Cognitive impairment is common in multiple sclerosis (MS), yet difficult to detect during routine neurologic examination. Therefore, brief screening tests that identify patients who may benefit from a more thorough assessment or treatment are needed. We investigated the utility of the Symbol Digit Modalities Test (SDMT) as a screen for cognitive dysfunction because it can be administered and scored in about 5 minutes. One hundred MS patients and 50 healthy controls, matched on demographic variables, participated in the study. Examination procedures included the neuropsychological (NP) tests included in the Minimal Assessment of Cognitive Function in MS (MACFIMS) battery. Patients were considered impaired if they performed one and a half standard deviations below controls on two or more MACFIMS variables, excluding the SDMT. Bayesian statistics showed that a total score of 55 or lower on the SDMT accurately categorized 72% of patients, yielding sensitivity of 0.82, specificity of 0.60, positive predictive value (PPV) of 0.71, and negative predictive value (NPV) of 0.73. These results suggest that the effectiveness of the SDMT as a screen for cognitive impairment in MS is roughly equal to that of other psychometric and questionnaire methods.

The Minimal Assessment of Cognitive Function in Multiple Sclerosis (MACFIMS) is a consensus neuropsychological battery with established reliability and validity. One of the difficulties in implementing the MACFIMS in clinical settings is the reliance on manualized norms from disparate sources. In this study, we derived regression-based norms for the MACFIMS, using a unique data set to control for standard demographic variables (i.e., age, age2, sex, education). Multiple sclerosis (MS) patients (n = 395) and healthy volunteers (n = 100) did not differ in age, level of education, sex, or race. Multiple regression analyses were conducted on the performance of the healthy adults, and the resulting models were used to predict MS performance on the MACFIMS battery. This regression-based approach identified higher rates of impairment than manualized norms for many of the MACFIMS measures. These findings suggest that there are advantages to developing new norms from a single sample using the regression-based approach. We conclude that the regression-based norms presented here provide a valid alternative to identifying cognitive impairment as measured by the MACFIMS.

The relationship between perceived cognitive functioning and objective cognitive functioning was studied in 221 patients with multiple sclerosis. Perceptions of global cognitive functioning as well as perceptions of performance on specific cognitive tests were assessed. Patients' perceptions of global cognitive functioning in their daily lives were unrelated to their objective performance on the full cognitive test battery. However, patients' perceptions of their performance on specific tasks correlated with their objective performance on those tasks, even though they underestimated their performance on these tasks. The present study also evaluated predictors of patients' perceived cognitive functioning. Depression, anxiety, fatigue, and level of disability predicted perceptions of global cognitive functioning, whereas objective cognitive performance did not. These results add to our understanding of patients' expressed concerns regarding their cognitive functioning in the wake of multiple sclerosis, suggesting that such concerns should be interpreted with caution by clinicians.

Based on the assumption that cognitive impairment in MS is consistent with subcortical dementia, a battery of neuropsychological tests was assembled that included measures of executive function (Tower of London and Wisconsin Card Sorting Test), verbal learning and memory (a paired associates learning test), and speeded information processing (Stroop Color Word Interference Test). The battery was administered to patients with relapsing and primary progressive MS and to healthy controls. Differences between patients and controls occurred on several of the measures. However, when differences with respect to fatigue and depression were statistically controlled, the only differences that remained significant involved measures relating to the speed of information processing. Patients performed more slowly than controls, with the disparity being greater for relapsing patients than for those with primary progressive disease. The slowing was evident on measures of automatic as well as controlled processing and regardless of whether speed was an explicit feature of successful performance or recorded unobstrusively while the patient concentrated on planning a correct solution to a problem. Parallels were noted between cognitive slowing associated with MS and that of normal aging. (JINS, 2004, 10, 948–956.)

Cognitive impairment is more closely associated with physical disability than most previous studies indicate. This relationship appears to be stable throughout the duration of MS, although this conclusion is qualified by the cross-sectional design of the study. Further attention should be paid to cognitive impairment as a possible predictor of the rate of patients' physical decline.