In vivo animal models of cerebral vasospasm: A review

BACKGROUND: Cerebral vasospasm is delayed-onset cerebral arterial narrowing in response to blood clots left in the subarachnoid space after spontaneou s aneurysmal subarachnoid hemorrhage (SAH). Ideally, studies on the pathoge nesis and treatment of cerebral vasospasm in humans should be conducted usi ng human cerebral arteries. Because in vivo experiments using human vessels are not possible, and postmortem pathological examination of human arterie s in vasospasm provides only a limited amount of information, a number of a nimal models of vasospasm have been developed. METHODS: The literature was searched to find all references to in vivo anim al models of SAH and vasospasm. An online search of the medical database ME DLINE was initially performed using the key words "cerebral," "vasospasm," "subarachnoid," "hemorrhage," "animal," and "model." References were checke d to determine the first description of each in vivo animal model. RESULTS: Fifty-seven models of SAH and vasospasm were identified. These mod els used one of three techniques to simulate SAH: 1) an artery was puncture d allowing blood to escape and collect around the artery and its neighbors; 2) an artery was surgically exposed, and autologous blood obtained from an other site was placed around the artery; or 3) blood from another site was injected into the subarachnoid space and was allowed to collect around arte ries. Each technique has advantages and disadvantages. The majority of anim al models of SAH and vasospasm use intracranial arteries; however, extracra nial arteries have also been used recently in vasospasm experiments. These studies seem easier and less costly to perform, but concerns exist regardin g the physiological dissimilarity between systemic and cerebral arteries. CONCLUSION: The model of SAH and vasospasm used most frequently is the cani ne "two-hemorrhage" model, in which two injections of blood into the dog's basal cistern performed 48 hours apart result in greater arterial vasoconst riction than that effected by a single injection of blood. On the basis of its ability to accurately predict what occurs in human SAH, the best model of vasospasm seems to be the primate model in which a blood clot is surgica lly placed around the large cerebral vessels at the base of the monkey's br ain.