Transient global amnesia

INTRODUCTION — Transient global amnesia (TGA) is a clinical syndrome characterized by the acute onset of anterograde amnesia (the inability to form new memories) [1-3].

Patients with TGA frequently ask repetitive questions reflecting disorientation and may have variable inability to recall general or personal information (retrograde amnesia) while the episode lasts. During the episode of TGA, other cognitive functions are normal. Episodes are self-limited, and by definition resolve within 24 hours, with recovery of the symptoms of memory function, except for what happened during the episode. A minority of patients experience a few recurrent episodes. A single, definite etiology has not been determined, although epidemiologic and imaging data support several putative pathophysiologic processes, including vascular, migraine, epileptic, and psychogenic mechanisms. Management is directed at identifying the syndrome, diagnosing and treating other disorders that may present similarly with memory deficits, and reassuring patients and family of the benign prognosis of TGA.

This topic discusses TGA. The evaluation of memory impairment and memory impairment in other settings is discussed separately:

●(See "The mental status examination in adults", section on 'Memory'.)

●(See "Mild cognitive impairment: Epidemiology, pathology, and clinical assessment".)

●(See "Clinical features and diagnosis of Alzheimer disease", section on 'Memory impairment'.)

Other causes of transient neurologic symptoms are also discussed separately:

●(See "Definition, etiology, and clinical manifestations of transient ischemic attack".)

●(See "Initial evaluation and management of transient ischemic attack and minor ischemic stroke".)

●(See "Evaluation and management of the first seizure in adults".)

●(See "Nonepileptic paroxysmal disorders in adolescents and adults".)

EPIDEMIOLOGY

Incidence — In the general population, the incidence of TGA has been estimated as 5.2 to 10 per 100,000 per year [4-7]. It is uncommon in individuals less than 50 years of age; among those 50 years and older, the incidence is 23.5 to 32 per 100,000 per year.

Risk factors — TGA occurs primarily in older adults; the majority of episodes occur in individuals between the ages of 50 and 80 years, with a mean age of onset between 60 and 65 years [2,4]. The incidence of TGA does not differ according to gender [2,5,6].

Although cerebral ischemia is one of the postulated causes of TGA, it is controversial whether cerebrovascular risk factors increase the risk of TGA. Many case-control studies find no difference in the prevalence of atherosclerotic risk factors such as hypertension, diabetes, and hypercholesterolemia between TGA patients and age- and gender-matched controls [2,8-15]. By contrast, in one study based on the National Inpatient Sample (58 million hospital cases from 1998 to 2008), patients with TGA were almost twice as likely to have hypertension (51 versus 28 percent) and three times as likely to have hyperlipidemia (34 versus 13 percent) as the inpatient population overall [16]. In another large study comparing TGA patients, transient ischemic attack (TIA) patients, and matched controls, TGA patients were more likely than age- and sex-matched controls to have hyperlipidemia, previous ischemic stroke, and ischemic heart disease. However, compared with the TIA patients, those with TGA were less likely to have hypertension, diabetes, a history of ischemic stroke, and atrial fibrillation [17].

Migraine has been found to be strongly associated with TGA in some, but not all, case-control studies [2,9,10,15,18,19]. In the National Inpatient Sample, migraine was almost six times more prevalent in those with TGA [16]. Among patients with recurrent TGA events, 37.5 percent report migraine compared with 14 percent of those with no recurrence [20].

There are case reports of the familial occurrence of TGA [21]. This has not been reported in larger case series, and it has been suggested the familial cases were associated with familial migraine [22].

PATHOPHYSIOLOGY

Clinical neuroanatomy — The clinical symptoms of TGA suggest that the primary site of neurologic involvement is the mediobasal temporal lobe and hippocampus [7,23,24]. Functional neuroimaging studies provide evidence that this area is involved in the formation and retrieval of new episodic memories, the primary neurologic deficit in TGA [25,26]. Clinical neuroimaging studies also demonstrate that this area is pathologically affected in TGA. (See 'Neuroimaging findings' below.)

