CEREBRAL ischemia, although infrequent, is a potentially devastating complication of anesthesia and surgery. The exquisite vulnerability of the brain to cessation of blood flow has fostered a substantial investigative effort to identify pharmacologic agents that might reduce ischemic cerebral injury. Among these, anesthetics have long been considered logical candidates, given their ability to suppress cerebral metabolic rate, to antagonize glutamate-mediated excitotoxicity, and to enhance inhibitory synaptic transmission. Indeed, a large number of studies have shown that anesthetics can reduce ischemic injury in models of global, 1,2focal, 3–5and hemispheric ischemia. 6In fact, the magnitude of the neuroprotective efficacy of anesthetic agents is similar to that of antiglutamatergic agents. In most of these investigations, the recovery period after the initiation of ischemia has been relatively short, on the order of 1 to 5 days. Recent data have shown that postischemic neuronal injury is a dynamic process in which neurons continue to die for a long time after ischemia (at least several weeks). 7This begs the question of whether anesthetic neuroprotection, evident early after ischemia, is sustained after a much longer recovery period. This question is addressed by two meticulously conducted investigations by Bayona et al. 8and Elsersy et al. 9in this issue of the Journal.
Bayona et al. 8show that propofol infusion decreased infarction volume 3 days after insult in an endothelin-induced striatal ischemia model. When the animals were evaluated 3 weeks after ischemia, no difference between propofol-treated or control animals could be detected histologically. In the study of Elsersy et al. , 9isoflurane reduced neuronal injury within the hippocampus of rats subjected to global cerebral ischemia when the injury was evaluated 5 days later. However, when injury was evaluated 3 weeks or 3 months after insult, no difference in injury, either morphologically or neurologically, could be detected. The results of these studies are consistent with those published by Kawaguchi et al. , 10who demonstrated that isoflurane-mediated neuroprotection, apparent 2 days after ischemia, was not sustained 2 weeks later. This transient neuroprotection is by no means limited to anesthetic agents; sustained protection has not been achieved with the administration of the N -methyl-d-aspartate antagonist dizocilpine, 11,12alpha-amino-3-hydroxy-5-methyl-4-isoxazole propionic acid (AMPA) antagonist 2,3-dihydroxy-6-nitro-7-sulfamoyl-benzo(F)quinoxaline (NBQX), 13and postischemic hypothermia. 14Common to these investigations is the slow progression of injury such that neurons that were initially protected nonetheless underwent delayed death.
This Editorial View accompanies the following articles: Bayona NA, Gelb AW, Jiang Z, Wilson JX, Urquhart BL, Cechetto DF: Propofol neuroprotection in cerebral ischemia and its effects on low-molecular-weight antioxidants and skilled motor tasks. ANESTHESIOLOGY 2004; 100:1151–9; Elsersy H, Sheng H, Lynch JR, Moldovan M, Pearlstein RD, Warner DS: Effects of isoflurane versus fentanyl/nitrous oxide anesthesia on long-term outcome from severe forebrain ischemia in the rat. Anesthesiology 2004;100:1160 –6.
These studies highlight the importance of long-term neuronal viability as an endpoint in experimental studies of cerebral ischemia and pharmacologic neuroprotection. The aforementioned glutamate antagonists (N -methyl-d-aspartate and alpha-amino-3-hydroxy-5-methyl-4-isoxazole propionic acid receptor antagonists) entered clinical trials in stroke patients, in part based on very strong preclinical studies that demonstrated robust short-term neuroprotection; long-term neuroprotection was not systematically evaluated. Subsequent studies that used a recovery interval of greater than 2 weeks demonstrated no neuroprotective efficacy that could be attributed to glutamate antagonists. Although a number of factors contributed to the failure of the glutamate antagonists in the clinical setting, the lack of long-term neuroprotection by these agents was undeniably a major contributor. This inevitably leads to the question of what constitutes an appropriate postischemic time interval for the evaluation of neurologic outcome in studies of cerebral ischemia and neuroprotection. Clearly, the requisite recovery period is dependent on the experimental model of cerebral ischemia that is used. Nonetheless, the bulk of the available data indicate that, in studies in which neuroprotection is the primary outcome measure, a postischemic recovery interval of 2 weeks at a minimum, and preferably 4 weeks or longer, is needed to ensure that short-term neuroprotection is sustained. In fact, a cogent argument can be made that such a long-term recovery period should now be considered to be a standard for such studies. Of course, this standard should not apply to mechanistic studies in which considerably shorter postischemic intervals are essential.
