Initially, 8,437 patient records were screened via a search of the electronic medical records of patients who were hospitalized at our institution over an 11-year period between February 1, 2004 and November 31, 2013, and then followed up until November 31, 2014. In total, there were 157 patients with acute ischemic stroke (IS) involving a lateral medullary lesion, as identified on brain MRI scans within 7 days after stroke onset. All patients were at least 18 years of age (range, 30-91 years) and were followed up at our hospital during the study period. Patients whose brain MRI and conventional angiography were limited to evaluation of LMI lesions and arterial lesions, and for whom there was a lack of medical and clinical information and no outcome data, were excluded from the study. The baseline characteristics did not differ significantly between the study patients and those who were excluded (supplementary data). This study was approved by the Institutional Review Board of our medical institution. The requirement for informed consent was waived due to the retrospective nature of this study.

Baseline characteristics such as demographic data (age and sex) and conventional vascular risk factors [hypertension, diabetes mellitus, a past history of smoking, hyperlipidemia, heart disease, and a past history of stroke/transient ischemic attack (TIA)] were evaluated. Several laboratory findings related to inflammation, such as white blood cell (WBC) count, and serum concentrations of glucose and C-reactive protein (CRP), were also evaluated using routine laboratory methods. Blood samples were withdrawn after overnight fasting. The following definitions were used for clinical assessments:

1) Hypertension: a history of antihypertensive treatment, a systolic blood pressure of ≥140 mm Hg, or a diastolic blood pressure of ≥90 mm Hg.12

2) Hyperlipidemia: a history of antihyperlipidemia medication or a serum level of total cholesterol of >6.2 mmol/L (240 mg/dL).13

3) Diabetes mellitus: glycated hemoglobin (HbA1c) of ≥6.5%, fasting blood glucose of ≥126 mg/dL, or nonfasting blood glucose of ≥200 mg/dL, or the use of insulin or oral hypoglycemic drugs.14

4) Heart diseases: including arrhythmia, coronary artery disease, congestive heart failure, and valvular heart diseases.

5) Smoking: indicated if a patient was either a current smoker or an ex-smoker.

All of the patients were initially evaluated for stroke severity based on the National Institutes of Health Stroke Scale (NIHSS) neurological examinations, including sensory symptoms and ataxia, for which the functional status was evaluated by board-certified experts in neurology at the outpatient clinic or through a structured telephone interview by a trained nurse at 3 months and 1 year poststroke. These evaluations represented attempts to monitor the quality of inpatient stroke care. Initial dysphagia was identified based on electronic medical records using the initial bedside swallowing tests, evaluated by trained physicians within 1 day after admission to the hospital. This test included an evaluation of the level of consciousness, deglutition, voice change, evidence of coughing, drooling, oral motor control, control of bolus in the oral phase, and evidence of delay in the swallowing reflex. The patients were then managed by diet modification and swallowing therapy.1516

Functional outcomes were evaluated using the modified Rankin Scale (mRS), the scores for which were used to divide the patients into two groups: favorable outcome (mRS score ≤1) and unfavorable outcome (mRS score ≥2).17 The occurrence of clinical events during a 1-year period since the onset of LMI was evaluated; these events included IS, TIA, myocardial infarction, pneumonia, ICU treatment, and death. The following definitions were used:

1) IS: the presence of findings suggestive of stroke with a duration of ≥24 hours, or MRI-confirmed acute IS with a duration of <24 hours.

2) TIA: Acute loss of focal neurological signs of vascular origin, accompanied by the resolution of symptoms within 24 hours.

3) Death: sudden death occurring within a 1-year period since the onset of LMI as a result of vascular or nonvascular events due to underlying conditions such as cancer, pneumonia, or respiratory failure.

4) Pneumonia: radiologic evidence of consolidation or infiltration and at least one of the following conditions: granulocytosis, fever of >38℃ and/or cough, sputum, and shortness of breath.18

Based on the Trial of Org 10172 in Acute Stroke Treatment criteria and arterial dissection alone, cases of stroke were categorized as follows:19202122 1) large-artery atherosclerosis (LAA), 2) SVO (hypertension, no cardioembolic source and normal angiography), 3) cardioembolism, and 4) VA or PICA dissection.

Brain MRI/MR angiography (MRA) and conventional angiography were used to assess the LMI lesion and dissection of the VA and PICA. Brain MRI/MRA was performed using a 1.5-T or 3.0-T superconducting magnet system. VA or PICA dissections were defined as images of a 'double lumen,' 'intimal flap,' 'string-of-pearls appearance,' 'steno-occlusion,' and 'fusiform aneurysmal dilatation' on either brain MRI/MRA or conventional angiography.22 Based on previous reports, the LMI lesions were classified into three categories: rostral, middle, and caudal (Fig. 1).523 Two board-certified experts in neurology who were blinded to the clinical characteristics retrospectively reviewed and classified the findings on brain MRI/MRA and conventional angiography.

