Clinical and Prognostic Correlates of Stroke Subtype Misdiagnosis Within 12 Hours From Onset
Background and Purpose Pure motor hemiparesis and sensorimotor stroke syndromes are not accurate predictors of lacunar infarct when described in the first 12 hours of stroke onset. We evaluate here whether this inaccuracy of clinical diagnosis might have influenced the planning of patient management either in routine practice or in therapeutic trials.
Methods A consecutive hospital series of 517 first-ever ischemic hemispheric stroke patients presented lacunar or nonlacunar syndromes at the first examination within 12 hours of the event. A distinction was subsequently made, by means of a CT scan or autopsy performed within 15±2 days of stroke, between patients affected by lacunar or nonlacunar infarcts. We compared stroke risk factors, considered to be indicative of potential pathogenetic mechanisms, and the clinical outcome of lacunar infarct versus nonlacunar infarct patients and those of lacunar syndrome versus nonlacunar syndrome patients.
Results Two hundred nineteen patients (42%) presented a lacunar syndrome and 298 (58%) a nonlacunar syndrome, while 170 (33%) had lacunar infarcts and 347 (67%) nonlacunar infarcts. Lacunar infarct patients were more frequently associated with hypertension and a previous transient ischemic attack and less frequently with atrial fibrillation when compared with their nonlacunar infarct counterparts, whereas no differences were apparent between lacunar syndrome and nonlacunar syndrome patients. Logistic regression analysis showed that hypertension and a previous transient ischemic attack on the one hand and atrial fibrillation on the other were strongly correlated with the diagnosis of lacunar infarct and nonlacunar infarct, respectively, while no risk factor was correlated with the diagnosis of lacunar syndrome. Twenty-two percent of lacunar infarct patients and 68% of nonlacunar infarct subjects had a poor outcome (death plus disability of survivors) as opposed to 40% of lacunar syndrome and 63% of nonlacunar syndrome patients. Logistic regression selected age, severity of neurological deficit at entry, cardiopathies, diabetes, and lacunar infarct, but not lacunar syndrome, as predictors of outcome.
Conclusions The inaccurate clinical diagnosis of lacunar infarct made in the first 12 hours of stroke might lead to no distinction being made between stroke subgroups with potentially different pathogenetic mechanisms and prognostic estimates, thus negatively influencing the planning of patient management.
According to some authors, lacunar syndromes (ie, pure motor hemiparesis, pure sensory stroke, sensorimotor stroke, ataxic hemiparesis, and dysarthria–clumsy hand) may predict as many as 90% of lacunar infarcts.1 2 3 4 5 Hence, it is thought that the identification of lacunar infarcts is possible by the use of simple clinical criteria and that this may also help both daily patient management and therapeutic choices so that expensive and time-consuming investigations may be avoided.5 This would be particularly useful in pharmacological trials on acute ischemic stroke. In fact, given the presumed differences in pathogenesis and clinical evolution between lacunar and nonlacunar infarcts,6 the clinical identification of the former might guide patient enrollment. However, to make a reasonably accurate clinical diagnosis of lacunar infarct, one should in theory wait for the maximal expression of the neurological deficit, which means at least 96 hours after stroke onset.2 5 Unfortunately, this time frame is well beyond the therapeutic window, ie, the time interval between stroke onset and irreversible brain tissue damage, which is set at 6 to 8 hours in nonhuman primates7 and seems to be similar in humans.8 Therefore, one might ask whether this late, albeit accurate, classification of stroke subtypes is of any use.
In a previous study on a consecutive series of 517 first-ever ischemic hemispheric stroke patients examined within 12 hours (mean±SD, 6.1±3.2 hours) of the event, we observed that pure motor hemiparesis and sensorimotor stroke syndromes, described in this early phase, were able to predict lacunar infarcts in only 56% of cases.9 The same cohort is the object of the present study, in which we compared the frequency of risk factors for stroke and the clinical outcome of stroke subtypes defined on the one hand by the clinical presentation at entry and on the other by cerebral CT or autopsy. Our aim was to evaluate whether the low degree of accuracy of the clinical diagnosis might have influenced both the compilation of risk factor profiles, considered to be indicative of the potential pathogenetic mechanisms involved,10 and the prognostic estimates.
