Additional related information may be found at:
Neuropsychopharmacology: The Fifth Generation of Progress

Back to Psychopharmacology - The Fourth Generation of Progress

Parkinson's Disease

Amos D. Korczyn


Parkinson's disease (PD) is one of the common chronic diseases of old age. It is a prototypical disease in the sense that the understanding of its pathophysiology and treatment have advanced hand in hand at a very impressive rate during the past two or three decades. In this review, these developments will be discussed.



PD is primarily a disease of the motor system. It has a gradual onset, slowly progressing to eventual severe disability. The motor symptoms include tremor at rest, poverty or slowness of movement, rigidity, and loss of postural reflexes. None of these four primary manifestations are specific to PD, and therefore the clinical diagnosis can only be tentative. The slow evolution and the lack of other features (e.g., pyramidal, sensory, or marked autonomic disturbances) support the diagnosis although some vegetative symptoms, and particularly constipation, are common (26). As will be discussed below, the clinical diagnosis is also supported by a positive response to levodopa.

Several other brain diseases can mimic PD. The assumption that vascular brain disease can result in similar manifestations was favored several years ago, leading to the clinical designation of arteriosclerotic parkinsonism. This nosologic entity has been disfavored but has recently re-emerged. Its eventual destiny is unclear. Historically, the encephalitis pandemic of the 1920s resulted in a multitude of cases with postencephalitic parkinsonism. Six decades after the disappearance of new cases of lethargic encephalitis, patients with recent-onset PD are highly unlikely to be due to encephalitis.

Other disorders with extrapyramidal features resembling PD include progressive supranuclear palsy, olivopontocerebellar atrophy, and Shy-Drager syndrome, all of which can frequently be identified by specific clinical features. However, several reports indicate that the accuracy of the clinical diagnosis is limited and that as many as one-quarter or one-third of patients will be found at autopsy to have alternative diagnoses (24).

In addition to the motor abnormalities, patients with PD frequently have cognitive and affective disturbances. Depression is common in PD and in many patients predates the extrapyramidal features. The nosologic entity of the association of the motor and affective features is still unclear, but for reasons discussed below it is quite likely that depression should be regarded as one of the features of PD rather than one of its complications.

Dementia also commonly occurs in PD patients; prevalence data suggest that about 50% of PD cases have significant cognitive impairment. This too seems to be an integral part of the spectrum of clinical manifestations of PD.



The pathological hallmark of PD is intracellular inclusions called Lewy bodies. These occur inside neurons in the substantia nigra, presumably dopamine (DA)-containing cells. These inclusions probably accumulate in neurons undergoing degeneration. The number of DA neurons progressively diminishes in PD. It is important to note that only DA neurons in the substantia nigra whose axons are destined to go to the putamen (less so to the caudate) in the nigrostriatal tract are affected. Chemical analysis shows progressive loss of DA in the striatum; clinical symptoms first appear when DA content in the striatum is reduced by about 70%. This may imply that a long preclinical stage, of 20 years or more, occurs (19, 38, 46).

Other neurotransmitter systems are also affected in PD. These include norepinephrine (NE) loss in the cell bodies of the locus coeruleus, serotonin [5-hydroxytryptamine (5-HT)] loss in the raphe nuclei, and cholinergic cell loss in the nucleus basalis of Meynert. These deficiencies probably contribute to the affective and cognitive changes in PD but may also be involved in motor dysfunction. However, it is clear that the motor disturbances are primarily related to DA deficits, because replacement of endogenous DA can miraculously alleviate the motor disability.

Until recently, it was impossible to demonstrate the DA deficiencies during life. However, this was changed by the use of positron emission tomography (PET). Using radioactive tracer techniques it can be demonstrated that DA markers accumulation is reduced in the corpus striatum in PD, presumably because of the loss of DA terminals which take up these markers. At the present state, the PET is not sensitive enough to detect the progression of the disease (5), but easily correlates with the side of the clinical abnormalities in unilateral parkinsonism.



