To the Editor:—
Barr et al. 1have provided useful information about the pharmacokinetics and pharmacodynamics of midazolam and lorazepam when administered by infusion to intensive care patients. However, their conclusions may be misleading and are at variance with other published data.
Methodologically, it is notable that Barr et al. 1report a mean duration of lorazepam infusion more than twice that for midazolam (36.94 vs. 15.02 h, respectively), and that the SD of lorazepam infusion duration was approximately 31 h. These differences could easily explain the prolonged time to awakening and extubation associated with lorazepam administration. It is also unfortunate that the authors did not better standardize the analgesic regimens between the two groups, or at least report the doses of fentanyl administered. Differences in drug potency may have been related to the success of the associated analgesic regimen.
Swart et al. previously compared the pharmacology of midazolam and lorazepam in a randomized, double-blind trial of critically ill patients. 2Swart studied a larger and much more ill (Apache II score of approximately 26 vs. 9 in the Barr study) group of patients, one that may be more representative of the type of patients requiring continuous infusions of sedative drugs. In addition, patients received sedative infusions for a much longer duration of time in the Swart study (141 vs. 15 h). Although Swart et al.’ s pharmacokinetic analysis was not as sophisticated as Barr et al.’ s, they did report a longer elimination half life for lorazepam than midazolam (13.8 vs. 8.9 h, respectively). However, there was much more variability in midazolam's half life, with a range extending from 2.2–35.5 h. In contrast to Barr et al. , Swart found that the infusion rate of midazolam required to achieve a comparable level of sedation was approximately 15 times higher than for lorazepam, despite administration of comparable doses of fentanyl. In addition, Swart et al. found that patients achieved the desired level of sedation more frequently with lorazepam than midazolam. Finally, the average daily drug cost was approximately 10 times higher with midazolam than lorazepam.
A comparison of the two studies mentioned above reveals the complexity of sedation of critically ill patients. In the spirit of context-sensitive half-life, the pharmacokinetics of a particular drug may vary markedly with the duration of administration, severity of illness, and so on. Despite the evidence of prolonged sedation in association with lorazepam use in the intensive care unit presented by Barr et al. , lorazepam may be a more effective and cheaper agent than midazolam “in the long run.”