Featured Clinical Investigation
Prediction of Infarct Topography Using the Oxfordshire Community Stroke Project Classification of Stroke Subtypes
Amira R. A1-Buhairi, MB, ChB, FRCPC, Stephen J. Phillips, MBBS, FRCPC, Grant Llewellyn, MD, FRCPC, and Mohammed M.S. Jan, MB, ChB, FRCPC
Background: The Oxfordshire Community Stroke Project clinical classification of ischemic stroke syndromes has been shown to be predictive of important clinical outcomes, In this study, we examined the correlation between this dassification system and infarct topography on computed tomography (CT) of the brain. Method: A cohort of consecutive cases of acute ischemic stroke admitted to an acute stroke service during the 3-year period ending December 31, 1996 were identified from a prospec- tive stroke registry. Brain scans were reviewed b y a single neuroradiologist without knowledge of the clinical featu~s. Results: There were 418 patients with acute ischemic stroke who met the study admission criteria. Forty patients were excluded, 20 (5%) did not have a CT scan during the admission, and 20 scans were not available for review.
In 239 of 378 patients (63%), the brain scan revealed the lesion responsible for the clinical syndrome. In patients with positive scans, the positive predictive values of the clinical subtypes were: 86% (95% confidence interval 78-94) for the total anterior territory stroke syndrome, 96% (92-100) for the partial anterior territory stroke syndrome, 99% (97-100) for the lacunar stroke syndrome, and 100% for the posterior circulation stroke syndrome. Conclusion: The Oxfordshire Community Stroke Project classification of ischemic stroke syndromes usefully predicts infarct topography on CT scan. Key Words: Cerebral infarction--CT scan--Clinical stroke syndromes.
The ideal classification system for ischemic stroke w o u l d be simple, reproducible, a n d predictive of b o t h the u n d e r l y i n g pathophysiologic process and clinical out- c o m e s ) Several classification systems have been de- scribed, 2-6 b u t none is ideal. There is a trade-off b e t w e e n practicality and i n - d e p t h investigation.
From the Divisions of Neurology (A.R.A.-B., S.J.P., M.M.S.J.) and Neuroradiology (G.L.), Queen Elizabeth II Health Sciences Center and Dalhousie University, Halifax, NS, Canada.
Received January 12, 1998; accepted April 9,1998.
Address reprint requests to Stephen J. Phillips, MBBS, FRCPC, Division of Neurology, Queen Elizabeth II Health Sciences Center, 1796 Summer St, Halifax, NS B3H 3A7, Canada
Copyright 9 1998 by National Stroke Association 1052-3057 / 98 / 0705-001253.00 / 0
The detailed etiologic classification system d e v e l o p e d b y Johnson et al. 2 is time-consuming and requires detailed investigations. Both pathophysiological classifcation sys- tems devised b y the National Institute for Neurological Disorders and Stroke (NINDS), a n d the Stroke Data Bank 3-5 are heavily d e p e n d e n t on the results of investiga- tions. A n o t h e r classification system used in the Trial of Org 10172 in Acute Stroke Treatment (TOAST), required d a t a collected b y tests and some clinical criteria. 6 In that system, m o r e emphasis was given to the results of ancillary tests than to clinical assessment. Some of these tests are not readily available to m a n y c o m m u n i t y practi- tioners. Detailed investigation is impractical in c o m m u - nity hospitals, in treatment trials where there m a y be a n a r r o w therapeutic time window, and in epidemiological studies. Furthermore, e v e n after such detailed investiga-
Journal of Stroke and Cerebrovascular Diseases, Vol. 7, No. 5 (September-October), 1998: p p 339-343 339
