Share this article on:

InFocus: The Enigma of the New Stroke Guidelines

Roberts, James R MD

doi: 10.1097/01.EEM.0000395418.79399.7f
For stroke patients, it is unclear whether clinicians should target and manipulate cerebral perfusion pressure (CPP) or intracranial pressure (ICP)

For stroke patients, it is unclear whether clinicians should target and manipulate cerebral perfusion pressure (CPP) or intracranial pressure (ICP)

Attempting to decipher the most advantageous clinical approach in the ED for many intracranial hemorrhage issues is muddied by theory, conflicting or lacking data, nonintuitive clinical outcomes, unwarranted pessimism, and unbridled optimism. One would think that we would know exactly what to do by now, and we should have the pathophysiology well in hand. Unfortunately, we are left with only guidelines, which essentially are admissions that we don't know what to do. The recent guidelines from the American Heart Association (AHA) and the American Stroke Association (ASA) are vague and limited by rigorous data, and are rife with enigmas.

Guidelines for the Management of Spontaneous Intracerebral Hemorrhage: A Guideline for Healthcare Professionals from the American Heart Association/American Stroke Association

LB Morgenstern, Hemphill JC 3rd, et al



The promulgators of the new ICH guidelines address prehospital management, neuroimaging, ED management of potential seizures, and the control of hyperglycemia and fever.

Prehospital Management: While the guidelines underscore the dismal prognosis of patients with ICH, the AHA and ASA continue to reiterate that early aggressive management is of paramount importance. EMS' goal is to garner historical data, provide respiratory and cardiovascular support, and deliver patients to the ED expeditiously. EMS should provide advanced notice of the impending arrival of a potential stroke victim. This allows the ED to gear up and set in motion any stroke alerts or laboratory and CT protocols. The guidelines say nothing about the prehospital management of hypertension, specifically emphasizing rapid transport to the ED. There is a greater than 75 percent mortality rate for patients with ICH whose prehospital neurological score declines by more than six points, emphasizing that by the time the patient reaches the ED, the ravages of ICH have likely set the final prognosis.

General ED Management and Logistics: The guidelines stress that the ED must be prepared to treat ICH, or have a plan for rapid transfer to a tertiary care center. While stroke alerts usually have focused on ischemic strokes, the concepts are similar for ICH. In short, such patients are made a priority, available resources are marshaled, and appropriate consultants alerted. There are no specific guidelines addressing transfer criteria or timing constraints for ED interventions to commence.

Neuroimaging: Although vomiting, severe hypertension (more than 220 mm Hg systolic), severe headache, or decreased level of consciousness over a few minutes suggests ICH, the guidelines state that these are only nonspecific parameters. ED interventions for such patients are similar to those applied to many sick patients: adequate history, thorough physical exam, and appropriate testing.

The gold standard for identifying acute hemorrhage is still a noncontrast head CT. While an MRI may provide more information, it is acknowledged that the logistics of acquiring this test in the ED is not necessary for acute management. Prior research found that a CT obtained within three hours of ICH onset will demonstrate up to 38 percent incidence of hematoma expansion on follow-up CT, attesting to the rapid ongoing insult of a brain bleed. CT angiography and contrast-enhanced CT may identify patients at higher risk for ICH expansion, but these tests are not recommended as the initial study. It is presumed that tumors, aneurysms, arteriovenous malformation, or venous thrombosis can be treated more readily if one knows the nuances of the vasculature causing the stroke.

Management of Glucose and Temperature: While tight control of glucose levels was once thought to be beneficial and garnered clinical dogma status for many critically ill patients, this too has been dampened by scientific scrutiny. It is now assumed that high blood glucose predicts an increased risk of mortality and poor outcome, but it is a clinical marker rather than a causative agent. Because there is significant risk of hypoglycemia with micromanagement of glucose levels, the new guidelines conclude that the ideal glucose level remains to be clarified, but hypoglycemia should be avoided.