Neurons in the CA1 area of the hippocampus are known to be selectively vulnerable to cellular metabolic stress, specifically in the setting of hypoxia and ischemia [8,27,28]. This vulnerability may relate to sensitivity to cytotoxic glutamatergic release and uptake, or to the vascular anatomy of the hippocampus, as this area may represent a watershed between two major arteries.

In addition to hippocampal involvement, altered connectivity among other brain regions, especially those in the default mode network in regions of the orbitofrontal, parietal, and angular gyri, has been found in 20 TGA patients (compared with 38 matched controls) using diffusion tensor imaging and graph theory analysis [29]. Another study of 16 acute TGA patients, 20 patients with prior TGA, and 20 matched controls demonstrated reduced functional connectivity between the hippocampus and executive and salience network hubs (frontal and prefrontal regions, cingulate gyrus, parietal and subcortical regions) using resting state functional magnetic resonance imaging (MRI) [30].

Potential etiologic mechanisms — There is no consensus regarding a common underlying pathogenesis of TGA [7,31-37]. Hypotheses include a vascular etiology (arterial or venous), a migrainous phenomenon (or cortical spreading depression), epilepsy, and a psychogenic disorder. None of these hypotheses clearly and consistently explain the clinical features of TGA. It is possible that different mechanisms play a role in some but not all patients with TGA.

●Arterial ischemia – TGA shares some features with transient ischemic attack (TIA), in that both produce neurologic symptoms that resolve. Both TIA and TGA occur more commonly in older adults. Some (but not all) studies have found that cerebrovascular risk factors are prevalent in TGA (see 'Epidemiology' above). However, while patients with TIA have an increased risk for recurrent TIA and stroke, this risk does not appear to be increased in patients after TGA [4,10,15,38-40]. Findings from structural and functional neuroimaging studies provide data that both support and refute arterial ischemia as a pathogenesis for TGA. (See 'Neuroimaging findings' below.)

Some episodes of TGA do appear to be stroke related, but these appear to be a minority (a few percent) of patients with TGA. Such patients are in general considered to have an amnestic cerebrovascular event (ischemic stroke or TIA) and not to have idiopathic TGA. (See 'Differential diagnosis' below.)

●Venous congestion – Venous congestion was proposed as pathogenesis for TGA based, in part, on the observation that specific events appear to trigger TGA frequently [36]. These include Valsalva maneuver, in as much as 41 percent of cases [20]; vigorous activity of the arms; sympathetic activation from acute emotional events; and cold-water immersion (diving reflex). In this paradigm, increases in either intrathoracic pressure or venous return could lead to central venous congestion affecting the deep venous system that drains the dorsomedial thalamus and hippocampus [27,36].

Support for this theory came from a series of Doppler ultrasound studies that demonstrated inducible venous reflux in at least one internal jugular vein in a higher prevalence of TGA patients compared with controls [7,9,11,36,41,42]. The largest of these series included 142 patients with TGA and 40 controls [41]. Valvular insufficiency was found in 80 percent of the TGA patients and 25 percent of the controls. MRI studies have also reported retrograde venous flow in a higher percentage of TGA patients than controls [43,44]. These findings seemed particularly relevant in that the demonstration of retrograde flow was more common in TGA patients who reported an antecedent Valsalva maneuver, compared with those who did not [9,43].

Presumably, the internal jugular valvular insufficiency and propensity for retrograde venous flow is a premorbid condition that then predisposes patients to venous congestion and TGA. This hypothesis does not account for the low frequency of TGA with venous sinus thrombosis, the low frequency of recurrence, and the increased incidence with age [3,33].

●Migrainous phenomenon – Migraine can cause transient neurologic symptoms (eg, migraine aura and migraine equivalents) and has been proposed as a cause of TGA. The mechanism of migraine aura is understood to be a self-propagating wave of neuronal and glial depolarization, so-called cortical spreading depression, which may trigger glutamate release in the hippocampus [45]. (See "Pathophysiology, clinical manifestations, and diagnosis of migraine in adults".)