340 A.R. AL-BUHAIRI ET AL.
tions, the pathophysiological mechanism in as m a n y as 40% of infarcts remains undefined. 4
Bamford et al. 7-9 described four ischemic stroke syn- dromes in the Oxfordshire C o m m u n i t y Stroke Project (OCSP), a large prospective, community-based study in which 675 patients were registered, and of w h o m 545 (81%) had cerebral infarcts. The diagnosis was confirmed in 439 patients (65%) by computed tomography (CT) within 28 days of onset or necropsy. The presenting clinical syndromes were lacunar (LACS) in 133 patients, total anterior circulation stroke (TACS) in 85 patients, partial anterior circulation stroke (PACS) in 183 patients, posterior circulation stroke (POCS) in 124 patients, and u n k n o w n in 20 patients. The four syndromes have distinc- tive patterns of survival, dependencN and stroke recur- rence as shown in Table 1. 7 For example, in the TACS group, the chance of good functional outcome is poor with high early mortality and low recurrence rate; pa- tients in the PACS group have a high early risk of recurrent stroke; in the LACS group the case-fatality rate is low; and patients in the POCS group have the highest risk of recurrence.
In this study we examined the correlation between brain CT scan findings in patients presenting with acute ischemic stroke, and their allocated_ OCSP classification sub- type. O u r goal was to determine the sensitivity, specificity, and positive predictive values of the OCSP syndromes.
Method
The Queen Elizabeth II (QEII) Health Sciences Center is a multispecialty adult hospital providing primary care to a metropolitan area of about 350,000 people, as well as secondary and tertiary care for an additional 500,000 people. A cohort of consecutive cases of acute ischemic stroke admitted to the acute stroke service during the 3-year period ending December 31, 1996 were identified
from an acute stroke registry. This prospective database includes demographic data as well as information about stroke subtype, severity, management, and outcome.
Cases of stroke are defined as the s u d d e n onset of focal cerebral dysfunction, with symptoms lasting 24 hours or longer or leading to death of apparent vascular cause. 1~
Patients were included in this study if they presented with a stroke syndrome, required admission to hospital, and had a CT scan that excluded intracranial hemorrhage and nonstroke lesions.
The OCSP subtype was determined from the findings of the first neurological examination performed b y the responsible physician in the emergency department and recorded prospectively in the stroke registry. If the clinical findings could not be classified, the stroke was typed uncertain. The severity of stroke was estimated at the time of initial clinical assessment using the stroke severity scale used in the Extracranial/Intracranial Bypass Study and the North American Symptomatic Carotid Endarterec- t o m y TrialY
Brain CT scans were reviewed by a single neuroradiolo- gist, without knowledge of the OCSP type. The latest scan was used if the patient had more than one scan performed during the hospitalization for the index stroke. The most recent infarct was used if the patient had more than one lesion on the scan. Recent infarcts were distinguished from old ones b y the degree of hypodensity and atrophy, and by reviewing previous and follow-up scans. The CT findings were recorded as follows: complete middle cerebral artery (MCA) territory infarction (esti- mated -> 75% involvement of the MCA territory); partial MCA territory infarction (estimated < 75% involvement of the MCA territory); lacunar infarction (<-1.5 cm lesion in internal capsule or basal ganglia); posterior circulation infarction (lesions in occipital lobes, brain stem, or cerebel- lum); subcortical infarction (lesions confined to the corona radiata and centrum semiovale); anterior cerebral artery Table 1. The OCSP classification o f ischemic stroke syndromes
OCSP
classification Definitions
Case fatality rate in 1 year
Recurrent stroke in 1 year
Dependency at 1 year*
1. TACS
2. PACS
3. LACS 4. POCS
Combination of higher cerebra/dysfunction (e.g., dysphasia, dyscalculia, visuospatial disorder), homonymous field defect, and ipsllateral motor and/or sensory deficit of at least 2 areas of the face, arm, and leg
Two of the 3 components of the TACS, with higher cerebral dysfunction alone, or with a motor/sensory deficit more restricted than those classified as LACS (e.g., confined to one limb, or face & hand but not the whole arm).
Pure motor, sensory, sensory-motor stroke, or ataxic hemiparesis.
