Whether patients with Parkinson's disease develop cognitive impairments or improvements after ventral pallidotomy is still a debated issue.1 Recent studies produced contradictory findings which may have resulted from methodological factors such as differences in surgical techniques, neuropsychological assessments, duration of follow up, and the lack of evaluations of non-operated controls with Parkinson's disease.1

We assessed a consecutive series of 27 patients with Parkinson's disease who received unilateral pallidotomy using the microelectrode registration technique.2 Sixteen of these patients received a 3–6 month follow up evaluation, and 10 of them received a 12 month follow up evaluation. They were compared with a non-operated control group of 20 patients with Parkinson's disease matched for age, severity of extrapyramidal symptoms, and overall cognitive status who received the same neuropsychological evaluation at baseline and 12 months later. The neuropsychological examination included the Raven's progressive matrices, the Wisconsin card sorting test (WCST), the controlled oral word association test, the Buschke selective reminding test, the Benton visual retention test, the digit span, and the Perdue pegboard.

No significant differences between the pallidotomy and the control groups were found for age (years (SD) pallidotomy group 56.3 (6.9), control group 59.3 (7.9)), sex (pallidotomy group 50% women, control group: 50% women), years of education (years (SD) pallidotomy group10.7 (2.7), control group: 11.4 (4.1)), baseline levodopa equivalent dosage, and UPDRS total scores (table 1). All patients were right handed.

Sixteen patients with Parkinson's disease who underwent unilateral pallidotomy received a 3–6 month follow up. A repeated measures multivariate analysis of variance (MANOVA) for the neuropsychological variables comparing baseline versus 3–6 month follow up evaluation showed no significant overall time effect (F(7,56)=1.01; NS). There was a significant time effect for the Perdue pegboard test (F(1,14)=30.9; p<0.0001), with a significant improvement in manipulative dexterity over time. A repeated measures MANOVA for the neuropsychological variables comparing patients with either a left (n=7) or right (n=9) pallidotomy showed no significant group effect (F(1,7)=0.05; NS), time effect (F(1,7)=1.03; NS), or group×time interaction (F(8,56)=0.22; NS).

A repeated measures MANOVA for the neuropsychological variables for the 10 patients who had undergone pallidotomy (six right, four left) with a 12 month follow up and the 20 non-operated patients with Parkinson's disease did not show a significant group effect (F(1,23)=0.29; NS), time effect (F(1,23)=0.43; NS), or group× time interaction (F(7,161)=0.18; NS). On the other hand, there was a significant group×time interaction for the Perdue pegboard test (F(1,28)=8.84; p<0.01): the pallidotomy group showed a significant improvement during the follow up period, whereas the control group had a slight decline.

Most studies on the cognitive sequelae of pallidotomy could not show significant neuropsychological deficits after surgery,3and the only studies that to our knowledge included a non-operated Parkinson's disease control group (Perrine et al 4 and the present one) confirmed this finding. On the other hand, Lang et al 5reported some cognitive impairments after ventral pallidotomy; and differences in neuropsychological outcome measures may account for this discrepancy. We examined the neuropsychological sequelae of pallidotomy in a consecutive series of 16 patients with Parkinson's disease, 10 of whom had a 1 year follow up evaluation. When compared with a group of 20 patients with Parkinson's disease matched for MMSE scores and age who did not receive a pallidotomy, no significant between group differences were found in the rate of cognitive changes. On the other hand, the pallidotomy group showed a significant improvement on a task measuring manual dexterity compared with the control Parkinson's disease group. The question now arises as to why pallidotomy in Parkinson's disease does not produce significant cognitive deficits, given that some case reports described various intellectual problems after spontaneous pallidal lesions. Firstly, most lesion studies included patients with bilateral lesions, whereas pallidotomy is usually performed on one side only. The few reports of bilateral pallidotomy in Parkinson's disease described important cognitive sequelae in some of the patients. Secondly, pallidotomy usually produces a small and localised lesion, whereas spontaneous pallidal lesions are usually larger and often involve white matter tracts next to the pallidum. Finally, some of our pallidotomy patients were tested three or four times, compared with only two neuropsychological evaluations for the control group, which may have produced some learning effects.