©1997 by Guilherme L. Werneck. All rights reserved
Parkinsonism is a motor syndrome consisting of a group of symptoms—tremor, bradykinesia, rigidity, and postural instability or impaired postural reflexes—that result from striatal dopamine deficiency.1 Idiopathic Parkinson’s Disease (PD) is the most common cause of parkinsonism and is thought to account for more than two thirds of the cases.2 Parkinsonism can result from many other disorders such as: hereditary diseases (e.g. Wilson’s disease), multiple system atrophies (e.g. Shy-Drager syndrome), other degenerative diseases (e.g. Creutzfeldt-Jakob disease), other central nervous system disorders (e.g. tumor), infections (e.g. viral encephalitis), exposure to toxins (e.g. MPTP, carbon monoxide), metabolic disturbances (e.g. hepatocerebral encephalopathy), and the use of medications (e.g. neuroleptics).1,2,3 PD is generally accepted as a clinicopathological entity that is different from other causes of extrapyramidal signs.1 However, since there is no accurate biological marker for PD, the differential diagnosis of parkinsonism can be extremely arduous.1
The reported proportion of cases of parkinsonism induced by drugs varies from 8 to 51%.1,4 The relative frequency of drug-induced parkinsonism may be underestimated since drug reactions in older persons may suggest the presence of a new disease or may simply be attributed to the aging process.5 This misinterpretation is particularly likely when the drug-induced symptoms correlate almost perfectly with those of PD, an illness that occurs more frequently in the elderly.6 In addition, a global underestimation of adverse drug effects in the elderly is likely, since this group is underrepresented in premarketing clinical trials.7 The proportion of drug-induced parkinsonism may on the other hand be overestimated if some patients had previous subclinical idiopathic PD, and the onset of the classical symptoms coincided with the use of the incriminated drug.8 Either underestimation or overestimation can also occur depending on the selection of study patients and the criteria used to establish the putative causal relationship.
In such a complex context it is important to identify what are the specific pieces of evidence linking each incriminated drug with the onset of parkinsonism. Incorrect diagnosis of drug-induced parkinsonism would lead to unnecessary treatment with antiparkinsonian agents, which may not be effective in this situation and have themselves important adverse affects.6,9 If a drug is really responsible for parkinsonism, it should be used carefully in patients at higher risk. On the other hand, erroneously attributing an adverse effect to a drug can delay the start of the specific treatment. In addition, it would discard a possible beneficial drug from the therapeutic arsenal for the elderly.
Parkinsonism as an adverse effect of neuroleptics has been suggested since the early 60’s10 and now it seems to be well-established.11,12,13 For other drugs, such as cinnarizine and flunarizine, initial evidences appeared only in the middle 80’s.14,15,16 Despite the fact that both review papers available2,3 accept a causal relationship between cinnarizine/flunarizine and onset of parkinsonism, none describe the pieces of information that allow one to make such a conclusive inference. Although not available in the United States,17,18 cinnarizine and flunarizine are widely prescribed in some European and South American countries.19,20,21,22 If these drugs really induce parkinsonism, the impact in terms of number of affected people in these countries provides enough justification for evaluating this conjecture.
This paper aims to gather and describe the pieces of evidence incriminating cinnarizine and flunarizine as causal agents of parkinsonism, to examine some validity problems affecting the ability to make causal statements, and to suggest prospects for future investigations.
