Bone loss after bariatric surgery: discordant results between DXA and QCT bone density

Abstract

Several studies, using dual-energy X-ray absorptiometry (DXA), have reported substantial bone loss after bariatric surgery. However, profound weight loss may cause artifactual changes in DXA areal bone mineral density (aBMD) results. Assessment of volumetric bone mineral density (vBMD) by quantitative computed tomography (QCT) may be less susceptible to such artifacts. We assessed changes in BMD of the lumbar spine and proximal femur prospectively for 1 year using DXA and QCT in 30 morbidly obese adults undergoing Roux-en-Y gastric bypass surgery and 20 obese nonsurgical controls. At 1 year, subjects who underwent gastric bypass surgery lost 37 ± 2 kg compared with 3 ± 2 kg lost in the nonsurgical controls (p < 0.0001). Spine BMD declined more in the surgical group than in the nonsurgical group whether assessed by DXA (-3.3 versus -1.1%, p = 0.034) or by QCT (-3.4 versus 0.2%, p = 0.010). Total hip and femoral neck aBMD declined significantly in the surgical group when assessed by DXA (-8.9 versus -1.1%, p < 0.0001 for the total hip and -6.1 versus -2.0%, p = 0.002 for the femoral neck), but no changes in hip vBMD were noted using QCT. Within the surgical group, serum P1NP and CTX levels increased by 82% ± 10% and by 220% ± 22%, respectively, by 6 months and remained elevated over 12 months (p < 0.0001 for all). Serum calcium, vitamin D, and PTH levels remained stable in both groups. We conclude that moderate vertebral bone loss occurs in the first year after gastric bypass surgery. However, striking declines in DXA aBMD at the proximal femur were not confirmed with QCT vBMD measurements. These discordant results suggest that artifacts induced by large changes in body weight after bariatric surgery affect DXA and/or QCT measurements of bone, particularly at the hip.

Trial registration: ClinicalTrials.gov NCT01098942.

Keywords: BONE METABOLISM; DUAL-ENERGY X-RAY ABSORPTIOMETRY; GASTRIC BYPASS; OSTEOPOROSIS; QUANTITATIVE COMPUTED TOMOGRAPHY; SOFT TISSUE ARTIFACT.

Figure 1. Change in weight over 1 year (mean ± SEM)

Weight is presented for the gastric bypass surgical (solid line) and non-surgical (dotted line) groups. * indicates between-group difference by longitudinal regression p-value < 0.05. # indicates within-group difference by paired t-test p-value < 0.05.

Figure 2. Percent change in lumbar spine BMD over 1 year by DXA and QCT (mean ± SEM)

Percent change in DXA lumbar spine aBMD (panel A), QCT integral spine vBMD (panel B), and QCT trabecular spine vBMD (panel C) are shown for the gastric bypass surgical (solid line) and non-surgical (dotted line) groups. * indicates between-group difference by longitudinal regression p-value < 0.05. # indicates within-group difference by paired t-test p-value < 0.05.

Figure 3. Percent change in total hip BMD over 1 year by DXA and QCT (mean ± SEM)

Percent change in DXA total hip aBMD (panel A), QCT integral total hip vBMD (panel B), and QCT trabecular total hip vBMD (panel C) are shown for the gastric bypass surgical (black bars) and non-surgical (grey bars) groups. * indicates between-group difference by longitudinal regression p-value < 0.05. # indicates within-group difference by paired t-test p-value < 0.05.

Figure 4. Percent change in femoral neck BMD over 1 year by DXA and QCT (mean ± SEM)

Percent change in DXA femoral neck aBMD (panel A), QCT integral femoral neck vBMD (panel B), and QCT trabecular femoral neck vBMD (panel C) are shown for the gastric bypass surgical (black bars) and non-surgical (grey bars) groups. * indicates between-group difference by longitudinal regression p-value < 0.05. # indicates within-group difference by paired t-test p-value < 0.05.

Figure 5. Percent change in bone turnover markers over 1 year (mean ± SEM)

Percent change in bone turnover markers serum P1NP (panel A), serum osteocalcin (panel B), and serum CTX (panel C) are shown for the gastric bypass surgical (solid line) and non-surgical (dotted line) groups. * indicates between-group difference by longitudinal regression p-value < 0.05. # indicates within-group difference by paired t-test p-value < 0.05.

Figure 6. Change in metabolic bone labs over 1 year (mean ± SEM)

Serum calcium (panel A), serum 25-hydroxyvitamin D (panel B), and serum parathyroid hormone (panel C) are presented for the gastric bypass surgical (solid line) and non-surgical (dotted line) groups. * indicates between-group difference by longitudinal regression p-value < 0.05. # indicates within-group difference by paired t-test p-value < 0.05.