Reviews

Aneurysmal subarachnoid hemorrhage is a devastating condition, with an estimated death rate of 30% during the initial episode.^1,2 Approximately the same number of patients survive but leave the hospital with disabling neurologic deficits.³

However, better outcomes can be achieved by systems that are able to work as a team on the collective goal of quick intervention to secure the ruptured aneurysm, followed by the implementation of measures to minimize secondary brain injury. Although the search for new diagnostic, prognostic, and therapeutic modalities continues, it is clear that there exists no “silver bullet” that will help all patients. Instead, it is the systematic integration and application of small advances that will ultimately maximize the patient’s chances of survival and neurologic recovery.

This review focuses on the management of aneurysmal subarachnoid hemorrhage and its systemic and neurologic complications.

ANEURYSM IS THE MOST COMMON CAUSE OF SUBARACHNOID BLEEDING

Aneurysmal subarachnoid hemorrhage, ie, rupture of an intracranial aneurysm, flooding  the subarachnoid space with blood, affects about 24,000 Americans each year.^1,2 A ruptured aneurysm is the most common cause of subarachnoid hemorrhage, accounting for about 85% of cases. Less common causes include idiopathic benign perimesencephalic hemorrhage, arteriovenous malformation, dural arteriovenous fistula, and hemorrhagic mycotic aneurysm. These have their own natural history, pathophysiology, and specific treatment, and will not be addressed in this article.

Risk factors for aneurysmal subarachnoid hemorrhage include having a first-degree relative who had the disease, hypertension, smoking, and consuming more than 150 g of alcohol per week.⁴

CLINICAL PRESENTATION AND DIAGNOSIS

The key symptom of aneurysmal subarachnoid hemorrhage is the abrupt onset of severe headache that peaks in intensity over 1 hour,⁵ often described as “the worst headache of my life.” Headache is accompanied by brief loss of consciousness in 53% of cases (conversely, nearly half of patients maintain normal mental status), by nausea or vomiting in 77%, and by meningismus (neck pain or stiffness) in 35%.⁶

These clinical manifestations and risk factors have been incorporated into a decision rule:

Obtain brain imaging if the patient has acute headache reaching maximal intensity within 1 hour, associated with any of the following factors:

• Age 40 or older

• Neck pain or stiffness

• Witnessed loss of consciousness

• Onset during exertion

• “Thunderclap” headache (ie, instantly peaking pain)

• Limited neck flexion on examination.⁵

This decision rule has nearly 100% sensitivity for aneurysmal subarachnoid hemorrhage in clinical practice.⁵ All patients require brain imaging if they have a severe headache plus either abnormal neurologic findings (eg, a focal neurologic deficit) or a history of cerebral aneurysm.

Emergency physicians should have a low threshold for ordering noncontrast computed tomography (CT) of the head in patients with even mild symptoms suggesting aneurysmal subarachnoid hemorrhage. Failure to order CT is the most common diagnostic error in this situation.⁶ CT performed within 6 hours of headache onset is nearly 100% sensitive for this condition,⁷ but the sensitivity falls to 93% after the first 24 hours and to less than 60% after 5 days.⁸ In patients who have symptoms highly suggestive of aneurysmal subarachnoid hemorrhage but a normal CT, lumbar puncture is the next diagnostic step.

There are two alternatives to CT followed by lumbar puncture: ie, noncontrast CT followed by CT angiography,^9,10 and magnetic resonance imaging followed by magnetic resonance angiography. In patients with suspicious clinical symptoms but negative CT results, CT followed by CT angiography can rule out aneurysmal subarachnoid hemorrhage with a 99% probability.^9,10 However, CT followed by lumbar puncture remains the standard of care and carries a class I recommendation in the American Heart Association guidelines for ruling out subarachnoid hemorrhage.⁵

GRADING THE SEVERITY OF SUBARACHNOID HEMORRHAGE

Age, the thickness of the blood layer in the subarachnoid space, intraventricular hemorrhage and the findings of the neurologic examination at presentation are predictors of long-term outcomes in aneurysmal subarachnoid hemorrhage (Figure 1).

Different grading systems used in clinical practice are based on the findings on the initial neurologic examination and on the initial noncontrast CT (ie, the thickness of the blood, and whether intraventricular hemorrhage is present). Among the most widely used are those developed by Hunt and Hess¹² and by the World Federation of Neurological Surgeons¹¹ (WFNS), and the CT grading scales (Fisher¹³ or its modified version¹⁴)  (Tables 1 and 2). With either the Hunt and Hess scale or the WFNS scale, the higher the score, the worse the patient’s probable outcome. Scores on both Fisher scales correlate with the risk of angiographic vasospasm. The higher the grade, the higher the risk of angiographic vasospasm.

The VASOGRADE score—a combination of the WFNS score and the modified Fisher scale—stratifies patients at risk of delayed cerebral ischemia, allowing for a tailored monitoring strategy.¹⁵ There are three variations:

• VASOGRADE green—Modified Fisher 1 or 2 and WFNS 1 or 2

• VASOGRADE yellow—Modified Fisher 3 or 4 and WFNS 1, 2, or 3

• VASOGRADE red—WFNS 4 or 5. 

After the initial bleeding event, patients with aneurysmal subarachnoid hemorrhage are at high risk of delayed systemic and neurologic complications, with poor functional outcomes. Delayed cerebral ischemia holds the greatest risk of an unfavorable outcome and ultimately can lead to cerebral infarction, disability, and death.^6,7

INITIAL MANAGEMENT

After aneurysmal subarachnoid hemorrhage is diagnosed, the initial management (Figure 2) includes appropriate medical prevention of rebleeding (which includes supportive care, blood pressure management, and, perhaps, the early use of a short course of an antifibrinolytic drug) and early transfer to a high-volume center for securing the aneurysm. The reported incidence of rebleeding varies from 5% to 22% in the first 72 hours. “Ultra-early” rebleeding (within 24 hours of hemorrhage) has been reported, with an incidence as high as 15% and a fatality rate around 70%. Patients with poor-grade aneurysmal subarachnoid hemorrhage, larger aneurysms, and “sentinel bleeds” are at higher risk of rebleeding.16

Outcomes are much better when patients are managed in a high-volume center, with a specialized neurointensive care unit¹⁷ and access to an interdisciplinary team.¹⁸ Regardless of the initial grade, patients with aneurysmal subarachnoid hemorrhage should be quickly transferred to a high-volume center, defined as one treating at least 35 cases per year, and the benefit is greater in centers treating more than 60 cases per year.¹⁹ The higher the caseload in any given hospital, the better the clinical outcomes in this population.²⁰

Treating cerebral aneurysm: Clipping or coiling

Early aneurysm repair is generally considered the standard of care and the best strategy to reduce the risk of rebleeding. Further, early treatment may be associated with a lower risk of delayed cerebral ischemia²¹ and better outcomes.²²

Three randomized clinical trials have compared surgical clipping and endovascular repair (placement of small metal coils within the aneurysm to promote clotting).

The International Subarachnoid Aneurysm Trial²³ showed a reduction of 23% in relative risk and of 7% in absolute risk in patients who underwent endovascular treatment compared with surgery. The survival benefit persisted at a mean of 9 years (range 6–14 years), but with a higher annual rate of aneurysm recurrence in the coiling group (2.9% vs 0.9%).²⁴ Of note, this trial included only patients with aneurysms deemed suitable for both coiling and clipping, so that the exclusion rate was high. Most of the patients presented with good-grade (WFNS score 1–3), small aneurysms (< 5 mm) in the anterior circulation.

A single-center Finnish study²⁵ found no differences in rates of recovery, disability, and  death at 1 year, comparing surgery and endovascular treatment. Additionally, survival rates at a mean follow-up of 39 months were similar, with no late recurrences or aneurysmal bleeding.

Lastly, the Barrow Ruptured Aneurysm Trial26,27 found that patients assigned to endovascular treatment had better 1-year neurologic outcomes, defined as a modified Rankin score of 2 or less. Importantly, 37.7% of patients originally assigned to endovascular treatment crossed over to surgical treatment. The authors then performed intention-to-treat and as-treated analyses. Either way, patients treated by endovascular means had better neurologic outcomes at 1 year. However, no difference in the relative risk reduction in worse outcome was found on 3-year follow-up, and patients treated surgically had higher rates of aneurysm obliteration and required less aneurysm retreatment, both of which were statistically significant.

The question that remains is not whether to clip or whether to coil, but whom to clip and whom to coil.²⁸ That question must be answered on a patient-to-patient basis and requires the expertise of an interventional neuroradiologist and a vascular neurosurgeon—one of the reasons these patients are best cared for in high-volume centers providing such expertise.

MEDICAL PREVENTION OF REBLEEDING

Blood pressure management

There are no systematic data on the optimal blood pressure before securing an aneurysm. Early studies of hemodynamic augmentation in cases of ruptured untreated aneurysm reported rebleeding when the systolic blood pressure was allowed to rise above 160 mm Hg.^29,30 A recent study evaluating hypertensive intracerebral hemorrhage revealed better functional outcomes with intensive lowering of blood pressure (defined as systolic blood pressure < 140 mm Hg) but no significant reduction in the combined rate of death or severe disability.³¹ It is difficult to know if these results can be extrapolated to patients with aneurysmal subarachnoid hemorrhage. Current guidelines^3,32 say that before the aneurysm is treated, the systolic pressure should be lower than 160 mm Hg.

There is no specific drug of choice, but a short-acting, titratable medication is preferable. Nicardipine is a very good option, and labetalol might be an appropriate alternative.³³ Once the aneurysm is secured, all antihypertensive drugs should be held. Hypertension should not be treated unless the patient has clinical signs of a hypertensive crisis, such as flash pulmonary edema, myocardial infarction, or hypertensive encephalopathy.

Antifibrinolytic therapy

Risk factors: Family history, hypertension, smoking, heavy drinking

The role of antifibrinolytic therapy in aneurysmal subarachnoid hemorrhage is controversial and has been studied in 10 clinical trials. In a Swedish study,³⁴ early use of tranexamic acid (1 g intravenously over 10 minutes followed by 1 g every 6 hours for a maximum of 24 hours) reduced the rebleeding rate substantially, from 10.8% to 2.4%, with an 80% reduction in the mortality rate from ultra-early rebleeding. However, a recent Cochrane review that included this study found no overall benefit.³⁵

An ongoing multicenter randomized trial in the Netherlands will, we hope, answer this question in the near future.³⁶ At present, some centers would consider a short course of tranexamic acid before aneurysm treatment.

DIAGNOSIS AND TREATMENT OF COMPLICATIONS

Medical complications are extremely common after aneurysmal subarachnoid hemorrhage. Between 75% and 100% of patients develop some type of systemic or further neurologic derangement, which in turn has a negative impact on the long-term outcome.^37,38 In the first 72 hours, rebleeding is the most feared complication, and as mentioned previously, appropriate blood pressure management and early securing of the aneurysm minimize its risk.

NEUROLOGIC COMPLICATIONS

Hydrocephalus

Hydrocephalus is the most common early neurologic complication after aneurysmal subarachnoid hemorrhage, with an overall incidence of 50%.³⁹ Many patients with poor-grade aneurysmal subarachnoid hemorrhage and patients whose condition deteriorates due to worsening of hydrocephalus require the insertion of an external ventricular drain (Figure 1).

Up to 30% of patients who have a poor-grade aneurysmal subarachnoid hemorrhage improve neurologically with cerebrospinal fluid drainage.⁴⁰ An external ventricular drain can be safely placed, even before aneurysm treatment, and placement does not appear to increase the risk of rebleeding.^39,41 After placement, rapid weaning from the drain (clamping within 24 hours of insertion) is safe, decreases length of stay in the intensive care unit and hospital, and may be more cost-effective than gradual weaning over 96 hours.⁴²

Increased intracranial pressure

Intracranial hypertension is another potential early complication, and is frequently due to the development of hydrocephalus, cerebral edema, or rebleeding. The treatment of increased intracranial pressure does not differ from the approach used in managing severe traumatic brain injury, which includes elevating the head of the bed, sedation, analgesia, normoventilation, and cerebrospinal fluid drainage.

Hypertonic saline has been tested in several studies that were very small but nevertheless consistently showed control of intracranial pressure levels and improvement in cerebral blood flow measured by xenon CT.^43–47 Two of these studies even showed better outcomes at discharge.^43,44 However, the small number of patients prevents any meaningful conclusion regarding the use of hypertonic saline and functional outcomes.

Outcomes are much better when patients are managed in a high-volume center

Barbiturates, hypothermia, and decompressive craniectomy could be tried in refractory cases.⁴⁸ Seule et al⁴⁹ evaluated the role of therapeutic hypothermia with or without barbiturate coma in 100 patients with refractory intracranial hypertension. Only 13 patients received hypothermia by itself. At 1 year, 32 patients had achieved a good functional outcome (Glasgow Outcome Scale score 4 or 5). The remaining patients were severely disabled or had died. Of interest, the median duration of hypothermia was 7 days, and 93% of patients developed some medical complication such as electrolyte disorders (77%), pneumonia (52%), thrombocytopenia (47%), or septic shock syndrome (40%). Six patients died as a consequence of one of these complications.

Decompressive craniectomy can be life-saving in patients with refractory intracranial hypertension. However, most of these patients will die or remain severely disabled or comatose.⁵⁰

Seizure prophylaxis is controversial

Seizures can occur at the onset of intracranial hemorrhage, perioperatively, or later (ie, after the first week). The incidence varied considerably in different reports, ranging from 4% to 26%.⁵¹ Seizures occurring perioperatively, ie, after hospital admission, are less frequent and are usually the manifestation of aneurysm rebleeding.²⁴

The question is not whether to clip or coil, but whom to clip and whom to coil

Seizure prophylaxis remains controversial, especially because the use of phenytoin is associated with increased incidence of cerebral vasospasm, infarction, and worse cognitive outcomes after aneurysmal subarachnoid hemorrhage.^52,53 Therefore, routine prophylactic use of phenytoin is not recommended in these patients,³ although the effect of other antiepileptic drugs is less studied and less clear. Patients may be considered for this therapy if they have multiple risk factors for seizures, such as intraparenchymal hematoma, advanced age (> 65), middle cerebral artery aneurysm, craniotomy for aneurysm clipping, and a short course (≤ 72 hours) of an antiepileptic drug other than phenytoin, especially while the aneurysm is unsecured.³

Levetiracetam may be an alternative to phenytoin, having better pharmacodynamic and kinetic profiles, minimal protein binding, and absence of hepatic metabolism, resulting in a very low risk of drug interaction and better tolerability.^54,55 Because of these advantages, levetiracetam has become the drug of choice in several centers treating aneurysmal subarachnoid hemorrhage in the United States.

Addressing this question, a survey was sent to 25 high-volume aneurysmal subarachnoid hemorrhage academic centers in the United States. All 25 institutions answered the survey, and interestingly, levetiracetam was the first-line agent for 16 (94%) of the 17 responders that used prophylaxis, while only 1 used phenytoin as the agent of choice.⁵⁶

A retrospective cohort study by Murphy-Human et al⁵⁷ showed that a short course of levetiracetam (≤ 72 hours) was associated with higher rates of in-hospital seizures compared with an extended course of phenytoin (eg, entire hospital stay). However, the study did not address functional outcomes.⁵⁷

Continuous electroencephalographic monitoring may be considered in comatose patients, in patients requiring controlled ventilation and sedation, or in patients with unexplained alteration in consciousness. In one series of patients with aneurysmal subarachnoid hemorrhage who received continuous monitoring, the incidence of nonconvulsive status epilepticus was 19%, with an associated mortality rate of 100%.⁵⁸

Continuous quantitative electroencephalography is useful to monitor and to detect angiographic vasospasm and delayed cerebral ischemia. Relative alpha variability and the alpha-delta ratio decrease with ischemia, and this effect can precede angiographic vasospasm by 3 days.^59,60

Delayed cerebral ischemia

Delayed cerebral ischemia is defined as the occurrence of focal neurologic impairment, or a decrease of at least 2 points on the Glasgow Coma Scale that lasts for at least 1 hour, is not apparent immediately after aneurysm occlusion, and not attributable to other causes (eg, hyponatremia, fever).⁶¹

Classically, neurologic deficits that occurred within 2 weeks of aneurysm rupture were ascribed to reduced cerebral blood flow caused by delayed large-vessel vasospasm causing cerebral ischemia.⁶² However, perfusion abnormalities have also been observed with either mild or no demonstrable vasospasm.⁶³ Almost 70% of patients who survive the initial hemorrhage develop some degree of angiographic vasospasm. However, only 30% of those patients will experience symptoms.

In addition to vasospasm of large cerebral arteries, impaired autoregulation and early brain injury within the first 72 hours following subarachnoid hemorrhage may play important roles in the development of delayed cerebral ischemia.⁶⁴ Therefore, the modern concept of delayed cerebral ischemia monitoring should focus on cerebral perfusion rather than vessel diameter measurements. This underscores the importance of comprehensive, standardized monitoring techniques that provide information not only on microvasculature, but also at the level of the microcirculation, with information on perfusion, oxygen utilization and extraction, and autoregulation.

Although transcranial Doppler has been the most commonly applied tool to monitor for angiographic vasospasm, it has a low sensitivity and negative predictive value.³⁷ It is nevertheless a useful technique to monitor good-grade aneurysmal subarachnoid hemorrhage patients (WFNS score 1–3) combined with frequent neurologic examinations (Figure 3).

CT angiography is a good noninvasive alternative to digital subtraction angiography. However, it tends to overestimate the degree of vasoconstriction and does not provide information about perfusion and autoregulation.⁶⁵ Nevertheless, CT angiography combined with a CT perfusion scan can add information about autoregulation and cerebral perfusion and has been shown to be more sensitive for the diagnosis of angiographic vasospasm than transcranial Doppler and digital subtraction angiography (Figure 4).

Patients with a poor clinical condition (WFNS score 4 or 5) or receiving continuous sedation constitute a challenge in monitoring for delayed neurologic deterioration. Neurologic examination is not sensitive enough in this setting to detect subtle changes. In these specific and challenging circumstances, multimodality neuromonitoring may be useful in the early detection of delayed cerebral ischemia and may help guide therapy.⁶⁷

Several noninvasive and invasive techniques have been studied to monitor patients at risk of delayed cerebral ischemia after subarachnoid hemorrhage.⁶⁶ These include continuous electroencephalography, brain tissue oxygenation monitoring (Ptio₂), cerebral microdialysis, thermal diffusion flowmetry, and near-infrared spectroscopy. Of these techniques, Ptio₂, cerebral microdialysis, and continuous electroencephalography (see discussion of seizure prophylaxis above) have been more extensively studied. However, most of the studies were observational and very small, limiting any recommendations for using these techniques in routine clinical practice.⁶⁸

Ptio₂ is measured by inserting an intraparenchymal oxygen-sensitive microelectrode, and microdialysis requires a microcatheter with a semipermeable membrane that allows small soluble substances to cross it into the dialysate. These substances, which include markers of ischemia (ie, glucose, lactate, and pyruvate), excitotoxins (ie, glutamate and aspartate), and membrane cell damage products (ie, glycerol), can be measured. Low Ptio₂ values (< 15 mm Hg) and abnormal mycrodialysate findings (eg, glucose < 0.8 mmol/L, lactate-to-pyruvate ratio > 40) have both been associated with cerebral ischemic events and poor outcome.⁶⁸

Preventing delayed cerebral ischemia

Oral nimodipine 60 mg every 4 hours for 21 days, started on admission, carries a class I, level of evidence A recommendation in the management of aneurysmal subarachnoid hemorrhage.^3,32,69 It improves clinical outcome despite having no effect on the risk of angiographic vasospasm. The mechanism of improved outcome is unclear, but the effect may be a neuroprotective phenomenon limiting the extension of delayed cerebral ischemia.⁷⁰

If hypotension occurs, the dose can be lowered to 30 mg every 2 hours. Whether to discontinue nimodipine in this situation is controversial. Of note, the clinical benefits of nimodipine have not been replicated with other calcium channel blockers (eg, nicardipine).⁷¹

Prophylactic hyperdynamic fluid therapy, known as “triple-H” (hypervolemia, hemodilution, and hypertension) was for years the mainstay of treatment in preventing delayed cerebral ischemia due to vasospasm. However, the clinical data supporting this intervention have been called into question, as analysis of two trials found that hypervolemia did not improve outcomes or reduce the incidence of delayed cerebral ischemia, and in fact increased the rate of complications.^72,73 Based on these findings, current guidelines recommend maintaining euvolemia rather than prophylactic hypervolemia in patients with aneurysmal subarachnoid hemorrhage.^3,32,69

TREATING DELAYED CEREBRAL ISCHEMIA

Hemodynamic augmentation

In patients with neurologic deterioration due to delayed cerebral ischemia, hemodynamic augmentation is the cornerstone of treatment. This is done according to a protocol, started early, involving specific physiologic goals, clinical improvement, and escalation to invasive therapies in a timely fashion in patients at high risk of further neurologic insult (Figure 5).

The physiologic goal is to increase the delivery of oxygen and glucose to the ischemic brain. Hypertension seems to be the most effective component of hemodynamic augmentation regardless of volume status, increasing cerebral blood flow and brain tissue oxygenation, with reversal of delayed cerebral ischemic symptoms in up to two-thirds of treated patients.^74,75 However, this information comes from very small studies, with no randomized trials of induced hypertension available.

The effect of a normal saline fluid bolus in patients suspected of having delayed cerebral ischemia has been shown to increase cerebral blood flow in areas of cerebral ischemia.⁷⁴ If volume augmentation fails to improve the neurologic status, the next step is to artificially induce hypertension using vasopressors. The blood pressure target should be based on clinical improvement. A stepwise approach is reasonable in this situation, and the lowest level of blood pressure at which there is a complete reversal of the new focal neurologic deficit should be maintained.^3,29

Inotropic agents such as dobutamine or milrinone can be considered as alternatives in patients who have new neurologic deficits that are refractory to fluid boluses and vasopressors, or in a setting of subarachnoid hemorrhage-induced cardiomyopathy.^76,77

Once the neurologic deficit is reversed by hemodynamic augmentation, the blood pressure should be maintained for 48 to 72 hours at the level that reversed the deficit completely, carefully reassessed thereafter, and the patient weaned slowly. Unruptured unsecured aneurysms should not prevent blood pressure augmentation in a setting of delayed cerebral ischemia if the culprit aneurysm is treated.^3,32 If the ruptured aneurysm has not been secured, careful blood pressure augmentation can be attempted, keeping in mind that hypertension (> 160/95 mm Hg) is a risk factor for fatal aneurysm rupture.

Endovascular management of delayed cerebral ischemia

When medical augmentation fails to completely reverse the neurologic deficits, endovascular treatment can be considered. Although patients treated early in the course of delayed cerebral ischemia have better neurologic recovery, prophylactic endovascular treatment in asymptomatic patients, even if angiographic signs of spasm are present, does not improve clinical outcomes and carries the risk of fatal arterial rupture.⁷⁸

SYSTEMIC COMPLICATIONS

Hyponatremia and hypovolemia

Aneurysmal subarachnoid hemorrhage is commonly associated with abnormalities of fluid balance and electrolyte derangements. Hyponatremia (serum sodium < 135 mmol/L) occurs in 30% to 50% of patients, while the rate of hypovolemia (decreased circulating blood volume) ranges from 17% to 30%.⁷⁹ Both can negatively affect long-term outcomes.^80,81

Decreased circulating blood volume is a well-described contributor to delayed cerebral ischemia and cerebral infarction after aneurysmal subarachnoid hemorrhage.^80–82 Clinical variables such as heart rate, blood pressure, fluid balance, and serum sodium concentration are usually the cornerstones of intravascular volume status assessment. However, these variables correlate poorly with measured circulating blood volumes in those with aneurysmal subarachnoid hemorrhage.^83,84

The mechanisms responsible for the development of hyponatremia and hypovolemia after aneurysmal subarachnoid hemorrhage are not completely understood. Several factors have been described and may contribute to the increased natriuresis and, hence, to a reduction in circulating blood volume: increased circulating natriuretic peptide concentrations,^85–87 sympathetic nervous system hyperactivation,⁸⁸ and hyperreninemic hypo-
aldosteronism syndrome.^89,90

Guidelines: Before treating the aneurysm, the systolic pressure should be < 160 mm Hg

Lastly, the cerebral salt wasting syndrome, described in the 1950s,⁹¹ was thought to be a key mechanism in the development of hyponatremia and hypovolemia after aneurysmal subarachnoid hemorrhage. In contrast to the syndrome of inappropriate antidiuretic hormone, which is characterized by hyponatremia with a normal or slightly elevated intravascular volume, the characteristic feature of cerebral salt wasting syndrome is the development of hyponatremia in a setting of intravascular volume depletion.⁹² In critically ill neurologic and neurosurgical patients, this differential diagnosis is very difficult, especially in those with aneurysmal subarachnoid hemorrhage in whom the clinical assessment of fluid status is not reliable. These two syndromes might coexist and contribute to the development of hyponatremia after aneurysmal subarachnoid hemorrhage.^92,93

Hoff et al^83,84 prospectively compared the clinical assessment of fluid status by critical and intermediate care nurses and direct measurements of blood volume using pulse dye densitometry. The clinical assessment failed to accurately assess patients’ volume status. Using the same technique to measure circulating blood volume, this group showed that calculation of fluid balance does not provide adequate assessment of fluid status.^83,84

Hemodynamic monitoring tools can help guide fluid replacement in this population. Mutoh et al⁹⁴ randomized 160 patients within 24 hours of hemorrhage to receive early goal-directed fluid therapy (ie, preload volume and cardiac output monitored by transpulmonary thermodilution) vs standard therapy (ie, fluid balance or central venous pressure). Overall, no difference was found in the rates of delayed cerebral ischemia (33% vs 42%; P = .33) or favorable outcome (67% vs 57%; P = .22). However, in the subgroup of poor-grade patients (WFNS score 4 or 5), early goal-directed therapy was associated with a lower rate of delayed cerebral ischemia (5% vs 14%; P = .036) and with better functional outcomes at 3 months (52% vs 36%; P = .026).

Fluid restriction to treat hyponatremia in aneurysmal subarachnoid hemorrhage is no longer recommended because of the increased risk of cerebral infarction due to hypovolemic hypoperfusion.⁸²

Prophylactic use of mineralocorticoids (eg, fludrocortisone, hydrocortisone) has been shown to limit natriuresis, hyponatremia, and the amount of fluid required to maintain euvolemia.^95,96 Higher rates of hypokalemia and hyperglycemia, which can be easily treated, are the most common complications associated with this approach. Additionally, hypertonic saline (eg, 3% saline) can be used to correct hyponatremia in a setting of aneurysmal subarachnoid hemorrhage.⁷⁹

Cardiac complications

Cardiac complications after subarachnoid hemorrhage are most likely related to sympathetic hyperactivity and catecholamine-induced myocyte dysfunction. The pathophysiology is complex, but cardiac complications have a significant negative impact on long-term outcome in these patients.⁹⁷

Electrocardiographic changes and positive cardiac enzymes associated with aneurysmal subarachnoid hemorrhage have been extensively reported. More recently, data from studies of two-dimensional echocardiography have shown that subarachnoid hemorrhage can also be associated with significant wall-motion abnormalities and even overt cardiogenic shock.^98–100

There is no specific curative therapy; the treatment is mainly supportive. Vasopressors and inotropes may be used for hemodynamic augmentation.

Pulmonary complications

Pulmonary complications occur in 20% to 30% of all aneurysmal subarachnoid hemorrhage patients and are associated with a higher risk of delayed cerebral ischemia and death. Common pulmonary complications in this population are mild acute respiratory distress syndrome (27%), hospital-acquired pneumonia (9%), cardiogenic pulmonary edema (8%), aspiration pneumonia (6%), neurogenic pulmonary edema (2%), and pulmonary embolism (1%).^101–103

SUPPORTIVE CARE

Hyperthermia, hyperglycemia, and liberal use of transfusions have all been associated with longer stays in the intensive care unit and hospital, poorer neurologic outcomes, and higher mortality rates in patients with acute brain injury.¹⁰⁴ Noninfectious fever is the most common systemic complication after subarachnoid hemorrhage.

Antipyretic drugs such as acetaminophen and ibuprofen are not very effective in reducing fever in the subarachnoid hemorrhage population, but should still be used as first-line therapy. The use of surface and intravascular devices can be considered when fevers do not respond to nonsteroidal anti-inflammatory drugs.

Fluid restriction to treat hyponatremia in aneurysmal subarachnoid hemorrhage is no longer recommended

Although no prospective randomized trial has addressed the impact of induced normothermia on long-term outcome and mortality in aneurysmal subarachnoid hemorrhage patients, fever control has been shown to reduce cerebral metabolic distress, irrespective of intracranial pressure.¹⁰⁵ Maintenance of normothermia (< 37.5°C) seems reasonable, especially in aneurysmal subarachnoid hemorrhage patients at risk of or with active delayed cerebral ischemia.¹⁰⁶

Current guidelines^3,32,69 strongly recommend avoiding hypoglycemia, defined as a serum glucose level less than 80 mg/dL, but suggest keeping the blood sugar level below 180 or 200 mg/dL.

At the moment, there is no clear threshold for transfusion in patients with aneurysmal subarachnoid hemorrhage. Current guidelines suggest keeping hemoglobin levels between 8 and 10 g/dL.³

Preventing venous thromboembolism

The incidence of venous thromboembolism after aneurysmal subarachnoid hemorrhage varies widely, from 1.5% to 18%.¹⁰⁷ Active surveillance with venous Doppler ultrasonography has found asymptomatic deep vein thrombosis in up to 3.4% of poor-grade aneurysmal subarachnoid hemorrhage patients receiving pharmacologic thromboprophylaxis.¹⁰⁸

In a retrospective study of 170 patients, our group showed that giving drugs to prevent venous thromboembolism (unfractionated heparin 5,000 IU subcutaneously every 12 hours or dalteparin 5,000 IU subcutaneously daily), starting within 24 hours of aneurysm treatment, could be safe.¹⁰⁹ Fifty-eight percent of these patients had an external ventricular drain in place. One patient developed a major cerebral hemorrhagic complication and died while on unfractionated heparin; however, the patient was also on dual antiplatelet therapy with aspirin and clopidogrel.¹⁰⁹

Current guidelines suggest that intermittent compression devices be applied in all patients before aneurysm treatment. Pharmacologic thromboprophylaxis with a heparinoid can be started 12 to 24 hours after aneurysm treatment.^3,109

A TEAM APPROACH

Patients with subarachnoid hemorrhage need integrated care from different medical and nursing specialties. The best outcomes are achieved by systems that can focus as a team on the collective goal of quick intervention to secure the aneurysm, followed by measures to minimize secondary brain injury.

The modern concept of cerebral monitoring in a setting of subarachnoid hemorrhage should focus on brain perfusion rather than vascular diameter. Although the search continues for new diagnostic, prognostic, and therapeutic tools, there is no “silver bullet” that will help all patients. Instead, it is the systematic integration and application of many small advances that will ultimately lead to better outcomes.

ACKNOWLEDGMENT

This work was supported by research funding provided by the Bitove Foundation, which has been supportive of our clinical and research work for several years.

Aneurysmal subarachnoid hemorrhage is a devastating condition, with an estimated death rate of 30% during the initial episode.^1,2 Approximately the same number of patients survive but leave the hospital with disabling neurologic deficits.³

However, better outcomes can be achieved by systems that are able to work as a team on the collective goal of quick intervention to secure the ruptured aneurysm, followed by the implementation of measures to minimize secondary brain injury. Although the search for new diagnostic, prognostic, and therapeutic modalities continues, it is clear that there exists no “silver bullet” that will help all patients. Instead, it is the systematic integration and application of small advances that will ultimately maximize the patient’s chances of survival and neurologic recovery.

This review focuses on the management of aneurysmal subarachnoid hemorrhage and its systemic and neurologic complications.

ANEURYSM IS THE MOST COMMON CAUSE OF SUBARACHNOID BLEEDING

Aneurysmal subarachnoid hemorrhage, ie, rupture of an intracranial aneurysm, flooding  the subarachnoid space with blood, affects about 24,000 Americans each year.^1,2 A ruptured aneurysm is the most common cause of subarachnoid hemorrhage, accounting for about 85% of cases. Less common causes include idiopathic benign perimesencephalic hemorrhage, arteriovenous malformation, dural arteriovenous fistula, and hemorrhagic mycotic aneurysm. These have their own natural history, pathophysiology, and specific treatment, and will not be addressed in this article.

Risk factors for aneurysmal subarachnoid hemorrhage include having a first-degree relative who had the disease, hypertension, smoking, and consuming more than 150 g of alcohol per week.⁴

CLINICAL PRESENTATION AND DIAGNOSIS

The key symptom of aneurysmal subarachnoid hemorrhage is the abrupt onset of severe headache that peaks in intensity over 1 hour,⁵ often described as “the worst headache of my life.” Headache is accompanied by brief loss of consciousness in 53% of cases (conversely, nearly half of patients maintain normal mental status), by nausea or vomiting in 77%, and by meningismus (neck pain or stiffness) in 35%.⁶

These clinical manifestations and risk factors have been incorporated into a decision rule:

Obtain brain imaging if the patient has acute headache reaching maximal intensity within 1 hour, associated with any of the following factors:

• Age 40 or older

• Neck pain or stiffness

• Witnessed loss of consciousness

• Onset during exertion

• “Thunderclap” headache (ie, instantly peaking pain)

• Limited neck flexion on examination.⁵

This decision rule has nearly 100% sensitivity for aneurysmal subarachnoid hemorrhage in clinical practice.⁵ All patients require brain imaging if they have a severe headache plus either abnormal neurologic findings (eg, a focal neurologic deficit) or a history of cerebral aneurysm.

Emergency physicians should have a low threshold for ordering noncontrast computed tomography (CT) of the head in patients with even mild symptoms suggesting aneurysmal subarachnoid hemorrhage. Failure to order CT is the most common diagnostic error in this situation.⁶ CT performed within 6 hours of headache onset is nearly 100% sensitive for this condition,⁷ but the sensitivity falls to 93% after the first 24 hours and to less than 60% after 5 days.⁸ In patients who have symptoms highly suggestive of aneurysmal subarachnoid hemorrhage but a normal CT, lumbar puncture is the next diagnostic step.

There are two alternatives to CT followed by lumbar puncture: ie, noncontrast CT followed by CT angiography,^9,10 and magnetic resonance imaging followed by magnetic resonance angiography. In patients with suspicious clinical symptoms but negative CT results, CT followed by CT angiography can rule out aneurysmal subarachnoid hemorrhage with a 99% probability.^9,10 However, CT followed by lumbar puncture remains the standard of care and carries a class I recommendation in the American Heart Association guidelines for ruling out subarachnoid hemorrhage.⁵

GRADING THE SEVERITY OF SUBARACHNOID HEMORRHAGE

Age, the thickness of the blood layer in the subarachnoid space, intraventricular hemorrhage and the findings of the neurologic examination at presentation are predictors of long-term outcomes in aneurysmal subarachnoid hemorrhage (Figure 1).

Different grading systems used in clinical practice are based on the findings on the initial neurologic examination and on the initial noncontrast CT (ie, the thickness of the blood, and whether intraventricular hemorrhage is present). Among the most widely used are those developed by Hunt and Hess¹² and by the World Federation of Neurological Surgeons¹¹ (WFNS), and the CT grading scales (Fisher¹³ or its modified version¹⁴)  (Tables 1 and 2). With either the Hunt and Hess scale or the WFNS scale, the higher the score, the worse the patient’s probable outcome. Scores on both Fisher scales correlate with the risk of angiographic vasospasm. The higher the grade, the higher the risk of angiographic vasospasm.

The VASOGRADE score—a combination of the WFNS score and the modified Fisher scale—stratifies patients at risk of delayed cerebral ischemia, allowing for a tailored monitoring strategy.¹⁵ There are three variations:

• VASOGRADE green—Modified Fisher 1 or 2 and WFNS 1 or 2

• VASOGRADE yellow—Modified Fisher 3 or 4 and WFNS 1, 2, or 3

• VASOGRADE red—WFNS 4 or 5. 

After the initial bleeding event, patients with aneurysmal subarachnoid hemorrhage are at high risk of delayed systemic and neurologic complications, with poor functional outcomes. Delayed cerebral ischemia holds the greatest risk of an unfavorable outcome and ultimately can lead to cerebral infarction, disability, and death.^6,7

INITIAL MANAGEMENT

After aneurysmal subarachnoid hemorrhage is diagnosed, the initial management (Figure 2) includes appropriate medical prevention of rebleeding (which includes supportive care, blood pressure management, and, perhaps, the early use of a short course of an antifibrinolytic drug) and early transfer to a high-volume center for securing the aneurysm. The reported incidence of rebleeding varies from 5% to 22% in the first 72 hours. “Ultra-early” rebleeding (within 24 hours of hemorrhage) has been reported, with an incidence as high as 15% and a fatality rate around 70%. Patients with poor-grade aneurysmal subarachnoid hemorrhage, larger aneurysms, and “sentinel bleeds” are at higher risk of rebleeding.16

Outcomes are much better when patients are managed in a high-volume center, with a specialized neurointensive care unit¹⁷ and access to an interdisciplinary team.¹⁸ Regardless of the initial grade, patients with aneurysmal subarachnoid hemorrhage should be quickly transferred to a high-volume center, defined as one treating at least 35 cases per year, and the benefit is greater in centers treating more than 60 cases per year.¹⁹ The higher the caseload in any given hospital, the better the clinical outcomes in this population.²⁰

Treating cerebral aneurysm: Clipping or coiling

Early aneurysm repair is generally considered the standard of care and the best strategy to reduce the risk of rebleeding. Further, early treatment may be associated with a lower risk of delayed cerebral ischemia²¹ and better outcomes.²²

Three randomized clinical trials have compared surgical clipping and endovascular repair (placement of small metal coils within the aneurysm to promote clotting).

The International Subarachnoid Aneurysm Trial²³ showed a reduction of 23% in relative risk and of 7% in absolute risk in patients who underwent endovascular treatment compared with surgery. The survival benefit persisted at a mean of 9 years (range 6–14 years), but with a higher annual rate of aneurysm recurrence in the coiling group (2.9% vs 0.9%).²⁴ Of note, this trial included only patients with aneurysms deemed suitable for both coiling and clipping, so that the exclusion rate was high. Most of the patients presented with good-grade (WFNS score 1–3), small aneurysms (< 5 mm) in the anterior circulation.

A single-center Finnish study²⁵ found no differences in rates of recovery, disability, and  death at 1 year, comparing surgery and endovascular treatment. Additionally, survival rates at a mean follow-up of 39 months were similar, with no late recurrences or aneurysmal bleeding.