In support of this theory is that in one study, TGA patients were five times more likely to have a migraine than controls [16]. Other, although not all, case-control studies have found a similar association [9,10,15,18,19,46] (see 'Risk factors' above). Some believe that the observed changes on diffusion-weighted MRI (DWI-MRI) and positron emission tomography (PET) are most compatible with migraine [33,47], while others disagree [36]. (See 'Neuroimaging findings' below.)

However, migraines tend to present in younger, not older, patients and are also a recurring phenomenon, in contrast with TGA, which usually occurs in older persons as a single event. Two exceptional case reports describe atypical patients with recurrent episodes of TGA that remitted with beta blockers or valproic acid, treatments often used in the prophylaxis of migraine [48,49].

●Epileptic phenomenon – While transient amnesia is a documented manifestation of some epileptic seizures, epilepsy is not believed to cause TGA. Seizures are recurring and brief (minutes rather than hours), and amnesia doesn't typically occur as an isolated phenomenon; moreover, cases of electroencephalography (EEG) monitoring during a TGA episode have not shown epileptic discharges [7,50,51].

Other clinical features that distinguish epileptic amnesia from TGA are described below. (See 'Differential diagnosis' below.)

●Psychosomatic disorder – When TGA was first described, a psychogenic origin was postulated. Most clinicians now believe that TGA is a neurologic disorder, but that psychological stressors and personality factors may play a contributing role [31]. Psychosomatic factors as a cause of or risk factor for TGA should not be confused with dissociative amnesia, which is a distinct entity. (See 'Differential diagnosis' below.)

Several studies have identified a higher prevalence of phobic personality traits, psychiatric or substance abuse disorders, and family history of psychiatric conditions in TGA compared with controls [2,38,52]. Further, as much as 23 percent of patients experience emotional stress as a trigger for the TGA episode [20]. A prospective study evaluating stress-related factors in TGA found that compared with controls, TGA patients were more sensitive to stress, were more anxious, and had less effective coping mechanisms to deal with stress [53].

A few patients with TGA have gone on to experience Takotsubo syndrome (TTS), a stress-induced cardiomyopathy [54]. In TTS, it is thought that emotional stress causes a catecholamine-induced cardiomyopathy. It is hypothesized that a similar mechanism may cause TGA. Resting functional MRI studies of TTS patients (without TGA) have shown decreased resting functional connectivity in parasympathetic and sympathetic networks, involving the amygdala, hippocampus, insula, cingulate, parietal and temporal lobe, and cerebellum, as well as decreased connectivity in the default mode network, involving the hippocampus, parahippocampus, and medial prefrontal region [55]. This regional neuroanatomic involvement overlaps with that hypothesized to play a role in TGA (see 'Clinical neuroanatomy' above). TTS is discussed separately. (See "Clinical manifestations and diagnosis of stress (takotsubo) cardiomyopathy".)

●Multiple causes – Because no single theory of its pathogenesis clearly applies to all patients with TGA, some speculate that, despite its relatively consistent clinical presentation, TGA is a syndrome that has multiple different etiologies [2,34,56,57].

CLINICAL PRESENTATION — Episodes of TGA are discrete events, usually with a well-defined and abrupt time of onset. Case series describe clinical features that appear to be relatively consistent [3,58].

Symptoms and signs — Prominent anterograde amnesia, an inability to form new memories, is the defining clinical feature. Patients are disoriented in time and often repetitively ask questions about the date or their environment, the so-called "broken-record" phenomenon. Immediate recall (eg, digit-span) is intact, but delayed recall (following an interval of at least several minutes during which mental rehearsal is prevented) is impaired [7]. Retrograde amnesia (recall of information learned before the onset of the episode) is common but more variable, and can extend back hours to days, weeks, and rarely years [8]. Semantic memory (knowledge of words and previously learned general information) remains intact, as does procedural or motor memory (playing a sport, tying shoelaces, or similar skilled actions) [3]. Patients are able to perform complex, previously learned tasks such as driving, cooking, lecturing, and playing a musical instrument that do not require new learning [1,7,58].

