Clinical Significance of Isolated Atypical Transient Symptoms in a Cohort With Transient Ischemic Attack
Background and Purpose—Contrary to typical transient symptoms (TS), atypical TS, such as partial sensory deficit, dysarthria, vertigo/unsteadiness, unusual cortical visual deficit, and diplopia, are not usually classified as symptoms of transient ischemic attack when they occur in isolation, and their clinical relevance is frequently denied.
Methods—Consecutive patients with recent TS admitted in our transient ischemic attack clinic (2003–2008) had systematic brain, arterial, and cardiac investigations. We compared the prevalence of recent infarction on brain imaging, major investigational findings (symptomatic intracranial or extracranial atherosclerotic stenosis ≥50%, cervical arterial dissection, and major source of cardiac embolism), and 1-year risk of major vascular events in patients with isolated typical or atypical TS and nonisolated TS, after exclusion of the main differential diagnoses.
Results—Among 1850 patients with possible or definite ischemic diagnoses, 798 (43.1%) had isolated TS: 621 (33.6%) typical and 177 (9.6%) atypical. Acute infarction on brain imaging was similar in patients with isolated atypical and typical TS but less frequent than in patients with nonisolated TS, observed in 10.0%, 11.5%, and 15.3%, respectively (P<0.0001). Major investigational findings were found in 18.1%, 26.4%, and 26.3%, respectively (P=0.06). One-year risk of a major vascular events was not significantly different in the 3 groups.
Conclusions—Transient ischemic attack diagnosis should be considered and investigated in patients with isolated atypical TS.
In the absence of highly sensitive positive diagnostic tools, our ability to clinically ascribe transient symptoms (TS) to an ischemic rather than to a nonischemic cause remains limited, as illustrated by the low interobserver agreement for the clinical diagnosis of transient ischemic attack (TIA), even among stroke specialists.1 This holds particularly true for isolated atypical symptoms as they can be the consequence of many diseases or mechanisms other than cerebral ischemic insult. Historical classification postulated that some atypical symptoms—such as vertigo/unsteadiness, diplopia, dysarthria, partial sensory deficit, and unusual cortical visual symptoms—were not acceptable symptoms of TIA if they occurred in isolation.2,3 Since then, our knowledge has improved with the widespread use of magnetic resonance imaging that is highly sensitive to brain ischemia and the recent publication of 2 large clinical studies.4,5 One study has shown that after exclusion of patients with obvious nonischemic diagnosis by a stroke specialist, a source of embolism at high risk of recurrence was found in ≈20% of patients with isolated atypical visual symptoms (eg, diplopia, lone bilateral blindness, and positive visual symptoms).4 The second study found that among patients with vertebrobasilar strokes, 16% were heralded by TIA,5 92% of which were transient isolated brain stem symptoms that did not fulfill the National Institute of Neurological Disorders and Stroke criteria.2 Altogether, these results suggest that isolated atypical TS should be considered as possible TIA symptoms, reinforcing stroke specialists’ daily experience and clinical practice, and keeping in mind that differential diagnoses are more frequent than in undisputable symptoms of brain ischemia, such as motor deficit or aphasia. If true, this would be an important message because the identification of patients with TIA is crucial, leading to urgent investigation and treatment that yields a 50% to 80% reduction in stroke risk.6–8 Unfortunately, the importance of this clinical problem in routine practice is largely denied. The ABCD2 score (7 point score calculated on the basis of Age, Blood Pressure, Clinical Symptom, Duration of Symptoms, Diabetes) designed to spot high-risk patients is of little help because patients with isolated atypical TS, except for isolated dysarthria, would systematically scored 0 on the item Clinical Symptom with a maximum score of 5 rather than 7.
In a large cohort of patients with suspected TIA, we sought to determine the clinical significance of isolated atypical TS and to what extent they could be related to TIA. To address this question, we compared the prevalence of the underlying source of embolism at high risk of recurrence and the prevalence of recent infarct on brain imaging, as well as ABCD2 score in patients with isolated atypical, isolated typical, and nonisolated TS assuming that if isolated atypical TS could be related to ischemia, the prevalence should be similar in the 3 groups. We further compared the 1-year risk of vascular events in all groups. Finally, we looked at the same variables for each pathogenic subtype of isolated atypical TS.
