Critical Care Medicine - List
http://www.pitt.edu/~crippen/

Lorazepam and midazolam in the intensive care unit: A randomized, prospective, multi center study of hemodynamics, oxygen transport, efficacy and cost. Cernaianu AC, DelRossi AJ, Flum DR et al. Crit Care Med 1996;24:222-228. (February 1996).

Reviewed by David Crippen.


This is a multi center open label study looking at a total of 95 "critically ill" mechanically ventilated trauma patients with pulmonary artery catheters. Mean age 58 years. Sedation started 1 to 2 hours after admission to the ICU. These patients were randomly assigned to receive a total of 8 hours of either (intermittent bolus) lorazepam or (continuous infusion) midazolam for ICU sedation. In addition, All patients received intermittent boluses of 1 - 2 mg of morphine for pain.

Levels of agitation/anxiety were measured using an informal scale:

Assessments of these parameters were made at 5 min, 30 min, 4 and 8 hours.

During the treatment period, there was no statistically significantly different changes in hemodynamic or metabolic variables. The quality of sedation was judged roughly equivalent by both nurses and patients but lorazepam was judged more cost efficient than midazolam.

Comment by David Crippen:

After reading the article comparing lorazepam and midazolam in the intensive care unit (1), I tried the author's calculations and found them to be incomprehensible.

For the lorazepam group the authors note doses of 0.044 mg/kg IV over 1 - 2 minutes, followed by 0.044 mg/kg intermittent injection titrated to clinical effect. Using this calculation, a typical 80 kg patient would be loaded with 3.5 mg IV, then given an additional 3.5 mg IV intermittently. This seems to be a lot more than the 1.6 mg +/- 0.1 mg / 8 hrs used to achieve the same degree of sedation as the midazolam group. The authors do not say how much was given, but patients presumably were dosed at least twice and probably three times since lorazepam lasts about three hours. Therefore, a typical 80 kg patient would have received a total of 10.5 mg in eight hours. Since most of our patients do well with under 2 mg IV every three hours, this seems like a lot of lorazepam. At the prices for lorazepam quoted in the paper ($1.85 per mg) The cost for eight hours of medication would be $19.42.

Conversely, the midazolam group was allegedly given 0.2 - 0.3 mg IV over 1 - 2 minutes. This is a homeopathically small dose. Perhaps they really mean 0.2 - 0.3 mg per kg which for an 80 kg person would be 16 - 24 mg bolus, a very steep dose. Or maybe they mean 2 - 3 mg IV, which would be about right although a little on the light side to compare with their very heavy dose of lorazepam, the ratio of which they state in their text is 1:4. Their price for midazolam is $1.73 per mg. So, figuring a 2 mg bolus and 2 mg per hour, the price would be $31.14. Although they did not perform the calculations, the midazolam costs about twelve dollars more for an eight hour period, not nearly the expense alluded to in the manuscript, and certainly not the therapeutic ratio stated.

The authors state in their conclusion: "The dose of midazolam required for sedation is much larger than the dose of lorazepam required for sedation, and midazolam is therefore less cost-efficient". Their calculations do not seem to support this conclusion. To further muddy the issue, both populations of patients received unmeasured boluses of morphine during the course of the study. Since narcotics potentiate benzodiazepines, the timing and frequency of these doses becomes very important and could possibly skew the results dramatically. Studies have been done in the past comparing doses of short acting sedatives directly to a Ramsay Scale level, demonstrating effective titratability (2). The authors ostensibly begin each group with Sedation Scores that do not show the patient is agitated (3.2 / 3.3 for each group respectively postulating 1 as agitated and 6 as unconscious). Following treatment, each group seemed to be sedated adequately according to a seemingly undefined "Patient Comfort Score", bearing no resemblence to the original sedation assessment.

Clinically, there are many roads that lead to agitation in the ICU, and each of these roads may require different drugs given by different routes, tailored to individual patients rather than a single maxim. Instead of advocating a flexible regime of the most appropriate sedative in the most effective route of administration for specific disorders such as pain, anxiety or discomfort, the authors seem to say that lorazepam works for all disorders and is cheaper than it's next competitor, ergo; it is the drug of choice for agitation in the ICU.

This article is yet another in a series of papers (3,4,5) supported by a grant from Wyeth-Ayerst Laboratories, the makers of lorazepam suggesting that Brand A is better than Brand B ( brought to you by the makers of Brand A). The authors of this paper do not speculate on why each of these drugs may have it's own indication and niche in a logical ICU drug regime. After reading their conclusion that apples and oranges are indeed, round, firm and have seeds in the middle, I wonder how this paper managed to get through the peer review process.

  1. Boyd O, Mackay CJ, Rushmer F, et al: Propofol or midazolam for short-term alterations in sedation. Canadian Journal of Anaesthesia. 40(12):1142-7, 1993 Dec.

  2. Lorazepam and midazolam in the intensive care unit: A randomized, prospective, multi center study of hemodynamics, oxygen transport, efficacy and cost. Cernaianu AC, DelRossi AJ, Flum DR et al. Crit Care Med 1996;24:222-228. (February 1996).

  3. Pohlman AS, Simpson KP, Hall JB: Continuous infusions of lorazepam versus midazolam for sedation during mechanical ventilatory support: A prospective, randomized study. Crit Care Med 1994;22:1241-1247.