Lastly, the Barrow Ruptured Aneurysm Trial26,27 found that patients assigned to endovascular treatment had better 1-year neurologic outcomes, defined as a modified Rankin score of 2 or less. Importantly, 37.7% of patients originally assigned to endovascular treatment crossed over to surgical treatment. The authors then performed intention-to-treat and as-treated analyses. Either way, patients treated by endovascular means had better neurologic outcomes at 1 year. However, no difference in the relative risk reduction in worse outcome was found on 3-year follow-up, and patients treated surgically had higher rates of aneurysm obliteration and required less aneurysm retreatment, both of which were statistically significant.

The question that remains is not whether to clip or whether to coil, but whom to clip and whom to coil.²⁸ That question must be answered on a patient-to-patient basis and requires the expertise of an interventional neuroradiologist and a vascular neurosurgeon—one of the reasons these patients are best cared for in high-volume centers providing such expertise.

MEDICAL PREVENTION OF REBLEEDING

Blood pressure management

There are no systematic data on the optimal blood pressure before securing an aneurysm. Early studies of hemodynamic augmentation in cases of ruptured untreated aneurysm reported rebleeding when the systolic blood pressure was allowed to rise above 160 mm Hg.^29,30 A recent study evaluating hypertensive intracerebral hemorrhage revealed better functional outcomes with intensive lowering of blood pressure (defined as systolic blood pressure < 140 mm Hg) but no significant reduction in the combined rate of death or severe disability.³¹ It is difficult to know if these results can be extrapolated to patients with aneurysmal subarachnoid hemorrhage. Current guidelines^3,32 say that before the aneurysm is treated, the systolic pressure should be lower than 160 mm Hg.

There is no specific drug of choice, but a short-acting, titratable medication is preferable. Nicardipine is a very good option, and labetalol might be an appropriate alternative.³³ Once the aneurysm is secured, all antihypertensive drugs should be held. Hypertension should not be treated unless the patient has clinical signs of a hypertensive crisis, such as flash pulmonary edema, myocardial infarction, or hypertensive encephalopathy.

Antifibrinolytic therapy

Risk factors: Family history, hypertension, smoking, heavy drinking

The role of antifibrinolytic therapy in aneurysmal subarachnoid hemorrhage is controversial and has been studied in 10 clinical trials. In a Swedish study,³⁴ early use of tranexamic acid (1 g intravenously over 10 minutes followed by 1 g every 6 hours for a maximum of 24 hours) reduced the rebleeding rate substantially, from 10.8% to 2.4%, with an 80% reduction in the mortality rate from ultra-early rebleeding. However, a recent Cochrane review that included this study found no overall benefit.³⁵

An ongoing multicenter randomized trial in the Netherlands will, we hope, answer this question in the near future.³⁶ At present, some centers would consider a short course of tranexamic acid before aneurysm treatment.

DIAGNOSIS AND TREATMENT OF COMPLICATIONS

Medical complications are extremely common after aneurysmal subarachnoid hemorrhage. Between 75% and 100% of patients develop some type of systemic or further neurologic derangement, which in turn has a negative impact on the long-term outcome.^37,38 In the first 72 hours, rebleeding is the most feared complication, and as mentioned previously, appropriate blood pressure management and early securing of the aneurysm minimize its risk.

NEUROLOGIC COMPLICATIONS

Hydrocephalus

Hydrocephalus is the most common early neurologic complication after aneurysmal subarachnoid hemorrhage, with an overall incidence of 50%.³⁹ Many patients with poor-grade aneurysmal subarachnoid hemorrhage and patients whose condition deteriorates due to worsening of hydrocephalus require the insertion of an external ventricular drain (Figure 1).

Up to 30% of patients who have a poor-grade aneurysmal subarachnoid hemorrhage improve neurologically with cerebrospinal fluid drainage.⁴⁰ An external ventricular drain can be safely placed, even before aneurysm treatment, and placement does not appear to increase the risk of rebleeding.^39,41 After placement, rapid weaning from the drain (clamping within 24 hours of insertion) is safe, decreases length of stay in the intensive care unit and hospital, and may be more cost-effective than gradual weaning over 96 hours.⁴²

Increased intracranial pressure

Intracranial hypertension is another potential early complication, and is frequently due to the development of hydrocephalus, cerebral edema, or rebleeding. The treatment of increased intracranial pressure does not differ from the approach used in managing severe traumatic brain injury, which includes elevating the head of the bed, sedation, analgesia, normoventilation, and cerebrospinal fluid drainage.

Hypertonic saline has been tested in several studies that were very small but nevertheless consistently showed control of intracranial pressure levels and improvement in cerebral blood flow measured by xenon CT.^43–47 Two of these studies even showed better outcomes at discharge.^43,44 However, the small number of patients prevents any meaningful conclusion regarding the use of hypertonic saline and functional outcomes.

Outcomes are much better when patients are managed in a high-volume center

Barbiturates, hypothermia, and decompressive craniectomy could be tried in refractory cases.⁴⁸ Seule et al⁴⁹ evaluated the role of therapeutic hypothermia with or without barbiturate coma in 100 patients with refractory intracranial hypertension. Only 13 patients received hypothermia by itself. At 1 year, 32 patients had achieved a good functional outcome (Glasgow Outcome Scale score 4 or 5). The remaining patients were severely disabled or had died. Of interest, the median duration of hypothermia was 7 days, and 93% of patients developed some medical complication such as electrolyte disorders (77%), pneumonia (52%), thrombocytopenia (47%), or septic shock syndrome (40%). Six patients died as a consequence of one of these complications.

Decompressive craniectomy can be life-saving in patients with refractory intracranial hypertension. However, most of these patients will die or remain severely disabled or comatose.⁵⁰

Seizure prophylaxis is controversial

Seizures can occur at the onset of intracranial hemorrhage, perioperatively, or later (ie, after the first week). The incidence varied considerably in different reports, ranging from 4% to 26%.⁵¹ Seizures occurring perioperatively, ie, after hospital admission, are less frequent and are usually the manifestation of aneurysm rebleeding.²⁴

The question is not whether to clip or coil, but whom to clip and whom to coil

Seizure prophylaxis remains controversial, especially because the use of phenytoin is associated with increased incidence of cerebral vasospasm, infarction, and worse cognitive outcomes after aneurysmal subarachnoid hemorrhage.^52,53 Therefore, routine prophylactic use of phenytoin is not recommended in these patients,³ although the effect of other antiepileptic drugs is less studied and less clear. Patients may be considered for this therapy if they have multiple risk factors for seizures, such as intraparenchymal hematoma, advanced age (> 65), middle cerebral artery aneurysm, craniotomy for aneurysm clipping, and a short course (≤ 72 hours) of an antiepileptic drug other than phenytoin, especially while the aneurysm is unsecured.³

Levetiracetam may be an alternative to phenytoin, having better pharmacodynamic and kinetic profiles, minimal protein binding, and absence of hepatic metabolism, resulting in a very low risk of drug interaction and better tolerability.^54,55 Because of these advantages, levetiracetam has become the drug of choice in several centers treating aneurysmal subarachnoid hemorrhage in the United States.

Addressing this question, a survey was sent to 25 high-volume aneurysmal subarachnoid hemorrhage academic centers in the United States. All 25 institutions answered the survey, and interestingly, levetiracetam was the first-line agent for 16 (94%) of the 17 responders that used prophylaxis, while only 1 used phenytoin as the agent of choice.⁵⁶

A retrospective cohort study by Murphy-Human et al⁵⁷ showed that a short course of levetiracetam (≤ 72 hours) was associated with higher rates of in-hospital seizures compared with an extended course of phenytoin (eg, entire hospital stay). However, the study did not address functional outcomes.⁵⁷

Continuous electroencephalographic monitoring may be considered in comatose patients, in patients requiring controlled ventilation and sedation, or in patients with unexplained alteration in consciousness. In one series of patients with aneurysmal subarachnoid hemorrhage who received continuous monitoring, the incidence of nonconvulsive status epilepticus was 19%, with an associated mortality rate of 100%.⁵⁸

Continuous quantitative electroencephalography is useful to monitor and to detect angiographic vasospasm and delayed cerebral ischemia. Relative alpha variability and the alpha-delta ratio decrease with ischemia, and this effect can precede angiographic vasospasm by 3 days.^59,60

Delayed cerebral ischemia

Delayed cerebral ischemia is defined as the occurrence of focal neurologic impairment, or a decrease of at least 2 points on the Glasgow Coma Scale that lasts for at least 1 hour, is not apparent immediately after aneurysm occlusion, and not attributable to other causes (eg, hyponatremia, fever).⁶¹

Classically, neurologic deficits that occurred within 2 weeks of aneurysm rupture were ascribed to reduced cerebral blood flow caused by delayed large-vessel vasospasm causing cerebral ischemia.⁶² However, perfusion abnormalities have also been observed with either mild or no demonstrable vasospasm.⁶³ Almost 70% of patients who survive the initial hemorrhage develop some degree of angiographic vasospasm. However, only 30% of those patients will experience symptoms.

In addition to vasospasm of large cerebral arteries, impaired autoregulation and early brain injury within the first 72 hours following subarachnoid hemorrhage may play important roles in the development of delayed cerebral ischemia.⁶⁴ Therefore, the modern concept of delayed cerebral ischemia monitoring should focus on cerebral perfusion rather than vessel diameter measurements. This underscores the importance of comprehensive, standardized monitoring techniques that provide information not only on microvasculature, but also at the level of the microcirculation, with information on perfusion, oxygen utilization and extraction, and autoregulation.

Although transcranial Doppler has been the most commonly applied tool to monitor for angiographic vasospasm, it has a low sensitivity and negative predictive value.³⁷ It is nevertheless a useful technique to monitor good-grade aneurysmal subarachnoid hemorrhage patients (WFNS score 1–3) combined with frequent neurologic examinations (Figure 3).

CT angiography is a good noninvasive alternative to digital subtraction angiography. However, it tends to overestimate the degree of vasoconstriction and does not provide information about perfusion and autoregulation.⁶⁵ Nevertheless, CT angiography combined with a CT perfusion scan can add information about autoregulation and cerebral perfusion and has been shown to be more sensitive for the diagnosis of angiographic vasospasm than transcranial Doppler and digital subtraction angiography (Figure 4).

Patients with a poor clinical condition (WFNS score 4 or 5) or receiving continuous sedation constitute a challenge in monitoring for delayed neurologic deterioration. Neurologic examination is not sensitive enough in this setting to detect subtle changes. In these specific and challenging circumstances, multimodality neuromonitoring may be useful in the early detection of delayed cerebral ischemia and may help guide therapy.⁶⁷

Several noninvasive and invasive techniques have been studied to monitor patients at risk of delayed cerebral ischemia after subarachnoid hemorrhage.⁶⁶ These include continuous electroencephalography, brain tissue oxygenation monitoring (Ptio₂), cerebral microdialysis, thermal diffusion flowmetry, and near-infrared spectroscopy. Of these techniques, Ptio₂, cerebral microdialysis, and continuous electroencephalography (see discussion of seizure prophylaxis above) have been more extensively studied. However, most of the studies were observational and very small, limiting any recommendations for using these techniques in routine clinical practice.⁶⁸

Ptio₂ is measured by inserting an intraparenchymal oxygen-sensitive microelectrode, and microdialysis requires a microcatheter with a semipermeable membrane that allows small soluble substances to cross it into the dialysate. These substances, which include markers of ischemia (ie, glucose, lactate, and pyruvate), excitotoxins (ie, glutamate and aspartate), and membrane cell damage products (ie, glycerol), can be measured. Low Ptio₂ values (< 15 mm Hg) and abnormal mycrodialysate findings (eg, glucose < 0.8 mmol/L, lactate-to-pyruvate ratio > 40) have both been associated with cerebral ischemic events and poor outcome.⁶⁸

Preventing delayed cerebral ischemia

Oral nimodipine 60 mg every 4 hours for 21 days, started on admission, carries a class I, level of evidence A recommendation in the management of aneurysmal subarachnoid hemorrhage.^3,32,69 It improves clinical outcome despite having no effect on the risk of angiographic vasospasm. The mechanism of improved outcome is unclear, but the effect may be a neuroprotective phenomenon limiting the extension of delayed cerebral ischemia.⁷⁰

If hypotension occurs, the dose can be lowered to 30 mg every 2 hours. Whether to discontinue nimodipine in this situation is controversial. Of note, the clinical benefits of nimodipine have not been replicated with other calcium channel blockers (eg, nicardipine).⁷¹

Prophylactic hyperdynamic fluid therapy, known as “triple-H” (hypervolemia, hemodilution, and hypertension) was for years the mainstay of treatment in preventing delayed cerebral ischemia due to vasospasm. However, the clinical data supporting this intervention have been called into question, as analysis of two trials found that hypervolemia did not improve outcomes or reduce the incidence of delayed cerebral ischemia, and in fact increased the rate of complications.^72,73 Based on these findings, current guidelines recommend maintaining euvolemia rather than prophylactic hypervolemia in patients with aneurysmal subarachnoid hemorrhage.^3,32,69

TREATING DELAYED CEREBRAL ISCHEMIA

Hemodynamic augmentation

In patients with neurologic deterioration due to delayed cerebral ischemia, hemodynamic augmentation is the cornerstone of treatment. This is done according to a protocol, started early, involving specific physiologic goals, clinical improvement, and escalation to invasive therapies in a timely fashion in patients at high risk of further neurologic insult (Figure 5).

The physiologic goal is to increase the delivery of oxygen and glucose to the ischemic brain. Hypertension seems to be the most effective component of hemodynamic augmentation regardless of volume status, increasing cerebral blood flow and brain tissue oxygenation, with reversal of delayed cerebral ischemic symptoms in up to two-thirds of treated patients.^74,75 However, this information comes from very small studies, with no randomized trials of induced hypertension available.

The effect of a normal saline fluid bolus in patients suspected of having delayed cerebral ischemia has been shown to increase cerebral blood flow in areas of cerebral ischemia.⁷⁴ If volume augmentation fails to improve the neurologic status, the next step is to artificially induce hypertension using vasopressors. The blood pressure target should be based on clinical improvement. A stepwise approach is reasonable in this situation, and the lowest level of blood pressure at which there is a complete reversal of the new focal neurologic deficit should be maintained.^3,29

Inotropic agents such as dobutamine or milrinone can be considered as alternatives in patients who have new neurologic deficits that are refractory to fluid boluses and vasopressors, or in a setting of subarachnoid hemorrhage-induced cardiomyopathy.^76,77

Once the neurologic deficit is reversed by hemodynamic augmentation, the blood pressure should be maintained for 48 to 72 hours at the level that reversed the deficit completely, carefully reassessed thereafter, and the patient weaned slowly. Unruptured unsecured aneurysms should not prevent blood pressure augmentation in a setting of delayed cerebral ischemia if the culprit aneurysm is treated.^3,32 If the ruptured aneurysm has not been secured, careful blood pressure augmentation can be attempted, keeping in mind that hypertension (> 160/95 mm Hg) is a risk factor for fatal aneurysm rupture.

Endovascular management of delayed cerebral ischemia

When medical augmentation fails to completely reverse the neurologic deficits, endovascular treatment can be considered. Although patients treated early in the course of delayed cerebral ischemia have better neurologic recovery, prophylactic endovascular treatment in asymptomatic patients, even if angiographic signs of spasm are present, does not improve clinical outcomes and carries the risk of fatal arterial rupture.⁷⁸

SYSTEMIC COMPLICATIONS

Hyponatremia and hypovolemia

Aneurysmal subarachnoid hemorrhage is commonly associated with abnormalities of fluid balance and electrolyte derangements. Hyponatremia (serum sodium < 135 mmol/L) occurs in 30% to 50% of patients, while the rate of hypovolemia (decreased circulating blood volume) ranges from 17% to 30%.⁷⁹ Both can negatively affect long-term outcomes.^80,81

Decreased circulating blood volume is a well-described contributor to delayed cerebral ischemia and cerebral infarction after aneurysmal subarachnoid hemorrhage.^80–82 Clinical variables such as heart rate, blood pressure, fluid balance, and serum sodium concentration are usually the cornerstones of intravascular volume status assessment. However, these variables correlate poorly with measured circulating blood volumes in those with aneurysmal subarachnoid hemorrhage.^83,84

The mechanisms responsible for the development of hyponatremia and hypovolemia after aneurysmal subarachnoid hemorrhage are not completely understood. Several factors have been described and may contribute to the increased natriuresis and, hence, to a reduction in circulating blood volume: increased circulating natriuretic peptide concentrations,^85–87 sympathetic nervous system hyperactivation,⁸⁸ and hyperreninemic hypo-
aldosteronism syndrome.^89,90

Guidelines: Before treating the aneurysm, the systolic pressure should be < 160 mm Hg

Lastly, the cerebral salt wasting syndrome, described in the 1950s,⁹¹ was thought to be a key mechanism in the development of hyponatremia and hypovolemia after aneurysmal subarachnoid hemorrhage. In contrast to the syndrome of inappropriate antidiuretic hormone, which is characterized by hyponatremia with a normal or slightly elevated intravascular volume, the characteristic feature of cerebral salt wasting syndrome is the development of hyponatremia in a setting of intravascular volume depletion.⁹² In critically ill neurologic and neurosurgical patients, this differential diagnosis is very difficult, especially in those with aneurysmal subarachnoid hemorrhage in whom the clinical assessment of fluid status is not reliable. These two syndromes might coexist and contribute to the development of hyponatremia after aneurysmal subarachnoid hemorrhage.^92,93

Hoff et al^83,84 prospectively compared the clinical assessment of fluid status by critical and intermediate care nurses and direct measurements of blood volume using pulse dye densitometry. The clinical assessment failed to accurately assess patients’ volume status. Using the same technique to measure circulating blood volume, this group showed that calculation of fluid balance does not provide adequate assessment of fluid status.^83,84

Hemodynamic monitoring tools can help guide fluid replacement in this population. Mutoh et al⁹⁴ randomized 160 patients within 24 hours of hemorrhage to receive early goal-directed fluid therapy (ie, preload volume and cardiac output monitored by transpulmonary thermodilution) vs standard therapy (ie, fluid balance or central venous pressure). Overall, no difference was found in the rates of delayed cerebral ischemia (33% vs 42%; P = .33) or favorable outcome (67% vs 57%; P = .22). However, in the subgroup of poor-grade patients (WFNS score 4 or 5), early goal-directed therapy was associated with a lower rate of delayed cerebral ischemia (5% vs 14%; P = .036) and with better functional outcomes at 3 months (52% vs 36%; P = .026).

Fluid restriction to treat hyponatremia in aneurysmal subarachnoid hemorrhage is no longer recommended because of the increased risk of cerebral infarction due to hypovolemic hypoperfusion.⁸²

Prophylactic use of mineralocorticoids (eg, fludrocortisone, hydrocortisone) has been shown to limit natriuresis, hyponatremia, and the amount of fluid required to maintain euvolemia.^95,96 Higher rates of hypokalemia and hyperglycemia, which can be easily treated, are the most common complications associated with this approach. Additionally, hypertonic saline (eg, 3% saline) can be used to correct hyponatremia in a setting of aneurysmal subarachnoid hemorrhage.⁷⁹

Cardiac complications

Cardiac complications after subarachnoid hemorrhage are most likely related to sympathetic hyperactivity and catecholamine-induced myocyte dysfunction. The pathophysiology is complex, but cardiac complications have a significant negative impact on long-term outcome in these patients.⁹⁷

Electrocardiographic changes and positive cardiac enzymes associated with aneurysmal subarachnoid hemorrhage have been extensively reported. More recently, data from studies of two-dimensional echocardiography have shown that subarachnoid hemorrhage can also be associated with significant wall-motion abnormalities and even overt cardiogenic shock.^98–100

There is no specific curative therapy; the treatment is mainly supportive. Vasopressors and inotropes may be used for hemodynamic augmentation.

Pulmonary complications

Pulmonary complications occur in 20% to 30% of all aneurysmal subarachnoid hemorrhage patients and are associated with a higher risk of delayed cerebral ischemia and death. Common pulmonary complications in this population are mild acute respiratory distress syndrome (27%), hospital-acquired pneumonia (9%), cardiogenic pulmonary edema (8%), aspiration pneumonia (6%), neurogenic pulmonary edema (2%), and pulmonary embolism (1%).^101–103

SUPPORTIVE CARE

Hyperthermia, hyperglycemia, and liberal use of transfusions have all been associated with longer stays in the intensive care unit and hospital, poorer neurologic outcomes, and higher mortality rates in patients with acute brain injury.¹⁰⁴ Noninfectious fever is the most common systemic complication after subarachnoid hemorrhage.

Antipyretic drugs such as acetaminophen and ibuprofen are not very effective in reducing fever in the subarachnoid hemorrhage population, but should still be used as first-line therapy. The use of surface and intravascular devices can be considered when fevers do not respond to nonsteroidal anti-inflammatory drugs.

Fluid restriction to treat hyponatremia in aneurysmal subarachnoid hemorrhage is no longer recommended

Although no prospective randomized trial has addressed the impact of induced normothermia on long-term outcome and mortality in aneurysmal subarachnoid hemorrhage patients, fever control has been shown to reduce cerebral metabolic distress, irrespective of intracranial pressure.¹⁰⁵ Maintenance of normothermia (< 37.5°C) seems reasonable, especially in aneurysmal subarachnoid hemorrhage patients at risk of or with active delayed cerebral ischemia.¹⁰⁶

Current guidelines^3,32,69 strongly recommend avoiding hypoglycemia, defined as a serum glucose level less than 80 mg/dL, but suggest keeping the blood sugar level below 180 or 200 mg/dL.

At the moment, there is no clear threshold for transfusion in patients with aneurysmal subarachnoid hemorrhage. Current guidelines suggest keeping hemoglobin levels between 8 and 10 g/dL.³

Preventing venous thromboembolism

The incidence of venous thromboembolism after aneurysmal subarachnoid hemorrhage varies widely, from 1.5% to 18%.¹⁰⁷ Active surveillance with venous Doppler ultrasonography has found asymptomatic deep vein thrombosis in up to 3.4% of poor-grade aneurysmal subarachnoid hemorrhage patients receiving pharmacologic thromboprophylaxis.¹⁰⁸

In a retrospective study of 170 patients, our group showed that giving drugs to prevent venous thromboembolism (unfractionated heparin 5,000 IU subcutaneously every 12 hours or dalteparin 5,000 IU subcutaneously daily), starting within 24 hours of aneurysm treatment, could be safe.¹⁰⁹ Fifty-eight percent of these patients had an external ventricular drain in place. One patient developed a major cerebral hemorrhagic complication and died while on unfractionated heparin; however, the patient was also on dual antiplatelet therapy with aspirin and clopidogrel.¹⁰⁹

Current guidelines suggest that intermittent compression devices be applied in all patients before aneurysm treatment. Pharmacologic thromboprophylaxis with a heparinoid can be started 12 to 24 hours after aneurysm treatment.^3,109

A TEAM APPROACH

Patients with subarachnoid hemorrhage need integrated care from different medical and nursing specialties. The best outcomes are achieved by systems that can focus as a team on the collective goal of quick intervention to secure the aneurysm, followed by measures to minimize secondary brain injury.

The modern concept of cerebral monitoring in a setting of subarachnoid hemorrhage should focus on brain perfusion rather than vascular diameter. Although the search continues for new diagnostic, prognostic, and therapeutic tools, there is no “silver bullet” that will help all patients. Instead, it is the systematic integration and application of many small advances that will ultimately lead to better outcomes.

ACKNOWLEDGMENT

This work was supported by research funding provided by the Bitove Foundation, which has been supportive of our clinical and research work for several years.

Aneurysmal subarachnoid hemorrhage is a devastating condition, with an estimated death rate of 30% during the initial episode.^1,2 Approximately the same number of patients survive but leave the hospital with disabling neurologic deficits.³

However, better outcomes can be achieved by systems that are able to work as a team on the collective goal of quick intervention to secure the ruptured aneurysm, followed by the implementation of measures to minimize secondary brain injury. Although the search for new diagnostic, prognostic, and therapeutic modalities continues, it is clear that there exists no “silver bullet” that will help all patients. Instead, it is the systematic integration and application of small advances that will ultimately maximize the patient’s chances of survival and neurologic recovery.

This review focuses on the management of aneurysmal subarachnoid hemorrhage and its systemic and neurologic complications.

ANEURYSM IS THE MOST COMMON CAUSE OF SUBARACHNOID BLEEDING

Aneurysmal subarachnoid hemorrhage, ie, rupture of an intracranial aneurysm, flooding  the subarachnoid space with blood, affects about 24,000 Americans each year.^1,2 A ruptured aneurysm is the most common cause of subarachnoid hemorrhage, accounting for about 85% of cases. Less common causes include idiopathic benign perimesencephalic hemorrhage, arteriovenous malformation, dural arteriovenous fistula, and hemorrhagic mycotic aneurysm. These have their own natural history, pathophysiology, and specific treatment, and will not be addressed in this article.

Risk factors for aneurysmal subarachnoid hemorrhage include having a first-degree relative who had the disease, hypertension, smoking, and consuming more than 150 g of alcohol per week.⁴

CLINICAL PRESENTATION AND DIAGNOSIS

The key symptom of aneurysmal subarachnoid hemorrhage is the abrupt onset of severe headache that peaks in intensity over 1 hour,⁵ often described as “the worst headache of my life.” Headache is accompanied by brief loss of consciousness in 53% of cases (conversely, nearly half of patients maintain normal mental status), by nausea or vomiting in 77%, and by meningismus (neck pain or stiffness) in 35%.⁶

These clinical manifestations and risk factors have been incorporated into a decision rule:

Obtain brain imaging if the patient has acute headache reaching maximal intensity within 1 hour, associated with any of the following factors:

• Age 40 or older

• Neck pain or stiffness

• Witnessed loss of consciousness

• Onset during exertion

• “Thunderclap” headache (ie, instantly peaking pain)

• Limited neck flexion on examination.⁵

This decision rule has nearly 100% sensitivity for aneurysmal subarachnoid hemorrhage in clinical practice.⁵ All patients require brain imaging if they have a severe headache plus either abnormal neurologic findings (eg, a focal neurologic deficit) or a history of cerebral aneurysm.

Emergency physicians should have a low threshold for ordering noncontrast computed tomography (CT) of the head in patients with even mild symptoms suggesting aneurysmal subarachnoid hemorrhage. Failure to order CT is the most common diagnostic error in this situation.⁶ CT performed within 6 hours of headache onset is nearly 100% sensitive for this condition,⁷ but the sensitivity falls to 93% after the first 24 hours and to less than 60% after 5 days.⁸ In patients who have symptoms highly suggestive of aneurysmal subarachnoid hemorrhage but a normal CT, lumbar puncture is the next diagnostic step.

There are two alternatives to CT followed by lumbar puncture: ie, noncontrast CT followed by CT angiography,^9,10 and magnetic resonance imaging followed by magnetic resonance angiography. In patients with suspicious clinical symptoms but negative CT results, CT followed by CT angiography can rule out aneurysmal subarachnoid hemorrhage with a 99% probability.^9,10 However, CT followed by lumbar puncture remains the standard of care and carries a class I recommendation in the American Heart Association guidelines for ruling out subarachnoid hemorrhage.⁵

GRADING THE SEVERITY OF SUBARACHNOID HEMORRHAGE

Age, the thickness of the blood layer in the subarachnoid space, intraventricular hemorrhage and the findings of the neurologic examination at presentation are predictors of long-term outcomes in aneurysmal subarachnoid hemorrhage (Figure 1).

Different grading systems used in clinical practice are based on the findings on the initial neurologic examination and on the initial noncontrast CT (ie, the thickness of the blood, and whether intraventricular hemorrhage is present). Among the most widely used are those developed by Hunt and Hess¹² and by the World Federation of Neurological Surgeons¹¹ (WFNS), and the CT grading scales (Fisher¹³ or its modified version¹⁴)  (Tables 1 and 2). With either the Hunt and Hess scale or the WFNS scale, the higher the score, the worse the patient’s probable outcome. Scores on both Fisher scales correlate with the risk of angiographic vasospasm. The higher the grade, the higher the risk of angiographic vasospasm.

The VASOGRADE score—a combination of the WFNS score and the modified Fisher scale—stratifies patients at risk of delayed cerebral ischemia, allowing for a tailored monitoring strategy.¹⁵ There are three variations:

• VASOGRADE green—Modified Fisher 1 or 2 and WFNS 1 or 2

• VASOGRADE yellow—Modified Fisher 3 or 4 and WFNS 1, 2, or 3

• VASOGRADE red—WFNS 4 or 5. 

After the initial bleeding event, patients with aneurysmal subarachnoid hemorrhage are at high risk of delayed systemic and neurologic complications, with poor functional outcomes. Delayed cerebral ischemia holds the greatest risk of an unfavorable outcome and ultimately can lead to cerebral infarction, disability, and death.^6,7

INITIAL MANAGEMENT

After aneurysmal subarachnoid hemorrhage is diagnosed, the initial management (Figure 2) includes appropriate medical prevention of rebleeding (which includes supportive care, blood pressure management, and, perhaps, the early use of a short course of an antifibrinolytic drug) and early transfer to a high-volume center for securing the aneurysm. The reported incidence of rebleeding varies from 5% to 22% in the first 72 hours. “Ultra-early” rebleeding (within 24 hours of hemorrhage) has been reported, with an incidence as high as 15% and a fatality rate around 70%. Patients with poor-grade aneurysmal subarachnoid hemorrhage, larger aneurysms, and “sentinel bleeds” are at higher risk of rebleeding.16

Outcomes are much better when patients are managed in a high-volume center, with a specialized neurointensive care unit¹⁷ and access to an interdisciplinary team.¹⁸ Regardless of the initial grade, patients with aneurysmal subarachnoid hemorrhage should be quickly transferred to a high-volume center, defined as one treating at least 35 cases per year, and the benefit is greater in centers treating more than 60 cases per year.¹⁹ The higher the caseload in any given hospital, the better the clinical outcomes in this population.²⁰

Treating cerebral aneurysm: Clipping or coiling

Early aneurysm repair is generally considered the standard of care and the best strategy to reduce the risk of rebleeding. Further, early treatment may be associated with a lower risk of delayed cerebral ischemia²¹ and better outcomes.²²

Three randomized clinical trials have compared surgical clipping and endovascular repair (placement of small metal coils within the aneurysm to promote clotting).

The International Subarachnoid Aneurysm Trial²³ showed a reduction of 23% in relative risk and of 7% in absolute risk in patients who underwent endovascular treatment compared with surgery. The survival benefit persisted at a mean of 9 years (range 6–14 years), but with a higher annual rate of aneurysm recurrence in the coiling group (2.9% vs 0.9%).²⁴ Of note, this trial included only patients with aneurysms deemed suitable for both coiling and clipping, so that the exclusion rate was high. Most of the patients presented with good-grade (WFNS score 1–3), small aneurysms (< 5 mm) in the anterior circulation.

A single-center Finnish study²⁵ found no differences in rates of recovery, disability, and  death at 1 year, comparing surgery and endovascular treatment. Additionally, survival rates at a mean follow-up of 39 months were similar, with no late recurrences or aneurysmal bleeding.

Lastly, the Barrow Ruptured Aneurysm Trial26,27 found that patients assigned to endovascular treatment had better 1-year neurologic outcomes, defined as a modified Rankin score of 2 or less. Importantly, 37.7% of patients originally assigned to endovascular treatment crossed over to surgical treatment. The authors then performed intention-to-treat and as-treated analyses. Either way, patients treated by endovascular means had better neurologic outcomes at 1 year. However, no difference in the relative risk reduction in worse outcome was found on 3-year follow-up, and patients treated surgically had higher rates of aneurysm obliteration and required less aneurysm retreatment, both of which were statistically significant.

The question that remains is not whether to clip or whether to coil, but whom to clip and whom to coil.²⁸ That question must be answered on a patient-to-patient basis and requires the expertise of an interventional neuroradiologist and a vascular neurosurgeon—one of the reasons these patients are best cared for in high-volume centers providing such expertise.

MEDICAL PREVENTION OF REBLEEDING

Blood pressure management

There are no systematic data on the optimal blood pressure before securing an aneurysm. Early studies of hemodynamic augmentation in cases of ruptured untreated aneurysm reported rebleeding when the systolic blood pressure was allowed to rise above 160 mm Hg.^29,30 A recent study evaluating hypertensive intracerebral hemorrhage revealed better functional outcomes with intensive lowering of blood pressure (defined as systolic blood pressure < 140 mm Hg) but no significant reduction in the combined rate of death or severe disability.³¹ It is difficult to know if these results can be extrapolated to patients with aneurysmal subarachnoid hemorrhage. Current guidelines^3,32 say that before the aneurysm is treated, the systolic pressure should be lower than 160 mm Hg.

There is no specific drug of choice, but a short-acting, titratable medication is preferable. Nicardipine is a very good option, and labetalol might be an appropriate alternative.³³ Once the aneurysm is secured, all antihypertensive drugs should be held. Hypertension should not be treated unless the patient has clinical signs of a hypertensive crisis, such as flash pulmonary edema, myocardial infarction, or hypertensive encephalopathy.

Antifibrinolytic therapy

Risk factors: Family history, hypertension, smoking, heavy drinking

The role of antifibrinolytic therapy in aneurysmal subarachnoid hemorrhage is controversial and has been studied in 10 clinical trials. In a Swedish study,³⁴ early use of tranexamic acid (1 g intravenously over 10 minutes followed by 1 g every 6 hours for a maximum of 24 hours) reduced the rebleeding rate substantially, from 10.8% to 2.4%, with an 80% reduction in the mortality rate from ultra-early rebleeding. However, a recent Cochrane review that included this study found no overall benefit.³⁵

An ongoing multicenter randomized trial in the Netherlands will, we hope, answer this question in the near future.³⁶ At present, some centers would consider a short course of tranexamic acid before aneurysm treatment.

DIAGNOSIS AND TREATMENT OF COMPLICATIONS

Medical complications are extremely common after aneurysmal subarachnoid hemorrhage. Between 75% and 100% of patients develop some type of systemic or further neurologic derangement, which in turn has a negative impact on the long-term outcome.^37,38 In the first 72 hours, rebleeding is the most feared complication, and as mentioned previously, appropriate blood pressure management and early securing of the aneurysm minimize its risk.

NEUROLOGIC COMPLICATIONS

Hydrocephalus

Hydrocephalus is the most common early neurologic complication after aneurysmal subarachnoid hemorrhage, with an overall incidence of 50%.³⁹ Many patients with poor-grade aneurysmal subarachnoid hemorrhage and patients whose condition deteriorates due to worsening of hydrocephalus require the insertion of an external ventricular drain (Figure 1).

Up to 30% of patients who have a poor-grade aneurysmal subarachnoid hemorrhage improve neurologically with cerebrospinal fluid drainage.⁴⁰ An external ventricular drain can be safely placed, even before aneurysm treatment, and placement does not appear to increase the risk of rebleeding.^39,41 After placement, rapid weaning from the drain (clamping within 24 hours of insertion) is safe, decreases length of stay in the intensive care unit and hospital, and may be more cost-effective than gradual weaning over 96 hours.⁴²

Increased intracranial pressure

Intracranial hypertension is another potential early complication, and is frequently due to the development of hydrocephalus, cerebral edema, or rebleeding. The treatment of increased intracranial pressure does not differ from the approach used in managing severe traumatic brain injury, which includes elevating the head of the bed, sedation, analgesia, normoventilation, and cerebrospinal fluid drainage.

Hypertonic saline has been tested in several studies that were very small but nevertheless consistently showed control of intracranial pressure levels and improvement in cerebral blood flow measured by xenon CT.^43–47 Two of these studies even showed better outcomes at discharge.^43,44 However, the small number of patients prevents any meaningful conclusion regarding the use of hypertonic saline and functional outcomes.

Outcomes are much better when patients are managed in a high-volume center

Barbiturates, hypothermia, and decompressive craniectomy could be tried in refractory cases.⁴⁸ Seule et al⁴⁹ evaluated the role of therapeutic hypothermia with or without barbiturate coma in 100 patients with refractory intracranial hypertension. Only 13 patients received hypothermia by itself. At 1 year, 32 patients had achieved a good functional outcome (Glasgow Outcome Scale score 4 or 5). The remaining patients were severely disabled or had died. Of interest, the median duration of hypothermia was 7 days, and 93% of patients developed some medical complication such as electrolyte disorders (77%), pneumonia (52%), thrombocytopenia (47%), or septic shock syndrome (40%). Six patients died as a consequence of one of these complications.

Decompressive craniectomy can be life-saving in patients with refractory intracranial hypertension. However, most of these patients will die or remain severely disabled or comatose.⁵⁰

Seizure prophylaxis is controversial

Seizures can occur at the onset of intracranial hemorrhage, perioperatively, or later (ie, after the first week). The incidence varied considerably in different reports, ranging from 4% to 26%.⁵¹ Seizures occurring perioperatively, ie, after hospital admission, are less frequent and are usually the manifestation of aneurysm rebleeding.²⁴

The question is not whether to clip or coil, but whom to clip and whom to coil

Seizure prophylaxis remains controversial, especially because the use of phenytoin is associated with increased incidence of cerebral vasospasm, infarction, and worse cognitive outcomes after aneurysmal subarachnoid hemorrhage.^52,53 Therefore, routine prophylactic use of phenytoin is not recommended in these patients,³ although the effect of other antiepileptic drugs is less studied and less clear. Patients may be considered for this therapy if they have multiple risk factors for seizures, such as intraparenchymal hematoma, advanced age (> 65), middle cerebral artery aneurysm, craniotomy for aneurysm clipping, and a short course (≤ 72 hours) of an antiepileptic drug other than phenytoin, especially while the aneurysm is unsecured.³

Levetiracetam may be an alternative to phenytoin, having better pharmacodynamic and kinetic profiles, minimal protein binding, and absence of hepatic metabolism, resulting in a very low risk of drug interaction and better tolerability.^54,55 Because of these advantages, levetiracetam has become the drug of choice in several centers treating aneurysmal subarachnoid hemorrhage in the United States.

