Lateral and Medial Medullary Infarction
A Comparative Analysis of 214 Patients
Background and Purpose— No large-scale study has ever compared the clinical and radiological features of lateral medullary infarction (LMI) and medial medullary infarction (MMI). The aim of this study was to investigate them through the use of cooperatively collected cases.
Methods— Medical information on all patients from 1996 to 2000 with medullary infarction (MI) proven by brain MR images at 35 stroke centers in the Tohoku district, Japan, was collected, and their clinical and radiological features were analyzed.
Results— A total of 214 cases of MI were registered. They included 167 cases (78%) of LMI, 41 (19%) of MMI, and 6 (3%) of LMI plus MMI. The mean age of onset and the male-to-female ratio were 60.7 years and 2.7:1 in LMI and 65.0 years and 3.6:1 in MMI, respectively. The middle medulla was most frequently affected in LMI, and the upper medulla was most frequently affected in MMI. Dissection of the vertebral artery was observed in 29% of LMI and 21% of MMI. Prognosis, assessed by the Barthel Index, was favorable in both LMI and MMI. Diabetes mellitus was more frequently associated with MMI than with LMI.
Conclusions— The present study surveyed a large number of MI cases and revealed that (1) the mean age of onset of MMI is higher than that of LMI, (2) the dissection of the vertebral artery is an important cause not only of LMI but also of MMI, and (3) diabetes mellitus is frequently associated with MMI.
- cerebral infarction
- diabetes mellitus
- magnetic resonance imaging
- medulla oblongata
- vertebral artery
Before the era of MRI, a definite diagnosis of medullary infarction (MI), especially medial medullary infarction (MMI), could only be achieved by autopsy studies.1–3 For this reason, only a limited number of cases could be examined. Since the advent of MRI, the clinical diagnosis of MI has been made possible, and several clinical studies of MI are available.4–18 However, clinical information of MI, especially MMI, is insufficient. For example, although Kim and Choi-Kwon 18 reported that MMI constituted 25% of all MI, other researchers believed that MMI is rare.15–17 With regard to vascular changes causing MI, the dissection of the vertebral artery (VA) is observed in a considerable number of patients with lateral medullary infarction (LMI).19–21 However, it is still controversial whether the VA dissection plays an important role in the etiology of MMI.14–17 It is also of importance to know whether there is a difference in the age of onset and the risk factors between LMI and MMI. To elucidate these issues, we compared the clinical and radiological findings between LMI and MMI from 214 patients with MI.
Subjects and Methods
The Tohoku district, which is located in northern Honshu island (the main island), Japan, consists of 6 prefectures and has a population of 9 817 589 according to the 2000 population census data of Japan. All cases of MI with the onset during the period from January 1996 to December 2000 were registered from board-certified neurosurgeons/neurologists at 35 major stroke centers in the Tohoku district, which belong to the Study Group of the Association of Cerebrovascular Disease in Tohoku. All MIs proven by MRI (0.5, 1.0, or 1.5 T) were enrolled in this study. The upper (rostral), middle, and lower (caudal) portions of the medulla oblongata were defined according to Bassetti et al.16 MR angiography (MRA)/conventional angiography were also performed in 129 (60%) of 214 patients with MI. According to the criteria reported,19–21 the images of “double lumen,” “intimal flap,” and “pearl and string sign” on brain MRI, MRA, or a conventional angiogram were accepted as the presence of vascular dissection. The topographical patterns of MIs were determined with the use of the schemes by Vuilleumir et al7 and Bassetti et al16: small midlateral, dorsolateral, inferolateral, large inferodorsolateral, dorsal, hemimedullary, paramedian, bilateral paramedian, and unilateral pyramidal. Neurological features were analyzed according to the topographical patterns. Also examined was the presence of the ipsilateral cerebellar hemispheric infarctions and vascular pathology on the MR images.
