Motor Complications in Parkinson's Disease

ABSTRACT

Management of motor complications in advanced Parkinson's disease (PD) can be challenging. The main complications are inadequate dopaminergic tone (“off” time and dose failures) and excess dopaminergic tone (dyskinesia). These motor complications increase as PD progresses. Changing the dose and timing of L-dopa is the main strategy for both scenarios. Reducing “off” time can also be achieved by the addition of adjunctive therapies (dopamine agonists, catechol-O-methyl transferase inhibitors, and monoamine oxidase-B inhibitors). Dyskinesia can improve with amantadine and possibly several other medications. Surgical interventions such as lesioning and deep brain stimulation are considered when pharmacological strategies for motor complications are not satisfactory.

KEYWORDS: :

• dyskinesia
• fluctuations
• Parkinson
• surgery
• motor complications

GENESIS OF MOTOR COMPLICATIONS IN PARKINSON'S DISEASE

Parkinson's disease (PD) is a common neurodegenerative disorder resulting in tremor, rigidity, bradykinesia, postural instability, and other motor and behavioral symptoms. There is diffuse pathology throughout the central and peripheral nervous system; however, loss of nigra dopaminergic neurons correlates with the cardinal motor signs that clinically diagnose PD. Treatment with dopaminergic medications, especially levodopa [a precursor to dopamine (DA)], improves motor dysfunction. Although initially satisfactory, continued treatment with levodopa results in motor complications, including reduced duration of response, dose failures, dose variability in response, and dyskinesia, in the majority of patients within 3–5 years, and when specifically observed for, in up to 20% within a single year [1, 2]. This occurs most rapidly in younger patients and in patients with more advanced cardinal features, especially rigidity [1]. Fluctuations can also be seen with DA agonists (longer-acting medicine that stimulate DA receptors) but are much more associated with L-dopa, the most potent symptomatic treatment for PD [3].

The etiology of the shortened duration of effect and inconsistency of response is probably due to pharmacokinetics and L-dopa's poor absorption through the duodenum [4]. Only about 5% of an orally ingested pill of L-dopa will eventually enter the brain as attrition is seen with gut absorption, peripheral metabolism, and competition to cross the blood–brain barrier [5, 6]. Therefore, small changes in gut absorption, which can be affected by food or gut motility, can result in large changes in effective brain L-dopa levels. Furthermore, L-dopa actively competes with other branched chain neutral amino acids to cross into the brain, so protein content in a meal can more specifically inhibit transportation into the brain. The serum T^1/2 of L-dopa is only 90 minutes so a short duration of effect is not surprising. When DA cell loss is less severe, both the short pharmacokinetics and inconsistent absorption are buffered by the fact that L-dopa is taken up into the dopaminergic cells through the DA transporter protein, and then released physiologically over time. Therefore, a single dose may improve symptoms for 6–8 hours despite being long eliminated from the serum. As the disease progresses and the presynaptic DA cells are further lost, the brain loses this buffering capacity, and clinical effect begins to more exactly follow the pharmacokinetics of the drug. Therefore, patients may derive only 1–2 hours of clinical effect, or none at all, if passage to the brain is interrupted at any stage.

The phenomenology and mechanisms of drug-induced dyskinesia are more complicated. Phenomenologically, dyskinesia usually consists of appendicular or neck choreiform and stereotypic movements. Dystonia and myoclonus may also occur, but are less common. Most common is a mixed chorea, stereotype with a nonpainful, nonforceful, dystonic component. Dyskinesia usually starts in the foot of the most affected side, although this often goes unrecognized by the patient [7]. Although dyskinesia can be dramatic and functionally incapacitating, most patients usually prefer dyskinesia to being “off” (no benefit from dopaminergic therapy), and often, the caregiver and physician are more concerned by their presence.

In the majority of patients, dyskinesia correlates with the highest levels of serum and cerebrospinal fluid levodopa, “peak dose dyskinesia”. This is referred to as the improvement-dyskinesia-improvement (I-D-I) pattern. A minority of patients however develop a pattern consisting of dyskinesia (as serum levels rise), followed by clinical improvement (at the highest serum levels), followed by dyskinesia (as levels fall), or just improvement followed by dyskinesia as the drug wears off. This dyskinesia-improvement-dyskinesia (D-I-D) pattern is particularly difficult to manage.

