Goal-directed therapy in osteoporosis

Abstract

Themanagement of osteoporosis is changing—for the better. Perhaps the most progressive recent step has been the assessment of fracture risk using clinical risk factors through tools such as the Fracture Risk Assessment Tool (FRAX). In addition to traditional intervention thresholds, such as those for osteoporosis defined by the WHO T-score or prior fracture, many guidelines now include recommendations for dual-energy X-ray absorptiometry (DXA) assessment or treatment for subgroups of patients identified to be at certain levels of fracture risk. In this regard, osteoporosis management is starting to follow risk-based strategies that have already been adopted in other disease areas, particularly in cardiovascular disease. In this issue’s Perspective article, Cummings and colleagues suggest that osteoporosis management should also embrace an additional strategy, that of goal-directed therapy. Although there is certainly merit in consideration and discussion of this proposal, the authors recognize that there are many challenges to be overcome. We would strongly argue that these challenges cannot be easily dismissed or glossed over in the desire to move forward—but we welcome the future discussions, debates, and most importantly, the research that will determine whether osteoporosis lends itself to goal-directed therapy. Themajor limitation in developing a goal-directed approach in osteoporosis is the choice of the parameter to define the goal. Cummings and colleagues imply that bone turnover markers (BTMs), bone mineral density (BMD), or fracture risk might be suitable parameters, but there remain many uncertainties and there is an ongoing debate about the optimal marker(s) of response in osteoporosis. Goal-directed therapy can only be considered when we are confident that we can accurately and consistently detect a therapeutic response and can demonstrate a strong correlation between that response and an improvement in clinical outcome. Such requirements are met in the fields of hypertension, hypercholesterolemia, diabetes, and critical care, in which treatment can be titrated to the desired level of effect. For example, in the treatment of hypercholesterolemia there is a strong relationship between the reduction in total cholesterol (TC) or low-density lipoprotein (LDL)-cholesterol (LDL-C) and mortality rates; in one meta-analysis, for every 1-mmol/L decrease in TC or LDL-C, there was a 24.5% or 28% reduction in coronary heart disease (CHD)-related mortality, respectively. Can we say the same about markers of response in osteoporosis? In the Perspective, a proposal is made that the goal could be a certain risk of fracture or level of BMD. It is certainly not unreasonable, given the availability of tools that calculate fracture risk, to consider a reduction to below a given threshold of risk as a goal for therapy. But how do we achieve this? The Perspective quite rightly draws attention to a number of unknowns about the interpretation of changes in risk factors during therapy. Other considerations are worth highlighting: a number of clinical risk factors can remain stable (eg, prior fracture status) or only progress in the direction of increased risk (eg, age, incident fracture in a fracture-naive patient). Others can move in the direction of lower risk, eg, stopping smoking or decreasing alcohol intake, though it is uncertain whether the expected risk reduction is actually achieved because there are no prospective studies to provide direct evidence of this. If such risk factors remain unchanged during therapy, then the only parameter in the risk models that can really show treatment-induced changes is BMD. So why not just propose a goal for BMD? As stated earlier, the role of BMD in monitoring therapeutic response remains controversial. In part, this reflects limitations in our ability to detect small within-person changes in BMD relative to the inherent measurement error. It also appears that the relationship between an increase in BMD and reduction in fracture risk in the setting of therapeutic response does not parallel the gradient of risk for BMD to predict fractures in untreated individuals. Many studies have shown that the relationship between change in BMD and fracture risk reduction is relatively weak COMMENTARY JBMR