Atraumatic subarachnoid hemorrhage represents a small proportion of strokes, but is a true medical emergency that results in significant morbidity and mortality. Making the diagnosis can be challenging and misdiagnosis can result in devastating consequences. There are several time-dependent diagnostic and management considerations for emergency physicians and other frontline providers. This article reviews the most up-to-date literature on the diagnostic workup of subarachnoid hemorrhage, avoiding misdiagnosis, and initial emergency department management recommendations.
Ruptured aneurysms are the most common cause of atraumatic subarachnoid hemorrhage and there are excellent endovascular and surgical treatments.
Misdiagnosis remains an important issue especially in well-appearing patients, for whom these treatments are the most effective.
Noncontrast computed tomography scan sensitivity approaches 100% if done within 6 hours of the headache onset and read by an experienced radiologist.
Beyond 6 hours, some other diagnostic testing is necessary and we recommend lumbar puncture.
Once the diagnosis of nontraumatic subarachnoid hemorrhage is made, there is a short checklist of next steps related to defining the offending vascular lesion, consulting a neurosurgeon, and starting treatments.
Nature of the Problem
Nontraumatic headache is the seventh most common complaint in patients presenting to the emergency department (ED). Although many causes of headache are benign, subarachnoid hemorrhage (SAH) is one serious etiology and is a true medical emergency in that time-dependent diagnostic and management techniques exist. , The clinical presentation in patients with SAH is highly variable and the classic “textbook” illness script of thunderclap headache, neck stiffness, and altered mental status may not occur. Misdiagnosis occurs in 12% to 50% of patients and is more common in patients with atypical presentations. , Misdiagnosis also accounts for a substantial proportion of return ED visits in patients discharged with a diagnosis of headache. Therefore, deciding which patients to evaluate for SAH is a critical issue for emergency physicians. Recent data have emerged that has shifted the diagnostic approach since the last publication in this series. This article reviews the most recent literature and guideline revisions in the diagnosis and ED management of atraumatic SAH.
Although trauma is the leading cause of SAH, the incidence of atraumatic SAH is estimated to range from 6 to 9 in 100,000. , These cases account for less than 1% of all ED patients with headache; however, the disease burden is high. One-third of patients die within weeks of the hemorrhage and the majority of those who survive suffer from cognitive impairment or long-term complications. Furthermore, one-half of patients who present with SAH are under age 55. In recent years, the incidence has found to be decreasing slightly with better blood pressure treatment and decreasing smoking rates, although regional variation continues to exist. ,
The most common cause of atraumatic SAH is a ruptured cerebral aneurysm, accounting for 85% of cases, followed by nonaneurysmal venous “perimesencephalic” hemorrhages and arteriovenous malformations. , , In the general population, approximately 2% harbor cerebral aneurysms, yet the vast majority of these never rupture. Other, less common causes of atraumatic SAH include amyloid angiopathy, hypertension, and reversible cerebral vasoconstriction syndrome. Risk factors for SAH owing to aneurysm rupture include smoking, hypertension, alcohol abuse, sympathomimetic drug use, having a first-degree relative with a cerebral aneurysm, female sex, and certain genetic conditions including type IV Ehlers–Danlos syndrome and polycystic kidney disease. ,
Patterns of Hemorrhage
The location of the hemorrhage pattern on a computed tomography (CT) scan can be useful in predicting the presence of an aneurysm, location of the aneurysm, whether or not it was traumatic, and the likelihood of an angiogram-negative SAH. Approximately 70% of aneurysms occur in the anterior communicating artery, posterior communicating artery, and middle cerebral artery. Blood from a ruptured aneurysm typically surrounds the basal cisterns. In traumatic SAH, the blood is typically located in areas of coup or contrecoup force, or higher up in the cerebral convexities. This difference in location may be particularly helpful in patients with SAH who may have fallen from syncope and present with a headache. Finally, convexal SAH has been described in various case series and the etiologies are believed to be related to reversible cerebral vasoconstriction syndrome in younger patients or cerebral amyloid angiopathy as opposed to aneurysmal in origin. , Patterns of blood are displayed in Fig. 1 .
