The first intravascular oxygenator and carbon dioxide (CO2) removal device (IVOX), conceived by Mortensen, was capable of removing 30% of CO2 production of an adult at normocapnia with a measurable reduction in ventilator requirements. Through studies of mathematical modeling, an ex vivo venovenous bypass circuit to model the human vena cava, animal models of severe smoke inhalation injury, and patients with acute respiratory failure, the practice of permissive hypercapnia has been established to enhance CO2 removal by IVOX. By allowing the blood pCO2 to rise gradually, the CO2 excretion by IVOX can be linearly increased in a 1:1 relationship. Experimental and clinical studies have shown that CO2 removal by IVOX can increase from 30 to 40 ml/min at a normal blood pCO2 to 80 to 90 ml/min at a pCO2 of 90 mm Hg. In addition, IVOX with permissive hypercapnia allows a significant reduction in minute ventilation and peak airway pressure. Active blood mixing to decrease the boundary layer resistance in the blood can significantly improve O2 transfer by up to 49% and CO2 removal by up to 35%. Design changes can also improve the performance of IVOX. Increased surface area with more fibers and enhanced mixing by increased fiber crimping in new prototypes of IVOX significantly increased CO2 removal. Other groups have used alternative designs to address the limited performance of intravascular gas exchange devices. With improved design and patient management, clinically meaningful gas exchange and reduction in mechanical ventilatory support may be achieved during treatment of severe respiratory failure.