Background A diagnosis of osteoporosis is achieved by demonstration of bone loss at either the hip or spine using bone densitometry. A subset of patients have bone loss at only one of these sites. Risk factors for osteoporosis are well known but it is unknown whether patients with exclusive bone loss at either site could be predicted. The relationship between exclusive bone loss and fracture has not been established.
Objectives To determine whether patients with bone loss exclusively at the spine or the hip have different predictors and whether they correlate with prevalent fractures.
Methods Patients referred to a district general hospital in North West England for bone densitometry between June 2004 and September 2010 were identified as either having osteoporosis (T score <-2.5) of either the spine or the hip but not both. Risk factors for osteoporosis were assessed and groups were compared before and after adjustment for age and gender using a logistic regression model. Risk factors identified include body mass index (BMI), steroid use, smoking history, rheumatoid arthritis (RA), alcohol excess, history of fracture, family history and maternal hip fracture.
Results Results are summarised in table 1. There were 711 patients with bone loss at the hip (OPHIP) and 1467 with bone loss at the spine (OPSPINE). Continuous variables are expressed as mean(standard deviation), dichotomous variables are expressed as a percentage. Results are shown before and after adjusting for the differences in age and gender between the two groups. After adjustment, the OPHIP group were shorter, heavier and had a higher prevalence of RA, steroid users and smokers. There were no statistically significant differences between the groups in prevalence of fracture at the femur(p 0.27, OR 9.28, CI 0.59, 6.67), pelvis (p 0.28, OR 1.74, CI 0.64, 4.72), forearm(p 0.94 OR 0.99, CI 0.79, 1.25), rib(p 0.22 OR 0.70, CI 0.4, 1.24), humerus(p 0.86, OR 1.04, CI 0.65, 1.66), spine (p 0.80, OR 0.96, CI 0.7, 1.10) or tibia/fibula (p 0.07, OR 1.38, CI 0.97, 1.96).
Conclusions We have identified several variables (height, weight, smoking status and history of rheumatoid arthritis and steroid use) which may be used to predict the site of bone loss. There was no difference between the two groups in prevalence of fracture or fracture at specific sites. Further investigation into determinants of exclusive bone loss at these sites is warranted. Use of the hip only as a predictor of fracture appears valid.
Disclosure of Interest None Declared