SAH has a wide variety of clinical presentations and, therefore, deciding which patients require evaluation is a critical issue for emergency physicians. Approximately 12% of patients die before reaching the hospital and up to 25% of patients die before admission to the hospital ward or intensive care unit. , For those presenting to the ED, one can consider the spectrum of presentation along a bell-shaped curve ( Fig. 2 ). On 1 side of the curve are critically ill patients who come in altered and oftentimes comatose or with clear-cut focal neurologic deficits. Although the diagnosis may not be clear cut initially, these patients will undergo an extensive workup. In the middle are the patients who present with the more classic symptoms with thunderclap headache, neck pain, and/or vomiting and syncope. Oftentimes the headache is exertional. The need for diagnostic workup in this group is clear. On the other side of the curve, however, are the less typical presentations. These patients will often describe symptoms of isolated neck pain, vomiting, sinusitis-type headache, headache with an elevated blood pressure, mildly altered level of consciousness, chest pain, or an abnormal electrocardiogram, all less classic symptoms but nevertheless well-described in the literature. , , It is this group of patients who tend to be misdiagnosed or have a delay in diagnosis, which in turn leads to treatment delays and poorer outcomes. Understanding the full spectrum of presentations is important for avoiding misdiagnosis.
As always, the physical examination begins with the vital signs, general appearance, and assessment of airway, breathing, and circulation. A focused physical examination should be performed once the patient is stabilized that includes a relevant neurologic examination. In patients who cannot give a full history or are comatose, ophthalmoscopic examination looking for retinal hemorrhages, a finding observed in approximately 10% of all patients with SAH, can be helful. This finding may be the only clue to the diagnosis in comatose patients. In many cases, the neurologic examination is nonfocal and, therefore, less useful for making the diagnosis. Abnormal examination findings can sometimes suggest the location of an offending aneurysm. Meningismus is oftentimes a later finding and should not be relied on in excluding the diagnosis.
In a large, prospective population of ED patients presenting to 6 Canadian EDs with thunderclap headache (defined as reaching maximum intensity within 1 hour), Perry and colleagues proposed the Ottawa Subarachnoid Rules. Of the 2000 patients included in this cohort, 130 (6.5%) had SAH, making the study population representative that seen in practice. The authors developed 3 clinical decision rules, all of which were 100% sensitive for SAH that would have decreased the rates of investigations from 83% (baseline) to between 64% and 74%. The rule can be applied in alert patients 15 years of age or older who have no recent head trauma, lack new neurologic deficits, and have no history of a prior aneurysm, brain tumor, or SAH. If 1 or more criteria are present, SAH cannot be ruled out and the patient requires a workup:
Age greater than 40 years
Complaint of neck pain or stiffness
Witnessed loss of consciousness
Onset with exertion
Limited neck flexion on examination
The Ottawa Subarachnoid Rules have since been internally and externally validated and are highly sensitive for ruling out SAH. It is important to remember that the differential diagnosis for thunderclap headache is broad and only 10% of these patients will have a SAH. , Emergency physicians must consider the differential diagnosis for this type of headache, which includes both serious and less serious etiologies, which in turn may require additional workup Box 1 summarizes other causes of thunderclap headache.
“Cannot miss,” emergent etiologies
Cervical or cranial artery dissection
Cerebral venous sinus thrombosis
Meningitis or encephalitis
Idiopathic intracranial hypertension
Temporal arteritis or central nervous system vasculitis
Acute narrow angle closure glaucoma
Spontaneous intracranial hypotension
Reversible cerebral vasoconstriction syndrome
Carbon monoxide poisoning
Mass lesions, including
In patients for whom SAH is suspected, the initial testing modality is brain imaging. This modality is most commonly a noncontrast CT scan, because it is most readily available in most EDs. MRI has extremely high sensitivity for hemorrhage, but the sensitivity early on is unknown and is not often possible as an initial imaging modality in the ED. When performed in the first 24 hours of headache onset, the sensitivity of CT scan for detecting subarachnoid blood ranges from 90% to 100% and approaches 100% when performed within the first 6 hours. The sensitivity decreases as time from onset to CT scan elapses owing to the dilution of blood by the normal flow of cerebrospinal fluid (CSF). Other factors that decrease sensitivity include technical issues (ie, older CT scanners <32 slices or patient movement), interpretation error, small-volume bleeds, and anemia. ,
There has been a growing body of literature demonstrating the sensitivity of CT scan approaches 100% when the patient is imaged within 6 hours of headache onset. In a 2016 meta-analysis of 8907 patients, the incidence of SAH on noncontrast brain CT performed within 6 hours and read by an attending radiologist in neurologically intact patients was 1.46 in 1000, with an overall sensitivity of 98.7%. Perry and colleagues sought to validate both the Ottawa SAH Rules and the 6-hour CT scan rule in a prospective study at 6 Canadian EDs after implementing education on both of these approaches. In their analysis of 3672 patients, they found the Ottawa Rule was 100% sensitive (95% confidence interval, 98.1%–100%), and the 6-hour head CT scan rule was 95.5% sensitive (95% confidence interval, 89.8%–98.5%) for SAH. Furthermore, hospital admission rates were significantly decreased after this was implemented. The 5 missed SAH on the 6-hour head CT scan were owing to a radiology misread, 2 incidental aneurysms, 1 nonaneurysmal cause, and 1 profoundly anemic patient. As a result of the growing body of evidence, the American College of Emergency Physicians 2019 Clinical Policy on the Evaluation and Management of Adult Patients presenting the ED with Acute Headache recommends that a noncontrast head CT scan performed within 6 hours of headache onset precludes the need for further diagnostic workup for SAH if the CT scan was performed with a modern generation scanner (Level B evidence). Before this clinical policy update, lumbar puncture was recommend to definitively rule out SAH. The authors recommended that the decision to invoke the 6 hour rule and defer lumbar puncture takes into account the limitations of CT scans and applies this only to cases in which the patient is neurologically intact, the CT scan is read by an experienced radiologist, the scan was performed with modern CT scanner (third generation or higher), the CT scan is of strong technical quality, and the patient is not anemic.
