Towards safe and effective use of medication: prediction scores for major fracture risk assessment in the general elderly population and in elderly patients with rheumatoid arthritis
04 / 2011 - 03 / 2015
Background Osteoporotic fractures are a major burden of disease in the elderly. Every year, more than 17,000 people (with a mean age of 75 years, and 73% with female gender) sustain a hip fracture in the Netherlands [CBO 2009,Pouwels 2009]. Extrapolating data from abroad would suggest that excess 1-year mortality is 20%-24% [Cummings 2002,Leibson 2002], and annual costs are 392 million euro or 0.6% of the 2005 Dutch healthcare budget [Poos 2008]. In elderly, psychotropic drugs and oral corticosteroids are frequently used, and associated with hip fracture. We have estimated their population atrributable risk to hip fracture as follows: anxiolytics/hypnotics: 16%-32%; antidepressants: 10%-16%; antipsychotics 2%-4%, and oral corticosteroids 3%-9%. The estimates are based on 2008 prevalence rates of medication use in patients aged 70+, and the following ranges of relative risk of hip fracture: 1.5-2.1 for anxiolytics/hypnotics users [Cumming 2003], 1.7-2.4 for antidepressant users [vd Brand 2009], 1.3-1.7 for antipsychotic users [Pouwels 2009,Hugenholtz 2005], and 1.3-1.9 for corticosteroid users [de Vries 2007]. In contrast to many other countries, a validated clinical score to predict risk of fracture in elderly patients, does not exist in the Netherlands. Identifying persons at high risk could help to reduce fracture rates, and associated mortality and morbidity. A major limitation of existing fracture prediction scores is, that they largely consist of non-modifiable risk factors such as age, gender, smoking status, rheumatoid arthritis and secondary osteoporosis. Three recent developments may now give the opportunity to develop a new concept of case-finding and fracture preventing strategies in patients at risk of osteoporotic fracture: 1. The World Health Organisation has developed FRAX (R), a validated clinical risk score that estimates absolute risks of fracture; 2. Several important, frequently prevalent genetic risk factors of fracture have been identified; 3. There is overwhelming evidence that elderly patients on psychotropic drugs (such as antidepressants, hypnotics/anxyolitics and antipsychotics) are at increased risk of (hip) fracture; In contrast to the risk factors from FRAX (R) or genetic risk factors, prescribing of CNS drugs (or their daily dose) may be modifiable, which offers an instrument for clinical interventions. Objectives 1. To determine -for the first time ever- the incidence and prevalence of fractures in the Netherlands. 2. To estimate the absolute 5-years risk of osteoporotic and hip fracture in the elderly. A prediction score will be developed and validated, integrating longitudinal information on exposure to corticosteroids and central nervous medication, genetic risk factors, and clinical risk factors. 3. To recalibrate the validated prediction score in patients with rheumatoid arthritis. 4. To determine predictors of over-/undertreatment with osteoporosis drugs, and psychotropic drugs. 5. To determine predictors of 1-year mortality and 1-year fracture risk, after a hip and/or osteoporotic fracture. 6. To dissemate the knowledge from this study that will contribute to new clinical guidances on fracture prevention. The costs of developing this first Dutch clinical score with validated risk factors of osteoporotic fracture and this project, are lower than the costs of 43 new hip fractures [Poos 2008]. Methods In the Netherlands, the general practitioner is the gatekeeper to the public healthcare system. They practice in the community outside the hospital, referring ambulatory patients to other medical disciplines for outpatient or inpatient care. Other medical disciplines comprise consultants and complementary specialties. They report their findings and actions back to the concerned GPs. GPs address approximately 90% of the medical problems presented to them (www.ipci.nl). We will use primary care data in a representative sample (n=1.2 million) of the total Dutch population. This data will be linked to the national hospitalisations registry, pharmacy dispensing data, and the national registry of mortality, on an individual patient level. In this representative sample, we will develop and validate prediction scores for hip and osteoporotic fracture, using an efficient two-gate design. Validated predictors from the WHO FRAX tool will be combined with data on exposure to psychotropic drugs, corticosteroids, and common (prevalence 50%-80%) genetic risk factors. The fracture risk score will be calibrated to a high-risk subpopulation with rheumatologist-confirmed rheumatoid arthritis. Using life table- and Cox regression analysis, the impact of treatment adherence, and under- & overtreatment with osteoporosis/psychotropic drugs on fracture risk and its associated mortality in elderly will be studied.