FOR PATIENTS IN the remitted phase of systemic lupus erythematosus (SLE), it is difficult to judge whether the psychotic symptom is primary or secondary to neuropsychiatric SLE (NPSLE). We report a woman with SLE showing psychosis, which was thought to be primary, but which was eventually found to be NPSLE. The patient has given the authors written informed consent to publish this letter.

A 52-year-old woman was diagnosed with SLE at the age of 42 with initial presentations of arthritis, malar rash, photosensitivity and oral ulcer. She was treated with hydroxychloroquine 200 mg/day and had remission of symptoms. She had had one major depressive episode at the age of 43 and a manic episode at 51, which had subsided under psychotropic treatment within 1 month. She was admitted to our psychiatric ward due to persecutory delusion, auditory and visual hallucination, and labile mood for 1 week.

As the patient had normal serological data, no systemic lupus signs, and no abnormality on brain computed tomography (CT), both the rheumatologist and psychiatrist favored the diagnosis of bipolar I disorder with psychotic features. To control her psychotic symptoms, we first prescribed aripiprazole 15 mg/day. However, her visual and tactile hallucinations were aggravated. She often witnessed and felt water dripping on her face from the dry ceiling. Two weeks later, we switched to amisulpride 800 mg/day but there was still no improvement. Meanwhile, another major depressive episode developed, which did not respond to lamotrigine 150 mg/day and bupropion 300 mg/day augmentation. The rheumatologist still favored primary psychosis because the autoimmune profiles (anticardiolipin antibodies [IgG and IgM], C3, C4 complements, antinuclear antibody, anti-Ro, anti-La, D-D dimmer) were still within normal limits. Nevertheless, her erythrocyte sedimentation rate (ESR) was 62 mm/h. On the third rheumatologic consultation, brain magnetic resonance imaging (MRI) revealed a 0.8-cm old infarction in the left thalamus and old leukoaraiosis in periventricular white matter. 99mTc-ECD brain single-photon emission computed tomography (SPECT) showed bilateral ventricle dilatation, and moderate to severe hypoperfusion in bilateral frontal and parietal cerebrum. The diagnosis was modified to NPSLE. On the 90th day of admission, she was transferred to the rheumatologic ward for immunosuppressant therapy. After a 3-day course of cyclophosphamide pulse therapy, she was discharged with dramatic improvement.

Although her lupus activity seemed in the remitted phase, NPSLE was most likely due to persistently vivid tactile and visual hallucination, the changeable and treatment-refractory course of psychosis and depression, abnormally high level of ESR as well as brain hypoperfusion on SPECT. The dramatic improvement under a 3-day course of cyclophosphamide pulse therapy supported our rationale. The moderate-to-severe hypoperfusion in bilateral frontal and parietal lobes could explain her psychotic and depressive symptoms. Patchy hypoperfusion in the parietal lobes and frontal lobes occurs in up to 80% and 65%, respectively, in SPECT in patients with NPSLE.1 The infarct and periventricular white matter changes are also common findings.1 Though her presentations did not fulfill the usual diagnostic criteria of SLE, the evidences strongly suggested NPSLE.

Although there is still no single neurodiagnostic tool proven to be definitive in diagnosing NPSLE,1 we suggest that, for a patient in the remitted phase of SLE showing psychosis, the brain MRI and SPECT could be beneficial in making the differential diagnosis and further management.


  1. Top of page