International guidelines for cardiopulmonary resuscitation recommend mild hypothermia (32-34 degrees C) for 12-24h in comatose survivors of cardiac arrest. To induce therapeutic hypothermia a variety of external and intravascular cooling devices are available. A cheap and effective method for inducing hypothermia is the infusion of large volume, ice-cold intravenous fluid. There are concerns regarding the effects of rapid infusion of large volumes of fluid on respiratory function in cardiac arrest survivors. We have retrospectively studied the effects of high volume cold fluid infusion on respiratory function in 52 resuscitated cardiac arrest patients. The target temperature of 32-34 degrees C was achieved after 4.1+/-0.5h (cooling rate 0.48 degrees C/h). During this period 3427+/-210 mL ice-cold fluid was infused. Despite significantly reduced LV-function (EF 35.8+/-2.2%) the respiratory status of these patients did not deteriorate significantly. On intensive care unit admission the mean PaO(2) was 231.4+/-20.6 mmHg at a F(i)O(2) of 0.82+/-0.03 (PaO(2)/F(i)O(2)=290.0+/-24.1) and a PEEP level of 7.14+/-0.31 mbar. Until reaching the target temperature of <or=34 degrees C the F(i)O(2) could be significantly reduced to 0.63+/-0.03 with unchanged PEEP level (7.23+/-0.36 mbar). Under these conditions the PaO(2)/F(i)O(2) ratio slightly decreased to 247.5+/-18.5 (P=0.0893). Continuing the saline infusion to achieve a body temperature of 33 degrees C, the F(i)O(2) could be further reduced with unchanged PEEP. The infusion of large volume, ice-cold fluid is an effective and inexpensive method for inducing therapeutic hypothermia. Resuscitation from cardiac arrest is associated with a deterioration in respiratory function. The infusion of large volumes of cold fluid does not cause a statistically significant further deterioration in respiratory function. A larger, randomized and prospective study is required to assess the efficacy and safety of ice-cold fluid infusion for the induction of therapeutic hypothermia.