Subjects and methods

Thirteen patients (four men, nine women) with clinical definite multiple sclerosis3 were recruited during an acute relapse. A relapse was defined as the development of new neurological symptoms or abrupt deterioration of existing symptoms lasting more than 24 hours within the past six weeks and confirmed on neurological examination. Patients were excluded if their corrected visual acuity was<6/12 or if they were unable to use a computer touch screen accurately. They were examined at the time of recruitment and about six weeks later after treatment with intravenous methylprednisolone. All patients had a neurological examination and physical disability was assessed using the Kurtzke expanded disability status scale (EDSS).4 Four patients had relapsing-remitting disease and nine had secondary progressive disease. Ten healthy controls (three men, seven women) selected to match the patients for age and estimated premorbid IQ, were also examined on two occasions. Informed consent was obtained from all subjects.

The hospital anxiety and depression questionnaire5 was administered to all subjects. Scores>10 on either the anxiety or depression subscale were indicative of “caseness”. The national adult reading test (NART) was used to estimate premorbid IQ.6 A battery of neuropsychological tests to assess attention and memory was administered on both occasions and included the following:

Paced auditory serial addition test (PASAT)7

Thirty one randomised digits were read out at two second intervals. The subject was instructed to add each digit to the one preceeding it. The score was the number of correct responses.

Stroop8

Names of colours, each printed in a different coloured ink, were presented on a computer screen. The score was the time taken for the subject to name the colours in which the words were printed.

Symbol digit substitution9

The subject was required to name numbers represented by nine different symbols according to a code shown on the computer screen. Eight trials were presented. The mean time taken per trial was recorded.

Digit span forwards10

The score was the longest sequence of digits that the subject could recall.

Story recall test11

The subject was required to recall a story read aloud to him or her immediately after and following an interval of 20 minutes.

Spatial span

This test was taken from the Cambridge neuropsychological test automated battery (CANTAB).12 It determined the longest sequence of squares lighting up on the computer screen that the subject could recall accurately.

Other memory tests from the CANTAB (spatial working memory, spatial recognition and pattern recognition tests) were also administered but no significant differences in performance between patients and controls were found.

Brain MRI was performed in the patients with multiple sclerosis using an NMR 1.5 Tesla GE Signa System on both occasions. Gadolinium-diethylenetriaminepentaacetic acid (Gd-DTPA; 0.1 mmole/kg) was administered intravenously and T1 weighted images with a TR of 540 and TE of 11 were obtained. Thirty six contiguous slices (3 mm thickness) were selected for measurement of lesion volume. A neuroradiologist (GQ) delineated the Gd enhanced lesion areas on hard copies. With reference to this, one rater (JF) used a semi-automated contouring technique13 to mark the lesions on the images displayed on a SUNSPARC station. A software lesion volume measurement programme was utilised to compute the total cerebral Gd enhanced lesion load.

STATISTICAL ANALYSIS

Mann-Whitney and Wilcoxon matched pairs signed ranks tests were used to examine between and within group differences. Correlation analysis was performed using Pearson’s correlation coefficient.

Results

There were no significant differences between patient and control groups for age (mean 37.15 years and 38.5 years respectively) and premorbid IQ (mean IQ 109.31 and 115.0 respectively). Patients were recruited within a mean of 3.42 weeks (range 1–6 weeks) of the onset of relapse. The mean interval between the two testings was not significantly different between patients and controls (47.54 days and 51.50 days respectively).

Physical disability as assessed on the EDSS scale, had improved significantly in patients at follow up (z=−2.67, p=0.008). EDSS scores improved in nine patients and remained unchanged in four. There were no significant differences in the improvement of physical disability between patients with relapsing-remitting disease and those with secondary progressive disease. Patients with multiple sclerosis scored significantly higher than controls only on the depression subscale (p=0.021) on initial testing although scores did not reach “caseness”. At follow up, the HAD depression (z=−2.27, p=0.023) and anxiety scores (z=−2.34, p=0.019) had significantly improved for the multiple sclerosis group but were unchanged in the control group.