Cinnarizine and flunarizine are piperazine derivatives with anti-histaminic and calcium antagonistic activity.20,22 Flunarizine is a di-fluorinated derivative of cinnarizine which was marketed at the beginning of the 80s, 12-14 years after cinnarizine.16,17 It has been shown that they have sedative actions on the vestibular labyrinth and decrease blood viscosity.20,22 These actions provided support for their use in the treatment of dizziness, vertigo, migraine and cerebral blood flow disturbances. Nevertheless, clinical trials have produced ambiguous conclusions about their efficacy in treating these disorders.23,24,25 Indeed, these drugs are frequently used as "placebos" in the treatment of poorly defined conditions in the elderly.17 The dose originally recommended for cinnarizine was 7.5-10mg/daily but most formulations today contain 25 to 75mg.21,24
One conceivable mechanism for the adverse effect invokes the similarity of these drugs’ chemical structure to that of the neuroleptic trifluoperazine, which blocks postsynaptic dopamine receptors.22 An alternative hypothesis suggests that calcium channel blockers may act presynaptically by decreasing the content of dopamine-b-hydroxylase, activity which appears to be regulated by protein-bound Ca++-calmodulin.19
Until recently, only a few experimental studies in animals were done to investigate these two hypotheses. In 1986, it was shown that another calcium channel blocker (nimodipine) could reduce the release and synthesis of striatal dopamine in the mouse (presynaptic action).26 In 1992, interactions of cinnarizine with the dopaminergic system at the presynaptic level were reported.27 In 1991, a postsynaptic dopamine receptor blocking effect of flunarizine was described.28
Apparently only one study in humans explored the effects of flunarizine and cinnarizine in the dopaminergic system.22 This study compared the striatal dopamine D2 receptor-binding potential in patients under treatment with flunarizine or cinnarizine with that of age-matched controls. The results showed that D2 receptor binding was reduced in patients taking these drugs in comparison to controls. The authors concluded that, although a presynaptic effect cannot be excluded, it seems likely that the neuroleptic-like action of flunarizine and cinnarizine (blockage of dopamine receptors) may explain their extrapyramidal side-effects.
Although these studies provided some biological support for incriminating these drugs, none correlates the findings with clinical symptoms in disease-free subjects. In the next section we will present the available clinical and epidemiological evidence supporting that cinnarizine/flunarizine can induce parkinsonism.
The earliest report of the association between the use of cinnarizine/flunarizine and the onset of parkinsonism was in 1984.14 Parkinsonism was diagnosed in five subjects taking flunarizine. The next study was published in a Spanish journal in 1985.15 The authors described eleven patients aged 65-83 in whom parkinsonism developed after treatment with cinnarizine at daily doses of 150mg for 6 to more than 36 months.15 In ten patients the withdrawal of the drug was followed by the disappearance of parkinsonism within one year.
In 1986, a new Spanish article reported a double-blind placebo controlled trial of cinnarizine 150mg daily in 20 patients with mildly incapacitating PD.29 During the study the patients remained under their usual anti-Parkinson therapy. There was a significant deterioration in the cinnarizine-treated group when comparing to placebo treated individuals regarding their score in the Webster scale, and in the time taken to turn over bed. There were no changes in the inability scale, nor in the time taken to walk ten meters. Although there was a significant increase in the writing time among the cinnarizine-treated group, there was not any statistical significant difference between groups.
The earliest study in English language was published in June 7, 1986, by Uruguayan neurologists.16 The authors studied 8 women and 4 men (mean age=64.7 years) who presented extrapyramidal motor signs after flunarizine treatment for 3 weeks to 15 months. No patients had taken neuroleptics. The parkinsonian signs "practically disappeared after 20 months after withdrawal of flunarizine."16 Only slight improvement was noticed in dyskinesia and akathisia.
In the same month there was a letter published in The Lancet from a member of Janssen Pharmaceutica, a company which sold flunarizine in Europe at this time.21 He argued that extrapyramidal reactions were unusual at the recommended daily dose and that they occurred primarily in patients with subclinical PD. He pointed out that the extrapyramidal symptoms during flunarizine treatment may not be drug-related since flunarizine is often prescribed to older patients, who are themselves at an increased risk of parkinsonism. He presented no data.
In August 2nd 1986, researchers from The Netherlands wrote a letter to The Lancet reporting one possible case of flunarizine-induced parkinsonism.30 Among 55 reports of adverse effects of flunarizine since 1983, one referred to a patient with depression and parkinsonian symptoms. A 68-year-old men with no previous extrapyramidal symptoms developed parkinsonism after one year of drug use (10mg/day) "to prevent cerebral hypoxic damage."30 Flunarizine was discontinued and over a period of several months all symptoms disappeared.