Addressing this question, a survey was sent to 25 high-volume aneurysmal subarachnoid hemorrhage academic centers in the United States. All 25 institutions answered the survey, and interestingly, levetiracetam was the first-line agent for 16 (94%) of the 17 responders that used prophylaxis, while only 1 used phenytoin as the agent of choice.⁵⁶

A retrospective cohort study by Murphy-Human et al⁵⁷ showed that a short course of levetiracetam (≤ 72 hours) was associated with higher rates of in-hospital seizures compared with an extended course of phenytoin (eg, entire hospital stay). However, the study did not address functional outcomes.⁵⁷

Continuous electroencephalographic monitoring may be considered in comatose patients, in patients requiring controlled ventilation and sedation, or in patients with unexplained alteration in consciousness. In one series of patients with aneurysmal subarachnoid hemorrhage who received continuous monitoring, the incidence of nonconvulsive status epilepticus was 19%, with an associated mortality rate of 100%.⁵⁸

Continuous quantitative electroencephalography is useful to monitor and to detect angiographic vasospasm and delayed cerebral ischemia. Relative alpha variability and the alpha-delta ratio decrease with ischemia, and this effect can precede angiographic vasospasm by 3 days.^59,60

Delayed cerebral ischemia

Delayed cerebral ischemia is defined as the occurrence of focal neurologic impairment, or a decrease of at least 2 points on the Glasgow Coma Scale that lasts for at least 1 hour, is not apparent immediately after aneurysm occlusion, and not attributable to other causes (eg, hyponatremia, fever).⁶¹

Classically, neurologic deficits that occurred within 2 weeks of aneurysm rupture were ascribed to reduced cerebral blood flow caused by delayed large-vessel vasospasm causing cerebral ischemia.⁶² However, perfusion abnormalities have also been observed with either mild or no demonstrable vasospasm.⁶³ Almost 70% of patients who survive the initial hemorrhage develop some degree of angiographic vasospasm. However, only 30% of those patients will experience symptoms.

In addition to vasospasm of large cerebral arteries, impaired autoregulation and early brain injury within the first 72 hours following subarachnoid hemorrhage may play important roles in the development of delayed cerebral ischemia.⁶⁴ Therefore, the modern concept of delayed cerebral ischemia monitoring should focus on cerebral perfusion rather than vessel diameter measurements. This underscores the importance of comprehensive, standardized monitoring techniques that provide information not only on microvasculature, but also at the level of the microcirculation, with information on perfusion, oxygen utilization and extraction, and autoregulation.

Although transcranial Doppler has been the most commonly applied tool to monitor for angiographic vasospasm, it has a low sensitivity and negative predictive value.³⁷ It is nevertheless a useful technique to monitor good-grade aneurysmal subarachnoid hemorrhage patients (WFNS score 1–3) combined with frequent neurologic examinations (Figure 3).

CT angiography is a good noninvasive alternative to digital subtraction angiography. However, it tends to overestimate the degree of vasoconstriction and does not provide information about perfusion and autoregulation.⁶⁵ Nevertheless, CT angiography combined with a CT perfusion scan can add information about autoregulation and cerebral perfusion and has been shown to be more sensitive for the diagnosis of angiographic vasospasm than transcranial Doppler and digital subtraction angiography (Figure 4).

Patients with a poor clinical condition (WFNS score 4 or 5) or receiving continuous sedation constitute a challenge in monitoring for delayed neurologic deterioration. Neurologic examination is not sensitive enough in this setting to detect subtle changes. In these specific and challenging circumstances, multimodality neuromonitoring may be useful in the early detection of delayed cerebral ischemia and may help guide therapy.⁶⁷

Several noninvasive and invasive techniques have been studied to monitor patients at risk of delayed cerebral ischemia after subarachnoid hemorrhage.⁶⁶ These include continuous electroencephalography, brain tissue oxygenation monitoring (Ptio₂), cerebral microdialysis, thermal diffusion flowmetry, and near-infrared spectroscopy. Of these techniques, Ptio₂, cerebral microdialysis, and continuous electroencephalography (see discussion of seizure prophylaxis above) have been more extensively studied. However, most of the studies were observational and very small, limiting any recommendations for using these techniques in routine clinical practice.⁶⁸

Ptio₂ is measured by inserting an intraparenchymal oxygen-sensitive microelectrode, and microdialysis requires a microcatheter with a semipermeable membrane that allows small soluble substances to cross it into the dialysate. These substances, which include markers of ischemia (ie, glucose, lactate, and pyruvate), excitotoxins (ie, glutamate and aspartate), and membrane cell damage products (ie, glycerol), can be measured. Low Ptio₂ values (< 15 mm Hg) and abnormal mycrodialysate findings (eg, glucose < 0.8 mmol/L, lactate-to-pyruvate ratio > 40) have both been associated with cerebral ischemic events and poor outcome.⁶⁸

Preventing delayed cerebral ischemia

Oral nimodipine 60 mg every 4 hours for 21 days, started on admission, carries a class I, level of evidence A recommendation in the management of aneurysmal subarachnoid hemorrhage.^3,32,69 It improves clinical outcome despite having no effect on the risk of angiographic vasospasm. The mechanism of improved outcome is unclear, but the effect may be a neuroprotective phenomenon limiting the extension of delayed cerebral ischemia.⁷⁰

If hypotension occurs, the dose can be lowered to 30 mg every 2 hours. Whether to discontinue nimodipine in this situation is controversial. Of note, the clinical benefits of nimodipine have not been replicated with other calcium channel blockers (eg, nicardipine).⁷¹

Prophylactic hyperdynamic fluid therapy, known as “triple-H” (hypervolemia, hemodilution, and hypertension) was for years the mainstay of treatment in preventing delayed cerebral ischemia due to vasospasm. However, the clinical data supporting this intervention have been called into question, as analysis of two trials found that hypervolemia did not improve outcomes or reduce the incidence of delayed cerebral ischemia, and in fact increased the rate of complications.^72,73 Based on these findings, current guidelines recommend maintaining euvolemia rather than prophylactic hypervolemia in patients with aneurysmal subarachnoid hemorrhage.^3,32,69

TREATING DELAYED CEREBRAL ISCHEMIA

Hemodynamic augmentation

In patients with neurologic deterioration due to delayed cerebral ischemia, hemodynamic augmentation is the cornerstone of treatment. This is done according to a protocol, started early, involving specific physiologic goals, clinical improvement, and escalation to invasive therapies in a timely fashion in patients at high risk of further neurologic insult (Figure 5).

The physiologic goal is to increase the delivery of oxygen and glucose to the ischemic brain. Hypertension seems to be the most effective component of hemodynamic augmentation regardless of volume status, increasing cerebral blood flow and brain tissue oxygenation, with reversal of delayed cerebral ischemic symptoms in up to two-thirds of treated patients.^74,75 However, this information comes from very small studies, with no randomized trials of induced hypertension available.

The effect of a normal saline fluid bolus in patients suspected of having delayed cerebral ischemia has been shown to increase cerebral blood flow in areas of cerebral ischemia.⁷⁴ If volume augmentation fails to improve the neurologic status, the next step is to artificially induce hypertension using vasopressors. The blood pressure target should be based on clinical improvement. A stepwise approach is reasonable in this situation, and the lowest level of blood pressure at which there is a complete reversal of the new focal neurologic deficit should be maintained.^3,29

Inotropic agents such as dobutamine or milrinone can be considered as alternatives in patients who have new neurologic deficits that are refractory to fluid boluses and vasopressors, or in a setting of subarachnoid hemorrhage-induced cardiomyopathy.^76,77

Once the neurologic deficit is reversed by hemodynamic augmentation, the blood pressure should be maintained for 48 to 72 hours at the level that reversed the deficit completely, carefully reassessed thereafter, and the patient weaned slowly. Unruptured unsecured aneurysms should not prevent blood pressure augmentation in a setting of delayed cerebral ischemia if the culprit aneurysm is treated.^3,32 If the ruptured aneurysm has not been secured, careful blood pressure augmentation can be attempted, keeping in mind that hypertension (> 160/95 mm Hg) is a risk factor for fatal aneurysm rupture.

Endovascular management of delayed cerebral ischemia

When medical augmentation fails to completely reverse the neurologic deficits, endovascular treatment can be considered. Although patients treated early in the course of delayed cerebral ischemia have better neurologic recovery, prophylactic endovascular treatment in asymptomatic patients, even if angiographic signs of spasm are present, does not improve clinical outcomes and carries the risk of fatal arterial rupture.⁷⁸

SYSTEMIC COMPLICATIONS

Hyponatremia and hypovolemia

Aneurysmal subarachnoid hemorrhage is commonly associated with abnormalities of fluid balance and electrolyte derangements. Hyponatremia (serum sodium < 135 mmol/L) occurs in 30% to 50% of patients, while the rate of hypovolemia (decreased circulating blood volume) ranges from 17% to 30%.⁷⁹ Both can negatively affect long-term outcomes.^80,81

Decreased circulating blood volume is a well-described contributor to delayed cerebral ischemia and cerebral infarction after aneurysmal subarachnoid hemorrhage.^80–82 Clinical variables such as heart rate, blood pressure, fluid balance, and serum sodium concentration are usually the cornerstones of intravascular volume status assessment. However, these variables correlate poorly with measured circulating blood volumes in those with aneurysmal subarachnoid hemorrhage.^83,84

The mechanisms responsible for the development of hyponatremia and hypovolemia after aneurysmal subarachnoid hemorrhage are not completely understood. Several factors have been described and may contribute to the increased natriuresis and, hence, to a reduction in circulating blood volume: increased circulating natriuretic peptide concentrations,^85–87 sympathetic nervous system hyperactivation,⁸⁸ and hyperreninemic hypo-
aldosteronism syndrome.^89,90

Guidelines: Before treating the aneurysm, the systolic pressure should be < 160 mm Hg

Lastly, the cerebral salt wasting syndrome, described in the 1950s,⁹¹ was thought to be a key mechanism in the development of hyponatremia and hypovolemia after aneurysmal subarachnoid hemorrhage. In contrast to the syndrome of inappropriate antidiuretic hormone, which is characterized by hyponatremia with a normal or slightly elevated intravascular volume, the characteristic feature of cerebral salt wasting syndrome is the development of hyponatremia in a setting of intravascular volume depletion.⁹² In critically ill neurologic and neurosurgical patients, this differential diagnosis is very difficult, especially in those with aneurysmal subarachnoid hemorrhage in whom the clinical assessment of fluid status is not reliable. These two syndromes might coexist and contribute to the development of hyponatremia after aneurysmal subarachnoid hemorrhage.^92,93

Hoff et al^83,84 prospectively compared the clinical assessment of fluid status by critical and intermediate care nurses and direct measurements of blood volume using pulse dye densitometry. The clinical assessment failed to accurately assess patients’ volume status. Using the same technique to measure circulating blood volume, this group showed that calculation of fluid balance does not provide adequate assessment of fluid status.^83,84

Hemodynamic monitoring tools can help guide fluid replacement in this population. Mutoh et al⁹⁴ randomized 160 patients within 24 hours of hemorrhage to receive early goal-directed fluid therapy (ie, preload volume and cardiac output monitored by transpulmonary thermodilution) vs standard therapy (ie, fluid balance or central venous pressure). Overall, no difference was found in the rates of delayed cerebral ischemia (33% vs 42%; P = .33) or favorable outcome (67% vs 57%; P = .22). However, in the subgroup of poor-grade patients (WFNS score 4 or 5), early goal-directed therapy was associated with a lower rate of delayed cerebral ischemia (5% vs 14%; P = .036) and with better functional outcomes at 3 months (52% vs 36%; P = .026).

Fluid restriction to treat hyponatremia in aneurysmal subarachnoid hemorrhage is no longer recommended because of the increased risk of cerebral infarction due to hypovolemic hypoperfusion.⁸²

Prophylactic use of mineralocorticoids (eg, fludrocortisone, hydrocortisone) has been shown to limit natriuresis, hyponatremia, and the amount of fluid required to maintain euvolemia.^95,96 Higher rates of hypokalemia and hyperglycemia, which can be easily treated, are the most common complications associated with this approach. Additionally, hypertonic saline (eg, 3% saline) can be used to correct hyponatremia in a setting of aneurysmal subarachnoid hemorrhage.⁷⁹

Cardiac complications

Cardiac complications after subarachnoid hemorrhage are most likely related to sympathetic hyperactivity and catecholamine-induced myocyte dysfunction. The pathophysiology is complex, but cardiac complications have a significant negative impact on long-term outcome in these patients.⁹⁷

Electrocardiographic changes and positive cardiac enzymes associated with aneurysmal subarachnoid hemorrhage have been extensively reported. More recently, data from studies of two-dimensional echocardiography have shown that subarachnoid hemorrhage can also be associated with significant wall-motion abnormalities and even overt cardiogenic shock.^98–100

There is no specific curative therapy; the treatment is mainly supportive. Vasopressors and inotropes may be used for hemodynamic augmentation.

Pulmonary complications

Pulmonary complications occur in 20% to 30% of all aneurysmal subarachnoid hemorrhage patients and are associated with a higher risk of delayed cerebral ischemia and death. Common pulmonary complications in this population are mild acute respiratory distress syndrome (27%), hospital-acquired pneumonia (9%), cardiogenic pulmonary edema (8%), aspiration pneumonia (6%), neurogenic pulmonary edema (2%), and pulmonary embolism (1%).^101–103

SUPPORTIVE CARE

Hyperthermia, hyperglycemia, and liberal use of transfusions have all been associated with longer stays in the intensive care unit and hospital, poorer neurologic outcomes, and higher mortality rates in patients with acute brain injury.¹⁰⁴ Noninfectious fever is the most common systemic complication after subarachnoid hemorrhage.

Antipyretic drugs such as acetaminophen and ibuprofen are not very effective in reducing fever in the subarachnoid hemorrhage population, but should still be used as first-line therapy. The use of surface and intravascular devices can be considered when fevers do not respond to nonsteroidal anti-inflammatory drugs.

Fluid restriction to treat hyponatremia in aneurysmal subarachnoid hemorrhage is no longer recommended

Although no prospective randomized trial has addressed the impact of induced normothermia on long-term outcome and mortality in aneurysmal subarachnoid hemorrhage patients, fever control has been shown to reduce cerebral metabolic distress, irrespective of intracranial pressure.¹⁰⁵ Maintenance of normothermia (< 37.5°C) seems reasonable, especially in aneurysmal subarachnoid hemorrhage patients at risk of or with active delayed cerebral ischemia.¹⁰⁶

Current guidelines^3,32,69 strongly recommend avoiding hypoglycemia, defined as a serum glucose level less than 80 mg/dL, but suggest keeping the blood sugar level below 180 or 200 mg/dL.

At the moment, there is no clear threshold for transfusion in patients with aneurysmal subarachnoid hemorrhage. Current guidelines suggest keeping hemoglobin levels between 8 and 10 g/dL.³

Preventing venous thromboembolism

The incidence of venous thromboembolism after aneurysmal subarachnoid hemorrhage varies widely, from 1.5% to 18%.¹⁰⁷ Active surveillance with venous Doppler ultrasonography has found asymptomatic deep vein thrombosis in up to 3.4% of poor-grade aneurysmal subarachnoid hemorrhage patients receiving pharmacologic thromboprophylaxis.¹⁰⁸

In a retrospective study of 170 patients, our group showed that giving drugs to prevent venous thromboembolism (unfractionated heparin 5,000 IU subcutaneously every 12 hours or dalteparin 5,000 IU subcutaneously daily), starting within 24 hours of aneurysm treatment, could be safe.¹⁰⁹ Fifty-eight percent of these patients had an external ventricular drain in place. One patient developed a major cerebral hemorrhagic complication and died while on unfractionated heparin; however, the patient was also on dual antiplatelet therapy with aspirin and clopidogrel.¹⁰⁹

Current guidelines suggest that intermittent compression devices be applied in all patients before aneurysm treatment. Pharmacologic thromboprophylaxis with a heparinoid can be started 12 to 24 hours after aneurysm treatment.^3,109

A TEAM APPROACH

Patients with subarachnoid hemorrhage need integrated care from different medical and nursing specialties. The best outcomes are achieved by systems that can focus as a team on the collective goal of quick intervention to secure the aneurysm, followed by measures to minimize secondary brain injury.

The modern concept of cerebral monitoring in a setting of subarachnoid hemorrhage should focus on brain perfusion rather than vascular diameter. Although the search continues for new diagnostic, prognostic, and therapeutic tools, there is no “silver bullet” that will help all patients. Instead, it is the systematic integration and application of many small advances that will ultimately lead to better outcomes.

ACKNOWLEDGMENT

This work was supported by research funding provided by the Bitove Foundation, which has been supportive of our clinical and research work for several years.

Can patients with atrial fibrillation  undergo a percutaneous procedure to reduce their risk of stroke, thereby eliminating the need for lifelong treatment with an oral anticoagulant drug? The data are preliminary, but this is an emerging option that physicians should be aware of.

We review here the current evidence and techniques aimed at isolating the left atrial appendage to prevent stroke, and we emphasize the need for continued systematic comparisons between oral anticoagulation and percutaneous treatment options.

NOVEL TREATMENTS ARE NEEDED

Atrial fibrillation is the most common cardiac arrhythmia,¹ affecting an estimated 1% to 2% of people worldwide. In 2001, an estimated 2.3 million persons in the United States had atrial fibrillation, and that number is expected to more than double by 2050.²

Atrial fibrillation independently increases the risk of stroke by a factor of 4 to 5.³ The American Heart Association ranks stroke as the fourth most common cause of death and the leading cause of disability in the United States.⁴ Atrial fibrillation accounts for 15% of strokes in people of all ages and 30% in those over age 80.⁵ Untreated, 2% to 5% of patients with atrial fibrillation suffer a stroke in any given year.⁶ Most of these strokes are cardioembolic, with thrombi originating in the left atrial appendage.⁷ Furthermore, it has been estimated^8,9 that patients with atrial fibrillation who have already had a stroke and cannot tolerate oral anticoagulants have an annual risk of stroke close to 12% and a relative risk of approximately 3.0 compared with those with atrial fibrillation and prior stroke who can tolerate anticoagulation.

Oral anticoagulation effectively prevents thromboembolic events associated with atrial fibrillation,¹⁰ but several factors limit its efficacy and applicability. The risk of bleeding complications, the need for frequent monitoring, and challenges with compliance create a large population of patients who would benefit from alternative approaches. Consequently, physicians have looked for other ways to prevent stroke—especially surgical and transcatheter procedures—that are not associated with an ongoing risk of hemorrhage and a lifelong need to take an anticoagulant.

THE LEFT ATRIAL APPENDAGE: A SITE OF CLOT FORMATION

The left atrial appendage is the most common site of thrombus formation, particularly in patients with nonvalvular atrial fibrillation. Nearly 90% of thrombi discovered in the left atrium form in the appendage.⁷ A study of 233 patients not on long-term anticoagulation revealed that after 48 hours of atrial fibrillation, 15% had a left atrial thrombus, and all but one of the thrombi were in the appendage.¹¹

Atrial fibrillation increases the risk of stroke by a factor of 4 to 5

Believed to function as a decompression chamber during left ventricular systole, the left atrial appendage is embryologically derived from the left wall of the primary atrium. It is in close proximity to the free wall of the left ventricle, and therefore its flow can vary with left ventricular function. Relative stasis due to its location and extensive trabeculations, especially in times of poor forward flow, make it a high-risk site for clot formation.¹²

ANTICOAGULATION: EFFECTIVE BUT IMPERFECT

In deciding whether a patient with atrial fibrillation should be prescribed anticoagulation therapy, the physician must balance the risk of stroke against the risk of bleeding. Several tools for assessing these two risks have been developed. Of note, some of the risk factors for stroke are the same as some of the risk factors for bleeding.

Calculating the risk of stroke

CHADS₂ and CHA2DS₂-VASc are the two most commonly used tools for assessing the risk of stroke, but only the newer CHA₂DS₂-VASc has received a class I recommendation (the highest) from the European Society of Cardiology (ESC).¹³

CHADS₂ risk factors are Congestive heart failure (1 point), Hypertension (1 point), Age 75 or older (1 point), Diabetes (1 point), and  Stroke or transient ischemic attack (2 points). Risk of stroke is considered low with a score of 0, intermediate with a score of 1, and high with a score of 2 or more. 

CHA₂DS₂-VASc risk factors are Congestive heart failure or left ventricular ejection fraction ≤ 40% (1 point), Hypertension (1 point), Age ≥ 75 (2 points), Age 65–74 (1 point), Diabetes mellitus (1 point), Stroke, transient ischemic attack, or thromboembolism (2 points), Vascular disease (1 point), and female Sex (1 point). Low risk is defined as a score of 0 for a man or, for a woman with no other risk factors, a score of 1. A score of 1 for a man indicates moderate risk, and a score of 2 or more is high risk.

Calculating the risk of bleeding

Tools for assessing bleeding risk include ATRIA2 and HAS-BLED,¹⁴ the latter carrying a class I recommendation from the ESC.¹³

HAS-BLED risk factors are Hypertension (1 point), Abnormal renal or liver function (1 point each), Stroke (1 point), Bleeding (1 point), Labile international normalized ratio (INR) (1 point), Elderly (age > 65) (1 point), and Drug or alcohol use (1 point each). The risk of bleeding is considered high with a score of 3 or higher.

Disadvantages of oral anticoagulation

Oral anticoagulation is the standard treatment for preventing stroke in patients with atrial fibrillation, and the vitamin K antagonist warfarin remains the foundation.

Though highly effective, warfarin requires close monitoring and frequent dose adjustments because of its numerous food and drug interactions. Bleeding risk and the challenge of frequent monitoring rule out treatment with warfarin in 14% to 44% of patients with atrial fibrillation.¹⁵ Even in “ideal” candidates, warfarin is underused, with one study reporting that only 38% of those with clinical indications for it had been prescribed warfarin, and of those for whom it had not been prescribed, 63% were also not taking aspirin.¹⁶ Moreover, a meta-analysis suggested that the average patient treated with warfarin has his or her INR in the therapeutic range only about 55% of the time.¹⁷

Newer, target-specific oral anticoagulants such as dabigatran (a direct thrombin inhibitor) and rivaroxaban and apixaban (both factor Xa inhibitors) do not require monitoring and have fewer drug interactions. But like warfarin, they also confer a risk of serious bleeding.^18–20 Most of the studies of these newer drugs have compared them with warfarin, with the preponderance of evidence showing them to be either noninferior or superior to warfarin for stroke reduction. But bleeding complication rates remain significant, apixaban having lower rates of major bleeding than dabigatran and rivaroxaban.

Untreated, 2%–5% of patients with atrial fibrillation will suffer a stroke in any given year

In a meta-analysis, Ruff et al²¹ concluded that the target-specific oral anticoagulants provide a favorable balance of risk and benefit. Compared with warfarin, these new drugs reduced the rate of stroke or systemic embolic events by 19%. There was also a significant reduction in rates of intracranial hemorrhage and all-cause mortality. The risk of major bleeding was similar to that with warfarin, though there was a higher risk of gastrointestinal bleeding with target-specific agents. These effects were consistent across a wide range of patients.

Given the difficulties, risks, and serious side effects of anticoagulant therapy, many patients stop taking these drugs soon after starting them, either on their own or on their physician’s recommendation. In the RE-LY trial (Dabigatran vs Warfarin in Patients With Atrial Fibrillation), 10% of patients receiving dabigatran and 17% of those receiving warfarin stopped the treatment within 1 to 2 years.²² In a similar trial of rivaroxaban vs warfarin in nonvalvular atrial fibrillation (the ROCKET-AF trial), 24% of those treated with rivaroxaban and 22% of those treated with warfarin stopped treatment during the study.¹⁹ In the ARISTOTLE trial (Apixaban vs Warfarin in Patients With Atrial Fibrillation), 25% of patients discontinued apixaban and 28% discontinued warfarin.²⁰

The results of these trials show a clear need for treatments without high attrition rates, since patients with atrial fibrillation need protection from stroke for the rest of their life.

SURGICAL CLOSURE AS AN ADD-ON TO OTHER PROCEDURES

If the patient is undergoing cardiac surgery for another reason, the surgeon can excise, suture, staple, or clip the left atrial appendage shut at the same time. Closure is recommended as part of valve replacement.⁸ In a 2008 retrospective study, left atrial appendage closure was successfully performed in 40% of those undergoing the procedure during cardiac surgery, and complete closure occurred more often with excision than with suture exclusion and stapler exclusion.²³ A study of patients who underwent ligation of the left atrial appendage during mitral value replacement found that 35% demonstrated incomplete closure as determined by transesophageal echocardiography.²⁴

Newer devices have shown more success. The AtriClip (AtriCure Inc., West Chester, OH) is a self-closing, implantable clip applied epicardially by either an open surgical or a minimally invasive technique.²⁵ Successful closure was confirmed in 60 of 61 patients at 90 days as determined by computed tomography or transesophageal echocardiography, and there were no adverse events related to implantation of the device.²⁵ The TigerPaw system (Terumo Cardiovascular Systems, Ann Arbor, MI)²⁶ is a fastener delivered surgically around the base of the ostium of the left atrial appendage. In an initial trial, 90 days after the procedure, transesophageal echocardiography showed no leaks in any of those who were examined (54 of 60 patients).

Amputation of the left atrial appendage is also considered part of the maze procedure for atrial fibrillation, in which the operator creates multiple small scars in the atria to prevent irregular impulses from being conducted.²⁷

Results of these surgical approaches have been mixed, as incomplete closure or clipping actually increases the risk of left atrial thrombus formation and embolization.²⁸ Moreover, these invasive surgical techniques are associated with significant periprocedural morbidity.²⁹ Because of the high risk of surgical complications, cardiac specialists have sought less invasive percutaneous procedures to manage stroke risk in patients with atrial fibrillation.

PERCUTANEOUS OCCLUSION

One option for closing the left atrial appendage is a percutaneous transseptal approach in which a plug is placed in the opening connecting the appendage to the rest of the atrium.

The PLAATO device

The Percutaneous LAA Transcatheter Occlusion (PLAATO) device (Appriva Medical Inc., Sunnyvale, CA) contains an expandable nitinol-covered cage designed to be placed in the orifice of the left atrial appendage. Over time, tissue grows into the device, entirely isolating the appendage from the rest of the atrium.

In 2002, Sievert et al³⁰ reported using this device in 15 patients. Subsequently, in a nonrandomized trial in patients with contraindications to lifelong anticoagulation, total occlusion was achieved in 108 of 111 patients, with no thrombosis or migration of the device at 6 months. The annual risk of stroke was 2.2%, a reduction in relative risk of 65% based on the CHADS₂ score.³¹

But despite this apparent success, the PLAATO device was discontinued for unspecified commercial reasons.

Amplatzer cardiac plug

Modeled after an atrial septal occluder, the Amplatzer cardiac plug (St. Jude Medical, St. Paul, MN) consists of a lobe and a disk connected by a central waist.

In 2011, Park et al³² published their initial experience implanting this device in patients who either could not tolerate or did not desire long-term anticoagulation. They reported a 96% closure rate (137 of 143 patients), but there were serious complications in 10 patients: 3 with ischemic stroke, 2 with device embolism, and 5 with pericardial effusions.

Warfarin remains the foundation of stroke prevention in atrial fibrillation

Urena et al³³ reported similar results in 52 patients with absolute contraindications to warfarin, with a 98.1% implantation rate. Patients were then maintained on either single or dual antiplatelet therapy at the discretion of the operator. At 20-month follow-up, there had been one stroke, one transient ischemic attack, and one major bleeding event. The leakage rate was 16.2% as determined by transesophageal echocardiography.

While initial results were promising, a clinical trial comparing this device and optimal medical treatment is currently on hold. Thus, there are no clear data comparing the Amplatzer device with oral anticoagulation.³⁴

The Watchman device

The Watchman device (Boston Scientific, Marlborough, MA), an evolution of the PLAATO device, is a self-expanding nitinol structure with fixation barbs and a polyethylene membrane to protect the atrium-facing side of the device (Figure 1).

A pilot trial reported successful implantation in 66 of 75 patients, though the device was found to migrate after placement in 5 of the first 16 patients using the original device and delivery system. The device was modified, and no further embolization of the device occurred.³⁵

The PROTECT-AF trial (Protection in Patients With Atrial Fibrillation)³⁶ was the first completed and published randomized controlled trial evaluating left atrial appendage closure using a device vs long-term warfarin therapy. This study randomized 707 people with nonvalvular atrial fibrillation from 59 centers worldwide to receive the Watchman device or a control treatment. The study included patients age 18 or older with nonvalvular atrial fibrillation who were able to tolerate warfarin therapy. Patients in the control group received warfarin for the duration of the study and were monitored every 2 weeks for a goal INR of 2 to 3, achieving a therapeutic INR 66% of the time. The device group was also treated with warfarin for 45 days to allow device endothelialization. Warfarin was discontinued if transesophageal echocardiography showed complete closure or significantly decreased flow around the device. Patients in the device group were then treated with aspirin and clopidogrel for 6 months, and then with aspirin indefinitely.

Incomplete closure or clipping actually increases the risk of thrombosis and embolization

At 1,065 patient-years of follow-up, PROTECT-AF showed that in patients with atrial fibrillation who were candidates for warfarin therapy, percutaneous left atrial appendage closure using the Watchman device reduced the rate of hemorrhagic stroke compared with warfarin and was noninferior to warfarin in terms of all-cause mortality and stroke. A 4-year follow-up to the PROTECT-AF trial found that receiving the Watchman was better than taking warfarin in terms of risk of cardiovascular death, stroke and other systemic embolization, and all-cause mortality. The adverse event rates were 2.3% in the device group and 3.8% in the control group, a 40% relative risk reduction in the Watchman group.³⁷

The PREVAIL trial (Prospective Randomized Evaluation of the WATCHMAN LAA Closure Device in Patients With Atrial Fibrillation vs Long-Term Warfarin Therapy) aimed to confirm the safety and efficacy of the Watchman device compared with long-term warfarin therapy.³⁸ The event rate (defined as 7-day occurrence of death, ischemic stroke, systemic embolism, and procedure- or device-related complications requiring major cardiovascular or endovascular intervention) was 2.2%. But the PREVAIL trial was unable to show that the device was noninferior to warfarin in terms of its second primary end point of stroke, systemic embolism, and cardiovascular or unexplained death at 18 months. When performed by physicians who were new to the procedure, the procedure was successful (ie, the device was successfully implanted) in 93.2%; the rate was slightly higher (96.3%) when performed by experienced implanters.

Safety data gathered in PREVAIL in conjunction with demonstrated efficacy from PROTECT-AF suggest that the Watchman device may be a safe and effective alternative to long-term oral anticoagulation in patients with nonvalvular atrial fibrillation.

In patients with contraindications to warfarin

Most of the published data have been about the efficacy of occlusion devices compared with long-term warfarin therapy. Unfortunately, the population that has not been studied extensively is patients who have contraindications to long-term oral anticoagulation, who would benefit the most from an occlusive device.

The ASA Plavix Feasibility Study (ASAP) focused on this population, specifically those who had a CHADS₂ score of 1 or higher and who were considered ineligible for warfarin, to determine whether closure using the Watchman device could be safely performed without a transition period with warfarin.³⁹ After device implantation, trial participants were given clopidogrel for 6 months and aspirin indefinitely. The trial enrolled 150 patients and followed them for a mean of 14.4 (± 8.6) months. In that time, there were four strokes, five pericardial effusions, and six instances of device-related thrombus by transesophageal echocardiography. Three of the strokes were ischemic (1.7% per year), which is a 77% reduction from the expected rate of 7.3% based on the CHADS₂ scores of the patient cohort.

These data suggest that implantation of the Watchman device may be appropriate for those who cannot tolerate warfarin even in the short term.

The Lariat system

The Lariat suture delivery device (SentreHeart, Inc., Redwood City, CA) is approved by the US Food and Drug Administration (FDA) for soft-tissue closure and has been used for percutaneous left atrial appendage closure. It uses a magnet-tipped wire that is passed to the epicardial side of the left atrial appendage via pericardial access to meet a second magnet-tipped wire introduced into the appendage via transseptal access. A “lasso” is then advanced over the epicardial guide wire and tightened down around the ostium of the left atrial appendage. This tool facilitates deployment of a nonabsorbable polyester suture, which effectively ligates off the appendage from the rest of the left atrium (Figure 2).⁴⁰ In theory, the Lariat’s epicardial approach could eliminate the need for short- and long-term anticoagulation, as there would be no foreign body left within the heart.

In an initial cohort of 89 patients in Poland,41 the investigators reported a 96% closure rate as determined by transesophageal echocardiography immediately after the procedure. At 1-year follow-up, there was 98% complete closure, including cases of incomplete closure detected earlier.41 Adverse events were limited, with only two cases of severe pericarditis, two strokes, and one pericardial effusion. These results were replicated in the United States in a cohort of 25 patients, with a 100% closure rate and no stroke events.⁴²

There have been three published case reports of left atrial clot formation after successful left atrial appendage ligation using the Lariat device.^43–45 These experiences further emphasize that closure does not necessarily confer instant stroke prevention, and there remains a need to investigate the need for routine imaging and possibly periprocedural anticoagulation after ligation.

More recently, Pillai et al⁴⁶ published their initial experience following 71 patients with echocardiograms 3 months after left atrial appendage closure using the Lariat device. They reported leaks in 6 of the 71 patients; five of the leaks were successfully closed using the Amplatzer Septal Occluder, and one was closed with a repeat Lariat procedure.

Although the Lariat system has been used in more than 2,000 patients worldwide (SentreHeart, personal communication), there has been no published systematic comparison between it and oral anticoagulation to date.

AN EMERGING OPTION

Established guidelines help determine which patients with atrial fibrillation should receive oral anticoagulant therapy. For patients who have absolute contraindications to oral anticoagulants or who are undergoing cardiac surgery, surgical ligation of the left atrial appendage is an option. But for those with contraindications to oral anticoagulation in both the short term and the long term, there is a growing body of evidence suggesting that a percutaneous intervention is at least noninferior to—and in some cases is superior to—warfarin. Figure 3 shows our recommendations for the steps to determine which patients would be most appropriate to consider for left atrial appendage closure.

Holmes et al⁴⁷ propose that we may now have enough evidence to support an expedited regulatory approval process of these occlusion devices. But there are still a number of areas in which further investigation is clearly needed before left atrial appendage occlusion devices can be widely adopted.

The trials discussed above had specific inclusion and exclusion criteria, and therefore, although they support percutaneous intervention, the generalizability of their results remains in question. Indeed, the patients in PROTECT-AF³⁶ had an average CHADS₂ score of only 2.2. This study also included only patients who were able to tolerate both aspirin and clopidogrel simultaneously for a significant amount of time. Hence, one cannot assume the results would be the same in patients who have strict contraindications to warfarin or any target-specific oral anticoagulant. Concern regarding the generalizability of the conclusions from PROTECT-AF and PREVAIL has led to mixed votes (three assessments to date) from the FDA Circulatory Device Panel.⁴⁸

In an encouraging review of cases, Gafoor et al⁴⁹ reported safe and efficacious occlusion in octogenarians using the devices mentioned above. These patients often pose the greatest challenge in initiating long-term anticoagulation because of the many drug-drug interactions and the risk of intracranial hemorrhage secondary to falls.

Further, while occlusion devices would clearly be useful for patients in whom traditional oral anticoagulation is not an option, the newer oral anticoagulants might complicate the picture somewhat. As shown by Ruff et al,²¹ the risk-benefit ratio of these target-specific oral anticoagulants is quite favorable and by some measurements is superior to that of warfarin. Could there be a group of patients who cannot take warfarin but could instead do well on one of the newer anticoagulants, thus alleviating the need for percutaneous intervention? As the newer oral anticoagulants become more commonly used, the cost-benefit analysis of implanting an occlusion device could shift.

We expect that percutaneous closure will someday be a viable and equal option for stroke prevention

Lastly, in this era of high-value medical care, one must consider the cost-effectiveness of these novel interventions. As with any new technology, the up-front cost of implantation is certainly greater than that of warfarin therapy. If device implantation can prevent a hospitalization from a major bleed secondary to warfarin use or prevent a catastrophic stroke due to untreated atrial fibrillation, then the cost-benefit analysis may be tipped in the other direction. As these devices become more widely available and physicians have more experience implanting them, the costs will likely decrease.

As with oral anticoagulation therapy, all interventions, whether surgical or percutaneous, carry a risk of bleeding and stroke. There remains no substitute for frank and clear discussions between the physician and patient regarding the risks and benefits of each approach.

While a growing body of evidence surrounds left atrial appendage occlusion devices, many questions remain. Notably, could these devices be used in patients who can tolerate oral anticoagulants? And if so, which subgroups would benefit most? Does occlusion or ligation of the left atrial appendage affect electrical connections between it and the left atrium, thereby lowering the burden of atrial fibrillation?

We expect that continued investigation of and experience with left atrial appendage closure devices will position them one day as a viable and equal option for preventing stroke in patients with atrial fibrillation.

Can patients with atrial fibrillation  undergo a percutaneous procedure to reduce their risk of stroke, thereby eliminating the need for lifelong treatment with an oral anticoagulant drug? The data are preliminary, but this is an emerging option that physicians should be aware of.

We review here the current evidence and techniques aimed at isolating the left atrial appendage to prevent stroke, and we emphasize the need for continued systematic comparisons between oral anticoagulation and percutaneous treatment options.

NOVEL TREATMENTS ARE NEEDED

Atrial fibrillation is the most common cardiac arrhythmia,¹ affecting an estimated 1% to 2% of people worldwide. In 2001, an estimated 2.3 million persons in the United States had atrial fibrillation, and that number is expected to more than double by 2050.²

Atrial fibrillation independently increases the risk of stroke by a factor of 4 to 5.³ The American Heart Association ranks stroke as the fourth most common cause of death and the leading cause of disability in the United States.⁴ Atrial fibrillation accounts for 15% of strokes in people of all ages and 30% in those over age 80.⁵ Untreated, 2% to 5% of patients with atrial fibrillation suffer a stroke in any given year.⁶ Most of these strokes are cardioembolic, with thrombi originating in the left atrial appendage.⁷ Furthermore, it has been estimated^8,9 that patients with atrial fibrillation who have already had a stroke and cannot tolerate oral anticoagulants have an annual risk of stroke close to 12% and a relative risk of approximately 3.0 compared with those with atrial fibrillation and prior stroke who can tolerate anticoagulation.

Oral anticoagulation effectively prevents thromboembolic events associated with atrial fibrillation,¹⁰ but several factors limit its efficacy and applicability. The risk of bleeding complications, the need for frequent monitoring, and challenges with compliance create a large population of patients who would benefit from alternative approaches. Consequently, physicians have looked for other ways to prevent stroke—especially surgical and transcatheter procedures—that are not associated with an ongoing risk of hemorrhage and a lifelong need to take an anticoagulant.