The pathology of dyskinesia is not entirely understood. Presynaptic DA cell loss is necessary as normal individuals will not develop dyskinesia with L-dopa [8]. A similar loss of buffering capacity probably accounts for the greater prevalence of dyskinesia as disease progresses. Without internalization of the exogenous L-dopa into presynaptic DA cells, a greater amount of DA directly stimulates the DA receptors. Positron emission tomography studies do indeed show a significant reduction in striatal ¹⁸F-fluorodopa in advanced and fluctuating, but not in stable nonfluctuating PD patients, and patients with dyskinesia have more rapid fluctuations in CNS DA concentrations, compared with stable responders [9]. This, however, cannot account for all the risk of dyskinesia. There is ample clinical and laboratory evidence suggesting that L-dopa changes the DA receptors and physiology abnormally to facilitate dyskinesia [8]. Clinically, even in advanced but untreated PD, it takes time for L-dopa to cause dyskinesia. If it were simply a loss of buffering, the first dose should cause dyskinesia as much as any other. Second, apomorphine, a powerful short-acting DA agonist that is not taken up into DA cells, and thus not affected by buffering capacity, can also cause dyskinesia in patients who already have dyskinesia from L-dopa [10]. Animal models of PD dyskinesia show a number of DA receptor second messenger changes after L-dopa administration. Much research specifically implicates DA type-1 receptors, as opposed to type 2/3 receptors. Foremost, D1 knockout mice will not develop dyskinesia whereas D2 and D3 knockout mice can [11]. Striatal neuronal plasticity changes, including changes in gamma-aminobutyric acid (GABA) receptors, are also greatly affected by dopaminergic tone, in conjunction with glutamatergic stimulation from the cortex [11]. This postsynaptic L-dopa priming aspect of dyskinesia development is the greatest theoretical rational for withholding L-dopa until absolutely necessary.

Clinically, total levodopa dose and duration of treatment usually correlate with the development of dyskinesia [2]. The mode of levodopa administration may also be important as more stable exogenous DA stimulation, or lack of pulsatile stimulation, also reduces the development of dyskinesia [3]. Animal studies demonstrate that intermittent dopaminergic infusions result in more dopaminergic sensitization when compared with continuous infusion [8]. Less controlled short-term human studies have also reported less dyskinesia and fluctuations with continuous infusions of levodopa compared with pulsatile oral administration. Specifically, Mouradian et al. showed that even 1 week of continuous intravenous levodopa, as opposed to oral pulsatile intake, reduced the threshold for the subsequent development of dyskinesia [12]. However, strategies to reduce pulsatility using different oral preparations of L-dopa have not clearly reduced fluctuations or dyskinesia [13, 14]. In reality, these strategies may not have adequately reduced serum L-dopa fluctuations, so the concept that continuous dopaminergic stimulation reduces fluctuations is not disproven.

MANAGEMENT OF WEARING OFF AND DOSE FAILURES

L-dopa

The most important strategy for reducing off time and other fluctuations with L-dopa is manipulation and fine adjustments of the L-dopa itself. L-dopa is a precursor to DA, converted to DA by dopa decarboxylation, and is usually available combined with a peripheral decarboxylase inhibitor such as carbidopa or benserazide, which inhibits dopa decarboxylase outside of the blood–brain barrier. There are many L-dopa preparations. Generic carbidopa/levodopa is dosed as 10 mg/100 mg, 25 mg/100 mg, and 25 mg/250 mg. Slow-release preparations take longer to clinical onset but have a longer clinical effect. They are hampered by inconsistent absorption but can be used, often in conjunction with immediate-release L-dopa. Mixed immediate-/controlled-release pills are in development. An orally dissolvable L-dopa preparation may have a faster onset in some patients [15]. Dissolving L-dopa in a slightly acidic liquid may also speed absorption. Intraduodenal/jejunal infusion of levodopa (Duodopa) forms an alternative route of administration that is available in some countries [16]. This continuous infusion of L-dopa effectively controls motor fluctuations and reduces off time in advanced cases. A number of other investigational manipulations of L-dopa are also underdevelopment. That said, nuanced manipulation of regular L-dopa is the most important method to reduce off time. This requires a careful assessment of the patient's “on”/“off” time, often aided by daily diaries. Most commonly, more frequent dosing is required. Up to 10 doses of L-dopa per day are not rare in advanced fluctuating patients. Dose changes to accommodate low endogenous DA levels (mid-afternoon) and compensate for the effects of food intake (especially protein) and longer durations of L-dopa abstinence (first morning dose), are often needed to optimize on time.