Research was exempted from by the responsible review committee (Comité de protection des personnes).
SOS-TIA is an outpatient TIA clinic located in a stroke unit.6 Primary care physicians, cardiologists, ophthalmologists, neurologists, and emergency physicians in Paris and its administrative regions can contact the SOS-TIA clinic to refer patients with suspected TIA 24/7 using a toll-free phone number (0800888AIT). Patients are admitted to the clinic if a trained nurse or a senior vascular neurologist confirms the suspicion of TIA after a telephone interview with the healthcare practitioner. Admission criteria at the interview are based on the presence of transient focal cerebral or visual function disturbances with sudden onset and apparent total resolution.
On admission, all patients were questioned, following a standardized questionnaire, about their symptoms during a face-to-face interview with a senior vascular neurologist. The questionnaire covered the following symptoms: motor deficit, sensory deficit, aphasia, dysarthria, vertigo/unsteadiness, transient monocular blindness, cortical visual deficit, and diplopia. Patients with sensory deficit were divided into 2 categories: partial sensory deficit (if abnormal sensation/deficit was located only in 1 limb or only on the face) or complete sensory deficit (if ≥2 limbs or 1 limb and the face were involved). Patients with cortical visual field defect were split into 2 categories: homonymous lateral hemianopia and unusual cortical visual symptoms (including lone bilateral blindness and bilateral positive visual phenomena). Patients were considered to have isolated symptoms if only 1 category of symptom occurred and nonisolated symptoms if ≥2 categories of symptoms occurred. Atypical symptoms of TIA included partial sensory deficit, dysarthria, vertigo/unsteadiness, unusual cortical visual deficit, and diplopia. Motor deficit, complete deficit, transient monocular blindness, aphasia, and homonymous lateral hemianopia were referred to as typical symptoms of TIA.
Investigations and TIA Classification
If a senior vascular neurologist clinically confirmed that a TIA diagnosis was certain or possible, patients entered a standardized evaluation process, including brain magnetic resonance imaging (or a default computed tomographic scan), extracranial and intracranial arterial evaluation (cervical duplex ultrasonography and transcranial Doppler, most often completed by magnetic resonance imaging or computed tomographic angiography), cardiac investigations (electrocardiography and echocardiography), and standard blood chemistry. Patients with visual symptoms underwent an ophthalmologic evaluation. Vascular risk factors were systematically reported. Patients were classified according to 4 final diagnoses: minor stroke (if a recent infarct was present on brain imaging), definite TIA (if an ischemic origin was confirmed but no infarct was present), possible TIA, and nonischemic diagnosis. Definite TIA was considered for patients with symptoms of sudden onset, complete recovery, fitting with an arterial territory, no infarct on brain imaging, and no other obvious diagnosis. Patients with transient focal neurological symptoms in whom clinical and radiological features were not sufficiently clear to affirm or rule out the diagnosis of definite TIA and with no other evident cause were classified as possible TIA. Nonischemic diagnoses were considered if there was an obvious disease explaining the symptoms or if symptom characteristics (onset and type) were not compatible with an arterial occlusion.
Definition of Major Investigational Findings
Major investigational findings were defined as underlying causes of TIA that carry a short-term high risk of recurrent stroke: symptomatic intracranial or extracranial atherosclerotic stenosis ≥50%, cervical arterial dissection, and a major source of cardiac embolism (according to grade C1 in the ASCOD [A, atherosclerosis; S, small vessel disease; C, cardiac pathology; O, other causes; and D, dissection system]).9
Neurologists or research nurses obtained follow-up information on the occurrence of any vascular events or death during face-to-face interviews or via telephone calls at 90 days and 1 year using a standardized questionnaire. If the patient could not be contacted, a close relative or their family practitioner was interviewed.
Prevalence of isolated TS was described in patients with a suspected TIA after exclusion of patients with nonischemic events. Patients with a minor stroke and those with a definite or possible TIA were divided into 3 groups according to the presence of TS (isolated atypical, isolated typical, and nonisolated). Continuous values were expressed as means±SDs and nominal variables as counts and percentages. Baseline characteristics were compared between the 3 groups by an ANOVA or Wilcoxon test for values that were not normally distributed. Categorical variables were compared using the χ2 test or Fisher exact test when the expected cell frequency was <5. The presence of major investigational findings was compared across the 3 groups using the χ2 or Fisher exact test. The 1-year combined outcome (stroke, myocardial infarction, or vascular death) was compared according to the 3 groups using the log-rank test. Patients who died from causes other than stroke or vascular disease were censored at the time of their death. All pairwise comparisons were performed with a Bonferroni correction. P<0.05 was considered to indicate a statistically significant difference. Data were analyzed using the SAS software package, release 9.3 (SAS Institute, Cary, NC).
