Objective To investigate the effects of rapid hypothermia induced via esophagus on intestinal mucous injury in early stage after cardiopulmonary resuscitation (CPR) in a swine model of cardiac arrest. Methods Twenty-seven male domestic pigs weighing (36±2)kg were utilized. The animals were randomly divided into 3 groups (n= 9 in each): normothermia group (NT group), surface cooling group (SC group), and esophageal cooling group (EC group). The pig model was established by 8 mins of untreated ventricular fibrillation and then 5 mins of CPR. At 5 mins after restoration of spontaneous circulation (ROSC), therapeutic hypothermia was applied by either an esophageal cooling device in the EC group or a surface cooling blanket in the SC group to reach a targeted temperature of 33 ℃ maintained for 24 h after ROSC, and then followed by warming up in a rate of 1℃/hr for 5 hrs. A normal temperature of (38.0±0.5)℃ was maintained throughout the experiment in the NT group. The core temperature was continuously monitored during a period of 30 h after ROSC. At 3 h, 6 h, 12 h, 24 h and 30 h after ROSC, intestinal fatty acid binding protein (IFABP) content and diamine oxidase (DAO) activity in serum were measured by ELISA. At 30 h after ROSC, the pigs were sacrificed, and then intestinal tissue was rapidly obtained for the determination of tumor necrosis factor-alpha (TNF-α) and interleukin-6 (IL-6) contents by ELISA, cell apoptosis by TUNEL, and caspase-3 expression by immunohistochemistry. Results The rate of temperature decrease was 2.8 ℃/h and the time required for target temperature was 102 min in the EC group, while the rate of temperature decrease was 1.5 ℃/h and the time consumed for target temperature was 185 mins in the SC group, which suggested the efficacy of cooling was significantly better in the EC group than that in the SC group (both P<0.05). Compared with the NT group, serum IFABP content and DAO activity were significantly decreased at 3 hrs after ROSC in the EC group and at 6 hrs after ROSC in the SC group. Compared with the SC group, serum IFABP content at 6 hrs after ROSC and DAO activity at 12 h after ROSC were significantly decreased in the EC group IFABP (pg/mL): (710±32) vs.(777±52) at 6 h, (870±49) vs. (960±64) at 12 h, (1 022±65)vs. (1 143±63) at 24 h, (882±71) vs. (1 006±45) at 30 h DAO (U/mL): (39.9±1.9) vs. (43.4±3.2) at 12 h, (30.6±2.4) vs.(34.0±3.1) at 24 h, (26.1±2.7) vs. (29.4±2.2) at 30 h, all P<0.05. In the intestinal tissue, TNF-α and IL-6 contents were significantly reduced, and cell apoptosis index and caspase-3 expression were significantly decreased in the SC and EC groups compared with the NT group. Additionally, inflammatory response and cell apoptosis in intestinal tissue were further significantly lesser in the EC group compared with the SC group TNF-α(pg/mL): (721±94) vs. (922±125); IL-6(pg/mL): (454±69) vs. (697±132); Apoptotic index(%): (6.2±2.6)vs.(12.8±3.0); caspase-3 expression (IOD): (8.9±1.6) vs. (15.9±1.9), all P<0.05. Conclusions In a swine model of cardiac arrest, rapid hypothermia could be successfully induced via esophagus and consequently produced a greater protective effect on post-resuscitation intestinal injury compared with the conventional surface cooling. The protective mechanisms are associated with the inhibition of inflammatory response and cell apoptosis. Key words: Cardiac arrest; Cardiopulmonary resuscitation; Therapeutic hypothermia; Surface cooling; Esophageal cooling; Intestinal mucous injury; Inflammatory response; Cell apoptosis