Fever can occur from CNS damage, particularly in patients with intraventricular hemorrhage. As with glucose, fever is likely a marker that is related to outcome, is not causative, and is an independent factor of poor prognosis. Because there are no data linking fever control with outcome nor a proven benefit from the current penchant for therapeutic cooling (as in cardiac arrest), maintaining normothermia is the best one can posit.

Seizure Control: Seizures occur within the first few weeks of ICH, usually near the onset, in about three percent to 20 percent of patients. Seizures may occur even when prophylactic anticonvulsants have been given. If one uses continuous EEG evaluation, a rather impressive number of subclinical seizures are identified (up to 31%), even if prophylactic anticonvulsants have been given. It is also very difficult to relate clinical seizures to a worsened neurological outcome or increased mortality, although such an association seems intuitive. Two studies have actually demonstrated a significantly higher mortality or morbidity rate when prophylactic seizure drugs, particularly phenytoin, have been administered. The risk-benefit analysis of prophylactic anticonvulsants is currently uncertain. These guidelines suggest treating clinical seizures or those with EEG findings of seizures if a patient has a change in mental status, but do not recommend seizure prophylaxis or a specific anti-epileptic treatment.

ICP Monitoring/Treatment: Although intracranial pressure monitoring is often performed for ICH, there are limited published data evaluating the frequency of elevated ICP or its management. The specific areas in the brain that may experience elevated ICP have not yet been identified, but ICP monitoring is still commonly employed. The guidelines state that elevated ICP is usually secondary to hydrocephalus from intraventricular hemorrhage, or mass effect from hematoma or surrounding edema. Patients with small hematomas or those lacking hydrocephalus do not usually require ICP monitoring. The following sentence from the guidelines sums it up: “The absence of published studies showing that management of elevated ICP impacts on ICH outcome makes the decision whether to monitor or treat elevated ICP unclear.” (Guidelines available at

Even in proven herniation, intraventricular hypertension, or hydrocephalus, monitoring and treating ICP “might” be considered; hardly a rousing endorsement. Ventricular drainage is given a “reasonable” recommendation in patients with a decreased level of consciousness.

Surgery for Clot Removal: Whether to remove an intracranial hematoma surgically remains controversial. Obviously, removing deep hematomas requires cutting through uninjured brain. Most trials of surgery for ICH exclude patients with cerebellar hemorrhage. It is generally considered that cerebellar hemorrhage producing brain stem compression or hydrocephalus will benefit from surgical removal of the hematoma; however, randomized trials have not been undertaken.

Comment: If you thought aggressive management of blood pressure, and immediate reversal of warfarin-induced coagulopathy were straightforward topics, consider other parameters of ED hemorrhagic stroke treatment. It's depressing enough that we don't even know if blood pressure control is good or bad, or whether manipulating hematoma size by reversing coagulopathy is universally beneficial. The new guidelines, however, attack other sacred cows that have been the mantra of unenlightened, dogmatic, and hubris-eluting clinicians. For those of you who still harbor the belief that strict control of glucose levels, ritualistic cooling, and the immediate institution of seizure prophylaxis are the gold standard for ICH, read on.

Prehospital Care: Hearing about an impending stroke patient from EMS is a good idea, although with a 20 percent incidence of stroke mimics arriving in the ED, it does not appear to be rational or standard to call a stroke alert, or inform the neurosurgeon or CT tech prior to patient arrival. One study demonstrated that EMS notification of potential tPA candidates shortened door-to-CT time by only seven minutes (47 vs. 40 minutes). (Prehosp Emerg Care 2008;12[4]:426.) A stroke-alert mentality is intuitively state-of-the-art, but that concept also has never been proven. Witness the information on rapid response teams in the hospital for other medical emergencies. Who would guess that having doctors rapidly respond to the general medical floor for ailing patients would not change morbidity or mortality? But this intuitive stance has yet to prove its worth in final outcome.