The pathophysiology of PD is still rather clouded (33). The dopaminergic denervation of the basal ganglia (and particularly the striatum) is obviously central in the basic movement abnormalities. The existence of jmotor loopsd involving the basal ganglia, subthalamic nucleus, thalamus and cortex was recently established. However, their function is poorly understood (2, 4). How the loss of dopamine causes tremor at rest, enhanced tone both at rest and during action, and bradykinesia or hypokinesia still needs to be explained. The pathophysiology of the fourth cardinal feature of PD, loss of postural reflexes, is totally unclear. Some gains were made through the use of kinematic studies, such as of arm trajectories. These quantify the defects and demonstrate some unexpected findings (e.g., regarding the importance of visual feedback), demonstrating how the jmotor loopsd incorporate sensory information (12).

The nigrostriatal pathway activates, in the corpus striatum, dopaminergic receptors. Four subtypes have been identified and while the most important seem to be of the D2 type, the role of the others, and particularly D1, in normal activity is unclear and it is therfore not completely established what is the functional correlate of activation of these receptors. In particular it is not clear whether activation of D1 receptors is important for the elicitation of dyskinesias or other motor complications occurring in advanced cases who are treated by levodopa.


There is no consensus on the pathogenesis of the disease. The fact that only a selected population of neurons die off may suggest the involvement of a toxin affecting DA cells. Drugs are known which can selectively damage catecholaminergic neurons—for example, 6-hydroxydopamine (6-OHDA). This substance is uptaken by the DA transporter (vide infra), and it is concentrated in DA cells and causes their degeneration. Because 6-OHDA does not cross the blood-brain barrier, it cannot account for human PD. But recently another chemical, MPTP, was identified as causing in humans a disorder quite similar to PD in most characteristics. The mechanisms of MPTP toxicity have been explored in depth and although there is no doubt that this chemical only accounts for very few cases of PD, the possible existence of MPTP-ioids has been explored (50). While epidemiological and toxicological studies have generally failed to support the theory of an environmental toxin, the possibility of endogenous production of a substance similar to MPTP in its mechanism of action is still actively explored (48).

If such endogenous substances account for PD, occurrence of the disease in only a minority of elderly humans may suggest a genetic abnormality responsible for the production of this MPTP-ioid. Preliminary data have strongly suggested that there is no genetic contribution, but there has recently been a reappraisal of these conclusions. Some families have been described in which genetic transmission may well account for the high frequency. These families constitute a small minority, but it is possible that in others, multigenic transmission (or perhaps mitochondrial mutations), may contribute to the occurrence of PD.

Excessive concentrations of excitatory amino acids, particularly glutamate, may be involved in causing irreversible neuronal damage (47). This is particularly relevant for PD because of the existence of strong glutamatergic innervation of the corpus striatum. The neurotoxicity is produced by activation of N-methyl-D-aspartate (NMDA) receptors, and competitive antagonists can limit this damage. Although the assumption that the development of PD is preceded, or accompanied, by excessive excitatory amino acid bombardment is presently speculative, preliminary data suggest that NMDA antagonists, like memantine, may slow the process (40).

It is quite likely that several factors may contribute to the occurrence of PD, and an important question focuses on whether some or all of them operate through one common pathway that leads to DA cell loss. Along these lines, several authors have suggested that a selective increase exists in lipid peroxidation in the substantia nigra in PD, which may lead to excessive production of free radicals; these may in turn result in cellular damage and death (3). Of particular relevance is that DA degradation, like that of 6-OHDA, may cause free radical formation. As will be discussed later, these hypotheses have important immediate therapeutic implications.

The substantia nigra and globus pallidum are rich in iron, and the iron concentration increases with age and particularly in PD. This may suggest involvement of this metal in the process of lipid peroxidation (3). ( Excitatory Amino Acid Neurotransmission, Development of Mesencephallic Dopamine Neurons in the Nonhuman Primate: Relationship to Survival and Growth Following Neural Transplantation, and Dopamine Receptors: Clinical Correlates, for related topics.)



The basic treatment of PD is by DA replacement using levodopa. Levodopa is absorbed from the gastrointestinal tract and transported through the blood-brain barrier by active amino acid transport mechanisms. In the brain, as well as in the periphery, levodopa is metabolized to DA by an enzyme, 1-amino acid decarboxylase. This enzyme can be blocked by the substances benserazide and carbidopa. Employing either of these inhibitors can prevent the peripheral conversion of levodopa to DA. Most patients today are treated by a combination of levodopa and one of those enzyme inhibitors. The aim of using this fixed dose drug combination is to prevent the peripheral conversion to DA, because otherwise DA may act in the periphery to produce undesirable side effects such as orthostatic hypotension and nausea (6, 7).

