Alternative classification and screening protocol for transitional lumbosacral vertebra in German shepherd dogs

The ventrodorsal pelvic and laterolateral lumbar spine radiographs of 228 German shepherd dogs were evaluated for existence and type of LTV. Morphology of transverse processes was used in classification of LTV in ventrodorsal radiographs. The relative length of sixth (L6) and seventh (L7) vertebrae (L6/L7) was used in characterization of these vertebrae in laterolateral radiographs. CT studies were available for 16 dogs, and they were used for more detailed characterization of different types of LTV. Non-parametric χ2 statistics, generalized logit model for multinomial data, and one-way analysis of variance was used for statistical analyses.In all, 92 (40%) dogs had a LTV, the most common type being separation of first spinous process from the median crest of the sacrum in 62 dogs (67% of LTV). Eight dogs had eight lumbar vertebrae. Those dogs with LTV had longer L7 in relation to L6 than dogs with normal lumbosacral junctions. When L6/L7 decreased by 0.1 units, the proportion of dogs belonging to the group with L8 was 14-fold higher than in the group with normal lumbosacral junctions. L8 resembled first sacral vertebra (S1) in length and position and was therefore classified as one type of LTV. With CT it was shown that categorizing LTV, based on shape and visibility of transverse processes seen in ventrodorsal radiographs, could be misleading.We suggest that L8 be included as a part of the LTV complex, and the laterolateral radiographs of the lumbar spine be considered as an addition to ventrodorsal projections in the screening protocols for LTV.