It is reasonable to educate EMS on at least taking the blood pressure prior to arrival so it can be communicated in their report. Checking prehospital glucose should be axiomatic, and rapid transport is always supported. The value of bypassing your community ED to take the patient to a university center, with all the bells and whistles has yet to be clarified. Unless the garden-variety ED and tertiary care centers are equidistant, I would vote for the nearest hospital concept. At least I can intubate, perform a CT scan, stop seizures, check the glucose, and evaluate other necessary parameters. Again, “rapid transfer” to a tertiary care center is a nice thought but difficult to accomplish. If you can find an accepting facility in Philadelphia, finding an ambulance is the next hurdle.

General ED Treatment: The patient with suspected ICH should be a top priority with regard to physician examination, CT scanning, early intervention with intubation when clinically indicated, and rapid basic laboratory tests to assess if the patient has a coagulopathy or an alternative diagnosis, such as hypoglycemia, epidural hematoma, or meningitis. The noncontrast head CT is still the gold standard, although the time by which most hospitals can obtain a fully read noncontrast head CT had not been standardized. My ballpark estimate is that it's probably possible to meet, greet, stabilize, gown, examine, and otherwise place the patient in top priority, in half an hour at best, if you have nothing else to do. Under the best of circumstances, transporting the patient to CT with all the proper precautions and monitoring probably makes a 60-minute turnaround time from arrival to a fully interpreted CT a reasonable expectation. The NINDS criterion for this is 25 minutes, not humanly possible in many community EDs without advanced EMS notification or a stroke alert policy. For that to happen, you probably have to go into the room alongside EMS, put all other patients on the back burner, personally call the CT scanner personnel, and pester the radiologist for a rapid interpretation.

Although not mentioned in these guidelines, raising the head of the bed to 30 degrees and avoiding all oral medications also appear standard. Raising the head of the bed is easy enough, but like more sophisticated interventions, I doubt this has ever been proven to affect morbidity or mortality. It will, however, likely effectuate some decrease in intracranial pressure. Likewise, if the patient requires intubation, the time-honored traditions of the neurointubation protocol, that mandate such Holy Grail status of lidocaine, fentanyl, barbiturates, and other machinations to blunt rising intracranial pressure from intubation, do not have proven benefits.

Glucose/Temperature: Just when we were gearing up for strict glycemic control for patients with anything from sepsis to ICH, that once seemingly brilliant idea has fallen by the wayside. Likely, the observed hyperglycemia is a response to the hemorrhage itself, similar to the WBC count as a response to systemic infection. While trying to manage markedly elevated glucose levels in the ED can be contemplated, this is largely a neurointensive care intervention. I have yet to give insulin in such a scenario.

The dreaded fever in patients with ICH is now also discounted as a major variable under the control of physicians. Other than to measure and lower temperature to normal in patients with brain injury, induced hyperthermia is not recommended. One wonders why therapeutic hypothermia for stroke victims is not as helpful as it is touted to be to save the brain following cardiac arrest; it seems similar to me. I am an advocate of hypothermia for post-cardiac arrest. While all are currently on the hypothermic bandwagon for victims of cardiac arrest, these zealots were also on the bandwagon for high-dose epinephrine, steroids for sepsis and spinal cord trauma, and a bevy of other seemingly too-good-to-be-true interventions.

Seizures: Most clinicians are convinced that a seizure with ICH is a devastating event, worsening outcome by further injuring the brain. It seems so logical. I can remember rushing to give phenytoin to patients with ICH, patting myself on the back that I actually did something useful, prevented seizures, and thwarted further certain brain damage. Not so fast with this intervention either. Although I am somewhat skeptical that prophylactic phenytoin is a real danger, this sacred cow also bit the dust. Per the AHA and ASA: “Use antiepileptic drugs only in patients with clinical seizures or EKG seizures in patients with a change in mental status.” Bravo for recommending EEG monitoring, but this is clearly not an ED intervention at this time. Lobar location and small volume of ICH are independent predictors of immediate seizures. (Epilepsia 2002;43[10]:1175.)

