Date: Monday, November 9, 2015
Session Type: ACR Concurrent Abstract Session
Session Time: 2:30PM-4:00PM
Background/Purpose: Neurocognitive impairment (NCI) is an important morbidity in childhood-onset systemic lupus erythematosus (cSLE); however, the gold standard formal neuropsychological assessment is difficult to access. Screening for NCI using computerized testing with the Pediatric Automated Neuropsychological Assessment Metrics (PedANAM) is a more feasible option. Four recently derived cognitive performance scores (CPS) use PedANAM raw data and differing statistical methods (weighted average, unweighted average, logistic regression and principal components analysis) to establish suggested cutoff values for patients requiring further evaluation. Our objectives were to examine external, concurrent, criterion and diagnostic validity of the PedANAM-CPS scores using a single-centre multiethnic cSLE cohort.
Methods: Patients were recruited within 18 months of cSLE onset, and all had a formal neuropsychological assessment battery and PedANAM testing on the same day. Validation of the PedANAM-CPS scores utilized sensitivity and specificity analyses in addition to other nonparametric statistical comparisons.
Results: 29 cSLE patients without premorbid NCI completed the study procedures. Median age at testing was 15.2 years, 83% were female. Six (21%) patients had NCI identified by formal testing. All four CPS scores were significantly different in patients with NCI, and predetermined cutoffs were reached only in patients with NCI (Table 1). Two of the 4 CPS scores (one using an unweighted average and the other a logistic regression model) had 100% sensitivity to detect NCI, while the CPS based on principal components analysis had the highest specificity (87%) for detecting NCI (Table 2). All four CPS scores were significantly correlated to 3 of the cognitive domains (visuoconstructive abilities, working memory and psychomotor speed), but not to attention and executive functioning as evaluated by formal neuropsychological testing.
Table 2: Sensitivity and Specificity of PedANAM-CPS for NCI
Normal Cognitive Function
Patients with NCI
|PedANAM-CPSmultiscore||3.79 (1.05 – 13.75)||7.37 (1.86 – 15.58)||0.04|
|PedANAM-CPS UWA*||0.45 (-2.47 – 0.86)||-0.75 (-1.20 – 0.04)||0.002|
|PedANAM-CPS Logit||-2.80 (-4.78 – 6.41)||-0.43 (-1.46 – 0.86)||0.003|
|PedANAM-CPS PCA*||1.17 (-8.14 – 2.45)||-2.33 (-3.11 – 0.53)||
|Sensitivity (%, 95% CI)||Specificity (%, 95% CI)|
|PedANAM-CPS multiscore||83.3 (43.7 – 97.0)||73.9 (53.5 – 87.5)|
|PedANAM-CPS UWA*||100 (61.0 – 100)||73.9 (53.5 – 87.5)|
|PedANAM-CPS Logit||100 (61.0 – 100)||56.5 (36.8 – 74.4)|
|PedANAM-CPS PCA*||83.3 (43.7 – 97)||87.0 (67.9 – 95.5)|
*UWA = unweighted average; PCA = principal components analysis
Conclusion: The 4 PedANAM-CPS scores have high sensitivity and specificity for NCI in cSLE patients in our multiethnic cohort. The CPS scores are easily calculated by any clinician following administration of the PedANAM test battery, and can be used to discriminate cSLE patients who require referral for a formal neuropsychological assessment.
To cite this abstract in AMA style:Nguyen J, Williams T, Silverman E, Levy DM. Cross-Validation of the Pediatric Automated Neuropsychological Assessment Metrics-Cognitive Performance Scores in the Screening of Neurocognitive Impairment in Childhood-Onset Systemic Lupus Erythematosus [abstract]. Arthritis Rheumatol. 2015; 67 (suppl 10). https://acrabstracts.org/abstract/cross-validation-of-the-pediatric-automated-neuropsychological-assessment-metrics-cognitive-performance-scores-in-the-screening-of-neurocognitive-impairment-in-childhood-onset-systemic-lupus-eryt/. Accessed October 16, 2021.
« Back to 2015 ACR/ARHP Annual Meeting
ACR Meeting Abstracts - https://acrabstracts.org/abstract/cross-validation-of-the-pediatric-automated-neuropsychological-assessment-metrics-cognitive-performance-scores-in-the-screening-of-neurocognitive-impairment-in-childhood-onset-systemic-lupus-eryt/