THE LEFT ATRIAL APPENDAGE: A SITE OF CLOT FORMATION

The left atrial appendage is the most common site of thrombus formation, particularly in patients with nonvalvular atrial fibrillation. Nearly 90% of thrombi discovered in the left atrium form in the appendage.⁷ A study of 233 patients not on long-term anticoagulation revealed that after 48 hours of atrial fibrillation, 15% had a left atrial thrombus, and all but one of the thrombi were in the appendage.¹¹

Atrial fibrillation increases the risk of stroke by a factor of 4 to 5

Believed to function as a decompression chamber during left ventricular systole, the left atrial appendage is embryologically derived from the left wall of the primary atrium. It is in close proximity to the free wall of the left ventricle, and therefore its flow can vary with left ventricular function. Relative stasis due to its location and extensive trabeculations, especially in times of poor forward flow, make it a high-risk site for clot formation.¹²

ANTICOAGULATION: EFFECTIVE BUT IMPERFECT

In deciding whether a patient with atrial fibrillation should be prescribed anticoagulation therapy, the physician must balance the risk of stroke against the risk of bleeding. Several tools for assessing these two risks have been developed. Of note, some of the risk factors for stroke are the same as some of the risk factors for bleeding.

Calculating the risk of stroke

CHADS₂ and CHA2DS₂-VASc are the two most commonly used tools for assessing the risk of stroke, but only the newer CHA₂DS₂-VASc has received a class I recommendation (the highest) from the European Society of Cardiology (ESC).¹³

CHADS₂ risk factors are Congestive heart failure (1 point), Hypertension (1 point), Age 75 or older (1 point), Diabetes (1 point), and  Stroke or transient ischemic attack (2 points). Risk of stroke is considered low with a score of 0, intermediate with a score of 1, and high with a score of 2 or more. 

CHA₂DS₂-VASc risk factors are Congestive heart failure or left ventricular ejection fraction ≤ 40% (1 point), Hypertension (1 point), Age ≥ 75 (2 points), Age 65–74 (1 point), Diabetes mellitus (1 point), Stroke, transient ischemic attack, or thromboembolism (2 points), Vascular disease (1 point), and female Sex (1 point). Low risk is defined as a score of 0 for a man or, for a woman with no other risk factors, a score of 1. A score of 1 for a man indicates moderate risk, and a score of 2 or more is high risk.

Calculating the risk of bleeding

Tools for assessing bleeding risk include ATRIA2 and HAS-BLED,¹⁴ the latter carrying a class I recommendation from the ESC.¹³

HAS-BLED risk factors are Hypertension (1 point), Abnormal renal or liver function (1 point each), Stroke (1 point), Bleeding (1 point), Labile international normalized ratio (INR) (1 point), Elderly (age > 65) (1 point), and Drug or alcohol use (1 point each). The risk of bleeding is considered high with a score of 3 or higher.

Disadvantages of oral anticoagulation

Oral anticoagulation is the standard treatment for preventing stroke in patients with atrial fibrillation, and the vitamin K antagonist warfarin remains the foundation.

Though highly effective, warfarin requires close monitoring and frequent dose adjustments because of its numerous food and drug interactions. Bleeding risk and the challenge of frequent monitoring rule out treatment with warfarin in 14% to 44% of patients with atrial fibrillation.¹⁵ Even in “ideal” candidates, warfarin is underused, with one study reporting that only 38% of those with clinical indications for it had been prescribed warfarin, and of those for whom it had not been prescribed, 63% were also not taking aspirin.¹⁶ Moreover, a meta-analysis suggested that the average patient treated with warfarin has his or her INR in the therapeutic range only about 55% of the time.¹⁷

Newer, target-specific oral anticoagulants such as dabigatran (a direct thrombin inhibitor) and rivaroxaban and apixaban (both factor Xa inhibitors) do not require monitoring and have fewer drug interactions. But like warfarin, they also confer a risk of serious bleeding.^18–20 Most of the studies of these newer drugs have compared them with warfarin, with the preponderance of evidence showing them to be either noninferior or superior to warfarin for stroke reduction. But bleeding complication rates remain significant, apixaban having lower rates of major bleeding than dabigatran and rivaroxaban.

Untreated, 2%–5% of patients with atrial fibrillation will suffer a stroke in any given year

In a meta-analysis, Ruff et al²¹ concluded that the target-specific oral anticoagulants provide a favorable balance of risk and benefit. Compared with warfarin, these new drugs reduced the rate of stroke or systemic embolic events by 19%. There was also a significant reduction in rates of intracranial hemorrhage and all-cause mortality. The risk of major bleeding was similar to that with warfarin, though there was a higher risk of gastrointestinal bleeding with target-specific agents. These effects were consistent across a wide range of patients.

Given the difficulties, risks, and serious side effects of anticoagulant therapy, many patients stop taking these drugs soon after starting them, either on their own or on their physician’s recommendation. In the RE-LY trial (Dabigatran vs Warfarin in Patients With Atrial Fibrillation), 10% of patients receiving dabigatran and 17% of those receiving warfarin stopped the treatment within 1 to 2 years.²² In a similar trial of rivaroxaban vs warfarin in nonvalvular atrial fibrillation (the ROCKET-AF trial), 24% of those treated with rivaroxaban and 22% of those treated with warfarin stopped treatment during the study.¹⁹ In the ARISTOTLE trial (Apixaban vs Warfarin in Patients With Atrial Fibrillation), 25% of patients discontinued apixaban and 28% discontinued warfarin.²⁰

The results of these trials show a clear need for treatments without high attrition rates, since patients with atrial fibrillation need protection from stroke for the rest of their life.

SURGICAL CLOSURE AS AN ADD-ON TO OTHER PROCEDURES

If the patient is undergoing cardiac surgery for another reason, the surgeon can excise, suture, staple, or clip the left atrial appendage shut at the same time. Closure is recommended as part of valve replacement.⁸ In a 2008 retrospective study, left atrial appendage closure was successfully performed in 40% of those undergoing the procedure during cardiac surgery, and complete closure occurred more often with excision than with suture exclusion and stapler exclusion.²³ A study of patients who underwent ligation of the left atrial appendage during mitral value replacement found that 35% demonstrated incomplete closure as determined by transesophageal echocardiography.²⁴

Newer devices have shown more success. The AtriClip (AtriCure Inc., West Chester, OH) is a self-closing, implantable clip applied epicardially by either an open surgical or a minimally invasive technique.²⁵ Successful closure was confirmed in 60 of 61 patients at 90 days as determined by computed tomography or transesophageal echocardiography, and there were no adverse events related to implantation of the device.²⁵ The TigerPaw system (Terumo Cardiovascular Systems, Ann Arbor, MI)²⁶ is a fastener delivered surgically around the base of the ostium of the left atrial appendage. In an initial trial, 90 days after the procedure, transesophageal echocardiography showed no leaks in any of those who were examined (54 of 60 patients).

Amputation of the left atrial appendage is also considered part of the maze procedure for atrial fibrillation, in which the operator creates multiple small scars in the atria to prevent irregular impulses from being conducted.²⁷

Results of these surgical approaches have been mixed, as incomplete closure or clipping actually increases the risk of left atrial thrombus formation and embolization.²⁸ Moreover, these invasive surgical techniques are associated with significant periprocedural morbidity.²⁹ Because of the high risk of surgical complications, cardiac specialists have sought less invasive percutaneous procedures to manage stroke risk in patients with atrial fibrillation.

PERCUTANEOUS OCCLUSION

One option for closing the left atrial appendage is a percutaneous transseptal approach in which a plug is placed in the opening connecting the appendage to the rest of the atrium.

The PLAATO device

The Percutaneous LAA Transcatheter Occlusion (PLAATO) device (Appriva Medical Inc., Sunnyvale, CA) contains an expandable nitinol-covered cage designed to be placed in the orifice of the left atrial appendage. Over time, tissue grows into the device, entirely isolating the appendage from the rest of the atrium.

In 2002, Sievert et al³⁰ reported using this device in 15 patients. Subsequently, in a nonrandomized trial in patients with contraindications to lifelong anticoagulation, total occlusion was achieved in 108 of 111 patients, with no thrombosis or migration of the device at 6 months. The annual risk of stroke was 2.2%, a reduction in relative risk of 65% based on the CHADS₂ score.³¹

But despite this apparent success, the PLAATO device was discontinued for unspecified commercial reasons.

Amplatzer cardiac plug

Modeled after an atrial septal occluder, the Amplatzer cardiac plug (St. Jude Medical, St. Paul, MN) consists of a lobe and a disk connected by a central waist.

In 2011, Park et al³² published their initial experience implanting this device in patients who either could not tolerate or did not desire long-term anticoagulation. They reported a 96% closure rate (137 of 143 patients), but there were serious complications in 10 patients: 3 with ischemic stroke, 2 with device embolism, and 5 with pericardial effusions.

Warfarin remains the foundation of stroke prevention in atrial fibrillation

Urena et al³³ reported similar results in 52 patients with absolute contraindications to warfarin, with a 98.1% implantation rate. Patients were then maintained on either single or dual antiplatelet therapy at the discretion of the operator. At 20-month follow-up, there had been one stroke, one transient ischemic attack, and one major bleeding event. The leakage rate was 16.2% as determined by transesophageal echocardiography.

While initial results were promising, a clinical trial comparing this device and optimal medical treatment is currently on hold. Thus, there are no clear data comparing the Amplatzer device with oral anticoagulation.³⁴

The Watchman device

The Watchman device (Boston Scientific, Marlborough, MA), an evolution of the PLAATO device, is a self-expanding nitinol structure with fixation barbs and a polyethylene membrane to protect the atrium-facing side of the device (Figure 1).

A pilot trial reported successful implantation in 66 of 75 patients, though the device was found to migrate after placement in 5 of the first 16 patients using the original device and delivery system. The device was modified, and no further embolization of the device occurred.³⁵

The PROTECT-AF trial (Protection in Patients With Atrial Fibrillation)³⁶ was the first completed and published randomized controlled trial evaluating left atrial appendage closure using a device vs long-term warfarin therapy. This study randomized 707 people with nonvalvular atrial fibrillation from 59 centers worldwide to receive the Watchman device or a control treatment. The study included patients age 18 or older with nonvalvular atrial fibrillation who were able to tolerate warfarin therapy. Patients in the control group received warfarin for the duration of the study and were monitored every 2 weeks for a goal INR of 2 to 3, achieving a therapeutic INR 66% of the time. The device group was also treated with warfarin for 45 days to allow device endothelialization. Warfarin was discontinued if transesophageal echocardiography showed complete closure or significantly decreased flow around the device. Patients in the device group were then treated with aspirin and clopidogrel for 6 months, and then with aspirin indefinitely.

Incomplete closure or clipping actually increases the risk of thrombosis and embolization

At 1,065 patient-years of follow-up, PROTECT-AF showed that in patients with atrial fibrillation who were candidates for warfarin therapy, percutaneous left atrial appendage closure using the Watchman device reduced the rate of hemorrhagic stroke compared with warfarin and was noninferior to warfarin in terms of all-cause mortality and stroke. A 4-year follow-up to the PROTECT-AF trial found that receiving the Watchman was better than taking warfarin in terms of risk of cardiovascular death, stroke and other systemic embolization, and all-cause mortality. The adverse event rates were 2.3% in the device group and 3.8% in the control group, a 40% relative risk reduction in the Watchman group.³⁷

The PREVAIL trial (Prospective Randomized Evaluation of the WATCHMAN LAA Closure Device in Patients With Atrial Fibrillation vs Long-Term Warfarin Therapy) aimed to confirm the safety and efficacy of the Watchman device compared with long-term warfarin therapy.³⁸ The event rate (defined as 7-day occurrence of death, ischemic stroke, systemic embolism, and procedure- or device-related complications requiring major cardiovascular or endovascular intervention) was 2.2%. But the PREVAIL trial was unable to show that the device was noninferior to warfarin in terms of its second primary end point of stroke, systemic embolism, and cardiovascular or unexplained death at 18 months. When performed by physicians who were new to the procedure, the procedure was successful (ie, the device was successfully implanted) in 93.2%; the rate was slightly higher (96.3%) when performed by experienced implanters.

Safety data gathered in PREVAIL in conjunction with demonstrated efficacy from PROTECT-AF suggest that the Watchman device may be a safe and effective alternative to long-term oral anticoagulation in patients with nonvalvular atrial fibrillation.

In patients with contraindications to warfarin

Most of the published data have been about the efficacy of occlusion devices compared with long-term warfarin therapy. Unfortunately, the population that has not been studied extensively is patients who have contraindications to long-term oral anticoagulation, who would benefit the most from an occlusive device.

The ASA Plavix Feasibility Study (ASAP) focused on this population, specifically those who had a CHADS₂ score of 1 or higher and who were considered ineligible for warfarin, to determine whether closure using the Watchman device could be safely performed without a transition period with warfarin.³⁹ After device implantation, trial participants were given clopidogrel for 6 months and aspirin indefinitely. The trial enrolled 150 patients and followed them for a mean of 14.4 (± 8.6) months. In that time, there were four strokes, five pericardial effusions, and six instances of device-related thrombus by transesophageal echocardiography. Three of the strokes were ischemic (1.7% per year), which is a 77% reduction from the expected rate of 7.3% based on the CHADS₂ scores of the patient cohort.

These data suggest that implantation of the Watchman device may be appropriate for those who cannot tolerate warfarin even in the short term.

The Lariat system

The Lariat suture delivery device (SentreHeart, Inc., Redwood City, CA) is approved by the US Food and Drug Administration (FDA) for soft-tissue closure and has been used for percutaneous left atrial appendage closure. It uses a magnet-tipped wire that is passed to the epicardial side of the left atrial appendage via pericardial access to meet a second magnet-tipped wire introduced into the appendage via transseptal access. A “lasso” is then advanced over the epicardial guide wire and tightened down around the ostium of the left atrial appendage. This tool facilitates deployment of a nonabsorbable polyester suture, which effectively ligates off the appendage from the rest of the left atrium (Figure 2).⁴⁰ In theory, the Lariat’s epicardial approach could eliminate the need for short- and long-term anticoagulation, as there would be no foreign body left within the heart.

In an initial cohort of 89 patients in Poland,41 the investigators reported a 96% closure rate as determined by transesophageal echocardiography immediately after the procedure. At 1-year follow-up, there was 98% complete closure, including cases of incomplete closure detected earlier.41 Adverse events were limited, with only two cases of severe pericarditis, two strokes, and one pericardial effusion. These results were replicated in the United States in a cohort of 25 patients, with a 100% closure rate and no stroke events.⁴²

There have been three published case reports of left atrial clot formation after successful left atrial appendage ligation using the Lariat device.^43–45 These experiences further emphasize that closure does not necessarily confer instant stroke prevention, and there remains a need to investigate the need for routine imaging and possibly periprocedural anticoagulation after ligation.

More recently, Pillai et al⁴⁶ published their initial experience following 71 patients with echocardiograms 3 months after left atrial appendage closure using the Lariat device. They reported leaks in 6 of the 71 patients; five of the leaks were successfully closed using the Amplatzer Septal Occluder, and one was closed with a repeat Lariat procedure.

Although the Lariat system has been used in more than 2,000 patients worldwide (SentreHeart, personal communication), there has been no published systematic comparison between it and oral anticoagulation to date.

AN EMERGING OPTION

Established guidelines help determine which patients with atrial fibrillation should receive oral anticoagulant therapy. For patients who have absolute contraindications to oral anticoagulants or who are undergoing cardiac surgery, surgical ligation of the left atrial appendage is an option. But for those with contraindications to oral anticoagulation in both the short term and the long term, there is a growing body of evidence suggesting that a percutaneous intervention is at least noninferior to—and in some cases is superior to—warfarin. Figure 3 shows our recommendations for the steps to determine which patients would be most appropriate to consider for left atrial appendage closure.

Holmes et al⁴⁷ propose that we may now have enough evidence to support an expedited regulatory approval process of these occlusion devices. But there are still a number of areas in which further investigation is clearly needed before left atrial appendage occlusion devices can be widely adopted.

The trials discussed above had specific inclusion and exclusion criteria, and therefore, although they support percutaneous intervention, the generalizability of their results remains in question. Indeed, the patients in PROTECT-AF³⁶ had an average CHADS₂ score of only 2.2. This study also included only patients who were able to tolerate both aspirin and clopidogrel simultaneously for a significant amount of time. Hence, one cannot assume the results would be the same in patients who have strict contraindications to warfarin or any target-specific oral anticoagulant. Concern regarding the generalizability of the conclusions from PROTECT-AF and PREVAIL has led to mixed votes (three assessments to date) from the FDA Circulatory Device Panel.⁴⁸

In an encouraging review of cases, Gafoor et al⁴⁹ reported safe and efficacious occlusion in octogenarians using the devices mentioned above. These patients often pose the greatest challenge in initiating long-term anticoagulation because of the many drug-drug interactions and the risk of intracranial hemorrhage secondary to falls.

Further, while occlusion devices would clearly be useful for patients in whom traditional oral anticoagulation is not an option, the newer oral anticoagulants might complicate the picture somewhat. As shown by Ruff et al,²¹ the risk-benefit ratio of these target-specific oral anticoagulants is quite favorable and by some measurements is superior to that of warfarin. Could there be a group of patients who cannot take warfarin but could instead do well on one of the newer anticoagulants, thus alleviating the need for percutaneous intervention? As the newer oral anticoagulants become more commonly used, the cost-benefit analysis of implanting an occlusion device could shift.

We expect that percutaneous closure will someday be a viable and equal option for stroke prevention

Lastly, in this era of high-value medical care, one must consider the cost-effectiveness of these novel interventions. As with any new technology, the up-front cost of implantation is certainly greater than that of warfarin therapy. If device implantation can prevent a hospitalization from a major bleed secondary to warfarin use or prevent a catastrophic stroke due to untreated atrial fibrillation, then the cost-benefit analysis may be tipped in the other direction. As these devices become more widely available and physicians have more experience implanting them, the costs will likely decrease.

As with oral anticoagulation therapy, all interventions, whether surgical or percutaneous, carry a risk of bleeding and stroke. There remains no substitute for frank and clear discussions between the physician and patient regarding the risks and benefits of each approach.

While a growing body of evidence surrounds left atrial appendage occlusion devices, many questions remain. Notably, could these devices be used in patients who can tolerate oral anticoagulants? And if so, which subgroups would benefit most? Does occlusion or ligation of the left atrial appendage affect electrical connections between it and the left atrium, thereby lowering the burden of atrial fibrillation?

We expect that continued investigation of and experience with left atrial appendage closure devices will position them one day as a viable and equal option for preventing stroke in patients with atrial fibrillation.

Can patients with atrial fibrillation  undergo a percutaneous procedure to reduce their risk of stroke, thereby eliminating the need for lifelong treatment with an oral anticoagulant drug? The data are preliminary, but this is an emerging option that physicians should be aware of.

We review here the current evidence and techniques aimed at isolating the left atrial appendage to prevent stroke, and we emphasize the need for continued systematic comparisons between oral anticoagulation and percutaneous treatment options.

NOVEL TREATMENTS ARE NEEDED

Atrial fibrillation is the most common cardiac arrhythmia,¹ affecting an estimated 1% to 2% of people worldwide. In 2001, an estimated 2.3 million persons in the United States had atrial fibrillation, and that number is expected to more than double by 2050.²

Atrial fibrillation independently increases the risk of stroke by a factor of 4 to 5.³ The American Heart Association ranks stroke as the fourth most common cause of death and the leading cause of disability in the United States.⁴ Atrial fibrillation accounts for 15% of strokes in people of all ages and 30% in those over age 80.⁵ Untreated, 2% to 5% of patients with atrial fibrillation suffer a stroke in any given year.⁶ Most of these strokes are cardioembolic, with thrombi originating in the left atrial appendage.⁷ Furthermore, it has been estimated^8,9 that patients with atrial fibrillation who have already had a stroke and cannot tolerate oral anticoagulants have an annual risk of stroke close to 12% and a relative risk of approximately 3.0 compared with those with atrial fibrillation and prior stroke who can tolerate anticoagulation.

Oral anticoagulation effectively prevents thromboembolic events associated with atrial fibrillation,¹⁰ but several factors limit its efficacy and applicability. The risk of bleeding complications, the need for frequent monitoring, and challenges with compliance create a large population of patients who would benefit from alternative approaches. Consequently, physicians have looked for other ways to prevent stroke—especially surgical and transcatheter procedures—that are not associated with an ongoing risk of hemorrhage and a lifelong need to take an anticoagulant.

THE LEFT ATRIAL APPENDAGE: A SITE OF CLOT FORMATION

The left atrial appendage is the most common site of thrombus formation, particularly in patients with nonvalvular atrial fibrillation. Nearly 90% of thrombi discovered in the left atrium form in the appendage.⁷ A study of 233 patients not on long-term anticoagulation revealed that after 48 hours of atrial fibrillation, 15% had a left atrial thrombus, and all but one of the thrombi were in the appendage.¹¹

Atrial fibrillation increases the risk of stroke by a factor of 4 to 5

Believed to function as a decompression chamber during left ventricular systole, the left atrial appendage is embryologically derived from the left wall of the primary atrium. It is in close proximity to the free wall of the left ventricle, and therefore its flow can vary with left ventricular function. Relative stasis due to its location and extensive trabeculations, especially in times of poor forward flow, make it a high-risk site for clot formation.¹²

ANTICOAGULATION: EFFECTIVE BUT IMPERFECT

In deciding whether a patient with atrial fibrillation should be prescribed anticoagulation therapy, the physician must balance the risk of stroke against the risk of bleeding. Several tools for assessing these two risks have been developed. Of note, some of the risk factors for stroke are the same as some of the risk factors for bleeding.

Calculating the risk of stroke

CHADS₂ and CHA2DS₂-VASc are the two most commonly used tools for assessing the risk of stroke, but only the newer CHA₂DS₂-VASc has received a class I recommendation (the highest) from the European Society of Cardiology (ESC).¹³

CHADS₂ risk factors are Congestive heart failure (1 point), Hypertension (1 point), Age 75 or older (1 point), Diabetes (1 point), and  Stroke or transient ischemic attack (2 points). Risk of stroke is considered low with a score of 0, intermediate with a score of 1, and high with a score of 2 or more. 

CHA₂DS₂-VASc risk factors are Congestive heart failure or left ventricular ejection fraction ≤ 40% (1 point), Hypertension (1 point), Age ≥ 75 (2 points), Age 65–74 (1 point), Diabetes mellitus (1 point), Stroke, transient ischemic attack, or thromboembolism (2 points), Vascular disease (1 point), and female Sex (1 point). Low risk is defined as a score of 0 for a man or, for a woman with no other risk factors, a score of 1. A score of 1 for a man indicates moderate risk, and a score of 2 or more is high risk.

Calculating the risk of bleeding

Tools for assessing bleeding risk include ATRIA2 and HAS-BLED,¹⁴ the latter carrying a class I recommendation from the ESC.¹³

HAS-BLED risk factors are Hypertension (1 point), Abnormal renal or liver function (1 point each), Stroke (1 point), Bleeding (1 point), Labile international normalized ratio (INR) (1 point), Elderly (age > 65) (1 point), and Drug or alcohol use (1 point each). The risk of bleeding is considered high with a score of 3 or higher.

Disadvantages of oral anticoagulation

Oral anticoagulation is the standard treatment for preventing stroke in patients with atrial fibrillation, and the vitamin K antagonist warfarin remains the foundation.

Though highly effective, warfarin requires close monitoring and frequent dose adjustments because of its numerous food and drug interactions. Bleeding risk and the challenge of frequent monitoring rule out treatment with warfarin in 14% to 44% of patients with atrial fibrillation.¹⁵ Even in “ideal” candidates, warfarin is underused, with one study reporting that only 38% of those with clinical indications for it had been prescribed warfarin, and of those for whom it had not been prescribed, 63% were also not taking aspirin.¹⁶ Moreover, a meta-analysis suggested that the average patient treated with warfarin has his or her INR in the therapeutic range only about 55% of the time.¹⁷

Newer, target-specific oral anticoagulants such as dabigatran (a direct thrombin inhibitor) and rivaroxaban and apixaban (both factor Xa inhibitors) do not require monitoring and have fewer drug interactions. But like warfarin, they also confer a risk of serious bleeding.^18–20 Most of the studies of these newer drugs have compared them with warfarin, with the preponderance of evidence showing them to be either noninferior or superior to warfarin for stroke reduction. But bleeding complication rates remain significant, apixaban having lower rates of major bleeding than dabigatran and rivaroxaban.

Untreated, 2%–5% of patients with atrial fibrillation will suffer a stroke in any given year

In a meta-analysis, Ruff et al²¹ concluded that the target-specific oral anticoagulants provide a favorable balance of risk and benefit. Compared with warfarin, these new drugs reduced the rate of stroke or systemic embolic events by 19%. There was also a significant reduction in rates of intracranial hemorrhage and all-cause mortality. The risk of major bleeding was similar to that with warfarin, though there was a higher risk of gastrointestinal bleeding with target-specific agents. These effects were consistent across a wide range of patients.

Given the difficulties, risks, and serious side effects of anticoagulant therapy, many patients stop taking these drugs soon after starting them, either on their own or on their physician’s recommendation. In the RE-LY trial (Dabigatran vs Warfarin in Patients With Atrial Fibrillation), 10% of patients receiving dabigatran and 17% of those receiving warfarin stopped the treatment within 1 to 2 years.²² In a similar trial of rivaroxaban vs warfarin in nonvalvular atrial fibrillation (the ROCKET-AF trial), 24% of those treated with rivaroxaban and 22% of those treated with warfarin stopped treatment during the study.¹⁹ In the ARISTOTLE trial (Apixaban vs Warfarin in Patients With Atrial Fibrillation), 25% of patients discontinued apixaban and 28% discontinued warfarin.²⁰

The results of these trials show a clear need for treatments without high attrition rates, since patients with atrial fibrillation need protection from stroke for the rest of their life.

SURGICAL CLOSURE AS AN ADD-ON TO OTHER PROCEDURES

If the patient is undergoing cardiac surgery for another reason, the surgeon can excise, suture, staple, or clip the left atrial appendage shut at the same time. Closure is recommended as part of valve replacement.⁸ In a 2008 retrospective study, left atrial appendage closure was successfully performed in 40% of those undergoing the procedure during cardiac surgery, and complete closure occurred more often with excision than with suture exclusion and stapler exclusion.²³ A study of patients who underwent ligation of the left atrial appendage during mitral value replacement found that 35% demonstrated incomplete closure as determined by transesophageal echocardiography.²⁴

Newer devices have shown more success. The AtriClip (AtriCure Inc., West Chester, OH) is a self-closing, implantable clip applied epicardially by either an open surgical or a minimally invasive technique.²⁵ Successful closure was confirmed in 60 of 61 patients at 90 days as determined by computed tomography or transesophageal echocardiography, and there were no adverse events related to implantation of the device.²⁵ The TigerPaw system (Terumo Cardiovascular Systems, Ann Arbor, MI)²⁶ is a fastener delivered surgically around the base of the ostium of the left atrial appendage. In an initial trial, 90 days after the procedure, transesophageal echocardiography showed no leaks in any of those who were examined (54 of 60 patients).

Amputation of the left atrial appendage is also considered part of the maze procedure for atrial fibrillation, in which the operator creates multiple small scars in the atria to prevent irregular impulses from being conducted.²⁷

Results of these surgical approaches have been mixed, as incomplete closure or clipping actually increases the risk of left atrial thrombus formation and embolization.²⁸ Moreover, these invasive surgical techniques are associated with significant periprocedural morbidity.²⁹ Because of the high risk of surgical complications, cardiac specialists have sought less invasive percutaneous procedures to manage stroke risk in patients with atrial fibrillation.

PERCUTANEOUS OCCLUSION

One option for closing the left atrial appendage is a percutaneous transseptal approach in which a plug is placed in the opening connecting the appendage to the rest of the atrium.

The PLAATO device

The Percutaneous LAA Transcatheter Occlusion (PLAATO) device (Appriva Medical Inc., Sunnyvale, CA) contains an expandable nitinol-covered cage designed to be placed in the orifice of the left atrial appendage. Over time, tissue grows into the device, entirely isolating the appendage from the rest of the atrium.

In 2002, Sievert et al³⁰ reported using this device in 15 patients. Subsequently, in a nonrandomized trial in patients with contraindications to lifelong anticoagulation, total occlusion was achieved in 108 of 111 patients, with no thrombosis or migration of the device at 6 months. The annual risk of stroke was 2.2%, a reduction in relative risk of 65% based on the CHADS₂ score.³¹

But despite this apparent success, the PLAATO device was discontinued for unspecified commercial reasons.

Amplatzer cardiac plug

Modeled after an atrial septal occluder, the Amplatzer cardiac plug (St. Jude Medical, St. Paul, MN) consists of a lobe and a disk connected by a central waist.

In 2011, Park et al³² published their initial experience implanting this device in patients who either could not tolerate or did not desire long-term anticoagulation. They reported a 96% closure rate (137 of 143 patients), but there were serious complications in 10 patients: 3 with ischemic stroke, 2 with device embolism, and 5 with pericardial effusions.

Warfarin remains the foundation of stroke prevention in atrial fibrillation

Urena et al³³ reported similar results in 52 patients with absolute contraindications to warfarin, with a 98.1% implantation rate. Patients were then maintained on either single or dual antiplatelet therapy at the discretion of the operator. At 20-month follow-up, there had been one stroke, one transient ischemic attack, and one major bleeding event. The leakage rate was 16.2% as determined by transesophageal echocardiography.

While initial results were promising, a clinical trial comparing this device and optimal medical treatment is currently on hold. Thus, there are no clear data comparing the Amplatzer device with oral anticoagulation.³⁴

The Watchman device

The Watchman device (Boston Scientific, Marlborough, MA), an evolution of the PLAATO device, is a self-expanding nitinol structure with fixation barbs and a polyethylene membrane to protect the atrium-facing side of the device (Figure 1).

A pilot trial reported successful implantation in 66 of 75 patients, though the device was found to migrate after placement in 5 of the first 16 patients using the original device and delivery system. The device was modified, and no further embolization of the device occurred.³⁵

The PROTECT-AF trial (Protection in Patients With Atrial Fibrillation)³⁶ was the first completed and published randomized controlled trial evaluating left atrial appendage closure using a device vs long-term warfarin therapy. This study randomized 707 people with nonvalvular atrial fibrillation from 59 centers worldwide to receive the Watchman device or a control treatment. The study included patients age 18 or older with nonvalvular atrial fibrillation who were able to tolerate warfarin therapy. Patients in the control group received warfarin for the duration of the study and were monitored every 2 weeks for a goal INR of 2 to 3, achieving a therapeutic INR 66% of the time. The device group was also treated with warfarin for 45 days to allow device endothelialization. Warfarin was discontinued if transesophageal echocardiography showed complete closure or significantly decreased flow around the device. Patients in the device group were then treated with aspirin and clopidogrel for 6 months, and then with aspirin indefinitely.

Incomplete closure or clipping actually increases the risk of thrombosis and embolization

At 1,065 patient-years of follow-up, PROTECT-AF showed that in patients with atrial fibrillation who were candidates for warfarin therapy, percutaneous left atrial appendage closure using the Watchman device reduced the rate of hemorrhagic stroke compared with warfarin and was noninferior to warfarin in terms of all-cause mortality and stroke. A 4-year follow-up to the PROTECT-AF trial found that receiving the Watchman was better than taking warfarin in terms of risk of cardiovascular death, stroke and other systemic embolization, and all-cause mortality. The adverse event rates were 2.3% in the device group and 3.8% in the control group, a 40% relative risk reduction in the Watchman group.³⁷

The PREVAIL trial (Prospective Randomized Evaluation of the WATCHMAN LAA Closure Device in Patients With Atrial Fibrillation vs Long-Term Warfarin Therapy) aimed to confirm the safety and efficacy of the Watchman device compared with long-term warfarin therapy.³⁸ The event rate (defined as 7-day occurrence of death, ischemic stroke, systemic embolism, and procedure- or device-related complications requiring major cardiovascular or endovascular intervention) was 2.2%. But the PREVAIL trial was unable to show that the device was noninferior to warfarin in terms of its second primary end point of stroke, systemic embolism, and cardiovascular or unexplained death at 18 months. When performed by physicians who were new to the procedure, the procedure was successful (ie, the device was successfully implanted) in 93.2%; the rate was slightly higher (96.3%) when performed by experienced implanters.

Safety data gathered in PREVAIL in conjunction with demonstrated efficacy from PROTECT-AF suggest that the Watchman device may be a safe and effective alternative to long-term oral anticoagulation in patients with nonvalvular atrial fibrillation.

In patients with contraindications to warfarin

Most of the published data have been about the efficacy of occlusion devices compared with long-term warfarin therapy. Unfortunately, the population that has not been studied extensively is patients who have contraindications to long-term oral anticoagulation, who would benefit the most from an occlusive device.

The ASA Plavix Feasibility Study (ASAP) focused on this population, specifically those who had a CHADS₂ score of 1 or higher and who were considered ineligible for warfarin, to determine whether closure using the Watchman device could be safely performed without a transition period with warfarin.³⁹ After device implantation, trial participants were given clopidogrel for 6 months and aspirin indefinitely. The trial enrolled 150 patients and followed them for a mean of 14.4 (± 8.6) months. In that time, there were four strokes, five pericardial effusions, and six instances of device-related thrombus by transesophageal echocardiography. Three of the strokes were ischemic (1.7% per year), which is a 77% reduction from the expected rate of 7.3% based on the CHADS₂ scores of the patient cohort.

These data suggest that implantation of the Watchman device may be appropriate for those who cannot tolerate warfarin even in the short term.

The Lariat system

The Lariat suture delivery device (SentreHeart, Inc., Redwood City, CA) is approved by the US Food and Drug Administration (FDA) for soft-tissue closure and has been used for percutaneous left atrial appendage closure. It uses a magnet-tipped wire that is passed to the epicardial side of the left atrial appendage via pericardial access to meet a second magnet-tipped wire introduced into the appendage via transseptal access. A “lasso” is then advanced over the epicardial guide wire and tightened down around the ostium of the left atrial appendage. This tool facilitates deployment of a nonabsorbable polyester suture, which effectively ligates off the appendage from the rest of the left atrium (Figure 2).⁴⁰ In theory, the Lariat’s epicardial approach could eliminate the need for short- and long-term anticoagulation, as there would be no foreign body left within the heart.

In an initial cohort of 89 patients in Poland,41 the investigators reported a 96% closure rate as determined by transesophageal echocardiography immediately after the procedure. At 1-year follow-up, there was 98% complete closure, including cases of incomplete closure detected earlier.41 Adverse events were limited, with only two cases of severe pericarditis, two strokes, and one pericardial effusion. These results were replicated in the United States in a cohort of 25 patients, with a 100% closure rate and no stroke events.⁴²

There have been three published case reports of left atrial clot formation after successful left atrial appendage ligation using the Lariat device.^43–45 These experiences further emphasize that closure does not necessarily confer instant stroke prevention, and there remains a need to investigate the need for routine imaging and possibly periprocedural anticoagulation after ligation.

More recently, Pillai et al⁴⁶ published their initial experience following 71 patients with echocardiograms 3 months after left atrial appendage closure using the Lariat device. They reported leaks in 6 of the 71 patients; five of the leaks were successfully closed using the Amplatzer Septal Occluder, and one was closed with a repeat Lariat procedure.

Although the Lariat system has been used in more than 2,000 patients worldwide (SentreHeart, personal communication), there has been no published systematic comparison between it and oral anticoagulation to date.

AN EMERGING OPTION

Established guidelines help determine which patients with atrial fibrillation should receive oral anticoagulant therapy. For patients who have absolute contraindications to oral anticoagulants or who are undergoing cardiac surgery, surgical ligation of the left atrial appendage is an option. But for those with contraindications to oral anticoagulation in both the short term and the long term, there is a growing body of evidence suggesting that a percutaneous intervention is at least noninferior to—and in some cases is superior to—warfarin. Figure 3 shows our recommendations for the steps to determine which patients would be most appropriate to consider for left atrial appendage closure.

Holmes et al⁴⁷ propose that we may now have enough evidence to support an expedited regulatory approval process of these occlusion devices. But there are still a number of areas in which further investigation is clearly needed before left atrial appendage occlusion devices can be widely adopted.

The trials discussed above had specific inclusion and exclusion criteria, and therefore, although they support percutaneous intervention, the generalizability of their results remains in question. Indeed, the patients in PROTECT-AF³⁶ had an average CHADS₂ score of only 2.2. This study also included only patients who were able to tolerate both aspirin and clopidogrel simultaneously for a significant amount of time. Hence, one cannot assume the results would be the same in patients who have strict contraindications to warfarin or any target-specific oral anticoagulant. Concern regarding the generalizability of the conclusions from PROTECT-AF and PREVAIL has led to mixed votes (three assessments to date) from the FDA Circulatory Device Panel.⁴⁸

In an encouraging review of cases, Gafoor et al⁴⁹ reported safe and efficacious occlusion in octogenarians using the devices mentioned above. These patients often pose the greatest challenge in initiating long-term anticoagulation because of the many drug-drug interactions and the risk of intracranial hemorrhage secondary to falls.

Further, while occlusion devices would clearly be useful for patients in whom traditional oral anticoagulation is not an option, the newer oral anticoagulants might complicate the picture somewhat. As shown by Ruff et al,²¹ the risk-benefit ratio of these target-specific oral anticoagulants is quite favorable and by some measurements is superior to that of warfarin. Could there be a group of patients who cannot take warfarin but could instead do well on one of the newer anticoagulants, thus alleviating the need for percutaneous intervention? As the newer oral anticoagulants become more commonly used, the cost-benefit analysis of implanting an occlusion device could shift.

We expect that percutaneous closure will someday be a viable and equal option for stroke prevention

Lastly, in this era of high-value medical care, one must consider the cost-effectiveness of these novel interventions. As with any new technology, the up-front cost of implantation is certainly greater than that of warfarin therapy. If device implantation can prevent a hospitalization from a major bleed secondary to warfarin use or prevent a catastrophic stroke due to untreated atrial fibrillation, then the cost-benefit analysis may be tipped in the other direction. As these devices become more widely available and physicians have more experience implanting them, the costs will likely decrease.

As with oral anticoagulation therapy, all interventions, whether surgical or percutaneous, carry a risk of bleeding and stroke. There remains no substitute for frank and clear discussions between the physician and patient regarding the risks and benefits of each approach.

While a growing body of evidence surrounds left atrial appendage occlusion devices, many questions remain. Notably, could these devices be used in patients who can tolerate oral anticoagulants? And if so, which subgroups would benefit most? Does occlusion or ligation of the left atrial appendage affect electrical connections between it and the left atrium, thereby lowering the burden of atrial fibrillation?

We expect that continued investigation of and experience with left atrial appendage closure devices will position them one day as a viable and equal option for preventing stroke in patients with atrial fibrillation.

In some patients, symptoms of depres­sion, psychosis, delirium, or dementia exist concomitantly with, or as a result of, an abnormal (elevated or low) serum cal­cium concentration that has been precipi­tated by an unrecognized endocrinopathy. The apparent psychiatric presentations of such patients might reflect parathyroid pathology—not psychopathology.

