Assessing 5-year incidence rates and determinants of osteoporotic fractures in primary care

To assess the gender and age-related 5-year incidence rates of osteoporotic fractures, and their related predictors, in a primary care setting.

We obtained information from the Health Search-CSD Longitudinal Patients Database (HSD). This is an Italian General Practice data repository which comprises information given by computer-based patient records of a selected group of over 900 Primary Care Physicians (PCPs). We selected all patients aged 50 to 85 years, who were actively included into the PCP's list at the beginning of the enrolment period (1st January 2002-31st December 2003). We excluded individuals who were registered in the PCPs' list for less than 1 year before the entry date (Index date) into the cohort, as well as those who were diagnosed with Paget disease or malignant neoplasm. Participants were followed up until the occurrence of osteoporotic fracture, one of the exclusion criteria, or the end of the study period.

The 5-year rates (per 1000 person-years) of any osteoporotic fracture were 11.56 (95% C.I. 11.33 to 11.77) among females, and 4.91 (95% C.I. 4.75 to 5.07) among males. For hip fractures, the overall incidence rates were 3.23 (95% C.I. 3.11 to 3.34) among females and 1.21 (95% C.I. 1.12 to 1.28) among males, respectively. Advanced age, history of fracture, use of corticosteroids, rheumatoid arthritis, BMI<=20, presence of osteoporosis, gastrointestinal and chronic hepatic disease, depression, chronic obstructive pulmonary disease, use of anticonvulsants and a higher number of co-medications, increased the risk of any osteoporotic fractures.

The use of primary care data confirms a higher incidence of osteoporotic fractures among females vs. males as well as in older individuals. Predictors of osteoporotic fractures were consistent with FRAX® algorithm. Given the clinical utility of a simple score for the assessment of absolute fracture risk among osteoporotic patients, its assessment and validation in the Italian HSD could potentially provide an applicable prediction tool.