Journal ofNeurology, Neurosurgery, and Psychiaty 1993;56:1265-1270 1265 J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.56.12.1265 on 1 December 1993. Downloaded from

PAPERS Pathogenetic and prognostic features of lacunar transient ischaemic attack syndromes

Gianluca Landi, Cristina Motto, Ermelinda Cella, Massimo Musicco, Susanna Lipari, Edoardo Boccardi, Mario Guidotti

Abstract studies reporting a significantly lower preva- Lacunar ischaemic syndromes are lence of emboligenic heart disease and of a well defined subgroup of ischaemic carotid among lacunar than . To determine whether a similar among other ischaemic strokes.3-5 Diagnosis subgroup can be identified among of lacunar ischaemic stroke syndrome has patients with transient ischaemic attacks prognostic implications, too, as these patients (TIAs) we studied prospectively 102 con- have a more favourable long term outcome secutive patients within 24 hours of their than those with other subtypes of cerebral first TIA. Based on their history they infarction.67 Transposition of the lacunar were classified as lacunar TIA syn- hypothesis to TIAs is not straightforward, as dromes (LTIAS; n = 45) if isolated motor their short duration usually precludes objec- or sensory symptoms or their combina- tive verification of symptoms; thus, diagnosis tion had involved at least two of three of a lacunar TIA syndrome would have to body parts (face, arm, leg), whereas all rest on the patient's history and on the clini- other subjects were grouped as non-lacu- cian's possibility to fit it retrospectively into nar TIA syndromes (NLTIAS; n = 57). one of the lacunar syndromes. Despite this All patients were investigated according limitation, two papers have endeavoured to to a standardised protocol and followed distinguish a subgroup with small vessel dis- up for an average of 51 1 months. ease also among TIAs. One study considered Cardiac and arterial sources of throm- patients with TIA who had an infarct at CT: boembolism were more frequent among the authors reported a significantly higher NLTIAS (p = 0 0001). Survival curve proportion of small deep (lacunar) infarcts on analysis demonstrated that LTIAS had a CT scan among patients with unilateral iso- significantly lower long term mortality lated motor and/or sensory symptoms consis- and incidence of major vascular events. tent with a lacunar syndrome than among In a multivariate regression analysis, the patients with symptoms suggestive of cortical type of TIA (that is, NLTIAS) was an dysfunction.8 In another study, limited to independent predictor of stroke or death. patients who had undergone angiography, http://jnnp.bmj.com/ LTIAS share the same distinct patho- subjects with a lacunar TIA syndrome had a Neurological Clinic, genetic and prognostic features of lacu- significantly lower prevalence of symptomatic Ospedale Policlinico; nar ischaemic stroke syndromes. These internal than those Milan, Italy findings have implications for manage- with a cortical TIA.9 We have studied an uns- G Landi C Motto ment of TIAs and for studies oftheir nat- elected population with TIA to determine E Cella ural history and treatment. whether lacunar TIA syndromes share the

