Introduction

While drug-induced parkinsonism is the best understood of the drug-induced movement disorders, it still holds a number of mysteries. It is generally thought that drugs that block dopamine D2 receptors or reduce dopamine transmission cause a syndrome that mimics the motor aspects of idiopathic PD (IPD) (1). However, there are medications, such as lithium and valproic acid, that produce parkinsonism without known effect on dopamine transmission (2,3). Another interesting and unanswered question is the explanation for the extremely wide spectrum of responses to dopamine blocking drugs. Some patients on small doses develop extreme parkinsonism while others on very high doses of similar potency drugs do not (personal observation). While this may be due to early unmasking of subclinical idiopathic PD, there is only mild support for this hypothesis (4). The second generation of antipsychotic drugs, also known as “atypical” neuroleptics, all share the property of dopamine D2 receptor blockade, yet two, quetiapine and clozapine, do not cause parkinsonism, while all the others do (5). An explanation for these observations is debated, with one major theory holding that the ratio of 5HT2-A serotonin receptor blockade to dopamine D2 receptor blockade is the important determinant (6), whereas the second hypothesis maintains that the amount of time the drug actually binds to the D2 receptor, the “fast-off” theory, is the explanation (7). And patients who develop parkinsonism on their neuroleptic often have it resolve without any treatment (1). The observation that “pseudoparkinsonism will often spontaneously remit six to eight weeks after first presenting” (8) is present in much of the early literature on neuroleptic extrapyramidal side effects (9). It has even been suggested that only 1% of NIP cases fail to return to normal after discontinuation of neuroleptic (10) and that these likely suffered from idiopathic PD; yet some patients develop parkinsonism that neither progresses nor regresses after years of being off the offending drug (11). Furthermore, although MIP is frequently treated with anti-PD medications, their efficacy is not clear (12). These issues remain unresolved.

Far and away the most common MIP is neuroleptic-induced parkinsonism (NIP). While it is commonly believed by authorities in the field that the second generation of antipsychotic drugs (SGA), the “atypical” neuroleptics, are less likely to cause this problem than the first generation, this is not supported by data (13–15), although several SGA have been shown to be less likely to cause this than haloperidol (16). Yet no differences were found between haloperidol and risperidone effects in neuroleptic naïve patients (17). The discrepancy between experience and the evidence base may lie in confounding factors inherent in studies of antipsychotic drugs in psychiatric patients, due to the long duration of parkinsonism even after all neuroleptics have been discontinued.

Clinical aspects of MIP

Parkinsonism is defined as an “akinetic-rigid” syndrome. It is a disorder that looks like IPD. Most published studies of NIP are from the early years of neuroleptics. Since most patients on antipsychotics are treated for primary psychotic disorders which tend to be life-long, these drugs usually cannot be discontinued. In addition, psychosis tends to be treated very soon after presentation, often before psychiatrists, or at least attending-level psychiatrists, see the patient, making contemporary studies of NIP difficult.

NIP typically develops over weeks (18) but may develop within days. Onset depends partly on the route of administration, the drug itself, and its dose. Extrapyramidal disorders tend to occur more quickly with injected preparations.

NIP is clinically indistinguishable from IPD in individual cases (4,19). While there are some differences, in general, between the motor syndrome of IPD and medication-induced parkinsonism (MIP), these are statistical in nature, applying to large groups, and cannot be applied to individuals to distinguish one condition from the other. Drug-induced parkinsonism is thought less likely to cause tremor than IPD (18,20,21), and is more likely to be symmetric. However, IPD may be without tremor and may be symmetric. In some reports, NIP was asymmetric in over half the cases (12,22).

No study has compared age and gender matched groups to determine if these statistical differences may be age- or gender-related, since idiopathic PD generally affects older people, and men more than women.

