Motor impairment in PD

Methods.

Results.

Of 173 patients, 50 became demented during a mean follow-up period of 3.6 ± 2.2 years. Baseline characteristics of the cohort are summarized in table 2. Patients that subsequently became demented were older, less educated, and had more severe motor signs at baseline than those who did not become demented. No significant differences were seen in gender, ethnicity, duration of PD, language in which the neuropsychological tests were administered (English or Spanish), total Hamilton Depression Rating Scale22 score, use of dopaminergic and anticholinergic medications, or l-dopa dosage. The patients that were excluded from the analysis due to missing data (n = 7) were not significantly different from the other patients in terms of gender, age, ethnicity, education, duration of PD, or language in which the neuropsychological tests were administered.

The total UPDRS motor score (range, 0 to 100) was associated with incident dementia in the Cox model (relative risk [RR] = 1.06; 95% CI, 1.04 to 1.09; p < 0.0001), independent of demographic and medication variables. This risk ratio means that a one-point difference in total UPDRS motor scores between two subjects with similar demographic profiles would be associated with a 6% increase in risk of dementia over the 3.6 year follow-up period. Both subscore A (RR = 1.07; 95% CI, 1.03 to 1.10; p = 0.0003) and subscore B (RR = 1.23; 95% CI, 1.14 to 1.32; p < 0.0001) were associated with incident dementia when entered separately into the model. When both were entered into the model, subscore B but not subscore A was significantly associated with incident dementia. Male gender, age at baseline, and education were also significantly associated with incident dementia in this model (table 3). The analysis was repeated using l-dopa dosage rather than use of dopaminergic medications as a covariate. Nineteen additional cases were lost due to missing information about l-dopa dosage; subscore B (RR = 1.17; 95% CI, 1.06 to 1.29; p = 0.002) was still associated with incident dementia but subscore A (RR = 1.04; 95% CI, 1.00 to 1.08; p = 0.07) was not. In a separate analysis including the six domains of the UPDRS, speech and axial impairment (subscore B) as well as bradykinesia (subscore A) were retained in the final model, although axial impairment approached significance (table 4).

Among the six UPDRS motor domains, axial impairment (Spearman’s rank correlation coefficient = 0.29; p < 0.001) (figure) and speech (Spearman’s rank correlation coefficient = −0.15; p = 0.03) were correlated with age. The motor domains, with the exception of tremor and rigidity, were correlated with duration of PD (p < 0.001). When the analysis of the correlation between age and the motor domains was repeated controlling for duration of PD and l-dopa dosage, the positive correlation between axial impairment and age was still significant (correlation coefficient = 0.32; p < 0.001), but the negative correlation between speech and age was not significant anymore (correlation coefficient = −0.13; p = 0.09). Each of the four tests of the UPDRS that were included in the axial impairment domain (rising from a chair, posture, gait, and postural stability) were also significantly correlated with age.

• Download figure
• Open in new tab
• Download powerpoint

Figure. Scatterplot and fit line of axial impairment versus age.

Discussion.

Our findings suggest that PD motor impairment mediated predominantly by nondopaminergic systems is associated with incident dementia. When the UPDRS total motor score was divided in two subscores according to relative responsiveness to dopaminergic therapy, motor impairment partially responsive or nonresponsive to l-dopa (subscore B) predicted the development of dementia in PD, whereas in the same model, motor impairment responsive to l-dopa throughout the course of the disease (subscore A) did not. Among six domains of extrapyramidal motor impairment in PD, speech and bradykinesia predicted the development of incident dementia, and axial impairment approached significance in the Cox model.

Among the motor domains believed to respond to l-dopa throughout the course of the disease, bradykinesia was the only one to be significantly associated with incident dementia. Indeed, bradykinesia has been associated with cognitive impairment in PD in many cross-sectional1–5 and longitudinal12,13 studies. Other evidence, however, suggests a stronger relationship between non–l-dopa-responsive motor manifestations and cognitive impairment in PD. In one study,23 none of the neuropsychological test scores was correlated with the l-dopa-responsive motor score (the difference between the motor score after withdrawal of l-dopa and the motor score on l-dopa treatment), but all were highly correlated with the motor score on l-dopa treatment (motor impairment not alleviated by l-dopa and presumed to be the result of nondopaminergic lesions). Moreover, cognitive impairment was poorly correlated with akinesia and rigidity, strongly correlated with gait disorder and dysarthria, and not correlated with tremor at all. In an analysis of the DATATOP cohort,24 a postural instability and gait disorder subtype of PD was associated with a more rapid disease progression and greater subjective intellectual impairment as compared with a tremor-dominant subtype of PD.

