This study evaluated the sputum penetration of cefpirome following slow intravenous infusion of 0.5 and 1.0 g using a comparative cross-over design to reduce variability. Five patients with chronic respiratory tract infections were randomized to receive either 0.5 g followed by 1.0 g, or by 1.0 g followed by 0.5 g cefpirome, by slow intravenous infusion over 1 h, with a 24-h wash-out period between each treatment. With the exception of one patient, sputum concentration correlated well with plasma concentration. Higher sputum levels of cefpirome were achieved following the higher dose.