Endocrine Abstracts

Introduction: One of the most common endocrine complications of anorexia nervosa (AN) is the decrease in bone mineral density. The authors evaluated the predictive factors of osteopenia and osteoporosis in AN patients admitted with low weight.

Patients and methods: Retrospective analysis of 45 patients admitted with AN between 2001 and 2015 in the Endocrinology department, corresponding to 63 admissions. Bone mineral density was classified according to WHO criteria. The relationship between clinical and analytical parameters and bone density was evaluated.

Results: Most patients were female (93.3%; n=42), with mean age 20.6±7.7 years. Restrictive subtype in 86.7% (n=39), with the remainder of the purgative type. On admission, mean BMI 14.3±1.6 kg/m² and percentage of fat mass 4.2±2.4%. Amenorrhea was present in 48.9% (n=22), euthyroid sick syndrome in 15.6% (n=7), hypogonadotrophic hypogonadism in 51.1% (n=23), Bone densitometry documented osteopenia and osteoporosis in 57.1 and 31%, respectively. Initial weight was statistically different between the densitometry results (normal, osteopenia and osteoporosis; P=0.04), without association with initial fat mass (IFM), measurements of FSH, LH, estradiol, total testosterone, TSH and fT4. Lumbar spine T-score was moderately correlated with initial weight (ρ=0.58), IFM (ρ=0.39), initial fat-free mass (IFFM) (ρ=0.55) and estradiol levels (ρ=0.37). Total femur T-score was moderately correlated with initial weight (ρ=0.48) and IFFM (ρ=0.47).

Discussion and conclusions: In this group of patients, low weight on admission was associated with osteopenia and osteoposoris. Initial weight was positively correlated with better results of densitometry in lumbar spine and total femur. The greater number of correlations with lumbar spine bone loss is in agreement with the earlier changes in the trabecular bone described in the literature. The low number of patients with normal densitometry (n=5) may have limited the association with other factors, as well as the statistic strength of the presented results.