Among pathologies prevalent in western societies, anorexia nervosa has increased over the last decade. Its effects on bone mass need to be defined, and prognostic factors, either clinical or biochemical, could aid clinicians in individual patient management. To determine which clinical and/or biochemical parameters could be related to bone mass status in adolescent female anorexia nervosa patients, 73 female patients were classified according to different stages of their illness and studied in terms of clinical and biochemical parameters and bone densitometric mineral content at lumbar spine. Patients (age 17.2 +/- 1.7 y, mean +/- SD) with Tanner pubertal stage 5, regular menstruation for more than 3 mo before the onset of secondary amenorrhea, and diagnosed with anorexia nervosa were consecutively studied and classified in three clinical situations: I) active phase (34 patients): undernourished and amenorrheic; II) weight recovered but still amenorrheic (20 patients); III) fully recovered (19 patients). Clinical data were recorded at the time of bone density measurement, concomitant with blood sample extraction for study of IGF-I, IGF-binding protein 3 (IGFBP-3), IGFBP-1, estradiol, sex hormone-binding globulin, dehydroepiandrosterone sulfate, prealbumin, amino-terminal propeptide of procollagen III, osteocalcin, bone alkaline phosphatase, carboxy-terminal propeptide of procollagen I, amino-terminal propeptide of procollagen I, carboxy-terminal telopeptide of collagen I, 25-OH-vitamin D, 1,25(OH)(2)-vitamin D, and parathormone. In addition, a 24-h urine collection was made for cortisol, GH, deoxypyridinoline, amino-terminal telopeptide of collagen I, and calcium and creatinine content analysis. IGF-I, estradiol, and biochemical bone formation markers were higher and IGFBP-1, sex hormone-binding globulin, and biochemical bone resorption markers were lower in the weight-recovered stages (stages II and III) compared with the active phase (stage I). Bone formation markers correlated positively with body mass index SD score and IGF-I, whereas bone resorption markers correlated negatively with body mass index SD score and estradiol. Although no statistically significant differences regarding lumbar spine bone mineral density SD score values were recorded among the three stages of the illness, the proportion of osteopenic patients was clearly lower among stage III patients. The actual bone mineral density was inversely related to the duration of amenorrhea and directly related to duration of postmenarcheal menses before amenorrhea. In addition, a subset of osteopenic patients (five of 19) in the fully clinically recovered group with accelerated bone turnover was identified. Normal circulating estrogen level exposure time predicts actual bone mineral density at lumbar spine in young adolescent anorexia nervosa patients. In addition to psychiatric and nutritional interventions, estrogen-deprivation periods must be shortened to less than 20 mo. Patients remaining osteopenic at full clinical recovery require additional follow-up studies.