Subjects and Methods
Patient selection has been described in detail elsewhere.9 We admit to our stroke unit patients affected by moderate to severe motor or sensorimotor deficits, with or without signs and symptoms of cortical involvement, presumably caused by a supratentorial carotid territory stroke. For the present study, all patients were seen at admission (in the stroke unit) by one of six neurologists (D.T., M.F., M. De M., M.L.S., E.M.Z., C.A.) who have a specific interest in acute stroke research. The neurological examination was performed according to a chart included in the patient’s medical record, which summarized all the items of a complete examination, including dominant and nondominant hemisphere cognitive functions. The examination took from 10 to 20 minutes, and patients were classified according to the initial clinical description.
The lacunar syndromes we observed were cases of pure motor hemiparesis and sensorimotor stroke, as described at hospital admission within 12 hours of the event. Thus, according to clinical presentation we distinguished the patients as presenting with lacunar or nonlacunar syndrome, while according to CT or autopsy data the patients were distinguished as affected by lacunar or nonlacunar infarct.
A lacunar infarct was defined as a subcortical sharply delineated lesion with a diameter equal to or less than 15 mm, observed on a CT scan or on autopsy performed within 15±2 days from stroke onset. In addition, permanently negative CT scans in patients with a neurological deficit lasting more than 24 hours were considered compatible with a lacunar infarct.9
At hospital admission, the neurological deficit was quantified with the use of the CNS,11 which has a global score ranging from 1.5 (maximum deficit) to 10 (absence of deficit). The patient’s medical history was investigated to detect one or several risk factors for stroke: hypertension, diabetes, atrial fibrillation, other potentially embolic cardiopathies,12 previous TIA, or cigarette smoking.
Patients were followed up for 30 days after stroke onset. During this period we took account of case-fatality rates and causes of death, which were divided into cardiac, cerebral, and others (eg, pulmonary embolism, bronchopneumonia, sepsis). We defined the clinical outcome of survivors by comparing the CNS score at the end of follow-up to that at hospital admission and taking a score of 7.5 as a threshold for self-sufficiency.13 Thus, the neurological status at the last examination was defined as (1) improved (final CNS score higher than that at entry and ≥7.5) or (2) impaired (final CNS score lower than or the same as that at entry or, if higher, not reaching the score of self-sufficiency).
Univariate analysis was performed with the use of the t test and χ2 test to compare demographic data, CNS score at entry, risk factors for stroke, and outcome of lacunar syndrome versus nonlacunar syndrome and of lacunar infarct versus nonlacunar infarct patients. A logistic regression analysis was then applied to the aforementioned variables to look for independent predictors of lacunar syndrome and lacunar infarct. Finally, we submitted the same variables to an additional logistic regression analysis by separately introducing first lacunar syndrome (yes=1; no=0) and then lacunar infarct (yes=1; no=0) in the model to determine whether the clinical or the CT classification of stroke subtypes was an independent predictor of outcome.
Of the 517 patients (299 men, 218 women; mean±SD age, 67±10 years) studied for 8 years, 219 (42%) presented a lacunar syndrome and 298 (58%) a nonlacunar syndrome at hospital admission, while 170 (33%) had a lacunar infarct and 347 (67%) a nonlacunar infarct on CT scan or autopsy.
Table 1⇓ shows the demographic data, the severity of neurological status at hospital admission defined by CNS score, and the frequency of risk factors of the stroke subtypes defined by clinical syndrome and by CT or autopsy. Both lacunar syndrome and lacunar infarct patients were less severely affected at entry (P<.0001) than their counterparts.
In regard to the distribution of risk factors in the patients’ medical history, no significant difference was found between lacunar syndrome and nonlacunar syndrome patients, whereas lacunar infarct subjects were more frequently associated with a previous TIA (P<.002) and hypertension (P<.05) and less frequently with atrial fibrillation (P<.01) when compared with their nonlacunar infarct counterparts.
Logistic regression did not select any variable predictive of lacunar syndrome, whereas a previous TIA (OR, 2.07; 95% CI, 1.27 to 3.37) and hypertension (OR, 1.48; 95% CI, 1.01 to 2.16) on the one hand and atrial fibrillation (OR, 0.55; 95% CI, 0.34 to 0.90) on the other were strongly correlated with the diagnosis of lacunar infarct and nonlacunar infarct, respectively.
Table 2⇓ shows 30-day outcome and causes of death in the four subgroups studied. At the end of follow-up, 87 lacunar syndrome (40%) and 187 nonlacunar syndrome patients (63%) were neurologically impaired or had died, as opposed to 37 lacunar infarct (22%) and 237 nonlacunar infarct patients (68%). In regard to causes of death, brain herniation prevailed among both nonlacunar syndrome and nonlacunar infarct patients, but in as many as 6 lacunar syndrome patients (35%) the cause of death was cerebral as opposed to none in the lacunar infarct subgroup.
