We have reported that percutaneous cardiopulmonary assist systems (PCPS) are effective in treating life-threatening cardiogenic shock that is intractable to treatment with intraaortic balloon pumping (IABP). However, there are few clinical indices that can be used to evaluate the effectiveness of PCPS. End-tidal carbon dioxide (ET-CO(2)) content reflects pulmonary blood flow. We monitored ET-CO(2) continuously and determined whether we could use it as a new index to evaluate the effectiveness of PCPS. Seventeen patients with cardiogenic shock were intubated and evaluated by ET-CO(2) monitoring during PCPS. The etiology of shock included acute myocardial infarction (n=10), acute myocarditis (n=2), recent coronary artery bypass graft (n=1), cardiac rupture (n=1), hypertrophic obstructive cardiomyopathy complicated by ventricular fibrillation (n=1), left atrial myxoma (n=1) and artificial valve malfunction (n=1). PCPS was extremely effective in 10 of 17 patients (58.8%), and they recovered from the cardiogenic shock. The remaining 7 patients did not recover from shock, and died during PCPS. Six of ten patients who recovered from shock were successfully weaned from PCPS and 4 patients had good long-term survival. In the cases where PCPS was effective, the ET-CO(2) measured soon after the beginning of PCPS was significantly higher than in the cases in which PCPS was ineffective. Furthermore, the ET-CO(2) content increased gradually with the improvement in hemodynamics. In contrast, ET-CO(2) content remained low if PCPS was not effective. The ET-CO(2) represents a useful predictor of survival or death and is also a good index for weaning in patients treated with PCPS.