Diagnosis of Transient Ischemic Attack by the Nonneurologist
A Validation Study
Background and Purpose Interobserver reliability of the diagnosis of transient ischemic attack (TIA) is low, and diagnosis of TIA made by nonneurologists is often erroneous. We sought to validate the diagnosis of TIA made by general practitioners (GPs) and by hospital emergency service physicians (emergency MDs).
Methods A list of 20 neurological symptoms was distributed to 20 GPs and 22 neurologists who graded the compatibility of each symptom with the TIA diagnosis. At least two neurologists validated TIA diagnoses made by GPs for patients under their care or by emergency MDs.
Results Compared with neurologists, GPs considered “confusion” and “unexplained fall” more often compatible with TIA and “lower facial palsy” and “monocular blindness” less often compatible with TIA. Validation of diagnosis by GP was confirmed in 10 patients (19%); 26 patients had strokes, and 16 (31%) had a noncerebrovascular disorder. Validation of diagnosis by emergency MD was confirmed in 4 patients (13%); 10 patients had strokes, and 17 (55%) had noncerebrovascular disorders. The most frequent conditions misdiagnosed as TIAs were transient disturbances of consciousness, mental status, and balance.
Conclusions The TIA concept is understood differently by neurologists and nonneurologists. GPs and emergency MDs often label minor strokes and several nonvascular transient neurological disturbances as TIAs. Until this misconception of TIA is changed, the term TIA should probably be avoided in the communication between referring physicians and neurologists. If not referred to a neurologist, one third to one half of patients labeled with a diagnosis of TIA will be inappropriately managed.
Depending as it does on patient interview, diagnosis of TIA relies on the patient's ability to verbalize his/her symptoms, on the receiver's ability to analyze its meaning, and on the examining physician's semantic and operational concept of TIA. Failure to recognize TIA or labeling other transient neurological disturbances as TIA has obvious, important consequences, with patients being denied appropriate therapy or given inappropriate investigation and treatment. Diagnosis of TIA is difficult and has limited interobserver reliability, even between neurologists.1 2 3 4 5 TIA diagnosis made by nonneurologists is often erroneous.5 6 7 8 Previous studies indicate that young patients and those with hemispheric TIAs are more likely to be referred to a hospital7 9 and that the likelihood of a correct TIA diagnosis made by a GP is higher in young patients, in those with symptoms of hemispheric ischemia, and in those with long or recurrent attacks.7
The objective of this investigation was to validate the diagnosis of TIAs made by nonneurologists and to identify factors that increase or decrease the likelihood of diagnostic errors.
Subjects and Methods
We defined TIA as a focal neurological deficit of sudden onset occurring in a cerebrovascular territory distribution and recovering without sequelae in less than 24 hours, with nonvascular causes excluded. Symptoms with gradual onset or “march,” very brief symptoms (<15 minutes), positive visual or motor phenomena, “shaking” spells, isolated diplopia or vertigo, unexplained falls, or paresthesias in the setting of hyperventilation were not considered TIAs.10 The combination of a focal and a nonfocal symptom (eg, hemiparesis and confusion) was considered compatible with the diagnosis of TIA.
Neurologists' and GPs' Concepts of TIA
To compare the semantic meaning of TIA for neurologists and GPs, a list of 20 transient neurological symptoms was distributed to 22 neurologists (including residents) attending the weekly departmental round at our department and to 20 GPs practicing in the Lisbon area who attended a session on cerebrovascular disorders. They were asked to grade (no; unlikely; uncertain; likely; yes) the compatibility of each symptom with the diagnosis of TIA. Positive (“likely” and “yes” responses) and negative (“no” and “unlikely” responses) options were compared between the two groups of physicians by Fisher's exact test.
Validation of GPs' Diagnoses of TIA
Data from an ongoing epidemiological study of TIA and stroke performed by 42 GPs in central and south Portugal11 were used, covering the period from April 1, 1993, to March 31, 1995. If a presumed TIA was observed or diagnosed by one of these GPs in a patient under his or her care, he notified the coordinating center, indicating the patient's age and sex, diagnosis (TIA or stroke), grade of diagnostic confidence (certain, probable), vascular risk factors, neurological symptoms, ancillary procedures performed, and treatment. GPs were encouraged to refer these TIA patients to the stroke outpatient clinic at our hospital, where a final diagnosis (TIA, stroke, noncerebrovascular disorder) was established by two neurologists. After reviewing the patient's history, physical and neurological examinations, and all ancillary procedures, the first neurologist wrote a case history and a diagnosis. These notes were reviewed by the second neurologist. If the information provided by the first neurologist seemed unclear or incomplete or there was a disagreement about the diagnosis, discussion followed until a consensus diagnosis was established.