Predictors of outcome selected by logistic regression were age, CNS score at entry, cardiopathies other than atrial fibrillation, and diabetes, but not lacunar syndrome. In contrast, when lacunar infarct was included in the model, it was selected as a predictor of outcome in addition to the other aforementioned variables (Table 3⇓).
The diagnosis of lacunar infarct on the basis of the clinical identification of lacunar syndrome, as described within 12 hours of the event, had proved to be inaccurate in a previous study on a consecutive hospital series of 517 first-ever ischemic hemispheric stroke patients.9 Recently, Lodder and colleagues5 pointed out some of the possible causes of the false clinical diagnosis of small deep infarcts in patients examined no earlier than 24 hours after stroke onset. We agree that speech disorders and above all inadequate testing of nondominant higher function deficits14 are undoubtedly the main factors in this misdiagnosis. Unfortunately, the weight of these routine neurological examination pitfalls is magnified in the hyperacute phase, when the frequent mental impairment and the poor collaboration of patients may hinder a thorough assessment of these functions. We therefore decided to evaluate whether the low degree of accuracy of the clinical classification of stroke subtypes might also have negatively influenced patient triage by leading to the compilation of erroneous risk factor profiles and/or by confounding the prognostic evaluation, thus resulting in inadequate therapeutic approaches. For this purpose, we chose the CT definition of lacunar infarcts as the gold standard and compared the risk factors and the clinical outcome of lacunar infarct versus nonlacunar infarct patients and those of lacunar syndrome versus nonlacunar syndrome patients.
In regard to risk factors, lacunar infarct patients were found to have a higher frequency of hypertension15 16 17 18 and previous TIAs15 and a lower frequency of atrial fibrillation than nonlacunar infarct patients,16 17 18 whereas no differences were found in the frequency of other cardiopathies, diabetes,3 18 or cigarette smoking. It is not our aim to infer pathogenetic hypotheses from these data nor to side with proponents19 or opponents20 of the “lacunar hypothesis.” However, it is worth noting that whereas lacunar infarct and nonlacunar infarct patients presented different risk factor profiles, as one could expect,6 no remarkable difference was found between lacunar syndrome and nonlacunar syndrome patients. This means that by grouping patients on clinical grounds, we would group patients with different risk factors (and therefore with potentially different pathogeneses) that might consequently benefit from different therapies. One could argue that in daily clinical practice, patient management, ie, what investigations are to be made and which treatments are to be given, depends on the patient’s own medical history and not on the stroke subtype into which we allocate them on clinical grounds.21 For instance, potential arterial or cardiac sources of emboli must be investigated if the medical history raises this possibility, even in patients presenting with lacunar syndrome.
On the other hand, how far and how actively to treat a patient are largely suggested by prognostic estimates, which in our patients were significantly biased by the clinical classification. In fact, 30 days after stroke, the case-fatality rate of lacunar syndrome patients was twice as high as that of lacunar infarct subjects, whereas as many as one third of the former and only one fifth of the latter were neurologically impaired. In contrast, nonlacunar syndrome and nonlacunar infarct patients had the same case-fatality rate and approximately the same frequency of residual neurological impairment. Therefore, it was the group of lacunar syndrome patients with nonlacunar infarcts that was most affected by the mistaken prognosis.
This conclusion is also emphasized by the observation that, in addition to age, severity of neurological status at hospital admission, cardiopathies, and diabetes, all well-known predictors of outcome,22 logistic regression also selected lacunar infarct as an independent variable strongly correlated with a good clinical evolution. In contrast, the allocation of patients to the group of those with or without lacunar syndrome did not provide any additional prognostic information.
In the context of pharmacological trials, the exclusion of lacunar strokes from randomization would in theory be justified both by the presumed differences in pathogenetic mechanisms, given that stroke trials are essentially aimed at treating patients with large-artery thromboembolic stroke, and by the very frequent spontaneous good outcome, which suggests that giving lacunar stroke patients expensive and potentially harmful treatments, such as thrombolytics, is unnecessary. Unfortunately, our data demonstrate that in an emergency setting such as that of stroke trials, it is misleading to rely solely on clinical presentation when identification of lacunar infarcts is attempted.