The following risk factors were assessed to determine the association with the infarction: age, sex, hypertension, hypercholesterolemia, diabetes mellitus, arrhythmia, and ischemic heart disease. Hypertension was defined as a history of antihypertensive medication use, a systolic blood pressure ≥140 mm Hg, or a diastolic blood pressure ≥90 mm Hg during the chronic stage of the stroke. Hypercholesterolemia was defined as a history of antihypercholesterolemia medication or a serum level of total cholesterol ≥5.69 mmol/L (220 mg/dL); diabetes mellitus was defined as the use of insulin or oral hypoglycemic drugs, HbA1c ≥6.5%, fasting blood glucose ≥7.78 mmol/L (140 mg/dL), or nonfasting blood glucose ≥11.11 mmol/L (200 mg/dL). To determine the outcome of MI, the activities of daily living of the patients were assessed twice with the Barthel Index, within 7 days of onset and at approximately 1 month after onset. To compare the contribution of the aforementioned risk factors for LMI and MMI, multiple logistic regression analysis was performed to determine the extent of the contribution of those risk factors to MMI relative to LMI. Age, sex, hypertension, diabetes mellitus, hypercholesterolemia, ischemic heart disease, arrhythmia, and cigarette smoking were used as the covariates to calculate odds ratios (ORs) for MMI relative to LMI. Continuous values, such as age and scores on the Barthel Index, were presented as mean±SD and analyzed by the Student t test. Categorized data were presented as a percentage and analyzed by the χ2 test. Statistical analysis was performed with the programmed package SPSS, version 11.0. P<0.05 was accepted as statistically significant.
The present study was approved by the Ethical Review Committee of Yamagata University.
During a period of 5 years (January 1996 to December 2000), a total of 214 patients (159 men and 55 women) with MI were treated at 35 major stroke centers in the Tohoku district, Japan. Of the 214 patients, 167 (78%) (122 men and 45 women) had LMI, 41 (19%) (32 men and 9 women) had MMI, and 6 (3%) (5 men and 1 woman) had LMI plus MMI (Babinski-Nageotte syndrome22). The male-to-female ratio was 2.7:1 in LMI and 3.6:1 in MMI. The mean age (±SD) of onset of MI was 61.3±12.4 years. The mean age (±SD) of onset of LMI and MMI was 60.7±12.4 years and 65.0±12.3 years, respectively; the onset age of MMI was significantly older than that of LMI (P=0.034) (Figure 1). There was no significant difference in the age of onset of MI between men and women.
On brain MR images, the lesions of LMI were most frequently located in the middle medulla (35%), and those of MMI were most frequently located in the upper medulla (56%) (Table 1). The χ2 test showed that there was a significant difference in the preferential location (upper, middle, or lower) of the infarcts between LMI and MMI (P=0.019). On MRA and/or conventional angiography, which were performed in 129 cases of MI, dissection of the VA with images of double lumen, intimal flap, and/or pearl and string sign was found in 31 cases (29%) of LMI (n=107), in 4 cases (21%) of MMI (n=19), and in 1 case (33%) of LMI plus MMI (n=3) (Figure 2). In MMI, VA dissection was observed only in paramedian infarct. The frequency of VA dissection was not significantly different between LMI and MMI (P=0.586) (Figure 2). Most of the remaining cases of LMI and MMI had atherosclerotic changes of VA on MRA and/or conventional angiography.
The neurological symptoms and signs are summarized in Table 2. The sensory disturbance of the extremities and face (89%), dysarthria (75%), vertigo/dizziness (73%), Horner’s syndrome (72%), cerebellar ataxia (69%), and diminished pharyngeal reflex (64%) were the major symptoms and signs of LMI. Lingual palsy was observed in 9% of LMI, in which the infarcts were extended to the dorsal region of the medulla. In MMI, on the other hand, motor weakness (93%) and sensory disturbance (68%) of the extremities were most frequently observed; tongue weakness was present in only 30% of MMI. Consciousness disturbance occurred in 24% of the patients at the onset of MMI. It was very mild and transient, ie, characterized by somnolence or confusion. The patient’s consciousness became clear soon after onset. The cause of the consciousness disturbance was unclear but may have been due to a transient insufficiency of systemic circulation or general dehydration. The clinical outcome assessed with the Barthel Index was favorable in both LMI and MMI. The only exception was an MMI patient who had a contralateral hypoplasia of the VA with subsequent multiple infarcts of the brain stem leading to death. The mean (±SD) scores of the Barthel Index were 63.4±33.3 in LMI and 57.6±38.3 in MMI within 7 days of onset and 85.0±25.5 in LMI and 78.8±30.6 in MMI at 1 month after onset. These scores of the Barthel Index were not significantly different between LMI and MMI (within 7 days of onset, P=0.357; 1 month after onset, P=0.202).
