A quiet revolution has been taking place over the past 25 years in the treatment of cerebral vascular diseases. The era of delayed aneurysm surgery is history. Prophylaxis and therapy for cerebral vasospasm have changed the face of the management of aneurysmal subarachnoid hemorrhage (SAH). Guglielmi detachable coil (GDC) embolization of aneurysms, introduced less than 10 years ago, has become an integral component of the arsenal for the treatment of cerebral aneurysms. N-butyl-cyanoacrylate (NBCA) glue embolization and stereotactic radiosurgery have so completely altered the landscape in the area of cerebral arteriovenous malformations that treatment approaches and outcomes bear little resemblance to the paradigms expounded by the current generation of retiring leaders in the field. The “Decade of the Brain” that has just passed has impressed on the national consciousness the need for emergency treatment of “brain attack.” This concept has dramatically altered the approach to patients with ischemic stroke and intracranial hemorrhagic diseases. Intravascular thrombolysis, neurological intensive care units, and critical care neurology are terms not found in textbooks of a quarter century ago.

It is intuitively evident that with the explosion of supportive technology and modern treatment paradigms for vascular disease that every medical center in the country cannot be ideally equipped for treating cerebrovascular diseases. Within the field of neurosurgery, similar advances are occurring in spine treatments, oncology, epilepsy, etc. The task of remaining current and technologically up to date in all areas is a daunting task.

The Case for Regionalization

Regionalization of neurovascular medicine is an eventuality that cannot be denied. The underlying premise behind regionalization is that optimal outcomes are achieved at greatest efficiency if patient care is restricted to relatively few dedicated cerebrovascular centers. The first issue is that the majority of practicing physicians in this country were trained in an era when the approach to neurovascular medicine bore little resemblance to the state of the art at the dawn of the 21st century. Change continues at a rapid and accelerating pace. Obviously, many physicians have kept up with progress and have constantly modernized treatment approaches. Many have not. Substandard treatment is not only tolerated, it has been well-documented.

Meanwhile, economics and local politics have led the medical staff of small hospitals to be reluctant to transfer patients to regional centers, even when the modern infrastructure for the treatment of cerebrovascular diseases does not exist. Unfortunately, numerous studies show significant differences in morbidity and mortality for the management of cerebrovascular diseases in highvolume centers with high-volume surgeons compared with low-volume surgeons and hospitals.^1–9 Mortality has also been demonstrated to be significantly lower at highvolume hospitals for elective abdominal aortic aneurysm repair, carotid endarterectomy, lower extremity arterial bypass surgery, coronary artery bypass surgery, coronary angioplasty, heart transplantation, pediatric cardiac surgery, pancreatic cancer surgery, esophageal cancer surgery, trauma care, and treatment of human immunodeficiency virus (HIV)/acquired immunodeficiency syndrome (AIDS).^3–5,10–1² The inference is inescapable: The management of cerebrovascular diseases is safer and less expensive at regional medical centers with specialized facilities and medical personnel.

Solomon et al demonstrated that in New York State hospitals from 1987 to 1993 there was a 43% reduction in mortality for aneurysm surgery in hospitals performing more than 30 craniotomies per year for cerebral aneurysm compared with the lower-volume hospitals (p <0.01).⁸ Taylor et al analyzed Medicare patients over an 8-year period from 1984 through 1991.^9,13 The mortality rate for patients treated surgically for aneurysmal SAH was inversely correlated with the annual number of craniotomies performed at a given institution. Johnson et al studied individual California hospital experiences and overall treatment volume of cerebral aneurysms as a predictor of outcome.⁷ The data on both endovascular treatment and surgical treatment was retrieved from a California statewide database of hospital discharges from January 1990 through December 1998. Hospital treatment volume was divided into quartiles to distribute patients into four groups of equal size (quartile 1: 0 to 15.7 cases per year; quartile 4: 68 to 107 cases per year). Adverse outcomes for unruptured aneurysms decreased with increasing hospital volume (from 33% in quartile 1 to 14% in quartile 4; p < 0.005).

Despite some of the inherent weaknesses associated with retrospective database-based studies, we recently took a second look at the New York State discharges database, this time focusing on patients treated between 1995 and 2000.^14–17 For each 10 additional procedures performed per year, hospital volume was associated with fewer adverse outcomes and lower in-hospital mortality for both unruptured cerebral aneurysms (adverse outcome odds ratio [OR] 0.89, p < 0.0001; mortality OR 0.94, p = 0.002) and aneurysmal subarachnoid hemorrhage (adverse outcome OR 0.94, p = 0.03; mortality OR 0.95, p = 0.005). Indeed, even among the top 10 hospitals in New York State, there was a surprising residual effect of volume on outcome. In the top hospitals, which treated nearly 50% of the state’s cases, the beneficial effect of institutional volume on adverse outcomes following treatment for both unruptured and ruptured aneurysms was about as large as it was for the overall state dataset. We also found that outcome following embolization of cerebral aneurysms is less affected by overall hospital volume than is clipping of cerebral aneurysms. This is not surprising given the fact that many have posited that aneurysm clipping may be more dependent on operator experience than aneurysm coiling. Although direct support for this contention has been difficult to demonstrate, data from the Cleveland Clinic does at least show that the number of aneurysm operations performed by each individual surgeon is a robust predictor of functional outcome (r = 0.99, p = 0.05).¹⁸

