Subarachnoid Hemorrhage

Juan Carlos Baez, Rush H. Chewning, Gerald Wyse and Kieran P.J. Murphy

When a patient has a severe or “thunderclap” headache of sudden onset, subarachnoid hemorrhage (SAH) must be considered and investigated. The most common symptom described by patients with SAH is headache, often the worst they have ever experienced and associated with a sense of doom. Less specific symptoms include stiff neck, change in level of consciousness, vomiting, and focal neurologic deficits, including cranial nerve palsies. Cranial nerve palsies can arise as a result of bleeding, compression from an aneurysm, or ischemia secondary to acute vasoconstriction immediately after aneurysmal rupture. The stereotypic third nerve palsy is manifested as an inferiorly rotated abducted eye with accompanying ptosis and pupil dilation. Patients often describe visual deterioration in the form of diplopia. Onset is sudden, and pain is a common complaint. Third nerve palsy often serves as an indicator of a posterior communicating artery aneurysm, although superior cerebellar and basilar artery aneurysms can less commonly cause the same symptoms.¹

Ruptured intracranial aneurysms account for approximately 85% of cases of nontraumatic SAH. Although some have suggested familial or genetic causes for the development of intracranial aneurysms, other factors such as smoking, heavy alcohol use, and hypertension play a larger role. Female sex, African American heritage, and hereditary disorders such as autosomal polycystic kidney disease constitute some of the genetic factors also linked to development of these vascular lesions.²

Other causes of SAH include cerebral arteriovenous malformation (AVM) and perimesencephalic hemorrhage. AVMs are complex bundles of arteriovenous connections without intervening capillaries. The abnormally high flow commonly results in secondary aneurysms. AVMs are less common than aneurysms as a cause of SAH and will not be discussed further. Perimesencephalic hemorrhage is manifested as blood in the cerebral cisterns, with a predominantly prepontine distribution, and accounts for 10% of all SAH cases. This benign condition is thought to be secondary to rupture of a prepontine vein and poses no threat to the patient.³

Multiple grading systems are used for subarachnoid hemorrhage. Two of the more commonly used include the Fisher scale and the Hunt and Hess classification. The Fisher scale scores SAH on computed tomography (CT) appearance and quantification of subarachnoid blood. Patients are in group 1 if no blood is detected. Group 2 consists of those with diffuse deposition of subarachnoid blood, no clots, and no layers of blood greater than 1 mm. Group 3 patients have localized clots or vertical layers of blood 1 mm or greater in thickness, whereas patients in group 4 have intracerebral or intraventricular clots or diffuse subarachnoid blood.⁴ The Fisher scale is a radiologic classification and should not be used to predict clinical outcomes.

The Hunt and Hess classification scale grades SAH on a scale from I to V based on clinical manifestations of the patient. A grade of I is given to patients with asymptomatic findings or mild headache. Grade II is assigned for cranial nerve palsies, moderate to severe headache, and nuchal rigidity. Grade III includes mild focal neurologic deficits, lethargy, or confusion. Grade IV consists of stupor, moderate to severe hemiparesis, and early decerebrate rigidity. Grade V is deep coma, decerebrate rigidity, and a moribund appearance.⁴ If the patient has a serious systemic disease such as hypertension, diabetes mellitus, or chronic obstructive pulmonary disease or if severe vasospasm occurs during arteriography, the Hunt and Hess grade is increased by one level.⁵ The Hunt and Hess system fails to adequately describe the patient’s stage of consciousness on admission, which ultimately proves to be the most reliable indicator of recovery. Consequently, interphysician variability exists when using this classification system. In addition to the Hunt and Hess system, use of the Glasgow Coma Scale, which attempts to objectively grade consciousness according to eye opening, verbal ability, and motor function, helps stratify patients and minimize variability among physicians.

Indications

Outcomes in cases of SAH are poor, with mortality rates ranging from 32% to 67%.⁶ Early diagnosis is crucial in minimizing the consequences associated with this condition, because up to 30% of fatalities in individuals with previous SAH result from rebleeding. Among survivors of SAH, 30% will have moderate to severe disability after the event. Unfortunately, between a fourth and half of cases of SAH are missed by the first evaluating physician. Initial evaluation of SAH typically involves the use of noncontrast-enhanced CT in the emergency department. CT scans are fast, readily available, and relatively inexpensive. CT has been demonstrated to have 93% to 95% sensitivity for SAH when conducted within 24 hours of initial bleed.⁷ A timely study is critical because 50% of SAH episodes are not visible on CT 1 week after the bleeding.

When clinical suspicion for SAH is high, negative CT findings cannot rule it out. The literature reports a 7% false-negative rate of head CT for the diagnosis of SAH in patients later found to have ruptured aneurysms. Because mortality with untreated SAH is exceedingly high, patients with negative CT findings but clinical suspicion for SAH should undergo additional diagnostic testing. A lumbar puncture (LP) is recommended as the next step in the diagnosis of SAH. Recent work published in the BMJ supports modern multislice head CT alone as adequate to exclude SAH if it is performed within 6 hours of ictus.⁸ This is being adopted in some emergency departments but not all, and will no doubt be controversial because of cases where LP found CT-negative SAH.

The typical LP involves sequentially drawing four aliquots of cerebrospinal fluid (CSF) from the L4-L5 interspace under local anesthesia of the patient’s back. Inspection of CSF for xanthochromia (yellowish tint of CSF, indicating hemolyzed blood) provides the best evidence of SAH. A perfectly performed LP in a patient with no SAH yields clear CSF with no red blood cells (RBCs) after violating the L2-L3 interspace. Common parlance refers to this as a “champagne tap,” and this result combined with the absence of xanthochromia reliably rules out SAH.⁹

Approximately 10% to 25% of the time, however, LP yields varying amounts of blood. The physician must then decide whether the blood represents SAH or perforation of the epidural venous plexuses and cauda equina vessels during the procedure. The “three-tube test” (named before four tubes became standard) is based on the principle that if the amount of blood present in the aliquots decreases between the first and fourth tubes, the blood is attributed to a traumatic tap. No consensus exists regarding what size decrement of blood indicates a traumatic LP. This results in uncertainty for physicians who must base therapeutic decisions on these results. Although multiple criteria have been used to differentiate between traumatic taps and true SAH, these criteria have been devised on the basis of traditional medical practice and not evidence-based medicine.

A study by Heasley et al., however, showed that a 25% reduction in RBCs between the first and fourth tubes—as one would expect to see in a traumatic tap—does not rule out SAH.⁹ Other authors have used different cutoffs for what constitutes a traumatic tap, but these cutoffs have also yielded unacceptably high false-positive and false-negative rates. Heasley’s group considers determination of traumatic taps via the three-tube test unreliable and recommend further invasive testing to determine the presence or absence of SAH with certainty if clinical suspicion remains high.

Additional CSF testing for D-dimer, clot formation, and opening pressure have all been studied in an attempt to diagnose SAH with greater accuracy, but the results have been inconclusive; these tests are rarely part of the LP procedure. In cases where a patient is complaining of the worst headache ever experienced, formal evaluation of the cerebral blood vessels is necessary subsequent to any LP that yields blood.