2017 ISAKOS Biennial Congress ePoster #1206

 

Identifying the Variation in Femoral Rotational Axes Across Demographic and Radiographical Measures in Pre-Operative Total Knee Arthroplasty Patients

Joshua Twiggs, BEng, Pymble, NSW AUSTRALIA
Caitlin Wilcox, Bachelor of Engineering, Sydney, NSW AUSTRALIA
Willy Theodore, BEng, Pymble, NSW AUSTRALIA
Elizabeth Kolos, Bachelor of Engineering, Sydney, NSW AUSTRALIA
David M. Dickison, MBBS(Hons), FRACS, FAOrthA, Sydney, NSW AUSTRALIA
Brett A. Fritsch, MBBS BSc(Med), FRACS, FAOrthA, Hunters Hill, NSW AUSTRALIA
Justin P. Roe, FRACS, A/Prof., Sydney, NSW AUSTRALIA
Brad Miles, BEng, Sydney, NSW AUSTRALIA

360 Knee Systems, Sydney, NSW, AUSTRALIA

FDA Status Cleared

Summary

This study characterised differences in the TEA and AP orientation to the PCA across native alignment, gender and demographics.

Abstract

Introduction

Femoral component rotation landmarks have traditionally referenced the Trans-Epicondylar Axis (TEA), Whiteside’s Anterior-Posterior Axis (AP), or the Posterior Condylar Axis (PCA). The TEA is a superior capture of the native kinematic flexion axis but is not as reproducible or accessible in surgery. This study characterised differences in the TEA and AP orientation to the PCA across native alignment, gender and demographics.

Methods

A patient specific CT segmentation process was used to analyse 877 pre-operative CT scans by two trained biomedical engineers each. Ethics was approved by Bellberry Human Research Ethics Committee, application number 2012-03-710. The population was 44% male and had an average age of 69.7 ± 8.6. 21.5% of the knees were valgus. Points defining the medial epicondylar sulcus, lateral epicondyle, posterior condylar axis and Whiteside’s anterior axis, along with the mechanical axis of the femur and tibia. Comparisons between the axial and long leg alignments obtained and demographic patient details were performed. Two tailed paired t-tests were used to determine significant differences between subsets, while trends were assessed with Spearman’s correlations.

Results

Of the population studied, the average coronal long leg alignment in CT scan was 4.2° ± 5.8° varus. The mean AP angle to the TEA was 0.7° ± 5.3° externally rotated, which was statistically different from 0°. (P<0.001) The mean TEA angle to the PCA was 1.8° ± 1.8° of external rotation and was statistically significantly different from the typically quoted 3° (p < 0.001).

Statistically significant differences were observed between the varus and valgus subpopulations for the TEA to PCA angle (p<0.001), with the varus mean angle being 1.6° ± 1.8° external rotation while the valgus mean angle was 2.7° ± 1.8°. Statistically significant differences also existed between the varus and valgus groups in terms of AP to TEA angle, with the varus group having a mean rotation of 0.9° ± 5.3° external rotation and the valgus 0.3 ° ± 5.1° internal rotation. Statistically significant gender differences also existed between the coronal long leg alignment (females trending to valgus) and TEA to PCA (females trending to greater external rotation).

Discussion

The mean TEA to PCA angle deviated from the standard reference of 3° external rotation knees defined in literature. This results were statistically significant but may not have clinical relevance in and of their own. More significant was the subgroup analysis by varus and valgus alignment of the knee, which showed statistically different means. These relationships were all clinically identifiable with regards to the patient’s morphology. The results showed a large spread of values that had variances in malalignment from the standard 3° reference. While subgrouping was able to explain some of the variance, the majority appears to be unaccounted for by the factors in this study.

Conclusion

This study presented results that showed patient specific variation when comparing the TEA and AP orientation to the PCA. The unaccounted for nature of these patient specific variations encourages a CT or MRI based approach to pre-surgical planning, given the challenges of repeatable intraoperative axis identification.