Background/Purpose:  Individuals with a greater medial to lateral tibial periarticular bone mineral density (M:L paBMD) and static malalignment have greater risk for osteoarthritis (OA) progression.  As these risk factors are inter-related, we aimed to evaluate whether M:L paBMD is associated with OA progression, independent of static alignment, which would suggest bone as a therapeutic target.

Methods:  This was a study nested within two ancillary studies to the Osteoarthritis Initiative (OAI); therefore the visit for dual x-ray absorptiometry (DXA) was the OAI 30 or 36 month visit, for static alignment was mostly the 12 month visit, and for OA progression (longitudinal medial joint space narrowing (mJSN)) between the 24 and 48 month visits.  Knee DXAs generated M:L paBMD using proprietary software.  Hip-knee-ankle (HKA) angles (static alignment) were measured on long limb films.  mJSN (OARSI grade) were assessed using PA semi-flexed radiographs.  Knees with 24-month mJSN of 3 were excluded as they were not eligible for mJSN worsening.

Only evaluating one knee per person, usually the right knee unless contraindicated, we generated Pearson’s correlations between HKA and M:L paBMD.  We also performed logistic regression with predictors of M:L pa BMD quartiles and outcomes of worsening of mJSN between the 24 and 48 month visits (including within OARSI grade worsening). Models were adjusted for HKA quartiles, age, sex, and BMI.  The Cochrane Armitage Test assessed for statistical significance of trends.  We performed a sensitivity analysis of only neutrally aligned knees, defined as HKA between -2° and 2°.

To further evaluate the influence of M:L paBMD, eliminating the influence of HKA, we used the residual of M:L paBMD regressed on HKA, representing the uncorrelated aspect of M:L paBMD unrelated to static alignment in logistic regression models.

Results: 419 participants were included, with a mean age of 63.8 +9.2 years; BMI of 29.4 +4.8 kg/m²; 52% male.  At the OAI 24 month visit, 60 (14%), 68 (16%), 158 (38%), 122 (29%), and 11 (3%) had Kellgren and Lawrence grades of 0 – 4 respectively.  45% of knees were neutrally aligned.

M:L paBMD and HKA were correlated; R = 0.64 (p<0.0001); greater M:L paBMD was associated with greater varus alignment.  Those with greater M:L paBMD had greater odds of mJSN progression, even in fully adjusted models (Table).  Sensitivity analyses of only neutrally aligned knees were similar.  Analyses evaluating M:L paBMD uncorrelated with HKA also supported a relationship with mJSN progression.

Conclusion: Higher relative paBMD is associated with greater structural progression, even after adjustment for static alignment, a known correlate of periarticular BMD and predictor of progression. These findings suggest that treatments targeting periarticular bone, not focused solely on static realignment, may ultimately be protective of OA progression. 

ACR Meeting Abstracts - https://acrabstracts.org/abstract/periarticular-bone-mineral-density-predicts-structural-progression-of-knee-osteoarthritis-independently-of-static-alignment/