Levodopa replacement is extremely effective in controlling much of the disability in PD. It is most efficacious against rigidity and hypokinesia, but tremor also responds. However, postural instability does not respond well to dopaminergic therapy. Because the progressive loss of DA neurons continues despite levodopa treatment, patients gradually require higher doses of the drug. These increments may cause significant problems, particularly peak-dose dyskinesias. Basically these reactions are to be expected because when brain DA concentrations are very high, the patient is in a state opposite to the baseline DA deficiency. The requirement to increase the dose will be obvious to the patient who becomes less and less mobile as time elapses since the last dose was ingested, manifested as end-of-dose hypokinesia.

Treating this stage is usually done by dividing the dose into several daily administrations. While initially three daily doses are sufficient, as the disease progresses six or more doses may be required. Additional mechanisms may be relevant. In normal subjects, levodopa will not produce dyskinesias (at least by doses used in PD). Presumably this is because presynaptic terminals in the corpus striatum take up any excessive DA and either store or degrade it to inactive metabolities. This mechanism will necessarily fail in PD because of the progressive loss of DA neurons and terminals (25).

The loss of this buffering capacity may be responsible also for the eventual and most problematic complication of therapy, the so-called "on-off" phenomenon. Patients fluctuate from being normal in their function, or even dyskinetic as a manifestation of excessive DA stimulation ("on"), to severe parkinsonian hypokinesia and rigidity ("off"). These fluctuations may initially be regular but later come on unexpectedly ("random on-off"). In addition to the loss of buffering capacity, pharmacokinetic factors (e.g., changes of levodopa serum concentration) may contribute to the occurrence of this disabling state (27, 35, 36). These fluctuations could be due to erratic absorption (possibly related partly to competition by amino acids derived from dietary proteins), distribution factors, or transportation across the blood-brain barrier. Attempts to reduce such fluctuations, which are of some benefit, include low-protein diet (23), gastric administration of levodopa at a constant rate (45), or by duodenal infusion (44), addition of vitamin C to facilitate absorption, controlled-release levodopa preparations (30), and administration of direct-acting water-soluble DA agonists (e.g., lisuride and apomorphine) subcutaneously, rectally, or intranasally (14, 20, 29, 53).

The revolutionary introduction of levodopa into the therapeutics of PD was so dramatic that its impact is unlikely to be superseded by another drug any time soon (56). However, as has been discussed above, this treatment does not solve all the problems (49).

One critical question relates to the time at which levodopa therapy should be initiated. The basic aim of levodopa therapy is to replace endogenous DA; it is thus a symptomatic therapy which, however, also masks to some extent the relentless progression of neuron cell loss. However, it is still unclear whether levodopa treatment itself accelerates or retards this loss. There are views suggesting that levodopa reduces the oxidative stress which results from the excessive burden on the remaining neurons. Alternatively, it is possible that the pharmacological concentrations of extrinsic levodopa will contribute to the formation of toxic free radicals inside neurons. Therefore, diverging views exist on whether levodopa should be started immediately upon diagnosing PD, or delayed as much as possible, with the aid of nondopaminergic therapy (9).

Monoamine oxidase (MAO), the enzyme which metabolizes several catecholamine and indole amines, exists in two forms. MAO-A metabolizes not only dopamine but also NE and 5-HT, whereas MAO-B does not metabolize either NE or 5HT. Selective inhibitors of MAO-B, and particularly deprenyl, are effective against MPTP toxicity. In PD patients, deprenyl provides symptomatic benefit (10, 18). This may be related to an amphetamine-like action in releasing DA from terminals or preventing DA reuptake or metabolism. Interestingly, a large study has shown that newly diagnosed PD patients can be maintained on deprenyl alone for a long period (51, 52).

There is significant controversy about the interpretation of this study, specifically whether the benefit results from a transient symptomatic effect of deprenyl or whether the drug actually slows down the progress of the disease (28, 37, 43). At present, many patients are being treated with deprenyl in addition to levodopa. The usefulness of this combination in retarding the progression of the disease has not been convincingly demonstrated.