Patients often appear disorganized, confused, and perseverative and may be mistaken as having an acute confusional state or delirium. However, in TGA, cognitive functions other than memory are generally spared [3,58]. In particular, both impaired level of alertness and disorientation to self exclude TGA. Language comprehension and speech are also normal. Certain executive functions, such as reverse digit sequences and category or initial letter verbal fluency, may be affected in the acute phase [59].

Associated symptoms may be reported. Between 10 to 40 percent of patients complain of headache during the spell [2,8,10,18]. Nausea, undefined dizziness, anxiety, paresthesia, and other symptoms are reported by some patients.

The physical neurologic exam is normal, without focal signs or symptoms [3].

Duration — The mean duration of episodes is approximately six hours [2,4,6]. Most clinical symptoms last a few hours and by definition resolve within 24 hours. In one series of 114 cases of TGA, all resolved within 12 hours [58].

The symptoms resolve gradually. As the patient recovers anterograde memory function, the retrograde amnesia also resolves completely or is restricted to a brief time immediately preceding the event. Patients are left with an amnestic gap for the episode, since they cannot recall what they were unable to learn during the period of anterograde amnesia.

Triggers and clinical setting — Antecedent, presumably triggering events are reported to occur in 30 to 90 percent of patients [2,4,8-10,12,15,18,58,60,61]. The variable proportion of those with TGA who have apparent event triggers is probably due to how systematically such events were solicited and what types of events were included as potential triggers.

Reported event triggers have included acute and chronic emotional events, physical activity (eg, strenuous exercise, sexual intercourse), Valsalva maneuver, intense pain, postural changes, medical procedures, high altitude, contact with hot or cold water, and changes in body temperature [2,8,61-66]. A retrospective analysis of 203 cases determined that 66 percent had antecedent events [20]. Among those with triggers, Valsalva maneuver was most common (41 percent), followed by stress (23 percent) and intercourse (11 percent). These observations have not been confirmed (or refuted) with case-control analysis. However, the relevance of Valsalva maneuver as a triggering event is supported by the fact that inducible retrograde flow in the internal jugular vein is more common in patients who reported a preceding Valsalva maneuver than in those who did not [12]. (See 'Potential etiologic mechanisms' above.)

Several studies suggest that more than half of TGA attacks begin in the morning, after the patient has arisen without symptoms [2,19,38]. A history of a similar prior episode is reported in a minority of cases (5 to 9 percent) [2,8,15].

NEUROIMAGING FINDINGS — A neuroimaging study is often performed in patients with TGA. (See 'Evaluation and treatment' below.)

●Computed tomography – A computed tomography (CT) scan of the brain is often performed in the emergency department setting and is usually normal or may show incidental findings [58].

●Magnetic resonance imaging – Brain magnetic resonance imaging (MRI) may be normal or show only incidental findings, especially within the first few hours. Several case series report the finding of one or more punctate lesions (mean lesion size 4mm) in the hippocampus on diffusion-weighted MRI (DWI-MRI), with an incidence that ranges from 57 to 88 percent (image 1) [2,7,8,13,47,56,61,67-74]. These lesions are associated with small reductions in the apparent diffusion coefficient (ADC) values and are most often unilateral and left-sided; right-sided and bilateral lesions are also reported [7,67,74]. In patients who had acute neuroimaging and neuropsychological evaluation, evidence of lateralization is congruent in both tests; that is, prominent verbal memory loss was associated with dominant-side hippocampal lesions, while more prominent visuospatial memory deficits were associated with nondominant hippocampal lesions [8,23].

The timing of MRI and imaging technique may impact the likelihood of demonstrating these lesions; they are less common in the hyperacute phase and become more prevalent after 12 hours, with a peak incidence between 12 and 72 hours [67,68,70,74,75]. These lesions often persist for 7 to 10 days [58,67,73,76,77] and resolve on follow-up imaging performed one to six months after the event [8,47,56,67,71,78]. Other studies emphasize that the sensitivity of MRI also depends on imaging technique, with detection increased by up to 88 percent when MRI is performed within a two- to three-day window, using higher diffusion gradient strength (B = 2000) and 2 to 3 mm thin slices [79-81].

