Background/Purpose: Previous studies on the effects of weight loss and exercise on bone mass show conflicting results and there are no studies of older adults with knee OA. We examined change in bone mineral density (BMD) in response to weight loss, with and without exercise, and to exercise alone  in older adults enrolled in the IDEA trial and determined whether BMD changes were related to the magnitude of fat mass loss and to changes in leptin.

Methods: 454 overweight and obese (avg BMI 34) adults (avg age 66yrs, 72% female) with symptomatic knee OA were randomized to 18 months of intensive dietary induced weight loss (D);  intensive weight loss plus exercise (D+E); or exercise control (E). The weight loss goal for the two diet groups was ≥ 10% of baseline body weight. The exercise intervention consisted of low to moderate intensity walking and resistance training 3 d/wk for 1 hr/d. A total of 399 (88%) participants completed the study. BMD at the hip (total hip=thip and femoral neck=FN) and spine, as well as total body fat mass, were measured by DXA and vitamin D levels (baseline only) were assessed in 392 subjects. Osteoporosis (OP) and osteopenia (OPE) were defined as location-specific t-scores <-2.5 and between -2.5 and -1, respectively.  A subset of 162 subjects had serum leptin levels measured. Treatment group means were compared using ANCOVA, adjusted for baseline dependent variable values, BMI, and gender. Corrections for multiple comparisons used Tukey's method. Linear associations of change measures used a general model, adjusted for baseline values, BMI, gender, and treatment arm.

Results: Mean weight loss was: D+E, 11.4%; D, 9.5%; E, 2.2%. BMD declined at the hip but not the spine in both weight loss groups (D= -0.024g/cm² thip; -.013 FN; both p<.0001 and D+E =-0.020 thip; -.012 FN; both p<.0001) but not the E group (-0.002;p=0.54 thip; -0.001; p=0.8 FN).  At baseline 10 subjects (2.5%) had OP (4 hip and 6 spine) and 45% had OPE in at least 1 site.  Following intervention, 9 subjects had OP and no subjects with OPE became OP. The mean serum vitamin D level at baseline was 29 ng/mL and baseline BMD values were not related to vitamin D levels.  The D and D+E groups but not the E group had significant declines in fat mass and leptin. Changes in hip (but not spine) BMD correlated positively with changes in fat mass (b=.003 g/cm²/kg,p<.0001) (Fig1) and changes in leptin(b=.001ng/ml/kg,p<.0001).

Conclusion:   To our knowledge, this is the longest study to date that analyzed change in BMD with intentional weight loss and the first in older adults with knee OA. Weight loss and the associated reduction in fat mass resulted in a decrease in total hip and femoral neck, but not spine, BMD. The addition of exercise did not prevent these declines. The bone loss was not severe enough to result in osteoporosis but the findings suggest patients with low BMD should be monitored when starting a weight loss interevention.