Considering that the characteristic motor features of Parkinson’s disease (PD) are caused by a loss of dopamine in the brain, it stands to reason that medications which decrease dopamine’s action would cause a parkinsonian condition. Drugs known to deplete dopamine, such as reserpine or tetrabenazine, certainly have this effect. However, the most common scenario in which drug-induced parkinsonism (DIP) arises is with postsynaptic dopamine receptor blocking agents such as antipsychotic or antiemetic medications. These drugs are covered elsewhere in this book.
Lesser known and often enigmatic is when parkinsonism is caused by medications that have no apparent action upon the dopamine system. Perhaps this is an indication that there is an occult antidopaminergic effect or that there is another lesser understood alternate mechanism that causes the clinical effect of parkinsonism. Such is the case with the medications we will discuss in this chapter. Most of the medications were designed for a specific purpose, such as blood pressure control in the case of the calcium channel blockers, but have caused a significant amount of morbidity in the form of DIP. DIP can also develop with exposure to naturally occurring substances used as therapeutic agents as well. For instance, lithium is a naturally occurring element that has been used since ancient times when its clinical effect became apparent, but there was no knowledge of pharmacology. Lithium probably exerts its psychiatric benefits through several mechanisms and may cause DIP through several as yet unknown mechanisms as well.
In this chapter, we will cover the nondopaminergic medications that induce parkinsonism. Table 12.1 summarizes such agents. Although antidepressants fall into this category, they will be covered elsewhere, with the exception of lithium.
|Category of medications||Class of medications||Subclass||Drug names|
Calcium channel antagonists
Calcium channel antagonists
etoposide, methotrexate, mitoxantrone,
The incidence and prevalence of DIP under these circumstances is difficult to quantify because DIP is often not recognized (1), and in some cases is extremely rare and unexpected. For example, while most physicians would know to attribute extrapyramidal side effects to neuroleptics, many may not know of the role that valproic acid has in causing parkinsonism. Furthermore, in situations where single or few cases have been reported it is hard to know for sure if there is a cause and effect relationship. A good example of this is the drug amlodipine which enjoys widespread use for hypertension and cardiac arrhythmias, and yet only two cases of parkinsonism have been reported. Nevertheless, several studies have attempted to examine the incidence of DIP. A French pharmacovigilance center tallied up all of its cases from 1993 to 2009. Parkinsonism was defined as having either resting tremor, rigidity, or akinesia (2). Dopamine antagonists constituted 54% of the DIP cases, antidepressants 8%, leaving a substantial percentage being due to a variety of nondopaminergic medications, suggesting that this type of DIP is of significant importance.
Calcium channel blockers: nondihydropyridine
The antihypertensive/antiarrhythmic medications diltiazem (3, 4) and verapamil (5, 6) are both widely used in the US, but were only implicated in causing DIP in a few case reports decades ago. On the other hand, two calcium channel blockers that are not approved in the US, cinnarizine and its derivative flunarizine, are well known to cause parkinsonism. These are utilized for migraine prophylaxis, vestibular disorders, and as adjunctive treatment for epilepsy in southern Europe and South America (7). DIP caused by these two calcium channel blockers is characterized by the presence of a symmetric akinetic-rigid syndrome which may or may not be accompanied by rest or postural tremor. Other associated features include akathisia and dyskinesia (8). Risk factors for developing DIP from calcium channel blockers include female gender and age over 60 (7, 9–11).
In one prospective study, 93 out of 101 patients who were prescribed flunarizine or cinnarizine developed parkinsonism (9). Some studies (9, 11) indicated a good outcome after the drugs were discontinued with all or almost all patients recovering over the course of months. Garcia-Ruiz et al. found a difference in outcome based on the phenotype of their parkinsonism; the akinetic-rigid symptoms were mostly reversible, but tremor tended to persist long after the drug was removed (7). Conversely, Negrotti et al., had only 13 patients in their study, but found that all continued to have nonprogressive extrapyramidal side effects seven years after withdrawal of the offending medications (10).
