Patient population
This was a prospective study of 191 SLE patients with or without NP syndromes who were admitted to the Aoyama Hospital of The Tokyo Women's Medical University from August 1994 through October 2003. All these patients had 4 or more revised ACR (formerly, the American Rheumatism Association) criteria for SLE [10]. At our institution, those patients suspected of or newly diagnosed as having SLE were typically admitted for systemic evaluation regardless of the severity of the disease, and were eligible for inclusion in the study. Previously diagnosed SLE patients whose disease (NP syndrome or other manifestations of SLE) flared were also enrolled in the study. A total of 269 patients with SLE gave informed consent for inclusion in this study, including their MRI examinations during the above period. Among these 269 patients, those who had non-SLE-related NP manifestations arising from infection, uremia, electrolyte imbalance, hypoxia, brain tumor, trauma, primary mental disease or drug use (n = 45) or past histories of NP involvement (n = 33) were excluded prior to MRI scans. The patients were excluded because we wanted to compare recently (i.e. within a month) diagnosed active CNS lupus patients to non-NPSLE patients and because the unrelated conditions could affect current symptoms or laboratory and MRI findings. At the time of admission to the hospital, each patient completed a standardized medical history, including medication use, and had physical examinations that included neurologic and rheumatologic examinations. Psychiatric examinations were employed when indicated. Serology profiling for each patient was performed using standard immunoassays. The activity of SLE was measured using the SLE Activity Index (SLEDAI) [11]. Treatment with corticosteroids or immunosuppressive drugs was instituted as indicated following completion of these evaluations. One non-NPSLE patient at admission, who later developed NPSLE, was reevaluated at the onset of the NP syndrome (a mood disorder in this patient) and was reclassified into the CNS group. In this patient, data from the reevaluation, including the results of MRI, were used. Subjects were classified into the CNS group or the non-CNS group according to the presence or absence of active CNS syndromes. The CNS group was then further classified into the neurologic disorders group consisting of patients with neurologic disorders with or without other NP syndromes, or the psychiatric disorders group comprising patients with psychiatric disorders with or without other NP syndromes [2, 4, 8, 12]. Thirteen patients were classified into both neurologic and psychiatric disorders groups. Detailed diagnostic criteria for these groups are described below. The study was approved by the Ethical Committee of our institution and the Helsinki declaration was followed throughout the study.
Diagnosis of CNS lupus
Although ACR nomenclature and case definitions include 12 CNS syndromes and 7 peripheral nervous system syndromes [2, 4, 8, 12], we included only the 12 CNS syndromes in the present study because of the substantial differences in anatomy, function, and clinical characteristics between the central and peripheral nervous systems. Slight or mild cognitive dysfunction without significant clinical impairment, as revealed by the detailed neuropsychological tests described below in "Evaluation of psychological impairment," was excluded from the CNS syndromes in our study. Using published normative data, patients in this category (n = 3) exhibited less than the median of normal controls by 2 standard deviation (SD) in 2 or more of areas of cognitive function without significant clinical impairment [13]. Tension headache, or episodic tension type headache, was also excluded from our study.
The final clinical diagnosis and classification of the various NP syndromes for inclusion in the study were made by an experienced rheumatologist (M. Hara) and psychiatrist (K. N.), according to the standardized ACR nomenclature and case definitions for neuropsychiatric lupus syndromes [2]. These decisions were based on the medical history and neuropsychological examinations by rheumatologists, an experienced neurologist (S. U.) and a psychiatrist (K. N.) and supported by conventional laboratory tests and appropriate complementary tests, including MRI, electroencephalography, and cerebral spinal fluid tests, and the clinical course. Cases before the ACR nomenclature and case definitions for NPSLE were published in 1999 were originally diagnosed and classified according to another criteria and later re-diagnosed retrospectively according to the ACR criteria.
Clinical improvement of CNS lupus was defined as either sustained complete recovery or recovery with minor residual deficits that no longer required hospitalization. Stabilization was defined as the status in which no new clinical (i.e., neurologic or psychiatric) abnormalities occurred, although the previous abnormalities remained. Deterioration was defined as the status in which previous neuropsychiatric symptoms were exacerbated or new ones developed during follow-up [14, 15].
Evaluation of psychological impairment
Cognitive function was evaluated using the Mini Mental Status Examination [16] and the ACR-suggested test battery [2], including the WAIS-R/Digit Span (Forward) [17], Trail Making Test (Part B) [18], WAIS-R/Digit Span (Backward) [17], Wisconsin Card Sorting Test [18], Rey Auditory-Verbal Learning Test [17], WAIS-R/Block Design [17], Animal Naming Test [18], WAIS-R/Similarities [17], Trail Making Test (Part A) [18], and WAIS-R/Digit Symbol Substitution Test [17]. Mood and behavioral dysfunction was assessed by clinical observation, patient history and standardized instruments, such as the Profile of Mood States [19]. However, these formal neuropsychological tests were not performed routinely in all patients and the diagnosis of cognitive, mood, or behavioral dysfunction was on the basis of clinical assessment using the ACR definitions [2], rather than formal neuropsychological testing, especially in cases where disturbance of consciousness, such as acute confusional state, created difficulty in taking the tests. The impact of disturbance on daily life and prior occupational and social functioning was determined from information provided by the patient or other informants.
Magnetic resonance imaging (MRI)
MRI was performed using a 1.5T MR scanner (Toshiba Medical Systems, Otawara, Tochigi, Japan) within a week from admission and within a month from onset of NP symptoms in the CNS group. Scans were aligned parallel to the axial plane through the anterior to posterior commissure and covered the entire brain in all sequences. T1-weighted, T2-weighted, and fluid-attenuated inversion recovery (FLAIR) images were acquired from all patients who had brain MRIs. The software, but not the hardware of the MR scanner was updated once in 2000, which improved its resolution but did not practically influence the sensitivity of this study. The MRIs were interpreted at the time of scanning by an experienced radiologist (T. Y.) who was not blinded to all the clinical and sequential information. MRI tests were defined as 'positive' when any abnormal intensity lesion was found. The abnormal MRI signals were classified into large abnormal MRI signals (ø ≥ 10 mm) or small abnormal MRI signals (ø < 10 mm) [15]. Mild brain atrophy, characterized by loss of brain volume, was not included among MRI abnormalities in the present study because it is not a well-established abnormality [6], even though it is the most frequent abnormal findings in SLE [9]. Moderate to severe brain atrophy was not seen at the time of enrolment in the study population. CNS lupus patients who had had abnormalities in initial MRIs were reevaluated by MRI at approximately 1 year or when clinical amelioration or deterioration was determined. MRI improvement was defined as a more than 50% decrease in the number or size of abnormal MRI signals. MRI stabilization was defined as occurring when no new abnormalities were detected, but previously detected abnormalities were unchanged or only slightly changed. Deterioration was defined as when the previously detected abnormalities became exacerbated or new ones developed during follow-up [15].
Statistical analyses
Diagnostic tests evaluated in the present study were correlated with the final clinical diagnosis for each. Two-group comparisons were analyzed using the Mann-Whitney U test for continuous variables and Fisher's exact test for categorical variables. Values of p < 0.05 were considered statistically significant. Accuracy, positive predictive value (PPV), and negative predictive value (NPV) of MRI and the relative risk of CNS lupus to MRI were also calculated. All statistical analyses were performed using SPSS statistical software (version 14.0J; SPSS Inc., Tokyo, Japan).