new research suggests.
In a paper published in Arthritis Care & Research, researchers assessed the validity of the Automated Neuropsychological Assessment Metrics (ANAM) test in 211 adult patients with systemic lupus erythematosus (SLE).
First author Oshrat E. Tayer-Shifman, MD, of the University of Toronto Lupus Clinic and coauthors wrote that current assessment of cognitive impairment in adults with SLE is done using the American College of Rheumatology neuropsychological battery (ACR-NB). However, this approach involves protected tests that require specialized personnel and takes around 1 hour to administer, as well as time for scoring and interpretation.
“For many clinics, these are notable barriers to accessing CI [cognitive impairment] assessment, as health care payers do not cover these costs,” the investigators wrote. And although briefer cognitive screening tools, such as the Montreal Cognitive Assessment, the Controlled Oral Word Association Test, and the Hopkins Verbal Learning Test–Revised, have been examined in studies of patients with SLE, “they also require specialized personnel for administration and interpretation and cannot be self-administered. In addition, their validity for the screening for CI in SLE has not been well established. Thus, there is an unmet need for a screening assessment for CI that is validated for SLE and that can be applied in an ambulatory clinic setting without specialized personnel.”
The full ANAM battery requires about 40 minutes and has been used to screen cognitive performance in a range of clinical contexts.
A total of 96 patients (46%) had CI and 52 (25%) did not, according to the ACR-NB, while the results were indeterminate in the remaining 63 (30%).
The study showed that patients without CI performed significantly better on the majority of the ANAM subtests in comparison with patients who have cognitive impairment. This was particularly evident on mean reaction time and the number of correct responses per minute (a measure of cognitive efficiency), but less so for percentage of correct responses and consistency of response speed.
“Three of the most affected cognitive domains in the CI patients in this cohort, as well as in previous studies, were learning and memory, visual spatial construction, and simple attention and speed of processing,” the researchers wrote.
The investigators created testing models using the subtests that were most discriminative for CI. The two best models included one encompassing the percentage of correct responses, consistency of response speed, and mean reaction time, as well as one encompassing these three factors as well as the number of correct responses per minute. The investigators then derived candidate ANAM composite indices from these two models. For one composite index that used 8 of the 15 ANAM subtests and included five of the six cognitive domains tested on the ACR-NB, a high and a low cutoff value gave an area under the curve of 79%, sensitivity of 80%-89%, specificity of 54%-70%, positive predictive value of 78%-83%, and negative predictive value of 65%-74%.
The composite index performed similarly well among patients with or without neuropsychiatric lupus.
“This approach not only enables us to use a cost-effective screening approach without specialized personnel, but we have reduced the duration of the ANAM battery itself. The ANAM [version 4 General Neuropsychological Screening] full battery requires approximately 40 minutes to administer, while our analyses enable us to limit the number of ANAM subtests used, shortening the testing duration to 20 minutes,” they wrote.
The investigators noted that the study included only individuals with sufficient English language ability to complete the tests, and they also excluded patients with indeterminate cognitive status. “We reasoned that they represented a nonhomogeneous group, and without a clear consensus on the definition of CI in SLE patients, we chose to concentrate on the more clearly defined CI SLE patients.”
The study was supported by grants from the Arthritis Society of Canada, Physician’s Services, the Kathi and Peter Kaiser Family, and the Lou and Marissa Rocca Family. One author was supported by the Arthritis Society and the Canadian Rheumatology Association. No conflicts of interest were declared.
SOURCE: Tayer-Shifman OE et al. Arthritis Care Res. 2019 Oct 18. doi: 10.1002/acr.24096.