Validation of Emergency MDs' Diagnoses of TIA
At our hospital, the emergency department serves a population of approximately 800 000 persons; approximately 600 patients are seen daily. Patients are first examined by a rotating emergency MD (either an internal medicine resident, a GP, or a nonstaff physician engaged to work only at the emergency department). A staff neurologist and a neurology resident are also on duty for 24-hour shifts. Suspected strokes or TIAs are always referred for neurological evaluation with an accompanying brief referral note, before admission or discharge is decided. For the period from September 1, 1994, to November 1, 1995, the seven participating neurologists registered on a special form all patients referred with diagnosis of TIA or possible TIA; their age, sex, and risk factors; discrepancy between the history elicited by the emergency MD and the neurologist; and the neurologist's diagnosis. Only patients for whom the emergency MD wrote a diagnosis on the referral note (either TIA or possible TIA) were included. These were presented by the neurologist who had examined the patient at the weekly stroke rounds, where a consensus diagnosis (TIA, stroke, or noncerebrovascular disorder) was established.
For studies of validation of diagnosis made by GP or validation of diagnosis made by emergency MD, sex, age (≤70 or >70 years), presence of vascular risk factors, grade of diagnostic confidence, and type of symptoms (motor and/or sensory versus other) were compared (χ2 and Fisher's tests) between the three groups of final diagnoses (TIA, stroke, and noncerebrovascular disorders).
Neurologists' and GPs' Concepts of TIA
Table 1⇓ summarizes neurologists' and GPs' responses to the questionnaire regarding symptoms that could be compatible with the diagnosis of TIA. Compared with neurologists, GPs considered “confusion” and “unexplained fall” as more frequently compatible and “lower facial palsy” and “monocular blindness” as less frequently compatible with the diagnosis of TIA.
GPs were particularly doubtful about some symptoms, such as “monocular visual loss,” “confusion,” and “unexplained fall,” whereas neurologists were rather uncertain about “unsteadiness” and “diplopia.” On the whole, GPs gave 14% more uncertain answers than neurologists (126 of 400 versus 79 of 440 possible answers; 95% CI, 8% to 19%).
Validation of GPs' Diagnoses of TIA
During the study period, the participating GPs diagnosed 103 TIAs, and 52 patients (50%) were referred for neurological evaluation. Referred TIA patients were predominantly males (referred, 31 males versus 21 females; nonreferred, 19 males versus 32 females; difference, 22%; 95% CI, 4% to 41%) and included more cases of probable TIA (19 versus 8; difference between proportions of certain and probable TIAs, 22%; 95% CI, 5% to 41%). There were no differences in age, risk factors, or type of symptoms between referred and nonreferred patients with TIA.
Diagnosis of TIA was confirmed in 10 patients (19%). TIA diagnosis was incorrect in 42 patients. In 26 patients (50%), symptoms/signs lasted longer than 24 hours, whereas 16 patients (31%) showed a noncerebrovascular disorder (Table 2⇓). Before discussion, there were only two disagreements between the two neurologists who formulated the final diagnosis.
No significant differences between the characteristics of patients with correct and incorrect diagnoses were identified. However, a comparison between patients with a cerebrovascular diagnosis (TIA or stroke) and those with a diagnosis of noncerebrovascular disorder showed that male sex (25 of 36 versus 6 of 16, P=.04) and motor/sensory symptoms (31 of 36 versus 6 of 16, P=.002) (Table 3⇓) were significantly more frequent among patients with cerebrovascular events.