Cinnarizine and flunarizine have antihistaminic, antiserotonergic, and antidopaminergic activity, so there are several theories as to why they cause parkinsonism. This could be a presynaptic process via loss of tyrosine hydroxylase (12) leading to dopamine depletion, by interfering with calcium-mediated neurotransmitter release (9), or it could be a postsynaptic process by blocking striatal dopamine receptors (9, 11).
Calcium channel blockers: dihydropyridine
The class of dihydropyridine calcium channel blockers includes amlodipine and manidipine, both indicated for the treatment of hypertension. There are only two reported cases of amlodipine-induced parkinsonism. A 68-year-old woman developed tremor and parkinsonism upon initiating amlodipine, then improved within months of medication discontinuation (13). Another woman, 83 years old, developed tremor and parkinsonism within two months after starting amlodipine, which then resolved a month after stopping it (14). Manidipine has not been linked to “de novo” DIP, but rather worsening in two patients with preexisting PD (15). Amlodipine is widely used in parkinsonian patients with hypertension. Based on these reports we are left to ponder whether they are worsening motor features. Only a larger prospective examination of these agents in PD will address the issue of whether these drugs should be contraindicated in PD.
ACE inhibitor (captopril)
One other medication that bears mention is the Angiotensin Converting Enzyme (ACE) inhibitor captopril. There are two case reports that implicate this medication in the induction of reversible tremor and parkinsonism (16, 17) or worsening of preexisting PD (18). Onset of parkinsonism and then resolution of symptoms took just days. This side effect is unexpected, since none of the other ACE inhibitors have been associated with DIP.
Amiodarone is a class III antiarrhythmic, typically prescribed for supraventricular tachycardia. It carries a variety of neurological side effects including neuropathy, ataxia, visual problems, fatigue, dizziness, and tremor (19). Neurological side effects tend to be dose-dependent and can be seen during the loading period or after long-term use. Tremor is the most common neurological symptom, with 30% of users exhibiting this complication (19). Most of the time, the tremor resembles essential tremor. It has a postural predilection and a frequency of 6 to 10 Hz (20). However, it can take on a distinctly parkinsonian character as well (21). Werner and Olanow reviewed the literature and described several cases of parkinsonism caused by amiodarone (22). Most cases did not resemble PD, but instead had nonspecific extrapyramidal syndromes featuring tremor, which interestingly included jaw tremor. Amiodarone can also cause a pure akinetic syndrome (23).
The amount of time from withdrawal to resolution of symptoms depends on the duration of exposure. Symptoms appearing with short courses of amiodarone only take days to improve, while longer courses may take several months (21). Tremor is a dose-dependent side effect and it can be reduced by lowering the dosage (20). Also, adding propranolol may be helpful in treating the postural component of tremor (20).
Aprindine is a lesser known antiarrhythmic medication which reportedly caused parkinsonism in a single case report (24). A 78-year-old man developed disorientation and parkinsonism after a month’s use then recovered a few days after withdrawal. The mechanism by which amiodarone and aprindine cause parkinsonism is unknown, but they have been shown to block D1 and D2 receptors as part of in vivo and in vitro studies (25). In addition, amiodarone has been shown to interfere with mitochondrial function (26), which could explain its side effect profile. There have been two amiodarone-induced parkinsonian cases with pathological reports and they showed discrepant findings. LeMair presented a case whose pathology demonstrated depigmentation of the substantia nigra, without Lewy Bodies (27), whereas Ishida’s case revealed a lack of Lewy Bodies, with normal pigmentation of the substantia nigra (28). The former case might suggest that amiodarone has a toxic effect on nigral neurons. Further study is needed.
Trimetazidine is an antiischemic medication that has been tied to parkinsonism. Two retrospective case series with 21 cases by Masmoudi (29) and 20 cases by Martí Massó (30) demonstrate this convincingly. Parkinsonism would develop over months to years and symptoms would generally improve weeks to months after withdrawal of the drug. This medication has the same piperazine core as the aforementioned calcium channel blockers, cinnarizine and flunarizine. Therefore, it is thought that the D2 receptor antagonism leads to the development of parkinsonism.