Patients were divided into three groups according to the pattern of Gd enhancement: group A: (n=4 (three secondary progressive disease, one relapsing-remitting disease)), no Gd enhancing lesions in the brain detected on either testing; group B (n=3 (all secondary progressive disease)), an increase in Gd enhancing lesion load found on the second testing; group C (n=6 (three relapsing-remitting disease, three secondary progressive disease)), a reduction in Gd enhanced lesion load found on the second testing. Patients with relapsing-remitting disease had a greater reduction in Gd enhanced lesion load than patients with secondary progressive disease (table 1).

Patients performed significantly worse than controls on initial testing in all tests of attention (table 2). At follow up, there was a significant improvement on the PASAT (z=−2.23, p=0.026) and a trend of improvement on the symbol digit substitution test (z=−1.84, p=0.065) in the multiple sclerosis group. Table 1 shows the individual scores of all patients.

The patient subgroups (A, B, C) did not differ significantly in their performance on the tests of attention. Although individual variation in performance was evident, there seemed to be a trend for greater improvement in these tasks in group C patients. The patients with relapsing-remitting disease in this group were less impaired than the patients with secondary progressive disease on initial testing and greater improvement on the PASAT was found at follow up. By contrast, there was little change in the group performance on these tasks for the patients who had no detectable Gd enhancing lesions and a slight deterioration in performance for the patients who had an increase in Gd enhanced lesion load (table 1). The control group’s performance on these tasks was unchanged at follow up.

There were significant group differences between patients and controls for the recall tasks (digit span, spatial span, immediate and delayed story recall) on initial testing. Group differences persisted for the immediate and delayed story recall tests at follow up (table 2). Performance did not change significantly for any group in these tests at follow up.

Table 1 shows the individual patient’s Gd enhanced lesion load. As we were interested in examining the relation between improvement in neuropsychological performance and Gd enhanced lesion load, only the six patients from group C were selected for the correlation analysis. We found that the improvement in the Stroop and symbol digit substitution test scores correlated significantly with reduction in Gd enhanced lesion load (r =−0.99, p<0.01 and r=−0.82, p<0.05 respectively).

Discussion

The results of this study highlight the variability of cognitive manifestations and lesion enhancement on MRI during acute relapses of multiple sclerosis. Some patients in our sample failed to show Gd enhancement on both occasions and there was little change in their cognitive performance. For the rest of the patients, variation in the Gd enhancing lesions was to some extent in parallel with the cognitive changes, particularly for the patients who had a reduction in Gd lesion load. Cognitive improvement seemed to be limited to attentional tasks whereas memory impairment remained unchanged. Psychiatric symptoms were unlikely to have contributed to the patients’ neuropsychological performance in view of the HAD scores failing to reach “caseness” and the lack of correlation with the neuropsychological test scores.

In group C, it was apparent that the patients with relapsing-remitting disease had a greater improvement in Gd enhanced lesion load than the patients with secondary progressive disease, in whom it was in parallel with their improvement on the attentional tasks. The patients with relapsing-remitting disease were also less cognitively impaired than those with secondary progressive disease at the outset, suggesting that those who are least cognitively impaired are most likely to improve after a relapse. This would also be in keeping with previous reports that cognitive deterioration is related to a progressive course.14 A somewhat unexpected finding is the divergent natural history of attention and memory deficits. Immediate and delayed recall deficits in the patients with multiple sclerosis did not change after remission. It is possible that the more entrenched memory deficits may be caused by different neuropathological processes such as axonal and myelin loss whereas attention deficits may be a result of transient inflammatory changes.The discrepancy between MRI cerebral lesion load and clinical relapse is illustrated by our finding that several patients had no enhancing lesions. In addition, the group B patients were found to have an increase in enhancing lesion load at follow up in the absence of any neurological deterioration. This may be because physical disability as measured by the EDSS is more closely related to spinal cord than cerebral lesions. It is also likely that Gd enhancement of acute lesions does not reflect the degree of demyelination or axonal loss, which are the likely substrates of functional impairment.

We acknowledge that this study was limited by the small sample size. None the less, our findings suggest that it may be possible to predict which cognitive deficits may be reversible and which patients are more likely to experience improvement during remission. This information would be relevant in selecting patients for clinical trials and interpreting the results of treatment.