The most commonly employed criteria for the diagnosis of parkinsonism, or a Parkinson syndrome, referred to as “the UK Brain Bank Criteria” (23), require bradykinesia as the fundamental abnormality, plus rest tremor, rigidity, or postural instability. Idiopathic Parkinson’s disease (IPD) is defined as a progressive Parkinson syndrome with unilateral onset, persistent asymmetry, rest tremor, L-Dopa responsiveness maintained for five years, the presence of L-Dopa dyskinesias, and the absence of features that would suggest an alternative diagnosis such as eye movement abnormalities, dementia, corticospinal tract abnormalities, etc. The standard assessment tool in psychiatric studies of parkinsonism is the Simpson Angus Scale (24), which is heavily weighted toward rigidity and does not include bradykinesia. Unlike most reports in the psychiatric literature, no minimum score on a Parkinson assessment tool is used to define IPD in neurological reports. Patients either meet the UK Brain Bank Criteria or not. The feature of facial masking, considered a reflection of akinesia, is not a cardinal feature.

Akinesia is technically different than bradykinesia, but is subsumed under the term for purposes of classification. Parkinson patients move more slowly than normal. It takes them longer to dress, bathe, and perform chores of all types, particularly those involving finger dexterity, like buttoning. On tasks involving rapidly performed repetitive movements they show reduced power (25), similar to myasthenics, but also reduced amplitude, and frequently suffer transient arrests of movements. Bradykinesia, the single required cardinal feature in the UK criteria, is usually the most debilitating problem in early IPD. Akinesia refers to the reduction in spontaneous movements that are typically part of the syndrome. Patients blink less than others, producing a “staring” expression, which may lead, paradoxically, to tearing, when the underlying problem is dry eyes. They smile less. They move less as well, so that they are somewhat like statues. In the waiting room they sit still, not commonly crossing their legs, rubbing their hair, crossing their arms, etc. They swallow less than others, which may lead to drooling. Rigidity is the resistance to passive movement. This can be difficult to assess since it requires relaxation of the limb, something many people have great difficulty achieving.

Patients may or may not appreciate their own rigidity, which they will describe as “stiff.” While this rigidity is often characterized as “cogwheel” because it is often ratchety, like a cogwheel, it often is not. The presence or absence of cogwheeling should not be construed as having any impact on the diagnosis of parkinsonism. The rigidity differs from that seen with spasticity in that it is not rate-dependent, nor is it associated with any change in reflexes. Interestingly, bradykinesia and akinesia, while related, may often be quite dissociated, with a very akinetic patient demonstrating very little slowness, and vice versa.

The tremor of parkinsonism is quite stereotypic, occurring at rest. It may occur with posture holding as well but resolves with movement. Complicating interpretation of tremor is the high prevalence of essential tremor and the occurrence of tremors due to medications. Lithium, in particular, frequently causes sustention and action tremor, independent of parkinsonism, as does valproic acid. Parkinsonian tremor most commonly affects the fingers or hands, usually starting on one side. Although frequently labeled a “pill rolling” tremor, because the tremors in the fingers, particularly the thumb, make it appear as if the affected hand is rolling a pill, the tremor often does not look like this, and may affect the hand rather than the fingers. Sometimes a single finger may be involved, most commonly the thumb, but usually all of the fingers are involved. After fingers and hands, the next most commonly involved body parts having tremor are the jaw and feet. Tremor of the tongue is not uncommon. Head tremor may occur but is very uncommon, whereas foot tremor is far more common in parkinsonism than in essential tremor. Vocal tremor does not occur.

The posture of parkinsonism is stooped. Patients tend to be flexed. They adopt a posture similar to that of an aged person, with flexion of the hip, shoulders and neck. They lose “associated movements,” namely armswing, when they walk. Their stride is reduced in length, and the heel strike becomes less angled. Walking slows. Balance may be impaired, but not usually unless the NIP is severe.