On the other hand, it has recently been demonstrated that focal hypodensities detected by brain CT scan performed within 5 hours of stroke onset are able to identify sizable subgroups of patients with a high frequency of early deterioration23 and poor outcome.23 24 These patients are unlikely to benefit from treatments presently under investigation and could consequently be excluded from trials to lower costs and improve the consistency of results. Moreover, it would be interesting to investigate whether it may be possible to improve the early prediction regarding which lacunar syndrome patients have lacunar infarcts by combining these early CT findings and the clinical data at admission.
In conclusion, an accurate clinical identification of lacunar infarcts is not yet possible in the acute stroke setting. This lack of accuracy in the early clinical diagnosis might lead to no distinction being made between stroke subgroups with different pathogenetic mechanisms and prognostic estimates, thus negatively influencing the planning of patient management. In this context, the patient’s medical history and imaging techniques are still an irreplaceable aid in the choice of therapeutic approaches.
Selected Abbreviations and Acronyms
|CNS||=||Canadian Neurological Scale|
|TIA||=||transient ischemic attack|
Reprint requests to Dr Danilo Toni, Department of Neurological Sciences, 1st Chair of Neurology, University “La Sapienza,” Viale dell’Università 30, 00185, Rome, Italy.
- Received February 28, 1995.
- Revision received June 15, 1995.
- Accepted June 16, 1995.
- Copyright © 1995 by American Heart Association
Bamford J, Sandercock P, Jones L, Warlow CP. The natural history of lacunar infarction: the Oxfordshire Community Stroke Project. Stroke. 1987;18:545-551.
Boiten J, Lodder J. Lacunar infarcts: pathogenesis and validity of the clinical syndromes. Stroke. 1991;22:1374-1378.
Arboix A, Martı̀-Vilalta JL. Lacunar syndromes not due to lacunar infarcts. Cerebrovasc Dis. 1992;2:287-292.
Lodder J, Bamford J, Kappelle J, Boiten J. What causes false clinical prediction of small deep infarcts? Stroke. 1994;25:86-91.
Adams RD, Victor M. Cerebrovascular diseases. In: Principles of Neurology. 5th ed. New York, NY: McGraw-Hill Publishing Co; 1993:669-748.
Toni D, Del Duca R, Fiorelli M, Sacchetti ML, Bastianello S, Giubilei F, Martinazzo C, Argentino C. Pure motor hemiparesis and sensorimotor stroke: accuracy of the very early clinical diagnosis of lacunar strokes. Stroke. 1994;25:92-96.
Warlow C. Disorders of the cerebral circulation. In: Walton J, ed. Brain’s Disease of the Nervous System. 10th ed. Oxford, England: Oxford Medical Publications; 1993:197-268.
Coté R, Hachinski VC, Shurvell BL, Norris IW, Wolfson C. The Canadian Neurological Scale: a preliminary study in acute stroke. Stroke. 1986;17:731-737.
Kittner SJ, Sharkness CM, Price TR, Plotnick GD, Dambrosia JM, Wolf PA, Mohr JP, Hier DB, Kase CS, Tuhrim S. Infarcts with a cardiac source of embolism in the NINCDS Stroke Data Bank: historical features. Neurology. 1990;40:281-284.
Toni D, Del Duca R, Sacchetti ML, Fiorelli M, Argentino C, Fieschi C. Clinical definition and CT confirmation of lacunar stroke. Cerebrovasc Dis. 1992;2:202. Abstract.
Bogousslavsky J, Van Melle G, Regli F. The Lausanne Stroke Registry: analysis of 1000 consecutive patients with first stroke. Stroke. 1988;19:1083-1092.
Chamorro A, Sacco RL, Mohr JP, Foulkes MA, Kase CS, Tatemichi TK, Wolf PA, Price TR, Hier DB. Clinical–computed tomographic correlations of lacunar infarction in the Stroke Data Bank. Stroke. 1991;22:175-181.
Sacco SE, Whisnant JP, Broderick JP, Philips JS, O’Fallon WM. Epidemiological characteristics of lacunar infarcts in a population. Stroke. 1991;22:1236-1241.
Bamford JM, Warlow CP. Evolution and testing of the lacunar hypothesis. Stroke. 1988;19:1074-1082.
Millikan C, Futrell N. The fallacy of the lacunar hypothesis. Stroke. 1990;21:1251-1257.
Caplan L. TIAs: we need to return to the question, What is wrong with Mr. Jones? Neurology. 1988;38:791-793.
Von Kummer R, Meyding-Lamadé U, Forsting M, Rosin L, Rieke K, Hacke W, Sartor K. Sensitivity and prognostic value of early CT in occlusion of the middle cerebral artery trunk. AJNR Am J Neuroradiol. 1994;15:9-15.