Standard Protocol Approvals, Registrations, and Patient Consents.
This study was approved by Ethics Committee of the Hospital District of Southwest Finland. Written informed consent was obtained from all cases in the study. Informed consent was not required for controls, because controls were included from retrospective register. All methods were performed in accordance with STROBE guidelines and Declaration of Helsinki.
KD patients who were diagnosed and treated in the catchment area of the Turku University Hospital (population of 887 000 citizens) from 1978 to 1995 were identified retrospectively by using diagnostic codes (International Classification Code-9, 446.1; International Classification Code-10, M30.3). Diagnosis were confirmed from the patient records for each patient according to American Heart Association (AHA) 2004 diagnostic criteria for complete KD.19 Patients with a current age of ≥ 25 years and a history of KD occurring in the childhood were included in this study. Age criteria was based on protocol for this cohort described in the earlier study.20 Patients with current age < 25 years, Marfans syndrome, Ehler-Danlos syndrome type IV, polycystic kidney disease or history of intracranial aneurysms or bleeding were excluded. Patients with a positive family history of intracranial aneurysms were also excluded.
There were 87 KD patients diagnosed between 1978 and 1995. From 87 KD patients, 27 were excluded because of age < 25 years. Based on a review of patient records, none of the 87 KD patients had been diagnosed with ischemic or hemorrhagic stroke before beginning the study enrollment year 2016. An invitation letter was sent to 60 patients who met the inclusion criteria and 40 of them were willing to participate in the study, and 20 refused. Prior to brain magnetic resonance imaging (MRI), patients were interviewed for past medical history (hypertension, diabetes mellitus, hypertension, migraine, hyperlipidemia, depression, history of stroke, neurological symptoms), medication, smoking, alcohol consumption and possible signs of heart or lung problems.
Of 40 cases, 37 had accurate information on which drug KD had been treated with and 37 patients had accurate information about echocardiographic data during the acute phase of KD. From 37 patients 22 were treated with intravenous immunoglobulin and 15 patients with aspirin only.
This was a population-based study, since all the patients were collected from our catchment area.
All patients who had undergone brain MRI for any reason (n= 39 993) between 2003 and 2020 in our tertiary hospital were reviewed to include migraine patients (controls). Of these patients, 1062 had migraine diagnosis (International Classification Code-10, G43.0-G43.3) in patient records. Of 1062 migraine patients, 68 were excluded because of intracranial tumor, history of acute brain infarction, sinus thrombosis or subarachnoid- /intracerebral hemorrhage. None of the controls had as history of KD, other vasculitis or brain diseases. From 994 migraine patients, controls were matched (4 controls to 1 case) randomly by age (±2 years at the time of the brain MRI) and sex with cases. Patient records were reviewed for hypertension, hypercholesterolemia, type 1- and type 2-diabetes, smoking and migraine subtype. Smoking was categorized as never smoker vs. current- or ex-smoker.
Brain MRI Data Acquisition
For cases, MRI scans were conducted on a Philips Ingenia 3T scanner (Philips Medical Systems, Best, the Netherlands). Axial 3D T2-weighted sequence with TR (Repetition Time) of 2500 ms (milliseconds), TE (Time Echo) of 250 ms, matrix of 352 x 352 and slice thickness of 1 mm (millimeters) was obtained. We also obtained a coronal 2D FLAIR (Fluid Attenuation Inversion Recovery) sequence with TR of 4800, TI (Time Inversion) of 1650 ms, TE of 285 ms, matrix of 352 x 352 and slice thickness of 3 mm, sagittal 3DT1 sequence with TR of 81 ms, TE of 3.7 ms, matrix of 320 x 320 and slice thickness of 1 mm was obtained as well as susceptibility-weighted sequence with TR of 20 ms, TE of 27 ms, matrix of 512 x 512 and slice thickness of 2 mm. These sequences were obtained to find and exclude any brain pathology. MR angiography using a axial Time-Of-Flight (TOF) sequence with TR of 23 ms, TE of 3.5 ms, matrix of 640 x 640 and slice thickness of 1.2 mm was performed to evaluate the arteries of the brains. TOF images were interpreted as such and also 3D reconstructions in two different planes were built and interpreted.
For controls, MRI scans were conducted with any available 1.5-3T scanners at our catchment area with a routine MRI protocol that includes the following sequences; T1- and T2-weighted sequences, susceptibility-weighted sequences, and FLAIR sequences. MRI scanner type and field strength for each control is presented in Supplemental Table S1.
Brain Imaging and Analysis
For controls and cases, two neuroradiologist (each with more than 10 years of experience in neuroradiology), blinded to case-control status and clinical data, evaluated independently the number, the location and the size of WMHs from the fluid-attenuated inversion recovery (FLAIR)-sequences and T2-weighted images. In addition, microbleeds and lacunes of presumed vascular origin were evaluated in the same blinded fashion. Conflicting interpretations between the two radiologists were resolved by consensus of the two interpreters.
Lesions ≥2 mm were categorized as WMH. Modified Scheltens’ visual rating scale was used to evaluate WMH burden,21, 22, 23 because WMHs located in the basal ganglia or brainstem were excluded according to neuroimaging strandards for WMHs24. Modified Scheltens’ visual rating scale provides scoring system for periventricular WMH (0-9 points) and deep white matter hyperintesities (0-24 points) (Supplemental Table S2).22, 23 WMH located less than 10 mm from the ventricles was categorized as periventricular WMH.25 Subcortical WMHs was categorized as deep WMHs.
Lacune of presumed vascular origin was categorized as round or ovoid, subcortical, fluid-filled cavity of between 3mm and 15 mm in diameter from T1-weighted, T2-weighted and FLAIR sequenses.24 Location of lacune was defined by vascular territory.24 Microbleeds were evaluated from susceptibility sequences and differiantiated from sponatenous intracerebral hemorrhages with T1-weighted and T2-weighted or FLAIR sequences.24
All analyses were performed using SPSS Statistics 27 (IBM Corp., Armonk, NY, USA).
Mean ages within cases and between cases and controls were compared with two-sample t-test. Percentage distribution of total, deep, and periventricular Scheltens’ score were compared between cases and controls by using Chi-square test and Fisher’s exact test to evaluate total, deep and periventricular WMH burden. Chi-square and Fisher’s exact test were also used to test the association of categorical variables with Scheltens’ score. Scheltens’ score was dichotomized to 0 vs ≥1 to compare prevalence and risk factors for total WMHs, deep WMHs and periventricular WMHs within cases and controls, and between cases and controls with chi-square and Fisher’s exact test. In those with positive WMH findings (Scheltens’ score ≥1), Scheltens’ score values were compared between cases and controls by using Mann-Whitney U-test. P-values less than 0.05 were considered as statistically significant. Missing data for each variable were excluded from the analyses.
Cohen’s kappa (k) analysis was used to evaluate inter-observer agreement for the WMH prevalence at the first evaluation round. Kappa value between 0.00-0.20 was defined as slight agreement, 0.21-0.40 fair agreement, 0.41-0.60 moderate agreement, 0.61-0.80 substantial agreement and 0.81-1.00 almost perfect agreement.26