The risk of fractures in new users of statins: A propensity-score matched sequential cohort study

Date of ISAC Approval: 
10/03/2017
Lay Summary: 
Reduced bone strength is a problem, which is particularly frequent in elderly people and in women after menopause (i.e. after the last menstrual period). In patients with reduced bone strength, fractures frequently happen after a simple fall or another minor accident. Statins are drugs which lower blood lipid levels and which are prescribed to prevent heart attacks and strokes. In the United Kingdom, approximately 11% of the population are prescribed statins. Experiments in animals suggest that statins may reduce bone fragility by strengthening the bone. In a previous CPRD-based study by our group, which was published in 2000, we observed a decreased rate of fractures in statin users. These results led to a controversial scientific debate, and various subsequent studies have found contradictory results. Several other studies also reported a decreased probability to break a bone in statin users, while more recent studies did not confirm this association. We aim to perform another study to address this same question (i.e. is the probability for a fracture the same in statin-users and non-users?) using the same data source, but using a more advanced, sophisticated study design.
Technical Summary: 
There has been an ongoing scientific discussion over the past two decades, about whether or not statins reduce the risk of fragility fractures. This discussion was triggered by results from observational studies, one of which was a CPRD-based nested case-control analysis from our group, in which we reported a 45% decreased relative risk of fractures in statin users. More recent studies, including some randomized clinical trials, did not reproduce this finding and reported a null-association. We therefore aim to re-assess the relative risk of fracture associated with new statin use (a prior 3 years statin-free history is required), using CPRD data in a propensity score (PS)-matched sequential cohort study. Statin initiators will be matched to non-initiators on their PS within 2 year entry blocks (to avoid time trend bias). We will follow all patients from the day of a first statin prescription for 5 years, or until they have a recorded fracture, change of exposure status, or until they are otherwise censored. We will perform subgroup analyses by sex, age, daily statin dose, and by duration of follow-up. Cox proportional hazard analyses will be performed to calculate hazard ratios (HR) with 95% confidence intervals (CI).
Health Outcomes to be Measured: 
We will investigate a composite outcome including all first-time ever fragility fractures, i.e. - Malleolus (ankle), foot - femur, hip - vertebrae - humerus, shoulder, lower arm, wrist, hand - undefined Patients will be followed until the first fracture code after cohort entry is observed.
Collaborators: 

Christoph Meier - Chief Investigator - University of Basel
Julia Spoendlin - Corresponding Applicant - University of Basel
Dr Susan S Jick - Collaborator - BCDSP - Boston Collaborative Drug Surveillance Program
Theresa Burkard - Collaborator - University of Basel