Presence of any of the following: ipsilateral cranial nerve palsy with contralatelN motor and/or sensory deficit, bilateral motor and/or sensory deficit, disorder of conjugate eye movement, cerebellar dysfunction without ipsilateral long-tract deficit or isolated homonymous field defect.
60%
16%
11%
19%
6%
17%
9%
20%
36%
29%
28%
19%
*Dependency = functionally dependent (Rankin grades 3-5).
Data from Bamford et al. v-9
(ACA) territory infarction; and combinations of anterior and middle, or middle and posterior cerebral artery territory infarction.
Analyses were performed using Epi Info, version 6.12 The relationship between the clinical and radiological findings is presented as sensitivity, specificity, positive predictive value (PPV), and the 95% confidence interval (CI) for the PPV. For these analyses, CT evidence of complete MCA territory infarction, A C A territory infarc- tion, or combined MCA and A C A territory infarction was considered indicative of a TACS; partial MCA infarcts and subcortical infarcts were considered PACS; lacunar in- farcts -<1.5 cm diameter in the internal capsule and basal ganglia were considered LACS; and infarcts in the brain- stem and cerebellum were considered POCS.
The analyses were performed in three ways: (1) assum- ing that the negative CT scans were concordant with the OCSP (best-case scenario), (2) excluding the negative scans, and (3) assuming that the negative CT scans were discordant with the OCSP (worst-case scenario).
Results
During the 3-year study period, 418 patients with acute ischemic stroke were admitted to the stroke service. Of these, 378 were included in this study; 20 (5%) were excluded because they did not have a CT scan, and 20 scans were not available for review. The main reason for not obtaining a CT scan was the severity of the stroke.
More than one CT scan was performed on 87 patients (23%). Twenty-two patients had cranial magnetic reso- nance imaging (MRI) performed during hospitalization.
Cranial CT scan did not reveal the presumed infarct responsible for the presenting symptoms and signs in 139 (37%) patients. The distribution of the OCSP subtypes and CT findings is shown in Table 2. Table 3 compares the CT positive and negative groups.
Of the 239 positive CT scans, 44 patients had complete MCA territory infarcts, 85 had partial MCA territory infarcts, 7 h a d subcortical infarcts, 69 had lacunar infarcts, and 33 had posterior circulation infarcts of which 6 were pontine, 8 cerebellar, and 19 hemispheric. Discrepancy between the assigned OCSP type and the CT findings was present in 8 cases of partial MCA infarcts and 4 cases of lacunar infarcts that were classified as TACS and PACS, respectively. Table 4 shows the sensitivity, specificity, PPV, and 95% CI of the PPV of the OCSP subtypes using CT as the gold standard.
Discussion
Our findings suggest that the OCSP dinical classifica- tion system has high correlation with neuroradiological findings in cerebral infarction. Our results support the assertion that this simple clinical classification system is useful in the evaluation of stroke patients.
Table 2. CT scan findings and OCSP subtypes for 378 cases of acute ischemic stroke
OCSP subtype
CT appearances TACS PACS LACS POCS UNCE Total Responsible
lesion not
shown 15 39 40 39 6 139
Lacunar infarct 0 4 65 0 0 69
Pontine infarct 0 0 0 6 0 6
Cerebellar
infarct 0 0 0 l 8 I 19
PCA infarct 0 0 0 8 0 8
Complete MCA
infarct 40 0 0 0 0 40
Partial MCA
infarct 8 72 0 0 5 85
Subcortical
infarct 1 5 1 0 0 7
ACA infarct 0 1 0 0 0 1
ACA and MCA
infarct 0 0 0 0 1 1
MCA and PCA
infarct 0 0 0 0 3 3
Total 64 121 106 71 16 378
NOTE. CT Appearances defined in methods section.
Abbreviations: ACA, anterior cerebral artery; MCA, middle cerebral artery; PCA, posterior cerebral artery; UNCE, uncertain.
