Significant osteoporosis affects over half of all women with anorexia nervosa (AN). The mechanisms of bone loss in this condition are not known, and estrogen administration alone has not been shown to prevent bone loss. Insulin-like growth factor I (IGF-I), a nutritionally dependent bone trophic hormone, is know to stimulate osteoblast function and collagen synthesis in vivo and in vitro. We hypothesized that short term administration of recombinant human IGF-I (rhIGF-I) would increase bone turnover in young women with AN. We studied 23 women, aged 18-29 yr (mean +/- SD, 23 +/- 4 yr) with AN. Spinal bone density was significantly reduced compared to that in age-matched controls (0.85 +/- 0.11 vs. 1.19 +/- 0.12 g/cm2 by dual energy x-ray absorptiometry; P < 0.001) and was below the normal mean in 54% of the women. Patients were randomized to receive rhIGF-I (100 or 30 micrograms/kg) or placebo sc twice a day for 6 days. Bone turnover was assessed at baseline and after 3 and 6 days of treatment using two markers of bone formation [osteocalcin (OC) and type I procollagen carboxyl-terminal propeptide (PICP)] and three specific markers of bone resorption [pyridinoline (PYRX), deoxypyridinoline (DPYRX), and N-telopeptide (NTX)]. Serum OC was reduced significantly (P < 0.001) in women with AN compared to normal premenopausal women (5.4 +/- 3.8 vs. 8.6 +/- 4.5 ng/mL) and correlated with percent fat mass (r = 0.60;P < 0.01) and body mass index (r = 0.50;P < 0.05). Markers of bone resorption were elevated significantly compared to normal levels [DPYRX, 18.2 +/- 7.0 vs. 11.4 +/- 5.2 nmol/mmol creatinine, (P < 0.001); NTX, 53.5 +/- 22.5 vs. 36.5 +/- 14.6 nmol BCE/mmol creatinine (P < 0.01)]. IGF-I levels were relatively low at baseline compared to those in age-matched controls (203 +/- 93 vs. 262 +/- 84 ng/mL;P < 0.01) and increased to 673 +/- 268 ng/mL [P < 0.05; 100 micrograms/kg twice daily (BID)] and 545 +/- 255 ng/mL (P < 0.05; 30 micrograms/kg BID). During short term administration of rhIGF-I at a dose of 100 micrograms/kg BID, there was a significant (P < 0.05) increase in markers of bone formation, as assessed by both PICP (147 +/- 33 to 303 +/- 187 ng/mL) and OC (5.3 +/- 3.8 to 10.9 +/- 7.4 ng/mL). There was also a significant (P < 0.05) increase in markers of bone resorption as assessed by PYRX (51.0 +/- 16.6 to 87.1 +/- 8.2 nmol/mmol creatinine) and DPYRX (17.3 +/- 4.5 to 26.3 +/- 3.7 nmol/mmol creatinine). The group randomized to receive short term administration of rhIGF-I at a dose of 30 micrograms/kg BID demonstrated a significant (P < 0.05) increase in PICP (110.9 +/- 47.0 to 134.8 +/- 43.2 ng/mL) and an insignificant increase in OC levels (4.5 +/- 3.2 to 6.8 +/- 5.9 ng/mL). However, markers of bone resorption were unchanged during rhIGF-I administration at this dose. Serum PTH and serum and urinary calcium were unchanged in both treatment groups compared to placebo levels. These data demonstrate that young women with anorexia nervosa have decreased markers of bone formation and increased bone resorption. This is the first demonstration that short term rhIGF-I administration increases markers of bone turnover in severely osteopenic women with AN. The effects of short term rhIGF-I on bone turnover are dose dependent. At a dose of 100 micrograms BID, rhIGF-I administration significantly stimulated both markers of bone formation and bone resorption. At a dose of rhIGF-I of 30 micrograms BID, there was an increase in one marker of bone formation, PICP, without a change in markers of bone resorption. Further studies are required to determine whether chronic administration of rhIGF-I can affect bone mass in young women with profound osteopenia due to anorexia nervosa.