Debate about the use of fluoride for the treatment of vertebral osteoporosis has centered not only on whether fluoride treatment decreases vertebral fractures, but also the interindividual vertebral bone mineral density (BMD) response, the potential for nonvertebral fractures, as well as side effects and tolerability. These effects may be dose dependent and, in this study, we examine the pharmacokinetics of sodium monofluorophosphate (MFP) in osteoporotic patients and relate this to changes in BMD. Plasma fluoride absorption curves were measured from 0 to 6 h after ingestion of MFP at baseline and during long-term dosing in 21 patients with vertebral osteoporosis (T scores < or = 2). BMD was measured at baseline and at 12 months at the lumbar spine (LS), femoral neck (FN), trochanter, and Ward's triangle. We found that fluoride elimination was inversely related to creatinine clearance. LS BMD increased from a median of 0.77 g/cm2 (range 0.69 to 0.99) at baseline to 0.88 g/cm2 (0.75 to 1.13) (p < 0.001) after 12 months. This equates to a median increase of 12% (range -1.2 to 37). Median femoral neck BMD decreased from 0.75 g/cm2 (0.62 to 0.94) at baseline to 0.69 g/cm2 (0.62 to 0.92) (p = 0.13) after 12 months. This equates to a decrease of -2% (-19 to 10). BMD at the other hip sites also decreased slightly. Changes in LS and FN BMD were not significantly related (r = 0.28, p = 0.29). The various pharmacokinetic parameters measured were not related to changes in LS BMD; however, there was an inverse relationship between trough fluoride concentration during long-term dosing and change in FN BMD. Further studies are required to see if this relationship can be used to monitor osteoporotic patients treated with fluoride and prevent significant decreases in FN BMD and possibly fractures at this site.