Arthritis & Rheumatism, Volume 60,
October 2009 Abstract Supplement

The 2009 ACR/ARHP Annual Scientific Meeting
Philadelphia October 16-21, 2009.

Active Shape Modeling of the Hip and Incident Osteoporotic Hip Fractures

Luker1,  K.R., Lynch2,  John A., Chaganti3,  R.K, Parimi4,  N., Nevitt2,  M., Gregory5,  J.S., Hillier6,  T.

UCSD, San Diego, CA
UCSF, San Francisco, CA
Aberdeen, United Kingdom
KPMG, Portland, OR
UCDMC, Sacramento, CA


Variations in femoral head shape have been reported to predict incident hip fractures. This study evaluated the proximal femur shape as a risk factor for incident hip fractures in a cohort of elderly women through the method of active shape modeling (ASM).


This analysis was performed using prospectively collected data from a cohort of white women >=65 years of age at baseline enrolled in the Study of Osteoporotic Fractures (SOF). Supine pelvic radiographs were obtained at the baseline SOF visit. A nested case-control study was performed. Hips were eligible for inclusion if they had no prevalent radiographic osteoarthritis (RHOA) or fracture in either hip at baseline. Case subjects had no RHOA baseline and had an incident hip fracture after the baseline visit (n=168). Control subjects were a random selection of subjects (n= 231) without RHOA in either hip at baseline or incident hip fractures during the follow-up period. The shape of the right proximal femur was outlined on a digitized baseline radiograph by two readers (KL, KC) using a statistical image analysis technique. Active shape modeling (ASM) was performed to generate 10 unique and independent modes of variations in the proximal femur shape, accounting for 95% of the variance in the shape of the proximal femurs (Gregory et al, Osteo. Inter, 2004). Three other measures of hip geometry associated with hip fracture, namely femoral neck width and length and femoral head diameter, were also measured. The association of proximal femur shape by ASM and other geometric measures were analyzed by logistic regression as independent predictors of incident hip fracture adjusting for covariates.


The incident hip fracture cases were older, weighed less and had lower hip bone mineral density than the control subjects (p<0.05). Point-to-point correlations between each film and the composite average shape were computed for the entire set of 399 images and were used to describe 10 modes of variation. Modes 1 (OR 1.62,95% CI: 1.26–2.07 p<0.001, per SD change in mode score) and 4 (OR 2.13, CI: 1.63–2.77, p<0.001) were found to carry a significantly increased risk for future hip fractures while modes 6 (0.49 95% CI: 0.38–0.65, p<0.001) and 10 (0.60, 95% CI: 0.47–0.78 ), P<0.001) were shown to have a significantly lower risk for incident hip. Femoral neck length was also associated with incident hip fracture, while femoral head diameter and femoral neck diameter were not significantly correlated.


These results suggest that variation in the relative size of the femoral head and neck are moderate determinants of incident hip fractures. The combination of hip shape, femoral neck length, and BMD may improve our ability to identify patients who may be at increased risk for osteoporotic hip fractures and implement appropriate therapy.

To cite this abstract, please use the following information:
Luker, K.R., Lynch, John A., Chaganti, R.K, Parimi, N., Nevitt, M., Gregory, J.S., et al; Active Shape Modeling of the Hip and Incident Osteoporotic Hip Fractures [abstract]. Arthritis Rheum 2009;60 Suppl 10 :594
DOI: 10.1002/art.25674

Abstract Supplement

Meeting Menu