Background/Purpose Obesity and metabolic disorders increase the risk of hand osteoarthritis (HOA).  This study aimed to determine i) clinical and radiographic features associated with HOA symptoms and structural severity ii) factors associated with HOA progression.  

Methods This is an ancillary study from an international 3-year, randomized, placebo-controlled phase III trial designed to evaluate strontium ranelate on the X-ray progression of knee OA (SEKOIA trial). A clinical assessment at baseline and at 3 years was performed.  Hand radiographs were scored by 2 reproducible readers (ICCs >0.8) for Kellgren-Lawrence (KL) and Verbruggen anatomical phase (Verb) scores. We included subjects with radiographic HOA defined by at least 2 joints with KL≥2. Symptoms were assessed using the overall Australian/Canadian (AUSCAN) score (normalized at 300) and the criterion hand pain and Functional index for HOA (FIHOA) ≥5. Radiographic HOA severity at baseline was assessed by global KL and Verb scorings. The longitudinal analysis was performed on the placebo group only to avoid potential biases due to a treatment effect. The clinical and radiographic progressions were defined as the changes of AUSCAN, KL or Verb scores between baseline and endpoint. Baseline age, gender, body mass index (BMI), clinical features and cardio-metabolic parameters were included in a multivariate linear or logistic regression model.

Results At baseline, 869 subjects (72 % women) with mean ± SD age of 64 ± 7 years and BMI 29.6 ±4.7 kg/m² had radiographic HOA. Multivariate analyses indicated that AUSCAN level was associated with menopause (p<0.0005), depression (p<0.01), a history of ischemic cardiopathy (p<0.05) and radiographic severity (p<0.0005) (Table). A similar association was found for “FIHOA ≥ 5 and pain” criterion (p<0.05 for all analyses). Accumulation of metabolic factors (hypertension, dyslipidemia, diabetes mellitus and obesity (BMI ≥30 kg/m²) was associated with the AUSCAN score (β=5.5 [0.3;10.9]; p<0.05). Radiographic severity was associated with age, obesity, menopause as well as knee KL grade after adjustment on confounders (Table). Similar associations were found for Verb score (p<0.05 for all analyzes). 307 HOA patients from the placebo group were followed for a mean duration of 31.5 ± 8.5 months. Obesity independently predicted Verb score variation (β=-1.3 [-2.2;-0.45]; p<0.01). Baseline KL score was the main factor associated with radiographic progression (KL or Verb). Baseline AUSCAN score was the main factor associated with subsequent clinical progression (β=-0.39 [-0.05;-0.03]; p<0.0001).

Conclusion In HOA, obesity and ischemic cardiopathy are associated with structural damages and symptoms respectively. Such results further delineate the metabolic OA phenotype.

Table. Factors associated with baseline HOA symptoms and radiographic severity. NT: not tested, NS: not significant