We found that one year of intentional weight loss in overweight and obese adults with type 2 diabetes resulted in greater bone loss at the total hip and femoral neck, but not at the spine or whole body. To our knowledge, this is the first trial to assess the effects of an intentional weight loss intervention on bone in adults with type 2 diabetes.
Studies in broader populations have reported bone loss at the total hip, spine and whole body with weight loss (11
). In obese patients undergoing bariatric surgery, bone loss appears to occur in proportion to the degree of weight loss (13
). Bone loss with weight loss is likely due to less skeletal loading and reduced anabolic signals from muscle and fat mass (14
). In observational studies, lower weight and intentional weight loss increase fracture risk (17
). This relationship between lower weight and fracture is strong for BMI<25, but is weaker with higher BMI (17
). In a small trial of a lifestyle intervention in obese older adults without diabetes, the intervention group had increased bone loss at the hip but not the spine or whole body, similar to our findings (19
). Spine BMD changes are difficult to assess in older adults because of osteophytes and other degenerative changes. Whole body BMD has a higher proportion of cortical bone than hip BMD and may not be as sensitive to changes in weight.
A limitation of this study is the use of DXA to assess BMD changes in the setting of weight change, particularly in obese participants. With obesity, the reproducibility of BMD is reduced which is likely to attenuate any real associations between changes in weight and BMD (20
). Changes in fat mass can affect the ability of DXA to identify bone edges and to accurately allow for the composition of soft tissue overlying bone, possibly introducing artificial changes in BMD (21
). In a recent study using a Hologic densitometer, fat mass added to volunteers reduced the measured mean spine BMD but did not alter the results for mean hip BMD (22
). In addition, with substantial weight loss, smaller hip circumference reduces the height of the bone above the densitometer table which may introduce fan beam magnification effects (23
). The net effect of changes in fat mass on bone measurements vary with skeletal site and scanner model. However, in our analyses considering decreases in fat mass and in lean mass separately, we found that both were associated with total hip bone loss, suggesting that the observed association between weight loss and bone loss in this trial is not simply an artifact of changes in fat mass.
The ILI group experienced greater weight loss as well as improved physical fitness and glycemic control compared with the DSE group. Look AHEAD was not designed to identify whether these changes had separate effects on outcomes. However, it is plausible that weight loss results in bone loss while improved glycemic control and physical fitness have a positive effect on BMD. When we sought evidence for independent effects by considering the intervention group separately, we found that weight loss was associated with total hip bone loss in the intervention group even after adjustment for changes in HbA1c and physical fitness. TZD use, associated with bone loss and fracture in other studies (24
), was greater in the DSE group during the first year and thus does not account for the differences in hip bone loss.
Improved HbA1c was associated with preservation of total hip BMD in men but not women. The relationship between glycemic control and changes in BMD has not been clearly established. In an uncontrolled trial, Gregorio et al. reported increased femoral neck bone mineral content in 50 type 2 diabetic adults with poor glycemic control after a year of treatment that improved control (3
). However, the investigators did not report on changes in weight. In Look AHEAD, most participants had good glycemic control at study entry (mean A1C ~ 7.2%) which may have limited any potential skeletal benefits of improved control.
Our analyses in the intervention group suggest that improved physical fitness during weight loss did not preserve bone. Physical activity was only measured on a subset of Look AHEAD participants using the Paffenbarger scale, but changes in physical activity over the first year of the trial correlated positively (R = 0.25, p<0.0001) with changes in physical fitness (9
). Results from studies of the effects of exercise during weight loss on bone loss have been conflicting (25
). One study in older type 2 diabetic adults reported that high-intensity resistance training reduced the effects of weight loss on total body BMD (4
). The exercise intervention in Look AHEAD was not as intensive and did not emphasize resistance training which may account for the different findings.
Bone loss with intentional weight loss is primarily a concern because of the strong relationship between low BMD and fracture risk. However, the additional bone loss experienced in the ILI group averaged a modest 1%. Thus, the initial year of weight loss in Look AHEAD may not result in a clinically important increase in fracture risk, especially as subsequent years will focus on weight maintenance rather than additional weight loss. On the other hand, effects of weight loss on bone density might be greater in subsequent years as any skeletal remodeling response may lag behind the weight loss stimulus. This report is limited to one year of follow-up. For comparison, in a U.S. cohort of older women, an annual loss of 0.9% total hip BMD over 8 years, a cumulative loss of about 7%, was associated with a relative risk of subsequent hip fracture of 1.29 (28
). Fracture events are not currently available for Look AHEAD, but are being collected and will be assessed at the end of the trial.
This study is the first to address the effects of a weight loss intervention on bone in a diabetic population. A strength of this study is the randomization of participants to the weight loss intervention. Thus, weight loss in the ILI group was generally intentional and not due to illness or frailty. A limitation of this study is the use of DXA in the setting of weight change. In addition, data were not collected on calcium or vitamin D supplement use. It is possible that the two groups used different amounts of these supplements. Another limitation is the lack of bone turnover markers to assess effects on bone resorption. These results may not apply to patients weighing more than 300 lbs, the weight limit for DXA measurements, or to patients with poor glycemic control.
In conclusion, one year of an intensive lifestyle intervention in overweight and obese men and women with type 2 diabetes that resulted in weight loss, improved glycemic control and enhanced physical fitness, also resulted in a modest increased bone loss at the hip. The extent to which this additional bone loss may increase fracture risk is not known. Further research, including fracture outcomes in Look AHEAD, is needed to elucidate the effects of intentional weight loss on fracture risk in older adults with type 2 diabetes.