Of 2398 consecutive patients with a suspected TIA admitted to our SOS-TIA clinic between January 2003 and December 2008, 1850 (77.1%) had a final possible or definite ischemic diagnosis. Of these, 798 (43.1%) had isolated symptoms, which were atypical in 177 (9.6%) and typical in 621 (33.6%). The Figure shows the distribution of patients with isolated and nonisolated TS according to symptom subtypes in the whole cohort (n=2398).
Final Diagnosis According to Symptom Categories
Of all 2398 patients with suspected TIA, the distribution of the final diagnosis was significantly different between the 3 groups (P<0.0001) and is reported in Table 1. There was wide heterogeneity between isolated atypical TS subgroups about the likelihood of cerebral ischemia (Table I in the online-only Data Supplement). Acute brain infarction was found in 22.5% of patients with dysarthria and in 12.5% of patients with partial sensory deficit, but in only 8.7% of those with diplopia and 1.4% of those with ataxia/vertigo. No patients with unusual cortical visual symptoms had an acute infarct on brain imaging. In comparison, acute infarction was found in 19.9% of patients with motor deficit, 13.2% of patients with aphasia mutism, 11.3% of patients with complete sensory deficit, and 3.4% of patients with transient monocular blindness. Table II in the online-only Data Supplement reports the final diagnosis of patients with TIA mimics according to clinical presentation (isolated typical TS, isolated atypical TS, and nonisolated TS). Significant differences were found for migraine, epilepsia, and peripheral neuropathy.
Characteristics of Patients With an Ischemic Diagnosis (Definite or Possible)
Patients with isolated atypical TS were older and had the lowest prevalence of previous recent TIA and ABCD2 score (Table 2). Patients with nonisolated TS were younger, had a lower prevalence of hypertension, a longer duration of symptoms, and a higher ABCD2 score. There were no significant differences in the main characteristics between patients with isolated atypical and typical TS except for ABCD2 score (χ2 test with Bonferroni correction). In the atypical TS group, an ABCD2 score ≥4 was less than half as frequent than in patients with typical TS.
Isolated Symptoms and Major Investigation Finding
A major source of embolism was found in 18.1% in patients with isolated atypical TS, which was less frequent than in patients with isolated typical TS (26.4%) or nonisolated TS (26.3%; P=0.06; Table 3). High-grade extracranial stenosis was significantly less frequent in patients with isolated atypical TS (5.6% versus 12.9% and 10.4% in patients with isolated typical and nonisolated TS, respectively). It remained significant when compared atypical with typical TS (χ2 test with Bonferroni correction) and after adjusting for age and sex (odds ratio, 0.38; 95% confidence interval, 0.19–0.76). This difference was driven mainly by patients with isolated diplopia (Table 4). Major findings of patients with isolated atypical symptom subtypes are reported in Table 4. As a comparison, major investigational finding(s) were found in 32.2% of patients with motor deficit, 24.3% of patients with aphasia, 21.6% of patients with transient monocular blindness, and 17.6% of patients with complete sensory deficit (data not shown).
Symptom Subtypes and Risk of Recurrent Vascular Events
A total of 1810 patients (97.8%) with a diagnosis of TIA (possible or definite) or minor stroke were followed up for a median of 13 months (interquartile range, 12–16 months). By 1 year, 52 strokes (37 within 90 days, 6 of which were fatal), 6 myocardial infarctions (2 fatal), and 6 other vascular deaths had occurred (Table 5). A further 14 nonvascular deaths and 5 deaths of unknown cause occurred. The 1-year risk of the combined outcome (stroke, myocardial infarction, or vascular death), as well as the risk of each individual component of our composite outcome, did not differ significantly between the 3 groups (P=0.07).