Several studies have looked at the interobserver reliabil- ity and the validity of the OCSP classification system. 13A4 In the study b y Lindley et al., 13 two clinicians indepen- dently assessed 85 patients with the diagnosis of stroke
Table 3. Comparison between the CT positive and negative groups
Positive Negative P
Variable CT CT value
No. (%) 239 (63) 139(37)
Mean age (y) 69 70 .8
Gender (% male) 60 50 .06
Mean time from stroke onset to first clinical assessment in
days 1.2 1.1 .2
Median Stroke Severity Score* 6 6 .1 OCSP subtype (%)
TACS 18 1 l .01
PACS 36 28 .2
LACS 28 29 .5
POCS 14 28 .001
UNCE 4 4
Mean time from spoke onset to
CT (days) 4.7 2.9 .001
*Stroke Severity S c o r e 11 determined at time of first clinical assessment.
342
Table 4. The correlation between the OCSP subtype and CT findings
OCSPsubtype Sensitivity Specificity PPV 95% CI TACS
Best case scenario 93 97 86 78-94 Positive scans only 91 95 82 72-92 Worst case scenario 91 93 63 51-75 PACS
Best case scenario 89 98 96 92-100 Positive scans only 84 97 94 88-99 Worst case scenario 84 85 64 56-72 LACS
Best case scenario 96 99 99 97-100 Positive scans only 94 99 98 95-100 Worst case scenario 94 87 61 52-70 POCS
Best case scenario 99 100 100 - - Positive scans only 97 100 100 - - Worst case scenario 97 98 45 34-56
within 10 days of hospital admission. They found that interobserver agreement was satisfactory (K 0.62; 95% CI, 0.45-0.78). Lindgren et al. 14 compared the clinical and neuroradiological findings in first-ever stroke. They in- cluded 228 patients, 28 with intracereberal hemorrhage.
In that stud)~ clinical evaluation was performed by a neurologist at the emergency department, and a second examination was done within I week of admission. They found significant relationship between the clinical stroke classification and features on brain imaging. Serial brain imaging (two CT scans and one MRI) were performed.
Despite serial brain imaging, the brain lesion responsible for the stroke syndrome was not identified in about 30%
of patients.
In our stud36 37% of the patients had negative CT scans.
This could be the result of the lesion being too small (suspected in 40 patients with lacunar infarcts), the scan being performed too early (60 patients were scanned within 48 hours of stroke onset), or because the lesion was in the brainstem (suspected in 37 patients). In other instances, cortical atrophy and subcortical leukaraiosis m a y have masked small infarcts. The distribution of stroke dinical subgroups in our study was similar to that reported in the original study by Bamford et al. 7-9
There are a few limitations to our study. First, the clinical assessment was not performed by single physi- cian; however, it has been shown that the OCSP classifica- tion has satisfactory interobserver reliability. 13 Secondly, CT scans were repeated inconsistently and at different times after the onset of symptoms; however, serial CT scanning in another study did not substantially increase the yield of positive scans. 14
In addition to its correlation with infarct topography, the OCSP is reasonably predictive of the pattern of underlying vascular patholog)5 i.e., major vessel occlu-
A.R. AL-BUHAIRI ET AL.
sion in TACS, no evidence of occlusion in LACS, and patent major vessels and probable MCA branch occlusion in PACS. is A simple and reliable clinical classification of ischemic stroke that predicts infarct topography, vascular pathology, and clinical outcome could find increasing application in trials of treatment for acute stroke. In these situations, time windows are short, CT scans are fre- quently negative, and more complicated classification systems that require investigations such as MR1, single- proton emission tomograph34 and transcranial Doppler, seriously impede recruitment of patients. The OCSP classification has been used successfully in the Interna- tional Stroke Trial, 16 the Clomethiazole Acute Stroke Study, ~7 and the Chinese Acute Stroke Trial. is
We conclude that the OCSP stroke classification system has high correlation with neuroradiological findings in cerebral infarction. The validity, reliability, and simplicity of the OCSP classification system make it a useful tool for the categorization of patients into prognostic groups in clinical trials and epidemiological studies. Also, it pro- vides information useful for planning the investigation of patients and the counseling of patients and their families in clinical practice.