Hypercalcemia and hypocalcemia often are related to a distinct spectrum of condi­tions, such as diseases of the parathyroid glands, kidneys, and various neoplasms including malignancies. Be alert to the pos­sibility of parathyroid disease in patients whose presentation suggests mental ill­ness concurrent with, or as a direct conse­quence of, an abnormal calcium level, and investigate appropriately.

The Table^1-9 illustrates how 3 clini­cal laboratory tests—serum calcium, serum parathyroid hormone (PTH), and phosphate—can narrow the differen­tial diagnosis when the clinical impres­sion is parathyroid-related illness. Seek endocrinology consultation whenever a parathyroid-associated ailment is discov­ered or suspected. Serum calcium is rou­tinely assayed in hospitalized patients; when managing a patient with treatment-refractory psychiatric illness, (1) always check the reported result of that test and (2) consider measuring PTH.

Case reports¹
Case 1: Woman with chronic depression. The patient was hospitalized while suicidal. Serial serum calcium levels were 12.5 mg/dL and 15.8 mg/dL (reference range, 8.2–10.2 mg/dL). The PTH level was elevated at 287 pg/mL (refer­ence range, 10–65 pg/mL).

After thyroid imaging, surgery revealed a parathyroid mass, which was resected. Histologic examination confirmed an adenoma.

The calcium concentration declined to 8.6 mg/dL postoperatively and stabilized at 9.2 mg/dL. Psychiatric symptoms resolved fully; she experienced a complete recovery.

Case 2: Man on long-term lithium mainte­nance. The patient was admitted in a delusional psychotic state. The serum calcium level was 14.3 mg/dL initially, decreasing to 11.5 mg/dL after lithium was discontinued. The PTH level was elevated at 97 pg/mL at admission, consis­tent with hyperparathyroidism.

A parathyroid adenoma was resected. Serum calcium level normalized at 10.7 mg/dL; psycho­sis resolved with striking, sustained improve­ment in mental status.

Full return to mental, physical health

The diagnosis of parathyroid adenoma in these 2 patients, which began with a psy­chiatric presentation, was properly made after an abnormal serum calcium level was documented. Surgical treatment of the endocrinopathy produced full remission and a return to normal mental and physi­cal health.

Although psychiatric manifestations are associated with an abnormal serum calcium concentration, the severity of those presen­tations does not correlate with the degree of abnormality of the calcium level.¹⁰

Disclosures
The authors report no financial relationships with any company whose products are mentioned in this article or with manufacturers of competing products.

In some patients, symptoms of depres­sion, psychosis, delirium, or dementia exist concomitantly with, or as a result of, an abnormal (elevated or low) serum cal­cium concentration that has been precipi­tated by an unrecognized endocrinopathy. The apparent psychiatric presentations of such patients might reflect parathyroid pathology—not psychopathology.

Hypercalcemia and hypocalcemia often are related to a distinct spectrum of condi­tions, such as diseases of the parathyroid glands, kidneys, and various neoplasms including malignancies. Be alert to the pos­sibility of parathyroid disease in patients whose presentation suggests mental ill­ness concurrent with, or as a direct conse­quence of, an abnormal calcium level, and investigate appropriately.

The Table^1-9 illustrates how 3 clini­cal laboratory tests—serum calcium, serum parathyroid hormone (PTH), and phosphate—can narrow the differen­tial diagnosis when the clinical impres­sion is parathyroid-related illness. Seek endocrinology consultation whenever a parathyroid-associated ailment is discov­ered or suspected. Serum calcium is rou­tinely assayed in hospitalized patients; when managing a patient with treatment-refractory psychiatric illness, (1) always check the reported result of that test and (2) consider measuring PTH.

Case reports¹
Case 1: Woman with chronic depression. The patient was hospitalized while suicidal. Serial serum calcium levels were 12.5 mg/dL and 15.8 mg/dL (reference range, 8.2–10.2 mg/dL). The PTH level was elevated at 287 pg/mL (refer­ence range, 10–65 pg/mL).

After thyroid imaging, surgery revealed a parathyroid mass, which was resected. Histologic examination confirmed an adenoma.

The calcium concentration declined to 8.6 mg/dL postoperatively and stabilized at 9.2 mg/dL. Psychiatric symptoms resolved fully; she experienced a complete recovery.

Case 2: Man on long-term lithium mainte­nance. The patient was admitted in a delusional psychotic state. The serum calcium level was 14.3 mg/dL initially, decreasing to 11.5 mg/dL after lithium was discontinued. The PTH level was elevated at 97 pg/mL at admission, consis­tent with hyperparathyroidism.

A parathyroid adenoma was resected. Serum calcium level normalized at 10.7 mg/dL; psycho­sis resolved with striking, sustained improve­ment in mental status.

Full return to mental, physical health

The diagnosis of parathyroid adenoma in these 2 patients, which began with a psy­chiatric presentation, was properly made after an abnormal serum calcium level was documented. Surgical treatment of the endocrinopathy produced full remission and a return to normal mental and physi­cal health.

Although psychiatric manifestations are associated with an abnormal serum calcium concentration, the severity of those presen­tations does not correlate with the degree of abnormality of the calcium level.¹⁰

Disclosures
The authors report no financial relationships with any company whose products are mentioned in this article or with manufacturers of competing products.

In some patients, symptoms of depres­sion, psychosis, delirium, or dementia exist concomitantly with, or as a result of, an abnormal (elevated or low) serum cal­cium concentration that has been precipi­tated by an unrecognized endocrinopathy. The apparent psychiatric presentations of such patients might reflect parathyroid pathology—not psychopathology.

Hypercalcemia and hypocalcemia often are related to a distinct spectrum of condi­tions, such as diseases of the parathyroid glands, kidneys, and various neoplasms including malignancies. Be alert to the pos­sibility of parathyroid disease in patients whose presentation suggests mental ill­ness concurrent with, or as a direct conse­quence of, an abnormal calcium level, and investigate appropriately.

The Table^1-9 illustrates how 3 clini­cal laboratory tests—serum calcium, serum parathyroid hormone (PTH), and phosphate—can narrow the differen­tial diagnosis when the clinical impres­sion is parathyroid-related illness. Seek endocrinology consultation whenever a parathyroid-associated ailment is discov­ered or suspected. Serum calcium is rou­tinely assayed in hospitalized patients; when managing a patient with treatment-refractory psychiatric illness, (1) always check the reported result of that test and (2) consider measuring PTH.

Case reports¹
Case 1: Woman with chronic depression. The patient was hospitalized while suicidal. Serial serum calcium levels were 12.5 mg/dL and 15.8 mg/dL (reference range, 8.2–10.2 mg/dL). The PTH level was elevated at 287 pg/mL (refer­ence range, 10–65 pg/mL).

After thyroid imaging, surgery revealed a parathyroid mass, which was resected. Histologic examination confirmed an adenoma.

The calcium concentration declined to 8.6 mg/dL postoperatively and stabilized at 9.2 mg/dL. Psychiatric symptoms resolved fully; she experienced a complete recovery.

Case 2: Man on long-term lithium mainte­nance. The patient was admitted in a delusional psychotic state. The serum calcium level was 14.3 mg/dL initially, decreasing to 11.5 mg/dL after lithium was discontinued. The PTH level was elevated at 97 pg/mL at admission, consis­tent with hyperparathyroidism.

A parathyroid adenoma was resected. Serum calcium level normalized at 10.7 mg/dL; psycho­sis resolved with striking, sustained improve­ment in mental status.

Full return to mental, physical health

The diagnosis of parathyroid adenoma in these 2 patients, which began with a psy­chiatric presentation, was properly made after an abnormal serum calcium level was documented. Surgical treatment of the endocrinopathy produced full remission and a return to normal mental and physi­cal health.

Although psychiatric manifestations are associated with an abnormal serum calcium concentration, the severity of those presen­tations does not correlate with the degree of abnormality of the calcium level.¹⁰

Disclosures
The authors report no financial relationships with any company whose products are mentioned in this article or with manufacturers of competing products.

Suicidologists and legal experts implore clinicians to document their suicide risk assessments (SRAs) thoroughly. It’s difficult, however, to find practical guid­ance on how to write a clinically sound, legally defensible SRA.

The crux of every SRA is written justifica­tion of suicide risk. That justification should reveal your thinking and present a well-reasoned basis for your decision.

Reasoned vs right
It’s more important to provide a justifica­tion of suicide risk that’s well-reasoned rather than one that’s right. Suicide is impossible to predict. Instead of predic­tion, legally we are asked to reasonably anticipate suicide based on clinical facts. In hindsight, especially in the context of a courtroom, decisions might look ill-considered. You need to craft a logical argu­ment, be clear, and avoid jargon.

Convey thoroughness by covering each component of an SRA. Use the mnemonic device CAIPS to help the reader (and you) understand how a conclusion was reached based on the facts of the case.

Chronic and Acute factors. Address the chronic and acute factors that weigh heavi­est in your mind. Chronic factors are condi­tions, past events, and demographics that generally do not change. Acute factors are recent events or conditions that potentially are modifiable. Pay attention to combina­tions of factors that dramatically elevate risk (eg, previous attempts in the context of acute depression). Avoid repeating every factor, especially when these are documented else­where, such as on a checklist.

Imminent warning signs for suicide. Address warning signs (Table 1),¹ the nature of current suicidal thoughts (Table 2), and other aspects of mental status (eg, future ori­entation) that influenced your decision. Use words like “moreover,” “however,” and “in addition” to draw the reader’s attention to the building blocks of your argument.

Protective factors. Discuss the protective factors last; they deserve the least weight because none has been shown to immunize people against suicide. Don’t solely rely on your judgment of what is protective (eg, chil­dren in the home). Instead, elicit the patient’s reasons for living and dying. Be concerned if he (she) reports more of the latter.

Summary statement. Make an explicit statement about risk, focusing on imminent risk (ie, the next few hours and days). Avoid a “plot twist,” which is a risk level inconsis­tent with the preceding evidence, because it suggests an error in judgment. The Box gives an example of a justification that follows the CAIPS method.

Additional tips
Consider these strategies:
   • Bolster your argument by explicitly addressing hopelessness (the strongest psy­chological correlate of suicide); use quotes from the patient that support your decision; refer to consultation with family members and colleagues; and include pertinent nega­tives to show completeness2 (ie, “denied sui­cide plans”).
   • Critically resolve discrepancies between what the patient says and behavior that suggests suicidal intent (eg, a patient who minimizes suicidal intent but shopped for a gun yesterday).
   • Last, while reviewing your justification, imagine that your patient completed suicide after leaving your office and that you are in court for negligence. In our experience, this exercise reveals dangerous errors of judg­ment. A clear and reasoned justification will reduce the risk of litigation and help you make prudent treatment plans.


Disclosures
The authors report no financial relationships with any companies whose products are mentioned in this article or with manufacturers of competing products.

Suicidologists and legal experts implore clinicians to document their suicide risk assessments (SRAs) thoroughly. It’s difficult, however, to find practical guid­ance on how to write a clinically sound, legally defensible SRA.

The crux of every SRA is written justifica­tion of suicide risk. That justification should reveal your thinking and present a well-reasoned basis for your decision.

Reasoned vs right
It’s more important to provide a justifica­tion of suicide risk that’s well-reasoned rather than one that’s right. Suicide is impossible to predict. Instead of predic­tion, legally we are asked to reasonably anticipate suicide based on clinical facts. In hindsight, especially in the context of a courtroom, decisions might look ill-considered. You need to craft a logical argu­ment, be clear, and avoid jargon.

Convey thoroughness by covering each component of an SRA. Use the mnemonic device CAIPS to help the reader (and you) understand how a conclusion was reached based on the facts of the case.

Chronic and Acute factors. Address the chronic and acute factors that weigh heavi­est in your mind. Chronic factors are condi­tions, past events, and demographics that generally do not change. Acute factors are recent events or conditions that potentially are modifiable. Pay attention to combina­tions of factors that dramatically elevate risk (eg, previous attempts in the context of acute depression). Avoid repeating every factor, especially when these are documented else­where, such as on a checklist.

Imminent warning signs for suicide. Address warning signs (Table 1),¹ the nature of current suicidal thoughts (Table 2), and other aspects of mental status (eg, future ori­entation) that influenced your decision. Use words like “moreover,” “however,” and “in addition” to draw the reader’s attention to the building blocks of your argument.

Protective factors. Discuss the protective factors last; they deserve the least weight because none has been shown to immunize people against suicide. Don’t solely rely on your judgment of what is protective (eg, chil­dren in the home). Instead, elicit the patient’s reasons for living and dying. Be concerned if he (she) reports more of the latter.

Summary statement. Make an explicit statement about risk, focusing on imminent risk (ie, the next few hours and days). Avoid a “plot twist,” which is a risk level inconsis­tent with the preceding evidence, because it suggests an error in judgment. The Box gives an example of a justification that follows the CAIPS method.

Additional tips
Consider these strategies:
   • Bolster your argument by explicitly addressing hopelessness (the strongest psy­chological correlate of suicide); use quotes from the patient that support your decision; refer to consultation with family members and colleagues; and include pertinent nega­tives to show completeness2 (ie, “denied sui­cide plans”).
   • Critically resolve discrepancies between what the patient says and behavior that suggests suicidal intent (eg, a patient who minimizes suicidal intent but shopped for a gun yesterday).
   • Last, while reviewing your justification, imagine that your patient completed suicide after leaving your office and that you are in court for negligence. In our experience, this exercise reveals dangerous errors of judg­ment. A clear and reasoned justification will reduce the risk of litigation and help you make prudent treatment plans.


Disclosures
The authors report no financial relationships with any companies whose products are mentioned in this article or with manufacturers of competing products.

Suicidologists and legal experts implore clinicians to document their suicide risk assessments (SRAs) thoroughly. It’s difficult, however, to find practical guid­ance on how to write a clinically sound, legally defensible SRA.

The crux of every SRA is written justifica­tion of suicide risk. That justification should reveal your thinking and present a well-reasoned basis for your decision.

Reasoned vs right
It’s more important to provide a justifica­tion of suicide risk that’s well-reasoned rather than one that’s right. Suicide is impossible to predict. Instead of predic­tion, legally we are asked to reasonably anticipate suicide based on clinical facts. In hindsight, especially in the context of a courtroom, decisions might look ill-considered. You need to craft a logical argu­ment, be clear, and avoid jargon.

Convey thoroughness by covering each component of an SRA. Use the mnemonic device CAIPS to help the reader (and you) understand how a conclusion was reached based on the facts of the case.

Chronic and Acute factors. Address the chronic and acute factors that weigh heavi­est in your mind. Chronic factors are condi­tions, past events, and demographics that generally do not change. Acute factors are recent events or conditions that potentially are modifiable. Pay attention to combina­tions of factors that dramatically elevate risk (eg, previous attempts in the context of acute depression). Avoid repeating every factor, especially when these are documented else­where, such as on a checklist.

Imminent warning signs for suicide. Address warning signs (Table 1),¹ the nature of current suicidal thoughts (Table 2), and other aspects of mental status (eg, future ori­entation) that influenced your decision. Use words like “moreover,” “however,” and “in addition” to draw the reader’s attention to the building blocks of your argument.

Protective factors. Discuss the protective factors last; they deserve the least weight because none has been shown to immunize people against suicide. Don’t solely rely on your judgment of what is protective (eg, chil­dren in the home). Instead, elicit the patient’s reasons for living and dying. Be concerned if he (she) reports more of the latter.

Summary statement. Make an explicit statement about risk, focusing on imminent risk (ie, the next few hours and days). Avoid a “plot twist,” which is a risk level inconsis­tent with the preceding evidence, because it suggests an error in judgment. The Box gives an example of a justification that follows the CAIPS method.

Additional tips
Consider these strategies:
   • Bolster your argument by explicitly addressing hopelessness (the strongest psy­chological correlate of suicide); use quotes from the patient that support your decision; refer to consultation with family members and colleagues; and include pertinent nega­tives to show completeness2 (ie, “denied sui­cide plans”).
   • Critically resolve discrepancies between what the patient says and behavior that suggests suicidal intent (eg, a patient who minimizes suicidal intent but shopped for a gun yesterday).
   • Last, while reviewing your justification, imagine that your patient completed suicide after leaving your office and that you are in court for negligence. In our experience, this exercise reveals dangerous errors of judg­ment. A clear and reasoned justification will reduce the risk of litigation and help you make prudent treatment plans.


Disclosures
The authors report no financial relationships with any companies whose products are mentioned in this article or with manufacturers of competing products.

Lisdexamfetamine, approved by the FDA in 2007 for attention-deficit/hyperactivity disorder (ADHD), has a new indication: binge eating disorder (BED) (Table 1). BED is characterized by recurrent episodes of consuming a large amount of food in a short time. A prodrug of amphet­amine, lisdexamfetamine is a Schedule-II controlled substance, with a high potential for abuse and the risk of severe psychologi­cal or physical dependence.

Lisdexamfetamine is not indicated for weight loss or obesity.

Dosage
For BED, the initial dosage of lisdexamfet­amine is 30 mg/d in the morning, titrated by 20 mg/d per week to the target dos­age of 50 to 70 mg/d. Maximum dosage is 70 mg/d. Morning dosing is recommended to avoid sleep disturbance.

Efficacy
The clinical efficacy of lisdexamfetamine was assessed in two 12-week parallel group, flexible-dose, placebo-controlled trials in adults with BED (age 18 to 55). Primary efficacy measure was the num­ber of binge days per week. Both studies had a 4-week dose-optimization period and an 8-week dose-maintenance period and followed the same dosage protocol. Patients began treatment at 30 mg/d and after 1 week were titrated to 50 mg/d; increases to 70 mg/d were made if clini­cally necessary and well tolerated. Patients were maintained on the optimized dos­age during the 8-week dose-maintenance period. A dosage of 30 mg/d did not produce a statistically significant effect, but 50 mg/d and 70 mg/d dosages were statistically superior to placebo. Patients taking lisdexamfetamine also had greater improvement on the Clinical Global Impression—Improvement scores, 4-week binge cessation, and greater reduction in the Yale-Brown Obsessive Compulsive Scale Modified for Binge Eating score.

The prescribing information does not state if lisdexamfetamine should be con­tinued long-term for treating BED.

Adverse reactions
In controlled trials, 5.1% of patients receiv­ing lisdexamfetamine for BED discontin­ued the drug because of an adverse event, compared with 2.4% of patients receiv­ing placebo. The most common adverse reactions in BED studies were dry mouth (36%), insomnia (20%), decreased appetite (8%), increased heart rate (8%), constipation (6%), and feeling jittery (6%). In trials of children, adolescents, and adults with ADHD, decreased appetite was more com­mon (39%, 34%, and 27%, respectively) than in BED trials (Table 2). Anaphylactic reac­tions, Stevens-Johnson syndrome, angio­edema, and urticaria have been described in postmarketing reports.

The safety of lisdexamfetamine for BED has not been studied in patients age <18, but has been studied in patients with ADHD.

Contraindications
Do not give lisdexamfetamine to patients who have a known hypersensitivity to amphetamine products or other ingredi­ents in lisdexamfetamine capsules.

Lisdexamfetamine is contraindicated in patients who are taking a monoamine oxi­dase inhibitor, because of a risk of hyper­tensive crisis.
 

Related Resources
• Wilens TE. Lisdexamfetamine for ADHD. Current Psychiatry. 2007;6(6):96-98,105.
• Peat CM, Brownley KA, Berkman ND, et al. Binge eating dis­order: evidence-based treatments. Current Psychiatry. 2012; 11(5):32-39.

Lisdexamfetamine, approved by the FDA in 2007 for attention-deficit/hyperactivity disorder (ADHD), has a new indication: binge eating disorder (BED) (Table 1). BED is characterized by recurrent episodes of consuming a large amount of food in a short time. A prodrug of amphet­amine, lisdexamfetamine is a Schedule-II controlled substance, with a high potential for abuse and the risk of severe psychologi­cal or physical dependence.

Lisdexamfetamine is not indicated for weight loss or obesity.

Dosage
For BED, the initial dosage of lisdexamfet­amine is 30 mg/d in the morning, titrated by 20 mg/d per week to the target dos­age of 50 to 70 mg/d. Maximum dosage is 70 mg/d. Morning dosing is recommended to avoid sleep disturbance.

Efficacy
The clinical efficacy of lisdexamfetamine was assessed in two 12-week parallel group, flexible-dose, placebo-controlled trials in adults with BED (age 18 to 55). Primary efficacy measure was the num­ber of binge days per week. Both studies had a 4-week dose-optimization period and an 8-week dose-maintenance period and followed the same dosage protocol. Patients began treatment at 30 mg/d and after 1 week were titrated to 50 mg/d; increases to 70 mg/d were made if clini­cally necessary and well tolerated. Patients were maintained on the optimized dos­age during the 8-week dose-maintenance period. A dosage of 30 mg/d did not produce a statistically significant effect, but 50 mg/d and 70 mg/d dosages were statistically superior to placebo. Patients taking lisdexamfetamine also had greater improvement on the Clinical Global Impression—Improvement scores, 4-week binge cessation, and greater reduction in the Yale-Brown Obsessive Compulsive Scale Modified for Binge Eating score.

The prescribing information does not state if lisdexamfetamine should be con­tinued long-term for treating BED.

Adverse reactions
In controlled trials, 5.1% of patients receiv­ing lisdexamfetamine for BED discontin­ued the drug because of an adverse event, compared with 2.4% of patients receiv­ing placebo. The most common adverse reactions in BED studies were dry mouth (36%), insomnia (20%), decreased appetite (8%), increased heart rate (8%), constipation (6%), and feeling jittery (6%). In trials of children, adolescents, and adults with ADHD, decreased appetite was more com­mon (39%, 34%, and 27%, respectively) than in BED trials (Table 2). Anaphylactic reac­tions, Stevens-Johnson syndrome, angio­edema, and urticaria have been described in postmarketing reports.

The safety of lisdexamfetamine for BED has not been studied in patients age <18, but has been studied in patients with ADHD.

Contraindications
Do not give lisdexamfetamine to patients who have a known hypersensitivity to amphetamine products or other ingredi­ents in lisdexamfetamine capsules.

Lisdexamfetamine is contraindicated in patients who are taking a monoamine oxi­dase inhibitor, because of a risk of hyper­tensive crisis.
 

Related Resources
• Wilens TE. Lisdexamfetamine for ADHD. Current Psychiatry. 2007;6(6):96-98,105.
• Peat CM, Brownley KA, Berkman ND, et al. Binge eating dis­order: evidence-based treatments. Current Psychiatry. 2012; 11(5):32-39.

Lisdexamfetamine, approved by the FDA in 2007 for attention-deficit/hyperactivity disorder (ADHD), has a new indication: binge eating disorder (BED) (Table 1). BED is characterized by recurrent episodes of consuming a large amount of food in a short time. A prodrug of amphet­amine, lisdexamfetamine is a Schedule-II controlled substance, with a high potential for abuse and the risk of severe psychologi­cal or physical dependence.

Lisdexamfetamine is not indicated for weight loss or obesity.

Dosage
For BED, the initial dosage of lisdexamfet­amine is 30 mg/d in the morning, titrated by 20 mg/d per week to the target dos­age of 50 to 70 mg/d. Maximum dosage is 70 mg/d. Morning dosing is recommended to avoid sleep disturbance.

Efficacy
The clinical efficacy of lisdexamfetamine was assessed in two 12-week parallel group, flexible-dose, placebo-controlled trials in adults with BED (age 18 to 55). Primary efficacy measure was the num­ber of binge days per week. Both studies had a 4-week dose-optimization period and an 8-week dose-maintenance period and followed the same dosage protocol. Patients began treatment at 30 mg/d and after 1 week were titrated to 50 mg/d; increases to 70 mg/d were made if clini­cally necessary and well tolerated. Patients were maintained on the optimized dos­age during the 8-week dose-maintenance period. A dosage of 30 mg/d did not produce a statistically significant effect, but 50 mg/d and 70 mg/d dosages were statistically superior to placebo. Patients taking lisdexamfetamine also had greater improvement on the Clinical Global Impression—Improvement scores, 4-week binge cessation, and greater reduction in the Yale-Brown Obsessive Compulsive Scale Modified for Binge Eating score.

The prescribing information does not state if lisdexamfetamine should be con­tinued long-term for treating BED.

Adverse reactions
In controlled trials, 5.1% of patients receiv­ing lisdexamfetamine for BED discontin­ued the drug because of an adverse event, compared with 2.4% of patients receiv­ing placebo. The most common adverse reactions in BED studies were dry mouth (36%), insomnia (20%), decreased appetite (8%), increased heart rate (8%), constipation (6%), and feeling jittery (6%). In trials of children, adolescents, and adults with ADHD, decreased appetite was more com­mon (39%, 34%, and 27%, respectively) than in BED trials (Table 2). Anaphylactic reac­tions, Stevens-Johnson syndrome, angio­edema, and urticaria have been described in postmarketing reports.

The safety of lisdexamfetamine for BED has not been studied in patients age <18, but has been studied in patients with ADHD.

Contraindications
Do not give lisdexamfetamine to patients who have a known hypersensitivity to amphetamine products or other ingredi­ents in lisdexamfetamine capsules.

Lisdexamfetamine is contraindicated in patients who are taking a monoamine oxi­dase inhibitor, because of a risk of hyper­tensive crisis.
 

Related Resources
• Wilens TE. Lisdexamfetamine for ADHD. Current Psychiatry. 2007;6(6):96-98,105.
• Peat CM, Brownley KA, Berkman ND, et al. Binge eating dis­order: evidence-based treatments. Current Psychiatry. 2012; 11(5):32-39.

Too few psychiatrists. A growing number of patients. A new federal law, technological advances, and a genera­tional shift in the way people communicate. Add them together and you have the perfect environment for telepsy­chiatry—the remote practice of psychiatry by means of tele­medicine—to take root (Box 1). Although telepsychiatry has, in various forms, been around since the 1950s,¹ only recently has it expanded into almost all areas of psychiatric practice.

Here are some observations from my daily work on why I see this method of delivering mental health care is poised to expand in 2015 and beyond. Does telepsychiatry make sense for you?

Lack of supply is a big driver
There are simply not enough psychiatrists where they are needed, which is the primary driver of the expansion of telepsychiatry. With 77% of counties in the United States reporting a shortage of psychiatrists² and the “graying” of the psychiatric workforce,³ a more efficient way to make use of a psychiatrist’s time is needed. Telepsychiatry elimi­nates travel time and allows psychiatrists to visit distant sites virtually.

The shortage of psychiatric practitioners that we see today is only going to become worse. The Patient Protection and Affordable Care Act of 2010 includes mental health care and substance abuse treatment among its 10 essential benefits; just as important, new rules arising from the Mental Health Parity and Addiction Equity Act of 2008 limit restrictions on access to mental health care when insurance provides such cover­age.⁴ These legislative initiatives likely will lead to increased demand for psychiatrists in all care settings—from outpatient consults to acute inpatient admissions.


Why so attractive an option?
The shortage of psychiatrists creates limita­tions on access to care. Fortunately, telemed­icine has entered a new age, ushered in by widely available teleconferencing technol­ogy. Specialists from dermatology to surgery currently are using telemedicine; psychia­try is a good fit for telemedicine because of (1) the limited amount of “touch” required to make a psychiatric assessment, (2) signifi­cant improvements in video quality in recent years, and (3) a decrease in the stigma associ­ated with visiting a psychiatrist.

A generation raised on the Internet is entering the health care marketplace. These consumers and clinicians are accustomed to using video for many daily activities, and they seek health information from the Web. Visiting a psychiatrist through teleconferenc­ing isn’t strange or alienating to this genera­tion; their comfort with technology allows them to have intimate exchanges on video.

Subspecialty particulars
The earliest adopters, not surprisingly, are in areas where the strain of shortage has been felt most, with pediatric, geriatric, and correctional psychiatrists leading the way. In these fields, a substantial literature supports the use of telepsychiatry from a number of practice perspectives.

Pediatric psychiatry. The literature shows that children, families, and clinicians are, on the whole, satisfied with telepsychia­try.⁵ Children and adolescents who have been shown to benefit from telepsychia­try include those with depression,⁶ post­traumatic stress disorder, and eating disorders.⁷ Based on a case series, some authors have asserted that telepsychiatry might be preferable to in-person treatment (Box 2).⁸

Correctional psychiatry. Clinicians work­ing in correctional psychiatry have been at the forefront of experimentation with tele­psychiatry. The technology is a natural fit for this setting:  
   • Prisons often are located in remote locations.  
   • Psychiatrists can be reluctant to pro­vide on-site services because of safety concerns.

With correctional telepsychiatry, not only are patient outcomes comparable with in-person psychiatry, but the cost of delivering care can be significantly lower.¹¹ With the U.S. Department of Justice reporting that 50% of inmates have a diagnosable mental disorder, including substance abuse,¹² the need for access to a psychiatrist in the cor­rectional system is acute.

Telepsychiatry can confidently be pro­vided in a number of settings:  
   • emergency rooms  
   • nursing homes  
   • offices of primary care physicians  
   • in-home care.

Clinical services in these settings have been offered, studied, and reviewed.¹³

Can confidentiality and security be assured?
As with any new medical tool, the risk and benefits must be weighed care­ fully. The most obvious risk is to privacy. Telepsychiatry visits, like all patient encounters, must be secure and confiden­tial. Given the growing suspicion among the public and professionals who use com­puters that all data are at risk, clinicians must take appropriate cautions and, at the same time, warn patients of the risks. Readily available videoconferencing soft­ware, such as Skype, does not provide the level of security that patients expect from health care providers.¹⁴

Other common concerns about telepsy­chiatry are stable access to videoconferenc­ing and the safety from hackers of necessary hardware. Medical device companies have created hardware and software for use in telepsychiatry that provide a Health Insurance Portability and Accountability Act-compliant high-quality, stable, video­conferencing visit.

Do patients benefit?
Clinically, patients have fared well when they receive care through telepsychiatry. In some studies, however, clinicians have expressed some dissatisfaction with the technology¹³— understandable, given the value that psychi­atry traditionally has put on sitting with the patient. As Knoedler¹⁵ described it, making the switch to telepsychiatry from in-person contact can engender loneliness in some phy­sicians; not only is patient contact shifted to videoconferencing, but the psychiatrist loses the supportive environment of a busy clinical practice. Knoedler also pointed out that, on the other hand, telepsychiatry offers practi­tioners the opportunity to evaluate and treat people who otherwise would not have men­tal health care.

Obstacles—practical, knotty ones
Reimbursement and licensing. These are 2 pressing problems of telepsychiatry, although recent policy developments will help expand telepsychiatry and make it more appealing to physicians:
   • Medicare reimburses for telepsychiatry in non-metropolitan areas.
   • In 41 states, Medicaid has included tele­psychiatry as a benefit.¹⁶
   • Nine states offer a specific medical license for practicing telepsychiatry¹⁷ (in the remaining states, a full medical license must be obtained before one can provide telemedi­cine services).
   • The Joint Commission has included lan­guage in its regulations that could expedite privileging of telepsychiatrists.¹⁸

Even with such advancements, problems with licensure, credentialing, privacy, secu­rity, confidentiality, informed consent, and professional liability remain.¹⁹ I urge you to do your research on these key areas before plunging in.

Changes to models of care. The risk that telepsychiatry poses to various models of care has to be considered. Telepsychiatry is a dramatic innovation, but it should be used to support only high-quality, evidence-based care to which patients are entitled.²⁰ With new technology—as with new medi­cations—use must be carefully monitored and scrutinized.

Although evidence of the value of telepsy­chiatry is growing, many methods of long-distance practice are still in their infancy. Data must be collected and poor outcomes assessed honestly to ensure that the “more-good-than-harm” mandate is met.
 

Good reasons to call this shift ‘inevitable’
The future of telepsychiatry includes expansion into new areas of practice. The move to providing services to patients where they happen to be—at work or home— seems inevitable:
   • In rural areas, practitioners can com­municate with patients so that they are cared for in their homes, without the expense of transportation.
   • Employers can invest in workplace health clinics that use telemedicine ser­vices to reduce absenteeism.
   • For psychiatrists, the ability to provide services to patients across a wide region, from a single convenient location, and at lower cost is an attractive prospect.

To conclude: telepsychiatry holds potential to provide greater reimburse­ment and improved quality of life for psy­chiatrists and patients: It allows physicians to choose where they live and work, and limits the number of unreimbursed com­mutes, and gives patients access to psychi­atric care locally, without disruptive travel and delays.

Bottom Line
The exchange of medical information from 1 site to another by means of electronic communication has great potential to improve the health of patients and to alleviate the shortage of psychiatric practitioners across regions and settings. Pediatric, geriatric, and correctional psychiatry stand to benefit because of the nature of the patients and locations.

Related Resources
• American Telemedicine Association. Practice guidelines for video-based online mental health services. http://www. americantelemed.org/docs/default-source/standards/practice-guidelines-for-video-based-online-mental-health-services. pdf?sfvrsn=6. Published May 2013. Accessed February 10, 2015.
• Freudenberg N, Yellowlees PM. Telepsychiatry as part of a com­prehensive care plan. Virtual Mentor. 2014;16(12):964-968.
• Kornbluh R. Telepsychiatry is a tool that we must exploit. Clinical Psychiatry News. August 7, 2014. http://www. clinicalpsychiatrynews.com/home/article/telepsychiatry-is-a-tool-that-we-must-exploit/28c87bec298e0aa208309fa 9bc48dedc.html.
• University of Colorado Denver. Telemental Health Guide. http:// www.tmhguide.org.

Disclosure
Dr. Kornbluh reports no financial relationships with any company whose products are mentioned in this article or with manufacturers of competing products.

Too few psychiatrists. A growing number of patients. A new federal law, technological advances, and a genera­tional shift in the way people communicate. Add them together and you have the perfect environment for telepsy­chiatry—the remote practice of psychiatry by means of tele­medicine—to take root (Box 1). Although telepsychiatry has, in various forms, been around since the 1950s,¹ only recently has it expanded into almost all areas of psychiatric practice.

Here are some observations from my daily work on why I see this method of delivering mental health care is poised to expand in 2015 and beyond. Does telepsychiatry make sense for you?

Lack of supply is a big driver
There are simply not enough psychiatrists where they are needed, which is the primary driver of the expansion of telepsychiatry. With 77% of counties in the United States reporting a shortage of psychiatrists² and the “graying” of the psychiatric workforce,³ a more efficient way to make use of a psychiatrist’s time is needed. Telepsychiatry elimi­nates travel time and allows psychiatrists to visit distant sites virtually.

The shortage of psychiatric practitioners that we see today is only going to become worse. The Patient Protection and Affordable Care Act of 2010 includes mental health care and substance abuse treatment among its 10 essential benefits; just as important, new rules arising from the Mental Health Parity and Addiction Equity Act of 2008 limit restrictions on access to mental health care when insurance provides such cover­age.⁴ These legislative initiatives likely will lead to increased demand for psychiatrists in all care settings—from outpatient consults to acute inpatient admissions.


Why so attractive an option?
The shortage of psychiatrists creates limita­tions on access to care. Fortunately, telemed­icine has entered a new age, ushered in by widely available teleconferencing technol­ogy. Specialists from dermatology to surgery currently are using telemedicine; psychia­try is a good fit for telemedicine because of (1) the limited amount of “touch” required to make a psychiatric assessment, (2) signifi­cant improvements in video quality in recent years, and (3) a decrease in the stigma associ­ated with visiting a psychiatrist.

A generation raised on the Internet is entering the health care marketplace. These consumers and clinicians are accustomed to using video for many daily activities, and they seek health information from the Web. Visiting a psychiatrist through teleconferenc­ing isn’t strange or alienating to this genera­tion; their comfort with technology allows them to have intimate exchanges on video.

Subspecialty particulars
The earliest adopters, not surprisingly, are in areas where the strain of shortage has been felt most, with pediatric, geriatric, and correctional psychiatrists leading the way. In these fields, a substantial literature supports the use of telepsychiatry from a number of practice perspectives.

Pediatric psychiatry. The literature shows that children, families, and clinicians are, on the whole, satisfied with telepsychia­try.⁵ Children and adolescents who have been shown to benefit from telepsychia­try include those with depression,⁶ post­traumatic stress disorder, and eating disorders.⁷ Based on a case series, some authors have asserted that telepsychiatry might be preferable to in-person treatment (Box 2).⁸

Correctional psychiatry. Clinicians work­ing in correctional psychiatry have been at the forefront of experimentation with tele­psychiatry. The technology is a natural fit for this setting:  
   • Prisons often are located in remote locations.  
   • Psychiatrists can be reluctant to pro­vide on-site services because of safety concerns.

With correctional telepsychiatry, not only are patient outcomes comparable with in-person psychiatry, but the cost of delivering care can be significantly lower.¹¹ With the U.S. Department of Justice reporting that 50% of inmates have a diagnosable mental disorder, including substance abuse,¹² the need for access to a psychiatrist in the cor­rectional system is acute.

Telepsychiatry can confidently be pro­vided in a number of settings:  
   • emergency rooms  
   • nursing homes  
   • offices of primary care physicians  
   • in-home care.

Clinical services in these settings have been offered, studied, and reviewed.¹³

Can confidentiality and security be assured?
As with any new medical tool, the risk and benefits must be weighed care­ fully. The most obvious risk is to privacy. Telepsychiatry visits, like all patient encounters, must be secure and confiden­tial. Given the growing suspicion among the public and professionals who use com­puters that all data are at risk, clinicians must take appropriate cautions and, at the same time, warn patients of the risks. Readily available videoconferencing soft­ware, such as Skype, does not provide the level of security that patients expect from health care providers.¹⁴

Other common concerns about telepsy­chiatry are stable access to videoconferenc­ing and the safety from hackers of necessary hardware. Medical device companies have created hardware and software for use in telepsychiatry that provide a Health Insurance Portability and Accountability Act-compliant high-quality, stable, video­conferencing visit.

Do patients benefit?
Clinically, patients have fared well when they receive care through telepsychiatry. In some studies, however, clinicians have expressed some dissatisfaction with the technology¹³— understandable, given the value that psychi­atry traditionally has put on sitting with the patient. As Knoedler¹⁵ described it, making the switch to telepsychiatry from in-person contact can engender loneliness in some phy­sicians; not only is patient contact shifted to videoconferencing, but the psychiatrist loses the supportive environment of a busy clinical practice. Knoedler also pointed out that, on the other hand, telepsychiatry offers practi­tioners the opportunity to evaluate and treat people who otherwise would not have men­tal health care.