S Lipari same distinct pathogenetic and prognostic on September 30, 2021 by guest. Protected copyright. National Research U, Neurol Neurosurg Psychiatry 1993;56:1265-1270) features of lacunar ischaemic stroke syn- Council, Institute of dromes. Advanced Biomedical Technologies; Milan, Italy The customary distinction between transient M Musicco ischaemic attacks (TIAs) and ischaemic Patients and methods Department of strokes, based conventionally on duration of This prospective study includes a consecutive Neuroradiology, to or 24 series of visited from 1 September Ospedale Niguarda; symptoms up beyond hours, conveys patients Ailan, Italy no information on the pathogenesis or prog- 1983 to 31 August 1990 by one of the study E Boccardi nosis of the attack. Subdivision of ischaemic neurologists at the emergency room of the Division ofNeurology, strokes according to their clinical manifesta- Policlinico Hospital in Milan within 24 hours Ospedale Valduce, tions allows identification of lacunar stroke of their first TIA. This is the only emergency Como, Italy syndromes.' These well defined constellations room in the city's inner district where all M Guidotti with Correspondence to: of neurological signs are generally caused by patients presenting symptoms suggestive Dr Gianluca Landi, Clinica small deep cerebral infarcts, the lacunes, of receive free neuro- Neurologica, Ospedale the and it is Policlinico, Via F Sforza 35, which in turn are attributed to a distinct vas- logical evaluation around clock, 20122 Milano, Italy. culopathy leading to occlusion of small perfo- often resorted to for urgent specialist exami- Received 5 October 1992 rating .2 The hypothesis that lacunar nation bypassing the lengthy waiting lists of and in revised form TIA as an 19 March 1993. ischaemic strokes are caused by small vessel the local health service. We defined Accepted 26 March 1993 disease has been strengthened by several acute focal loss of cerebral or ocular function 1266 Landi, Motto, Cella, Musicco, Lipani, Boccardi, Guidotti J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.56.12.1265 on 1 December 1993. Downloaded from attributed to ischaemia with symptoms lasting symptoms. They were further subdivided into less than 24 hours. Isolated vertigo, diplopia, superficial and deep; the latter were diag- or , drop attacks, tran- nosed as lacunar infarcts if their maximum sient global amnesia, as well as focal symp- diameter did not exceed 15 mm on no more toms associated with migrainous headache or than two adjacent 10 mm tissue sections. spreading with a pattern suggestive of Cerebral angiography was carried out in 24 epilepsy or migraine were not accepted as evi- potential candidates for carotid endarterec- dence of TIA. We excluded patients with pre- tomy (14 with LTIAS and 10 with NLTIAS). vious stroke, even if it had occurred in the Four of them were randomised in the interval between TIA and emergency room European carotid surgery trial and two of evaluation. A detailed history of the attack them underwent carotid endarterectomy (one was obtained, and the clinical features of all with LTIAS and one with NLTIAS); no patients were reviewed and discussed with a other patient had carotid endarterectomy. second study neurologist to improve diagnos- There were no angiographic or surgical com- tic interobserver reliability.10 Based on his- plications. tory, patients were subdivided in lacunar TIA Emergency room examination was taken as syndromes (LTIAS) and non-lacunar TIA zero time from which event-free survival was syndromes (NLTIAS). LTIAS were diag- measured. All patients were followed up for nosed if symptoms of isolated unilateral at least 12 months or until death (average, motor or sensory deficit or their combination 51i1 months), and all survivors were re- had involved at least two of three body parts examined by one of us in September 1991. (face, arm, leg), partially or completely. Antiplatelet treatment with (300 to Occurrence of dysarthria did not exclude a 500 mg/day) or with ticlopidine (500 mg/day) LTIAS if the symptom was attributed to was generally prescribed with the exception of supranuclear weakness of the mouth or five patients with NLTIAS who received oral tongue. All other episodes consistent with anticoagulants. We considered patients as diagnosis of TIA, as established by accepted treated with antithrombotic drugs if they had criteria," were grouped as NLTIAS. One taken this therapy for at least half of their patient who had experienced both types of total follow up time. Control of vascular risk attack before evaluation was included in the factors was pursued in all cases. Occurrence NLTIAS group. of new cerebrovascular events (which were All patients were screened for the presence classified as TIAs or strokes according to of vascular risk factors: (blood duration of symptoms up to or beyond 24 pressure repeatedly higher than 160/90 mmHg hours) was recorded; a CT scan was per- or regular use of antihypertensive drugs); dia- formed whenever possible in case of stroke. betes (fasting blood sugar higher than Stroke disability was evaluated at approxi- 140 mg/100 ml or regular use of antidiabetic mately 3 months by the modified Rankin drugs); smoking (at least 10 cigarettes daily scale,"3 and strokes were classified as follows: during the previous 6 months), hyperlipi- not disabling (Rankin grades 0-1); partially daemia (fasting plasma cholesterol and/or disabling (grades 2-3); severely disabling triglycerides higher than 240 mg/100 ml and (grades 4-5); or fatal. Occurrence of myocar- 170 mg/i 00 ml, respectively, or regular use of dial infarction was also recorded. Diagnosis lipid-lowering drugs). A continuous wave was accepted if at least two of the following Doppler study of the extracranial vessels was criteria were fulfilled: typical chest pain, con- http://jnnp.bmj.com/ performed as described by Buedingen et al.12 cordant ECG changes, and enzymatic alter- ECG, chest radiograph, and cardiological ations. Causes of death were ascertained by evaluation were carried out in all cases; reviewing clinical records and necropsy depending on specific diagnostic needs, M- reports; if the cause remained uncertain, rela- mode and B-mode echocardiograms were tives and attending physicians were also con- obtained in 28 cases (18 with LTIAS and 10 sulted. Vascular deaths included those due to

