Clinical Course and Recurrence in Transient Global Amnesia: A Study From the TEMPiS Telestroke Network

Transient global amnesia (TGA) is a clinical syndrome characterized by a sudden, reversible loss of anterograde and retrograde memory for up to around 24 hours.1, 2 New information cannot be memorized during the acute episode, which typically results in repetitive questions by the patients.3 The main clinical characteristic of TGA is that consciousness and procedural memory remain unaffected.1, 2 The clinical diagnosis of TGA is based on the criteria of Hodges and Warlow1, 2 and the exclusion of important differential diagnoses.4 TGA may be induced by external triggers such as cold water, physical exercise, or emotional stress, pain, or sexual activity, but may also occur spontaneously.3

The cause of TGA is still not well understood, and various potential pathomechanisms have been considered. Functional MRI has revealed the involvement of the hippocampal cornu ammonis 1 (CA1) region in amnestic episodes.5 This is consistent with the clinical syndrome but does not explain the detailed pathophysiology underlying it. One potential mechanism discussed is the mechanism of “cortical spreading depression,” which is also observed in migraine.6 Moreover, the susceptibility of CA1 neurons to stress factors suggests that excitotoxicity via mechanisms such as glutamine release could contribute.7, 8 Furthermore, vascular mechanisms such as focal ischemia in the hippocampus or venous congestion causing a transient intracranial hypertension are discussed as potential mechanisms, supported by the appearance of hyperintense hippocampal changes indicating diffusion restriction on diffusion-weighted imaging (DWI) MRI.9, 10

Given the acute onset of the TGA, symptoms may lead to an initial differential diagnosis of acute ischemic stroke; for example, an embolic stroke in the posterior cerebral artery territory. TGA potentially has the form of a “stroke mimic,” and affected subjects may therefore acutely present at a specialized emergency or stroke department. The Telemedical Project for Integrated Stroke Care (TEMPiS) performed in southeast Bavaria, Germany started as a pilot project in February 2003 and has emerged as the standard of care in rural southeastern Bavaria.11

This multicenter telemedical stroke support network provides diagnoses and/or therapeutic interventions to patients with acute neurological symptoms. Thus, patients with suspicion of acute stroke including “stroke mimics” such as TGA can be diagnosed in specialized stroke departments or in community hospitals with telemedical support provided by experienced neurologists when there is no on-site neurologist. Identifying the clinical factors (including concomitant vascular conditions) associated with TGA is important for patient counseling and for managing vascular risk factors and their relationship with TGA recurrence. However, knowledge on these aspects is currently limited, since the reported rates of TGA recurrence have varied widely from 3% to 26% depending on the follow-up durations in different studies.12 A recent published review described a recurrence rate of 13.5% in TGA patients, and discussed a possible relationship between recurrence and a history of migraine and/or depression.13

The aims of the present cohort study based on the multicenter TEMPiS network data were as follows: 1) to identify concomitant vascular risk factors as well as other preexisting diseases in subjects diagnosed with TGA, 2) to characterize concomitant cerebrovascular disease, including the presence of atherosclerotic large-vessel disease, 3) to characterize TGA episodes using clinical and MRI measures, and 4) to determine the frequencies of recurrences in TGA and identify potential factors associated with TGA recurrence.

This retrospective analysis was applied to patients assessed by stroke experts of the TEMPiS telestroke network and patients who presented at the Department of Neurology and Stroke Center (University of Regensburg, medbo Bezirksklinikum Regensburg, Regensburg, Germany) due to acute neurological symptoms and were eventually diagnosed with TGA. We collected clinical data from the medical records of patients, with the diagnosis of TGA established by the involved neurologist (at the treating center or by telemedical support via TEMPIS) between 2009 and 2018. The study was approved by the ethics committee of the University of Regensburg (IRB No. 17-949-101).

