Biochemical markers of bone turnover are often measured in patients treated with antiresorptive agents to monitor the effects of therapy. In order for a change in these markers to clearly indicate treatment effect, the change in the markers must exceed the amount of spontaneous variation typically seen with no treatment. Based on the measured long-term variability of markers in untreated patients, we defined a minimum significant change (MSC), that is, a change that was sufficiently large that it was unlikely to be due to spontaneous variability. We also examined the changes in markers of bone turnover in subjects treated with pamidronate to see how often observed changes in turnover after treatment exceeded the MSC. We found that urinary markers of bone resorption are best measured on 2-hour fasting samples, because results on random urine showed poor precision and less decline with therapy. We also found that of all the markers, urinary N-telopeptide cross-links (NTX) had the greatest decline after therapy (58%), although it also had the highest long-term variability (29.5%). The marker that most often showed a decline with treatment that exceeded the MSC was serum bone-specific alkaline phosphatase where 74% of observed changes exceeded the MSC. Other markers that often showed a decline with treatment that exceeded the MSC were 2-hour fasting urine NTX and free deoxypyridinoline, where 57% and 48%, respectively, of changes in therapy exceeded the MSC. The ideal marker would combine the large decline after treatment characteristic of NTX (60-70%) with the good precision of bone-specific alkaline phosphatase.