Armed with the data from New York State, we then looked to see what the effect would be of regionalizing aneurysm care in New York State by referring patients to the top 10 busiest centers (?35 cases/year [35 to 133]). For unruptured aneurysms, there we predicted a 27% annual improvement in the adverse outcome rate and a 14% improvement in mortality. For aneurysmal subarachnoid hemorrhage, the reductions would be 12% and 11%, respectively. In absolute numbers, the reduction in poor outcome or death per year would 18 patients with unruptured aneurysms and 50 patients with ruptured aneurysms. If fewer centers were used with even higher volume, the predicted reductions in suffering would be even greater. If one extrapolates these data to the United States as a whole, regionalization of cerebral aneurysm treatment alone would eliminate 5700 poor outcomes or deaths each year in the United States.^2,19 One could only guess at the potential cost savings. Add stroke, carotid artery disease, and arteriovenous malformations to the mix, and the case for a serious national health care initiative is on the table.

Given the rapidity of the projected growth in endovascular procedures, without regionalization, there will be a real shortage of qualified neuroendovascular surgeons. Even as the number of these specialists grows, the main driving force will be the treatment of acute stroke rather than orphan diseases such as aneurysms and arteriovenous malformations (AVMs). Therefore, as the field of endovascular surgery matures, individual practitioners will treat fewer and fewer aneurysms and AVMs in much the same way as the neurosurgical workforce as a whole has come to treat less and less intracranial disease. Eventually, it is possible that only those endovascular surgeons in busy centers will be able to demonstrate the lowest complication rates for rarer diseases. Currently there is no direct evidence that this will be the case for endovascular coiling, but only 19 (7%) of New York State’s 257 hospitals actually coil aneurysms, compared with 113 (44%) in which clipping is performed. In California from 1990 to 1998, only 11 centers performed GDC coiling, whereas 167 centers provided open surgical services for these lesions.⁷ As endovascular techniques filter to the community, it will obviously be necessary to examine whether results with common diseases such as cerebral ischemia vary to the same degree as results with lesscommon diseases like aneurysms. What is certain is that there will likely be a need for several types of cerebrovascular referral centers as time goes on. There will be quaternary centers treating low-frequency diseases that can stand minor delays in therapy related to transportation, and there will be tertiary centers where services are provided for common diseases that cannot bear any delay at all in referral. Even a large state like New York might have 10 centers treating aneurysms, but 100 centers providing acute pharmacologic and mechanical thrombolysis.

Cerebrovascular Disease Center

The heart of a modern center for the treatment of cerebrovascular diseases is a neuroscience intensive care unit (NICU). Ideally, the ICU will focus only on patients with neurological diseases. This type of focus guarantees that the medical staff is trained and devoted to treating neurological illness. Such a unit could also be part of a more general ICU but there would have to be staff commitment to the clinical neurosciences. Physicians trained in critical care neurology staffing the unit 24 hours/day are an essential component. These physicians could be neurology residents, neurosurgery residents, trained intensive care specialists, or other physicians attuned to the cardiopulmonary management of patients with central nervous system diseases. The most successful models of NICUs are in major medical centers with full-time neurology and neurosurgery residents, and thus the capacity to staff the unit on a 24-hour basis with involved, welltrained physician support staff. This staffing usually consists of a resident on call at night with backup from a fellow and senior staff member. Resident work hours’ restrictions, the movement among neurology training programs to limit clinical inpatient experiences in favor of outpatient and laboratory training, and the growing evidence that ICU patients should be continuously, visually observed by attending-level physicians make the NICU an increasingly complex and expensive program to administer.

Nursing is also an essential aspect of the NICU. Nurses at the bedside are the ones that have to be able to recognize the earliest signs of neurological deterioration that often initiate a chain reaction of treatment that can make the difference between a good and poor outcome. General ICU nurses without a special interest in neurological diseases usually are not well trained in recognizing important subtle changes in the neurological examination. These nurses may also not be able to institute critical therapy even before the physician staff becomes available. With national nursing shortages and the increasing opportunities open to nurses in management, public health and as inpatient and outpatient physician extenders, nursing turnover in ICUs has increased in the last decade, making the experienced neuroscience ICU nurse an endangered species. Assembling and maintaining teams of these nurses is increasingly complicated and expensive.

Neuroradiological facilities have obvious implications for the management of cerebrovascular diseases. 24-hour access to sophisticated computerized tomography (CT) scanning (CT angiography [CTA], CT perfusion [CTP], Xenon-CT etc.), magnetic resonance imaging (MRI), and cerebral angiography are essential components of any fully equipped modern cerebrovascular center. Availability also includes radiology technicians and fully trained neuroradiologists capable of performing invasive diagnostic procedures, reconstructing critical data sets, and reading the films as they become available at all hours of the day and night.