Generally, the lesions seen in TGA are not typical of those seen in transient ischemic attack (TIA), which are larger, are more likely to be observed in the first few hours of the onset of symptoms, and evolve into a permanent T2 abnormality [7,8,37,56,82,83]. However, the ability of MRI to clearly distinguish between TIA and TGA has not been formally assessed. One investigation observed that the lesions observed in TGA have higher ADC values than are usual for TIA-associated lesions [67]. Another study found that the lesions in TGA were not associated with perfusion deficits, as would be expected for an ischemic etiology [75]. Magnetic resonance spectroscopy demonstrates a lactate peak in these lesions, consistent with, but not specific for, an arterial ischemic etiology [28].

●Other modalities – Single-photon emission computed tomography (SPECT) scanning in cases of TGA has shown variable results, usually hypoperfusion in one or both temporal lobes [2,33]. Hyperperfusion or involvement of other brain areas is uncommonly reported [84,85]. These findings do not indicate whether the alteration in cerebral blood flow is a primary event or a secondary phenomenon related to changes in cerebral metabolism from another etiology [31].

During the TGA attack, 18-F fluorodeoxyglucose (FDG)-PET scans have shown abnormalities in several brain regions including the hippocampus, amygdala, lentiform nucleus, and frontal lobe [2,23,31,86]. However, findings appear to be consistent with stable cerebral blood flow and a reduction in regional oxygen uptake, suggesting altered neuronal activity rather than a vascular event [31].

INITIAL MANAGEMENT

Evaluation and treatment — The initial evaluation and management of patients during a TGA episode focus on excluding other diagnoses and should include the following:

●If the patient is symptomatic on presentation, the patient should be observed in the hospital (inpatient or emergency department) until the amnesia resolves, the evaluation for other causes is complete, and the patient can be discharged safely to the care of family or friends.

Fever is not seen with TGA and suggests an underlying infection or encephalitis. Marked vital sign abnormalities also require explanation with the caveat that anxiety from the event may cause mild tachycardia or hypertension. (See 'Differential diagnosis' below.)

●A detailed neurologic examination should be performed in order to demonstrate that the memory loss is an isolated finding. The following exam features are normal in patients with TGA; abnormalities suggest an alternative diagnosis as noted:

•Extraocular movements (abnormalities may be seen in brainstem stroke or Wernicke encephalopathy [WE]).

•Visual fields (deficits may be seen in temporal lobe stroke).

•Attention span as assessed by spelling "world" backward, for example (impaired in delirium due to toxic-metabolic encephalopathy or other cause).

•Knowledge of personal identity (lost in dissociative amnesia).

•Gait (abnormalities may be seen in stroke or WE).

•Other higher cognitive functions including procedural memory, such as showing how to tie shoelaces or fold a letter and put it in an envelope; visuospatial perception, such as drawing a clock; language function, such as naming, reading, and command-following (deficits in any of these suggest neurologic impairment beyond TGA).

The examination of these mental status domains is presented in more detail separately. (See "The mental status examination in adults", section on 'Mental status examination'.)

●Diagnostic testing includes oxygenation status, serum electrolytes, glucose, and a toxicology screen. These are performed to exclude serious medical conditions that can produce a delirium that might be mistaken for TGA.

●We suggest that patients receive thiamine 500 mg intravenously.

While patients with WE typically present with a global encephalopathy rather than a restricted amnestic syndrome, it is possible that a mild form of WE might be confused with TGA, and no test provides immediate confirmation or exclusion of WE. Thiamine treatment is benign and can be life-saving for WE. (See "Wernicke encephalopathy", section on 'Treatment'.)

●We also suggest a neuroimaging study in all patients, preferably noncontrast brain magnetic resonance imaging (MRI) with diffusion-weighted imaging (DWI), to exclude acute ischemia, head trauma, and other pathologies [82]. Contrast administration is not required. In the acute setting, if MRI is not available, computed tomography (CT) should be performed, particularly in patients who do not meet the diagnostic criteria. Some clinicians perform neuroimaging more selectively, for example, in patients who do not meet all diagnostic criteria and/or those who have vascular risk factors.