Valproic acid (VPA) is well known to cause a reversible form of parkinsonism in pediatric populations (31–36). Features include marked bradykinesia and bradyphrenia which is sometimes associated with a reversible atrophic appearance of the brain called “pseudoatrophy.” VPA-induced parkinsonism is also a well described clinical entity in adult populations (37). VPA is best known as a cause of nonparkinsonian postural hand tremor that has been estimated to occur in up to 45% of users and is not dose- or serum level–related (38). A smaller percentage of patients have been identified to develop actual parkinsonian features. In addition to causing de novo DIP, VPA was shown to worsen symptoms in those with preexisting parkinsonism (39). The studies that have examined the frequency of DIP from VPA have primarily looked at an exclusively epileptic patient population. There were disparate rates of parkinsonism described (39–43), with the highest estimates finding parkinsonism in 10% of those exposed. The incidence in the elderly treated for bipolar disorder remains unknown.
The diagnosis can sometimes be challenging in the elderly since this is the demographic in which one would expect to see PD, and normal aging is associated with mild parkinsonism. VPA-induced parkinsonism can present as an akinetic-rigid state or, as mixed postural/rest tremors, and it can look identical to PD with a unilateral bradykinesia and rest tremor. The parkinsonism that VPA causes is not a consequence of medication overdosage or toxicity, as low doses and therapeutic serum drug concentrations can result in this complication. For instance, one patient from our own series was on a dose of only 250 mg per day when she developed her symptoms (44). Further complicating the distinction between VPA-induced parkinsonism and PD is the fact that VPA-induced parkinsonism takes months to years to develop, so symptoms do not appear upon initiation of the offending agent as more typically seen with neuroleptics. Furthermore, symptoms typically require weeks to months to resolve after discontinuation of the VPA, so it takes time to assess for clinical improvement and clarify the diagnosis. The average duration in our series was seven months (44). Yet another confounding factor to distinguish between VPA-induced parkinsonism and PD is the fact that it is responsive to levodopa (44, 45) and one can even see levodopa-induced dyskinesia (44).
The mechanism by which VPA causes parkinsonism is unknown. Like other causes of DIP, it displays normal dopamine transporter density on SPECT scan (41), hence it does not appear to be toxic to presynaptic dopaminergic neurons. Based on the fact that this condition can be levodopa responsive, one would not expect a postsynaptic dopaminergic blockade to be the cause. Theories with regard to the mechanism of development of DIP include GABAergic effects, since VPA is known to increase GABAergic transmission (38), which is significant in the basal ganglia, including the substantia nigra. This may result in a net inhibition of dopaminergic activity (46). Other theories include the slow accumulation of a toxic metabolite of VPA in the substantia nigra (47), inhibition of mitochondrial metabolism (48), and inhibition of neurite outgrowth (49).
There are only two cases in the literature of parkinsonism associated with phenytoin. The first case described a 68-year-old man with seizures as a result of cranial injury who developed parkinsonism and tremors within one month of initiating the drug. Symptoms resolved over 6 months after switching to an alternative antiepileptic medication (50). The other case describes parkinsonism, tremor, and cerebellar ataxia within two weeks of phenytoin exposure, but his drug level was exceptionally high, at 40 mg/dL (51). Both patients had a history of severe cerebral trauma, so perhaps phenytoin requires the substrate of brain injury to cause parkinsonism.
Lithium is known to cause a variety of movement disorders including tremor, parkinsonism, and chorea (52). Like VPA, lithium causes a postural and action high frequency tremor that is distinct from PD, similar to essential tremor, and is thought to be a form of enhanced physiologic tremor (53). In fact, it might be the most common form of drug-induced tremor seen (52), with an incidence as high 33%–65% (54). The variance among different estimates of incidence is the result of lithium frequently being prescribed with other tremor-producing medications. The tremor can occur even when the drug levels are therapeutic.
