We studied the percutaneous losses of the new inhaled anesthetic, desflurane (I-653), and of isoflurane and halothane during anesthetic administration and elimination in seven healthy male volunteers. Anesthesia was induced and maintained with midazolam, thiopental, and fentanyl. We administered 70% N2O for 30 min, and then administered 2% desflurane, 0.4% isoflurane, and 0.2% halothane concurrently with 65% N2O for 30 min. Inspired, end-tidal, and mixed-expired gas samples were collected during administration of the volatile agents and for 5-7 days of elimination. The right arm and hand of each subject was enclosed in a sealed glass cylinder having a port at each end, one for sampling and both for flushing with N2 after anesthetic administration and every 15 min thereafter. We sampled gases from the cylinder during administration and for the 150 min of elimination and analyzed their anesthetic concentrations by gas chromatography. The surface area of the enclosed portion of the arm was measured, and the total body surface area was calculated. All values were normalized to (i.e., divided by) the end-tidal (alveolar) concentration at the end of administration. During administration, percutaneous loss of halothane was 3.5 times that of desflurane and 2 times that of isoflurane. During elimination, the loss of halothane was 6 times and 2 times greater than the loss of desflurane and isoflurane, respectively. Percutaneous loss of halothane significantly exceeded that of isoflurane. The elimination values included an estimate of elimination after 150 min. The percutaneous loss of each anesthetic was 2- to 3-fold greater during elimination than administration.(ABSTRACT TRUNCATED AT 250 WORDS)