In this study we made the following observations: 1) most TGA episodes occurred during autumn and winter, and most patients reported a potential trigger, 2) approximately 40% of the patients who underwent DWI-MRI showed characteristic DWI-positive lesions in the hippocampi, and arteriosclerotic large-vessel disease was present in more than half of the patients investigated by neurovascular ultrasonography, 3) arterial hypertension was frequently observed in patients with TGA, while most patients did not have a history of cerebro- or cardiovascular disease, 4) approximately half of the patients were hypertensive at admission during the acute TGA, and 5) approximately one-fifth of the patients experienced recurrence of TGA, and at the time point of the index event these patients were older and more frequently had atrial fibrillation, while other concomitant diseases (migraine and hypothyroidism) were more common in subjects without recurrence.

TGA patients are generally aged between 50 and 70 years, but TGA may also occur in both younger and older subjects.12 We did not observe any differences regarding the occurrence of TGA between males and females, which is consistent with the findings of some previous studies, while other studies found a female predominance in TGA occurrence.12, 19 Females seem to experience TGA more often after an emotional trigger, while in males physical stressors are present more often.12, 19

Emotional stress and physical activity are frequently reported to be the most common trigger factors of TGA, which is consistent with our results.3, 4, 20 The underlying reasons remain unclear, but changes in body homeostasis and in emotional states as well as individual factors have been discussed.3, 12

Seasonal variations in TGA cases have been reported previously.21, 22 We similarly found a predominance of TGA in the colder months (autumn/winter). Akkawi et al.22 found an association between low ambient temperature and the occurrence of TGA, which is consistent with our results.

Regarding neuroimaging, MRI was performed in two-thirds of the clinical TGA cases, and approximately 40% of these subjects had an acute DWI lesion in the hippocampus. The reported rate at which DWI hyperintensity is found in TGA patients has varied markedly among studies (12%–86%) depending on the sensitivity and slice thickness of the scans.18 The timing of MRI is also relevant, since the detection rate increases if scans are performed approximately 48 hours after symptom onset.18, 23 In our cohort, MRI scans were mostly performed within 3 days. Sedlaczek et al.23 detected a small, punctate DWI lesion in more than 80% of cases, mostly on the left side (approximately 60% of cases), with the remaining 20% of cases showing this on the right side or bilaterally. They similarly found that lesions were rarely visible during the hyperacute phase, while after 48 hours the detection rate increased.23

Approximately 50% of our subjects showed elevated blood pressure on admission, and approximately 60% had a diagnosis of arterial hypertension at the time of the clinical onset. The pathophysiological cause of this remains unclear. Melo et al.24 found in a case-control study in 1992 that hypertension and migraine were significantly more common in TGA patients than in normal controls. More recently, Rogalewski et al.25 implicated a strong association between acute hypertensive peaks and TGA, especially in patients not adapted to chronic hypertension. A vascular cause of the disease has therefore been hypothesized. It remains unclear if the hypertensive episode triggers TGA itself or if hypertension is a reaction to the situation caused by the amnesia, since TGA patients may be aware of their situation, potentially leading to a restlessness, which could be accompanied by elevated blood pressure. Further studies of pathophysiological relations are required to explain this correlation. Overall, the association of elevated blood pressure with TGA is clinically relevant for patient management and could potentially lead to the first diagnosis of arterial hypertension in affected subjects. However, arterial hypertension and elevated blood pressure were not associated with TGA recurrence.

In this study we further aimed to understand the frequencies of recurrences and identify the potential associated risk factors. In our cohort migraine was not associated with TGA recurrence, which contrasts with the findings of other studies, especially those of Liampas et al.,7 Morris et al.,26 Alessandro et al.27 In our study the presence of preexisting or ongoing migraine was assessed directly in the questionnaires, with answers from the patients being cross-checked with the hospitalization medical records during the TGA attack. This procedure increased the reliability of detecting migraine in the TGA patients. In our cohort the prevalence of migraine in patients without recurrence (14%) was slightly lower than that of 20% found by Morris et al.,26 while we observed no patient with a history of migraine in our recurrence group, which contrasts markedly with the rate of 36.4% found by those authors in patients with recurrence. The reasons for these contradictory findings regarding the association between migraine and TGA, and especially TGA recurrence, remain unclear and so should be investigated in further studies.

