Atypical antipsychotics (AAs), which are widely prescribed to manage psychiatric and behavioral symptoms in elderly adults in nursing home settings, are associated with an increased risk of bone fracture. The mechanism by which AAs increase fracture risk is not well understood, although recent evidence suggests molecular mechanisms that result in bone loss. ?-blockers (BBs), which are widely prescribed to nursing home residents to manage cardiac disease and high blood pressure, have demonstrated effects to reduce fracture risk. The size of the risk reduction has been found to be dependent on BB class, with ?1 selective blockers showing the greatest protective effect in most studies. The protective effect of BBs on fracture risk is also supported by animal studies, some of which have been completed by fellow team member and project leader in this COBRE (K. Motyl, Project 4). Our primary hypothesis is that concurrent BB use in incident AA users will result in a reduction in fracture risk, and we further hypothesize that the magnitude of this effect will vary with ?1 selectivity. We propose a large observational study to measure the effect of concurrent BB use on fracture risk in elderly nursing home residents initiating AAs. Our clinical study will be supported by concomitant studies (K. Motyl) addressing bone physiology in a mouse model. We will analyze data from a large, national database of nursing home resident data with linked clinical characteristics from the Minimum Data Set and diagnosis and drug data from Medicare Parts A, B, and D for the majority of U.S. long- stay residents. In our first specific aim, we will test the hypothesis that patients initiating AA (termed AA initiates) in the nursing home with concurrent exposure to BBs will have a reduced risk of fracture. Our primary outcome will be hip fracture, and major osteoporotic fractures will be a secondary outcome. We will also analyze falls as another secondary outcome that lies along the causal pathway to fracture. We will make appropriate adjustments for imbalanced patient characteristics between BB users and non-users using propensity score methods, and use a novel time-to-event method that allows for modeling time-dependent exposure to BBs to allow for precise estimates of the effect of BB use. In our second specific aim we will estimate the within-class variation in the effect of BB on fractures and falls as a function of ?1-selective vs. non-selective class. At the conclusion of this study we will discover whether BB exposure mitigates fracture risk associated with AA treatment, and show how that effect varies by class of BB drug. This study will inform animal model research and provide preliminary data for BB-based mechanisms for the treatment of AA-induced fracture risk in the nursing home setting.
Showing the most recent 10 out of 28 publications