Validation of Emergency MDs' Diagnoses of TIA
During the study period, 31 patients with diagnosis of TIA (4 patients) or possible TIA (27 patients) were referred for neurological evaluation by the emergency MDs. Consensus diagnosis after discussion between the participating neurologists confirmed the TIA diagnosis in only 4 patients (13%). Ten cases (32%) were in fact strokes, and 17 patients (55%) had noncerebrovascular diseases (Table 2⇑). In 13 patients (42%), symptoms reported in the referral note were either incorrect or incomplete, but adequacy of the elicited history was not determinant for TIA diagnostic accuracy. The final diagnosis of patients with correct history amounted to 2 TIAs, 2 strokes, and 9 others, whereas the final diagnosis of those with incorrect history was 2 TIAs, 4 strokes, and 7 others. No significant differences between the characteristics of patients with a correct or wrong TIA diagnosis were identified. Motor/sensory complaints were significantly more frequent among patients with a final diagnosis of TIA or stroke than in those with a diagnosis of noncerebrovascular disorders (9 of 14 versus 4 of 17, P=.03).
Our results confirm previous reports demonstrating a low interobserver agreement for TIA diagnosis.1 2 3 4 5 6 7 8 The majority of previous studies on the interobserver reliability of TIA diagnosis compared the diagnoses of two or more neurologists.1 2 3 4 5 The present study adds to previous investigations information on the diagnostic accuracy of those physicians who will most likely be the first medical contact of TIA patients, ie, GPs and emergency MDs. We found that 81% and 87%, respectively, of the TIA diagnoses made by GPs and emergency MDs could not be confirmed by neurologists with an expertise in stroke; 31% and 55%, respectively, of these “false TIAs” corresponded to a noncerebrovascular disorder. Both these percentages of errors and types of misdiagnosis show some similarities with those reported in the few studies that investigated these issues.
In the Cooperative Study of hospital frequency and character of TIAs,5 the charts of all patients with suspected TIA were reviewed by institutional investigators; a final diagnosis of TIA was ruled out in 30%. Prevailing conditions misdiagnosed as TIA were postural hypotension, syncope, vertigo, seizures, anxiety, confusion, and complete stroke. Stereotyped and carotid TIAs, presence of carotid bruits, examination during an attack, and examiner's level of neurological training (by decreasing order of diagnostic reliability: neurologist, nonneurology staff, neurology resident, house officer) increased diagnostic accuracy.
In the Oxfordshire Community Stroke Project,6 317 (62%) of 512 patients with supposed TIA referred by GPs had noncerebrovascular disorders such as migraine, syncope, vertigo, epilepsy, and transient global amnesia. The number of supposed TIAs lasting more than 24 hours or persisting was not indicated. In the Oxfordshire Community Stroke Project, amaurosis fugax was predominantly common among hospital-referred TIAs.9 In our study, GPs were more reluctant than neurologists to relate monocular visual loss to transient ischemia.
Quik-van Milligen et al7 mailed a questionnaire that included 10 TIA cases to a random sample of 10% of all GPs in the Netherlands. Eighty percent were accurately diagnosed. A correct diagnosis was predominantly made in patients with transient hemispheric ischemia, longer or recurrent attacks, in patients younger than 65 years, and in those without nonspecific symptoms. Because the response rate did not exceed 59%, it is reasonable to infer that had all GPs answered, the percentage of accurate diagnoses would be lower.
In an epidemiological study of TIA and minor strokes performed in central Spain,8 where a methodological design similar to our study of TIA diagnoses made by GPs was used, GPs referred 193 patients with supposed TIA for neurological evaluation. Sixty-four had a final diagnosis of noncerebrovascular disorders, including vertigo, syncope, epilepsy, and transient global amnesia. It is not stated how many minor strokes were labeled TIAs by the GPs.
Neurologists often disagree on TIA diagnosis; however, upon discussion a high degree of agreement can be reached.4 For both settings (GP and emergency MD), final diagnosis was always decided by the consensus of at least two neurologists. Hankey et al9 showed that hospital-referred and community TIA patients differ in items of age, sex, persisting complaints, prevalence of cigarette smokers and atrial fibrillation, level of hematocrit, and time for assessment. Similarly, in our study referred TIA patients were more often males. Except for the degree of diagnostic confidence, no other differences were noted between referred and nonreferred TIA patients, probably because referral was encouraged and neurological consultation was readily available.
