Background/Purpose:  
There
is an unmet need for a screening test of cognitive function that can be
administered in clinic in patients with SLE. We aimed to determine the: 1) validity
of Montreal Cognitive Assessment (MoCA) and Mini-Mental State Examination
(MMSE) as screening tests of Cognitive Impairment (CI) in SLE and 2) associated
factors with CI.

Methods:  
Consecutive
patients followed at a single centre and seen since Feb 2014 were included. Screening
tests were administered by 2 trained assessors: Hopkins Verbal Learning
Test-Revised (HVLT-R), Controlled Oral Word Association Test (COWAT) via
telephone interview and MoCA and MMSE via face-to-face assessment.

Patients
completed Centre of Epidemiologic Studies Depression Scale (CES-D), Beck
Anxiety Inventory (BAI), and Reynolds Intellectual Screening Test (RIST), and
scores were compared in patients with and without CI.

Sensitivity/Specificity,
Positive Predictive Value (PPV)/Negative Predictive Value (NPV) and Likelihood
Ratio (LR) of MoCA and MMSE in detecting CI (HVLT-R external construct) were
determined. Pearson correlation of MoCA and MMSE with HVLT-R were studied. Regression
analyses were performed to test for possible associations with CI.

Results:  
Of
86
patients, 45%
had CI using MoCA, 40% using HVLT-R, 16% using COWAT and 14% using MMSE.
Sensitivity was higher for MoCA (65%) compared to MMSE (21%), though MMSE was
more specific (90%) than MoCA (67%). PPV and LR were similar in MoCA and MMSE (PPV:
56% & 58%; LR: 2.0 & 2.1, respectively), but NPV was higher in MoCA (75%)
than MMSE (64%). HVLT-R correlated with MoCA (r=0.43, p<0.0001). Attention,
delayed recall and language domains of MoCA were lower than normal (Figure 1).

Univariate
analyses:
CI patients had higher CES-D
scores (ie. more depressive symptoms) than those without CI (22.1±13.7 vs. 15.9±10.6,
p=0.02). This did not hold in multivariate analysis. There were no
significant differences in BAI scores, diabetes, cardiovascular events,
dyslipidemia, hypertension, smoking status or antiphospholipid antibodies in
patients with and without CI.

Multivariate
analyses:
Patients with CI had shorter SLE disease duration than patients without CI (OR=0.94;
CI 95% 0.90-0.99; p=0.01). Each one year of SLE follow up reduces the
probability of CI by 6%. More education years and high RIST score were
protective against CI (OR=0.81; CI 95% 0.66-0.98; p=0.03 and OR=0.93; CI 95%
0.88-0.98; p=0.006, respectively). Each one-year increase in education
decreases the chance of developing CI by 19%. Each 1 score increase in RIST
decreases the chance of developing CI by 7.5%.

Conclusion:
CI
among SLE patients was highly prevalent (45%) using MoCA. Considerations of cost/administrative
burden and appropriate psychometric properties make MoCA the preferential
screening test for CI in SLE. Shorter SLE disease duration increases risk while
education and high intelligence scores protect against CI.