In this large cohort of consecutive patients with transient clinical symptoms admitted in a TIA clinic, those with isolated TS accounted for ≈40% of our population, but the frequency of isolated atypical TS was low (9% of the whole cohort). However, because of the design of our study, it is not possible to conclude whether these symptoms are actually infrequent or whether they are under-represented because many of the physicians did not refer those patients because a TIA diagnosis was deemed to be unlikely. Both could be true as it is relatively uncommon to see patients presenting with isolated atypical symptoms as the only sign of brain infarction probably because the chance to have an ischemic lesion limited to a small area is low. Not surprisingly, and despite a telephone interview conducted to limit admission of patients with a diagnosis other than TIA, twice as many patients with isolated atypical versus isolated typical TS were considered to have a nonischemic event after specialized neurological evaluation (Table 1). However, the proportion of patients considered to have a definite or possible ischemic event remained high in the atypical TS group (≈80%) and, most importantly, the proportion of patients with an acute ischemic lesion on brain imaging was similar in the 3 groups (10%–15%), mainly driven by patients with dysarthria and partial sensory symptom, confirming that a TIA diagnosis should be considered in patients with isolated atypical TS after the main differential diagnoses have been ruled out. These patients also deserve emergency evaluation and treatment because positive diffusion-weighted imaging (DWI) is a risk factor for imminent stroke occurrence.10 This was emphasized by the presence of major investigational findings at high risk of recurrence in ≈1 in 5 patients presenting with isolated atypical TS. Of note, the prevalence of a major investigational finding was higher in patients with isolated typical TS and in those with nonisolated TS than in patients with atypical TS, suggesting that despite careful investigation, some patients with isolated atypical TS could have been misclassified as TIA.
We did not find any significant difference in the 1-year risk of stroke, myocardial infarction, or vascular death between the typical and atypical TS groups despite the finding that baseline ABCD2 score differed significantly between the groups, with an ABCD2 score ≥4 being more less than half as frequent in patients with isolated atypical versus isolated typical TS. This is explained, at least in part, by the fact that C in the ABCD2 system denotes unilateral weakness and speech disturbance, which are except for dysarthria, typical symptoms.
Importantly, we found heterogeneity between the predefined isolated atypical TS subtypes. Nonischemic diagnosis was more frequent in patients with ataxia/vertigo (60.6%), unusual cortical visual symptoms (41.2%), and partial sensory deficit (29.2%) than in patients with diplopia (17.4%) or dysarthria (5.0%). To further determine the relationship between isolated atypical TS and TIA, we determined the frequency of positive DWI for each isolated symptom subtype and found a high rate of positive DWI lesions in patients with transient isolated dysarthria (22.5%) and partial sensory symptom (12.5%), which was similar to, or higher than, the rate observed in patients with isolated typical symptom subtypes (Table I in the online-only Data Supplement). Conversely, DWI was negative in a large majority of patients with isolated diplopia (91.3%), ataxia/vertigo (98.6%), or unusual cortical visual symptoms (100%). Lower DWI sensitivity to small lesions or to lesions located in the posterior fossa could partially explain this discrepancy. No perfusion-weighted imaging was systematically performed in these cases, which is a limitation of this study. However, incorrect attribution of symptoms to transient arterial obstruction is also possible, as illustrated by the higher proportion of patients classified as having possible TIA in the event of isolated diplopia (29.0%) or ataxia/vertigo (18.3%) in comparison to patients with other isolated or nonisolated TS. As previously mentioned, because DWI is frequently negative in TS deemed to be of ischemic origin, we also looked at the prevalence of major investigational findings in isolated atypical TS subtypes. The presence of a major investigational finding in patients with acute transient neurological deficit compatible with a vascular territory (in which differential diagnoses have been reasonably ruled out) is a strong argument to ascribe the event to an ischemic cause. We found at least 1 major investigational finding in ≈1 in 5 patients among the different subtypes of isolated atypical TS except in patients with diplopia, a result not so different to that found in patients with nonisolated TS and in those with isolated typical TS. Patients with isolated diplopia had a notably lower rate of major investigational finding, suggesting, once again, that some of these patients could have been misclassified despite careful neurological and ophthalmologic examination.