In our cohort, hypothyroidism was also more common in the s-TGA group, with no patient in the r-TGA group having hypothyroidism as a concomitant disease. Related studies are scarce, although the meta-analysis of Liampas et al.13 and the study of Alessandro et al.27 found no association between hypothyroidism and TGA either in patients with or without recurrence. Overall, the role of thyroid function in the context of TGA remains to be resolved.

We observed a higher rate of atrial fibrillation in the r-TGA group than in the s-TGA group, which was not found in those two previous studies.13, 27 Alessandro et al.27 found atrial fibrillation in 4% of their total cohort, with all cases in the nonrecurrent-TGA group. Thus, the potential role of atrial fibrillation in the context of TGA recurrence has yet to be clarified. Potential age-related vascular mechanisms related to atrial fibrillation may be shared by TGA, but given the small number of subjects, our findings need to be interpreted with caution. Also, the r-TGA group was older at the time point of the index event.

Another study commonly found atrial fibrillation in TGA patients with a new DWI hyperintensity in MRI, leading to the suggestion of an increased risk of subsequent stroke, while in our cohort there were no differences in neuroimaging results between patients with and without recurrence.28

Oliveira et al.29 recently performed a retrospective investigation of risk factors predicting recurrence in TGA. Those authors observed an overall high rate of TGA recurrence of 27.1%, which was associated with female sex, depression, shorter episode, and a hippocampal hyperintensity on brain MRI. In contrast, arterial hypertension was not related to recurrence, but was overall prevalent in the reported cohort. This is consistent with the findings of our slightly larger study, which also extends the findings of Oliveira et al.29 by identifying other concomitant conditions associated with TGA recurrence.

Our study has some limitations that need to be discussed. The study had a retrospective design and only about 50% of the subjects identified and contacted agreed to participate and returned the questionnaire, which reduced the size of the total cohort. Also, some detailed data were not available for all of the included subjects; however, as stated in the Results section, the frequency of missing data was overall low. Our cohort was smaller than that in the study of Morris et al.26 on factors associated with the risk of recurrent TGA. However, our study included a thoroughly assessed and well-described cohort from a well-established specialized neurovascular network, which extends the previously reported findings especially with respect to vascular comorbidities and risk factors. The follow-up duration differed somewhat in the two groups, being slightly longer in the r-TGA group. This was especially due to those subjects who reported a previous diagnosis of a TGA episode, while the six subjects with r-TGA during follow-up showed a shorter median follow-up duration of 16 months. While this difference in follow-up duration might have influenced the subject categorization, these differences were only moderate. Finally, while not every subject underwent MRI or ultrasonography, more than 70% received both imaging modalities, and so these subgroups can be regarded as representative.

Our study has several strengths. The total cohort was relatively large, comparable to those in previous studies,6, 7, 20 and was well characterized. The diagnosis of TGA was established by experienced neurologists within a large stroke network, and a re-evaluation of individual diagnoses was performed for this study. Sending standardized questionnaires to patients made a median follow-up of 32.5 months possible, allowing a longitudinal course over more than 2 years with analyses of TGA recurrences and potential risk factors related to recurrence.

In conclusion, a recurrence rate of 17% was observed in our cohort, and high rates of arterial hypertension and elevated blood pressure on admission in patients with TGA were observed. The awareness of this is of clinical relevance, suggesting the need for thorough blood pressure monitoring in TGA subjects. Regarding recurrences, migraine and hypothyroidism were more common in subjects with s-TGA, while higher age and atrial fibrillation were more common in patients with r-TGA. The pathophysiological mechanisms underlying our findings remain unclear and so further studies are needed.

The online-only Data Supplement is available with this article at https://doi.org/10.3988/jcn.2022.0368.

SUPPLEMENTARY MATERIAL