Origins of the Concept of Vasospasm
- angiographic vasospasm
- brain ischemia
- cerebral aneurysms
- delayed cerebral ischemia
- subarachnoid hemorrhage
Celsus may have described subarachnoid hemorrhage (SAH) when he wrote in A.D. 30, “The characteristic marks… are strong shivering, nervous relaxation, dimness of sight, delirium, vomiting together with a suppression of voice, besides these symptoms there is a violent pain chiefly about the temples or occiput”.1 Hippocrates in Aphorisms on Apoplexy probably described a case: “When persons in good health are suddenly seized with pains in the head, and straightway are laid down speechless, and breathe with stertor, they die in seven days.” Neglecting the patient who would in those days die rapidly without articulating their history, could this be a description of “vasospasm,” or in current terminology, delayed cerebral ischemia (DCI)?
Aneurysms were well known to the ancients, but those located intracranially eluded detection by virtue of being encased in the cranium and not palpable or large and easily accessible to clinical or postmortem assessment. McDonald and Korb2 credited Biumi with describing in 1778 the first pathologically verified case of ruptured intracranial aneurysm. Weir wrote that Gull may have been the first to describe DCI in a 30-year-old female in 1859: “while walking, she suddenly called out, ‘Oh, my head’ and put up her left hand. She vomited, and her friend thought, fainted. After a brief interval she partially recovered, and was able to walk back to her residence with the support of two men. When admitted to the hospital at noon the following day, only a slight impression could be made by any attempt to rouse her. The right arm was quite paralysed, the muscles flaccid; the right leg in the same condition….”3
The text goes on to describe improvement in her condition such that on day 4 she could talk and eat. However, on the fifth day she deteriorated and died and was found to have SAH in the left Sylvian fissure at autopsy. There was softening of the left hemisphere, blood in the ventricles, and 2 middle cerebral artery aneurysms, one of which was ruptured (Figure).3–22
Hunter4 was one of the earliest to state that “There are 3 states in which an artery is found, viz. first, the natural pervious state; second, the stretched; and third, the contracted state…. The contracted state of an artery arises from the action of the muscular power….” Much of their knowledge was based on observation of trauma-induced constriction of arteries, a finding that led to theories that cerebral angiographic vasospasm was because of mechanical factors, that persisted even into the 1980s. Systemic arteries were observed to be innervated and much early work also focused on these nerves being important in vasospasm. Writing in 1728, Lancisi thought vasospasm caused aneurysms of peripheral arteries by increasing the pressure proximal to the spasm. Although I have not identified the basis for his conclusions, Luys23 wrote that the cerebral circulation was necessary for consciousness but also, in turn, the brain regulated its own circulation in that neuronal activity seemed to be associated with increased blood flow. By 1911, Cow24 wrote that chemicals such as carbon dioxide and amyl nitrite had been known for years, probably since the mid 19th century, to lower blood pressure by dilating arteries and that other substances constricted them.5 Cerebral vasospasm was used to explain transient ischemic attacks and other transient neurological phenomena.
Although he was not the first to study brain vasomotor responses, Florey6 observations in 1925 of constriction of cat cerebral arteries in vivo in response to electric, mechanical, and chemical stimulation—barium chloride was a powerful constrictor and nitrites were vasodilators, would later become an important basis for early experimental models of angiographic vasospasm.
The invention of lumbar puncture by Quinke in 1891 led to the ability to diagnose SAH antemortem. Around the same time, Symonds25 considered Froin26 thesis important in giving early description of changes in the cerebrospinal fluid after SAH, findings which remain correct to this day. The color changes from bright red to cherry red ≈5 days after SAH, the erythrocyte count declines after the third day and the supernatant fluid becomes increasingly yellow after the first day. The clinical course of untreated SAH, cerebrospinal fluid findings, etiologies, and key prognostic factors were startlingly well described and known even at that time, the main shortcomings being lack of methods to image what was occurring antemortem as well as limited understanding of the molecular and biochemical basis for the changes.
It was not until blood vessels could be visualized that vasospasm could be observed. Antunes7 reported that Haschek and Lindenthal27 were among the first to visualize blood vessels on radiographs by injecting opaque solutions into them in cadavers in 1895 and Berberich and Hirsch28 obtained extremity angiograms in 1923. Moniz, however, is credited with developing cerebral angiography. Robertson8 may have been the first to recognize the association of cerebral infarctions with ruptured aneurysms. In 5 pathologically examined cases, the infarctions are suggestive of DCI and of the types of infarctions or cortical damage seen under thick subarachnoid blood clots. He hypothesized the infarctions were caused by temporary arterial spasm because the arteries did not contain thrombi or emboli. Subsequent autopsy series by Birse and Tom29 identified focal, diffuse, patchy cortical necrosis in some cases that can now be observed radiologically. In 1964, Crompton30 associated cerebral infarctions at autopsy with the location of subarachnoid blood and with angiographic vasospasm.