Catechol-O-Methyl Transferase Inhibitors

Even with dopa decarboxylase inhibitors (carbidopa), only 5%–10% of each levodopa dose reaches the brain. Most of the rest is metabolized peripherally by catechol-O-methyl transferase (COMT). COMT inhibitors (entacapone and tolcapone) lead to a 20%–50% increase in levodopa half-life, and a 25%–50% increase in its concentration versus time curve [17–19] (Table 1). In clinical studies, they increase on time by 10%–25% over placebo when added to L-dopa in patients with “off” time [18, 19]. Tolcapone probably has a more robust effect but is less used due to several cases of fatal hepatic toxicity. Liver function test monitoring, aspartate aminotransferase (AST) and alanine aminotransferase (ALT), is required for tolcapone but not for entacapone. The more common adverse events (AEs) include diarrhea (more common with tolcapone) and fluid decolorization (more common with entacapone). Other AEs are generally the result of increased dopaminergic stimulation (dyskinesia, hallucinations, sedation, etc.). Combination pills with entacapone/levodopa/carbidopa are available in many countries with L-dopa doses of 50, 75, 100, 125, 150, and 200 mg. COMT inhibitors are only used in conjunction with L-dopa.

TABLE 1.  Medications used to increase “on” time in patients taking L-dopa

Drug or class	Mechanism	Dose	Specific side effects	Comment
DA agonists (see Table 2)	Direct stimulation of DA receptors	(see Table 2)	Nausea, vomiting, hypotension, edema, impulse control disorders	Mechanism thought to be independent of L-dopa
Rasagiline	MAO-B inhibitor	0.5–1 mg/day	Minimal	Contraindicated with meperidine and some other drugs
Selegiline	MAO-B inhibitor	5–10 mg/day	Minimal	Contraindicated with meperidine and some other drugs
Orally disintegrating selegiline	MAO-B inhibitor	1.25–2.5 mg	Minimal	Contraindicated with meperidine and some other drugs, dissolves on tongue
Entacapone	COMT inhibitor	200 mg with each dose of L-dopa	Body fluid decolorization, diarrhea	Available in L-dopa combination pills
Tolcapone	COMT inhibitor	100 mg 4×/day	Rare hepatic failure, diarrhea	Liver function monitoring

Monoamine-B Oxidase Inhibitors

Monoamine-B oxidase (MAO-B) inhibitors block the metabolism of DA into dihydroxyphenylacetic acid (DOPAC). Nonspecific MAO inhibitors also block the metabolism of serotonin, tyramine, and norepinephrine, which when combined with other medications or ingested tyramine can result in a life-threatening sympathetic discharge (serotonin syndrome and the “cheese reaction”).

Three available MAO-B preparations have increased “on” time in controlled trials: selegiline, a sublingual preparation (Zydis selegiline), and rasagiline. Oral selegiline is hepatically metabolized into l-amphetamine and l-methamphetamine, as well as desmethylselegiline. Although both amphetamine metabolites further augment DA synaptic concentrations, they are potential pro-oxidants that may offset any potentially beneficial antioxidant properties of selegiline. The sublingual preparation of selegiline bypasses that metabolism. Rasagiline is metabolized to compounds, without affinity for MAO-B, with possible neuroprotective properties. All three drugs modestly increase on time [17, 20, 21], and rasagiline and selegiline are also used as monotherapy. AEs are mild and consistent with increased dopaminergic tone (increased dyskinesia, etc.). Dietary restrictions are no longer recommended with MAO-B inhibitors. In the only large comparative trial evaluating different classes of agents to reduce “off” time, rasagiline and entacapone demonstrated equal efficacy [20].