In patients with a negative head CT scan more than 6 hours after the onset of headache, lumbar puncture to assess for the presence of red blood cells (RBCs) or xanthochromia will definitively rule out SAH. Both the number of RBCs and the presence of xanthochromia are a function of time from the bleed. Xanthochromia, or the yellowish hue resulting from the breakdown of hemoglobin, is detected by visual inspection (in most North American laboratories) or spectrophotometry and is highly sensitive for SAH. , As with CT scan, there are timing considerations one must consider with the lumbar puncture. RBCs appear very early in the course after the hemorrhage occurs, within the first few hours. Xanthrochromia may take up to 12 hours to develop, but is present by 6 hours. Thus, assessing the CSF for both RBCs and the presence of xanthochromia is recommended.
Emergency physicians must be aware of the possibility of a traumatic lumbar puncture when interpreting CSF results. These occur when blood from local trauma or a venous plexus contaminates the fluid and are estimated to happen in 16% to 31% of lumbar punctures. , There is no universally accepted method of evaluating RBC clearance to establish the likelihood of traumatic lumbar puncture versus SAH. On the surface, a commonly held belief has held that if there are fewer RBCs in tube 4 than tube 1, the likelihood of a traumatic lumbar puncture is high and the probability of SAH is lower. In a prospective analysis of 1739 patients with headache who underwent lumbar puncture to rule out SAH, the presence of fewer than 2000 × 10 6 /L RBCs in addition to absence of xanthochromia excluded the diagnosis of aneurysmal SAH, with a sensitivity of 100%. Smaller studies support a less conservative assumption that less than 500 RBCs in tube 4 along with a decrease from tube 1 rules out SAH. ,
For patients in whom the CSF analysis is suggestive of SAH, the next steps in the workup and management include vascular imaging and neurosurgical consultation. In the past several decades, most commonly using CT angiography (CTA) has become the imaging modality of choice because it is noninvasive and highly sensitive for cerebral aneurysms. , If no source is visible on the CTA, then digital subtraction angiography should be considered. There are some advocates of a CT scan/CTA approach as opposed to lumbar puncture in those patients who present greater than 6 hours after ictus. Although this strategy avoids the potential diagnostic challenge of traumatic tap as well as other complications of the lumbar puncture, there is a lack of robust data to support the superiority of this method. The 2019 American College of Emergency Physicians Clinical Policy thus makes a Level C recommendation to perform either an lumbar puncture or CTA after a negative noncontrast brain CT scan in patients for whom SAH is still suspected. If forgoing the lumbar puncture and opting for a CTA instead, it is important to understand the limitations and consequences of this path. For one, a CTA will not detect subarachnoid blood and may miss up to 15% of those cases that are not aneurysmal. Because of its extremely high sensitivity, CTA may detect incidental aneurysms that are not in fact the cause of the headache and commit the patient to further diagnostic workups and in turn health care costs. Additional and potentially unnecessary vascular imaging also exposes the patient to higher doses of radiation exposure and contrast, both of which can have downstream consequences. The authors recommend using this approach only in those patients in whom there is high concern for SAH and lumbar puncture is not feasible (ie, patient refusal, anticoagulation use and inability to administer a reversal agent). A suggested diagnostic algorithm is displayed in Fig. 3 .