R e f e r e n c e s
1. Bamford J. Clinical examination in diagnosis and subclas- sification of stroke. Lancet 1992;339:400-405.
2. Johnson CJ, Kitten SJ, McCarter RJ, et al. lnterrater reliability of an etiologic classification of ischemic stroke.
Stroke 1995;26:46-51.
3. National Institute of Neurological Disorders and Stroke.
Classification of cerebrovascular diseases III. Stroke 1990;
21:637-676.
4. Sacco RL, Ellenberg JH, Mohr JP, et al. Infarct of undeter- mined cause: The NINDS Stroke Data Bank. Ann Neurol 1989;25:382-390.
5. Kunitz SC, Gross CR, Heyman A, et al. The Pilot Stroke Data Bank: Definition, design, data. Stroke 1984;15:740- 746.
6. Adam HP Jr, Bendixen BH, Kappelle LJ, et al., and the TOAST Investigators. Classification of subtype of acute ischemic stroke. Definitions for use in a multicenter clinical trial. Stroke 1993;24:35-41.
7. Bamford J, Sandercock P, Dennis M, et al. Classification and natural history of clinically identifiable subtypes of cerebral infarction. Lancet 1991;337:1521-1526.
8. Bamford J, Sandercock P, Dennis M, et al. A prospective study of acute cerebrovascular disease in the community:
The Oxfordshire Community Stroke Project 1981-86 (methodology, demography and incident cases of first- ever stroke). J Neurol Neurosurg Psychiatry 1988;51:1373- 1380.
9. Bamforfl J, Sandercock P, Dennis M, et al. A prospective study of acute cerebrovascular disease in the community:
The Oxfordshire Community Stroke Project 1981-86 (inci- dence, case fatality rates and overall outcome at one year). J Neurol Neurosurg Psychiatry 1990;53:16-22.
10. Bontia R. Epidemiology of stroke. Lancet 1992;8:342-346.
PREDICTION OF INFARCT TOPOGRAPHY
11. The EC/IC Bypass Study Group. The International Coop-
erative Study of E• Arterial By-
pass (EC / IC Bypass Study). Stroke 1985;16:397-406.
12. Dean A, Dean J, Burton A, et al. Epi Info: A general- purpose microcomputer program for public health infor- mation systems. Am J Prev Med 1991;7:178-182.
13. Lindley RI, Warlow CP, Wardlaw JM, et al. Interobserver reliability of a clinical classification of acute cerebral infarction. Stroke 1993;24:1801-1804.
14. Lindgren A, Norrving Bo, Rudling O, et al. Comparison of clinical and neuroradiological findings in first-ever stroke. Stroke 1994;25:1371-1377.
15. Naylor AR, Sandercock PAG, Sellar RJ, et al. Patterns of vascular pathology in acute first-ever cerebral infarction.
Scot Med J 1993;38:41-44.
16. International Stroke Trial Study Collaborative group. The International Stroke Trial (IST): A randomised trial of aspirin, subcutaneous heparin, both, or neither among 19,435 patients with acute ischaemic stroke. Lancet 1997;
349:1569-1581.
17. Wahlgren NG, Clomethiazole Acute Stroke Study CoUabo- rative Group. The Clomethiazole Acute Stroke Study (CLASS): Efficacy results in a subgroup of 545 patients with total anterior circulation syndrome. Stroke 1998;29:
287.
18. CAST (Chinese Acute Stroke Trial) Collaborative Group.
CAST: Randomised placebo-controlled trial of early aspi- rin use in 20,000 patients with acute ischemic stroke.
Lancet 1997;349:1641-1649.
