Investigation of the Reliability of Saddle Fitters to Determine the Position of the Last Thoracic Vertebra of Horses Using Palpation Techniques: A Pilot Study.

Published guidelines indicate an English saddle tree should not extend beyond the 18th thoracic vertebra (T18). The study aimed to assess reliability of saddle fitters (SFs) to identify the T18 spinous process (SP). Part 1 investigated agreement between T18 (T18SF) as identified by three SFs using palpation and a veterinary surgeon (VS) using radiography (T18VS) in seven horses. Saddle fitter 1 and SF2 palpated the lumbosacral joint and counted cranially six SPs, whereas SF3 followed the rib curvature toward the dorsal midline. In part 2, SF1 and SF2 identified T18 by counting cranially five SPs in seven horses on two occasions. Agreement between SFs and VS was assessed using t tests and Bland-Altman plots. Interrater and intrarater reliability were estimated using intraclass correlation coefficients. In part 1, SF1 and SF2 found T18SF 4.3 cm (±4.1 and 4.0 cm, respectively) cranial to T18VS. Mean difference between T18SF3 and T18VS was 0.1 ± 4.9 cm (95% CI: -9.5 cm, 9.6 cm, P = .976). When counting cranially five SPs, mean difference between T18SF1 and T18VS was -1.5 ± 3.4 cm (95% CI: -8.3 cm; 5.1 cm; P = .265) and T18SF2 and T18VS was -0.3 cm ± 4.5 cm (95% CI: -8.8 cm; 8.5 cm; P = .847). Interrater reliability was "good" (ICC = 0.798). Intrarater reliability was "excellent" for SF1 (ICC = 0.905) and "good" for SF2 (ICC = 0.847). Counting cranially five SPs from the lumbosacral joint, when coupled with observation of the rib position and curvature should ensure a saddle is not placed beyond T18.