The accuracy of historical height loss for the detection of vertebral fractures in postmenopausal women

Abstract

Historical height loss (HHL) can be calculated as the difference between a patient's tallest recalled height (TRH) and the current measured height (MH). We have examined the accuracy of HHL as a clinical test for the detection of prevalent vertebral fractures. Subjects were postmenopausal women aged 50 or older who had been referred for specialist assessment of osteoporosis risk (n=323; average age 66.0+/-9.2 years; range 50-92 years). MH was determined using a wall-mounted stadiometer. The presence of prevalent vertebral fractures was assessed by radiographic morphometry, with fracture defined as a vertebral height ratio<0.8. The positive likelihood ratio (LR+) for fracture was relatively flat until HHL>6.0 cm. With HHL from 6.1 to 8.0 cm, the LR+ was 2.8 [95% confidence interval (95%CI), 1.3, 6.0]. When HHL was >8.0 cm, the LR+ was 9.8 (95% CI, 3.0, 31.8). The area under the receiver operating characteristics curve for the ability of HHL to detect fracture was 0.66 (95% CI, 0.59, 0.72). At HHL>6.0 cm, sensitivity was 30% (95% CI, 22, 37%), and specificity was 94% (95% CI, 90, 97%). The positive predictive value was relatively low across a range of theoretical prevalence, rising above 80% only at very high prevalence rates (>50%). In contrast, the negative predictive value was high at the prevalence rates seen in most clinical practice, and dropped below 80% only when the prevalence exceeded 25%. This study shows that HHL<or=6.0 cm rules out prevalent vertebral fracture with a high degree of accuracy; patients with HHL>6.0 cm should have spine radiographs to examine for the presence of vertebral fractures.