The pathophysiology of cold injury is still controversial. An inflammatory process has been implicated as the underlying mechanism and certain anti-inflammatory substances such as ibuprofen and acetylsalicylic acid have been used in the clinical treatment of frostbite injury. It has been postulated that the progressive ischemic necrosis is secondary to excessive thromboxane A2 production, which upsets the normal balance between prostacyclin (prostaglandin I2) and thromboxane A2. It was aimed to clarify the pathophysiology of cold injury in this study. Twenty-one New Zealand White rabbits, each weighing 1.2 to 2.9 kg, were divided into control (n = 10) and frostbitten (n = 11) groups the randomly. The rabbit ears in the frostbitten group were subjected to cold injury, and the levels of thromboxane A2 (as thromboxane B2) and of prostaglandin I2 (as 6-keto-prostaglandin F1alpha) and the number of inflammatory cells (polymorphonuclear leukocytes and mast cells) were measured in normal and frostbitten skin of rabbit ears. The levels of 6-keto prostaglandin F1alpha and thromboxane B2, the stable metabolites of prostaglandin I2 and thromboxane A2, respectively, were increased in a statistically significant way (p < 0.002) by frostbite injury; however, thromboxane B2 increased more than 6-keto prostaglandin F1alpha. Polymorphonuclear leukocytes and mast cells, absent in normal skin, were present in the frostbitten skin. There was a statistically significant (p < 0.01) correlation between the time a rabbit ear was maintained at below -10 degrees C and skin survival and between the weights of rabbits and skin survival (p < 0.024). All these findings suggest that inflammation is involved in frostbite injury; a decrease in prostaglandin I2/thromboxane A2 ratio could be one of the factors leading to necrosis; the bigger the animal, the better its ability to counter frostbite.