In conventional pharmacological studies, intersubject differences within an animal strain are normally neglected, leading to variations in pharmacological outcomes in response to the same stimulus. Using two classical experimental models, the Streptozotocin (STZ)-induced diabetic model of Wistar rats and the high-energy, diet-induced obesity model of Sprague-Dawley rats, we demonstrate that the different outcomes of STZ or diet intervention are closely associated with variation in predose (baseline) urinary metabolic profiles of the rats. The pharmacometabonomic analysis of predose metabolic profiles indicates that the intersubject difference is, to a great extent, associated with gut-microbiota, which predisposes different pathophysiological outcomes upon diet alteration or chemical stimulus. We hypothesize that there may exist an important association between observations from these two models and the obese/diabetic human population in that subtle variations in metabolic phenotype may predetermine different systems' responses to xenobiotic perturbation, ultimately leading to varied pathophysiological processes. Results from two independent models also suggest that the pharmacometabonomics approach is of great importance in the study of pharmacology and clinical drug evaluations, where endogenous metabolite signatures of predose individuals should be taken into consideration to minimize intersubject difference and the resulting variation in the postdose pharmacological outcomes.