Table 3 shows the topography of the MIs based on horizontal MRI sections, which identified 6 distinct subgroups of LMI. Representative neurological findings of Wallenberg syndrome,23 consisting of sensory disturbance of the face and body, Horner’s sign, and cerebellar ataxia, appeared in each subgroup with similar frequencies. Sensory disturbance usually showed a crossed (ipsilateral face and contralateral body) or unilateral (contralateral face and body) pattern. Pain loss restricted to the facial region was a rare presentation, at <8%, as shown in Table 3. Of the 157 cases of LMI with information of sensory function, 40 cases (25%) had a central pain syndrome (thermal hypesthesia with touch and thermal allodynia) without any correlation with a specific topographical subgroup. Horner’s syndrome usually showed ipsilateral miosis and ptosis, but anhidrosis was relatively unusual (17% of LMI cases). Of the 167 cases of LMI, 41 cases (25%) had limb paresis. Eleven cases (7%) had an ipsilateral spastic hemiplegia, consistent with Opalski syndrome. In our series, no case had cruciform paralysis or facial pain, which was reported by Fisher et al.24 Other presentations such as nystagmus, dysphagia, and vertigo were noted in each subgroup of MI, as shown in Table 3, without any specific predominance. On the other hand, dysphagia was found more often in the infarcts of the rostral than caudal medulla among various clinical presentations.
In our series the VA alterations were predominant in all topographical subgroups of LMI (Table 3). Among them, small midlateral and olivary infarcts had no posterior inferior cerebellar artery (PICA) lesions, while dorsal and large inferodorsolateral infarcts had PICA occlusions more often than the other subgroups. Cerebellar hemispheric infarctions were associated with 34 of 167 LMI cases (20%), 6 of 41 MMI cases (15%), and 1 of 6 hemimedullary infarcts (Babinski-Nageotte syndrome) (17%). On the other hand, pure medullary infarctions were seen in 76 of 167 patients (46%) with LMI, 15 of 41 (37%) with MMI, and 1 of 6 (17%) with LMI plus MMI. Cerebellar infarcts were noted predominantly in dorsal infarcts (42%) of LMI. Inferolateral infarction rarely had cerebellar infarcts (Table 3).
MMI also showed topographical subgroups: paramedian (23 cases), bilateral paramedian (6 cases), and unilateral pyramidal (10 cases) infarcts were identified. There were 2 cases of unclassified pattern. Bilateral paramedian infarct showed bilateral paralysis of 4 limbs, while other topographical subgroups of MMI usually had contralateral hemiplegia. Patterns and extent of sensory disturbance were poorly correlated with the subgroups of MMI. Similar to the cases of LMI, Horner’s sign usually consisted of ipsilateral miosis and ptosis.
Figure 3 illustrates the results of the multiple logistic regression analysis to show the contribution of medical conditions to MMI relative to LMI. Among several covariates such as age, sex, hypertension, diabetes mellitus, hypercholesterolemia, arrhythmia, ischemic heart disease, and cigarette smoking, only age and diabetes mellitus were independent risk factors for MMI relative to LMI; the OR was 1.467 (95% CI, 1.017 to 2.115; P=0.041) for age and 2.476 (95% CI, 1.104 to 5.549; P=0.028) for diabetes mellitus. Other variables were not shown to be independent risk factors for MMI relative to LMI. In MMI, there was no significant difference in the frequency of diabetes mellitus between rostral and caudal medulla.
In the present study a total of 214 cases of MI, including 167 cases of LMI and 41 cases of MMI, were examined; this number is the largest reported thus far. For MMI in particular, most reports have dealt with 1 or a few cases of MMI.1–3,11–13 Some larger studies included 18 and 14 cases of MMI (Kim et al14,18) and 11 cases of MMI (Toyoda et al15 and Kumral et al17). In the series of Kim and Choi-Kwon,18 it was reported that 75% of MIs were LMI and 25% were MMI. In the present series of 214 cases of MI, 78% had LMI and 19% had MMI. These results indicate that the ratio of LMI and MMI was approximately 3:1 to 4:1. These previous studies and ours also showed a male preponderance in both LMI and MMI, with the male-to-female ratio being approximately 3:1. With respect to the age of onset, Kim and Choi-Kwon18 reported that there was no difference in the mean age of onset between LMI and MMI. In contrast, we observed a statistically significant difference in the mean age of onset between LMI and MMI; the mean age (65.0 years) of MMI patients was approximately 5 years older than that (60.7 years) of LMI patients. This discrepancy seems to result from the small number of patients examined by Kim and Choi-Kwon18 (a total of 55 patients with MI, including 41 with LMI, 14 with MMI, and 1 with LMI plus MMI).
