Differential Effects of Low-Weight, Amenorrhea and Exercise on Bone in Adolescent Girls

Presentation Number: SAT 298
Date of Presentation: April 1st, 2017

Nurgun Kandemir*1, Meghan Slattery2, Kathryn E Ackerman3, Shreya Tulsiani2, Vibha Singhal1, Anne Klibanski4 and Madhusmita Misra1
1Massachusetts General Hospital/Harvard Medical School, Boston, MA, 2Massachusetts General Hospital, Boston, MA, 3Boston Children's Hospital/ Massachusetts General Hospital and Harvard Medical School, 4Massachusetts General Hospital and Harvard Medical School, Boston, MA


Background: Whereas low body weight and hypogonadism are deleterious to bone, mechanical loading has positive effects. We have reported low bone mineral density (BMD), impaired bone structure, and increased fracture risk in low-weight, amenorrheic girls with anorexia nervosa (AN) and normal-weight, oligo-amenorrheic athletes (OA) engaged in weight-bearing sports. However, effects of body weight vs. gonadal status vs. mechanical loading on site and compartment specific bone parameters remain to be clarified.

Objective:To evaluate effects of body weight, gonadal status and repetitive weight-bearing exercise on bone parameters, and describe high risk groups for fracture.

Methods: 391 females 14-21.9 years old were included, 215 with AN, 85 OA and 91 normal-weight eumenorrheic controls (C). The AN group was further classified according to duration of amenorrhea (< 1 year or ≥ 1 year). Areal BMD of the whole body less head (WBLH), lumbar spine and total hip was determined using DXA. BMD Z-scores were calculated for sex, age and race using reference curves from the BMD in Childhood Study, and adjusted for height Z-scores. Fracture history (stress and non-stress) was obtained. Volumetric BMD (vBMD), bone geometry and structure were assessed at the distal radius and tibia using high resolution peripheral quantitative CT (HR-pQCT).

Results: Groups did not differ for age and height Z-scores; BMI was lower in AN than OA and C (p<0.0001). AN had lower WBLH and hip BMD Z-scores than OA and C regardless of amenorrhea duration (p<0.0001 for all). AN (both groups) and OA had lower spine BMD Z-scores than C (p<0.0001, p<0.01). However, AN with ≥ 1 y of amenorrhea had lower spine BMD Z-scores than OA (p=0.02), whereas AN with <1 y of amenorrhea did not differ from OA. OA did not differ from C for WBLH and hip BMD Z-scores. AN and OA had higher fracture rates than C (p=0.03, p=0.009). Prevalence of stress fracture was higher in OA than AN and C (p<0.0001 for both); AN had more non-stress fractures than OA (p=0.02). HR-pQCT of the non-weight-bearing radius showed lower cortical area and thickness and lower total vBMD in AN and OA vs. C (p≤ 0.02 for all); cortical vBMD was lower in OA than C (p=0.02). At the weight-bearing tibia, AN had lower measures for most tibial parameters than OA and C; OA did not differ from C, except for lower cortical vBMD (p=0.0005).

Conclusion: Low weight and amenorrhea (in AN) are deleterious to bone at all sites and both bone compartments. Normal-weight OA have lower spine BMD and lower cortical vBMD at weight-bearing and non-weight-bearing sites compared to C (indicative of effects of oligo-amenorrhea), while most other bone parameters do not differ (indicative of the compensating effect of normal-weight and/or mechanical loading). Yet, the stress fracture rate is higher in OA than AN and C, indicating that bone measures in OA need to be stronger than in C to avoid fractures from weight-bearing activity.


Nothing to Disclose: NK, MS, KEA, ST, VS, AK, MM