Estrogen stimulates osteoblastic collagen production in vitro, but whether the same stimulation takes place in vivo is still unknown. To test the stimulatory effects of a combined estrogen-gestagen regimen in vivo we monitored serum levels of the carboxy-terminal propeptide of human type I procollagen (S-PICP) in a group of 12 osteoporotic women over a 150 week treatment period. Spinal bone mineral content (BMC) increased to a maximum of 5% over pretreatment values around week 90. Serum alkaline phosphatase (S-AP) and serum bone gla protein (S-BGP) both fell from initial values of 220 U/liter and 39 ng/ml, respectively, to 146 U/liter (p less than 0.01) and 27.2 ng/ml (NS) around week 60 and remained reduced over the remaining treatment period. S-PICP also fell from 117 to 68 micrograms/liter at week 60 and 70 micrograms/ml at week 150 (P less than 0.01). This is equal to a reduction to 32 +/- 10% pretreatment levels. The reduction in S-PICP was not significantly different from that of the other two markers of bone formation (S-AP and S-BGP). Thus, provided the metabolic clearance of PICP remains unaltered after hormone replacement therapy, no major stimulation of osteoblastic collagen type I synthesis was demonstrable during estrogen-gestagen treatment in this population of osteoporotic women. The changes in bone markers seen in this study are therefore consistent with an estrogen-mediated reduction in the frequency of remodeling activation. Because of the reduction in bone turnover and methodologic limitations of bone marker assays, however, smaller increases in the amount of bone formed per activation could remain undetectable.