While transient ischemic attack (TIA) or ischemic stroke is an uncommon cause of TGA symptoms, particularly those with a carefully documented, otherwise normal neurologic examination, this diagnosis is associated with increased risk of future stroke, which can be mitigated if properly investigated and managed. (See "Initial evaluation and management of transient ischemic attack and minor ischemic stroke".)

●The need for further testing varies depending on the circumstances, such as how typical the event is for TGA, the presence of vascular risk factors, and whether the ictus was observed. (See 'Diagnosis' below.)

Patients with recurrent or brief episodes, or activity suggesting motor automatisms, should be evaluated with electroencephalography (EEG) for possible epilepsy [7]. (See 'Differential diagnosis' below and "Electroencephalography (EEG) in the diagnosis of seizures and epilepsy".)

Diagnosis — TGA is a clinical diagnosis; there are no confirmatory diagnostic tests. For a patient who presents with an acute episode of anterograde amnesia, a diagnosis of TGA is confidently made based on the clinical syndrome if all of the following features are present [1,58,87]:

●Abrupt onset of memory difficulty, witnessed by a reliable observer

●Prominent anterograde amnesia, usually accompanied by repetitive questions

●Some retrograde amnesia and mild difficulty on executive tasks may be present; other cognitive domains appear normal

●No antecedent head trauma or alteration of consciousness during the episode

●No loss of personal identity

●No focal neurologic symptoms or signs during or after the episode

●Clinical symptoms last at least 1 hour and resolve within 24 hours

An accurate, witnessed description of the onset and antecedent events is critical for establishing a diagnosis of TGA. A thorough mental status evaluation and neurologic examination is also necessary since the pattern of cognitive deficits and presence or absence of other neurologic abnormalities are critical for differentiating TGA from other conditions. Some published diagnostic criteria also require that there is no history of epilepsy or new recent history of seizures or currently prescribed antiseizure medications [58].

When diagnostic criteria are not met, then there is an imperative to consider entities in the differential diagnosis and, when appropriate, exclude with specific testing as discussed in the next section.

Differential diagnosis — It is usually, but not always, possible to distinguish TGA with its relatively isolated deficit of amnesia from the more global impairments seen in patients with delirium. (See 'Symptoms and signs' above and 'Evaluation and treatment' above.)

However, in some patients, it can be difficult to differentiate TGA from an acute confusional state, such as when a patient with TGA is agitated and does not cooperate with examination. In these cases, it is important to fully evaluate the patient for potential causes of acute delirium including anoxia, metabolic encephalopathy (including WE), intoxication, withdrawal syndromes, encephalitis, or other neurologic or systemic causes. In these patients, investigations should include a full work-up for delirium and structural brain disease that should include lumbar puncture if blood work and MRI are not revealing. In addition, a duration of more than 24 hours excludes the diagnosis of TGA; such patients should undergo a similar evaluation. (See "Diagnosis of delirium and confusional states" and "Acute toxic-metabolic encephalopathy in adults".)

For other patients with more typical symptoms of TGA, amnestic stroke, epileptic amnesia, and dissociative disorder should be considered:

●TIA or stroke – Isolated amnesia without other focal neurologic deficits is a rare manifestation of TIA or stroke, estimated between 0.2 and 1.2 percent in different case series [82]. The presence of vascular risk factors should raise concern for this diagnosis [10,82].

There have been well-described stroke syndromes that produce acute amnesia similar to that seen in TGA; this is often, although not always, accompanied by other neurologic symptoms [88-92]. There are case reports of patients with a clinical syndrome compatible with TGA who had evidence of ischemic stroke on imaging studies with lesions in the mesial temporal lobe [82,88,91,93,94], thalamus [82,95], caudate nucleus [82,89], cingulate gyrus [92], and splenium of the corpus callosum [82,90]. While the neurologic examination is sometimes reported to be otherwise normal in these cases, visual field abnormalities are common in strokes involving the temporal lobe [91]. In one study of 20 patients with ischemic stroke in which acute amnesia was the predominant symptom, a diagnosis of ischemic stroke was not considered initially in five patients who were presumed to have TGA [82]. In retrospect, just two of these patients met all clinical TGA criteria, with other patients having other minor neurologic signs on detailed examination and/or a prolonged duration of symptoms.