  4. Deppe SA, Sipperly ME, Sargent AI: Intravenous lorazepam as an amnestic and anxiolytic agent in the intensive care unit: A prospective study. Crit Care Med 1994;22:1248-1252.

  5. Hadbavny AM: Am J Hosp Pharm Vol 50, April 1993: 660-661.Promotion of Cost effective benzodiazepine sedation.

Rainer G. Gedeit:

I agree, this paper makes me wonder about a few things.

  1. As previously stated, the methods section is unclear on the dose of midazolam given. Is it 0.2-0.3 mg or mg/kg.

  2. The assessment scales are impossible to quantify, since these variables are subjective to both patient and observer.

  3. Reading the paper, I wonder why there are more than a page and a half of graphs that show nothing that a simple statement or table would have shown. No mention of number of boluses given in the lorazepam group.

  4. How many patients were used to obtain the "patient comfort score"? And is this a valid measurement of what the patient really experienced. How can you be sure that the patient understands what 8 hour time period you are asking about, since I assume a critically ill intubated patient would still be receiving sedation after the study period.

  5. Math - An 80 kg patient would maximally receive a 3.5 mg bolus of lorazepam and then if it was required every hour, a maximum of 3.5 mg X 8 hours = 27 mg. A total of 30.5 mg. A patient receiving the maximum dose of lorazepam would receive 0.3mg X 80 kg = 24 mg and then 3 mg/hr X 8 hours = 24 mg. For a total of 48 mg. If the same patient received the minimum dose, this would be 24 mg. With those kinds of numbers it seems to me that the midazolam group was designed to fail. Since the lorazepam patients probably never got dosed every hour.

  6. Would lower doses of midazolam give a similar effect? Can't tell from this study.

The conclusion reached by this study is that both drugs seem to work at the doses given. Anything else is speculation, and should have been stated as such.

David Crippen:

And another thing, there is the little matter of what you are getting for your expenditure. Now we enter the "opinion and experience" zone.

"Boluses" of sedatives, by their nature, tend to act like the furnace in your house. When it reaches an uncomfortable level, the heater kicks on and runs the temp level quickly to a level that is actually too warm. Then it kicks off and the temperature slowly drifts down through the "comfort zone" until it kicks out the bottom again and the house becomes too cold. Then the cycle begins again. Giving a bolus of a sedative puts a lot of drug into the blood supply at the same time, then it slowly decays until it falls out the bottom and it's time for another bolus. A fairly big problem in critically ill folks is "too much too soon" (possibility of oversedation and ventilatory consequences) and "falling out the bottom". This is where the patient is unprotected and cardiac ischemia develops as well as increased intracranial pressure and other nasties.

Continuous infusions of analgesics and sedatives are a very effective method of avoiding the "valleys" inherent in bolus medication therapies that initiate a "peak" of therapeutic action followed by a variable period of "valley" where the patient has little or no drug effect. Current literature suggests that high risk cardiac patients are jeopardized by relatively brief periods analgesia ineffectiveness ( Mangano DT, Siliciano D, Hollerberg M et al. Postoperative myocardial ischemia: Therapeutic trials using intensive analgesia following surgery. Anesthesiology 76;342-353, 1992). Intermittent periods of sympathetic stimulation due to ineffective analgesia and sedation can cause relatively profound deleterious effects on compromised myocardium. Continuous intravenous infusions of short acting agents such as midazolam, propofol and fentanyl allow titration of plasma level effects to a fluctuating baseline of pain, anxiety and discomfort. This real time titration of natural fluctuations may occur with minimum hemodynamic and respiratory suppression. Increased costs of newer short acting agents are justified if complications are avoided as patients achieve more effective analgesia and sedation, avoiding blanket effects of less selective regimes.

The authors cite a paper by Pohlman, et al (Pohlman AS, Simpson KP, Hall JB: Continuous infusions of lorazepam versus midazolam for sedation during mechanical ventilatory support: A prospective, randomized study. Criti Care Med 1994;22:1241-1247.) in which they challenge the assertion (not very convincingly, IMHO) that continuous infusions are of any benefit. I should note that the Pohlman paper was also supported by a financial grant from the folks who have a vested interested in finding Brand X "better" than Brand Z, which brings me to my final comment.

Errington Thompson:

The study is a piece of crap. Not only for the reasons that you state but also for several other reasons which I mention in one of my replies. Other considerations - there are a large number of trauma patients in this study - what percentage of the truama patients were admitted with EtOH, benzo's, etc.? was it the same in both groups? It seems to me that the trauma pt admitted with cocaine, marj., and an EtOH level of 350 would suck up the benzo's much more that the 75 yo who had a carotid. This study is not worth the time that you took to analyze it.

I would not be so critical of CCM. There are plenty of studies which are published in major journals which are trash. The prestigious England Journal published the steriods in spine trauma study several years ago - crap! Several articles have been published which have shown that steriods are of no benefit in penetrating trauma and blunt trauma. That same journal published the vent. weaning article last year from Spain - crap. In this article vent changes were only made twice a day!

If you send a paper to the right reviewer who is not meticulous; if the paper is politically correct for that journal - - - crap can and does get published.