Among the six motor domains, only axial impairment was significantly correlated with age in our study. The relationship between gait and posture impairment and age has been demonstrated in other studies,4,25 and patients with young-onset PD are less likely to complain of difficulty walking as an early symptom compared with patients with late-onset PD.26 Thus, age may act as a confounder and at least partly explain the strong correlation of gait and postural disturbances with cognitive impairment in univariate analyses. In one previous analysis,13 rigidity, posture, gait, “bradykinesia of hands,” “upper extremity swing,” and “hoarseness and weakness of speech” were significantly associated with dementia in the univariate analysis. However, only bradykinesia and speech impairment were significant predictors of dementia in a logistic regression model controlling for age. In our analysis, the axial impairment contribution to the Cox model controlling for age and other covariates approached significance. Extrapyramidal manifestations resembling parkinsonian features have long been recognized in normal aging, including “an attitude of general flexion,” rigidity, poverty of movement, bradykinesia, and gait disorders.27 However, the unique correlation between axial impairment and age in the setting of PD suggests a combined effect of the disease and the aging process.

The strengths of this study include the large cohort of community-dwelling patients with PD followed for a relatively long period with neurologic and neuropsychological evaluations, and its distinct methodologic approach. Nevertheless, the implications of our findings for the contribution of dopaminergic and nondopaminergic mechanisms to the development of dementia in PD rest upon the assumption that nondopaminergic lesions play a predominant role in speech and axial impairment in the middle and late stages of PD, and underlie the declining l-dopa responsiveness over the course of the disease.14,16 Although in a recent study bradykinesia was the clinical sign of PD with the highest correlation with the degree of nigrostriatal dopaminergic deficit measured by fluorodopa PET,28 pathologic, pharmacologic, and experimental animal data suggest that dopamine is also related to rigidity and rest tremor.29,30 Evidence from human parkinsonism due to 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP) also indicates that dopamine deficiency may underlie the three cardinal signs of PD.31 Other weaknesses of this study include the lack of autopsy data allowing for clinical-pathologic correlation, and the fact that our cohort consisted of patients with a mean disease duration at baseline of 6.3 years. Therefore, our results may not be applicable to patients with early PD.

The role of nondopaminergic structures in the cognitive manifestations of PD is supported by pathologic studies demonstrating greater involvement of the basal forebrain32,33 and locus ceruleus34,35 in patients with PD and dementia compared with those with PD without dementia. Loss of neurons in the nucleus basalis of Meynert has been observed in PD independently of Alzheimer-type cortical pathology.36,37 Nondopaminergic structures have also been implicated in PD motor manifestations. Cell loss in the pedunculopontine nucleus containing cholinergic neurons (Ch5 cholinergic cell group38) has been shown in PD and, to a greater extent, in progressive supranuclear palsy.39–41 That the pedunculopontine nucleus may play a role in the parkinsonian gait disorder is supported by animal experimental studies showing its relationship to a mesencephalic locomotor region,42,43 and by its connections to the thalamus, basal ganglia, subthalamic nucleus, and substantia nigra, as well as to other sites in the brainstem and spinal cord.44,45 The locus ceruleus (noradrenergic), in turn, has been implicated in the motor phenomenon of freezing46,47 and, through a facilitatory influence on nigrostriatal and mesolimbic dopaminergic pathways, in locomotion behavior based on animal studies.48,49

We believe that patients with PD with a greater risk of developing dementia may exhibit a more widespread neurotransmitter deficiency resulting from subcortical pathology prior to the onset of dementia. The brainstem nondopaminergic structures (locus ceruleus and pedunculopontine nucleus) may be involved in both motor impairment (subscore B) and the development of dementia in PD, or degenerate in parallel with the cholinergic system of the basal forebrain. Moreover, axial motor impairment in PD may be the result of a combined effect of the disease and the aging process in brainstem nondopaminergic structures. To further clarify these issues, clinical-pathologic correlation exploring cognitive function and specific motor manifestations of PD and dopaminergic as well as nondopaminergic anatomic structures is needed.