A variety of stroke mechanisms have been implicated. Well-documented episodes of TGA-like symptoms have occurred in settings associated with acute cerebrovascular thromboembolism, such as vertebral artery angioplasty and stenting [96], coronary angiography [82], basilar artery thrombosis [97], aortic dissection [98], cardiogenic embolism [82], and hypercoagulable conditions [99-101]. In one series of 13 patients, cardiogenic embolism was the most common stroke mechanism, and most strokes involved the posterior circulation or were multiterritorial [82].

Brain MRI with DWI is performed in most cases of TGA to exclude TIA and stroke [82]. (See "Initial evaluation and management of transient ischemic attack and minor ischemic stroke", section on 'Brain imaging'.)

●Seizure, epileptic amnesia – Features that suggest this diagnosis include a short duration of the attack (less than one hour, usually a few to several minutes), occurrence on awakening from sleep, and multiple attacks at the time of presentation [10,51,58,102-105]. Rare cases of transient epileptic amnesia have been reported to last hours or days [102]. Common accompanying symptoms, which may not be volunteered by patients, include olfactory hallucinations and motor automatisms [103]. Amnestic seizures are less likely than TGA to be accompanied by anxiety and repetitive questioning, and during the ictus, retrograde amnesia is more prominent than anterograde amnesia [105,106]. Amnestic seizures are believed to arise from the mesial temporal lobe; some patients with epileptic amnesia also have interictal memory deficits [104,107]. Ictal and postictal MRI abnormalities can overlap with the findings observed with TGA. (See "Magnetic resonance imaging changes related to acute seizure activity".)

In general, we consider this diagnosis and obtain an EEG when symptoms are suggestive of seizure, when episodes are recurrent, or when a lapse of consciousness at the onset cannot be confidently ruled out by the history. Normal or nonspecific changes on an interictal EEG do not exclude the diagnosis. An EEG performed during sleep or with sleep deprivation may improve sensitivity. Prolonged EEG monitoring may be helpful; in some patients the diagnosis is made when recurrent events cease during an antiseizure medication trial [108]. (See "Focal epilepsy: Causes and clinical features", section on 'Mesial temporal lobe epilepsy' and "Electroencephalography (EEG) in the diagnosis of seizures and epilepsy".)

●Psychogenic (dissociative) amnesia – Dissociative amnesia should not be confused with TGA. Psychogenic amnesia is characterized by prominent retrograde autobiographical amnesia, sometimes even for one's own name, and is often associated with a fugue state, features which are atypical for TGA [109-111]. Dissociative amnesia is described in detail separately. (See "Dissociative amnesia: Epidemiology, pathogenesis, clinical manifestations, course, and diagnosis", section on 'Clinical manifestations'.)

LONG-TERM TREATMENT AND PROGNOSIS

Management — Treatment is not required for TGA. Because of the frightening nature of the symptoms, reassurance of the patient and family that the condition is benign and self-limited is necessary. The condition usually does not recur, and the patient does not need to be restricted from driving unless events are recurrent.

It is prudent to manage stroke risk factors according to current guidelines; however, unless the evaluation is specifically suggestive of an ischemic etiology (eg, magnetic resonance imaging [MRI] findings consistent with acute ischemia and/or additional neurologic deficits), we do not perform an evaluation for stroke subtype or prescribe antithrombotic therapy unless there is an alternative indication to do so [3]. (See "Overview of secondary prevention of ischemic stroke".)

Recurrence — Recurrence rates range from 2.9 to 26.3 percent in studies with variable lengths of follow-up; the annual rate is estimated to be 2.5 to 5.8 percent [2,4,10,12,15,18,31,38,58,112]. Frequent, multiple episodes would bring into question the diagnosis of TGA.

Long-term neurologic outcomes — There is no increased risk of mortality, epilepsy, stroke, or dementia following TGA as compared with age-matched control subjects [4,10,15,38-40,113].