with NLTIAS), and 24 hour ECG monitor- stroke, myocardial infarction, cardiac failure, on September 30, 2021 by guest. Protected copyright. ing in eight cases (two with LTIAS and six peripheral arterial disease, and ruptured aor- with NLTIAS). Potential cardioembolic tic , as well as sudden presumed sources were diagnosed in patients with atrial cardiac deaths. fibrillation, endocarditis, mitral valve disease (including mitral valve prolapse in patients STATISTICAL METHODS under 45 years of age), left ventricular Statistical analyses were performed using aneurysm, or thrombus. Ischaemic heart dis- Epilog Plus software package.'4 The strength ease was diagnosed in patients with a typical of association between the considered vari- history of angina or acute myocardial infarc- ables and the type of TIA (LTIAS or tion, or with clear cut ECG evidence of previ- NLTIAS) was calculated by means of odds ous myocardial infarction. A CT scan was ratio (OR),15 taking LTIAS as reference cate- performed as part of the emergency room gory, and their statistical significance was evaluation, and was repeated after 5 to 8 evaluated by the x2 test. Confidence intervals days; all scans were reviewed by a neuroradi- (CI) of OR were calculated as suggested by ologist (EB) who was blind with respect to Cornfield.'6 The significance of the difference clinical diagnosis of LTIAS or NLTIAS. between means was tested with two tailed t Cerebral infarcts were defined as circum- tests for unpaired data. Analyses of the cumu- scribed hypodense lesions and were consid- lative time dependent probability of event ered as congruous if appropriate to the side of free survival were carried out by Kaplan- Pathogenetic andprognosticfeatures oflacunar transient ischaemic attack syndromes 1267 J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.56.12.1265 on 1 December 1993. Downloaded from Table 1 Age, sex, vascular risk factors, and results ofcardiological and instrumental type, as opposed to four out of 10 among evaluation according to TIA subtpe NLTIAS (OR = 0-22, 95% CI 0-01-2'42, p Lacunar Non-lacunar = 0 3 1). Overall, 1 1 patients had a CT incon- TM TM syndromes* syndromes gruous infarct, six with LTIAS (five lacunar (n = 45) (n = 57) OR (95% CI) and one non-lacunar infarct) and five with Age (>65 years) 24 (53%) 39 (68%) 1 90 (0 85-4 24) NLTIAS (four lacunar and one non-lacunar Male sex 28 (62%) 30 (53%) 1-48 (0 60-3 68) infarct). The prevalence of ischaemic heart Hypertension 28 (62%) 29 (51%) 0-63 (0-25-1-56) 7 (16%) 8 (14%) 0-89 (0 26-3 07) disease was similar in the two groups of Smoking 27 (60%) 25 (44%) 0-52 (0-21-1-28) patients, but emboligenic heart disease was Hyperlipidaemia 19 (42%) 24 (42%) 1 00 (0-72-1-37) Infarct at CT scan 12 (27%) 13 (23%) 0-81 (029-228) significantly more frequent among NLTIAS Congruous infarct 8 (18%) 10 (18%) 0-98 (0 35-2-74) (p = 0-004). Occlusion or more than 50% Ischaemic heart disease 8 (18%) 10 (18%) 0-98 (0 43-2 23) Emboligenic heart disease 2 (4%) 16 (28%) 8-39 (1-81-38-90) stenosis of the symptomatic arterial district Stenosis/occlusion of symptomatic 4 (9%) 14 (25%) 3-34 (0 90-12 30) was demonstrated by angiography in four of neck vessel (Doppler findings) Total with ascertained 5 (11%)t 29 (51%)t 8-29 (270-25 40) 10 patients with NLTIAS and in one of 14 thromboembolic source with LTIAS. However, since angiography *Reference category. was performed only in a selected group of tOne patient had concomitance of emboligenic heart disease and stenosis of symptomatic neck patients, we assessed the prevalence of vascu- vessel. lar lesions by means of Doppler ultrasounds. NLTIAS had a higher frequency of obstruc- tive lesions of the symptomatic neck vessel than LTIAS; among patients without emboli- Meier method, and the strength of associa- genic heart disease, only