The notion that lithium can cause parkinsonism has been a controversial and complicated one. Certainly, when it does truly cause parkinsonism, it does so rarely. But patients who are prescribed lithium usually take it for bipolar disorder, often with neuroleptic drugs. Furthermore, much of the early literature on the matter probably overdiagnosed parkinsonism based on the physical exam findings of postural tremor and cogwheel phenomena without true rigidity. Cogwheeling is a nonspecific exam finding that does not necessarily indicate rigidity and is noted in patients with all types of postural tremor, probably as an extension of the tremor itself.
The first two papers on the matter examined randomly selected lithium users for “cogwheel rigidity” (55, 56), but it was not made clear if this represented true parkinsonism. In addition, in the first paper it was not clear if any were treated concomitantly with neuroleptics. Kane et al. then reported on 38 patients with therapeutic levels of lithium, two of whom had parkinsonian features, but both had been on neuroleptics in the year prior (57). It has been shown that neuroleptic-induced parkinsonism can last that long or even longer after neuroleptic discontinuation, complicating interpretation (1). Asnis evaluated the largest series of patients, 97, but again focused on cogwheeling instead of true parkinsonian signs (58).
The more solid evidence that lithium causes parkinsonism comes from more recent case reports, as summarized by Factor (52). Cogwheeling along with such parkinsonian traits as bradykinesia, rigidity, resting tremor, masked facies, and postural instability solidified the diagnosis of parkinsonism in these cases. All patients were on chronic lithium therapy, some with therapeutic, others with toxic serum levels. One can draw the conclusion based on these cases that the duration of lithium treatment, older age, and higher therapeutic levels of lithium appear to be risk factors for the development of parkinsonism. However, dopamine transporter SPECT scanning was not available for these cases, so it is possible some or all actually had PD.
Lithium’s mechanism of action for improving bipolar disorder is unknown, and the same is true for inducing parkinsonism. Lithium affects many different neurotransmitters and modulates several biochemical pathways. For instance, lithium has been shown to block inositol phospholipid hydrolysis, part of a second-messenger system which has been theorized to explain lithium’s benefit in affective disorder. This may be related to its upstream effect of blocking the stimulus-based release and the enhancement of reuptake of norepinephrine in animal models. Lithium also affects serotonin, GABA, glutamate, and several neuropeptides. Animal studies yield conflicting data for how lithium might cause parkinsonism, with lithium exerting contrasting effects on the dopaminergic system (52).
Causal relationships are difficult to determine when there are many medications given simultaneously. This especially becomes murky when it is against a background of an acute underlying illness. Such is the case when trying to attribute parkinsonism to an agent used in an organ transplant recipient. However, there are several convincing case reports of the calcineurin-inhibitor immunosuppressant agent cyclosporine (also referred to as cyclosporine A) causing a reversible form of parkinsonism (59–63). Cyclosporine is better known to cause action tremor (occurs in approximately 20%), and can also cause cerebellar ataxia, myoclonus, seizures, headache, cortical blindness, visual hallucinations, and encephalopathy (61, 63).
Cases of cyclosporine-induced parkinsonism varied in duration of exposure, ranging from days to years. The clinical syndrome includes rest tremor, bradykinesia, hypomimia, and rigidity. This could develop even with therapeutic serum levels of the drug. Generally resolution was fairly rapid, over days to weeks, upon stopping the medication. The parkinsonian symptoms in these cases sometimes responded to levodopa. Unfortunately, none of these cases included functional dopamine transporter imaging.
Amphotericin B is prescribed for life-threatening fungal infections and frequently used alongside chemotherapeutic agents in immunosuppressed patients with leukemia. There are reports of amphotericin itself causing parkinsonism (64). Amphotericin may have some synergistic effects with the chemotherapeutic agent cytosine arabinoside (65, 66, 67) to cause parkinsonism. This parkinsonian syndrome is characterized by bradykinesia, rigidity, and resting tremor, which can resolve over days to months after discontinuation of the offending medication. Levodopa appears to be of no benefit.