In the same month another letter in The Lancet reported findings from a survey on senile dementia carried out in the Republic of San Marino (June 84 to July 85).31 Twenty-one persons out of 398 examined and interviewed were on treatment with flunarizine. Seven of them presented depression and bradykinesia mimicking PD, but none had rigidity or rest tremor. Flunarizine was withdrawn from 6 of these patients. Four of them were put on amitriptyline, one started amineptine and two remained untreated. Those given amitriptyline recovered completely after three months, but the others did not.
In October of the same year a communication from Spanish researchers was published in The Lancet.17 Using two different spontaneous reporting systems operating in Spain, they were able to identify 14 reports of extrapyramidal symptoms associated with either cinnarizine (8 reports) or flunarizine (6 reports). The duration of treatment with these drugs varied from 2 days to 4.5 years. The range of daily doses varied from 45-180mg for cinnarizine (median=150mg), and from 10-20mg for flunarizine (median=10mg). The effect of dechallenge was assessed for 12 subjects, and improvement occurred in seven (58%) of them.
In 1987 the effect of cinnarizine on motor function was evaluated in a randomized double-blind study of 20 patients with a diagnosis of PD.20 After an initial basal clinical evaluation, patients were randomly allocated to two treatment groups—cinnarizine (75mg/twice a day) vs. placebo with identical formulation and prescription—and followed for three months. All patients were under antiparkinsonian treatment. At the end of the study, there was a significant worsening of mobility and time for turn in bed in the group taking cinnarizine. No modification in the time required to walk 10 meters and in the time to write a sentence were noted. No significant variation was detected in any response in the placebo group.
In 1988 another study from the Spanish group of researchers at Barcelona used spontaneous reporting systems of adverse drug reactions to investigate the role of cinnarizine/flunarizine in the onset of parkinsonism.32 Seventy reports associated parkinsonism or tremor with the use of cinnarizine. Of those with parkinsonism (n=39), twenty-four (62%) recovered after the drug withdrawal, 7(18%) did not recover, and for 8 patients information was missing. Among those with tremor (n=31), 18(58%) recovered after the withdrawal of the drug, 6(19%) did not recover, and no information was available for 7 patients.
Dechallenge of the drug was also used as the main argument in favor of the drug-induced parkinsonism hypothesis in an Argentinean study.33 The authors studied a series of 101 patients (mean age=69.1) who developed abnormal movements secondary to treatment with cinnarizine and/or flunarizine. Remission of the symptoms occurred in 94 patients after discontinuation of the drug(s).
Diagnostic criteria for cinnarizine-induced parkinsonism (CIP) were proposed in 1991.19 A diagnostic on CIP should be established if a patient had: "(a) presence of at least two cardinal signs (tremor at rest, rigidity, bradykinesia, and postural instability); (b) no concomitant administration of other drugs potentially capable of causing parkinsonism; (c) complete disappearance of any sign of parkinsonism following withdrawal of the offending drug after at least 1 year of follow-up; and (d) no reappearance of parkinsonian symptoms after withdrawal of anti-parkinsonian medication [...] after three months."19 The authors conducted a study with 24 cases that fulfilled these criteria. They used these cases to investigate differences in the age of onset, disease duration and family history of tremor in a case-control study. Controls were PD patients routinely attending the out-patient clinic (group 1), and hospitalized nonneurological patients (group 2).They found that the age of onset of parkinsonism in the case group exceeded that of PD patients by 10 years. Referral to that specialized clinic were prompt in the case group (mean=1.7 years) vs. a mean of 5 years in the PD group. A history of either tremor alone or tremor and PD were significantly more prevalent in the CIP group. The authors conclude that aging plus a background of genetically determined essential tremor are risk factors for developing CIP.