three out of 43 tion of these probabilities with possible prog- LTIAS had pathological Doppler findings, as nostic variables was evaluated by hazard ratio opposed to 13 of 41 NLTIAS (OR = 6-19; (HR).'7 Cox's model was employed to carry 95% CI 1-50-23-51; p = 0-009). Overall, out multivariate regression analysis.'8 patients with NLTIAS had a 4-5 times higher prevalence of cardiac or arterial sources of thromboembolism than LTIAS (p = 0-0001). Results Data on follow up are presented in table 2, A total of 102 patients with first TIA within and the figure shows the Kaplan-Meier sur- the previous 24 hours were seen during the vival curves for major vascular events and study period, 45 (44%) with LTIAS and 57 their combinations. Since patients with (56%) with NLTIAS. All had experienced LTIAS were on average younger than those TLIAs of the , and none reported iso- with NLTIAS, and a greater proportion of lated . Twenty four (53%) of them had taken antithrombotic drugs, results the 45 LTIAS had presented isolated motor of survival curve analysis have been expressed symptoms, three (7%) isolated sensory symp- as HR after adjusting for age and for treat- toms, and 18 (40%) a combination of senso- ment with antithrombotic drugs. Prognosis of rimotor symptoms. The average age of LTIAS was significantly more favourable with patients was 64-7 years (range 24-83) for regard to vascular mortality and to all consid- those with LTIAS and 68&7 (range 43-89) for ered combinations of major events; overall those with NLTIAS (t = 1-76, p = 0-08). The mortality was likewise lower in this group = = prevalence of vascular risk factors is reported (HR 5-2; 95% CI 1-5-18-1; p 0-01). http://jnnp.bmj.com/ in table 1. No significant difference was LTIAS also had a lower incidence of strokes, observed between the two groups. Results of which were less often disabling or fatal than cardiological and instrumental evaluations are those experienced by NLTIAS (table 2). A also reported in table 1. Although the propor- multivariate regression analysis including, tion of infarcts on CT scan was rather similar besides type of TIA, all variables listed in in the two groups, six of the eight congruous table 1 identified three independent prognos- infarcts among LTIAS were of the lacunar tic factors associated with occurrence of stroke or death, namely non-lacunar type of on September 30, 2021 by guest. Protected copyright. TIA (HR = 2-55, 95% CI 1-12-5-82; p = 0.02), diabetes (HR = 2-68; 95% CI 1-09-6 52; p = 0-03), and emboligenic heart Table 2 Follow up results disease (HR = 2-55; 95% CI 1-06-6-12; p = Lacunar Non-lacunar 0-03). TMA TM syndromes syndromes (n = 45) (n = 57) Mean duration (months) 56-6 46-7 Discussion Treated with antithrombotic drugs 35 (78%) 36 (63%) Recruitment of patients at the emergency Stroke: 6 (13%) 13 (23%)* room is unusual in studies on it not disabling 3 2 TIA, yet partially disabling 1 2 helped to obviate some potential drawbacks severely disabling 1 5 fatal 1 4 of our investigation. Firstly since most Myocardial infarction 1 (2%) 8 (14%)t patients presented without having been previ- fatal 0 7 examined a it Other vascular death 0 (0%) 2 (4%)t ously by physician, probably Non-vascular death 2 (4%) 5 (9%) reduced the referral bias inherent to hospital Overall number with major vascular 7 (16%) 20 (35%) based studies. In fact, the age and sex distrib- events (stroke, myocardial infarction, vascular death) ution of our population and the prevalence of vascular risk factors were more similar to *One patient with a previous myocardial infarction. tOne patient with a previous stroke. those observed in a community recruited 1268 Landi, Motto, Cella, Musicco, Lipani, Boccardi, Guidotti J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.56.12.1265 on 1 December 1993. Downloaded from