Bassetti et al16 reported that there was only 1 patient with MMI among their 7 MMI patients, who had an infarct in the upper part of the medulla. On the other hand, Kim et al14 and Toyoda et al15 observed that the upper medulla was most frequently involved in MMI (12 of 18 cases by Kim et al and 8 of 11 cases by Toyoda et al). The results of our study were consistent with those of Kim et al14 and Toyoda et al.15 In LMI, Kim et al6 reported that the lesions were located in the upper medulla in 8 patients, in the upper and middle medulla in 4, in the middle medulla in 8, in the middle and lower medulla in 4, and in the lower medulla in 9; therefore, LMI lesions were almost equally distributed in the upper, middle, or lower portion of the medulla among their 33 LMI patients. In our 167 patients with LMI, the lesions were most frequently encountered in the middle medulla, the second most frequently in the upper medulla, and the least in the lower medulla. The statistical analysis revealed that the preferential location (upper, middle, or lower) of the lesions was significantly different between LMI and MMI (P=0.019), ie, MMI affected most frequently the upper medulla, and LMI affected most frequently the middle medulla.
In the vascular changes causing MI, atherosclerosis of the VA and its branches has been known to be the major cause of MI. In addition, the dissection of the VA has been shown to be an important cause of LMI.19–21 In MMI, however, the dissection of the VA was reported to be unusual; there was no patient with VA dissection among the 18 patients with MMI reported by Kim et al14 and only 1 patient with traumatic dissection of the VA among the 11 patients with MMI by Toyoda et al.15 However, Bassetti et al16 observed VA dissection in 3 of 7 patients with MMI, although the location of the VA dissection was extracranial in their cases. In the present study we detected intracranial VA dissection in 29% of LMI patients and 21% of MMI patients; the frequency of VA dissection was not significantly different between LMI and MMI (P=0.586). Three patients had VA dissection during physical exercise (baseball, golf/badminton, or chiropractic manipulation). No traumatic dissection of the VA other than the aforementioned physical exercise was found in our series. The criteria for the presence of vascular dissection that we used were the images of double lumen, intimal flap, and/or pearl and string sign on MRA and/or conventional angiography. These criteria are fairly strict; as a result, some cases of MMI with VA dissection may have been regarded as having no VA dissection (false-negative result). Therefore, >21% of patients with MMI would have been classified as having VA dissection. The present study revealed that dissection of the VA is an important cause not only of LMI but also of MMI.
Another implication of the present study is reconfirmation of the wide spectrum of LMI in topographical patterns. In addition to the topographical subgroups of LMI reported by Vuilleumier et al,7 we identified an unrecognized pattern that showed an infarction restricted to the inferior olivary nucleus. The lesion never involved the surface of the medulla, suggesting that the perforator occlusion would result in the unique topographical pattern. Furthermore, our study disclosed the frequency of topographical subgroups of LMI. In our series, dorsolateral (36%) and inferolateral (27%) infarctions were frequent, while small midlateral (11%), large inferodorsolateral (9%), dorsal (7%), and olivary infarcts (9%) were relatively few.
As shown in Figure 3, age and diabetes mellitus were independent and significant risk factors for the occurrence for MMI relative to LMI, with ORs of 1.467 and 2.476, respectively. The other factors were not significant, independent risk factors contributing to the MMI. The previous reports also described a high prevalence of diabetes mellitus in patients with MMI (11 of 18 cases of MMI by Kim et al14 and 6 of 11 cases of MMI by Toyoda et al15), although they did not analyze it or comment on it. In an autopsy study, the frequency of severe atherosclerosis of the intracranial VA was reported to be greater in diabetic patients than in nondiabetic ones.25 This may explain why the prevalence of diabetes mellitus is high in MI. However, it remains unclear why diabetes mellitus is more frequent in MMI than in LMI. The medial portion of the medulla is supplied mainly by the anteromedial medullary arteries, and the lateral portion of the medulla is supplied mainly by the lateral medullary arteries.16 At the level of the lower medulla, the anteromedial arteries arise from the anterior spinal artery. In the upper medulla, the anteromedial arteries usually arise from the VA. The lateral medullary arteries arise from the VA and the PICA. Although we do not provide a plausible explanation for the high frequency of diabetes mellitus in MMI at present, one can speculate that the anteromedial medullary arteries may possibly be more susceptible to the condition of diabetes mellitus than the lateral medullary arteries.
In conclusion, the present study revealed that (1) the mean age of onset of MMI is higher than that of LMI, (2) the dissection of the VA is an important cause not only of LMI but also of MMI, and (3) diabetes mellitus is frequently associated with MMI. The present results will provide a new research field of MMI and diabetes mellitus.