There were several contemporaneous descriptions of angiographic vasospasm.9,10,31 Falconer31 thought it might be secondary to the contrast medium and does not seem to have been convinced as to its existence. The most well-known initial description was by Ecker and Riemenschneider9 who described 6 cases of reversible intradural arterial narrowing in patients with untreated, ruptured aneurysms. An early hint at the time course was their observation that spasm was only observed on angiograms done within 23 days of SAH. They wrote that the function of arterial smooth muscle was unknown as was whether the arteries narrowed because of the smooth muscle or elastic tissue in their walls. It was already known that mechanical stimulation, such as after gunshot wounds or other traumatic injuries, of systemic and cerebral arteries led to constriction. The arteries also were known to be innervated, leading the authors to postulate that angiographic vasospasm was because of mechanical effects of the SAH or perhaps because of activation of perivascular nerves. They speculated that patients who did not develop vasospasm would have continued bleeding from the aneurysm and die and thus that vasospasm was life-saving. However, they also suggested it could be detrimental by restricting blood flow to the brain. Ecker32 presented the angiographic findings of vasospasm at the Cushing meeting in 1951. He later wrote that senior members were shaking their heads in disbelief during his presentation. There was 1 man who smiled, as if understanding. Ecker later wrote: “I addressed the rest of my remarks to him. In discussion, after some of the older men denied the existence of cerebral arterial spasm…. I went to the unknown smiling man in the front to thank him for his encouragement. He answered, smiling, ‘I don’t speak English.’”9,32
Pool33 described the typical clinical picture of DCI in 1958. Treatment focused on topical application or intra-arterial injections of papaverine or procaine at the time of surgery. These treatments arose out of experiments in animals showing that these drugs reversed acute, mechanically induced spasm of cerebral arteries. Because papaverine was known to act directly on smooth muscle, and procaine on nerves, they interpreted this as evidence that angiographic vasospasm was because of neural mechanisms and to smooth muscle contraction.33 They considered the possibility that it was something in the blood but thought this was an inadequate explanation because they sometimes observed vasospasm at craniotomy after the SAH had cleared away.
From the late 1950s until even in to the 1980s, controversies persisted about SAH, including about timing of surgery. Rebleeding was widely recognized as a cause of death but early surgery was particularly hazardous. Intracranial aneurysm surgery often was thought best performed 7 to 10 days after SAH, during what is now known to be the worst time when there is a high risk of DCI.1 Without knowing that angiographic vasospasm did not begin until 3 or 4 days after SAH, they misinterpreted its appearance after aneurysm surgery as being postoperative spasm because of the surgery. In any case, by the late 1950s, it seems to have been recognized that angiographic vasospasm could be associated with focal neurological deficits and progress to coma and death, and that earlier reports had not usually distinguished this DCI from rebleeding. Walton1 wrote “Hemiparesis and other neurologic signs in the extremities cleared up completely in 14 patients during their stay in hospital; probably in these cases neurologic deficit resulted from temporary ischaemia or from compression….” Surgery in the setting of unrecognized angiographic vasospasm combined with use of induced hypotension both preoperatively to prevent rebleeding, and also intraoperatively to reduce the risk of intraoperative aneurysm rupture, often indirect surgery by carotid ligation and that patients were usually volume depleted led to predictably poor results.
It had been known for >20 years that chemicals, including serotonin, derived from blood, contracted cerebral arteries but it was Echlin34 who conducted extensive experiments building on the previous studies of cerebral arterial constriction in response to mechanical and chemical stimulation. He noted cerebral arterial constrictions in monkeys in response to fresh blood, leading him to advocate a search for substances in blood responsible for angiographic vasospasm. Experimental models of vasospasm that initially involved topical applications of drugs or blood to exposed cerebral arteries were later changed to include injections of blood into the subarachnoid space, inducing SAH by perforating cerebral arteries or by placing clotted blood into the subarachnoid space. Numerous studies showed these tended to produce more prolonged vasospasm, but the reason for this was not clarified until computed tomographic scan studies in humans confirmed the role of persistent SAH in causing angiographic vasospasm. Nevertheless, Kennady11 was already studying methods to speed clearance of subarachnoid blood in dogs. Alksne et al35 used plasmin in experimental SAH for the same purpose in studies coinciding with efforts to remove SAH surgically and predating the tissue plasminogen activator era.
