Study design and patients
The Tokyo Women’s Medical University (TWMU) Stroke Registry is an ongoing, prospective, single-center, observational study that enrolled patients with acute ischemic stroke or transient ischemic attack (TIA) within one week of onset, hospitalized at our center. 14,15 The study conforms to the ethical guidelines of the 1975 Declaration of Helsinki, Ethical Guidelines for Epidemiological Research by the Japanese government, and Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines. The study protocol was approved by the ethics committee of the Tokyo Women’s Medical University Hospital (approval no. 2955-R2). Written informed consent was obtained from all patients. The TWMU Stroke Registry is registered at UMIN000031913 (https://upload.umin.ac.jp). The data supporting the findings of this study are available from the corresponding author upon reasonable request.
Of 882 patients consecutively assessed for eligibility between December 2013 and January 2020, ten were excluded as their final diagnosis was a stroke mimic, more than one week had elapsed after stroke onset, or duplicate registrations were found. In addition, we excluded six patients who had no lipid or glucose profile data at baseline, leaving 866 patients (stroke, n = 781; TIA, n = 85) in the sample (Fig. 1).
All strokes and TIAs were diagnosed by board-certified stroke neurologists based on neurological and radiological findings. Patient data were collected for demographics, clinical symptoms during the qualifying event, medical history, investigations (including standard blood chemistry, brain and cerebral artery imaging, and cardiac work-up), management (medical treatment, revascularization procedure, and surgery), and the occurrence of clinical events after the qualifying event, using a structured case report form. Upon admission, neurological symptoms were assessed using the National Institutes of Health Stroke Scale (NIHSS) score.
TyG index
The TyG index was calculated as ln (fasting TG [mg/dL] × fasting glucose [mg/dL]/2). 2
The TG and glucose levels were assessed after overnight fasting within 48 hours of admission. We divided the patients into three groups according to tertiles of the TyG index: tertile 1, < 8.48; tertile 2, 8.48–9.01; and tertile 3, > 9.01.
Evaluation of vascular diseases
Intracranial arteries were examined using time-of-flight magnetic resonance angiography (n = 823) and/or computed tomography angiography (n = 183). The narrowest diameter of each stenosed vessel was measured and divided by the diameter of the normal vessel proximal to the lesion, or distal to the lesion if the proximal artery was diseased. Significant intracranial artery stenosis (ICAS) was defined as ≥ 50% stenosis or occlusion. ICAS was considered symptomatic if the stenosis was ipsilateral to the index stroke/TIA.
Extracranial carotid atherosclerosis was examined using ultrasonography (n = 813), and/or computed tomography angiography (n = 105), and/or time-of-flight magnetic resonance angiography (n = 78). We defined significant extracranial artery stenosis (ECAS) as the presence of atherosclerotic stenosis of ≥ 50% or occlusion according to the European Carotid Surgery Trial criteria. 16 ECAS was considered symptomatic if the stenosis was ipsilateral to the index stroke/TIA.
Aortic atherosclerosis was assessed using transesophageal echocardiography (n = 260). Mobile plaques were defined as mobile components seen swinging on their peduncles. An ulcerative plaque was diagnosed as a discrete indentation of the luminal surface of the plaque with a base width and maximum depth of at least 2 mm each. Complex aortic atheroma were defined as any plaque ≥ 4 mm in thickness or a plaque with ulceration or mobile components. 17
Cerebral small vessel disease was assessed using fluid-attenuated inversion recovery images acquired using a 1.5T magnetic resonance imaging scanner (Philips Ingenia 1.5T, Siemens Magnetom Avanto fit 1.5T) (n = 821). Degrees of periventricular hyperintensity and deep white matter hyperintensity were rated according to a published definition (0–4 for each hemisphere).18 The higher score between the right and left sides was used in the analysis.
Etiologic subtype
An etiologic subtype of the index event was assigned to each patient using the atherosclerosis, small vessel disease, cardiac pathology, other definite causes, and dissection (ASCOD) grading system. 19 The ASCOD system categorizes five predefined phenotypes. Each of the phenotypes is graded according to if the disease is: (1) likely causal; (2) potentially causal but uncertain; (3) present but is unlikely to be causal, and (0) absent. We defined stroke of atherothrombotic origin as ASCOD grade A1 or A2, which corresponds to the presence of symptomatic or potentially symptomatic ipsilateral ICAS or ECAS or complex aortic atheroma. Similarly, small vessel disease (lacunar stroke) and cardioembolic stroke were defined as ASCOD grades S1 or S2, and C1 or C2, respectively.
Follow-up and outcomes
Follow-up visits were scheduled at three months, one year, and three years after enrollment. The present study reports the one-year outcomes. At each follow-up visit, treatments, occurrence of clinical events, and modified Rankin Scale scores were recorded. If the patient could not be contacted for follow-up, a relative or caregiver was interviewed via telephone. The primary outcome was a composite of major adverse cardiovascular events (MACE), including nonfatal stroke (either ischemic or hemorrhagic), nonfatal acute coronary syndrome, major peripheral artery disease, and vascular death. Vascular death was defined as fatal acute coronary syndrome, fatal stroke, and other cardiovascular deaths including pulmonary embolism and sudden cardiac death. The secondary outcomes included stroke subtype and all-cause mortality.
Statistical analysis
Quantitative variables are expressed as mean (standard deviation) in cases of normal distribution, or median (interquartile range). Qualitative variables are expressed as frequencies (percentages). Comparisons were made between multiple groups using one-way analysis of variance or the Mann–Whitney U test for quantitative variables, the χ2 test for categorical variables, and the log-rank test for censored variables. Event rates were estimated using the Kaplan–Meier method. Cox proportional hazard regression models were used to evaluate the associations between the TyG index and risk of recurrent vascular events by calculating hazard ratios and 95% confidence intervals. All variables with P < 0.10 in univariate analysis (i.e., sex, body mass index, hypertension, dyslipidemia, diabetes mellitus, atrial fibrillation, current smoking, excessive alcohol, and etiologic subtype of index event) were included in the multivariable adjustments. Data for patients with no information at one year were censored at the time of the last available follow-up. For a given outcome, patients who died from causes other than the outcome were censored at the time of death. Events that occurred after the one-year follow-up period were not included in the current analysis. Sensitivity analyses were conducted to assess the effect of TyG on vascular risk according to the etiological subtype of the index event. The area under the receiver operating characteristic curve was used to determine the predictive value of the TyG index for MACE. For all analyses, statistical significance was set at P < 0.05.