100- 100-

80- ...... 80- a) ...... -o I 0 U) I...... a1) ISI- 0 60- 60- I cnU) 0 0 2 U) 6 40- a) 40- in

HR = 2.0 -I HR = 2.4 LTIAS ) 20- LTIAS 20- 95% Cl: 0.7 - 5.3 NLTIAS 95% CI :1.1 -5.4 -111,11NLTIAS p=0.19 p=0.03

6 I 7-- I-l c 2 4 6 8 2 4 6 8 Years Years 45 38 25 13 1 45 38 25 13 1 No at risk No at risk 57 39 27 9 0 57 39 27 9 0 L1 100- 100-711 ..I- ...... 11 111-...... I...... 80- = 80- I ...I er ...... I ..... "I'l""I'll""I'll'll""I'll""I'll",'ll""Il. U) -)CD -0 60- ...... 6 60- 0 ...... c6 a) I...... 0 0 U) 40- 40- a)EU) L-U) HR = 2.7 HR = 10.3 U) LTIAS LTIAS 20- 95% C: 1.2-6.3 NLTIAS 8 20- 95% C: 1.3-80.1 NLTIAS a- p = 0.02 p=0.03 n 0 I I--l 2 4 6 8 0 2 4 6 8 Years Years 45 38 25 13 1 45 39 28 15 1 No at risk No at risk 57 39 27 9 0 57 43 30 10 0

[ 100- I'll -1 co ...... a) ...... 80- k.. ci I U) - ...... I.... co 11 60- I...... http://jnnp.bmj.com/ 1- 60- 0 4 -

U) - HR = 5.0 LTIAS -0 20- 95% Cl: 1.4 - 17.3 NLTIAS a) p = 0.01 a) - 0~' 0.

2 4 6 8 on September 30, 2021 by guest. Protected copyright. Years 45 38 26 14 No at risk .1111. 57 40 27 9 0

Figure Kaplan-Meier survival curves showing the percentages surviving: (A) free ofstroke (ignoring those dyingfor other causes); (B) free of stroke or death; (C) free of disabling stroke-that is, Rankin grade 2 or worse-or vascular death (ignoring those dyingfor other causes); (D) free of stroke, myocardial infarction or death; (E) free of vascular death (ignoring those dyingfor other causes). Hazard ratios (HR) and their 95% CIs have been calculated after adjustmentfor age andfor treatment with antithrombotic drugs. LTIAS = lacunar TIA syndromes; NLTIAS = non-lacunar TIA syndromes.