Smith and Robertson must be credited with their vision to host the first of what became known as the vasospasm meetings in Jackson, Mississippi, in 1972. It was attended by 18 investigators. Smith was best known for investigating structural changes in the cerebral arteries in response to SAH, the role of heparin and in his disciple’s work, such as Clower et al,36 whose work on platelets and the endothelium has been recently rediscovered.37 Robertson, with White,38 also made many contributions to understanding vascular contractions after SAH. Subsequent vasospasm meetings attracted up to several hundred participants. Much of the initial experimental work focused on the role of innervation of the cerebral arteries, catecholamines and their antagonists as treatments for vasospasm, none of which panned out. There also was a Princeton Conference that year, where events that played out in 1975 between Millikan39 and neurosurgeons, including Thor Sundt and Charles Drake began. First, Millikan,39 a major figure in vascular neurology and the editor of Stroke, reviewed the literature on angiographic vasospasm and SAH as well as patients treated at the Mayo clinic in Rochester. He concluded there was no clinical presentation that was consistently present with angiographic vasospasm, that vasospasm had no effect on mortality because of SAH and that there was no relationship between complications of surgery and vasospasm. Millikan39 emphasized the role of arterial thrombosis in producing ischemia after SAH. It was implied vasospasm was sometimes used by surgeons to explain surgical complications as a cause of poor outcomes. It would generally be thought now that only the first of the 3 points is correct in that not all patients with moderate-to-severe angiographic vasospasm have DCI. This was already known. Sundt, an equally prominent figure in neurosurgery, who may have been a direct target of Millikan39 article (D. Piepgras, personal communication, October 28, 2014), had written that the correlation between angiographic vasospasm and clinical condition was incomplete.40 The Millikan39 article included data on patients managed at the Mayo clinic and at that time, such articles needed to be registered with the clinic and there would be an offer made to include as authors the other physicians whose patients were being described. This did not occur. Sundt was livid and declared to the chair of neurosurgery that “he goes or I go.” The Mayo personnel committee reviewed the situation and restrictions were placed on Millikan,39 which led to his departure to the University of Utah. He admitted to some inaccuracies in the data.39 Sundt died at 62 of cancer in 1992, a disease he thought he might have contracted from radiation from cerebral blood flow studies he conducted in service to his patients. Millikan39 outlived him, dying at 1995 in 2011.
At the same time, Fisher12 wrote the definitive description of DCI, first published in short form and was among the first to use a term such as delayed ischemic deficits. He noted that it developed 3 to 13 days after SAH, was reversible in 50% of cases and that angiographic vasospasm had to be severe to produce DCI. He assumed causation based on association. Investigators in Japan published virtually identical data.41 They emphasized the potential role of subarachnoid clot in sustaining angiographic vasospasm, discovered the key role of erythrocytes in the pathogenesis of vasospasm and use of cisternal drains to remove the clot. Treatment focused on the schizophrenic administration of intravenous cocktails of vasodilators, including isoproterenol, phentolamine, sodium nitroprusside, reserpine, and aminophylline with vasoconstrictors to maintain blood pressure.
Weir et al17 clarified the time course of vasospasm in man by measuring cerebral artery diameters on 627 angiograms from 293 patients with aneurysms. Vasospasm was present 3 to 12 days after SAH, and maximal around 7 to 8 days. Within 4 years of the invention of computed tomography, Japanese neurosurgeons showed that angiographic vasospasm was associated with the volume and location of subarachnoid blood.14,15 This was formalized by Fisher et al,16 in one of the most cited articles about SAH.
Two medical developments ended the decade of the 1970s. Kosnik and Hunt13 knew that Denny–Brown had reported that raising blood pressure could alleviate ischemic neurological deficits. In 1976, they specifically reported improvement in neurological deficits in patients with SAH and surgically repaired aneurysms, with intravenous fluids and induced hypertension. In those days when patients developed hyponatremia, they were fluid restricted. They commented that these patients may be frequently volume depleted. In 1979, Maroon and Nelson18 documented this for the first time. The pendulum swung from induced hypotension and fluid restriction to induced hypertension and volume loading and continues to be used to varying degrees in the absence of any level 1 evidence.
