Results

There were 250 (155 women) consecutive stroke patients admitted over a 12 month period; median age 76 (range 39–95) years. Two hundred and five patients had had a hemispheric stroke (45% in the right hemisphere). On admission 57 (27.8%) had an abnormal score on the motor subscale of the GCS and were unable to be assessed fully.

Sixty six patients had a first right hemispheric stroke and normal conscious level. Of these, 27 (40.9%) had neglect as defined by a total BIT score<129.

All 27 patients (13 women) were seen within a week of their stroke (median three days). Median age was 73 (range 57–85) years.

Twenty three (85.2%) of these neglect patients had a hemiparesis; three (11.1%) patients monoparesis, and one (3.7%) patient a visual field deficit and neglect without paresis. Seven (25.9%) patients had no evidence of any visual field deficit. The median Barthel score was 7; range 2–20 for these neglect patients.

Brain CT was carried out on 16 of the 27 patients. One patient had a normal scan, 13 had infarcts, and two patients had evidence of an intracerebral haemorrhage.

For the 27 patients with abnormal BIT scores on admission, there was a wide variation in the total BIT scores (mean BIT 56.3, range 10-126). During the follow up period, there was a progressive improvement in the mean total BIT scores (SD), one month 96.5 (38.3); two months 110 (36.7); three months 121.3 (28.6).

The three cancellation tests showed errors across both the left and right hemispace. There was a progressive improvement in visuospatial neglect across the left to right hemispace (table). During the three month follow up period, recovery occurred across the right and left hemispaces (figure).

• Download figure
• Open in new tab
• Download powerpoint

Letter cancellation. Recovery in the first three months across visual field sectors.

On admission the 27 neglect patients had a mean Barthel score of 6.9 (SD 3.5). There was no significant correlation between Barthel and BIT score on admission. However, at one, two, and three months there was a significant correlation (Spearman correlation r=0.642, p=0.001; r=0.623, p=0.003; r=0.636, p=0.003; respectively) between recovery from neglect and recovery in function (Barthel score).

At the end of the study period, six patients still had evidence of neglect. Seven (25.9%) patients had been discharged home, nine (27%) were still in hospital undergoing rehabilitation, three (11.1%) were in long term care, and eight (29.6%) had died. On admission the best test of visuospatial neglect predicting discharge was the line cancellation test; no other factor seemed to predict recovery. High scores on this test were significantly correlated with discharge home (Spearman’s rank correlation r=0.4217, p=0.028).

Discussion

Visuospatial neglect is an important adverse prognostic factor for rehabilitation after stroke.2-4 Our study found a lower prevalence of neglect (40.9%) than other studies.13 We excluded patients with an abnormal conscious level one week after stroke, and a significant proportion of these patients may have had unrecognised visuospatial neglect. It is well established however, that persisting coma is of importance for early mortality,14and such patients cannot participate in detailed neuropsychological evaluations.

During the study there was a progressive improvement in the cohort’s mean total BIT score. Recovery seemed to take place throughout the three month period but was greatest in the first month.

All patients received standard physiotherapy and occupational therapy regimens on the wards. It is unclear whether the amount of therapy given to patients influenced recovery. However, it is postulated that each impairment should have a specific remedy—that is, attention training for unilateral neglect.15

Not all the patients with visuospatial neglect underwent CT. Studies have suggested that awareness should be thought of in terms of complex corticosubcortical neural circuits. This is supported by imaging studies showing functional derangement in cerebral regions far removed from, but connected with, the structurally damaged areas.16

The three cancellation tests (line, star, and letter) showed that errors are made in both the right and left hemispaces. The SDs were large and the group’s mean results showed a continuous increment in accuracy from the far left column to the far right column. Patients were inevitably free to move their eyes; however, the test sheet was body centred and our results are consistent with previous findings.17 18

On admission, star and line cancellation tests had a correct score of over 85% in the far right hemispace. By contrast, the letter cancellation test score was 59.6% for the far right sector. A patient’s ability to detect stimuli is affected by the number of distracting stimuli and the difficulty of visual discrimination.19 20 The letter cancellation test has a greater number of distracting stimuli and requires controlled information processing which is more demanding of attention.21

In our study the application of these simple tests seems to predict outcome. A high score on the line cancellation test on admission was associated with recovery of visuospatial neglect and discharge home. By contrast poor scores on a test with no distracting stimuli (line cancellation) suggest a more severe degree of neglect and therefore a worse prognosis.

During the follow up period there was recovery across both the right and left hemispaces. Whereas this recovery was greatest in the right hemispace, errors were still being made in these sectors. Visuospatial neglect is postulated to be a lateralised attentional problem rather than a sensory disorder.22 23 Robertson, however, has previously found that patients with neglect have a general difficulty in deploying attention in space.7 Alternatively, there may be a lateral gradient of attention across both hemispaces. Therefore attention is biased rightward, regardless of the absolute location of the target.24

On admission, there was no relation between severity of neglect and functional impairment. This reflects the many factors that may influence activities after stroke. Therefore there was no relation between function and BIT score.

Subsequently, a significant correlation between improvement in neglect and function occurred, illustrating that recovery from visuospatial neglect and recovery of daily living activities are likely to be closely linked or associated. The continuing presence of visuospatial neglect may adversely affect the ability of the stroke patient to adapt to their neurological impairments.

In conclusion, acute stroke patients who present with visuospatial neglect have a high probability of improving. The continuing presence of visuospatial neglect may be an adverse factor for rehabilitation and specific measures may be required to aid their rehabilitation.