DA Agonists

DA receptor agonists mimic the effect of DA by binding directly with the postsynaptic DA receptors, thus bypassing the degenerating presynaptic nigrostriatal neurons. DAs are useful both as monotherapy in mild to moderate PD and as adjunctive therapy to levodopa in later disease. They vary in relative DA receptor affinities, pharmacokinetics, delivery systems, and chemical structure (Table 2). Extended-release preparations are also available for pramipexole and ropinirole. Ergot-derived DAs (bromocriptine, cabergoline, lisuride, and pergolide) are seldom used now due to their increased association with pleural, pericardial, and peritoneal effusion, and/or cardiac valvular fibrosis. In studies, DAs improve on time by 1–2.4 hours over placebo [17]. Extended-release ropinirole demonstrated the greatest improvement over placebo but this may have resulted from methodological considerations [22]. Adverse effects of DA agonists are more common than with the other adjunctive therapies and include nausea, vomiting, postural hypotension, hallucinations/psychosis, excessive daytime somnolence and sudden sleep episodes, peripheral edema (often years after initiation), and impulse control disorders such as gambling, increased spending, and hypersexuality. In all of these examples, the AE usually resolves quickly upon dose reduction or discontinuations.

TABLE 2.  Properties of various available DA agonists

	Route	Daily dose dosing	T½ (hours)	D1	D2	D3	α2	Distinctive AEs
Apomorphine	SC, SL	Variable	0.5–1.5	+++	+++	0	Not Known	Stomatitis (SL), acute hypotension
Bromocriptine	PO	7.5–40 mg TID/QID	3–8	+	+++	+	+	Ergot side effects
Pergolide	PO	0.5–6.0 mg TID	27	+++	+++	+	++	Ergot side effects
Pramipexole	PO	0.25–4.5 mg TID	8–12	0	+++	+++	+
Ropinirole	PO	0.75–24 mg TID	6	0	+++	+++	0
Rotigotine	Patch	2–8 mg/day	4–6	+	++	+++	+	Patch site irritation
Cabergoline	PO	0.25–4.0 mg QD/QOD	60–100	+	+++	0	+	Ergot side effects

Note: PO, oral; SC, subcutaneous; SL, sublingual.

Apomorphine

Apomorphine is a powerful DA agonist that is administered by subcutaneous bolus doses or continuous infusion. People with frequent “off” episodes or dose failures are suitable for intermittent bolus injections. The threshold dose to achieve an “on” state is established during a supervised initial injection session, where the patient is trained to use a prefilled apomorphine injection system. The degree of clinical effect approximates that of L-dopa in fluctuating patients but the onset of action is usually 5–10 minutes and the total duration of effect only 45–90 minutes [23, 24]. Therefore, it is used differently than other adjunct therapies that increase the duration of L-dopa effect. Apomorphine is used as “rescue therapy” when a rapid “on” is required. In some European countries, apomorphine can also be administered continuously by a portable syringe driver connected via a butterfly cannula sited in the abdominal wall or subcutaneous tissue of the thighs. A programmable pump delivers 50–120 mg of apomorphine over the waking day or the entire 24-hour period [25]. This therapy is difficult to use but when tolerated very effectively treats fluctuating PD, often in place of L-dopa.

TREATMENTS FOR DYSKINESIA

As with managing “off” time, the initial consideration for treating dyskinesia is the manipulation of L-dopa dosing. In general, fractionation of doses (smaller amount more frequently) is effective. Dyskinesia is less common with the initial morning dose so this dose often can be higher. Slow-release L-dopa preparations generally need to be jettisoned as they are too unpredictable to fine tune in the face of dyskinesia.

End of dose or D-I-D patterned dyskinesia may not respond to this dose fractioning strategy. In fact, L-dopa increase, to avoid the wearing off, is occasionally effective in the D-I-D setting. If dose adjustments are not satisfactory, several other medications have been shown to improve dyskinesia without worsening underlying PD symptoms.

