Follow-up of bone mineral density in 27 cases of anorexia nervosa.

Cortical and trabecular bone loss can lead to osteoporosis in chronic forms of anorexia nervosa (AN). As there is some debate about the reversibility of this condition, we performed a longitudinal follow-up study of 27 cases in which clinical, biological, X-ray and lumbar and femoral neck dual photon absorptiometry examinations were conducted every 6 months for up to 30 months. Three groups were distinguished: G1, untreated amenorrheic AN (N = 14, total follow-up 126 months); G2, effectively treated AN (N = 11, total follow-up 192 months), with two subgroups: fluoride (N = 5) and estrogen (N = 6); and G3, remitting AN with normalization of the gonadic function (N = 2, total follow-up 36 months). Results were adjusted for each patient to a 6-month variation. Semestrial variations in lumbar bone mineral density (BMD) were -2.1 +/- 1.3%, +2.8 +/- 1.5%, and -0.3 +/- 1.3% (mean +/- SEM), respectively for G1, G2 and G3; those for femoral neck BMD semestrial variations were -5.9 +/- 2.1%, -3.8 +/- 1.2% and -1.0 +/- 0.6%. Femoral neck and lumbar BMD variations for G1 were mainly correlated positively with bone-forming markers (serum osteocalcin, alkaline phosphatase) and negatively with initial lumbar BMD. Estrogen alone increased lumbar BMD by +1.4 +/- 2.3% every 6 months but did not stabilize femoral neck BMD (-3.5 +/- 1.4%). Fluoride increased lumbar BMD by 4.8 +/- 1.8%. Both lumbar and femoral neck BMD were stabilized in the remission group (-0.3 +/- 1.3% and -1.0 +/- 0.6%), despite half of the follow-up time with amenorrhea. In conclusion, untreated AN is associated with a marked trabecular and cortical bone loss (4-10% per year), which can lead to osteoporotic fractures. In prevention of bone loss, the efficacy of estrogen is difficult to investigate in AN, even with a well-controlled trial. Our study could provide argument that, when the observance of this preventive treatment is assessed, lumbar BMD can be stabilized in chronic forms of AN.