Patients
We analyzed 161 patients with SLE who visited the Osaka Medical College Hospital and underwent a brain MRI between April 2012 and June 2018. The inclusion criteria were (1) age of >16 years, (2) SLE diagnosis according to the ACR 1982 revised criteria for SLE22, and (3) no obvious evidence of any other disease that could explain SLE-like symptoms. The exclusion criteria were (1) abnormalities on brain MRI, such as hydrocephalus, neoplasm, or metal deposition and (2) history of brain trauma or operation. Of the 161 patients, 142 were women (88.2%). Only female patients were included in this study to standardize patient background characteristics. Of the 142 patients, 110 had undergone a brain MRI to screen brain lesions, whereas 32 had undergone brain MRI to evaluate the cause of neuropsychiatric SLE. Organic brain lesions, including ischemic stroke, were identified in five patients. Seventy-nine patients had undergone brain MRI on more than one occasion (two times in 39 patients, three times in 11, five times in seven, six times in five, seven times in four, and >10 times in four). A follow-up brain MRI was conducted to evaluate the status of CVDs and other central nervous system lesions. Of the 142 patients, 27 had antiphospholipid syndrome23. Patients having either anti-cardiolipin antibodies of ≥10.0 U/mL, anti-cardiolipin-β2-glycoprotein complex antibodies of ≥3.5 U/mL, or lupus anticoagulant of ≥1.3 were considered aPL antibody positive. The control group consisted of 216 female patients with non-stroke-related diseases who underwent brain MRI at the Osaka Medical College Hospital and Kyoto City Hospital between January 2012 and October 2018. This included 27 patients with Parkinson syndrome, 23 patients with non-autoimmune epilepsy, 18 patients with migraine without stroke, 16 patients with peripheral neuropathy, 15 patients with non-stroke-related movement disorders, 12 patients with spinocerebellar degeneration, 5 patients with myasthenia gravis, 5 patients with aseptic meningitis, 3 patients with Guillain–Barré syndrome, 3 patients with myositis, 2 patients with amyotrophic lateral sclerosis (ALS), and 87 patients with neurological symptoms. ALS patients were confirmed to have no abnormal signal in the pyramidal tracts. We excluded brain MRI images of patients with multiple sclerosis, neuromyelitis optica spectrum disorders, and encephalitis because of the difficulty in discriminating abnormal signal intensities from ischemic changes on MRI. Clinical information was retrospectively retrieved from the medical records. This study was conducted according to the 2013 Helsinki Declaration; the study protocol was approved by the Osaka Medical College Ethics Committee. The requirement for informed consent was waived because of the study’s retrospective nature and the removal of all individual patient identifiers from the data (Approval number # 609).
Clinical findings and laboratory examinations
Baseline information, including age at initial brain MRI, sex, history of smoking, diabetes mellitus, hypertension, and dyslipidemia, were collected for all patients. Smoking status was evaluated through interviews, and the patients were classified as current, past, or never smokers. Hypertension was defined as a resting blood pressure of ≥140/90 mmHg or current use of antihypertensive agents24. Diabetes mellitus was judged according to the World Health Organization criteria25 or by current medication history. Dyslipidemia was defined as low-density lipoprotein cholesterol levels of ≥160 mg/dL, triglyceride levels of ≥150 mg/dL, or current medication history26.
In the SLE group, baseline information on age at diagnosis and the titers of ANA, anti-Sm antibodies, and aPL antibodies at the initial brain MRI were collected. Additionally, we collected the history of antiplatelet and anticoagulant drug treatment at the initial brain MRI. In all patients with SLE, administration of corticosteroids was started immediately after the diagnosis of SLE. The patients with SLE continuously received a combination therapy of corticosteroids and a variety of immunosuppressive drugs. The duration of corticosteroid treatment almost matched with disease duration. For the assessment of autoimmune abnormality in SLE at the initial brain MRI, SLEDAI was evaluated27. The titers of ANA and anti-double-stranded DNA antibodies were measured by ELISA. Anti-Sm antibodies ≥10.0 U/mL were considered positive.
Assessment of WMH on brain MRI
Diffusion-, T2-, FLAIR, and T2*-weighted sequences (slice thickness: 2 mm) were obtained using 3.0T MRI (GE Signa HDxt). For analyzing WMH, the grades of PVH and DWMH were evaluated by the Fazekas rating scale28. PVH was evaluated on FLAIR-weighted transverse images covering the anterior horn and body of the lateral ventricles. DWMH was evaluated on FLAIR-weighted transverse images covering the semioval center. The initial brain MRI was used as the baseline image. To assess temporal changes in WMH, we evaluated the deterioration of WMH grades using the brain MRI data from 68 patients aged <60 years with SLE who repeatedly underwent brain MRI.
Statistical analysis
Categorical and continuous demographic variables in the SLE and control groups were analyzed via Fisher’s exact test and Wilcoxon rank-sum test, respectively. The ratio of patients with WMH was analyzed via Fisher’s exact test, and the grades of WMH were analyzed via the chi-squared test. The age-adjusted ratios of WMH were conducted by the direct method to the control group, and the data were analyzed using the Cochran-Mantel–Haenszel procedure. We divided the SLE and control groups into three groups based on their age at the initial brain MRI and evaluated the differences between the corresponding sub-groups using Fisher’s exact test to evaluate the occurrence of WMH between different age groups. Univariate analysis was conducted to determine the risk factors for WMH in the SLE group using Fisher’s exact test; the results are shown as OR and 95% CI. Subsequently, variables associated with p values of <0.25 in the univariate analysis were included in the multivariate logistic regression model to identify factors associated with WMH. We analyzed the progression of WMH grades using Kaplan–Meier curves with the log-rank test. Values are expressed as the median and IQR; p values of <0.05 were considered indicative of statistical significance. Statistical analysis was conducted using JMP Pro 14 software (SAS Institute Inc., Cary, NC, USA).