Fleckenstein,42 a German pharmacologist, discovered calcium channel antagonists. He recounted that in 1963, German pharmaceutical companies supplied him with drugs they had synthesized and found to be coronary vasodilators. He found that in cardiac muscle they mimicked the effects of calcium withdrawal. The key event was when in 1969, Bayer provided him with the first dihydropyridine compound, nifedipine, that turned out also to be a vasodilator. Another dihydropyridine, nimodipine, was developed specifically to be a cerebral vasodilator that would have less effect on blood pressure and be useful for treatment of cerebral disorders.43 After multiple preclinical studies, including a number showing that it had no effect on angiographic vasospasm, as well as human clinical trials, it was eventually approved by the US Food and Drug Administration in December, 1988, after having been available in Europe since 1985. The Food and Drug Administration concluded that it reduced the effect of vasospasm at causing cerebral infarctions and poor outcomes, but did not affect angiographic vasospasm. Its mechanism of action was unknown, and remains so today. Nothing has achieved this since, despite trials on ≈32 other drugs.44
Balloon angioplasty was first used to treat angiographic vasospasm by Zubkov et al.19 Technology improved the safety of the procedure and also spurned use of superselective infusions of vasodilator drugs. The efficacy of endovascular treatments has never been proven, and they are not entirely consistent with the theory that angiographic vasospasm is an epiphenomenon that does not by itself cause DCI.
Thirty years after the 1972 Princeton Conference, was it planned to have a debate over whether preventing vasospasm improves outcome after SAH?45 Long-time friends Macdonald and Zhang sparred for the issue and the polls showed it was a draw. Nobody lost their jobs this time and Diringer45 concluded essentially the same thing as was known in 1972. Thus, the controversies over angiographic vasospasm persist to this day, but they have changed their focus. There is no disputing the existence of angiographic vasospasm but its contribution to DCI continues to be questioned. The studies of the endothelin antagonist, clazosentan, which reduced angiographic vasospasm but had no effect on clinical outcome, fueled the controversy.21 Sixty years after description of angiographic vasospasm, much has been learned and outcomes from SAH are better but as Diringer45 wrote: “the debate has fostered the recognition that we must broaden our field of view and consider that DCI results from a complex interaction among multiple processes.”
The initial studies of the pathology and pathogenesis of SAH focused on the brain and were followed by study of isolated blood vessels. History shows that neither component can be neglected. Current theory is that poor outcome after SAH results from an interaction of the initial vascular and brain effects of the SAH (early brain injury) followed by delayed neurological deterioration that may be because of DCI.44 DCI, in turn, is thought to be a result of interaction of multiple processes, including early brain injury, angiographic vasospasm, cortical spreading ischemia, microthromboembolism, delayed neuronal apoptosis, dysautoregulation, and capillary transit time heterogeneity.44 What are the challenges for the future? One will be understanding the value of continued animal experimentation and delineation of the pathophysiology of DCI in the aftermath of continued lack of translation to humans of treatments that are supposedly efficacious in animals. Possible reasons why drugs, such as clazosentan, statins, and magnesium did not improve outcome have been listed but elude precise determination.44 A lesson from history may be that of antifibrinolytics that were efficacious at their target action of reducing rebleeding, but had no effect on outcome as a result of their propensity to increase DCI and possibly thromboembolic complications. As we gain knowledge that increases the understanding, and complexity, of the disease, simple statistics suggests larger groups of patients need to be studied. Efforts to do this are underway and have identified lack of common data elements for SAH and led to a project to develop them.46 Centers less than a hundred miles apart think rescue therapy works at one and that it does not at another. This could be addressed in a randomized clinical trial. However, we are sometimes surprisingly unwilling to question our theories, and the current trend is toward comparative effectiveness research and collection of registry data. Cooperative studies and registries of SAH were conducted beginning decades ago and led to some advances in treatment of SAH. This author, however, would recommend focusing energies on more randomized clinical trials.47
Dr Macdonald receives grant support from the Physicians Services Incorporated Foundation, Brain Aneurysm Foundation, Canadian Institutes for Health Research, and the Heart and Stroke Foundation of Canada; and is Chief Scientific Officer of Edge Therapeutics, Inc.
- Received November 4, 2014.
- Revision received November 4, 2014.
- Accepted December 30, 2014.
- © 2015 American Heart Association, Inc.
- Walton JN.
- Gull WM.
- Hunter J.
- Florey H.
- Robertson EG.
- Zubkov I,
- Nikiforov BM,
- Shustin VA.
- Etminan N,
- Vergouwen MD,
- Ilodigwe D,
- Macdonald RL.
- Macdonald RL,
- Kassell NF,
- Mayer S,
- Ruefenacht D,
- Schmiedek P,
- Weidauer S,
- et al
- Vergouwen MD,
- Vermeulen M,
- van Gijn J,
- Rinkel GJ,
- Wijdicks EF,
- Muizelaar JP,
- et al
- Luys J.
- Symonds CP.
- Froin G.
- Haschek E,
- Lindenthal OT.
- Crompton MR.
- Falconer MA.
- Diringer MN.