One strength of our study was the careful and systematic collection of data on symptoms. The prospective evaluation of a large number of consecutive patients with TS who underwent systematic brain, arterial, and cardiac evaluation is another important strength of our study. TIA event classification, performed by a senior vascular neurologist, was based on predefined strict criteria and not simply on stroke neurologist impression.7 Importantly, patient admission in our TIA clinic was not based on symptom subtypes (we did not exclude a priori patients with atypical TS) or demographic characteristics. The main criteria for TIA clinical admission were acute onset and full recovery of neurological symptoms deemed to be a TIA and exclusion of patients with an obvious non-TIA diagnosis that was easily made by telephone interview with the referring primary care physician (eg, typical recurrent positive visual symptoms associated with headache in young adults). As a consequence, we had a large panel of different symptoms and a low rate of nonischemic events observed in our cohort in comparison with the rates found in other studies.11,12 However, our study has other limitations. Despite the large number of patients included, the sample size of each isolated atypical TS subgroup is small, limiting the reliability of the results for some symptom subtypes. Moreover, as in every TIA study, because TIA is a clinical diagnosis, and despite classification following predefined criteria, misclassification of patients (TIA versus mimics) is difficult to avoid. Tools are required to improve TIA diagnosis in these patients, especially for those with diplopia, ataxia/vertigo, and unusual cortical visual symptoms. Our study was not population based, although the catchment area was the entire region of Grand Paris, and we know from diagnosis related-group data that we see ≈1 in 5 patients with TIA from this region.
Isolated atypical TS were infrequent in patients admitted in our TIA clinic. However, these symptoms were frequently associated with an underlying cause of stroke at high risk of recurrence, and 1-year brain infarction was as frequent in these patients as in patients with isolated typical TS. This holds particularly true for isolated dysarthria and partial sensory symptoms. In light of a recent study showing the prevalence of atypical TS before vertebrobasilar ischemic stroke,5 our results, if confirmed, call for a revision of the list of symptoms consistent with TIA to better educate patients and nonstroke specialists.
Editorial support (limited to editing for style, referencing, and figure editing) was provided by Sophie Rushton-Smith, PhD (MedLink Healthcare Communications). Drs Lavallée and Amarenco participated in the conception and design of the study. Dr Lavallée, L. Sissani, and Dr Amarenco analyzed and interpreted the data. Drs Lavallée, Cabrejo, Meseguer, Guidoux, Klein, Touboul, and Amarenco provided study material or recommended patients. L. Sissani and J. Labreuche performed the statistical analysis. Drs Lavallée and Amarenco wrote the article. All authors reviewed and approved the article.
Sources of Funding
Funding for this study was provided in part by SOS-ATTAQUE CEREBRALE and supported the Département Hospitalo-Universitaire Fibrosis Inflammation Remodeling of Université Paris-Diderot, France.
Dr Amarenco has received research grant for related research with Pfizer, AstraZeneca, Sanofi, and BMS and the French government, honoraria from Pfizer, AstraZeneca, Bayer, Kowa, Fibrogen, and GSK and has speaking engagement with Pfizer and Bayer to disclose. The other authors report no conflicts.
Guest Editor for this article was Gregory W. Albers, MD.
The online-only Data Supplement is available with this article at http://stroke.ahajournals.org/lookup/suppl/doi:10.1161/STROKEAHA.117.016743/-/DC1.
- Received January 18, 2017.
- Revision received March 6, 2017.
- Accepted March 8, 2017.
- © 2017 American Heart Association, Inc.
- Kraaijeveld CL,
- van Gijn J,
- Schouten HJ,
- Staal A
- 2.↵Advisory Council for the National Institute of Neurological and Communicative Disorders and Stroke, National Institutes of Health. A classification and outline of cerebrovascular diseases II. Stroke. 1975;6:564–616.
- Lavallée PC,
- Cabrejo L,
- Labreuche J,
- Mazighi M,
- Meseguer E,
- Guidoux C,
- et al
- Rothwell PM,
- Giles MF,
- Chandratheva A,
- Marquardt L,
- Geraghty O,
- Redgrave JN,
- et al
- Amarenco P,
- Bogousslavsky J,
- Caplan LR,
- Donnan GA,
- Wolf ME,
- Hennerici MG
- Merwick A,
- Albers GW,
- Amarenco P,
- Arsava EM,
- Ay H,
- Calvet D,
- et al
- Dennis MS,
- Bamford JM,
- Sandercock PA,
- Warlow CP