cohort than to those of hospital referred pop- include patients in the study within 24 hours ulations.19 Secondly, history could be of symptom onset and thus to identify events obtained shortly after the attack, allowing manifesting very soon after TIA, thereby early prospective differentiation of lacunar enhancing completeness of follow up, as illus- from non-lacunar episodes, not biased by trated by occurrence of four strokes during results of subsequent examinations. Finally, the first week. On the other hand, none of emergency room evaluation enabled us to our patients had experienced amaurosis Pathogenetic andprognostic features oflacunar transient ischaemic attack syndromes 1269 J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.56.12.1265 on 1 December 1993. Downloaded from fugax, a condition which may harbour a bet- that these same features characterise a similar ter prognosis than do TIAs of the brain,20 as subgroup among patients with TIA, who by these patients were first evaluated by an oph- definition experience complete resolution of thalmologist and only later referred to the symptoms within 24 hours, supports the neurologist. validity of the lacunar hypothesis. We there- Our criteria for diagnosis of LTIAS, which fore suggest transposition of the term "lacu- ensure an excellent degree of reliability nar" from its original pathological setting25 to among different observers,21 were similar to a clinical one, to designate attacks which those of Kappelle et al,8 who in addition manifest with specific patterns of signs or included in this subgroup patients with symp- even symptoms. Although it cannot be toms restricted to one body part. Hankey and excluded that, in a given patient, an LTIAS is Warlow9 adopted stricter criteria, as they due to thromboembolism from arterial or car- required involvement of all three body parts diac sources, the low prevalence of stenosis of or, in patients with brachiofacial or bra- the symptomatic neck vessel and of emboli- chiocrural symptoms, that the whole of each genic heart disease observed among patients body part was affected, information which with LTIAS, resembling those reported in may be difficult to ascertain from history. asymptomatic individuals of similar age,2S29 is Moreover, they considered only right handed consistent with the possibility that these patients with left hemispheric TIA who had abnormalities represent general markers of attempted to speak during the episode, thus atherosclerosis rather than factors involved in allowing them to evaluate occurrence of the pathogenesis of the attack. Careful history aphasia. While these criteria increased the taking, a simple and widely applicable sensitivity of the diagnosis of lacunar TIA method not dependent on sophisticated tech- syndrome, they also reduced its field of nologies, thus allows identification of a sub- applicability to only 26-5% of the authors' group of patients with TIA who may benefit original TIA population. At variance with the from a distinct diagnostic and therapeutic two abovementioned studies we did not approach and who should be considered sep- exclude patients with TIA in a vertebrobasilar arately in studies on the natural history and distribution, since lacunar syndromes may treatment of this condition. result from lesions, and because there is no definite evidence that posterior circulation TIAs have a peculiar pathogenesis and prognosis.22 Moreover, as we wanted to 1 Bamford JM, Sandercock P, Jones L, Warlow C. The nat- verify whether clinical history alone could ural history of lacunar infarction: the Oxfordshire Community Stroke Project. Stroke 1987;18:545-51. identify a pathogenetically and prognostically 2 Boiten J, Lodder J. Lacunar infarcts. Pathogenesis and homogeneous subgroup out of an unselected validity of clinical syndromes. Stroke 1991;22:1374-8. 3 Lodder J, Bamford JM, Sandercock P, Jones L, Warlow population with TIA, we did not reclassify C. Are hypertension or cardiac likely causes our patients according to CT scan results. In of lacunar infarction? Stroke 1990;21:375-8 1. 4 Tegeler CH, Shi F, Morgan T. Carotid stenosis in lacunar fact, 44% of our patients with "positive" CT stroke. Stroke 1991;22:1124-8. scan had clinically silent lesions, implying 5 Kappelle LJ, Koudstaal PJ, van Gijn J, Ramus L, Keunen J. Carotid angiography in patients with lacunar infarc- also that appropriately sited lesions may be tion: a prospective study. Stroke 1988;19:1093-6. mistakenly associated with occurrence of 6 Bamford JM, Sandercock P, Dennis M, Burn J, Warlow TIA. It is nonetheless remarkable that the C. Classification and natural history of clinically identi-