I am disappointed that these guidelines do not take a stand on which anticonvulsants to use when seizures occur. While phenytoin was slammed by a few studies, it's still universally used. It may be better to use benzodiazepines, seemingly always safe and effective, to treat a witnessed seizure, and let the ICU take over.

Surgery: One should certainly call a neurosurgeon when a massive ICH is confirmed. It's probably not a bad idea to give them a heads-up on critical patients with a high suspicion while the patient is still being processed. All EPs should know that cerebellar hemorrhage is a red flag for neurosurgical consultation because most agree that such pathology can cause rapid brain stem compression and herniation. Whether one should fish out the large intracerebral clot or put in an intracranial pressure monitor in the ED is controversial, but local tradition will likely take precedence. I find it very interesting that the guidelines include a table for treating elevated ICP in patients with head trauma, with a suggestion that the algorithm applies to ICH. Mannitol, craniotomy, hypothermia, and barbiturate coma are suggested ICH interventions, but, of course, none has been rigorously studied. To be fair, the AHA and ASA gives lukewarm endorsement to these head trauma guidelines, viewing them as “management principles” with only a “reasonable” recommendation.

Summary: Now that I am completely confused by the new guidelines, I believe that I have a clear understanding of the current state-of-the-art and knowledge of ICH. Simply stated, we don't appear to know much for sure, data are evolving, and we have precious little to offer the patient. Those things that we embark upon with religious fervor and loosely brand standard of care are merely reasonable intuitive interventions that might be considered in some patients.

Despite the erudite tone, this document hardly defines standard of care for the ED. Because there is not a single emergency medicine author, I vehemently protest, and note that some of the recommendations and standards are just not possible in the real world. A careful reading reveals multiple cautions, a lot of hedging, and plethora of “may/might be considered.” After all is said and done, there are only three recommendations give Class 1, Level of Evidence A. But buried deep in a table is a phrase that level of Evidence C, a “consensus of experts,” is considered “standard of care.” None of the issues and challenges that frustrate, confuse, scare, or make us feel better as an emergency physician are currently set in stone by these new guidelines.

Comments about this article? Write to EMN

Click and Connect!Access the links in this article by reading it

Part 3 in a Series

2010 AHA and ASA Guidelines for Treating ICH

2010 AHA and ASA Guidelines for Treating ICH

Back to Top | Article Outline

Seizures in Intracranial Hemorrhage

Patients with ICH are at a modest risk (2%-18%) for early seizures. Lobar location and small-volume ICH are independent predictors of immediate seizures. Prophylactic antiepileptic drugs have been recommended for patients with lobar hemorrhage under the 2007 guidelines from the AHA and ASA (Circulation 2007;116[16]:e391) and by the European Stroke Initiative (Cerebrovasc Dis 2006;22[4]:294).

One might intuit that seizures are bad, and would worsen ICH morbidity and mortality so seizure prophylaxis should be provided. There is no clear evidence, however, that seizures actually worsen the outcome in ICH, and there is evidence that antiepileptic drugs can impair recovery in patients with other types of brain injury, such as ischemic stroke.

According to the National Institute of Neurologic Diseases and Stroke, “the incidence, impact, and role of anticonvulsants [are] poorly understood in the presence of ICH, and the clinical benefit of prophylactic [antiepileptic drug] administration in patients with ICH needs to be assessed.” (Stroke 2005;36[3]:e23.) This is a curious conclusion because antiepileptic drugs given empirically in the ED for patients with acute ICH has been relatively commonplace and supported by clinical practice.