Epidemiology

The reported prevalence of NIP varies enormously, with figures generally ranging from 5%–90% (26). The long-term schizophrenia project from Nithsdale, Scotland, has reported figures varying from 27% (27) in 1992 to 35% (14) ten years later. These studies, while community based, involved small numbers of patients; however, the 20-year follow-up study from 2002 included only 136 patients. Two inpatient studies reported a prevalence of 40% in Holland (28) and 20% (29) in Curacao, both with under 200 patients. A study of 1559 patients in 6 psychiatric hospitals in Italy reported that 66% had parkinsonism, as defined by a threshold score on the Simpson Angus Scale (30). Yet a recent study involving over 1400 patients receiving intramuscular or oral olanzapine for over a year found that only 0.3% had parkinsonism (31). A door-to-door study in Spain found that the prevalence of MIP was equal to that of IPD (32).

Clinical course and treatment

Known risk factors for the development of parkinsonism include the choice of drug, with the “high”-potency neuroleptics having the greatest risk, and the age and gender of the patient. Genetic influences are being explored as well (33). Older people were at greater risk than younger (whereas younger are at greater risk for acute dystonic reactions) and women at greater risk than men (18). The observation that women are at greater risk is counterintuitive as women are typically given lower doses than men, and men are about 30% more likely than women to develop IPD than women.

How quickly parkinsonism develops varies considerably. Even in early reports where the patients had not ever been exposed to neuroleptics the time to onset was quite variable. One author (34) noted onset in the great majority by three weeks. Another report (35) described the majority being affected within the first week. The largest study (18) had 90% of patients affected by day 72,with half the patients affected by 30 days. A common issue for contemporary neurologists is in diagnosing NIP in an upper middle aged or elderly person who has been on neuroleptics for decades. Oftentimes the history suggests that the patient tolerated the drugs without a problem until recently, when the parkinsonian effects developed. The question then is whether the patient has simply become more sensitive to the drug side effects, as elderly people are, or whether the sensitivity is entirely due to the development of IPD.

The natural course of MIP is also unknown. A 1954 report (36) stated that the syndrome resolves, untreated, within 2 months with patients staying on their medication. Marsden also reported that NIP often resolves (37). A 1971 study found that almost half still had tremor after remaining on trifluoperazine 3 years (38). A study of NIP in the elderly (4) associated it with a high death rate and found that most surviving patients became and remained incapacitated. They also (4) found that 2/3 had resolution within 7 weeks of stopping the drug but that 11% did not by 18 months. A report following a cohort of long-term patients in a state psychiatric hospital found that parkinsonism increased over the 10 years of observation (39), with all patients still taking neuroleptics.

How long patients remain parkinsonian when lithium and valproic acid are stopped is unknown (see chapter 12). In general, when neuroleptics are stopped, the syndrome resolves over weeks to months, but prolonged parkinsonism lasting 18 months is not rare (10).

Most reports indicate that NIP is not a risk factor for the later development of tardive dyskinesia (29,39,40), but this is not clear. There are reports that neuroleptic parkinsonism increases the risk of the later development of idiopathic Parkinson’s disease (4,41). A report on NIP in patients seen at a geriatrics clinic in England (4) found that 25% were found to have IPD at 41 months after drug discontinuation. This latter observation seems intuitive, since PD patients are extremely sensitive to neuroleptic side effects.

Recognition

Parkinsonism is a common but underrecognized disorder. Although few studies have assessed this, it is common experience for neurologists to see patients referred for an evaluation only when parkinsonism is advanced, or tremor occurs. In two studies in Rhode Island, one in a nursing home used to teach geriatric medicine to physicians, approximately 50% of patients were found to have significant features of parkinsonism which were not identified in the nursing home medical record, either by the physician or the staff at any level, including physical therapists (55). In many cases parkinsonism was due to medication side effects. A second study by the same group reviewed all neurological inpatient consultations performed in a teaching hospital. All patients had been seen by at least four nonneurologist physicians, and often by more. Again, in only about 50% of patients with significant parkinsonism was the parkinsonism recognized (56). Most of these patients did not suffer from IPD, but the parkinsonism caused or was contributory to impaired gait in all the nursing home residents, and led to falls and hospitalization for many of the acute medically hospitalized patients.