Another study investigated a possible role of inherited susceptibility in cinnarizine/flunarizine-induced parkinsonism (CFIP).34 They defined a case of CFIP based on the presence of two of the cardinal symptoms of parkinsonism plus, "(1) absence of personal history of extrapyramidal disorders preceding the treatment with any of the two drugs, (2) onset of symptoms in the course of treatment with cinnarizine and/or flunarizine, (3) no concurrent treatment with other drug known to produce extrapyramidal syndromes."34 Age-matched spouses of PD patients were taken as controls. All cases were women in contrast to only 24% of the controls. Prevalence of referred PD and/or essential tremor were five times higher among the relatives of CFIP patients than among relatives of controls.
Although the evidence brought out by these studies seem to be convincing, there exist many methodological issues to be addressed before taking it for a fact.
The experimental studies suggested two biological mechanisms for CFIP: (1) a presynaptic effect (e.g. decreasing dopamine synthesis and/or release) and, (2) a postsynaptic effect (e.g. blocking dopamine receptors). The only human study available for evaluation claimed to show blockage of dopaminergic receptors by flunarizine and/or cinnarizine. It is not exactly true. What they did show was that patients taking these drugs had a reduced D2 receptor binding potential in comparison to controls. Indeed, half of the patients had already parkinsonian signs at the start of the study, which can explain their poor performance in comparison to healthy controls.
Most evidence comes from case series studies. Cinnarizine and flunarizine are mainly prescribed to patients at higher risk of developing extrapyramidal signs (e.g. elderly). Therefore, it is not surprising that a relative high percentage of patients with parkinsonism signs were taking cinnarizine and/or flunarizine in countries in which these drugs are widely prescribed.
Dechallenge in case series studies was the most common way of inferring causality. The rationale is that the disappearance of parkinsonian signs after removal of the drug is a strong evidence for a causal relationship. However, in some studies the disappearance of symptoms actually occurred under treatment with other drugs.31 Interpretation would be difficult even if the remission was evaluated under no other concurrent treatment. For instance, it is plausible to assume that most patients were also taking an antiparkinsonian agent (e.g. levodopa). Levodopa can cause dyskinesias in 80% of the patients after 1 year of treatment.9 Disappearance of symptoms after removal of cinnarizine/flunarizine could simply be related to the simultaneous suspension of levodopa.
The few controlled studies also had flaws. The two double-blind placebo trials involved only 20 subjects each. Point estimates can be very unstable in this situation. In addition, the study subjects were PD patients. It turns out that one cannot test whether these drugs cause parkinsonism in this setting, but only worsening of symptoms. One of these studies have failed to find association for some symptoms although it did find it for others. Interestingly they have found significant results only for those symptoms in which the groups were clearly different in their means at baseline (not significantly different in statistical terms, but with such lack of power to detect differences...). In addition, the cinnarizine group was taking higher doses of levodopa which may had caused some of the motor function worsening.
The last two studies compared PD patients with a group of individuals supposed to have CFIP. They found some difference in the age of onset of disease and in the prevalence of hereditary factors. This may suggest that these are really two different disorders, which acts in favor of the drug-induced hypothesis. On the other hand it may only reflect differences in the presentation of PD in two groups selected by different characteristics. In addition, in one of the studies the groups were clearly not comparable in terms of gender.34 Women may better recall information of health status of parents. Apart of age, confounding factors were never taken into account in these studies.
It is not to say that there is no evidence for parkinsonism induced by cinnarizine and/or flunarizine. There is sufficient evidence for being careful in using these drugs in the elderly, especially because they are not essential drugs and their efficacy is not well-established. Clearly there is a need for further research on this subject. A first approach should be a case-control study to investigate whether the incidence of antiparkinsonian therapy changes with exposure to cinnarizine and/or flunarizine. In general, cases would be patients receiving anti-parkinsonian drugs. Controls would be individuals coming from the same source population giving rise to cases but who had not been prescribed an antiparkinsonian drug. This approach was used is other studies investigating drug-induced parkinsonism by neuroleptics and metoclopramide.6,12
Return to the DDIL Homepage
This
document has been provided through the Drugs and Devices Information Line,
http://www.hsph.harvard.edu/Organizations/DDIL/ddil.html