COMPANION articles in this issue of Anesthesiology deal with the pharmacology of dexmedetomidine in humans and compare its sedative, ventilatory, and analgesic properties to those of the potent opioid narcotic remifentanil.1,2The marketing authorization label for dexmedetomidine, a highly selective α2-adrenoceptor agonist, stipulates its use for sedation in mechanically ventilated patients, and it is within this clinical context that these articles should be considered. Because the use of remifentanil has recently been validated in this clinical setting,3it is a more relevant comparator than it may seem at first blush.
The best method of sedating the mechanically ventilated patient in the intensive care unit has vexed clinicians who have virtually thrown the pharmacopoeia at this problem. However, we still encounter patients for whom the first-line combination of a γ-aminobutyric acid-mediated compound, such as propofol or a benzodiazepine, together with an opioid narcotic, does not accomplish the goal of safely providing sedation with cardiorespiratory stability that facilitates weaning from the ventilator.4
The findings of Hsu et al. ,1using a sophisticated pharmacokinetic approach but possibly controversial analytical methodology, build on a collection of studies extending over more than 15 yr that established the benign effect of dexmedetomidine5or clonidine6on ventilation, especially when compared to an opiate narcotic. Hsu et al. 1commented on the fact that the introduction of carbon dioxide (in the hypercarbic ventilatory response phase of the study) resulted in an arousal from dexmedetomidine-sedated and at times deeply asleep subjects; a similar “awakening” was noted when the subjects were observed in their “pseudonatural” sleep phase of the study protocol. The fact that even deeply sedated patients receiving dexmedetomidine can be easily aroused has been noted before,7albeit by auditory and tactile stimuli, and draws attention to a recent rodent study that establishes the similarity of the neurologic substrates involved in the hypnotic state produced by dexmedetomidine and that which occurs during non-rapid eye movement sleep.8In a functional magnetic resonance imaging crossover study in human volunteers, dexmedetomidine induced no significant difference in the blood flow signal compared with that seen in the natural sleep state.9These findings contrast with those seen during treatment with γ-aminobutyric acid-mediated hypnotic/sedative agents such as benzodiazepines, in which a qualitatively different pattern of neuronal activity was found in humans.9
Is the similarity between dexmedetomidine-induced sleep and non-rapid eye movement sleep necessarily good for sedation in the intensive care unit setting? What's “good” about a good night's sleep? Although reparative and restorative functions are facilitated by the neuroendocrine milieu that accompanies natural sleep, the salubrious properties of sleep are usually considered only in the context of the morbidity and even mortality of the sleep-deprived state.10The intensive care unit setting is not conducive to a good night's sleep; in fact, the typical nonsurgical intensive care unit patient has less than 2 h of encephalographic sleep within a 24-h epoch.11,12It is hypothesized, although not yet proven, that sleep deprivation is pathogenically involved in the development of delirium and psychotic reactions that occur with a frequency of 60–80% in mechanically ventilated patients.13Although it seems intuitive that avoidance of sleep deprivation can be best provided by drugs that most closely resemble the neurobiology and physiology of natural sleep, this has not yet been confirmed.
In the accompanying article, Cortinez et al. 2compared the analgesic effects of systemically administered dexmedetomidine and remifentanil in humans using an experimental heat pain model.2Clinical trials reveal that α2agonists produce significant analgesia in humans when administered by the intrathecal or epidural routes14; however, the analgesic action of systemically administered α2agonists, assessed by a reduction in the requirement for postoperative opiate narcotics, is modest at best and may be confounded by the coexistent sedative effect.15Human experimental pain studies examining the analgesic profile of systemically administered α2agonists paint an inconsistent picture. Although pain intensity decreased modestly in experiments using the cold pressor test,16only moderate attenuation of the unpleasantness of pain was reported in a model of ischemic pain, no reduction in pain was observed in studies using noxious heat or electricity, and no antihyperalgesic or antiallodynic effects were detected in models of secondary mechanical hyperalgesia.17,18How can the attenuation of heat-evoked pain by dexmedetomidine, now reported by Cortinez et al. ,2be reconciled with these earlier studies?16–18
The authors are to be commended for the use of advanced pharmacokinetic-pharmacodynamic modeling techniques, but differences in methodology and analysis must be considered further. Interpretation of analgesic drug studies may be confounded by the placebo effect, unblinding of subjects, and carryover phenomena. The study by Cortinez et al. 2does not control for placebo effects and blinding (the two of eight subjects who received placebo were excluded from the final analysis). Significant carryover effects may have resulted because all subjects studied during the dexmedetomidine infusion had first received remifentanil. Even though the investigators allowed time for washout to be effected as evidenced by the return of the visual analog score to baseline, enough opiate narcotic may still be present to produce the well-described synergistic analgesic interaction with α2agonist.19
The analgesic effect at each drug concentration was quantified by plotting individual noxious heat-versus -pain intensity functions, an approach that allows a more comprehensive characterization of an analgesic drug profile than algorithms examining a single pain intensity (e.g. , pain threshold).20However, it is not possible to determine whether their sigmoid Emax model best fits their findings, because the raw data are not provided. Psychophysical experiments suggest that the relation between increments in noxious heat and visual analog pain scores is best described by an exponential function21; therefore, an alternative modeling approach has been advocated to take these issues into consideration.22Limitations inherent to the sigmoid Emax model may explain why the findings of Cortinez et al. differ from those of other studies.
How can the findings in these two articles be applied for the sedation of mechanically ventilated patients? It is likely that weaning from the ventilator can be accomplished with less agitation in patients continuously treated with dexmedetomidine than in patients whose sedative drugs may have to be discontinued to avoid ventilatory depression. The ease with which dexmedetomidine-sedated patients can be aroused may facilitate a “daily wake-up” routine that has been shown to improve outcome significantly in mechanically ventilated patients.23The similarity between dexmedetomidine-induced hypnosis and natural sleep may maintain cognitive and immunologic function that deteriorates in sleep-deprived states. Until otherwise demonstrated, it may be prudent to include an opiate narcotic to enhance modest analgesic effects of systemically administered dexmedetomidine when pain is likely to be a significant component of a patient in the intensive care unit. Without randomized clinical trials with an appropriate comparator, each of these conclusions should be considered speculative.