The CHANT study analyzed data from the placebo arm of an international multicenter randomized trial on neuroprotective agents, and found that prophylactic antiepileptic drugs were associated with a poor outcome following ICH. (Neurocrit Care 2009;11[1]:38.) The antiepileptic drug used was primarily phenytoin, but a few received valproic acid or lamotrigine. Barbiturates or benzodiazepines were not used, although they are commonly suggested antiepileptic drugs. While the drugs were effective in stopping seizures post-ICH, the prophylactic intervention was strongly statistically associated with severe disability and death, independent of other significant predictors of poor outcome. Because most patients receive phenytoin, the results may not be generalized to other anti-seizure medications. Other studies have demonstrated that seizures are not a factor in early neurological deterioration in patients with ICH (Neurology 2004;63[3]:461), and phenytoin may be detrimental (Stroke 2009;40[12]:3810).

The bottom line is that numerous observational trials and prospective studies do not support the contention that clinical seizures worsen long-term outcome in patients with ICH. There is concern that using empiric phenytoin, in particular, to prevent potential seizures may be detrimental or otherwise impair recovery. The 2010 AHA-ASA guidelines differ from the 2007 guidelines where prophylactic anticonvulsant use was not cautioned. Overall, the rationale for prophylactic antiepileptic drug use in ICH is questionable. This is based on the conclusion that clinical seizures are actually quite rare following ICH, they have not proven to be detrimental, and the use prophylactic antiepileptic drugs can be associated with poor outcome.

To further confuse clinicians, a study of EEG monitoring found a 28 percent incidence of seizures (nonconvulsive and convulsive) in patients with ICH. (Neurology 2003;60[9]:1441.) In this study, post-hemorrhagic seizures were associated with a worsening neurological function determined by stroke scales, and seizures were associated with progressive CT midline shifts after ICH. The EEG monitoring detected four times as many subclinical seizures compared with clinical seizures. Perhaps the length of subclinical seizing accounted for the worse outcome, but other investigators have not corroborated the results. Importantly, this report did not investigate the use of antiepileptic drug treatment.

Reader Feedback: Readers are invited to ask specific questions and offer personal experiences, comments, or observations on InFocus topics. Literature references are appreciated. Pertinent responses will be published in a future issue. Please send comments to Dr. Roberts requests feedback on this month's column, especially personal experiences with successes, failures, and technique.

Dr. Roberts: I have found that Dermabond works well at repairing nail bed lacerations and also at securing the nail after it has been reinserted under the cuticle (proximal nail fold). In both cases, using medical grade adhesives is far easier than suturing, and at least in the former, it has similar functional and cosmetic results. (J Hand Surg Am 2008;33[2]:250.) Securing the nail to the nail bed with these adhesives is particularly useful when the nail is split in pieces. (J Hand Surg Am 2000;25[5]:979.) — Peter Gruber, MD, Bronx, NY

Dr. Roberts responds: Your approach has been adopted by others, and it seems to work. Thanks for the reminder and the references.

Back to Top | Article Outline

Intracranial Monitoring and Treatment

  • Patients with GCS score of 8 or lower, those with clinical evidence of transtentorial herniation, or those with significant IVH or hydrocephalus might be considered for ICP monitoring and treatment. A cerebral perfusion pressure of 50 to 70 mm Hg may be reasonable to maintain depending on the status of cerebral autoregulation. (Class IIb; Evidence Level C; new recommendation.)
  • Ventricular drainage as treatment for hydrocephalus is reasonable in patients with decreased level of consciousness. (Class IIa; Evidence Level B; new recommendation.)
  • For most patients with ICH, the usefulness of surgery is uncertain. (Class IIb; Evidence Level C; new recommendation.) Specific exceptions to this recommendation follow.
  • Although theoretically attractive, no clear evidence at present indicates that ultra-early removal of supratentorial ICH improves functional outcome or mortality rate. Very early craniotomy may be harmful due to increased risk of recurrent bleeding. (Class III; Evidence Level B; revised from the previous guidelines.)

Source: Guidelines for the Management of Spontaneous Intracerebral Hemorrhage: A Guideline for Healthcare Professionals from the American Heart Association/American Stroke Association. Stroke 2010;41(9):2108.

Dr. Roberts

Dr. Roberts

© 2011 Lippincott Williams & Wilkins, Inc.