Comparison of radiographic and anatomic femoral varus angle measurements in normal dogs

Objective— To determine if the clinically practiced method of radiographic femoral varus angle (R-FVA) measurement is repeatable, reproducible, and accurate.
Study Design— Radiographic and anatomic study.
Animals/Sample Population— Normal Walker hound cadavers (n=5) and femora (n=10).
Methods— Cadavers were held in dorsally-recumbent and torso-elevated positions as 3 craniocaudal radiographs were made of each femur, by each of 2 different technicians. Femora were then harvested for direct measurement of anatomic femoral varus angle (A-FVA). R-FVA was measured on each radiograph by each of 3 examiners on 3 separate occasions. Intra-observer (repeatability) and inter-observer (reproducibility) variance in R-FVA measurement and the strength of relationship between R-FVA and A-FVA (accuracy) were determined.
Results— Mean (±SD) A-FVA was 5.2±2.1° (range, 2.4–8.2°). Mean (±SD) R-FVA was 5.8±1.0° (range, 2.7–9.6°). Intra-observer variance (range: 11–16%) and inter-observer variance (16%) were acceptable. The strength of relationship between measured R-FVA and A-FVA (maximum adjusted R²<0) was unacceptably low regardless of observer, patient position, or radiographic technician.
Conclusion— R-FVA measurement was repeatable and reproducible, but not statistically accurate in predicting A-FVA in these 5 normal Walker hounds. The detected inaccuracy may be real or the result of a selection bias for normal dogs obscuring the true relationship.
Clinical Relevance— R-FVA may not be an accurate method of femoral varus measurement in dogs with A-FVA<10°. Using Slocum's criteria for distal femoral osteotomy (R-FVA>10°), the procedure would not have been erroneously performed in any of the normal dogs of this study.