However, despite the acute resolution of clinical symptoms, some studies suggest that the effects of TGA may be longer lasting and less benign than previously thought. Follow-up neuropsychologic testing in some case series of TGA patients suggests that subclinical residual deficits may be common one or more years later, while others have found no significant deficits [8,114,115]. Case-control studies have also had mixed results. In one study, 55 patients, evaluated one year after an episode of TGA, had worse performance on tests evaluating memory and attention compared with a group of 80 age-matched controls [116]. One-third of the TGA patients met criteria for amnestic mild cognitive impairment. However, another smaller case-control study reported no significant difference in neuropsychological test results between patients and controls at a mean of three years after the TGA event [117]. A meta-analysis of effect sizes from 25 studies found that impairments in anterograde memory and other cognitive spheres persisted up to five days but had completely resolved in the long-term phase (5 to 30 days) [59].

Two case series document a total of five patients with primary progressive aphasia (PPA) who experienced one or more episodes of TGA up to 10 years before, or 2 years after, the language impairment began [118,119]. This raises the question of whether TGA predisposes to PPA and also whether left temporal lobe degeneration contributes to increased vulnerability to TGA [120].

SUMMARY AND RECOMMENDATIONS

●Transient global amnesia (TGA) is an uncommon, benign, self-limited, and relatively selective disturbance of anterograde amnesia that occurs in middle-aged and older subjects. While some studies find that migraine and cerebrovascular risk factors are associated with TGA, the pathogenesis of this disorder remains uncertain. (See 'Epidemiology' above and 'Pathophysiology' above.)

●The event is characterized by a profound disruption in anterograde memory; patients are typically disoriented to place and will repeatedly ask where they are and why they are there, the so-called "broken-record" phenomenon. Retrograde amnesia is variably present. There may be deficits in executive function, but most other cognitive functions are intact. Patients do not experience altered consciousness or lose self-awareness. Resolution typically occurs after several hours, and always in less than 24 hours. (See 'Clinical presentation' above.)

●Clinical features of TGA are generally distinct from other metabolic encephalopathy, drug withdrawal, and other causes of delirium. Other causes of an acute amnestic syndrome include cerebrovascular ischemia, seizure, and psychogenic (dissociative) amnesia. (See 'Differential diagnosis' above.)

●Patients with TGA should be observed until the episode resolves. Investigations in the acute care setting should include laboratory tests to exclude intoxication, infection, and metabolic derangements.

We suggest that symptomatic patients with TGA receive thiamine 500 mg intravenously (Grade 2C).

In patients who do not meet diagnostic criteria for TGA, or who do not evolve as one would expect, we suggest performing brain magnetic resonance imaging (MRI) with diffusion-weighted imaging (DWI) to exclude stroke or other structural causes, and other diagnostic testing as above. (See 'Evaluation and treatment' above.)

●TGA is a clinical diagnosis, and no biomarkers or neuroimaging findings are diagnostic. Patients receive the diagnosis of TGA based upon the presence of the following clinical features (see 'Diagnosis' above):

•Abrupt onset of memory difficulty, witnessed by a reliable observer

•Prominent anterograde amnesia, usually accompanied by repetitive questions

•Some retrograde amnesia and mild difficulty on executive tasks may be present; other cognitive domains appear normal

•No antecedent head trauma or alteration of consciousness during the episode

•No loss of personal identity

•No focal neurologic symptoms or signs during or after the episode

•Clinical symptoms last at least 1 hour and resolve within 24 hours

●Patients with evidence of cerebral ischemia should be referred for additional evaluation and management of cerebrovascular risk factors. (See "Initial evaluation and management of transient ischemic attack and minor ischemic stroke".)

●The prognosis after idiopathic TGA is good, with a low risk of recurrence or other subsequent neurologic events. No treatment or driving restriction is required. (See 'Long-term treatment and prognosis' above.)

ACKNOWLEDGMENT — The editorial staff at UpToDate would like to acknowledge Patricia H Davis, MD, and Sarah Kremen, MD, who contributed to an earlier version of this topic review.