LOS ANGELES — A prediction rule combining five easily obtainable risk factors distinguishes with high sensitivity women at high risk of developing osteoporotic fractures within the next 3 years, Dr. Idris Guessous reported at the annual meeting of the Society of General Internal Medicine.
The Swiss Evaluation of the Methods of Measurement of Osteoporotic Fracture Risk (SEMOF) study was a 3-year, prospective, multicenter study (n = 623) that computed a prediction score using low heel ultrasound stiffness index (SI), older age, fracture history, recent fall, and missed chair test to predict osteoporotic hip fractures and other nonvertebral fractures.
The objective of the study was to compute a prediction rule that would identify women at high risk of osteoporotic fracture in general, or a hip fracture in particular, within the next 3 years, said Dr. Guessous of the University Hospital of Lausanne, Switzerland.
The heel bone ultrasonometer (Lunar Corp., Madison, Wisc.) was chosen because it is simple, inexpensive, noninvasive, and transportable. Of 7,114 Swiss women who responded to a mailed request to participate, 6,174 women between 70 and 85 years old were enrolled. Exclusion criteria included previous hip fracture, bilateral hip replacement, renal failure, and active cancer.
Bone SI was calculated using quantitative ultrasound of the heel, broadband ultrasound attenuation, and speed of sound as the inputs. SI is expressed as a percentage of values obtained by the manufacturer in a young adult population. Osteoporotic fractures were defined as hip, wrist, or arm breaks that occurred spontaneously or secondary to falling from standing height or lower despite a low level of trauma.
The investigators included baseline characteristics (age, weight, height, body mass index), known risk factors for osteoporosis (fracture history, history of maternal hip fracture, current smoking habits, early menopause, surgical menopause), fall (history of recent fall, missed chair test), and SI as parameters to develop a prediction score. The investigators then used bootstrap methods to evaluate the stability of the score, Dr. Guessous said.
Mean follow-up was 2.8 years (17,546 person-years). Five risk factors were independent, significant predictors of the incidence of osteoporotic fractures: age older than 75, SI greater than 78%, history of any prior fracture, history of a fall during the last 12 months, and missed chair test (not being able to rise from a chair three successive times without using one's arms).
The investigators assigned a score to each of the five significant predictors: age, up to 3; SI, up to 7.5; history of fall within past 12 months, 1.5; fracture history, 1; and positive chair test, 1. Thus, the maximum prediction score is 14 points. The cutoff score to discriminate women at high risk of fracture with 90% sensitivity is 4.5. With this cutoff, 1,464 women (23.7%) were considered at low risk of hip fracture (score less than 4.5), and 4,710 (76.3%) were considered at high risk (score at least 4.5). Among these high risk women, 60 (1.3%) experienced an osteoporotic hip fracture. In contrast, 6 (0.4%) of the low risk women experienced such a fracture.
The main limitation of this predictor rule is that at a sensitivity of 90%, the specificity was only 24%.