In Reply: Dr Taguchi raises the hypothesis that in our study symptomatic women with a prevalent vertebral fracture at baseline might be more likely to have received treatment and thus have had a lower risk of developing incident fractures. However, when we adjusted all models for baseline use of estrogen and any use of estrogen, bisphosphonate, or selective estrogen receptor modulator over the follow-up period, the results were essentially the same. Although the prevalence of back pain due to vertebral fractures is unknown1 and we do not have information on clinical spine fractures, we performed an additional analysis stratifying women with a prevalent vertebral fracture by whether or not they self-reported back pain at baseline. In our study of 2680 women, vertebral fracture incidence rates were 14.2% for women without a prevalent vertebral fracture (n = 2286), 36.5% for women with a prevalent vertebral fracture and no back pain (n = 104), and 43.1% for women with a prevalent vertebral fracture and back pain (n = 290). Compared with women without a prevalent vertebral fracture, the multivariate adjusted OR (including adjustment for lumbar spine bone mineral density) of an incident vertebral fracture in women with a prevalent vertebral fracture and no back pain was 2.52 (95% confidence interval [CI], 1.58-4.02); in women with a prevalent vertebral fracture and back pain, the OR was 3.51 (95% CI, 2.63-4.69) (P for interaction = .18). Thus, there was no evidence that presence of back pain modified the association between prevalent vertebral fractures and incident vertebral fractures.
Cauley JA, Lui L, Hillier TA. Prevalent Vertebral Fracture and the Risk of Incident Vertebral Fracture—Reply. JAMA. 2008;299(14):1666-1667. doi:10.1001/jama.299.14.1666-c