fiable subtypes of . Lancet 1991;337: http://jnnp.bmj.com/ proportion of congruous CT lacunar infarcts 1521-6. 7 Landi G, Cella E, Boccardi E, Musicco M. Lacunar ver- among our LTIAS (75%) and NLTIAS sus non lacunar infarcts: pathogenetic and prognostic (40%) were equal to those observed by differences. J Neurol Neurosurg Psychiatry 1992;55: 441-5. Kappelle et al.8 8 Kappelle LJ, van Latum JC, Koudstaal PJ, van Gijn J. Our study demonstrates that cardiac and Transient ischaemic attacks and small-vessel disease. Lancet 1991;337:339-41. arterial sources of thromboembolism are 9 Hankey G, Warlow C. Lacunar transient ischaemic associated with the occurrence of NLTIAS. attacks: a clinically useful concept? Lancet 1991;337: 335-8. The same features characterise non-lacunar 10 Koudstaal PJ, van Gijn J, Staal A, Duivenvoorden HJ, on September 30, 2021 by guest. Protected copyright. strokes.3-5 Reports on the long term outcome Gerritsma JGM, Kraaijeveld CL. Diagnosis of transient ischaemic attacks: improvement of interobserver agree- of stroke agree that patients with lacunar ment by a check-list in ordinary language. Stroke ischaemic stroke syndromes have a more 1986;17:723-8. 11 The ad hoc committee on the classification and outline of benign prognosis than those with other stroke cerebrovascular disease. II. Stroke 1975;6:566-616. subtypes.6 Our study indicates that the 12 Buedingen HJ, von Reutern GM, Freund HJ. Doppler- Sonographie der extrakraniellen Hirnarterien. Stuttgart: same holds true for LTIAS. Furthermore, Georg Thieme, 1982. results of our multivariate regression analysis 13 van Swieten JC, Koudstaal PJ, Visser MC, Schouten HJA, van Gijn J. Interobserver agreement for the assessment demonstrate that lacunar TIA represents an ofhandicap in stroke patients. Stroke 1988;19:604-7. independent prognostic variable, which 14 Epilogplus: Statistics packagefor epidemiology and dinical tri- als. Pasadena, CA: Epicenter software. 1985. remains significantly associated with a 15 Mantel M, Haenszel W. Statistical aspects of the analysis favourable outcome even if other risk factors of data from retrospective studies of disease. J Nad Cancer Inst 1959;22:719-48. are considered. Again, the same findings 16 Cornfield J. A statistical problem arising from retrospec- apply to lacunar ischaemic stroke.7 tive studies. In: Neyman J, ed. Proceedings of the third Berkeley symposium, IV. Berkeley: University of The validity of the lacunar hypothesis California Press, 1956:133-48. relating lacunar ischaemic stroke syndromes 17 Kaplan EL, Meier P. Non-parametric estimation from incomplete observations. J Am Stat Assoc 1958;53: to small vessel disease has been questioned 457-81. on the assumption that their pathogenetic 18 Cox DR. Regression model and life-tables. Y R Statist Soc B 1972;34:187-220. and prognostic features are not peculiar but 19 Hankey GJ, Slattery JM, Warlow CP. The prog,nosis of simply those of "small strokes".24 Our finding hospital referred transient ischaemic attacks. Y Neurol 1270 Landi, Motto, Celia, Musicco, Lipani, Boccardi, Guidotti J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.56.12.1265 on 1 December 1993. Downloaded from Neurosurg Psychiatty 1991;54:793-802. esis. Stroke 1990;21:1251-7. 20 Dennis MS, Bamford JM, Sandercock PAG, Warlow CP. 25 Fisher CM. Lacunes: small, deep cerebral infarcts. A comparison of risk factors and prognosis for transient 1965;15:774-84. ischemic attacks and minor ischemic strokes. The 26 Ostrander LD, Brandt RL, Kjelsberg MO, Epstein FH. Oxfordshire Community Stroke Project. Stroke 1989;20: Electrocardiographic findings among the adult popula- 1494-9. tion of a total natural community, Tecumseh, 21 Landi G, Candelise L, Celia E, Pinardi G. Interobserver Michigan. Circulation 1965;31:888-98. reliability of the diagnosis of lacunar transient ischemic 27 Kannel WB, Abbott RD, Savage DD, McNamara PM. attack (TIA with lacunar syndrome). Cerebrovasc Dis Epidemiologic features of chronic : the 1992;2:297-300. Framingham Study. N EnglJ Med 1982;306: 1018-22. 22 Cartlidge NEF, Whisnant JP, Elveback LR. Carotid and 28 Faris AA, Poser CM, Wilmore DW, Agnew CH. vertebro-basilar transient ischemic attacks. Mayo Clin Radiologic visualization of neck vessels in healthy men. Proc 1977;52:117-20. Neurology 1963;13:386-96. 23 Sacco SE, Whisnant JP, Broderick JP, Phillips SJ, 29 Harrison MJG, Marshall J. Angiographic appearance of O'Fallon WM. Epidemiological characteristics of lacu- carotid bifurcation in patients with completed stroke, nar infarcts in a population. Stroke 1991;22:1236-41. transient ischaemic attacks, and cerebral tumour. BMJ7 24 Millikan CH, Futrell N. The fallacy of the lacune hypoth- 1976;1:205-7.