Amantadine

Amantadine is an aliphatic primary amine antiviral agent, which has been used in the treatment of PD for almost 30 years. The drug has multiple CNS mechanisms. It increases DA release from presynaptic terminals, presumably at the DA transporter protein, is an anticholinergic, and has modest N-methyl-D-aspartate (NMDA) antagonist properties via blockade of open ion channels, resulting in diminished glutamatergic activity [26]. Glutamate antagonism may directly benefit dyskinesia associated with levodopa use by reducing glutamatergic outflow from the subthalamic nucleus (STN) and/or cortex.

Amantadine has improved L-dopa-induced dyskinesia in several controlled trials and is commonly used in clinical practice [27–29]. The typical dose is between 200 and 800 mg/day in two to four divided doses. It should be dosed to coincide with the presence of dyskinesia. The medicine is well tolerated but may cause mild nausea and anticholinergic side effects. It usually causes livedo reticularis, a lattice-like, mottling discoloration of the skin, seen especially around the shin. This is largely cosmetic but may be accompanied by edema. Resolution may take weeks after drug discontinuation. Psychosis is a relatively common problem with amantadine, especially in older and demented patients.

Other Glutamatergic Drugs

Memantine is an uncompetitive, low-affinity, open-channel blocker that enters the NMDA-type glutamate receptor preferentially when it is excessively open and has a relatively “fast off” so that it does not substantially accumulate in the channel to interfere with normal glutamate transmission. The drug may help cognition in PD [30–32] and has been reported to improve dyskinesia [33]. Dextromethorphan, a noncompetitive NMDA antagonist and sigma-1 receptor agonist, has also improved dyskinesia in one controlled trial [34].

Clozapine

Clozapine, an atypical D₄ blocker antipsychotic, effectively treats PD-associated hallucinations, improves PD tremor, and improves dyskinesia, all without compromising motor function [35–37]. The medicine's most common adverse effects include sedation, which can be greatly beneficial at night, confusion, and sialorrhea. In approximately 1% of cases, a potentially fatal granulocytopenia may occur, usually within the first year of use. Therefore, weekly complete blood counts are initially required, which greatly limits the drug's use in PD except for those with refractory psychosis. Other “atypical” antipsychotic medications have been reported to improve dyskinesia, but all except quetiapine (seldom effective for dyskinesia) will worsen the cardinal motor features of the disease.

Levetiracetam

Recently, levetiracetam has shown mixed results in three placebo-controlled dyskinesia trials [38–40]. This antiseizure medication has several mechanisms of action: reduction of high-voltage N-type calcium channels, opposition of the action of GABA(A) antagonists, reduction of the effect of voltage-gated potassium channels, and inhibition of hypersynchrony. The drug is well tolerated at doses up to 2,000 mg/day. No or slight worsening of PD signs was seen in the trials.

Miscellaneous

Several other medications also occasionally improve dyskinesia [17]. Benzodiazepines reduce movements, especially when exacerbated by anxiety. It certainly can cause sedation, confusion, and worsen balance, so utility is limited. In our experience, injections of botulinum toxin can improve head and neck dyskinesia but not limb or trunk movements. Finally, if fluctuations, including dyskinesia, cannot be adequately controlled with pharmacological manipulation, then surgical treatments may be considered.

SURGICAL TREATMENTS FOR PD

Stereotactic lesions into the thalamus and globus pallidus internus (GPi) were first performed for PD more than 50 years ago. After L-dopa and other agents became available, these procedures were largely abandoned, but reemerged in the mid-1990s as treatments for dopaminergic fluctuations, especially dyskinesia. Lesioning of the GPi (pallidotomy) markedly reduces dyskinesia and moderately improves the cardinal features of PD, most notably rigidity [41]. Patients who had a pallidotomy generally continued their PD medications with much less dyskinesia. In fact, medication doses often modestly increased to minimize “off” time. A correctly placed pallidotomy had minimal side effects, although lesioning of the adjacent internal capsule could result in dysarthria or other weakness. Bilateral pallidotomies modestly improved clinical benefit but markedly increased the risk of dysarthria and were usually not performed. Lesions of the thalamus continued to improve tremor, but lesioning the STN was riskier and never widely adopted.