Obstacles—practical, knotty ones
Reimbursement and licensing. These are 2 pressing problems of telepsychiatry, although recent policy developments will help expand telepsychiatry and make it more appealing to physicians:
   • Medicare reimburses for telepsychiatry in non-metropolitan areas.
   • In 41 states, Medicaid has included tele­psychiatry as a benefit.¹⁶
   • Nine states offer a specific medical license for practicing telepsychiatry¹⁷ (in the remaining states, a full medical license must be obtained before one can provide telemedi­cine services).
   • The Joint Commission has included lan­guage in its regulations that could expedite privileging of telepsychiatrists.¹⁸

Even with such advancements, problems with licensure, credentialing, privacy, secu­rity, confidentiality, informed consent, and professional liability remain.¹⁹ I urge you to do your research on these key areas before plunging in.

Changes to models of care. The risk that telepsychiatry poses to various models of care has to be considered. Telepsychiatry is a dramatic innovation, but it should be used to support only high-quality, evidence-based care to which patients are entitled.²⁰ With new technology—as with new medi­cations—use must be carefully monitored and scrutinized.

Although evidence of the value of telepsy­chiatry is growing, many methods of long-distance practice are still in their infancy. Data must be collected and poor outcomes assessed honestly to ensure that the “more-good-than-harm” mandate is met.
 

Good reasons to call this shift ‘inevitable’
The future of telepsychiatry includes expansion into new areas of practice. The move to providing services to patients where they happen to be—at work or home— seems inevitable:
   • In rural areas, practitioners can com­municate with patients so that they are cared for in their homes, without the expense of transportation.
   • Employers can invest in workplace health clinics that use telemedicine ser­vices to reduce absenteeism.
   • For psychiatrists, the ability to provide services to patients across a wide region, from a single convenient location, and at lower cost is an attractive prospect.

To conclude: telepsychiatry holds potential to provide greater reimburse­ment and improved quality of life for psy­chiatrists and patients: It allows physicians to choose where they live and work, and limits the number of unreimbursed com­mutes, and gives patients access to psychi­atric care locally, without disruptive travel and delays.

Bottom Line
The exchange of medical information from 1 site to another by means of electronic communication has great potential to improve the health of patients and to alleviate the shortage of psychiatric practitioners across regions and settings. Pediatric, geriatric, and correctional psychiatry stand to benefit because of the nature of the patients and locations.

Related Resources
• American Telemedicine Association. Practice guidelines for video-based online mental health services. http://www. americantelemed.org/docs/default-source/standards/practice-guidelines-for-video-based-online-mental-health-services. pdf?sfvrsn=6. Published May 2013. Accessed February 10, 2015.
• Freudenberg N, Yellowlees PM. Telepsychiatry as part of a com­prehensive care plan. Virtual Mentor. 2014;16(12):964-968.
• Kornbluh R. Telepsychiatry is a tool that we must exploit. Clinical Psychiatry News. August 7, 2014. http://www. clinicalpsychiatrynews.com/home/article/telepsychiatry-is-a-tool-that-we-must-exploit/28c87bec298e0aa208309fa 9bc48dedc.html.
• University of Colorado Denver. Telemental Health Guide. http:// www.tmhguide.org.

Disclosure
Dr. Kornbluh reports no financial relationships with any company whose products are mentioned in this article or with manufacturers of competing products.

Too few psychiatrists. A growing number of patients. A new federal law, technological advances, and a genera­tional shift in the way people communicate. Add them together and you have the perfect environment for telepsy­chiatry—the remote practice of psychiatry by means of tele­medicine—to take root (Box 1). Although telepsychiatry has, in various forms, been around since the 1950s,¹ only recently has it expanded into almost all areas of psychiatric practice.

Here are some observations from my daily work on why I see this method of delivering mental health care is poised to expand in 2015 and beyond. Does telepsychiatry make sense for you?

Lack of supply is a big driver
There are simply not enough psychiatrists where they are needed, which is the primary driver of the expansion of telepsychiatry. With 77% of counties in the United States reporting a shortage of psychiatrists² and the “graying” of the psychiatric workforce,³ a more efficient way to make use of a psychiatrist’s time is needed. Telepsychiatry elimi­nates travel time and allows psychiatrists to visit distant sites virtually.

The shortage of psychiatric practitioners that we see today is only going to become worse. The Patient Protection and Affordable Care Act of 2010 includes mental health care and substance abuse treatment among its 10 essential benefits; just as important, new rules arising from the Mental Health Parity and Addiction Equity Act of 2008 limit restrictions on access to mental health care when insurance provides such cover­age.⁴ These legislative initiatives likely will lead to increased demand for psychiatrists in all care settings—from outpatient consults to acute inpatient admissions.


Why so attractive an option?
The shortage of psychiatrists creates limita­tions on access to care. Fortunately, telemed­icine has entered a new age, ushered in by widely available teleconferencing technol­ogy. Specialists from dermatology to surgery currently are using telemedicine; psychia­try is a good fit for telemedicine because of (1) the limited amount of “touch” required to make a psychiatric assessment, (2) signifi­cant improvements in video quality in recent years, and (3) a decrease in the stigma associ­ated with visiting a psychiatrist.

A generation raised on the Internet is entering the health care marketplace. These consumers and clinicians are accustomed to using video for many daily activities, and they seek health information from the Web. Visiting a psychiatrist through teleconferenc­ing isn’t strange or alienating to this genera­tion; their comfort with technology allows them to have intimate exchanges on video.

Subspecialty particulars
The earliest adopters, not surprisingly, are in areas where the strain of shortage has been felt most, with pediatric, geriatric, and correctional psychiatrists leading the way. In these fields, a substantial literature supports the use of telepsychiatry from a number of practice perspectives.

Pediatric psychiatry. The literature shows that children, families, and clinicians are, on the whole, satisfied with telepsychia­try.⁵ Children and adolescents who have been shown to benefit from telepsychia­try include those with depression,⁶ post­traumatic stress disorder, and eating disorders.⁷ Based on a case series, some authors have asserted that telepsychiatry might be preferable to in-person treatment (Box 2).⁸

Correctional psychiatry. Clinicians work­ing in correctional psychiatry have been at the forefront of experimentation with tele­psychiatry. The technology is a natural fit for this setting:  
   • Prisons often are located in remote locations.  
   • Psychiatrists can be reluctant to pro­vide on-site services because of safety concerns.

With correctional telepsychiatry, not only are patient outcomes comparable with in-person psychiatry, but the cost of delivering care can be significantly lower.¹¹ With the U.S. Department of Justice reporting that 50% of inmates have a diagnosable mental disorder, including substance abuse,¹² the need for access to a psychiatrist in the cor­rectional system is acute.

Telepsychiatry can confidently be pro­vided in a number of settings:  
   • emergency rooms  
   • nursing homes  
   • offices of primary care physicians  
   • in-home care.

Clinical services in these settings have been offered, studied, and reviewed.¹³

Can confidentiality and security be assured?
As with any new medical tool, the risk and benefits must be weighed care­ fully. The most obvious risk is to privacy. Telepsychiatry visits, like all patient encounters, must be secure and confiden­tial. Given the growing suspicion among the public and professionals who use com­puters that all data are at risk, clinicians must take appropriate cautions and, at the same time, warn patients of the risks. Readily available videoconferencing soft­ware, such as Skype, does not provide the level of security that patients expect from health care providers.¹⁴

Other common concerns about telepsy­chiatry are stable access to videoconferenc­ing and the safety from hackers of necessary hardware. Medical device companies have created hardware and software for use in telepsychiatry that provide a Health Insurance Portability and Accountability Act-compliant high-quality, stable, video­conferencing visit.

Do patients benefit?
Clinically, patients have fared well when they receive care through telepsychiatry. In some studies, however, clinicians have expressed some dissatisfaction with the technology¹³— understandable, given the value that psychi­atry traditionally has put on sitting with the patient. As Knoedler¹⁵ described it, making the switch to telepsychiatry from in-person contact can engender loneliness in some phy­sicians; not only is patient contact shifted to videoconferencing, but the psychiatrist loses the supportive environment of a busy clinical practice. Knoedler also pointed out that, on the other hand, telepsychiatry offers practi­tioners the opportunity to evaluate and treat people who otherwise would not have men­tal health care.

Obstacles—practical, knotty ones
Reimbursement and licensing. These are 2 pressing problems of telepsychiatry, although recent policy developments will help expand telepsychiatry and make it more appealing to physicians:
   • Medicare reimburses for telepsychiatry in non-metropolitan areas.
   • In 41 states, Medicaid has included tele­psychiatry as a benefit.¹⁶
   • Nine states offer a specific medical license for practicing telepsychiatry¹⁷ (in the remaining states, a full medical license must be obtained before one can provide telemedi­cine services).
   • The Joint Commission has included lan­guage in its regulations that could expedite privileging of telepsychiatrists.¹⁸

Even with such advancements, problems with licensure, credentialing, privacy, secu­rity, confidentiality, informed consent, and professional liability remain.¹⁹ I urge you to do your research on these key areas before plunging in.

Changes to models of care. The risk that telepsychiatry poses to various models of care has to be considered. Telepsychiatry is a dramatic innovation, but it should be used to support only high-quality, evidence-based care to which patients are entitled.²⁰ With new technology—as with new medi­cations—use must be carefully monitored and scrutinized.

Although evidence of the value of telepsy­chiatry is growing, many methods of long-distance practice are still in their infancy. Data must be collected and poor outcomes assessed honestly to ensure that the “more-good-than-harm” mandate is met.
 

Good reasons to call this shift ‘inevitable’
The future of telepsychiatry includes expansion into new areas of practice. The move to providing services to patients where they happen to be—at work or home— seems inevitable:
   • In rural areas, practitioners can com­municate with patients so that they are cared for in their homes, without the expense of transportation.
   • Employers can invest in workplace health clinics that use telemedicine ser­vices to reduce absenteeism.
   • For psychiatrists, the ability to provide services to patients across a wide region, from a single convenient location, and at lower cost is an attractive prospect.

To conclude: telepsychiatry holds potential to provide greater reimburse­ment and improved quality of life for psy­chiatrists and patients: It allows physicians to choose where they live and work, and limits the number of unreimbursed com­mutes, and gives patients access to psychi­atric care locally, without disruptive travel and delays.

Bottom Line
The exchange of medical information from 1 site to another by means of electronic communication has great potential to improve the health of patients and to alleviate the shortage of psychiatric practitioners across regions and settings. Pediatric, geriatric, and correctional psychiatry stand to benefit because of the nature of the patients and locations.

Related Resources
• American Telemedicine Association. Practice guidelines for video-based online mental health services. http://www. americantelemed.org/docs/default-source/standards/practice-guidelines-for-video-based-online-mental-health-services. pdf?sfvrsn=6. Published May 2013. Accessed February 10, 2015.
• Freudenberg N, Yellowlees PM. Telepsychiatry as part of a com­prehensive care plan. Virtual Mentor. 2014;16(12):964-968.
• Kornbluh R. Telepsychiatry is a tool that we must exploit. Clinical Psychiatry News. August 7, 2014. http://www. clinicalpsychiatrynews.com/home/article/telepsychiatry-is-a-tool-that-we-must-exploit/28c87bec298e0aa208309fa 9bc48dedc.html.
• University of Colorado Denver. Telemental Health Guide. http:// www.tmhguide.org.

Disclosure
Dr. Kornbluh reports no financial relationships with any company whose products are mentioned in this article or with manufacturers of competing products.

“Mr. Smith seems somewhat confused today” is one of the most serious and concerning pre-visit reports you can receive from your staff or the patient’s family. Such a descriptor can be confusing—pardon the pun—not only for the patient, but to even seasoned mental health providers.

The term confusion can be code for diagnoses ranging from delirium^a to a progressive neurocognitive disorder (NCD) such as major NCD due to Alzheimer’s disease (AD), or even a more challenging prob­lem such as beclouded dementia (delirium superimposed on demen­tia/NCD). It is essential for all mental health professionals to have an evidence-based approach when encountering signs or symptoms of confusion.

^aICD-10 code R41.0 encompasses Confusion, Other Specified Delirium, or Unspecified Delirium.

CASE REPORT
Ms. T, age 62, has hypothyroidism and bipolar I disorder, most recently depressed, with comorbid generalized anxiety disorder. She has been tak­ing lithium, 600 mg/d, to control her mood symptoms. Her daughter-in-law reports that Ms. T has been exhibiting increasing signs of confusion. During the office evaluation, Ms. T minimizes her symptoms, only describ­ing mild issues with forgetfulness while cooking and concern over increas­ing anxiety. Her daughter-in-law plays a voicemail message from earlier in the week, in which Ms. T’s speech is halting, disorganized, and in a word, confused. I decide to use the mnemonic decision chart MR. MIND (Table 1) to get to the bottom of her recent confusion.

Measure cognition
It is nice to receive advanced warning about a cognitive change or a change in activities of daily living; however, many patients present with subtle, sub-acute changes that are more difficult to assess. When encountering a broad symptom such as “confusion”—which has an equally broad differential diagnosis—systematic assess­ment of the current cognitive state com­pared with the patient’s baseline becomes the first order of business. However, this requires that the patient has had a baseline cognitive assessment.

In my practice, I often administer one of the validated neurocognitive screening instruments when a patient first begins care—even a brief test such as the Mini- Cog (3-item recall plus clock drawing test), which is comparable to longer screening tests at least for NCD/dementia.¹ During a presentation for confusion, a more detailed neurocognitive assessment instrument would be recommended, allowing one to marry the clinical impression with a validated, objective measure. Formal neu­ropsychological testing by a clinical neuro­psychologist is the gold standard, but such testing is time-consuming and expensive and often not readily available. The screen­ing instrument I use for a more thorough evaluation depends on the clinical scenario.

The Six-Item Screener is used in some emergency settings because it is short but boasts a higher sensitivity than the Mini- Cog (94% vs 75%) with similar specificity when screening for cognitive impairment.² The Mini-Mental State Examination (MMSE) is a valuable instrument, although, recently, the Saint Louis University Mental Status Examination has been thought to be better at detecting mild NCD than the MMSE; more data are needed to substan­tiate this claim.³ The Montreal Cognitive Assessment is another validated screening tool that has been shown to be superior to the MMSE in terms of screening for mild cognitive impairment.⁴ The best delirium-specific assessment tool is the Confusion Assessment Method (Table 2).⁵

Review medications
A review of the medication list is not just a Joint Commission mandate (medication reconciliation during each encounter) but is important whenever confusion is noted. Polypharmacy can be a concern, but is not as concerning as the class of medication prescribed, particularly anticholinergic and sedative medications in patients age >65. The Drug Burden Index can be helpful in assessing this risk.⁶ Medications such as the benzodiazepine-receptor agonists, tri­cyclic antidepressants, and antipsychotics should be discontinued if possible, keeping in mind that the addition or subtraction of medications must be done prudently and only after reviewing the evidence and in consultation with the patient. A detailed medication review is as important for confused outpatients as it is for an inpatient case (steps 2 and 3 of the inpatient algo­rithm outlined in Table 3).⁷

In Ms. T’s case, the primary concern on her medication list was that her medical team was prescribing levothyroxine, 112 mcg/d, and desiccated thyroid (combination thy­roxine and triiodothyronine in the form of 20 mg Armour Thyroid), despite a lack of data for such combination therapy. Earlier, I had discontinued lorazepam, leaving lithium, 600 mg/d, quetiapine, 400 mg/d, and escitalopram, 10 mg/d, as her remain­ing psychotropics. Her other medications included atorvastatin, 40 mg/d, for hyper-lipidemia and metformin, 750 mg/d, for type 2 diabetes mellitus.

Medical illness
An organic basis must rank high in the dif­ferential diagnosis if medications are not the culprit. There are myriad medical disorders that can lead to confusion (Table 4).⁸ In an outpatient psychiatric setting, labo­ratory and radiology testing might not be readily available. It then becomes impor­tant to collaborate with a patient’s medical team if any of the following are met:
   •there is high suspicion of a medical cause
   •there could be delays in performing a medical workup
   •a physical examination is needed.

Other laboratory testing could be valu­able depending on the clinical scenario. These include tests such as:
   •drug level monitoring (lithium, val­proic acid, etc.) to assess for toxicity
   •HIV and rapid plasma reagin for sus­pected sexually transmitted infections
   •vitamin levels in patients with poor nutrition or post bariatric surgery
   •erythrocyte sedimentation rate or C-reactive protein, or both, if there are signs of inflammation
   •bacterial culture if blood or tissue infec­tion is a concern.

Esoteric tests include ceruloplasmin (Wilson’s disease), heavy metals screen, and even tests such as anti-gliadin anti­bodies because the prevalence of gluten sensitivity and celiac disease appear to be on the rise and have been associated with neuropsychiatric problems including encephalopathy.¹⁰

Brain imaging is an important consider­ation when a medical differential diagnosis for confusion is formulated. Unfortunately, there is little evidence-based guidance as to when brain imaging should be performed, often leading to overuse of tests such as CT, especially in emergency settings when con­fusion is noted. From a clinical standpoint, a head CT scan often is best ordered for patients who demonstrate an acute change in mental status, are age >70, are receiving anticoagulation, or have sustained trauma to the head. The key concern would be intracranial hemorrhage. However, some data suggest that the best use of head CT is for patients who have an impaired level of consciousness or a new focal neurologic deficit.¹¹

Apart from more acute changes, a brain MRI study is more helpful than a head CT when evaluating the brain parenchyma for more sub-acute diagnoses such as multiple sclerosis or a brain tumor. T2-weighted hyperintensities seen on an MRI are thought to predict an increased risk of stroke, dementia, and death.

Their discov­ery should prompt a detailed evaluation for risk factors of stroke and dementia/NCD.¹²

In Ms. T’s case, she was taking lithium, so it was logical to obtain a trough lith­ium level 12 hours after the last dose and to check kidney function (serum creati­nine to estimate the glomerular filtration rate), which were in the therapeutic/nor­mal range. Her serum lithium level was 0.7 mEq/L. Brain imaging was not ordered, but several other labs (CMP, CBC, hemoglo­bin A1c [HgbA1c], and TSH) were drawn. These labs were notable for HgbA1c of 5.1% (normal <5.7%) and TSH of 0.5 mIU/L (normal level, 1.5 mIU/L), which is low for someone taking thyroid replacement.

I requested that Ms. T stop Armour Thyroid to address the suppressed TSH. I also requested that she stop metfor­min because, although hypoglycemia from metformin monotherapy is uncom­mon, it can happen in older patients. Hypoglycemia associated with metformin also can occur in situations when caloric intake is deficient or when metformin is used in combination with other drugs such as sulfonylureas (ie, glipizide), beta-adrenergic blocking drugs, angiotensin-converting enzyme inhibitors, or even nonsteroidal anti-inflammatory drugs.¹³

Identifying overlapping psychiatric (or psychological) illness
Symptoms of depression, anxiety, psycho­sis, and even dissociation can present as con­fusion. The term pseudodementia describes patients who exhibit cognitive symptoms consistent with NCD but could improve once the underlying mood, thought, anxi­ety, or personality disorder is treated.

For example, a patient with depression typically exhibits neurovegetative symp­toms—such as poor sleep or appetite— amotivation, and low energy. All of these can lead to abrupt-onset cognitive changes, which are a hallmark of pseudodementia rather than the more insidious pattern of mild NCD. In cases of pseudodementia, neurocognitive testing will show impair­ment that often rapidly improves after the primary psychiatric (or psychological) issue is rectified. Making a diagnosis of pseu­dodementia at the initial presentation is difficult because neurocognitive tests such as the MMSE often fail to separate depres­sion from true cognitive changes.¹⁴ Such a diagnosis typically requires hindsight. Yet, one must also keep in mind that pseudode­mentia may be part of a NCD prodrome.¹⁵

Conversion disorder as well as the disso­ciative disorders and substance-related dis­orders are notorious for causing confusion. In Ms. T’s case, pseudodementia stemming from her underlying bipolar disorder and anxiety figured prominently in the differ­ential diagnosis, but she did not have any other overt psychopathology, personality disorder, or signs of malingering to further complicate her picture.

Notebook. I recommend that my patients keep a small notebook to record medical data ranging from blood pressure and gly­cemic measurements to details about sleep and dietary intake. Such data comprise the necessary metrics to properly assess target conditions and then track changes once treatment is initiated. This exercise not only yields much-needed detail about the patient’s condition for the clinician; the act of journaling also can be therapeutic for the writer through a process known as experi­mental disclosure, in which writing down one’s thoughts and observations has a posi­tive impact on the writer’s physical health and psychology.¹⁶

Diagnosis. The first rule in medicine (perhaps the second, behind primum non nocere) is to determine what you are treat­ing before beginning treatment (decernite quid tractemus, prius cura ministrandi, for Latin buffs). This means trying to fash­ion the best diagnostic label, even if it is merely a place-holder, while assessment of the confused state continues. DSM-5 has attempted to remove stigma from several neuropsychiatric disorders. On the cog­nition front, the new name for dementia is “neurocognitive disorder (NCD),” the umbrella term that focuses on the decline from a previous level of cognitive func­tioning. NCD has been divided into mild or major cognitive impairment headings either “with” or “without behavioral dis­turbance” subspecifiers.¹⁷

Aside from NCD, there are several other diagnoses in the differential for confusion. Delirium remains the most prominent and focuses on disturbances in attention and orientation that develops over a short period of time, with a change seen in an additional cognitive domain, such as memory, but not in the context of a severely reduced level of arousal such as coma. Subjective cognitive impairment (SCI) is when subjective complaints of cog­nitive impairment are hallmark compared with objective findings—with evidence suggesting that the presence of SCI could predict a 4.5 times higher rate of develop­ing mild cognitive impairment (MCI) over 7 years.¹⁸ MCI was originally used to describe the early prodrome of AD, minus functional decline.


Treatment
After even a provisional diagnosis comes the final, all-important challenge: treating the neuropsychiatric symptoms (NPS) of the confused patient. NPS are nearly universal in NCD/delirium throughout the course of illness. There are no FDA-approved treat­ments for the NPS associated with these conditions. In terms of treating delirium, the best approach is to treat the underlying medical condition. For control of behavior, which can range from agitated to psychotic to hypoactive, nonpharmacotherapeutic interventions are paramount; they include making sure that the patient is at the appro­priate level of care, which, for the confused outpatient, could mean hospitalization. Ensuring proper nutrition, hydration, sen­sory care (hearing aids, glasses, etc.), and stability in ambulation must be done before considering pharmacotherapy.

Antipsychotic use has been the mainstay of drug treatment of behavioral dyscontrol. Haloperidol has been the traditional go-to medication because there is no evidence that low-dose haloperidol (<3 mg/d) has any different efficacy compared with the atypical antipsychotics or has a greater fre­quency of adverse drug effects. However, high-dose haloperidol (>4.5 mg/d) was associated with a greater incidence of adverse effects, mainly parkinsonism, than atypical antipsychotics.¹⁹ Neither the typi­cal nor atypical antipsychotics have shown mortality benefit—the real outcome mea­sure of interest.

In terms of treating major (or minor) NCD, there are only 2 FDA-approved medication classes: cholinesterase inhibi­tors (donepezil, galantamine, rivastig­mine, etc.) and memantine. However, these medication classes—even when combined together—have only shown marginal benefit in terms of improving cognition. Worse, even when given early in the course of illness they do not reduce the rate of NCD. For pseudodementia, selec­tive serotonin reuptake inhibitors (SSRIs) and serotonin-norepinephrine reuptake inhibitors tend to form the mainstay of treating underlying depression or anxiety leading to cognitive changes. Preliminary data suggest that some SSRIs might improve cognition in terms of process­ing speed, verbal learning, and memory.²⁰ More studies are needed before definitive conclusions can be drawn.

For the confused patient, a personalized therapeutic program, in which multiple interventions are considered at once (tar­geting all areas of the patient’s life) is gain­ing research traction. For example, a novel, comprehensive program involving mul­tiple modalities designed to achieve meta­bolic enhancement for neurodegeneration (MEND) recently has shown robust benefit for patients with AD, MCI, and SCI.²¹ Using an individual approach to improve diet, activity, sleep, metabolic status including body mass index, and several other mark­ers that affect neural plasticity, researchers demonstrated symptom improvement in 9 of 10 study patients.

Yet, some of the interventions, such as the use of statins for hyperlipidemia, remain controversial, with some studies suggesting that they help cognition,^22,23 and others showing no association.²⁴ The researchers caution that further research is warranted before costly dementia pre­vention trials with statins are undertaken. It does not appear that there are current MEND-type research projects in delirium but it’s to be hoped that we will see these in the future.

In the case of Ms. T, the cause of delir­ium vs mild NCD was thought to be mul­tifactorial. Discontinuing Armour Thyroid and metformin—symptoms of hypoglyce­mia emerged as a leading concern—were simple adjustments that led to resolution of the most concerning elements of her confusion. She continued her other psy­chotropics, although there might be mild residual cognitive issues that warrant close observation.

Related Resources
• Lin JS, O’Connor E, Rossum RC, et al. Screening for cognitive impairment in older adults: an evidence update for the U.S. Preventive Services Task Force. Rockville, MD: Agency for Healthcare Research and Quality (US); 2013.
• Grover S, Kate N. Assessment scales for delirium: a review. World J Psychiatry. 2012;2(4):58-70.

Drug Brand Names
Atorvastatin • Lipitor                           Lithium • Eskalith, Lithobid
Donepezil • Aricept                              Lorazepam • Ativan
Escitalopram • Lexapro                        Memantine • Namenda
Flumazenil • Romazicon                       Metformin • Glucophage
Galantamine • Razadyne                      Naloxone • Narcan
Glipizide • Glucotrol                             Physostigmine • Antilirium
Haloperidol • Haldol                             Quetiapine • Seroquel
Levothyroxine • Levoxyl, Synthroid       Rivastigmine • Exelon
Lithium • Eskalith, Lithobid                   Valproic acid • Depakene

Disclosure
Dr. Raj is a speaker for Actavis Pharmaceuticals, AstraZeneca, and Merck.

“Mr. Smith seems somewhat confused today” is one of the most serious and concerning pre-visit reports you can receive from your staff or the patient’s family. Such a descriptor can be confusing—pardon the pun—not only for the patient, but to even seasoned mental health providers.

The term confusion can be code for diagnoses ranging from delirium^a to a progressive neurocognitive disorder (NCD) such as major NCD due to Alzheimer’s disease (AD), or even a more challenging prob­lem such as beclouded dementia (delirium superimposed on demen­tia/NCD). It is essential for all mental health professionals to have an evidence-based approach when encountering signs or symptoms of confusion.

^aICD-10 code R41.0 encompasses Confusion, Other Specified Delirium, or Unspecified Delirium.

CASE REPORT
Ms. T, age 62, has hypothyroidism and bipolar I disorder, most recently depressed, with comorbid generalized anxiety disorder. She has been tak­ing lithium, 600 mg/d, to control her mood symptoms. Her daughter-in-law reports that Ms. T has been exhibiting increasing signs of confusion. During the office evaluation, Ms. T minimizes her symptoms, only describ­ing mild issues with forgetfulness while cooking and concern over increas­ing anxiety. Her daughter-in-law plays a voicemail message from earlier in the week, in which Ms. T’s speech is halting, disorganized, and in a word, confused. I decide to use the mnemonic decision chart MR. MIND (Table 1) to get to the bottom of her recent confusion.

Measure cognition
It is nice to receive advanced warning about a cognitive change or a change in activities of daily living; however, many patients present with subtle, sub-acute changes that are more difficult to assess. When encountering a broad symptom such as “confusion”—which has an equally broad differential diagnosis—systematic assess­ment of the current cognitive state com­pared with the patient’s baseline becomes the first order of business. However, this requires that the patient has had a baseline cognitive assessment.

In my practice, I often administer one of the validated neurocognitive screening instruments when a patient first begins care—even a brief test such as the Mini- Cog (3-item recall plus clock drawing test), which is comparable to longer screening tests at least for NCD/dementia.¹ During a presentation for confusion, a more detailed neurocognitive assessment instrument would be recommended, allowing one to marry the clinical impression with a validated, objective measure. Formal neu­ropsychological testing by a clinical neuro­psychologist is the gold standard, but such testing is time-consuming and expensive and often not readily available. The screen­ing instrument I use for a more thorough evaluation depends on the clinical scenario.

The Six-Item Screener is used in some emergency settings because it is short but boasts a higher sensitivity than the Mini- Cog (94% vs 75%) with similar specificity when screening for cognitive impairment.² The Mini-Mental State Examination (MMSE) is a valuable instrument, although, recently, the Saint Louis University Mental Status Examination has been thought to be better at detecting mild NCD than the MMSE; more data are needed to substan­tiate this claim.³ The Montreal Cognitive Assessment is another validated screening tool that has been shown to be superior to the MMSE in terms of screening for mild cognitive impairment.⁴ The best delirium-specific assessment tool is the Confusion Assessment Method (Table 2).⁵

Review medications
A review of the medication list is not just a Joint Commission mandate (medication reconciliation during each encounter) but is important whenever confusion is noted. Polypharmacy can be a concern, but is not as concerning as the class of medication prescribed, particularly anticholinergic and sedative medications in patients age >65. The Drug Burden Index can be helpful in assessing this risk.⁶ Medications such as the benzodiazepine-receptor agonists, tri­cyclic antidepressants, and antipsychotics should be discontinued if possible, keeping in mind that the addition or subtraction of medications must be done prudently and only after reviewing the evidence and in consultation with the patient. A detailed medication review is as important for confused outpatients as it is for an inpatient case (steps 2 and 3 of the inpatient algo­rithm outlined in Table 3).⁷

In Ms. T’s case, the primary concern on her medication list was that her medical team was prescribing levothyroxine, 112 mcg/d, and desiccated thyroid (combination thy­roxine and triiodothyronine in the form of 20 mg Armour Thyroid), despite a lack of data for such combination therapy. Earlier, I had discontinued lorazepam, leaving lithium, 600 mg/d, quetiapine, 400 mg/d, and escitalopram, 10 mg/d, as her remain­ing psychotropics. Her other medications included atorvastatin, 40 mg/d, for hyper-lipidemia and metformin, 750 mg/d, for type 2 diabetes mellitus.

Medical illness
An organic basis must rank high in the dif­ferential diagnosis if medications are not the culprit. There are myriad medical disorders that can lead to confusion (Table 4).⁸ In an outpatient psychiatric setting, labo­ratory and radiology testing might not be readily available. It then becomes impor­tant to collaborate with a patient’s medical team if any of the following are met:
   •there is high suspicion of a medical cause
   •there could be delays in performing a medical workup
   •a physical examination is needed.

Other laboratory testing could be valu­able depending on the clinical scenario. These include tests such as:
   •drug level monitoring (lithium, val­proic acid, etc.) to assess for toxicity
   •HIV and rapid plasma reagin for sus­pected sexually transmitted infections
   •vitamin levels in patients with poor nutrition or post bariatric surgery
   •erythrocyte sedimentation rate or C-reactive protein, or both, if there are signs of inflammation
   •bacterial culture if blood or tissue infec­tion is a concern.

Esoteric tests include ceruloplasmin (Wilson’s disease), heavy metals screen, and even tests such as anti-gliadin anti­bodies because the prevalence of gluten sensitivity and celiac disease appear to be on the rise and have been associated with neuropsychiatric problems including encephalopathy.¹⁰

Brain imaging is an important consider­ation when a medical differential diagnosis for confusion is formulated. Unfortunately, there is little evidence-based guidance as to when brain imaging should be performed, often leading to overuse of tests such as CT, especially in emergency settings when con­fusion is noted. From a clinical standpoint, a head CT scan often is best ordered for patients who demonstrate an acute change in mental status, are age >70, are receiving anticoagulation, or have sustained trauma to the head. The key concern would be intracranial hemorrhage. However, some data suggest that the best use of head CT is for patients who have an impaired level of consciousness or a new focal neurologic deficit.¹¹

Apart from more acute changes, a brain MRI study is more helpful than a head CT when evaluating the brain parenchyma for more sub-acute diagnoses such as multiple sclerosis or a brain tumor. T2-weighted hyperintensities seen on an MRI are thought to predict an increased risk of stroke, dementia, and death.

Their discov­ery should prompt a detailed evaluation for risk factors of stroke and dementia/NCD.¹²

In Ms. T’s case, she was taking lithium, so it was logical to obtain a trough lith­ium level 12 hours after the last dose and to check kidney function (serum creati­nine to estimate the glomerular filtration rate), which were in the therapeutic/nor­mal range. Her serum lithium level was 0.7 mEq/L. Brain imaging was not ordered, but several other labs (CMP, CBC, hemoglo­bin A1c [HgbA1c], and TSH) were drawn. These labs were notable for HgbA1c of 5.1% (normal <5.7%) and TSH of 0.5 mIU/L (normal level, 1.5 mIU/L), which is low for someone taking thyroid replacement.

I requested that Ms. T stop Armour Thyroid to address the suppressed TSH. I also requested that she stop metfor­min because, although hypoglycemia from metformin monotherapy is uncom­mon, it can happen in older patients. Hypoglycemia associated with metformin also can occur in situations when caloric intake is deficient or when metformin is used in combination with other drugs such as sulfonylureas (ie, glipizide), beta-adrenergic blocking drugs, angiotensin-converting enzyme inhibitors, or even nonsteroidal anti-inflammatory drugs.¹³

Identifying overlapping psychiatric (or psychological) illness
Symptoms of depression, anxiety, psycho­sis, and even dissociation can present as con­fusion. The term pseudodementia describes patients who exhibit cognitive symptoms consistent with NCD but could improve once the underlying mood, thought, anxi­ety, or personality disorder is treated.

For example, a patient with depression typically exhibits neurovegetative symp­toms—such as poor sleep or appetite— amotivation, and low energy. All of these can lead to abrupt-onset cognitive changes, which are a hallmark of pseudodementia rather than the more insidious pattern of mild NCD. In cases of pseudodementia, neurocognitive testing will show impair­ment that often rapidly improves after the primary psychiatric (or psychological) issue is rectified. Making a diagnosis of pseu­dodementia at the initial presentation is difficult because neurocognitive tests such as the MMSE often fail to separate depres­sion from true cognitive changes.¹⁴ Such a diagnosis typically requires hindsight. Yet, one must also keep in mind that pseudode­mentia may be part of a NCD prodrome.¹⁵

Conversion disorder as well as the disso­ciative disorders and substance-related dis­orders are notorious for causing confusion. In Ms. T’s case, pseudodementia stemming from her underlying bipolar disorder and anxiety figured prominently in the differ­ential diagnosis, but she did not have any other overt psychopathology, personality disorder, or signs of malingering to further complicate her picture.

Notebook. I recommend that my patients keep a small notebook to record medical data ranging from blood pressure and gly­cemic measurements to details about sleep and dietary intake. Such data comprise the necessary metrics to properly assess target conditions and then track changes once treatment is initiated. This exercise not only yields much-needed detail about the patient’s condition for the clinician; the act of journaling also can be therapeutic for the writer through a process known as experi­mental disclosure, in which writing down one’s thoughts and observations has a posi­tive impact on the writer’s physical health and psychology.¹⁶

Diagnosis. The first rule in medicine (perhaps the second, behind primum non nocere) is to determine what you are treat­ing before beginning treatment (decernite quid tractemus, prius cura ministrandi, for Latin buffs). This means trying to fash­ion the best diagnostic label, even if it is merely a place-holder, while assessment of the confused state continues. DSM-5 has attempted to remove stigma from several neuropsychiatric disorders. On the cog­nition front, the new name for dementia is “neurocognitive disorder (NCD),” the umbrella term that focuses on the decline from a previous level of cognitive func­tioning. NCD has been divided into mild or major cognitive impairment headings either “with” or “without behavioral dis­turbance” subspecifiers.¹⁷

Aside from NCD, there are several other diagnoses in the differential for confusion. Delirium remains the most prominent and focuses on disturbances in attention and orientation that develops over a short period of time, with a change seen in an additional cognitive domain, such as memory, but not in the context of a severely reduced level of arousal such as coma. Subjective cognitive impairment (SCI) is when subjective complaints of cog­nitive impairment are hallmark compared with objective findings—with evidence suggesting that the presence of SCI could predict a 4.5 times higher rate of develop­ing mild cognitive impairment (MCI) over 7 years.¹⁸ MCI was originally used to describe the early prodrome of AD, minus functional decline.


Treatment
After even a provisional diagnosis comes the final, all-important challenge: treating the neuropsychiatric symptoms (NPS) of the confused patient. NPS are nearly universal in NCD/delirium throughout the course of illness. There are no FDA-approved treat­ments for the NPS associated with these conditions. In terms of treating delirium, the best approach is to treat the underlying medical condition. For control of behavior, which can range from agitated to psychotic to hypoactive, nonpharmacotherapeutic interventions are paramount; they include making sure that the patient is at the appro­priate level of care, which, for the confused outpatient, could mean hospitalization. Ensuring proper nutrition, hydration, sen­sory care (hearing aids, glasses, etc.), and stability in ambulation must be done before considering pharmacotherapy.

Antipsychotic use has been the mainstay of drug treatment of behavioral dyscontrol. Haloperidol has been the traditional go-to medication because there is no evidence that low-dose haloperidol (<3 mg/d) has any different efficacy compared with the atypical antipsychotics or has a greater fre­quency of adverse drug effects. However, high-dose haloperidol (>4.5 mg/d) was associated with a greater incidence of adverse effects, mainly parkinsonism, than atypical antipsychotics.¹⁹ Neither the typi­cal nor atypical antipsychotics have shown mortality benefit—the real outcome mea­sure of interest.

In terms of treating major (or minor) NCD, there are only 2 FDA-approved medication classes: cholinesterase inhibi­tors (donepezil, galantamine, rivastig­mine, etc.) and memantine. However, these medication classes—even when combined together—have only shown marginal benefit in terms of improving cognition. Worse, even when given early in the course of illness they do not reduce the rate of NCD. For pseudodementia, selec­tive serotonin reuptake inhibitors (SSRIs) and serotonin-norepinephrine reuptake inhibitors tend to form the mainstay of treating underlying depression or anxiety leading to cognitive changes. Preliminary data suggest that some SSRIs might improve cognition in terms of process­ing speed, verbal learning, and memory.²⁰ More studies are needed before definitive conclusions can be drawn.

For the confused patient, a personalized therapeutic program, in which multiple interventions are considered at once (tar­geting all areas of the patient’s life) is gain­ing research traction. For example, a novel, comprehensive program involving mul­tiple modalities designed to achieve meta­bolic enhancement for neurodegeneration (MEND) recently has shown robust benefit for patients with AD, MCI, and SCI.²¹ Using an individual approach to improve diet, activity, sleep, metabolic status including body mass index, and several other mark­ers that affect neural plasticity, researchers demonstrated symptom improvement in 9 of 10 study patients.

Yet, some of the interventions, such as the use of statins for hyperlipidemia, remain controversial, with some studies suggesting that they help cognition,^22,23 and others showing no association.²⁴ The researchers caution that further research is warranted before costly dementia pre­vention trials with statins are undertaken. It does not appear that there are current MEND-type research projects in delirium but it’s to be hoped that we will see these in the future.