The first description of idiopathic progressive the modes of presentation of motor neuron disease bulbar palsy and which may occasionally remain confined to the bulbar region. Given the advanced disease, a brain- Furukawa has recently recounted the lesser known stem tumour seems unlikely in the absence of sensory contributions to clinical neurology of Sir Charles Bell signs and if the bulbar palsy was due to myasthenia who first described the numb chin syndrome' and the gravis, ocular and limb symptoms would be expected. phantom phenomenon2 in his book on the nervous Bell's life is well documented but little is known of system.3 We would like to report an account of a the referring physician. Robinson was born in patient with progressive bulbar palsy (PBP) who was Lancashire and graduated Doctor of Medicine at referred to Bell, is included in the same book and pre- Edinburgh on 12 September 1800. He was president dates the writings of Duchenne de Boulogne,4 who is of the Royal Medical Society, founded in 1734 by a usually credited with the first descriptions.5 group of medical students after the successful acquisi- The patient's details are contained in a letter to Bell tion of a fresh body for dissection. The founders from Robert RW Robinson, of Preston, England, initially met in taverns ".... with the intention that dated 21 July 1825. He asked for advice about "an a dissertation in English or Latin on some medical subject unmarried lady, nearly seventy years of age, who has ... should be composed and read"6 and later meetings enjoyed uninterrupted good health up to the present were attended by Charles Darwin. There is no evi- illness". He states: dence that Robinson had an earlier interest in neurol- From the first of her complaint to the present moment ogy as he obtained a doctorate based on a urological she has been free from headach (sic) and from pain, subject (Disputatio medica inauguralis de vesicea, ure- numbness, or debility of the limbs. The vision and hearing thraeque morbis). He became a Licentiate of the are natural; the appetite good; the bowels regular, and the College of Physicians (London) in 1807 and returned sleep natural .... Somefew months ago she had some dif- to his native Lancashire where he was an influential ficulty in the and in member of the board of management of the Preston using tongue expressing particular http://jnnp.bmj.com/ words. This difficulty has gradually increased, and now Dispensary, founded in 1809. Unfortunately 169 she cannot protrude the tongue, or even move it. She has years after his original account, the aetiology of PBP lost her speech altogether. The tongue itself is soft and remains obscure. but it retains its sense taste and The ANDREW M CHANCELLOR pulpy; of feeling. deg- Department ofNeurology lutition is impaired and occasionally she is distressed with a University Hospital of Wales, sense of suffocation, in attempting to swalow food, which CardiffCF4 4XW, UK she is now obliged to do with great care. She afnot hack JOHN D MITCHELL up any thingfrom the throat, nor draw any thingfrom the Department ofNeurology, posterior nares by a back draught. The features of the face Royal Preston Hospital, on September 30, 2021 by guest. Protected copyright. are quite natural, and the skin retains its feeling. The Sharoe Lane, Preston, UK saliva occasionallyflows from the mouth." ROBERT J SWINGLER Bell noticed the similarity between this patient's Department ofNeurology, condition and the syndrome observed in a dog after Dundee Royal Infirmary, section of the lower cranial nerves and concluded that Dundee, UK the patient was suffering from "a paralytic affection of the ninth nerve" and noted that the "function of thefifth nerve was entire." Unfortunately no follow up or Correspondence to: Dr Chancellor necropsy results are available to tell us if the disease became more widespread, neither is there any infor- mation about the deep tendon reflexes, as this physical 1 Furukawa T. Charles Bell's description of numb chin syn- sign was not introduced into the neurological exami- drome. Neurology 1988;38:331. nation until 1875. Bell advised nauseating medicines 2 Furukawa T. Charles Bell's description of the phantom and leeches under the other reme- phenomenon in 1830. Neurology 1990;40:1830 mastoid, amongst 3 Bell C. The nervous system of the human body: As explained dies. in a series ofpapers read before the Royal Society ofLondon. This description would rival any modem case With an appendix of cases and consultations on nervous dis- report for conveying clearly the clinical signs, the eases. London: Longman, 1836: 390-1. 4 Duchenne de Boulogne G. Paralysie musculaire progres- important negative features for the diagnosis and the sive de la langue, du voile du palais et des 1evres. Archiv disability resulting from this tragic affliction. We Gin Mid 1860;2:283-96. would argue that this patient's illness, which combines 5 Morton LT. A medical bibliography. An annotated checklist oftext illustrating the history ofmedicine. 4th edn. London: anarthria, impaired deglutition and tongue movement Butler and Tanner, 1983:603-46. with ptyalism, probably with mixed upper and lower 6 Holmes JD. Early years of the medical society of motor neuron signs, is due to idiopathic PBP, one of Edinburgh. Univ Edin J 1968;23:333-40.