In response to the limitations of lesioning therapies, deep brain stimulation (DBS) was developed. This involves stereotactic implantation of an electrical lead into the desired area [usually STN, GPi, or ventral intermediate nucleus (VIM) for PD patients]. A wire is then run to an implanted pulse generator (IPG) with a battery. A number of different leads and IPG/batteries now exist. A high-frequency pulsed current is administered. The actual mechanism by which this effects local nerve function is debated but the clinical global effect is similar to that of a lesion in that brain area.

The major advantage of DBS over ablative procedures is that the stimulating electrodes (multiple electrodes are placed on each lead) and electrical parameters (frequency of stimulation, pulse width, and voltage or current) can be adjusted and “customized” to the needs of the individual patients. Furthermore, stimulation side effects, if they occur, are reversible, and bilateral stimulation is possible without the fear of permanent bulbar AEs. In some patients, ablative thalamotomy or pallidotomy on one side may be combined with or followed by DBS on the opposite side [42]. There are numerous technical issues regarding surgical and subsequent adjustment techniques beyond the scope of this text [43].

Proper patient selection for DBS is critical. Most feel that DBS should be reserved for patients with motor complications, as the observed motor improvement with STN DBS is not greater than that of optimized L-dopa. A number of positive and negative predictors of response have emerged [44, 45]. Younger age fairly consistently predicts better overall response. There is no clear age cutoff, but results generally diminish after age 70. This of course may be at odds with the principle of waiting for fluctuations. A dramatic clinical improvement with L-dopa also predicts STN DBS response. Relative contraindications include any dementia, true “on” medicine balance problems, or marked residual problem aside from dyskinesia while “on”, as these generally will not improve with DBS. Therefore, the ideal DBS candidate is a young onset, badly fluctuating patient, but with minimal “on” cardinal PD features, cognitively normal, and otherwise healthy. This represents only a minority of PD patients. Some do advocate DBS earlier in the disease citing potential improvement in quality of life measures compared with best medical management, but at the cost of more serious AEs [46].

The ideal target (STN vs. GPi) is also debated. Although historically STN DBS has been more commonly performed, each case should be individualized. Although each case should be individualized, STN generally allows more reduction of PD medications [47]. GPi DBS may relatively improve cognitive and mood outcomes. STN, which is more commonly performed, is perceived to improve “off” motor scores more than GPi, and GPi perceived to improve dyskinesia more; however, neither of these findings are consistently seen in studies. Both procedures are shown to improve “off” motor scores, on/off time ratio, dyskinesia, and quality of life [46, 47].

Perioperative surgical complications of course depend on the surgical team, but can also be unexplained. Hemorrhage and stroke are the most serious and occur in 1–5% of surgeries [48]. Lead infection or device damage, such as a fractured wire, can occur at any time. The most common postactivation AEs of STN DBS include dysarthria and worsening of balance. Weight gain is common and a temporary increase in dyskinesia is occasionally seen. Perhaps, more concerning is subtle but consistently observed worsening in cognitive functioning, especially executive measures. Cognitive decline may be less with GPi DBS [47] and VIM DBS.

Overall, in the proper patient and with skilled surgical and postsurgical management, DBS can dramatically improve the lives of patients. Unfortunately, DBS usually only helps the motor symptoms of PD, and there is no evidence that DBS slows the progressive course of PD, especially the nonmotor symptoms that can disable later-stage PD patients.

CONCLUSIONS

Management of motor complications in advanced PD can be challenging. Clinical trials and evidence-based medicine may offer some clues but ideal management required marked attention to the patient and a very individualized approach. Manipulation of L-dopa dosing, usually more frequent dosing, and adjunctive therapy with MAO-B inhibitors, DA agonists, or COMT inhibitors can improve “on” time. Dyskinesia is treated with L-dopa manipulation (smaller dose more frequently and elimination of extended-release preparations), the addition of amantadine, and possibly several other medications, or surgery. Strategies to optimize “on/off” time must be balanced against side effects of these strategies and the management of nonmotor features of PD and other medical problems.

Declaration of interest: The author reports no conflicts of interest. The author alone is responsible for the content and writing of the article. Speaker for TEVA, GSK, Allergan, Merz, Ipsen, Novartis Grant support Takeda, Allergan, Bayer, and HSG.