Monitoring individual response to hormone replacement therapy with bone markers

Hormone replacement therapy (HRT) induces a rapid decrease in biochemical markers of bone turnover that correlate with a subsequent increase in bone mineral density (BMD). To determine the utility of bone markers in the management of postmenopausal women receiving HRT, we analyzed the relationship between changes in four markers (serum osteocalcin and bone alkaline phosphatase [BAP], serum and urinary C-telopeptide of type I collagen [CTX]) and changes in spine BMD in 569 women treated for 2 years with different doses of a matrix transdermal 17beta-estradiol patch in two placebo-controlled trials. Using a logistic regression model, we found that both the percent change from baseline and the actual value of resorption markers at 3 and 6 months of treatment were predictive of BMD response at 2 years. Comparable results were obtained with formation markers at 6 months only. We determined the sensitivity, probably of positive BMD response, and corresponding cutoff value of markers at 3 and 6 months with a specificity set at a level of 0.90, so that <10% of women classified with markers as responders, i.e., as having a subsequent increase in BMD at 2 years >/=2.26%, would be false positive. All markers provided a high probability of positive BMD response ranging from 0.82 to 0.91, with a sensitivity higher for resorption than for formation markers, and sometimes improved in a model combining the percent change and the actual value of marker under HRT. For example, a decrease in serum CTX >/= 33% at 3 months of HRT provided a 68% sensitivity and 87% probability of positive BMD response at 2 years for a 90% specificity. At 6 months, a decrease in urinary CTX >/= 53% provided a 68% sensitivity and 91% probability of a positive BMD response for a 90% specificity. Half of false-negative cases at 3 months will be correctly identified by a subsequent urinary CTX measurement at 6 months. We conclude that the short-term change in bone markers reflects long-term changes of BMD in postmenopausal women treated with HRT. Our data suggest that bone turnover markers can be used to monitor the BMD response to HRT at the individual level. Whether such monitoring could improve long-term compliance to HRT should be tested prospectively.