The ready availability of neurological consultation favored referral of patients with an uncertain diagnosis, thus increasing the likelihood of false TIA diagnosis, to a greater extent than if the GPs had less access to specialist consultation, as occurs in actual clinical practice. However, such bias was absent from the study validating emergency MD diagnosis, since a neurologist was routinely on duty for 24 hours, and all patients with suspected TIA had to be examined by a neurologist. Whereas GPs were requested to produce a stroke or TIA diagnosis, such a condition was obviously not present for the emergency MD, whose study was performed under “natural” conditions.
It is known that TIA diagnosis relies on an accurate clinical history of the attack and the skill with which the history is elicited and interpreted, rather than the physical examination.12 In our study, the clinical history of the attack was often incorrect or incomplete when collected by the emergency MD. However, that was not a significant factor likely to affect the accuracy of the diagnosis, and neither were age, sex, vascular risk factors, or history of previous stroke or TIA.
For both physician groups, the likelihood of a noncerebrovascular transient disorder being misdiagnosed as a TIA was higher if the patient complained of symptoms other than motor or sensory defects. Transient disturbances of consciousness, higher nervous function, or balance presented special difficulties, a fact that confirms the results of the questionnaire showing that the semantic meaning of TIA was different for neurologists than it was for GPs. GPs were more uncertain about their options and considered some “nonfocal” symptoms compatible with TIA diagnosis. The results of the GPs' and emergency MDs' diagnostic validation studies showed that their concept of TIA also included “small strokes,” both those lasting more than 24 hours and those that exhibited persisting symptoms or signs, although they caused no major disability. Because secondary prevention management is similar for TIA and nondisabling stroke,13 14 15 this distinction may be regarded as of no pragmatic consequence, and it is questionable whether the term TIA is still of use, or whether TIA and minor, nondisabling stroke are in fact a continuum.16 17 18 If this is probably true for secondary prevention management, in acute stroke treatment19 a clear distinction should be made between TIAs, whose defect clears in a few hours,20 21 and established stroke. Because of this confusion between TIA and minor stroke, for the purpose of epidemiological and secondary prevention studies, patients with TIAs diagnosed by a nonneurologist should be screened as possible stroke cases.
Disagreement and incorrect diagnosis of TIA were mainly due to a different concept of TIA between neurologists and nonneurologists. This misconception is likely to be changed over time by adequate medical information and education, stressing that TIA is defined as transient focal symptoms or signs of ischemic origin. The arbitrary criteria of 24 hours' maximal duration should probably be shortened to 1 or 4 hours, because (1) 50% and 90% of TIAs, respectively, remit within these time intervals21 ; (2) the distinction between ischemic events with transient events (TIA) and persisting events (stroke) would became more clear; and (3) these time limits will be of operational value for acute stroke interventions. Meanwhile, the term TIA should probably be avoided in the communication between neurologists and nonneurologists.
Our study also shows that a readily available neurological consultation can considerably increase the quality of care received by patients with transient neurological symptoms. If not referred to a neurologist, one third to one half of the patients diagnosed with TIA will be the object of inappropriate investigation and treatment and denied adequate therapy. Current recommendations for TIA management include hospital admission and a costly workup,22 23 while many of the nonvascular transient neurological disorders are benign and require no hospital admission or expensive ancillary examinations. Referral to a neurologist,24 particularly of those patients without motor or sensory symptoms, can be cost-effective by decreasing the number of inappropriate investigations and admissions.
Selected Abbreviations and Acronyms
|emergency MD||=||hospital emergency service physician|
|TIA||=||transient ischemic attack|
This study was supported in part by Junta Nacional de Investigação Cientifica e Tecnológica (grant STRADA/C/SAU/353/92). The following GPs (Centro de Saúde [Health Care Center] shown in parentheses) referred patients for the study: Carlos Reis (Sines); Ana Pereira (Panoias, Ourique); Deolinda Diniz (Almada); João Pereira (Sacavém); Mário Silva (Carregueira); Mário Ferreira (Lapa, Lisbon); Carlos Roque (Entroncamento); Sergio Serra (Cartaxo); Maria José Rosa (Barreiro); Anabela Lima (São Nicolau, Lisbon); José Ricardo (Arronches); Maria Fernanda Gouvea (Ponte de Sor); Maria Gracinda Rodrigues (Ponte de Sor); Vitor Palmeira (Tavira); António Rosa (Baixa da Banheira); Maria José Carmo (Castro Marim); and Maria Helena Ferro (Estoril, Cascais).
- Received July 13, 1996.
