Background/Purpose: Falls in the elderly can lead to fragility
fractures. Fall prevention targeted at those with high risk for falls could
help reduce fractures. Although the fracture risk assessment tool, FRAX, does
not directly incorporate fall risk, it does include some risk factors for falls.
We examined whether FRAX could predict fall risk in older ambulatory women.

Methods: We studied 125 community-dwelling
women, age ≥ 65 years, in whom 124 had femoral neck bone mineral density (FN
BMD, g/cm²) and US-FRAX scores (% 10-yr risk for major osteoporotic
fractures [FRAX-OP] and hip fractures [FRAX-Hip]) reliably determined at
baseline. Falls were tracked longitudinally every two weeks over 1 year by
self-reported questionnaire. We examined fall risk for FRAX-OP and FRAX-Hip
scores separately. For analyses, we created 4 groups for each FRAX score, using
FRAX score cutoffs for osteoporosis [OP] treatment (≥20% for FRAX-OP and
≥3% for FRAX-Hip) as the initial basis for groupings. We then examined
whether FRAX scores (by groups and also as a continuous variable) predicted
falls using an Andersen-Gill model, adjusted for age and prior fall history.

Results: The median (range) age, FN BMD and
FRAX scores of women were 77 (65-92 years), 0.81 (0.54-1.20 g/cm²),
14.2 (3.3-56.9%, FRAX-OP) and 3.5 (0.1-45.8%, FRAX-Hip), respectively. There
were 28 women (23%) with FRAX-OP ≥20% and 67 (53%) with FRAX-Hip
≥3%. There were 47 women (38%) who reported a fall in the year prior to
their study visit. Compliance with all fall tracking questionnaires was 97.5%. Over
follow-up, 73 women fell, 38 of whom had more than one fall, for a total of 157
falls over 125 person-years of follow-up [p-y] (1.3 falls/p-y). There was a
trend for higher fall risk with greater FRAX score group (FRAX-OP, p=0.17; FRAX-Hip,
p=0.02) [Table]. But the association appeared non-linear when using FRAX
as a continuous variable in analyses, with decreasing fall risk in women having
the very highest (>25%, FRAX-OP and >10%, FRAX-Hip) scores [data not
shown]. Women with some of the highest FRAX scores were more likely to have a
parent with hip fracture, increasing the FRAX score, but which may not
necessarily have influenced fall risk. They also reported being less active, which
may have reduced their likelihood for falling over follow-up.

Conclusion: We found a higher risk of future falls
in older women with greater FRAX scores, including at some scores below the recommendations
for OP treatment. That said, the non-linear association we observed at the
highest FRAX scores suggests that information beyond FRAX would still be
necessary to distinguish those at high vs. low fall risk.