EQUINE | Abstracts STUDY DESIGN: Prospective, observational cross-sectional study. METHODS: Thoroughbred racehorses competing between 2012 and 2015 were examined on-course no earlier than 30 min after racing. Examinations were recorded and graded blindly by experienced veterinarians using a 0-4 scale. Linear mixed effect models were used for analysis of continuous response variables with horse name in- corporated as a random effect to account for repeated sampling and horse variability. Generalised estimating equations were used for analysis of binary responses. Performance variables were examined in 2 models, com- paring EIPH grade 0 to grades 1-4, and EIPH grade ≤2 compared with EIPH grades ≥3, controlling race factors that could influence performance. RESULTS: There were 3794 observations collected from 1567 hors- es. EIPH was detected in 55.1% of observations. Horses with grade 4 EIPH were significantly more likely to have a lower finishing positi ۈ[[\\\Z[B[\\Z[HXH[H\][ۜ[BXXHX\[XY\X\[\XH\[\H\ݙ\H\ HوHXH[ܜ\]]RT ܘYH K[Z[\\X][ۜ\BY[[[[]ܜ\]RTܘYH8iL][[\[܈\ܛX[H[\\Yܜ\]RT8iPRSSRUUSӔ[Y[\[\KӐTSӎZ[[\]HY[[ܜYH\\X]Y][\[܈XH^H\ܛX[H[\[][ۈو܋BYYXZܜ\˂L\]Z[H]M[[X[YH][X][HX\XHو[X[B\[\[[[HܜKXX]YK[\]Y[\Y\ۈ\ۚH][LKX\[^\BXXPԓS[\]\ JHوH[\[ H[\BY]]\[ U H[\H[[ۈ]\Hو[Y[\[ܜ\ˈ[KX\X[\XYۛX[Y\\XH[܂\\]]X[X[ۈ\H[YY\ۈ[XYۛB[X]X[Hܜ\]Hو\H[˂ВPUTHؚX]Hو\YH\]\Z[HHXKBXHو\[\\وH[U[\[B[\H[][ۈو\]Z[H\[ۜ[\\H\KB[˂QHTQӎ[Z\Y^\[Y[[YKQU^\]ܜXX[XY\[\U[[LY]Yܜ\ˈH\]ܜ\H[[H\BYۙY[XZ]\HY[YU܈BY[YUۈ\[[H\YۙYܜ\˂H[\H[XY]H [][ۈو\YY][H [ K\[H[[H KH[ H[[ZZX[ B[ KHZ[[][HوY[ܘ\وXX[\Z[Y][X]HH\[Hو\YY][B][H\][TSHU[\X\ٝ[H[XY[̍[MHX\XJKH[\X\ٝ[H[XY[L̍[ HX\XJK[][X][ۈ]Y[HU[[\\XH[HوX\ٝ[U[X[ۜˈ[][X][ۈ]Y[H[(\]Z[HX[\]H8(