In our previous double-blind trial, we reported that clodronate reduced the incidence of bone lesions, fractures, pain and hypercalcaemia in multiple myeloma. Recently, it has been assumed that the antiresorptive effect of bisphosphonates on the osteoclasts is mediated through the osteoblasts. We therefore determined, in 244 patients of the same trial, serum assays of aminoterminal propeptide of type I procollagen (PINP) and type I collagen degradation product (ICTP). PINP is an early synthesis product of proliferating osteoblasts, in comparison to the alkaline phosphatase (AP) which is secreted by differentiated osteoblasts during the maturation phase of collagen. ICTP circulates in serum when old bone is resorbed. Our results indicate that after 25 months, the PINP levels decreased in the clodronate group (from 68.9 +/- 4.4 micrograms/l to 37.2 +/- 3.5 micrograms/l; P < 0.001) but not in the control group (from 61.5 +/- 3.2 micrograms/l to 69.3 +/- 7.5 micrograms/l; P < NS). The fall in the ICTP levels was markedly steeper in the patients receiving clodronate (from 8.38 +/- 0.80 micrograms/l to 4.58 +/- 0.32 micrograms/l; P < 0.01) than placebo (from 7.84 +/- 0.53 micrograms/l to 6.45 +/- 0.95 micrograms/l; P = NS). A significant difference between the study groups was seen at 4 months in the PINP, at 7 months in the ICTP and at 13 months in the AP levels, suggesting that clodronate affected through the proliferating osteoblasts, the osteoclasts, and through the osteoclasts, the differentiated osteoblasts. High baseline ICTP, PINP and AP levels indicated a poor prognosis. The decrease of the markers by clodronate was more marked in survivors than in non-survivors.