Objective: To evaluate the effect of annual infusions of zoledronic acid (ZA) with or without a single injection of methylprednisolone, compared to placebo, on quantitative magnetic resonance imaging 3-D bone area and bone shape in participants with symptomatic knee osteoarthritis (OA).

Methods: This was a post-hoc analysis of the ZAP2 trial. Active appearance modelling was used to assess bone area (mm2) and femur bone shape (B-score) in 262 participants (mean 61.8 ± 8.0 years, 51% female) at baseline, 6, and 24 months. Radiographic joint space narrowing (JSN) was measured at baseline. An 'OA shape' was defined as a B-score of >1.96.

Results: At baseline 65% of participants demonstrated an OA shape. Treatment with ZA plus methylprednisolone but not ZA alone, compared to placebo, was associated with significantly slower expansion in bone area at the medial femoral (-33.9 mm2, 95% confidence interval [CI] -61.8 to -6.0) and lateral femoral (-22.0 mm², 95%CI -40.7 to -3.4) compartments over 24 months. B-score increased in all groups, with no significant between-group differences. There were significant interactions of JSN (grade 0 vs grade 1-2) and B-score (≤1.96 vs >1.96) with treatment effect on bone area (p < 0.05), such that ZA plus methylprednisolone slowed the expansion of medial and lateral femoral bone area over 24 months in participants with JSN grade 1-2 or a B-score of >1.96.

Conclusions: ZA plus methylprednisolone may retard expansion of bone area over 24 months, but ZA alone may not. Neither ZA with or without methylprednisolone slowed progression of bone shape over 6 or 24 months.