Although the mechanisms that underlie the progression of injury in the ischemic brain have not been clarified, neuronal apoptosis undoubtedly plays a role. 15–18In models of focal ischemia, apoptotic neurons are localized within the outer boundary of the evolving in-farct. 16Administration of agents that inhibit apoptosis, such as the protein synthesis inhibitor cycloheximide, mitigate the propagation of injury. 19By contrast, isoflurane does not prevent apoptosis. 20Kawaguchi et al. have shown the increase in the size of cerebral infarction in isoflurane-treated animals subjected to focal ischemia parallels the appearance of markers of apoptosis. 20Indeed, the broad spectrum caspase inhibitor zVAD-fmk (benzyloxycarbonyl-Val-Ala-Asp-fluoromethylketone) and the caspase 8 inhibitor IETD-fmk (z-Ile-Glu-Thr-Asp-(IETD)-fluoromethylketone) 21decrease neuronal apoptosis, prevent infarct extension, and sustain isoflurane-mediated protection. Prevention of apoptosis clearly presents a target for future interventions.
The development of inflammation in the postischemic brain has been well documented and may well be an important initiator of the apoptotic process. Expression of a variety of adhesion receptors leads to the recruitment of platelets and leukocytes within the cerebral circulation. 22Experimental therapies directed against adhesion molecules have met with some success. Of considerable importance is the activation of microglia by interleukin-1. 23The use of interleukin-1 antagonists decreases cerebral injury in a variety of stroke models, and clinical trials to evaluate their efficacy in humans are under way. Microglial activation can be demonstrated as late as 2 months after ischemia. 7This sustained micro-glial activation has been taken as evidence of a chronic encephalopathic process, initiated by ischemia, that can lead to ongoing neuronal loss. 7It is therefore not surprising that neuroprotective agents, administered only briefly prior to or after ischemia, do not provide long-term neuroprotection. In this regard, cerebral ischemia may be thought of as a chronic inflammatory disorder, and the achievement of sustained neuroprotection may well require chronic antiinflammatory therapy.
An interesting aspect common to the investigations of both Bayona et al. 8and Elsersy et al. 9is the finding that the size of the lesion decreased with longer recovery periods. Endothelin induced striatal infarction was smaller 3 weeks after insult than 3 days after insult. Similarly, the number of viable neurons within the CA1 sector of the hippocampus increased in both isoflurane and fentanyl-nitrous oxide groups 3 months after ischemia. It is now quite apparent that the brain is a highly plastic organ and that a substantial turnover of neurons in the adult brain occurs. The generation of neurons and glia from progenitor cells, neurogenesis, has been well characterized. 24,25In fact, a number of investigators have shown that neurons that develop from progenitor cells can be found in injured brain after focal and global ischemia. 26–28This process is clearly an attempt at regeneration by the injured brain. Although the administration of anesthetics during the ischemic interval does not appear to interfere with neurogenesis, additional experimental work to clarify the time course of neurogenesis and its modulation by anesthetics is needed. The reparative processes, however, add an interesting twist to the problem of evaluating the neuroprotective efficacy of therapeutic interventions. The need for long-term survival studies is now clear. However, the gradual improvement in neurologic function in control subjects would tend to reduce the difference between treated and untreated subjects. A plausible result of this is that an agent that does reduce injury may be considered to be ineffective if long-term differences between treated and untreated subjects are obscured by regenerative processes. As noted by Elsersy et al. , 9long-term neuroprotection studies may have to incorporate a means by which the contribution of ongoing neurogenesis is separated from the survival of “protected” neurons.
It is clear that the pathophysiology of cerebral ischemia is complex. A number of diverse processes are initiated by the ischemic insult. Couple this diversity with temporal and spatial differences among these processes, and it rapidly becomes apparent that a single pharmacologic intervention is unlikely to result in sustained neuroprotection in all (and probably not any) cell populations. A combination of different approaches that target specific stages of the evolution of ischemic injury may be required. Simply stated, the magazine of our cerebral protection “gun” is empty. Currently, there are no magic bullets.