Direct-acting DA agonists are important in the treatment of PD. These include apomorphine (14, 20, 29), bromocriptine, pergolide, and lisuride, but newer agents such as ropinirole and cabergoline are being introduced (39, 41). Theoretically, the use of such agents could be advantageous. In initial stages, they relieve the excessive burden of remaining neurons without being subject to metabolism into toxic free radicals inside DA neurons as has been hypothesized for levodopa (42, 54). In later stages, it is easier to maintain constant levels at receptor sites because these drugs do not depend on active transport in the gut and through blood-brain barrier. Particularly cabergoline, which has a very long biological half-life, may be advantageous in PD patients who develop motor fluctuations.

However, these drugs have significant drawbacks and side effects.Their potency is lower than that of levodopa and they can rarely be used as monotherapy, except in the initial stages of the disease. As ergoline derivatives, they are not very specific and interact with several subtypes of DA receptors as well as with 5-HT receptors. D1 stimulation may contribute to the occurrence of dyskinesias, while 5-HT and D4 stimulation may result in hallucinations and other psychiatric manifestations. Ropinirole, a novel synthetic non-ergoline derivative, is specific to DA (particularly D2) receptors and may therefore be advantageous. In addition, DA agonists act also at the periphery, and this may contribute to significant side effects such as orthostatic hypotension and nausea.

The inactivation of DA after its release into the synaptic cleft involves both reuptake by DA terminals and metabolism by catechol O-methyl transferase (COMT). The reuptake is performed by specialized DA transporter molecules in the membrane. Inhibitors of this transporter, as well as those of COMT, may prolong the action of DA and thus potentially could affect the response fluctuations (32).

Surgical interventions of PD include ablative and transplanting approaches. Targets for functional stereotactic neurosurgical lesions, which reduce tremor, are the ventrolateral thalamus and the posteroventral pallidum. There has been extensive interest in transplanting DA tissue into the caudate or putamen in PD. Originally, autologous adrenal tissue was used, but the benefits, if any, were less than the risks (13). This approach was discarded, and more recently dopaminergic transplants were used where the tissue was removed from aborted fetal midbrains. It is difficult to assess the success of this approach, because frequently the patients who were recruited had a poor prognosis to start with and also because this intervention is associated with a high placebo factor (55).


The prevalence of frank dementia in PD is far greater than that in the general population. PD dementia may be preceded by mild memory loss, transient confusional episodes, or hallucinosis. The progression of the cognitive decline is unrelated to that of the motor disability, and the only robust predictor for the development of dementia is the patient's age. Clinically, the dementia of PD differs from that of Alzheimer's disease (AD). PD patients rarely develop dysfunctions of the isocortical association areas, such as dysphasia or agnosia, and resemble a "frontal" type of dementia. But while the differentiation between cortical and subcortical dementia is of theoretical importance, it cannot be applied to individual PD patients who may develop a clinical AD-like picture. Cell loss in PD is not limited to DA neurons. The degeneration of cholinergic neurons in the nucleus basalis of Meynert, as well as 5-HT, NE, and somatostatin damage, was documented. These deficiencies are similar to those observed in AD and therefore suggest similarities in pathogenesis and treatment, as well as a clinical overlap.

During the past decade it became obvious that Lewy bodies are not limited to the substantia nigra in PD, but may occur in a widespread distribution in the cortex. Diffuse Lewy body disease is a pathological entity whose clinical correlates have not yet been defined. Patients commonly have cognitive decline and Parkinsonian features, and either one may dominate the picture.

Treatment of the cognitive changes of PD is unsatisfactory. The cholinergic defect suggests that drugs with antimuscarinic action may be detrimental, and these include of course not only specific antiparkinsonian agents such as benzhexol or trihexyphenidyl but also antidepressants such as amitryptiline. Contrariwise, cholinomimetic agents such as tacrine, presently used in AD, may increase the motor disability. Treatment of hallucinations and delusions similarly poses difficult problems because the use of D2 blockers may well cause exacerbation of the motor symptomatology. Recently the use of clozapine, a specific D4 blocker, was suggested as an efficacious treatment of this condition (16). In addition, if the cognitive decline and the motor deterioration both result from similar processes of neurotoxicity, it is possible that deprenyl may retard both. (TheEffects of Neuroleptics on Plasma Homovanallic Acid and Neurocognitive Functioning in Paitents with Schizophrenia: An Overview, for related topics.)


