RESUMO
BACKGROUND: The AnaConDa filter permits administration of volatile anesthetic without the use of an anesthesia machine. It is intended for use in the intensive care unit. METHODS: We studied the AnaConDa reflection filter on the bench and in anesthetized patients. The bench analysis used a test lung, a gas analyzer, an intensive care ventilator, the AnaConDa filter, and a syringe pump. We studied a range of tidal volume, respiratory rate, and positive end-expiratory pressure values. We simulated errors during syringe refilling and patient transportation. In 15 anesthetized patients, we used the AnaConDa with constant ventilation variables, a constant sevoflurane infusion rate (4-5 mL/h), and two consecutive fresh gas flow levels. RESULTS: In the bench study, the expired volatile anesthetic fraction decreased linearly with respiratory frequency at constant minute ventilation, and decreased markedly in a hyperbolical manner when tidal volume increased at a constant respiratory rate. Changing the positive end-expiratory pressure level and inspiration/expiration ratio did not modify the AnaConDa's performance. Several safety failures were observed: refilling caused a transient change in AnaConDa output because of a pumping effect, and a standard Luer lock made it possible to connect the halogenate syringe on an IV infusion line. In anesthetized patients, reducing fresh gas flow from 8 to 1 L/min led to a median 40% increase in the expired volatile anesthetic fraction. CONCLUSIONS: This study shows that the device is generally reliable, but that there are several conditions under which it might deliver more anesthetic than intended.