Simplified Scoring System Differentiates CNS Infections From Neuropsychiatric SLE

A simplified scoring system may help physicians in differentiating diagnoses of life-threatening central nervous system (CNS) infections from neuropsychiatric systemic lupus erythematosus (NPSLE) in patients with SLE, according to study results published in Arthritis Research & Therapy.

Researchers conducted a cohort study of 8491 patients with SLE in a tertiary referral center in China to identify the clinical features and etiology of CNS infections in patients with SLE and to assess the diagnostic features differentiating CNS infections from NPSLE.

Central nervous system infections were identified in 1.12% of patients (n=95; 81 women) from the total cohort; mean age at SLE diagnosis was 31.0±13.9 years, and mean age of CNS infection onset was 34.6±13.7 years. In terms of prior treatment, 36.8% of patients in the CNS infection group vs 7.4% of patients in the NPSLE group had been treated with pulse glucocorticoids (average daily dose in the prior 6 months, 43.5±44.2 mg vs 21.8±37.5 mg, respectively; P <.001). The 2 groups also differed on rates of having received ≥1 disease-modifying antirheumatic drug in the prior 6 months (70.5% vs 37.9%, respectively, P <.001).

The most common clinical manifestations of CNS infections were fever, headache, and acute confusional state; all clinical manifestations occurred more frequently in the CNS infections group than in the NPSLE group.

At the inpatient treatment center, the prevalence of CNS infections with SLE was 1.4%±1.4% within a 5-year interval frame; the rate of NPSLE was 17.6%±4.6%. Prevalence of CNS infections did not drastically change whereas that of NPSLE increased from 13.1% to 21.4% during this time interval.

Central nervous system infections were categorized into 3 subgroups, according to etiologic findings: bacterial (n=27), cryptococcal (n=18), and mycobacterial (n=11). Compared with patients with cryptococcal infections, those with bacterial or mycobacterial infections were more likely to manifest consciousness disturbance or signs of meningeal irritation. Nearly half of all deaths (13 of 30 patients) occurred in the bacterial infection subgroup, which had a 37.0% first-year mortality rate compared with 16.7% and 27.3% in the cryptococcal and mycobacterial subgroups, respectively.

In the multivariate logistic regression analysis, compared with patients in the NPSLE group, those in the CNS infections group had a longer disease duration, more frequent pyrexia, and polymorphonuclear leukocytosis in the cerebrospinal fluid (CSF) with significantly decreased CSF glucose.

“Together, these results may be used as clues to distinguish CNS infections from NPSLE,” the researchers wrote.

Researchers conducted a univariate analysis and multivariate logistic stepwise regression and established several benchmarks as vital risk factors for discriminating CNS infections from NPSLE: longer disease duration, fever, CSF, polymorphonuclear ratio, significantly decreased CSF glucose, and absence of hypocomplementemia. In addition, CSF examinations, intracranial pressure, white blood cell counts, and protein levels were included for their clinical significance in the study. These 8 items were used to create a simplified scoring system (SSS-8) to assist in the quick recognition of CNS infections in patients with SLE.

A total of 75 cases of patients from the CNS infections and NPSLE groups were combined into 1 group. Researchers applied the SSS-8 scoring system to a verification group with the remaining 20 patients and identified a cutoff value of 4, sensitivity of 85%, specificity of 85%, and area under the curve of 0.93 (95% CI, 0.86-1.00); CNS infections were discriminated from NPSLE using this method.

“The proposed… SSS-8 [scoring system]… may help clinicians promptly, and more adequately, distinguish CNS infections from NPSLE,” the researchers concluded.

Reference

Jiang M, Shi X, Gao X, et al. Clinical features of central nervous system infections and experience in differential diagnosis from neuropsychiatric lupus erythematosus in a cohort of 8491 patients with systemic lupus erythematosus. Arthritis Res Ther. 2019;21(1):189.