- Revision received September 17, 1996.
- Accepted September 19, 1996.
- Copyright © 1996 by American Heart Association
Tomasello F, Mariani F, Fieschi C, Argentino C, Bono G, De Zanche, Inzitari D, Martini A, Perrone P, Sangiovanni G. Assessment of inter-observer differences in the Italian Multicenter Study on reversible cerebral ischemia. Stroke. 1982;13:32-35.
Kraaijeveld CL, van Gijn J, Schouten HJA, Staal A. Interobserver agreement for the diagnosis of transient ischemic attacks. Stroke. 1984;15:723-725.
Koudstaal PJ, van Gijn J, Staal A, Duivenvoorden HJ, Gerritsma JGM, Kraaijeveld CL. Diagnosis of transient ischemic attacks: improvement of interobserver agreement by a check-list in ordinary language. Stroke. 1986;17:723-728.
Dennis MS, Bamford JM, Sandercock PAG, Warlow CP. Incidence of transient ischemic attacks in Oxfordshire, England. Stroke. 1989;20:333-339.
Quik-van Milligen MLT, Kuyvenhoven MM, de Melker RA, Touw-Otten FWMM, Koudstaal PJ, van Gijn J. Transient ischemic attacks and the general practitioner: diagnosis and management. Cerebrovasc Dis. 1992;2:102-106.
Jempere AP, Duarte J, Cabazas C, Claveria LE. Incidence of transient ischemic attacks and minor ischemic strokes in Segovia, Spain. Stroke. 1996;27:667-671.
Hankey GJ, Slattery JM, Warlow CP. The prognosis of hospital-referred transient ischaemic attacks. J Neurol Neurosurg Psychiatry. 1991;54:793-802.
Millikan CH, Ad Hoc Committee on Cerebrovascular Disease: a classification and outline of cerebrovascular disease, II. Stroke. 1975;6:565-616.
Ferro JM, Falcão I, Rodrigues AG. Community-hospital stroke registry in central-south Portugal: methodology and first-year results. In: Program and abstracts of Epidemiology and Prevention of Stroke; May 29-31, 1994; Umea, Sweden: 22. Abstract.
Hankey GJ, Warlow CP. Clinical features and differential diagnosis. In: Warlow CP, van Gijn J, eds. Transient Ischaemic Attacks of the Brain and Eye. London, England; WB Saunders Co Ltd; 1994:76-127.
Antiplatelet Trialists' Collaboration. Collaborative overview of randomised trials of antiplatelet therapy, I: prevention of death, myocardial infarction, and stroke by prolonged antiplatelet therapy in various categories of patients. BMJ. 1994;308:81-106.
Caplan LR. TIAs: we need to return to the question, ‘What is wrong with Mr. Jones?’ Neurology. 1988;38:791-793.
Koudstaal PJ, van Gijn J, Frenken CWGM, Hijdra A, Lodder J, Vermeulen M, Bulens C, Franke CL. TIA, RIND, minor stroke: a continuum, or different subgroups? J Neurol Neurosurg Psychiatry. 1992;55:95-97.
Loeb C, Gaadolfo C. Transient ischemic neurologic deficits. In: Vinken PJ, Bruyn GW, Klawans HL, Toole JF, eds. Handbook of Clinical Neurology, Vol 53. Amsterdam, Netherlands: Elsevier Science Publishers; 1988:257-290. Revised series 9.
Levy DE. How transient are transient ischemic attacks? Neurology. 1988;38:674-677.
Werdelin L, Juhler M. The course of transient ischemic attacks. Neurology. 1988;38:677-680.
Hankey GJ, Warlow CP. Cost-effective investigation of patients with suspected transient ischaemic attacks. J Neurol Neurosurg Psychiatry. 1992;55:171-176.
Feinberg WM, Albers GW, Barnett HJM, Biller J, Caplan LR, Carter LP, Hart RG, Hobson RW, Kronmal RA, Moore WS, Robertson JT, Adams HP, Mayberg M. Guidelines for the management of transient ischemic attacks: from the Ad Hoc Committee on Guidelines for the Management of Transient Ischemic Attacks of Stroke Council of the American Heart Association. Stroke. 1994;25:1320-1335.
Ringel SP. Future neurology workforce: the right kind and number of neurologists. Neurology. 1996;46:897-900.