Stroke Centers of the Study Group of the Association of Cerebrovascular Disease in Tohoku, Japan, that contributed to this study are as follows: Aomori Rosai Hospital, Hatinohe City Hospital, Akita Prefectural Cerebrovascular Research Center, Iwate Medical University Hospital, Iwate Prefectural Central Hospital, Iwate Prefectural Fukuoka Hospital, Iwate Prefectural Kuji Hospital, Morioka Red Cross Hospital, General Hanamaki Hospital, Iwate Rosai Hospital, Tohoku University Hospital, National Sendai Hospital, Kohnan Hospital, Ishinomaki Red Cross Hospital, Senseki Hospital, National Miyagi Hospital, Furukawa Municipal Hospital, Kesen-numa General Hospital, Southern Tohoku General Hospital, Sendai East Neurosurgical Hospital, Sendai Tokusyukai Hospital, Saiseikai Yamagata Hospital, San-yudo Hospital, Okitama Public General Hospital, Yamagata Prefectural Central Hospital, Yonezawa Municipal Hospital, Yamagata Prefectural Nihonkai Hospital, Kitamurayama General Hospital, Yamagata University Hospital, Yamagata City Hospital Saiseikan, Yamagata Prefectural Kahoku Hospital, Fukushima Medical University Hospital, Fukushima Red Cross Hospital, Shirakawa Kosei Hospital, Takeda General Hospital.
The authors thank the members of the Study Group of the Association of Cerebrovascular Disease in Tohoku for providing the MI data in Tohoku district.
A complete list of the participants in the Study Group of the Association of Cerebrovascular Disease in Tohoku appears in the Appendix.
- Received June 10, 2003.
- Revision received November 23, 2003.
- Accepted December 2, 2003.
Kase CS, Varakis JN, Stafford JR, Mohr JP. Medial medullary infarction from fibrocartilaginous embolism to the anterior spinal artery. Stroke. 1983; 14: 413–418.
Norrving B, Cronqvist S. Lateral medullary infarction: prognosis in an unselected series. Neurology. 1991; 41: 244–248.
Kim JS, Lee JH, Suh DC, Lee MC. Spectrum of lateral medullary syndrome: correlation between clinical findings and magnetic resonance imaging in 33 subjects. Stroke. 1994; 25: 1405–1410.
Vuilleumier P, Bogousslavsky J, Regli F. Infarction of lower brainstem: clinical, aetiological and MRI-topographical correlations. Brain. 1995; 118: 1013–1025.
Kim JS. Sensory symptoms in ipsilateral limbs/body due to lateral medullary infarction. Neurology. 2001; 57: 1230–1234.
Aydogdu I, Ertekin C, Tarlaci S, Turman B, Kiylioglu N, Secil Y. Dysphagia in lateral medullary infarction (Wallenberg’s syndrome): an acute disconnection syndrome in premotor neurons related to swallowing activity? Stroke. 2001; 32: 2081–2087.
Sawada H, Seriu N, Udaka F, Kameyama M. Magnetic resonance imaging of medial medullary infarction. Stroke. 1990; 21: 963–966.
Toyoda K, Hasegawa Y, Yonehara T, Oita J, Yamaguchi T. Bilateral medial medullary infarction with oculomotor disorders. Stroke. 1992; 23: 1657–1659.
Tyler KL, Sandberg E, Baum KF. Medial medullary syndrome and meningovascular syphilis: a case report in an HIV-infected man and a review of the literature. Neurology. 1994; 44: 2231–2235.
Kim JS, Kim HG, Chung CS. Medial medullary syndrome: report of 18 new patients and a review of the literature. Stroke. 1995; 26: 1548–1552.
Toyoda K, Imamura T, Saku Y, Oita J, Ibayashi S, Minematsu K, Yamaguchi T, Fujishima M. Medial medullary infarction: analyses of eleven patients. Neurology. 1996; 47: 1141–1147.
Bassetti C, Bogousslavsky J, Mattle H, Bernasconi A. Medial medullary stroke: report of seven patients and review of the literature. Neurology. 1997; 48: 882–890.
Kim JS, Choi-Kwon S. Sensory sequelae of medullary infarction: differences between lateral and medial medullary syndrome. Stroke. 1999; 30: 2697–2703.
Hosoya T, Watanabe N, Yamaguchi K, Kubota H, Onodera Y. Intracranial vertebral artery dissection in Wallenberg syndrome. AJNR Am J Neuroradiol. 1994; 15: 1161–1165.
Hosoya T, Adachi M, Yamaguchi K, Haku T, Kayama T, Kato T. Clinical and neuroradiological features of intracranial vertebrobasilar artery dissection. Stroke. 1999; 30: 1083–1090.
Nakane H, Okada Y, Sadoshima S, Fujishima M. Babinski-Nageotte syndrome on magnetic resonance imaging. Stroke. 1991; 22: 272–275.
Currier RD, Giles CL, DeJong RN. Some comments on Wallenberg’s lateral medullary syndrome. Neurology. 1961; 11: 778–791.