Exactly how frequently depression occurs in PD is a question that is extremely difficult to answer. There is quite a spectrum of answers, which diverge depending on (a) the criteria used to diagnose depression, (b) possible inclusion or exclusion of demented patients or those with parkinsonism due to causes other than PD (e.g., vascular etiology and progressive supranuclear palsy), and (c) the severity of their neurological impairment. In addition, referral bias to specialized centers may result in excessive numbers of depressed patients. However, and regardless of these factors, it is safe to conclude that depression is rather common in PD. Because depression is potentially treatable, this conclusion is of significant importance (8).

Several tests are available for diagnosing depression. These include neuropsychological evaluations, self-reports, projection tests, and others. However, while all these tests have important roles in research, none is superior to the clinical assessment by a competent clinician. Nor is such a test likely to ever be developed, because the manifestations of PD are quite varied.

The clinical evaluation of the affective state of PD patients may be difficult because the motionless face, the slowness of movement, and the bradyphrenia may create an erroneous impression of depression. The distinction from depressive motor retardation is obviously very important.

Decisions about the therapeutic approaches should be based not solely, perhaps not even primarily, on an objective measure but rather on the context and repercussions of the affective state of the patient.

Based on the above, every patient with PD must be assessed for possible depressive symptomatology, and adequate consideration should be given to the therapeutic implications.

The therapeutic considerations regarding the depression in PD may be quite unlike those for major depression. In the latter situation, massive treatment with tricyclic antidepressants is recommended, with the expected benefits occurring only several weeks later. If not efficacious, other treatment options exist, eventually leading to the employment of electroconvulsive therapy (ECT) (1, 11). However, in the parkinsonian patient who is depressed, less aggressive therapy is usually sufficient, and high doses may in fact cause intolerable side effects.

The present knowledge of therapeutic options for parkinsonian depression is limited because of the scarcity of drug evaluations in this condition, let alone of comparative studies of different agents.

Tricyclic antidepressants (TCAs) have marked antimuscarinic effects. These are potentially advantageous for the PD patient because they reduce the motor symptoms, particularly the tremor. Another feature of TCAs is their anxiolytic action, and of course this is helpful in those patients manifesting anxiety symptomatology. The third relevant feature is the soporific effect of TCAs which is of significant value in those suffering from insomnia (although some patients respond to TCAs with increased alertness and restlessness) (17).

Because of their supposed mechanism of actionnamely, blockade of the reuptake of 5-HT and NE by nerve terminalsthese drugs are of potential value in the depression of PD (31, 35). Previous attempts to differentiate depressive symptomatology related to 5-HT as opposed to NE dysfunction have been largely unrewarding. In addition, most drugs block the reuptake of both neurotransmitters. For example, while amitriptyline mainly blocks 5-HT reuptake, its active metabolite, nortriptyline, preferentially blocks NE reuptake.

The antimuscarinic action of TCAs, already alluded to, may well lead to disorientation and confusion. This is particularly true when patients with more limited cognitive reserves are being treated (i.e., those with imminent or actual dementia), when relatively high doses are prescribed, or when employed together with antiparkinsonian drugs with antimuscarinic actions (18). Although these newer drugs do have specific actions, it remains to be demonstrated that this is of practical significance.

Selective 5-HT reuptake blockers include clomipramine, fluvoxamine, and fluoxetine. Clomiprimine is a tricyclic derivative, while the others have novel structures. Fluoxetine and fluvoxamine lack antimuscarinic actions and thus may be particularly useful in those patients for whom the use of anticholinergic drugs is contraindicated.

The use of monoamine oxidase inhibitors (MAOIs) is of course well established for the treatment of depression; although they have a bad reputation regarding safety, they continue to be used. Ever since it was realized that DA deficiency is responsible for PD, attempts were made to treat it by MAOIs, but the response was limited. It is probably true that MAOIs can successfully be used in PD patients who are depressed, with expected mild benefits also in the motor function.

The use of ECT is reserved to patients with severe depression. Although previous reluctance to use this treatment in the elderly seems to have been excessive, there is only anecdotal information on its use in PD. Some case reports suggested improvement in both affective and motor symptomatology (1, 11).

Meager data exist suggesting an independent antidepressant action of levodopa. Bromocriptine is also reputed to have some antidepressant activity, although, again, this depends on unconfirmed, uncontrolled observations (21). Lastly, some anticholinergic drugs used in the treatment of PD have mood-elevating actions. This is particularly true for orphenadrine.

published 2000