In Reply:—

We appreciate the points raised by Dr. Deem. Our manuscript addresses the differences in infusion duration. 1To briefly recapitulate: this was a double-blind, randomized study. Occasionally randomization fails to divide patient covariates evenly between groups. The smaller the study, and the more covariates considered, the more likely it is that not all covariates will be evenly divided between groups. Since randomization failed to provide similar durations of infusions, we used a model-based approach to draw clinical inference from the fundamental PK/PD characteristics. Those inferences support our conclusions. The half-lives reported for midazolam (10 h) and lorazepam (16 h) in our manuscript are consistent with the published PK of midazolam 2–7and lorazepam. 7–12 

As explained in the manuscript, fentanyl was administered by target-controlled infusion, set to 1.5 ng/ml. Because this was a randomized study, the concept of “standardizing analgesic regimens between groups” is irrelevant. There was no difference in the fentanyl administered to the two groups, nor in the “success of the associated analgesic regimen.”

Swart et al . reported that 5 of 6 patients with delayed midazolam elimination had been treated with erythromycin for more than 2 days. 7Obviously such patients should be treated with drugs not metabolized by CYP 3A4, such as lorazepam. Absent those patients, the variability in midazolam and lorazepam reported by Swart was similar to what we observed in our patients, none of who received erythromycin. We cannot explain why Swart et al . were unable to achieve the desired level of sedation with midazolam, despite average  infusions rates of 16 mg/hr. In our experience, adequate sedation, including complete unconsciousness, can be achieved with either drug, and at far lower doses of midazolam than reported by Swart.

As Dr. Deem notes, Swart et al  report a 15-fold difference in infusion rates between midazolam and lorazepam. This is much higher than in our study or in other published comparisons of midazolam and lorazepam in the intensive care unit. 13,14Of note, the dosing differences for midazolam and lorazepam reported by Swart et al . correspond exactly with the concentration differences of drugs in their syringes: 0.33 mg/ml of lorazepam versus  5 mg/ml of midazolam. Swart et al . titrated to deep levels of sedation, where drug effect is difficult to assess precisely. In this study design, lack of precise titration to drug effect would be expected to produce a 15-fold potency difference by default. We believe this is the most likely explanation for their anomalous results.

Our manuscript documents that both midazolam and lorazepam are effective drugs for intensive care unit sedation, and provides guidelines in administering them to achieve comparable results. We did not address which drug was cheaper in the long run, which is a complex question involving far more than drug acquisition costs.

Barr J, Zomorodi K, Bertaccini EJ, Shafer SL, Geller E: A double-blind, randomized comparison of IV lorazepam versus  midazolam for sedation of ICU patients via  a pharmacologic model. A nesthesiology 2001; 95: 286–98
Dirksen MS, Vree TB, Driessen JJ: Clinical pharmacokinetics of long-term infusion of midazolam in critically ill patients: Preliminary results. Anaesth Intensive Care 1987; 15: 440–4
Maitre PO, Funk B, Crevoisier C, Ha HR: Pharmacokinetics of midazolam in patients recovering from cardiac surgery. Eur J Clin Pharmacol 1989; 37: 161–6
Boulieu R, Lehmann B, Salord F, Fisher C, Morlet D: Pharmacokinetics of midazolam and its main metabolite 1-hydroxymidazolam in intensive care patients. Eur J Drug Metab Pharmacokinet 1998; 23: 255–8
Zomorodi K, Donner A, Somma J, Barr J, Sladen R, Ramsay J, Geller E, Shafer SL: Population pharmacokinetics of midazolam administered by target controlled infusion for sedation following coronary artery bypass grafting. A nesthesiology 1998; 89: 1418–29
Bjorkman S, Wada DR, Berling BM, Benoni G: Prediction of the disposition of midazolam in surgical patients by a physiologically based pharmacokinetic model. J Pharm Sci 2001; 90: 1226–41
Swart EL, van Schijndel RJ, van Loenen AC, Thijs LG: Continuous infusion of lorazepam versus medazolam in patients in the intensive care unit: Sedation with lorazepam is easier to manage and is more cost-effective. Crit Care Med 1999; 27: 1461–5
Greenblatt DJ, Comer WH, Elliott HW, Shader RI, Knowles JA, Ruelius: Clinical pharmacokinetics of lorazepam. III. Intravenous injection. Preliminary results. J Clin Pharmacol 1977; 17 (8–9): 490–4
HW Kyriakopoulos AA, Greenblatt DJ, Shader RI: Clinical pharmacokinetics of lorazepam: a review. J Clin Psychiatry 1978; 39: 16–23
Greenblatt DJ, Shader RI, Franke K, MacLaughlin DS, Harmatz JS, Allen MD, Werner A, Woo E: Pharmacokinetics and bioavailability of intravenous, intramuscular, and oral lorazepam in humans. J Pharm Sci 1979; 68: 57–63
Greenblatt DJ: Clinical pharmacokinetics of oxazepam and lorazepam. Clin Pharmacokinet 1981; 6: 89–105
Aaltonen L, Kanto J, Arola M, Iisalo E, Pakkanen A: Effect of age and cardiopulmonary bypass on the pharmacokinetics of lorazepam. Acta Pharmacol Toxicol 1982; 51: 126–31
Pohlman AS, Simpson KP, Hall JB: Continuous intravenous infusions of lorazepam versus midazolam for sedation during mechanical ventilatory support: A prospective, randomized study. Crit Care Med 1994; 22: 1241–7
Krosner SM, Lin HM, Nagy K: Efficacy and cost: Comparison of lorazepam and midazolam for continuous IV sedation in critically injured trauma patients. Crit Care Med 1994; 22: A54