P077 Bone mineral density predicts fractures when measured in more than one area: results from a large observational cohort

Abstract

Background/Aims

Osteoporosis is characterised by a reduction in bone mineral density (BMD). Low BMD is correlated with increased fracture risk. It is unclear whether a BMD value at a specific site is superior for predicting fracture. The Fracture Risk Assessment (FRAX™) tool is used to evaluate fracture risk, using the BMD at the non-dominant hip; focussing on the neck mean and neck total. This is despite a T score of -2.5 or below in the lumbar spine defining a patient as osteoporotic. This study aimed to determine whether the other measures of BMD that are noted are as predictive of fracture.

Methods

Patients referred for routine dual-energy x-ray absorptiometry (DEXA) scans at the Royal Lancaster Infirmary from 2004 to 2019 were studied. BMD data from lumbar vertebrae (L1-L4) and anatomical regions of the left and right femur were collected from DEXA images. A logistic regression model was fitted using fracture as a dependent variable and BMD values from the lumbar vertebrae, right and left femur as predictors, adjusted for age at scan and sex. Goodness of fit was assessed using a receiver operating characteristic (ROC) curve.

Results

A total of 31,546 patients were included in the analysis (82.9% female), of whom 37.6% had sustained a fracture, and the mean age of the cohort was 64.9 (standard deviation 12.9). Table 1 shows the results of the logistic models. Areas under the receiver operating characteristic curve (AUC) demonstrated the positive predictive ability of a reduction in BMD in the lumbar vertebrae and regions of the femur and fracture risk. BMD data from the four lumbar vertebrae were just as predictive of fracture as the right and left femur.

Conclusion

The study findings demonstrate that low BMD in all areas measured were equally predictive of fracture. When the non-dominant hip is not measurable, using other areas from a DEXA scan could be used to predict fracture risk. Research evaluating change in BMD over time could alter this observation.

Disclosure

U. Nadeem: None. M. Bukhari: None. Z. Sultan: None.

P077 Table 1.

Odds ratio, 95% confidence interval and AUC value corresponding to the region of interest

Region of interest	Odds ratio	95% confidence interval	AUC value
L1	0.84	[0.82, 0.85]	0.63
L2	0.84	[0.82, 0.85]	0.64
L3	0.98	[0.97, 0.98]	0.64
L4	0.85	[0.84, 0.86]	0.64
Right femoral neck	0.63	[0.52, 0.77]	0.64
Right femoral shaft	0.16	[0.14, 0.19]	0.64
Trochanteric area of right femur	0.11	[0.09, 0.13]	0.64
Upper area of right femoral neck	0.79	[0.64, 0.96]	0.64
Lower area of right femoral neck	0.14	[0.12, 0.16]	0.64
Left femoral neck	0.06	[0.05, 0.08]	0.63
Left femoral shaft	0.17	[0.14, 0.19]	0.63
Trochanteric area of left femur	0.10	[0.09, 0.12]	0.63
Upper area of left femoral neck	0.08	[0.06, 0.1]	0.63
Lower area of left femoral neck	0.13	[0.11, 0.15]	0.64