association between knee malalignment in 20s and knee CC in GOAL participants with knee OA, but not in GOAL participants without knee OA (Table 3). The lack of significance in the latter subset may be due to a low prevalence of CC and knee malalignment. In those with knee OA, knee CC, any knee malalignment in 20s, and varus knee malalignment in 20s were present in 122 (11.1%), 142 (14.2%), and 113 (11.6%) participants, respectively. On the other hand, in those without knee OA, knee CC, any knee malalignment in 20s, and varus knee malalignment in 20s was only present in 40 (3.8%), 51 (5.2%), and 34 (3.5%) participants, respectively. The subgroup analysis with the prevalence of TF joint OA standardized to the Zoetermeer survey confirmed that early life varus knee malalignment is indeed an independent risk factor for the development of knee CC, and that this is likely to be observed in general population.

There are several caveats to this study. First, this is a case–control study executed by reconstituting the study groups of GOAL, which was primarily designed to identify risk factors for severe large joint OA. Therefore, this sample does not represent the general population. However, the association between knee malalignment in 20s and knee CC persisted in GOAL participants standardized to the Zoetermeer study for prevalence of TF joint OA (19), suggesting a true association. Moreover, we confirmed the well-established risk factors for CC, such as age and OA, suggesting face validity for this study (11, 20). Second, as there is a high prevalence of knee OA and related risk factors, the result may be confounded. However, we adjusted for knee OA and related risk factors to minimize any bias. The adjusted ORs for associations between knee malalignment in 20s and knee CC were similar across the 2 strata according to knee OA status, suggesting a homogeneous effect. Third, there were only 162 knee CC cases, which reduces study power. Moreover, information about knee malalignment in 20s, weight in 20s, and occupational exposure were self-reported and potentially subject to inaccuracy and recall bias. Finally, we used person-based rather than joint-based analysis due to the nature of the data. Because of such caveats, it is desirable for this observation to be confirmed in other studies using joint-based analysis.

In conclusion, early life varus knee malalignment is associated with knee CC. This is not restricted to the mechanically loaded TF compartment, and is independent of age, knee OA, and other risk factors.

AUTHOR CONTRIBUTIONS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. SUBJECTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. ROLE OF THE STUDY SPONSOR
  9. REFERENCES

All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. Dr. Abhishek had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study conception and design. Abhishek, Sally Doherty, Muir, Zhang, Michael Doherty.

Acquisition of data. Sally Doherty, Muir, Michael Doherty.

Analysis and interpretation of data. Abhishek, Maciewicz, Zhang, Michael Doherty.

ROLE OF THE STUDY SPONSOR

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. SUBJECTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. ROLE OF THE STUDY SPONSOR
  9. REFERENCES

AstraZeneca funded the collection of the Genetics of Osteoarthritis and Lifestyle (GOAL) database as a collaborative scientific project, with Dr. Rose A. Maciewicz the lead scientist for AstraZeneca on this project. The company was not involved in the design of the present study and did not need to give approval of the manuscript content. Publication of the article was not contingent on the approval of AstraZeneca. Dr. Maciewicz, however, continues in the role of an individual collaborator and author of projects involving GOAL.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. SUBJECTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. ROLE OF THE STUDY SPONSOR
  9. REFERENCES