In the case of Ms. T, the cause of delir­ium vs mild NCD was thought to be mul­tifactorial. Discontinuing Armour Thyroid and metformin—symptoms of hypoglyce­mia emerged as a leading concern—were simple adjustments that led to resolution of the most concerning elements of her confusion. She continued her other psy­chotropics, although there might be mild residual cognitive issues that warrant close observation.

Related Resources
• Lin JS, O’Connor E, Rossum RC, et al. Screening for cognitive impairment in older adults: an evidence update for the U.S. Preventive Services Task Force. Rockville, MD: Agency for Healthcare Research and Quality (US); 2013.
• Grover S, Kate N. Assessment scales for delirium: a review. World J Psychiatry. 2012;2(4):58-70.

Drug Brand Names
Atorvastatin • Lipitor                           Lithium • Eskalith, Lithobid
Donepezil • Aricept                              Lorazepam • Ativan
Escitalopram • Lexapro                        Memantine • Namenda
Flumazenil • Romazicon                       Metformin • Glucophage
Galantamine • Razadyne                      Naloxone • Narcan
Glipizide • Glucotrol                             Physostigmine • Antilirium
Haloperidol • Haldol                             Quetiapine • Seroquel
Levothyroxine • Levoxyl, Synthroid       Rivastigmine • Exelon
Lithium • Eskalith, Lithobid                   Valproic acid • Depakene

Disclosure
Dr. Raj is a speaker for Actavis Pharmaceuticals, AstraZeneca, and Merck.

“Mr. Smith seems somewhat confused today” is one of the most serious and concerning pre-visit reports you can receive from your staff or the patient’s family. Such a descriptor can be confusing—pardon the pun—not only for the patient, but to even seasoned mental health providers.

The term confusion can be code for diagnoses ranging from delirium^a to a progressive neurocognitive disorder (NCD) such as major NCD due to Alzheimer’s disease (AD), or even a more challenging prob­lem such as beclouded dementia (delirium superimposed on demen­tia/NCD). It is essential for all mental health professionals to have an evidence-based approach when encountering signs or symptoms of confusion.

^aICD-10 code R41.0 encompasses Confusion, Other Specified Delirium, or Unspecified Delirium.

CASE REPORT
Ms. T, age 62, has hypothyroidism and bipolar I disorder, most recently depressed, with comorbid generalized anxiety disorder. She has been tak­ing lithium, 600 mg/d, to control her mood symptoms. Her daughter-in-law reports that Ms. T has been exhibiting increasing signs of confusion. During the office evaluation, Ms. T minimizes her symptoms, only describ­ing mild issues with forgetfulness while cooking and concern over increas­ing anxiety. Her daughter-in-law plays a voicemail message from earlier in the week, in which Ms. T’s speech is halting, disorganized, and in a word, confused. I decide to use the mnemonic decision chart MR. MIND (Table 1) to get to the bottom of her recent confusion.

Measure cognition
It is nice to receive advanced warning about a cognitive change or a change in activities of daily living; however, many patients present with subtle, sub-acute changes that are more difficult to assess. When encountering a broad symptom such as “confusion”—which has an equally broad differential diagnosis—systematic assess­ment of the current cognitive state com­pared with the patient’s baseline becomes the first order of business. However, this requires that the patient has had a baseline cognitive assessment.

In my practice, I often administer one of the validated neurocognitive screening instruments when a patient first begins care—even a brief test such as the Mini- Cog (3-item recall plus clock drawing test), which is comparable to longer screening tests at least for NCD/dementia.¹ During a presentation for confusion, a more detailed neurocognitive assessment instrument would be recommended, allowing one to marry the clinical impression with a validated, objective measure. Formal neu­ropsychological testing by a clinical neuro­psychologist is the gold standard, but such testing is time-consuming and expensive and often not readily available. The screen­ing instrument I use for a more thorough evaluation depends on the clinical scenario.

The Six-Item Screener is used in some emergency settings because it is short but boasts a higher sensitivity than the Mini- Cog (94% vs 75%) with similar specificity when screening for cognitive impairment.² The Mini-Mental State Examination (MMSE) is a valuable instrument, although, recently, the Saint Louis University Mental Status Examination has been thought to be better at detecting mild NCD than the MMSE; more data are needed to substan­tiate this claim.³ The Montreal Cognitive Assessment is another validated screening tool that has been shown to be superior to the MMSE in terms of screening for mild cognitive impairment.⁴ The best delirium-specific assessment tool is the Confusion Assessment Method (Table 2).⁵

Review medications
A review of the medication list is not just a Joint Commission mandate (medication reconciliation during each encounter) but is important whenever confusion is noted. Polypharmacy can be a concern, but is not as concerning as the class of medication prescribed, particularly anticholinergic and sedative medications in patients age >65. The Drug Burden Index can be helpful in assessing this risk.⁶ Medications such as the benzodiazepine-receptor agonists, tri­cyclic antidepressants, and antipsychotics should be discontinued if possible, keeping in mind that the addition or subtraction of medications must be done prudently and only after reviewing the evidence and in consultation with the patient. A detailed medication review is as important for confused outpatients as it is for an inpatient case (steps 2 and 3 of the inpatient algo­rithm outlined in Table 3).⁷

In Ms. T’s case, the primary concern on her medication list was that her medical team was prescribing levothyroxine, 112 mcg/d, and desiccated thyroid (combination thy­roxine and triiodothyronine in the form of 20 mg Armour Thyroid), despite a lack of data for such combination therapy. Earlier, I had discontinued lorazepam, leaving lithium, 600 mg/d, quetiapine, 400 mg/d, and escitalopram, 10 mg/d, as her remain­ing psychotropics. Her other medications included atorvastatin, 40 mg/d, for hyper-lipidemia and metformin, 750 mg/d, for type 2 diabetes mellitus.

Medical illness
An organic basis must rank high in the dif­ferential diagnosis if medications are not the culprit. There are myriad medical disorders that can lead to confusion (Table 4).⁸ In an outpatient psychiatric setting, labo­ratory and radiology testing might not be readily available. It then becomes impor­tant to collaborate with a patient’s medical team if any of the following are met:
   •there is high suspicion of a medical cause
   •there could be delays in performing a medical workup
   •a physical examination is needed.

Other laboratory testing could be valu­able depending on the clinical scenario. These include tests such as:
   •drug level monitoring (lithium, val­proic acid, etc.) to assess for toxicity
   •HIV and rapid plasma reagin for sus­pected sexually transmitted infections
   •vitamin levels in patients with poor nutrition or post bariatric surgery
   •erythrocyte sedimentation rate or C-reactive protein, or both, if there are signs of inflammation
   •bacterial culture if blood or tissue infec­tion is a concern.

Esoteric tests include ceruloplasmin (Wilson’s disease), heavy metals screen, and even tests such as anti-gliadin anti­bodies because the prevalence of gluten sensitivity and celiac disease appear to be on the rise and have been associated with neuropsychiatric problems including encephalopathy.¹⁰

Brain imaging is an important consider­ation when a medical differential diagnosis for confusion is formulated. Unfortunately, there is little evidence-based guidance as to when brain imaging should be performed, often leading to overuse of tests such as CT, especially in emergency settings when con­fusion is noted. From a clinical standpoint, a head CT scan often is best ordered for patients who demonstrate an acute change in mental status, are age >70, are receiving anticoagulation, or have sustained trauma to the head. The key concern would be intracranial hemorrhage. However, some data suggest that the best use of head CT is for patients who have an impaired level of consciousness or a new focal neurologic deficit.¹¹

Apart from more acute changes, a brain MRI study is more helpful than a head CT when evaluating the brain parenchyma for more sub-acute diagnoses such as multiple sclerosis or a brain tumor. T2-weighted hyperintensities seen on an MRI are thought to predict an increased risk of stroke, dementia, and death.

Their discov­ery should prompt a detailed evaluation for risk factors of stroke and dementia/NCD.¹²

In Ms. T’s case, she was taking lithium, so it was logical to obtain a trough lith­ium level 12 hours after the last dose and to check kidney function (serum creati­nine to estimate the glomerular filtration rate), which were in the therapeutic/nor­mal range. Her serum lithium level was 0.7 mEq/L. Brain imaging was not ordered, but several other labs (CMP, CBC, hemoglo­bin A1c [HgbA1c], and TSH) were drawn. These labs were notable for HgbA1c of 5.1% (normal <5.7%) and TSH of 0.5 mIU/L (normal level, 1.5 mIU/L), which is low for someone taking thyroid replacement.

I requested that Ms. T stop Armour Thyroid to address the suppressed TSH. I also requested that she stop metfor­min because, although hypoglycemia from metformin monotherapy is uncom­mon, it can happen in older patients. Hypoglycemia associated with metformin also can occur in situations when caloric intake is deficient or when metformin is used in combination with other drugs such as sulfonylureas (ie, glipizide), beta-adrenergic blocking drugs, angiotensin-converting enzyme inhibitors, or even nonsteroidal anti-inflammatory drugs.¹³

Identifying overlapping psychiatric (or psychological) illness
Symptoms of depression, anxiety, psycho­sis, and even dissociation can present as con­fusion. The term pseudodementia describes patients who exhibit cognitive symptoms consistent with NCD but could improve once the underlying mood, thought, anxi­ety, or personality disorder is treated.

For example, a patient with depression typically exhibits neurovegetative symp­toms—such as poor sleep or appetite— amotivation, and low energy. All of these can lead to abrupt-onset cognitive changes, which are a hallmark of pseudodementia rather than the more insidious pattern of mild NCD. In cases of pseudodementia, neurocognitive testing will show impair­ment that often rapidly improves after the primary psychiatric (or psychological) issue is rectified. Making a diagnosis of pseu­dodementia at the initial presentation is difficult because neurocognitive tests such as the MMSE often fail to separate depres­sion from true cognitive changes.¹⁴ Such a diagnosis typically requires hindsight. Yet, one must also keep in mind that pseudode­mentia may be part of a NCD prodrome.¹⁵

Conversion disorder as well as the disso­ciative disorders and substance-related dis­orders are notorious for causing confusion. In Ms. T’s case, pseudodementia stemming from her underlying bipolar disorder and anxiety figured prominently in the differ­ential diagnosis, but she did not have any other overt psychopathology, personality disorder, or signs of malingering to further complicate her picture.

Notebook. I recommend that my patients keep a small notebook to record medical data ranging from blood pressure and gly­cemic measurements to details about sleep and dietary intake. Such data comprise the necessary metrics to properly assess target conditions and then track changes once treatment is initiated. This exercise not only yields much-needed detail about the patient’s condition for the clinician; the act of journaling also can be therapeutic for the writer through a process known as experi­mental disclosure, in which writing down one’s thoughts and observations has a posi­tive impact on the writer’s physical health and psychology.¹⁶

Diagnosis. The first rule in medicine (perhaps the second, behind primum non nocere) is to determine what you are treat­ing before beginning treatment (decernite quid tractemus, prius cura ministrandi, for Latin buffs). This means trying to fash­ion the best diagnostic label, even if it is merely a place-holder, while assessment of the confused state continues. DSM-5 has attempted to remove stigma from several neuropsychiatric disorders. On the cog­nition front, the new name for dementia is “neurocognitive disorder (NCD),” the umbrella term that focuses on the decline from a previous level of cognitive func­tioning. NCD has been divided into mild or major cognitive impairment headings either “with” or “without behavioral dis­turbance” subspecifiers.¹⁷

Aside from NCD, there are several other diagnoses in the differential for confusion. Delirium remains the most prominent and focuses on disturbances in attention and orientation that develops over a short period of time, with a change seen in an additional cognitive domain, such as memory, but not in the context of a severely reduced level of arousal such as coma. Subjective cognitive impairment (SCI) is when subjective complaints of cog­nitive impairment are hallmark compared with objective findings—with evidence suggesting that the presence of SCI could predict a 4.5 times higher rate of develop­ing mild cognitive impairment (MCI) over 7 years.¹⁸ MCI was originally used to describe the early prodrome of AD, minus functional decline.


Treatment
After even a provisional diagnosis comes the final, all-important challenge: treating the neuropsychiatric symptoms (NPS) of the confused patient. NPS are nearly universal in NCD/delirium throughout the course of illness. There are no FDA-approved treat­ments for the NPS associated with these conditions. In terms of treating delirium, the best approach is to treat the underlying medical condition. For control of behavior, which can range from agitated to psychotic to hypoactive, nonpharmacotherapeutic interventions are paramount; they include making sure that the patient is at the appro­priate level of care, which, for the confused outpatient, could mean hospitalization. Ensuring proper nutrition, hydration, sen­sory care (hearing aids, glasses, etc.), and stability in ambulation must be done before considering pharmacotherapy.

Antipsychotic use has been the mainstay of drug treatment of behavioral dyscontrol. Haloperidol has been the traditional go-to medication because there is no evidence that low-dose haloperidol (<3 mg/d) has any different efficacy compared with the atypical antipsychotics or has a greater fre­quency of adverse drug effects. However, high-dose haloperidol (>4.5 mg/d) was associated with a greater incidence of adverse effects, mainly parkinsonism, than atypical antipsychotics.¹⁹ Neither the typi­cal nor atypical antipsychotics have shown mortality benefit—the real outcome mea­sure of interest.

In terms of treating major (or minor) NCD, there are only 2 FDA-approved medication classes: cholinesterase inhibi­tors (donepezil, galantamine, rivastig­mine, etc.) and memantine. However, these medication classes—even when combined together—have only shown marginal benefit in terms of improving cognition. Worse, even when given early in the course of illness they do not reduce the rate of NCD. For pseudodementia, selec­tive serotonin reuptake inhibitors (SSRIs) and serotonin-norepinephrine reuptake inhibitors tend to form the mainstay of treating underlying depression or anxiety leading to cognitive changes. Preliminary data suggest that some SSRIs might improve cognition in terms of process­ing speed, verbal learning, and memory.²⁰ More studies are needed before definitive conclusions can be drawn.

For the confused patient, a personalized therapeutic program, in which multiple interventions are considered at once (tar­geting all areas of the patient’s life) is gain­ing research traction. For example, a novel, comprehensive program involving mul­tiple modalities designed to achieve meta­bolic enhancement for neurodegeneration (MEND) recently has shown robust benefit for patients with AD, MCI, and SCI.²¹ Using an individual approach to improve diet, activity, sleep, metabolic status including body mass index, and several other mark­ers that affect neural plasticity, researchers demonstrated symptom improvement in 9 of 10 study patients.

Yet, some of the interventions, such as the use of statins for hyperlipidemia, remain controversial, with some studies suggesting that they help cognition,^22,23 and others showing no association.²⁴ The researchers caution that further research is warranted before costly dementia pre­vention trials with statins are undertaken. It does not appear that there are current MEND-type research projects in delirium but it’s to be hoped that we will see these in the future.

In the case of Ms. T, the cause of delir­ium vs mild NCD was thought to be mul­tifactorial. Discontinuing Armour Thyroid and metformin—symptoms of hypoglyce­mia emerged as a leading concern—were simple adjustments that led to resolution of the most concerning elements of her confusion. She continued her other psy­chotropics, although there might be mild residual cognitive issues that warrant close observation.

Related Resources
• Lin JS, O’Connor E, Rossum RC, et al. Screening for cognitive impairment in older adults: an evidence update for the U.S. Preventive Services Task Force. Rockville, MD: Agency for Healthcare Research and Quality (US); 2013.
• Grover S, Kate N. Assessment scales for delirium: a review. World J Psychiatry. 2012;2(4):58-70.

Drug Brand Names
Atorvastatin • Lipitor                           Lithium • Eskalith, Lithobid
Donepezil • Aricept                              Lorazepam • Ativan
Escitalopram • Lexapro                        Memantine • Namenda
Flumazenil • Romazicon                       Metformin • Glucophage
Galantamine • Razadyne                      Naloxone • Narcan
Glipizide • Glucotrol                             Physostigmine • Antilirium
Haloperidol • Haldol                             Quetiapine • Seroquel
Levothyroxine • Levoxyl, Synthroid       Rivastigmine • Exelon
Lithium • Eskalith, Lithobid                   Valproic acid • Depakene

Disclosure
Dr. Raj is a speaker for Actavis Pharmaceuticals, AstraZeneca, and Merck.

Air-filled pulmonary lesions commonly detected on chest computed tomography. Cystic lung lesions should be distinguished from other air-filled lesions to facilitate diagnosis. Primary care physicians play an integral role in the recognition of cystic lung disease.

The differential diagnosis of cystic lung disease is broad and includes isolated pulmonary, systemic, infectious, and congenital etiologies.

Here, we aim to provide a systematic, stepwise approach to help differentiate among the various cystic lung diseases and devise an algorithm for diagnosis. In doing so, we will discuss the clinical and radiographic features of many of these diseases:

• Lymphangioleiomyomatosis
• Birt-Hogg-Dubé syndrome
• Pulmonary Langerhans cell histiocytosis
• Interstitial pneumonia (desquamative interstitial pneumonia, lymphocytic interstitial pneumonia)
• Congenital cystic lung disease (congenital pulmonary airway malformation, pulmonary sequestration, bronchogenic cyst) Pulmonary infection
• Systemic disease (amyloidosis, light chain deposition disease, neurofibromatosis type 1).

STEP 1: RULE OUT CYST-MIMICS

A pulmonary cyst is a round, circumscribed space surrounded by an epithelial or fibrous wall of variable thickness.¹ On chest radiography and computed tomography, a cyst appears as a round parenchymal lucency or low-attenuating area with a well-defined interface with normal lung.¹ Cysts vary in wall thickness but  usually have a thin wall (< 2 mm) and occur without associated pulmonary emphysema.¹ They typically contain air but occasionally contain fluid or solid material.

A pulmonary cyst can be categorized as a bulla, bleb, or pneumatocele.

Pulmonary cysts can be categorized as bullae, blebs, or pneumatoceles

Bullae are larger than 1 cm in diameter, sharply demarcated by a thin wall, and usually accompanied by emphysematous changes in the adjacent lung.¹

Blebs are no larger than 1 cm in diameter, are located within the visceral pleura or the subpleural space, and appear on computed tomography as thin-walled air spaces that are contiguous with the pleura.¹ The distinction between a bleb and a bulla is of little clinical importance, and is often unnecessary.

Pneumatoceles are cysts that are frequently caused by acute pneumonia, trauma, or aspiration of hydrocarbon fluid, and are usually transient.¹

Mimics of pulmonary cysts include pulmonary cavities, emphysema, loculated pneumothoraces, honeycomb lung, and bronchiectasis (Figure 1).²

Pulmonary cavities differ from cysts in that their walls are typically thicker (usually > 4 mm).³

Emphysema differs from cystic lung disease as it typically leads to focal areas or regions of decreased lung attenuation that do not have defined walls.¹

Honeycombing refers to a cluster or row of cysts, 1 to 3 mm in wall thickness and typically 3 to 10 mm in diameter, that are associated with end-stage lung fibrosis.¹ They are typically subpleural in distribution and are accompanied by fibrotic features such as reticulation and traction bronchiectasis.¹

Bronchiectasis is dilation and distortion of bronchi and bronchioles and can be mistaken for cysts when viewed en face.¹

Loculated pneumothoraces can also mimic pulmonary cysts, but they typically fail to adhere to a defined anatomic unit and are subpleural in distribution.

STEP 2: CHARACTERIZE THE CLINICAL PRESENTATION

Clinical signs and symptoms of cystic lung disease play a key role in diagnosis (Table 1). For instance, spontaneous pneumothorax is commonly associated with diffuse cystic lung disease (lymphangioleiomyomatosis and Birt-Hogg-Dubé syndrome), while insidious dyspnea, with or without associated pneumothorax, is usually associated with the interstitial pneumonias (lymphocytic interstitial pneumonia and desquamative interstitial pneumonia).

In addition, congenital abnormalities of the lung can lead to cyst formation. These abnormalities, especially when associated with other congenital abnormalities, are often diagnosed in the prenatal and perinatal periods. However, some remain undetected until incidentally found later in adulthood or if superimposing infection develops.

Primary pulmonary infections can also cause parenchymal necrosis, which in turn cavitates or forms cysts.⁴

Lastly, cystic lung diseases can occur as part of a multiorgan or systemic illness in which the lung is one of the organs involved. Although usually diagnosed before the discovery of cysts or manifestations of pulmonary symptoms, they can present as a diagnostic challenge, especially when lung cysts are the initial presentation.bsence of amyloid fibrils.

In view of the features of the different types of cystic lung disease, adults with cystic lung disease can be grouped according to their typical clinical presentations (Table 2):

• Insidious dyspnea or spontaneous pneumothorax
• Incidentally found cysts or recurrent pneumonia
• Signs and symptoms of primary pulmonary infection
• Signs and symptoms that are primarily nonpulmonary.

Insidious dyspnea or spontaneous pneumothorax

Insidious dyspnea or spontaneous pneumothorax can be manifestations of lymphangioleiomyomatosis, Birt-Hogg-Dubé syndrome, pulmonary Langerhans cell histiocytosis, desquamative interstitial pneumonia, or lymphocytic interstitial pneumonia.

Lymphangioleiomyomatosis is characterized by abnormal cellular proliferation within the lung, kidney, lymphatic system, or any combination.⁵ The peak prevalence is in the third to fourth decades of life, and most patients are women of childbearing age.⁶ In addition to progressive dyspnea on exertion and pneumothorax, other signs and symptoms include hemoptysis, nonproductive cough, chylous pleural effusion, and ascites.^7,8

Birt-Hogg-Dubé syndrome is caused by germline mutations in the folliculin (FLCN) gene.⁹ It is characterized by skin fibrofolliculomas, pulmonary cysts, spontaneous pneumothorax, and renal cancer.¹⁰

Pulmonary Langerhans cell histiocytosis is part of the spectrum of Langerhans cell histiocytosis that, in addition to the lungs, can also involve the bone, pituitary gland, thyroid, skin, lymph nodes, and liver.¹¹ It occurs almost exclusively in smokers, affecting individuals in their 20s and 30s, with no gender predilection.^12,13 In addition to nonproductive cough and dyspnea, patients can also present with fever, anorexia, and weight loss,¹³ but approximately 25% of patients are asymptomatic.¹⁴

Desquamative interstitial pneumonia is an idiopathic interstitial pneumonia that, like pulmonary Langerhans cell histiocytosis, is seen almost exclusively in current or former smokers, who account for about 90% of patients with this disease. It affects almost twice as many men as women.^15,16 The mean age at onset is 42 to 46.^15,16 In addition to insidious cough and dyspnea, digital clubbing develops in 26% to 40% of patients.^16,17

Lymphocytic interstitial pneumonia is another rare idiopathic pneumonia, usually associated with connective tissue disease, Sjögren syndrome, immunodeficiencies, and viral infections.^18­–21 It is more common in women, presenting between the 4th and 7th decades of life, with a mean age at diagnosis of 50 to 56.^18,22 In addition to progressive dyspnea and cough, other symptoms include weight loss, pleuritic pain, arthralgias, fatigue, night sweats, and fever.²³

In summary, in this clinical group, lymphangioleiomyomatosis and Birt-Hogg-Dubé syndrome should be considered when patients present with spontaneous pneumothorax; those with Birt-Hogg-Dubé syndrome also present with skin lesions or renal cancer. In patients with progressive dyspnea and cough, lymphocytic interstitial pneumonia should be considered in those with a known history of connective tissue disease or immunodeficiency. Pulmonary Langerhans cell histiocytosis typically presents at a younger age (20 to 30 years old) than desquamative interstitial pneumonia (smokers in their 40s). Making the distinction, however, will likely require imaging with computed tomography.

Incidentally found cysts or recurrent pneumonia

Incidentally found cysts or recurrent pneumonia can be manifestations of congenital pulmonary airway malformation, pulmonary sequestration, or bronchogenic cyst.

Congenital pulmonary airway malformation, of which there are five types, is the most common pulmonary congenital abnormality. It accounts for up to 95% of cases of congenital cystic lung disease.^24,25 About 85% of cases are detected in the prenatal or perinatal periods.²⁶ Late-onset congenital pulmonary airway malformation (arising in childhood to adulthood) presents with recurrent pneumonia in about 75% of cases and can be misdiagnosed as lung abscess, pulmonary tuberculosis, or bronchiectasis.²⁷

Pulmonary sequestration, the second most common pulmonary congenital abnormality, is characterized by a portion of lung that does not connect to the tracheobronchial tree and has its own systemic arterial supply.²⁴ Intralobar sequestration, which shares the pleural investment with normal lung, accounts for about 80% of cases of pulmonary sequestration.^28–30 In addition to signs or symptoms of pulmonary infection, patients with pulmonary sequestration can remain asymp-
tomatic (about 25% of cases), or can present with hemoptysis or hemothorax.^28–30 In adults, the typical age at presentation is between 20 and 25.^29,30

Bronchogenic cyst is usually life-threatening in children. In adults, it commonly causes cough and chest pain.³¹ Hemoptysis, dysphagia, hoarseness, and diaphragmatic paralysis can also occur.^32,33 The mean age at diagnosis in adults is 35 to 40.^31,32

In summary, most cases of recurrent pneumonia with cysts are due to congenital pulmonary airway malformation. Pulmonary sequestration is the second most common cause of cystic lung disease in this group. Bronchogenic cyst is usually fatal in fetal development; smaller cysts can go unnoticed during the earlier years and are later found incidentally as imaging abnormalities in adults.

Signs and symptoms of primary pulmonary infections

Signs and symptoms of primary pulmonary infections can be due to Pneumocystis jirovecii pneumonia or echinococcal infections.

P jirovecii pneumonia commonly develops in patients with human immunodeficiency virus infection and low CD4 counts, recipients of hematologic or solid-organ transplants, and those receiving immunosuppressive therapy (eg, glucocorticoids or chemotherapy).

Echinococcal infections (with Echinococcus granulosus or multilocularis species) are more common in less-developed countries such as those in South America or the Middle East, in China, or in patients who have traveled to endemic areas.³⁴

In summary, cystic lung disease in patients with primary pulmonary infections can be diagnosed by the patient’s clinical history and risk factors for infections. Those with human immunodeficiency virus infection and other causes of immunodeficiency are predisposed to P jirovecii pneumonia. Echinococcal infections occur in those with a history of travel to an endemic area.

Primarily nonpulmonary signs and symptoms

If the patient has primarily nonpulmonary signs and symptoms, think about pulmonary amyloidosis, light chain deposition disease, and neurofibromatosis type 1.

Pulmonary amyloidosis has a variety of manifestations, including tracheobronchial disease, nodular parenchymal disease, diffuse or alveolar septal pattern, pleural disease, lymphadenopathy, and pulmonary cysts.⁴

Light chain deposition disease shares some clinical features with amyloidosis. However, the light chain fragments in this disease do not form amyloid fibrils and therefore do not stain positively with Congo red. The kidney is the most commonly involved organ.⁴

Neurofibromatosis type 1 is characterized by collections of neurofibromas, café-au-lait spots, and pigmented hamartomas in the iris (Lisch nodules).³⁵

In summary, patients in this group typically present with complications related to systemic involvement. Those with neurofibromatosis type 1 present with ophthalmologic, dermatologic, and neurologic manifestations. Amyloidosis and light chain deposition disease most commonly involve the renal system; their distinction will likely require tissue biopsy and Congo-red staining.

STEP 3: CHARACTERIZE THE RADIOGRAPHIC FEATURES

Characterization of pulmonary cysts and their distribution plays a key role in the diagnosis. Radiographically, cystic lung diseases can be subclassified into two major categories according to their cystic distribution:

• Discrete (focal or multifocal)
• Diffuse (unilobular or panlobular).^2,3

Discrete cystic lung diseases include congenital abnormalities, infectious diseases, and interstitial pneumonias.^2,3

Diffuse, panlobular cystic lung diseases include lymphangioleiomyomatosis, pulmonary Langerhans cell histiocytosis, Birt-Hogg-Dubé syndrome, amyloidosis, light chain deposition disease, and neurofibromatosis type 1.^7,13,36–39

In addition, other associated radiographic findings play a major role in diagnosis.

Cysts in patients presenting with insidious dyspnea or spontaneous pneumothorax

Lymphangioleiomyomatosis. Cysts are seen in nearly all cases of advanced lymphangioleiomyomatosis, typically in a diffuse pattern, varying from 2 mm to 40 mm in diameter, and uniform in shape (Figure  2A).^7,8,40–42

Other radiographic features include vessels located at the periphery of the cysts (in contrast to the centrilobular pattern seen with emphysema), and chylous pleural effusions (in about 22% of patients).⁴⁰ Nodules are typically not seen with lymphangioleiomyomatosis, and if found represent type 2 pneumocyte hyperplasia.

Pulmonary Langerhans cell histiocytosis. Nodules measuring 1 to 10 mm in diameter and favoring a centrilobular location are often seen on computed tomography. Pulmonary cysts occur in about 61% of patients.^13,43 Cysts are variable in size and shape (Figure 2B), in contrast to their uniform appearance in lymphangioleiomyomatosis. Most cysts are less than 10 mm in diameter; however, they can be up to 80 mm.^13,43 Early in its course, nodules may predominate in the upper and middle lobes. Over time, diffuse cysts become more common and can be difficult to differentiate from advanced smoking-induced emphysema.⁴⁴

Birt-Hogg-Dubé syndrome. Approximately 70% to 100% of patients with Birt-Hogg-Dubé syndrome will have multiple pulmonary cysts detected on computed tomography. These cysts are characteristically basal and subpleural in location, with varying sizes and irregular shapes in otherwise normal lung parenchyma (Figure 2C).^36,45,46

Desquamative interstitial pneumonia. Pulmonary cysts are present on computed tomography in about 32% of patients.⁴⁷ They are usually round and less than 20 mm in diameter.⁴⁸ Ground-glass opacity is present in almost all cases of desquamative interstitial pneumonia, with a diffuse pattern in 25% to 44% of patients.^16,17,47

Pulmonary cysts occur in up to two-thirds of those with lymphocytic interstitial pneumonia. Cysts are usually multifocal and perivascular in distribution and have varying sizes and shapes (Figure 2D).²² Ground-glass opacity and poorly defined centrilobular nodules are also frequently seen. Other computed tomographic findings include thickening of the bronchovascular bundles, focal consolidation, interseptal lobular thickening, pleural thickening, and lymph node enlargement.²²

In summary, in this group of patients, diffuse panlobular cysts are due to lymphangioleiomyomatosis, pulmonary Langerhans cell histiocytosis, or Birt-Hogg-Dubé syndrome. Cysts due to lymphangioleiomyomatosis have a diffuse distribution, while those due to pulmonary Langerhans cell histiocytosis tend to be upper-lobe-predominant and in the early stages are associated with stellate centrilobular nodules. Cysts in Birt-Hogg-Dubé syndrome tend to be subpleural and those due to lymphocytic interstitial pneumonia are perivascular in distribution.

Cysts that are incidentally found or occur in patients with recurrent pneumonia

Congenital pulmonary airway malformation types 1, 2, and 4 (Figure 3A, 3B). Cysts are typically discrete and focal or multifocal in distribution, but cases of multilobar and bilateral distribution have also been reported.^27,49 The lower lobes are more often involved.⁴⁹ Cysts vary in size and shape and can contain air, fluid, or both.^27,49 Up to 50% of cases can occur in conjunction with pulmonary sequestration.⁵⁰

Pulmonary sequestration displays an anomalous arterial supply on computed tomography (Figure 3C). Other imaging findings include mass lesions (49%), cystic lesions (29%), cavitary lesions (12%), and bronchiectasis.³⁰ Air trapping can be seen in the adjacent lung. Lower lobe involvement accounts for more than 95% of total cases of sequestration.³⁰ The cysts are usually discrete or focal in distribution. Misdiagnosis of pulmonary sequestration is common, and can include pulmonary abscess, pneumonia, bronchiectasis, and lung cancer.³⁰

Bronchogenic cyst. Cyst contents generally demonstrate water attenuation, or higher attenuation if filled with proteinaceous/mucoid material or calcium deposits; air-fluid levels are seen in infected cysts.³² Intrapulmonary cysts have a predilection for the lower lobes and are usually discrete or focal in distribution.^31,32 Mediastinal cysts are usually homogeneous, solitary, and located in the middle mediastinum.³² Cysts vary in size from 20 to 90 mm, with  a mean diameter of 40 mm.³¹

In summary, in this group of cystic lung diseases, characteristic computed tomographic findings will suggest the diagnosis—air-filled cysts of varying sizes for congenital pulmonary airway malformation and anomalous vascular supply for pulmonary sequestration. Bronchogenic cysts will tend to have water or higher-than-water attenuation due to proteinaceous-mucoid material or calcium deposits.

Cysts in patients with signs and symptoms of primary pulmonary infections

P jirovecii pneumonia. Between 10% and 15% of patients have cysts, and about 18% present with spontaneous pneumothorax.⁵¹ Cysts in P jirovecii pneumonia vary in size from 15 to 85 mm in diameter and tend to occur in the upper lobes (Figure 4A).^51,52

Echinococcal infection. Echinococcal pulmonary cysts typically are single and located more often in the lower lobes (Figure 4B).^53,54 Cysts can be complicated by air-fluid levels, hydropneumothorax, or pneumothorax, or they can turn into cavitary lesions.

The diagnoses of these pulmonary infections are usually made by clinical and computed tomographic findings and depend less on detecting and characterizing lung cysts. Patients with P jirovecii pneumonia tend to have bilateral perihilar ground-glass opacities, while air-fluid levels suggest echinococcal infections. Cysts in this group of patients tend to be discrete or focal or multifocal in distribution, and vary in size.

Cysts in patients with primarily nonpulmonary signs and symptoms

Amyloidosis. Cyst formation is rare in amyloidosis.⁴ When present, cysts can be diffuse and scattered in distribution, in varying sizes (usually < 30 mm in diameter) and irregular shapes (Figure 5).^55,56

Pulmonary light chain deposition disease usually presents as linear opacities and small nodules on chest computed tomography. Numerous cysts that are diffuse in distribution and have no topographic predominance can also be present. They can progress in number and size and coalesce to form irregular shapes.⁵⁷

Neurofibromatosis type 1. In neurofibromatosis type 1, the most common radiographic presentations are bibasilar reticular opacities (50%), bullae (50%), and ground glass opacities (37%).⁵⁸ Well-formed cysts occur in up to 25% of patients and tend to be diffuse and smaller (2 to 18 mm in diameter), with upper lobe predominance.^58,59

In summary, in this group of patients, bibasilar reticular and ground-glass opacities suggest neurofibromatosis type 1, while nodules and linear opacities suggest amyloidosis or light chain deposition disease. Cysts tend to be diffuse with varying sizes.

STEP 4: PUT IT ALL TOGETHER

Diagnosis in insidious dyspnea or spontaneous pneumothorax

For patients who present with insidious dyspnea or spontaneous pneumothorax, the diagnosis of cystic lung disease can be made by characterizing the distribution, size, and shape of the cysts (Table 3).

Diffuse, panlobular distribution. Cystic lung diseases with this pattern include lymphangioleiomyomatosis, pulmonary Langerhans cell histiocytosis, and Birt-Hogg-Dubé syndrome. In this group, cysts that are uniform in size and regular in shape are invariably due to lymphangioleiomyomatosis. Those with variable size and irregular shapes can be due to pulmonary Langerhans cell histiocytosis or Birt-Hogg-Dubé syndrome. Patients with pulmonary Langerhans cell histiocytosis tend to be smokers and their cysts tend to be upper- lobe-predominant. Those with Birt-Hogg-Dubé syndrome will likely have renal cancer or skin lesions; their cysts tend to be basilar and subpleural in distribution.

Cysts that are focal or multifocal and unilobular are due to lymphocytic interstitial pneumonia or desquamative interstitial pneumonia. Patients with lymphocytic interstitial pneumonia tend to have underlying connective tissue disease; those with desquamative interstitial pneumonia are almost always smokers. The definitive diagnosis for lymphocytic interstitial pneumonia or desquamative interstitial pneumonia can require a tissue biopsy.

Diagnosis in patients with incidentally found cysts or recurrent pneumonia

In those who present with incidentally found cysts or recurrent pneumonia, suspicion for a congenital lung malformation should be raised. Patients with a type 1, 2, or 4 congenital pulmonary airway malformation typically have air-filled cysts in varying sizes; those with pulmonary sequestration have an anomalous arterial supply in addition to cysts that are usually located in the lower lobes. Bronchogenic cysts tend to be larger, with attenuation equal to or greater than that of water, and distinguishing them from congenital pulmonary airway malformation will likely require surgical examination.

Diagnosis in patients with signs and symptoms of pulmonary infections

Patients with signs and symptoms of pulmonary infections should be investigated according to clinical risk factors for P jirovecii pneumonia or echinococcal infections.

Diagnosis in patients with primarily nonpulmonary presentations

The distinction between amyloidosis and neurofibromatosis type 1 can be made by the history and the clinical examination. However, a  definitive diagnosis of amyloidosis or light chain deposition disease requires tissue examination for the presence or absence of amyloid fibrils.

Air-filled pulmonary lesions commonly detected on chest computed tomography. Cystic lung lesions should be distinguished from other air-filled lesions to facilitate diagnosis. Primary care physicians play an integral role in the recognition of cystic lung disease.

The differential diagnosis of cystic lung disease is broad and includes isolated pulmonary, systemic, infectious, and congenital etiologies.

Here, we aim to provide a systematic, stepwise approach to help differentiate among the various cystic lung diseases and devise an algorithm for diagnosis. In doing so, we will discuss the clinical and radiographic features of many of these diseases:

• Lymphangioleiomyomatosis
• Birt-Hogg-Dubé syndrome
• Pulmonary Langerhans cell histiocytosis
• Interstitial pneumonia (desquamative interstitial pneumonia, lymphocytic interstitial pneumonia)
• Congenital cystic lung disease (congenital pulmonary airway malformation, pulmonary sequestration, bronchogenic cyst) Pulmonary infection
• Systemic disease (amyloidosis, light chain deposition disease, neurofibromatosis type 1).

STEP 1: RULE OUT CYST-MIMICS

A pulmonary cyst is a round, circumscribed space surrounded by an epithelial or fibrous wall of variable thickness.¹ On chest radiography and computed tomography, a cyst appears as a round parenchymal lucency or low-attenuating area with a well-defined interface with normal lung.¹ Cysts vary in wall thickness but  usually have a thin wall (< 2 mm) and occur without associated pulmonary emphysema.¹ They typically contain air but occasionally contain fluid or solid material.

A pulmonary cyst can be categorized as a bulla, bleb, or pneumatocele.

