Fifty horses treated for metabolic trouble on long distance endurance rides were matched to 100 horses participating in the same rides and finishing the ride in good condition. The coupling was based on the breed, the age, the gender and the career. All the clinical parameters noted at each veterinary check before and during the ride were compared between the 2 groups of horses using the odds ratio and the c2 test. The statistical analysis demonstrated that all parameters were useful at a moment during the ride. A high respiratory rate, early signs of dehydration and a recovery time higher than 10 minutes appeared as significant risk factors. This experimental work is based on a limited group of horses and presents recognised methodological bias. A larger study would be needed to confirm the results.