Student's t-test and Pearson's χ² test were used to compare differences in continuous and categorical variables, respectively, between the two groups. Multiple logistic regression analysis was also performed to identify correlations between age, conventional vascular risk factors, initial symptoms, and the pathophysiology of stroke, and functional outcomes using statistically significant variables (p<0.05). In addition, odds ratios (ORs) and 95% confidence intervals (CIs) were calculated by adjusting those variables known to have a correlation with functional outcomes, including patient age, conventional vascular risk factors, initial symptoms, clinical complications, pathophysiology of the stroke, laboratory finding such as WBC count, serum CRP, and LMI lesions. All the tests were two-sided, and the threshold of statistical significance was set at p<0.05. All statistical analyses were conducted using SPSS software (Version 21.0, SPSS Inc., Chicago, IL, USA).

In total, 157 patients with acute LMI were enrolled; these patients were aged 60.1±12.3 years (mean±SD; range, 30-91 years) and showed a male predilection (73.2%). The demographic characteristics, conventional vascular risk factors, initial clinical symptoms, laboratory findings, and types of stroke are presented in Table 1 relative to functional outcome at 1 year. Of the 157 patients, 93 (59.2%) had favorable outcome and 64 (40.8%) had an unfavorable outcome. The mean age was significantly higher in the unfavorable outcome group than in the favorable outcome group (p=0.003), as were the prevalence rates of hypertension and hyperlipidemia (p=0.008 and p<0.001, respectively). The NIHSS score did not differ significantly between the two groups [mean, 2 (range 1-3) vs. 2 (range 1-4); p=0.437]. The most common type of stroke was LAA in both groups (61.3% and 62.5%, for the favorable and unfavorable outcome groups, respectively). The prevalence of cardioembolism was significantly higher in the unfavorable outcome group than in the unfavorable outcome group, whereas that of SVO was significantly higher in the latter (p=0.016) (Table 1). At 1 year after LMI onset, five patients had suffered a stroke/TIA arising from LAA (data not shown). The serum level of CRP was significantly higher in the unfavorable outcome group than in the favorable outcome group (p=0.036).

The prevalence of dysphagia at the time of stroke was significantly higher in the unfavorable outcome group (p=0.017) (Table 1); however, the prevalence rates of sensory symptoms and ataxia did not differ significantly between the two groups (Table 1). The baseline characteristics, stroke types, and outcome are presented in Table 2 relative to the presence of dysphagia. None of the baseline characteristics or conventional risk factors differed significantly between those with and without dysphagia, and LAA was the most common stroke type in both groups. Cardioembolism was significantly more common among those with dysphagia, whereas SVO was significantly more common among those without dysphagia (p=0.006) (Table 2).

The mortality and prevalence of pneumonia appeared to be higher among those with dysphagia, but the difference was not statistically significant (p=0.190 and 0.423, respectively). However, the serum level of CRP, WBC count, and frequency of an unfavorable prognosis were all significantly higher among those with dysphagia than among those without dysphagia (p=0.004, 0.028, and 0.017, respectively) (Table 2).

The clinical symptoms/signs and prognosis relative to the lesion type are presented in Table 3. Although dysphagia appeared to be more prevalent in the middle medulla, the difference was not statistically significant (p=0.061); however, sensory symptoms were significantly more prevalent in the caudal region (p=0.026). In addition, LAA was the most common of all of the LMI lesions (p=0.143). Cardioembolism was the most common LMI lesion in the rostral region (16.7%, 5.3%, and 0% in the rostral, middle, and caudal regions, respectively).

There was a nonsignificant tendency toward a greater rate of stroke/TIA recurrence and pneumonia with rostral medulla LMIs (p=0.665 and 0.557, respectively). However, ICU treatment and mortality were encountered significantly more frequently among patients with LMIs in the rostral region as compared with the other two regions (p=0.002 and 0.002, respectively). The rostral medulla tended to be a less common LMI location than either the middle or caudal medullae in the favorable outcome group, but again the differences did not reach statistical significance (p=0.135 and 0.186, respectively) (Table 3).

At 1 year post-LMI, the frequencies of pneumonia and ICU treatment were significantly higher in the unfavorable outcome group than in the favorable outcome group (p=0.025 and 0.001, respectively) (Table 1). Seven of the patients died of medical conditions such as septic shock and underlying critical diseases (gastric cancer and hematologic cancer); there were no cardiovascular deaths (Table 1). The mean age was significantly lower in the favorable outcome group than in the unfavorable outcome group (57.7±11.0 years vs. 63.6±13.4 years, p=0.003).

The impact of several factors on functional outcome at 1 year was investigated. Conditional logistic regression analysis revealed that older age and initial dysphagia were independently associated with unfavorable outcome at 1 year (older age: OR=1.04, 95% CI=1.001-1.087, p=0.049; initial dysphagia: OR=2.46, 95% CI=1.04-5.84, p=0.041). In addition, hyperlipidemia had a profound impact in the unfavorable outcome group compared to the other variables (Table 4).