Pulmonary cysts can be categorized as bullae, blebs, or pneumatoceles

Bullae are larger than 1 cm in diameter, sharply demarcated by a thin wall, and usually accompanied by emphysematous changes in the adjacent lung.¹

Blebs are no larger than 1 cm in diameter, are located within the visceral pleura or the subpleural space, and appear on computed tomography as thin-walled air spaces that are contiguous with the pleura.¹ The distinction between a bleb and a bulla is of little clinical importance, and is often unnecessary.

Pneumatoceles are cysts that are frequently caused by acute pneumonia, trauma, or aspiration of hydrocarbon fluid, and are usually transient.¹

Mimics of pulmonary cysts include pulmonary cavities, emphysema, loculated pneumothoraces, honeycomb lung, and bronchiectasis (Figure 1).²

Pulmonary cavities differ from cysts in that their walls are typically thicker (usually > 4 mm).³

Emphysema differs from cystic lung disease as it typically leads to focal areas or regions of decreased lung attenuation that do not have defined walls.¹

Honeycombing refers to a cluster or row of cysts, 1 to 3 mm in wall thickness and typically 3 to 10 mm in diameter, that are associated with end-stage lung fibrosis.¹ They are typically subpleural in distribution and are accompanied by fibrotic features such as reticulation and traction bronchiectasis.¹

Bronchiectasis is dilation and distortion of bronchi and bronchioles and can be mistaken for cysts when viewed en face.¹

Loculated pneumothoraces can also mimic pulmonary cysts, but they typically fail to adhere to a defined anatomic unit and are subpleural in distribution.

STEP 2: CHARACTERIZE THE CLINICAL PRESENTATION

Clinical signs and symptoms of cystic lung disease play a key role in diagnosis (Table 1). For instance, spontaneous pneumothorax is commonly associated with diffuse cystic lung disease (lymphangioleiomyomatosis and Birt-Hogg-Dubé syndrome), while insidious dyspnea, with or without associated pneumothorax, is usually associated with the interstitial pneumonias (lymphocytic interstitial pneumonia and desquamative interstitial pneumonia).

In addition, congenital abnormalities of the lung can lead to cyst formation. These abnormalities, especially when associated with other congenital abnormalities, are often diagnosed in the prenatal and perinatal periods. However, some remain undetected until incidentally found later in adulthood or if superimposing infection develops.

Primary pulmonary infections can also cause parenchymal necrosis, which in turn cavitates or forms cysts.⁴

Lastly, cystic lung diseases can occur as part of a multiorgan or systemic illness in which the lung is one of the organs involved. Although usually diagnosed before the discovery of cysts or manifestations of pulmonary symptoms, they can present as a diagnostic challenge, especially when lung cysts are the initial presentation.bsence of amyloid fibrils.

In view of the features of the different types of cystic lung disease, adults with cystic lung disease can be grouped according to their typical clinical presentations (Table 2):

• Insidious dyspnea or spontaneous pneumothorax
• Incidentally found cysts or recurrent pneumonia
• Signs and symptoms of primary pulmonary infection
• Signs and symptoms that are primarily nonpulmonary.

Insidious dyspnea or spontaneous pneumothorax

Insidious dyspnea or spontaneous pneumothorax can be manifestations of lymphangioleiomyomatosis, Birt-Hogg-Dubé syndrome, pulmonary Langerhans cell histiocytosis, desquamative interstitial pneumonia, or lymphocytic interstitial pneumonia.

Lymphangioleiomyomatosis is characterized by abnormal cellular proliferation within the lung, kidney, lymphatic system, or any combination.⁵ The peak prevalence is in the third to fourth decades of life, and most patients are women of childbearing age.⁶ In addition to progressive dyspnea on exertion and pneumothorax, other signs and symptoms include hemoptysis, nonproductive cough, chylous pleural effusion, and ascites.^7,8

Birt-Hogg-Dubé syndrome is caused by germline mutations in the folliculin (FLCN) gene.⁹ It is characterized by skin fibrofolliculomas, pulmonary cysts, spontaneous pneumothorax, and renal cancer.¹⁰

Pulmonary Langerhans cell histiocytosis is part of the spectrum of Langerhans cell histiocytosis that, in addition to the lungs, can also involve the bone, pituitary gland, thyroid, skin, lymph nodes, and liver.¹¹ It occurs almost exclusively in smokers, affecting individuals in their 20s and 30s, with no gender predilection.^12,13 In addition to nonproductive cough and dyspnea, patients can also present with fever, anorexia, and weight loss,¹³ but approximately 25% of patients are asymptomatic.¹⁴

Desquamative interstitial pneumonia is an idiopathic interstitial pneumonia that, like pulmonary Langerhans cell histiocytosis, is seen almost exclusively in current or former smokers, who account for about 90% of patients with this disease. It affects almost twice as many men as women.^15,16 The mean age at onset is 42 to 46.^15,16 In addition to insidious cough and dyspnea, digital clubbing develops in 26% to 40% of patients.^16,17

Lymphocytic interstitial pneumonia is another rare idiopathic pneumonia, usually associated with connective tissue disease, Sjögren syndrome, immunodeficiencies, and viral infections.^18­–21 It is more common in women, presenting between the 4th and 7th decades of life, with a mean age at diagnosis of 50 to 56.^18,22 In addition to progressive dyspnea and cough, other symptoms include weight loss, pleuritic pain, arthralgias, fatigue, night sweats, and fever.²³

In summary, in this clinical group, lymphangioleiomyomatosis and Birt-Hogg-Dubé syndrome should be considered when patients present with spontaneous pneumothorax; those with Birt-Hogg-Dubé syndrome also present with skin lesions or renal cancer. In patients with progressive dyspnea and cough, lymphocytic interstitial pneumonia should be considered in those with a known history of connective tissue disease or immunodeficiency. Pulmonary Langerhans cell histiocytosis typically presents at a younger age (20 to 30 years old) than desquamative interstitial pneumonia (smokers in their 40s). Making the distinction, however, will likely require imaging with computed tomography.

Incidentally found cysts or recurrent pneumonia

Incidentally found cysts or recurrent pneumonia can be manifestations of congenital pulmonary airway malformation, pulmonary sequestration, or bronchogenic cyst.

Congenital pulmonary airway malformation, of which there are five types, is the most common pulmonary congenital abnormality. It accounts for up to 95% of cases of congenital cystic lung disease.^24,25 About 85% of cases are detected in the prenatal or perinatal periods.²⁶ Late-onset congenital pulmonary airway malformation (arising in childhood to adulthood) presents with recurrent pneumonia in about 75% of cases and can be misdiagnosed as lung abscess, pulmonary tuberculosis, or bronchiectasis.²⁷

Pulmonary sequestration, the second most common pulmonary congenital abnormality, is characterized by a portion of lung that does not connect to the tracheobronchial tree and has its own systemic arterial supply.²⁴ Intralobar sequestration, which shares the pleural investment with normal lung, accounts for about 80% of cases of pulmonary sequestration.^28–30 In addition to signs or symptoms of pulmonary infection, patients with pulmonary sequestration can remain asymp-
tomatic (about 25% of cases), or can present with hemoptysis or hemothorax.^28–30 In adults, the typical age at presentation is between 20 and 25.^29,30

Bronchogenic cyst is usually life-threatening in children. In adults, it commonly causes cough and chest pain.³¹ Hemoptysis, dysphagia, hoarseness, and diaphragmatic paralysis can also occur.^32,33 The mean age at diagnosis in adults is 35 to 40.^31,32

In summary, most cases of recurrent pneumonia with cysts are due to congenital pulmonary airway malformation. Pulmonary sequestration is the second most common cause of cystic lung disease in this group. Bronchogenic cyst is usually fatal in fetal development; smaller cysts can go unnoticed during the earlier years and are later found incidentally as imaging abnormalities in adults.

Signs and symptoms of primary pulmonary infections

Signs and symptoms of primary pulmonary infections can be due to Pneumocystis jirovecii pneumonia or echinococcal infections.

P jirovecii pneumonia commonly develops in patients with human immunodeficiency virus infection and low CD4 counts, recipients of hematologic or solid-organ transplants, and those receiving immunosuppressive therapy (eg, glucocorticoids or chemotherapy).

Echinococcal infections (with Echinococcus granulosus or multilocularis species) are more common in less-developed countries such as those in South America or the Middle East, in China, or in patients who have traveled to endemic areas.³⁴

In summary, cystic lung disease in patients with primary pulmonary infections can be diagnosed by the patient’s clinical history and risk factors for infections. Those with human immunodeficiency virus infection and other causes of immunodeficiency are predisposed to P jirovecii pneumonia. Echinococcal infections occur in those with a history of travel to an endemic area.

Primarily nonpulmonary signs and symptoms

If the patient has primarily nonpulmonary signs and symptoms, think about pulmonary amyloidosis, light chain deposition disease, and neurofibromatosis type 1.

Pulmonary amyloidosis has a variety of manifestations, including tracheobronchial disease, nodular parenchymal disease, diffuse or alveolar septal pattern, pleural disease, lymphadenopathy, and pulmonary cysts.⁴

Light chain deposition disease shares some clinical features with amyloidosis. However, the light chain fragments in this disease do not form amyloid fibrils and therefore do not stain positively with Congo red. The kidney is the most commonly involved organ.⁴

Neurofibromatosis type 1 is characterized by collections of neurofibromas, café-au-lait spots, and pigmented hamartomas in the iris (Lisch nodules).³⁵

In summary, patients in this group typically present with complications related to systemic involvement. Those with neurofibromatosis type 1 present with ophthalmologic, dermatologic, and neurologic manifestations. Amyloidosis and light chain deposition disease most commonly involve the renal system; their distinction will likely require tissue biopsy and Congo-red staining.

STEP 3: CHARACTERIZE THE RADIOGRAPHIC FEATURES

Characterization of pulmonary cysts and their distribution plays a key role in the diagnosis. Radiographically, cystic lung diseases can be subclassified into two major categories according to their cystic distribution:

• Discrete (focal or multifocal)
• Diffuse (unilobular or panlobular).^2,3

Discrete cystic lung diseases include congenital abnormalities, infectious diseases, and interstitial pneumonias.^2,3

Diffuse, panlobular cystic lung diseases include lymphangioleiomyomatosis, pulmonary Langerhans cell histiocytosis, Birt-Hogg-Dubé syndrome, amyloidosis, light chain deposition disease, and neurofibromatosis type 1.^7,13,36–39

In addition, other associated radiographic findings play a major role in diagnosis.

Cysts in patients presenting with insidious dyspnea or spontaneous pneumothorax

Lymphangioleiomyomatosis. Cysts are seen in nearly all cases of advanced lymphangioleiomyomatosis, typically in a diffuse pattern, varying from 2 mm to 40 mm in diameter, and uniform in shape (Figure  2A).^7,8,40–42

Other radiographic features include vessels located at the periphery of the cysts (in contrast to the centrilobular pattern seen with emphysema), and chylous pleural effusions (in about 22% of patients).⁴⁰ Nodules are typically not seen with lymphangioleiomyomatosis, and if found represent type 2 pneumocyte hyperplasia.

Pulmonary Langerhans cell histiocytosis. Nodules measuring 1 to 10 mm in diameter and favoring a centrilobular location are often seen on computed tomography. Pulmonary cysts occur in about 61% of patients.^13,43 Cysts are variable in size and shape (Figure 2B), in contrast to their uniform appearance in lymphangioleiomyomatosis. Most cysts are less than 10 mm in diameter; however, they can be up to 80 mm.^13,43 Early in its course, nodules may predominate in the upper and middle lobes. Over time, diffuse cysts become more common and can be difficult to differentiate from advanced smoking-induced emphysema.⁴⁴

Birt-Hogg-Dubé syndrome. Approximately 70% to 100% of patients with Birt-Hogg-Dubé syndrome will have multiple pulmonary cysts detected on computed tomography. These cysts are characteristically basal and subpleural in location, with varying sizes and irregular shapes in otherwise normal lung parenchyma (Figure 2C).^36,45,46

Desquamative interstitial pneumonia. Pulmonary cysts are present on computed tomography in about 32% of patients.⁴⁷ They are usually round and less than 20 mm in diameter.⁴⁸ Ground-glass opacity is present in almost all cases of desquamative interstitial pneumonia, with a diffuse pattern in 25% to 44% of patients.^16,17,47

Pulmonary cysts occur in up to two-thirds of those with lymphocytic interstitial pneumonia. Cysts are usually multifocal and perivascular in distribution and have varying sizes and shapes (Figure 2D).²² Ground-glass opacity and poorly defined centrilobular nodules are also frequently seen. Other computed tomographic findings include thickening of the bronchovascular bundles, focal consolidation, interseptal lobular thickening, pleural thickening, and lymph node enlargement.²²

In summary, in this group of patients, diffuse panlobular cysts are due to lymphangioleiomyomatosis, pulmonary Langerhans cell histiocytosis, or Birt-Hogg-Dubé syndrome. Cysts due to lymphangioleiomyomatosis have a diffuse distribution, while those due to pulmonary Langerhans cell histiocytosis tend to be upper-lobe-predominant and in the early stages are associated with stellate centrilobular nodules. Cysts in Birt-Hogg-Dubé syndrome tend to be subpleural and those due to lymphocytic interstitial pneumonia are perivascular in distribution.

Cysts that are incidentally found or occur in patients with recurrent pneumonia

Congenital pulmonary airway malformation types 1, 2, and 4 (Figure 3A, 3B). Cysts are typically discrete and focal or multifocal in distribution, but cases of multilobar and bilateral distribution have also been reported.^27,49 The lower lobes are more often involved.⁴⁹ Cysts vary in size and shape and can contain air, fluid, or both.^27,49 Up to 50% of cases can occur in conjunction with pulmonary sequestration.⁵⁰

Pulmonary sequestration displays an anomalous arterial supply on computed tomography (Figure 3C). Other imaging findings include mass lesions (49%), cystic lesions (29%), cavitary lesions (12%), and bronchiectasis.³⁰ Air trapping can be seen in the adjacent lung. Lower lobe involvement accounts for more than 95% of total cases of sequestration.³⁰ The cysts are usually discrete or focal in distribution. Misdiagnosis of pulmonary sequestration is common, and can include pulmonary abscess, pneumonia, bronchiectasis, and lung cancer.³⁰

Bronchogenic cyst. Cyst contents generally demonstrate water attenuation, or higher attenuation if filled with proteinaceous/mucoid material or calcium deposits; air-fluid levels are seen in infected cysts.³² Intrapulmonary cysts have a predilection for the lower lobes and are usually discrete or focal in distribution.^31,32 Mediastinal cysts are usually homogeneous, solitary, and located in the middle mediastinum.³² Cysts vary in size from 20 to 90 mm, with  a mean diameter of 40 mm.³¹

In summary, in this group of cystic lung diseases, characteristic computed tomographic findings will suggest the diagnosis—air-filled cysts of varying sizes for congenital pulmonary airway malformation and anomalous vascular supply for pulmonary sequestration. Bronchogenic cysts will tend to have water or higher-than-water attenuation due to proteinaceous-mucoid material or calcium deposits.

Cysts in patients with signs and symptoms of primary pulmonary infections

P jirovecii pneumonia. Between 10% and 15% of patients have cysts, and about 18% present with spontaneous pneumothorax.⁵¹ Cysts in P jirovecii pneumonia vary in size from 15 to 85 mm in diameter and tend to occur in the upper lobes (Figure 4A).^51,52

Echinococcal infection. Echinococcal pulmonary cysts typically are single and located more often in the lower lobes (Figure 4B).^53,54 Cysts can be complicated by air-fluid levels, hydropneumothorax, or pneumothorax, or they can turn into cavitary lesions.

The diagnoses of these pulmonary infections are usually made by clinical and computed tomographic findings and depend less on detecting and characterizing lung cysts. Patients with P jirovecii pneumonia tend to have bilateral perihilar ground-glass opacities, while air-fluid levels suggest echinococcal infections. Cysts in this group of patients tend to be discrete or focal or multifocal in distribution, and vary in size.

Cysts in patients with primarily nonpulmonary signs and symptoms

Amyloidosis. Cyst formation is rare in amyloidosis.⁴ When present, cysts can be diffuse and scattered in distribution, in varying sizes (usually < 30 mm in diameter) and irregular shapes (Figure 5).^55,56

Pulmonary light chain deposition disease usually presents as linear opacities and small nodules on chest computed tomography. Numerous cysts that are diffuse in distribution and have no topographic predominance can also be present. They can progress in number and size and coalesce to form irregular shapes.⁵⁷

Neurofibromatosis type 1. In neurofibromatosis type 1, the most common radiographic presentations are bibasilar reticular opacities (50%), bullae (50%), and ground glass opacities (37%).⁵⁸ Well-formed cysts occur in up to 25% of patients and tend to be diffuse and smaller (2 to 18 mm in diameter), with upper lobe predominance.^58,59

In summary, in this group of patients, bibasilar reticular and ground-glass opacities suggest neurofibromatosis type 1, while nodules and linear opacities suggest amyloidosis or light chain deposition disease. Cysts tend to be diffuse with varying sizes.

STEP 4: PUT IT ALL TOGETHER

Diagnosis in insidious dyspnea or spontaneous pneumothorax

For patients who present with insidious dyspnea or spontaneous pneumothorax, the diagnosis of cystic lung disease can be made by characterizing the distribution, size, and shape of the cysts (Table 3).

Diffuse, panlobular distribution. Cystic lung diseases with this pattern include lymphangioleiomyomatosis, pulmonary Langerhans cell histiocytosis, and Birt-Hogg-Dubé syndrome. In this group, cysts that are uniform in size and regular in shape are invariably due to lymphangioleiomyomatosis. Those with variable size and irregular shapes can be due to pulmonary Langerhans cell histiocytosis or Birt-Hogg-Dubé syndrome. Patients with pulmonary Langerhans cell histiocytosis tend to be smokers and their cysts tend to be upper- lobe-predominant. Those with Birt-Hogg-Dubé syndrome will likely have renal cancer or skin lesions; their cysts tend to be basilar and subpleural in distribution.

Cysts that are focal or multifocal and unilobular are due to lymphocytic interstitial pneumonia or desquamative interstitial pneumonia. Patients with lymphocytic interstitial pneumonia tend to have underlying connective tissue disease; those with desquamative interstitial pneumonia are almost always smokers. The definitive diagnosis for lymphocytic interstitial pneumonia or desquamative interstitial pneumonia can require a tissue biopsy.

Diagnosis in patients with incidentally found cysts or recurrent pneumonia

In those who present with incidentally found cysts or recurrent pneumonia, suspicion for a congenital lung malformation should be raised. Patients with a type 1, 2, or 4 congenital pulmonary airway malformation typically have air-filled cysts in varying sizes; those with pulmonary sequestration have an anomalous arterial supply in addition to cysts that are usually located in the lower lobes. Bronchogenic cysts tend to be larger, with attenuation equal to or greater than that of water, and distinguishing them from congenital pulmonary airway malformation will likely require surgical examination.

Diagnosis in patients with signs and symptoms of pulmonary infections

Patients with signs and symptoms of pulmonary infections should be investigated according to clinical risk factors for P jirovecii pneumonia or echinococcal infections.

Diagnosis in patients with primarily nonpulmonary presentations

The distinction between amyloidosis and neurofibromatosis type 1 can be made by the history and the clinical examination. However, a  definitive diagnosis of amyloidosis or light chain deposition disease requires tissue examination for the presence or absence of amyloid fibrils.

Air-filled pulmonary lesions commonly detected on chest computed tomography. Cystic lung lesions should be distinguished from other air-filled lesions to facilitate diagnosis. Primary care physicians play an integral role in the recognition of cystic lung disease.

The differential diagnosis of cystic lung disease is broad and includes isolated pulmonary, systemic, infectious, and congenital etiologies.

Here, we aim to provide a systematic, stepwise approach to help differentiate among the various cystic lung diseases and devise an algorithm for diagnosis. In doing so, we will discuss the clinical and radiographic features of many of these diseases:

• Lymphangioleiomyomatosis
• Birt-Hogg-Dubé syndrome
• Pulmonary Langerhans cell histiocytosis
• Interstitial pneumonia (desquamative interstitial pneumonia, lymphocytic interstitial pneumonia)
• Congenital cystic lung disease (congenital pulmonary airway malformation, pulmonary sequestration, bronchogenic cyst) Pulmonary infection
• Systemic disease (amyloidosis, light chain deposition disease, neurofibromatosis type 1).

STEP 1: RULE OUT CYST-MIMICS

A pulmonary cyst is a round, circumscribed space surrounded by an epithelial or fibrous wall of variable thickness.¹ On chest radiography and computed tomography, a cyst appears as a round parenchymal lucency or low-attenuating area with a well-defined interface with normal lung.¹ Cysts vary in wall thickness but  usually have a thin wall (< 2 mm) and occur without associated pulmonary emphysema.¹ They typically contain air but occasionally contain fluid or solid material.

A pulmonary cyst can be categorized as a bulla, bleb, or pneumatocele.

Pulmonary cysts can be categorized as bullae, blebs, or pneumatoceles

Bullae are larger than 1 cm in diameter, sharply demarcated by a thin wall, and usually accompanied by emphysematous changes in the adjacent lung.¹

Blebs are no larger than 1 cm in diameter, are located within the visceral pleura or the subpleural space, and appear on computed tomography as thin-walled air spaces that are contiguous with the pleura.¹ The distinction between a bleb and a bulla is of little clinical importance, and is often unnecessary.

Pneumatoceles are cysts that are frequently caused by acute pneumonia, trauma, or aspiration of hydrocarbon fluid, and are usually transient.¹

Mimics of pulmonary cysts include pulmonary cavities, emphysema, loculated pneumothoraces, honeycomb lung, and bronchiectasis (Figure 1).²

Pulmonary cavities differ from cysts in that their walls are typically thicker (usually > 4 mm).³

Emphysema differs from cystic lung disease as it typically leads to focal areas or regions of decreased lung attenuation that do not have defined walls.¹

Honeycombing refers to a cluster or row of cysts, 1 to 3 mm in wall thickness and typically 3 to 10 mm in diameter, that are associated with end-stage lung fibrosis.¹ They are typically subpleural in distribution and are accompanied by fibrotic features such as reticulation and traction bronchiectasis.¹

Bronchiectasis is dilation and distortion of bronchi and bronchioles and can be mistaken for cysts when viewed en face.¹

Loculated pneumothoraces can also mimic pulmonary cysts, but they typically fail to adhere to a defined anatomic unit and are subpleural in distribution.

STEP 2: CHARACTERIZE THE CLINICAL PRESENTATION

Clinical signs and symptoms of cystic lung disease play a key role in diagnosis (Table 1). For instance, spontaneous pneumothorax is commonly associated with diffuse cystic lung disease (lymphangioleiomyomatosis and Birt-Hogg-Dubé syndrome), while insidious dyspnea, with or without associated pneumothorax, is usually associated with the interstitial pneumonias (lymphocytic interstitial pneumonia and desquamative interstitial pneumonia).

In addition, congenital abnormalities of the lung can lead to cyst formation. These abnormalities, especially when associated with other congenital abnormalities, are often diagnosed in the prenatal and perinatal periods. However, some remain undetected until incidentally found later in adulthood or if superimposing infection develops.

Primary pulmonary infections can also cause parenchymal necrosis, which in turn cavitates or forms cysts.⁴

Lastly, cystic lung diseases can occur as part of a multiorgan or systemic illness in which the lung is one of the organs involved. Although usually diagnosed before the discovery of cysts or manifestations of pulmonary symptoms, they can present as a diagnostic challenge, especially when lung cysts are the initial presentation.bsence of amyloid fibrils.

In view of the features of the different types of cystic lung disease, adults with cystic lung disease can be grouped according to their typical clinical presentations (Table 2):

• Insidious dyspnea or spontaneous pneumothorax
• Incidentally found cysts or recurrent pneumonia
• Signs and symptoms of primary pulmonary infection
• Signs and symptoms that are primarily nonpulmonary.

Insidious dyspnea or spontaneous pneumothorax

Insidious dyspnea or spontaneous pneumothorax can be manifestations of lymphangioleiomyomatosis, Birt-Hogg-Dubé syndrome, pulmonary Langerhans cell histiocytosis, desquamative interstitial pneumonia, or lymphocytic interstitial pneumonia.

Lymphangioleiomyomatosis is characterized by abnormal cellular proliferation within the lung, kidney, lymphatic system, or any combination.⁵ The peak prevalence is in the third to fourth decades of life, and most patients are women of childbearing age.⁶ In addition to progressive dyspnea on exertion and pneumothorax, other signs and symptoms include hemoptysis, nonproductive cough, chylous pleural effusion, and ascites.^7,8

Birt-Hogg-Dubé syndrome is caused by germline mutations in the folliculin (FLCN) gene.⁹ It is characterized by skin fibrofolliculomas, pulmonary cysts, spontaneous pneumothorax, and renal cancer.¹⁰

Pulmonary Langerhans cell histiocytosis is part of the spectrum of Langerhans cell histiocytosis that, in addition to the lungs, can also involve the bone, pituitary gland, thyroid, skin, lymph nodes, and liver.¹¹ It occurs almost exclusively in smokers, affecting individuals in their 20s and 30s, with no gender predilection.^12,13 In addition to nonproductive cough and dyspnea, patients can also present with fever, anorexia, and weight loss,¹³ but approximately 25% of patients are asymptomatic.¹⁴

Desquamative interstitial pneumonia is an idiopathic interstitial pneumonia that, like pulmonary Langerhans cell histiocytosis, is seen almost exclusively in current or former smokers, who account for about 90% of patients with this disease. It affects almost twice as many men as women.^15,16 The mean age at onset is 42 to 46.^15,16 In addition to insidious cough and dyspnea, digital clubbing develops in 26% to 40% of patients.^16,17

Lymphocytic interstitial pneumonia is another rare idiopathic pneumonia, usually associated with connective tissue disease, Sjögren syndrome, immunodeficiencies, and viral infections.^18­–21 It is more common in women, presenting between the 4th and 7th decades of life, with a mean age at diagnosis of 50 to 56.^18,22 In addition to progressive dyspnea and cough, other symptoms include weight loss, pleuritic pain, arthralgias, fatigue, night sweats, and fever.²³

In summary, in this clinical group, lymphangioleiomyomatosis and Birt-Hogg-Dubé syndrome should be considered when patients present with spontaneous pneumothorax; those with Birt-Hogg-Dubé syndrome also present with skin lesions or renal cancer. In patients with progressive dyspnea and cough, lymphocytic interstitial pneumonia should be considered in those with a known history of connective tissue disease or immunodeficiency. Pulmonary Langerhans cell histiocytosis typically presents at a younger age (20 to 30 years old) than desquamative interstitial pneumonia (smokers in their 40s). Making the distinction, however, will likely require imaging with computed tomography.

Incidentally found cysts or recurrent pneumonia

Incidentally found cysts or recurrent pneumonia can be manifestations of congenital pulmonary airway malformation, pulmonary sequestration, or bronchogenic cyst.

Congenital pulmonary airway malformation, of which there are five types, is the most common pulmonary congenital abnormality. It accounts for up to 95% of cases of congenital cystic lung disease.^24,25 About 85% of cases are detected in the prenatal or perinatal periods.²⁶ Late-onset congenital pulmonary airway malformation (arising in childhood to adulthood) presents with recurrent pneumonia in about 75% of cases and can be misdiagnosed as lung abscess, pulmonary tuberculosis, or bronchiectasis.²⁷

Pulmonary sequestration, the second most common pulmonary congenital abnormality, is characterized by a portion of lung that does not connect to the tracheobronchial tree and has its own systemic arterial supply.²⁴ Intralobar sequestration, which shares the pleural investment with normal lung, accounts for about 80% of cases of pulmonary sequestration.^28–30 In addition to signs or symptoms of pulmonary infection, patients with pulmonary sequestration can remain asymp-
tomatic (about 25% of cases), or can present with hemoptysis or hemothorax.^28–30 In adults, the typical age at presentation is between 20 and 25.^29,30

Bronchogenic cyst is usually life-threatening in children. In adults, it commonly causes cough and chest pain.³¹ Hemoptysis, dysphagia, hoarseness, and diaphragmatic paralysis can also occur.^32,33 The mean age at diagnosis in adults is 35 to 40.^31,32

In summary, most cases of recurrent pneumonia with cysts are due to congenital pulmonary airway malformation. Pulmonary sequestration is the second most common cause of cystic lung disease in this group. Bronchogenic cyst is usually fatal in fetal development; smaller cysts can go unnoticed during the earlier years and are later found incidentally as imaging abnormalities in adults.

Signs and symptoms of primary pulmonary infections

Signs and symptoms of primary pulmonary infections can be due to Pneumocystis jirovecii pneumonia or echinococcal infections.

P jirovecii pneumonia commonly develops in patients with human immunodeficiency virus infection and low CD4 counts, recipients of hematologic or solid-organ transplants, and those receiving immunosuppressive therapy (eg, glucocorticoids or chemotherapy).

Echinococcal infections (with Echinococcus granulosus or multilocularis species) are more common in less-developed countries such as those in South America or the Middle East, in China, or in patients who have traveled to endemic areas.³⁴

In summary, cystic lung disease in patients with primary pulmonary infections can be diagnosed by the patient’s clinical history and risk factors for infections. Those with human immunodeficiency virus infection and other causes of immunodeficiency are predisposed to P jirovecii pneumonia. Echinococcal infections occur in those with a history of travel to an endemic area.

Primarily nonpulmonary signs and symptoms

If the patient has primarily nonpulmonary signs and symptoms, think about pulmonary amyloidosis, light chain deposition disease, and neurofibromatosis type 1.

Pulmonary amyloidosis has a variety of manifestations, including tracheobronchial disease, nodular parenchymal disease, diffuse or alveolar septal pattern, pleural disease, lymphadenopathy, and pulmonary cysts.⁴

Light chain deposition disease shares some clinical features with amyloidosis. However, the light chain fragments in this disease do not form amyloid fibrils and therefore do not stain positively with Congo red. The kidney is the most commonly involved organ.⁴

Neurofibromatosis type 1 is characterized by collections of neurofibromas, café-au-lait spots, and pigmented hamartomas in the iris (Lisch nodules).³⁵

In summary, patients in this group typically present with complications related to systemic involvement. Those with neurofibromatosis type 1 present with ophthalmologic, dermatologic, and neurologic manifestations. Amyloidosis and light chain deposition disease most commonly involve the renal system; their distinction will likely require tissue biopsy and Congo-red staining.

STEP 3: CHARACTERIZE THE RADIOGRAPHIC FEATURES

Characterization of pulmonary cysts and their distribution plays a key role in the diagnosis. Radiographically, cystic lung diseases can be subclassified into two major categories according to their cystic distribution:

• Discrete (focal or multifocal)
• Diffuse (unilobular or panlobular).^2,3

Discrete cystic lung diseases include congenital abnormalities, infectious diseases, and interstitial pneumonias.^2,3

Diffuse, panlobular cystic lung diseases include lymphangioleiomyomatosis, pulmonary Langerhans cell histiocytosis, Birt-Hogg-Dubé syndrome, amyloidosis, light chain deposition disease, and neurofibromatosis type 1.^7,13,36–39

In addition, other associated radiographic findings play a major role in diagnosis.

Cysts in patients presenting with insidious dyspnea or spontaneous pneumothorax

Lymphangioleiomyomatosis. Cysts are seen in nearly all cases of advanced lymphangioleiomyomatosis, typically in a diffuse pattern, varying from 2 mm to 40 mm in diameter, and uniform in shape (Figure  2A).^7,8,40–42

Other radiographic features include vessels located at the periphery of the cysts (in contrast to the centrilobular pattern seen with emphysema), and chylous pleural effusions (in about 22% of patients).⁴⁰ Nodules are typically not seen with lymphangioleiomyomatosis, and if found represent type 2 pneumocyte hyperplasia.

Pulmonary Langerhans cell histiocytosis. Nodules measuring 1 to 10 mm in diameter and favoring a centrilobular location are often seen on computed tomography. Pulmonary cysts occur in about 61% of patients.^13,43 Cysts are variable in size and shape (Figure 2B), in contrast to their uniform appearance in lymphangioleiomyomatosis. Most cysts are less than 10 mm in diameter; however, they can be up to 80 mm.^13,43 Early in its course, nodules may predominate in the upper and middle lobes. Over time, diffuse cysts become more common and can be difficult to differentiate from advanced smoking-induced emphysema.⁴⁴

Birt-Hogg-Dubé syndrome. Approximately 70% to 100% of patients with Birt-Hogg-Dubé syndrome will have multiple pulmonary cysts detected on computed tomography. These cysts are characteristically basal and subpleural in location, with varying sizes and irregular shapes in otherwise normal lung parenchyma (Figure 2C).^36,45,46

Desquamative interstitial pneumonia. Pulmonary cysts are present on computed tomography in about 32% of patients.⁴⁷ They are usually round and less than 20 mm in diameter.⁴⁸ Ground-glass opacity is present in almost all cases of desquamative interstitial pneumonia, with a diffuse pattern in 25% to 44% of patients.^16,17,47

Pulmonary cysts occur in up to two-thirds of those with lymphocytic interstitial pneumonia. Cysts are usually multifocal and perivascular in distribution and have varying sizes and shapes (Figure 2D).²² Ground-glass opacity and poorly defined centrilobular nodules are also frequently seen. Other computed tomographic findings include thickening of the bronchovascular bundles, focal consolidation, interseptal lobular thickening, pleural thickening, and lymph node enlargement.²²

In summary, in this group of patients, diffuse panlobular cysts are due to lymphangioleiomyomatosis, pulmonary Langerhans cell histiocytosis, or Birt-Hogg-Dubé syndrome. Cysts due to lymphangioleiomyomatosis have a diffuse distribution, while those due to pulmonary Langerhans cell histiocytosis tend to be upper-lobe-predominant and in the early stages are associated with stellate centrilobular nodules. Cysts in Birt-Hogg-Dubé syndrome tend to be subpleural and those due to lymphocytic interstitial pneumonia are perivascular in distribution.

Cysts that are incidentally found or occur in patients with recurrent pneumonia

Congenital pulmonary airway malformation types 1, 2, and 4 (Figure 3A, 3B). Cysts are typically discrete and focal or multifocal in distribution, but cases of multilobar and bilateral distribution have also been reported.^27,49 The lower lobes are more often involved.⁴⁹ Cysts vary in size and shape and can contain air, fluid, or both.^27,49 Up to 50% of cases can occur in conjunction with pulmonary sequestration.⁵⁰

Pulmonary sequestration displays an anomalous arterial supply on computed tomography (Figure 3C). Other imaging findings include mass lesions (49%), cystic lesions (29%), cavitary lesions (12%), and bronchiectasis.³⁰ Air trapping can be seen in the adjacent lung. Lower lobe involvement accounts for more than 95% of total cases of sequestration.³⁰ The cysts are usually discrete or focal in distribution. Misdiagnosis of pulmonary sequestration is common, and can include pulmonary abscess, pneumonia, bronchiectasis, and lung cancer.³⁰

Bronchogenic cyst. Cyst contents generally demonstrate water attenuation, or higher attenuation if filled with proteinaceous/mucoid material or calcium deposits; air-fluid levels are seen in infected cysts.³² Intrapulmonary cysts have a predilection for the lower lobes and are usually discrete or focal in distribution.^31,32 Mediastinal cysts are usually homogeneous, solitary, and located in the middle mediastinum.³² Cysts vary in size from 20 to 90 mm, with  a mean diameter of 40 mm.³¹

In summary, in this group of cystic lung diseases, characteristic computed tomographic findings will suggest the diagnosis—air-filled cysts of varying sizes for congenital pulmonary airway malformation and anomalous vascular supply for pulmonary sequestration. Bronchogenic cysts will tend to have water or higher-than-water attenuation due to proteinaceous-mucoid material or calcium deposits.

Cysts in patients with signs and symptoms of primary pulmonary infections

P jirovecii pneumonia. Between 10% and 15% of patients have cysts, and about 18% present with spontaneous pneumothorax.⁵¹ Cysts in P jirovecii pneumonia vary in size from 15 to 85 mm in diameter and tend to occur in the upper lobes (Figure 4A).^51,52

Echinococcal infection. Echinococcal pulmonary cysts typically are single and located more often in the lower lobes (Figure 4B).^53,54 Cysts can be complicated by air-fluid levels, hydropneumothorax, or pneumothorax, or they can turn into cavitary lesions.

The diagnoses of these pulmonary infections are usually made by clinical and computed tomographic findings and depend less on detecting and characterizing lung cysts. Patients with P jirovecii pneumonia tend to have bilateral perihilar ground-glass opacities, while air-fluid levels suggest echinococcal infections. Cysts in this group of patients tend to be discrete or focal or multifocal in distribution, and vary in size.

Cysts in patients with primarily nonpulmonary signs and symptoms

Amyloidosis. Cyst formation is rare in amyloidosis.⁴ When present, cysts can be diffuse and scattered in distribution, in varying sizes (usually < 30 mm in diameter) and irregular shapes (Figure 5).^55,56

Pulmonary light chain deposition disease usually presents as linear opacities and small nodules on chest computed tomography. Numerous cysts that are diffuse in distribution and have no topographic predominance can also be present. They can progress in number and size and coalesce to form irregular shapes.⁵⁷

Neurofibromatosis type 1. In neurofibromatosis type 1, the most common radiographic presentations are bibasilar reticular opacities (50%), bullae (50%), and ground glass opacities (37%).⁵⁸ Well-formed cysts occur in up to 25% of patients and tend to be diffuse and smaller (2 to 18 mm in diameter), with upper lobe predominance.^58,59

In summary, in this group of patients, bibasilar reticular and ground-glass opacities suggest neurofibromatosis type 1, while nodules and linear opacities suggest amyloidosis or light chain deposition disease. Cysts tend to be diffuse with varying sizes.

STEP 4: PUT IT ALL TOGETHER

Diagnosis in insidious dyspnea or spontaneous pneumothorax

For patients who present with insidious dyspnea or spontaneous pneumothorax, the diagnosis of cystic lung disease can be made by characterizing the distribution, size, and shape of the cysts (Table 3).

Diffuse, panlobular distribution. Cystic lung diseases with this pattern include lymphangioleiomyomatosis, pulmonary Langerhans cell histiocytosis, and Birt-Hogg-Dubé syndrome. In this group, cysts that are uniform in size and regular in shape are invariably due to lymphangioleiomyomatosis. Those with variable size and irregular shapes can be due to pulmonary Langerhans cell histiocytosis or Birt-Hogg-Dubé syndrome. Patients with pulmonary Langerhans cell histiocytosis tend to be smokers and their cysts tend to be upper- lobe-predominant. Those with Birt-Hogg-Dubé syndrome will likely have renal cancer or skin lesions; their cysts tend to be basilar and subpleural in distribution.