At a median duration of 9.3 years after adolescent MBS, young adults had lower BMD at both weight bearing (TH, FN) and non-weight bearing (distal radius) skeletal sites when compared to young adults of similar age, sex, race and BMI who had not undergone MBS. While DXA BMD Z-scores at the FN and TH did not differ between RYGB vs. VSG participants, BMD Z-scores were lower in those who were farther out from either procedure, indicating a need for ongoing monitoring of skeletal health following adolescent MBS. Percent change in BMI since MBS and in the first year after MBS were not associated with any skeletal outcomes in this cohort, raising the possibility that additional mechanisms affect attained BMD after MBS.
Our findings at the hip were similar to those reported by Nimmala et al. who found that BMD Z-scores at the TH and FN significantly decreased over 12 months in adolescents who underwent VSG relative to non-surgical controls, but no significant change was noted at the WB (28). Mitchell et al. followed this cohort for another 12 months (27) and noted a reduction from baseline in FN BMD by 8.9% and TH BMD by 8.4% at 24 months post-surgery but found no significant change at the WB or 1/3 radius. However, in both studies, differences in BMD between VSG and non-surgical control groups were no longer significant after controlling for change in BMI. In contrast, we did not find that change in BMI since surgery or in the first year after surgery were associated with attained BMD a median of 9.3 years after MSB. The longer time since MSB in our study may have allowed opportunity for other factors to affect bone density. Also, we could not estimate loss of BMD in our cross-sectional study, preventing a direct comparison of findings. Our focus was BMD after adaptation to weight change compared to weight similar controls to assess the ability of bone to adapt to prevailing weight bearing forces.
BMD of trabecular and cortical bone by pQCT were consistent with what was observed with DXA. vBMD of trabecular bone at the 4% distal radius and tibia by pQCT were significantly lower in persons who had undergone MBS compared to those who had not, but there was no difference at the 30% distal radius by pQCT reflecting cortical bone. Using high resolution-pQCT (HR-pQCT), Nimmala et al. (28) and Mitchell et al. (27) also found that trabecular vBMD at the distal radius and distal tibia decreased over 12 months and 24 months in persons who underwent adolescent VSG. Though surgical type did not affect DXA BMD outcomes in our study, trabecular vBMD measured by pQCT was lower in RYGB vs. VSG recipients. PTH was higher in persons who had undergone RYGB than those who had undergone VSG, similar to results from adult MBS patients (11, 35), which may explain trabecular vBMD differences.
The clinical implications of lower trabecular bone vBMD at the distal tibia and radius in young adults who underwent adolescent MBS are unclear. Adults who undergo MBS have a greater risk of major fracture (including hip and distal radius) that increases over time: compared to incident fracture rates 5 years prior to MBS, the incidence rate ratio was 2.77 at < 3.0 years post-surgery and 3.78 at 3.1-5.0 years post-surgery (20). The increased fracture risk was associated with RYGB as opposed to VSG, but data are limited to small number of participants and potentially biased study designs (36, 37). In our cohort, there was no difference in the prevalence of ever experiencing a fracture between persons who had undergone MSB approximately 9 years prior and those that did not. Interestingly, persons who had undergone MBS were more likely to fracture after MSB than before. This may have been a consequence of bone loss after surgery, or it could have reflected increased physical activity after weight loss as physical activity is a predictor of fracture among youth (38, 39), though there were no group differences in reported physical activity. Participants in this study were relatively young and at low risk of hip and vertebral fracture. The greater periosteal circumference at the distal tibia in surgical vs. control groups may be a protective mechanism dispersing loading force on the long bone in the event of a fall.
Trabecular bone occupies most of the skeletal area within the vertebra. While no difference in DXA LS BMD between surgical vs. control groups was noted in our cohort, Huber et al reported declines in LS trabecular vBMD and vertebral body strength using quantitative CT at 24 months following adolescent VSG compared to non-surgical controls (40). This finding may explain the adjusted relative risk of 1.70 for clinical spine fractures following MBS in adults compared to non-obese controls (18). Further research to determine the impact of adolescent MBS on spinal bone health is necessary.
We first reported concerns about impaired bone health following adolescent RYGB in 2011 (23) as we found that WB BMD Z-scores decreased by nearly 1 SD in the first 12 months following surgery with continued though smaller declines between 12 to 24 months after surgery (23, 24). At all timepoints, mean WB BMD Z-scores remained above 0.0. BMI-matched controls were not included. In our current study, despite the lower BMD at the TH, FN, and distal radius in those who underwent MSB compared to controls, the BMD Z-scores for all skeletal sites except the WB were also at or above the median for their age, sex, and race. BMD is well known to increase with BMI as the skeleton adapts to support a greater body weight. It is not clear why WB BMD Z-scores in our current study were below 0.0, especially in the control group. Interestingly, the WB BMD Z-scores were also below 0.0 in the study by Nimmala et al. Percent body fat and visceral adiposity have been inversely associated with WB BMD in adults (41, 42) and adolescents (43, 44). Calibration of DXA machines used in this study and those used to generate the reference ranges should be considered as intermachine calibration differences are greater for the WB scans than for TH and LS scans (45).
About 50% of individuals were deficient in vitamin D at a median of 9.3 years after MBS, similar to what we found in the Teen-LABS cohort 5 years post-MBS (29). The prevalence of vitamin D deficiency was potentially even higher in controls whose vitamin D intake was lower than in MSB cases. We previously reported that the prevalence of elevated PTH rose from baseline to 5 years post-MBS among the RYGB group (29). These findings parallel that found in an adult cohort and suggest that RYGB may be associated with chronic disruptions of calcium homeostasis (15). PTH effects on bone turnover are complex. Intermittent increases in PTH stimulate bone formation, whereas sustained high levels of PTH lead to bone resorption (46). Hypocalcemia is the key stimulus of PTH release. In our study, estimated calcium intake was insufficient in all participants and thus would not explain group differences in BMD. We do not have measures of intestinal calcium absorption or urinary excretion, which would affect prevailing circulating calcium levels and calcium balance. Nonetheless, calcium intake is an important risk factor to address post-MBS.
Our study is limited in that presurgical BMD measurements were not obtained so we could not calculate bone loss. However, we compared the BMD of those with exposure to MBS in adolescence to a control group without such exposure but were of similar BMI. Mean BMD Z-scores were still within normal limits in all groups. However, we estimated a decrease in BMD Z-scores of 0.18 SD per year at the FN. If this rate continues over time, these young adults may have low BMD before age 50. This cohort underwent MBS after the completion of puberty (Tanner Stage IV-V and skeletal maturation of at least 95%). The 2018 American Society for Metabolic and Bariatric Surgery Pediatric Committee recommended removing this requirement for MBS (47). MBS in pre- and early pubertal children, who have not yet experienced peak bone accrual velocity, may have even greater effects on bone density.
In conclusion, young adults a median of 9.3 years after MBS as adolescents, bone density, especially at the FN, was lower than in young adults of similar BMI who had not undergone MSB. BMD Z-scores at the FN decreased with time since MSB and occurred independent of surgical type and percent weight loss. The clinical impact of BMD changes on osteoporotic fracture risk may not be fully appreciated until sixth decade of life.