Hallucinations in Parkinson disease in the prelevodopa era

Hallucinations occur in approximately one third of patients with Parkinson disease (PD) and may affect up to 40% in hospital-based series.¹ They include complex visual hallucinations, illusions, a sense of presence, and, less frequently, auditory, tactile, or olfactory sensations. Hallucinations may occur either as part of a confusional state or, more frequently, in the context of a clear sensorium with a chronic course.² Hallucinations in PD received growing attention when it became clear, starting from the 1970s, that chronic dopaminergic treatments (levodopa and, later, dopaminergic agonists) could trigger their onset. Hallucinations were thus commonly considered to be a drug-induced phenomenon, leading to designations such as “dopaminomimetic psychosis” or “levodopa psychoses,”³ although nondopaminergic drugs (e.g., anticholinergics) could induce similar complications. In recent years, reconsideration of the role of disease-related factors has identified cognitive impairment and possibly depression and age as contributory.¹ However, whether chronic hallucinations belong to the natural history of untreated PD remains undetermined. This issue has important pathophysiologic and possibly therapeutic implications. Studies of the natural history of PD in the prelevodopa era could be potentially useful to this issue because the confounding factors of dopaminergic stimulation and drug-treatment duration would have no influence. We therefore reviewed the early seminal clinical descriptions of PD from the 19th and early 20th century English-, French-, and German-language literature to examine descriptions of hallucinations in the prelevodopa era.

The “Charcot dogma”: Mental symptoms are end-stage behaviors or coincidental occurrences.

In his 1817 monograph An Essay on the Shaking Palsy, James Parkinson described delirium as a late feature of the disease but did not specifically mention hallucinations:

. . . and at the last, constant sleepiness, with slight delirium, and other marks of extreme exhaustion, announce the wished-for release.

Charcot, in an early article cowritten with Vulpian (1862),⁴ also mentioned the possible occurrence of cognitive impairment and delirium in the final phases of the disease:

Thus, a global decline and most markedly a loss of mobility and deterioration in memory and cognition show that the deficits are becoming more and more progressive . . . It hardly needs emphasizing that death is near when encephalopathy, a comatose state and incontinence develop and when sacral ulcers occur in a long-term patient reaching a severe stage of the illness.

Later, Charcot (1880) did not vary.

. . . then, sooner or later, the phase preceding death supervenes, known as the “terminal period” . . . At a given point, cognition blurs and memory fails. There is generalized weakness, patients become demented, and bedsores develop.⁵

At the same time, Vulpian explicitly denied that delusions or hallucinations could occur in PD.

Intelligence remains intact to the very end, so that no confusion, delusion or any sort of hallucination is evident.⁶

Gowers also recorded the possible occurrence of late dementia and here introduces an observation on delusions, although not specifically hallucinations:

The intellect may be unaffected throughout, except the irritability which usually accompanies the physical restlessness, or by mental depression (. . .). Pronounced mental symptoms are occasionally present, however, in the later stages of the disease, especially mental weakness and loss of memory, and they may occur early in its course. Very rarely, they are accompanied by a tendency to delusions and occasionally they amount to actual dementia.⁷

Charcot’s views were echoed by Wilhelm Erb, a prominent clinician from Heidelberg and a leading figure of early German neurology. In patients with paralysis agitans, he wrote,

mental and sensory functions remain completely intact . . . Intelligence and memory functions seem undisturbed. I have seen many of these patients doing big business without any problem; they are mentally fit. Despite the severity and hopelessness of their illness they often appear humorous and in good mood. Evidently, however, there are some exceptions of this rule.⁸

Despite extensive clinical descriptions on cardinal signs of disease progression, Charcot and Gowers never explicitly mentioned hallucinations in the course of PD. Charcot’s followers did allude to hallucinations, but only to emphasize their rarity, as illustrated by the following examples. Brissaud, the interim successor of Charcot at the Chair of Neurology at La Salpêtrière, reported the “rarity of hallucinations,” with no further comment, in one of his Leçons sur les maladies nerveuses.⁹ Souques, a late pupil of Charcot, was more specific in his 1921 report, emphasizing a lack of association:

As for the classic mental disorders, cognitive decline that covers the gamut from simple intellectual blunting to profound dementia and severe psychiatric disorders (depression, suicidal ideation, hallucinations etc), Parant, Ball and others have discussed them at length. However, these problems do not belong to paralysis agitans itself; when they occur, they are mere coincidences.¹⁰

German physicians remained equally skeptical. Reviewing the literature on psychological disturbances in Paralysis agitans, Runge, the author of a voluminous book chapter in a Handbook of Psychiatry, summarized his experience as follows:

Due to the additional occurrence of delusional ideas and sporadic hallucinations, depressive states in patients with Paralysis agitans sometimes do not appear in a “pure” form. But a thorough examination of a great number of patients with Paralysis agitans reveals that such symptoms (i.e., delusions and hallucinations) occur rather infrequently. Among our own 60 cases not a single one presented symptoms of paranoia. Acoustic hallucinations are predominant, but there are also cases of haptic hallucinations presumably due to abnormal bodily sensations . . . The existence of these clinical features cannot exclusively be attributed to Paralysis agitans.¹¹

In a similar manner, alluding to unidentified “early reports,” Gamper concluded, in the famous German Handbook on Neurology:

Whether hallucinations (in PD) turn up as frequently as early reports suggest (“Paranoia hallucinatoria,” “Delirium hallucinatorium,” etc.) is very questionable.¹²

Such a point of view was perpetuated by most authorities up to the 1940s, as exemplified by this quotation from Wilson:

Psychical symptoms are mostly limited to depression and irritability, the natural outcome, perhaps, of an incurable ailment; anything beyond these is to be ascribed to accompanying arteriosclerosis or some incidental condition.¹³

As a group, these early authors clearly considered that hallucinations did not belong to the natural history of PD and, although they did occasionally occur, they were either coincidental or secondary to the severely debilitated condition of late-stage disease.

Hallucinations do occur in particular instances.

Ball and Parant, the two authors pointedly criticized by Souques (above), countered the “Charcot dogma” starting with Ball’s seminal report “On insanity in Paralysis Agitans” published in 1882. Ball, the first recipient of the Chair of Psychiatry at the Faculté de Paris and Chief of the Sainte Anne Hospital of Paris, reported seven PD patients with various mental disorders, one of them having clear-cut hallucinations.¹⁴ The patient, aged 53, had had PD for 5 years with typical motor symptoms but also with distinctive behavioral disturbances:

The patient is in fact emotionally hypersensitive; he cries easily and his overall disposition tends toward the depressive. However, he also displays other more clearly defined psychiatric disorders. In the last six months, he has been hearing voices that insult and threatened him and even on occasion verbally abuse him. He perceives these enemies who surround his bed and place his life in jeopardy. In response, he violently trembles and tries to escape. These episodes only occur at night . . . He is absolutely convinced of the reality of his hallucinations and, more than once in the midst of this obsession, he tried to kill himself. His memory has been considerably affected and his intellectual faculties have definitely declined.

After a few months, while treated with hyoscyamine and atropine, the patient improved.

His pain and night-time hallucination have resolved. The patient is less depressed and no longer mentions his imaginary persecutors.

Ball offered the following conclusions:

1. Mental problems accompany paralysis agitans more frequently than commonly thought.

2. These always take the form of depressive symptomatology, most often a circumscribed depressive delusion [lypémanie] accompanied by suicidal impulses and by multiple hallucinations, but in some cases, dementia and mild stupor predominate.

3. They are almost always intermittent and seem to intensify in parallel with an aggravation of motor symptomatology, remitting when tremors subside.¹⁴

Soon after Ball, Parant, a psychiatrist in Toulouse (France), reported the observation of a patient who had development of parkinsonism and, several years later, dementia and psychiatric disturbances.¹⁵ The patient, of unspecified age, had had PD for approximately 12 years. Five years after disease onset, the patient had strange somatic hallucinations or illusions, including the “Alice in Wonderland” syndrome:

During the same period, he had strange delusions, often believing that he had more than two legs. He felt one in front of his body and a forth one behind him. Or, he felt his legs elongating in extraordinary proportion to the point that they could reach a wall located several feet away from him.

Eight years after onset, delusions developed, with themes of jealousy, persecution, and robbery.

When M. X was alone with his wife, he would claim that a third person, his enemy, had just positioned himself between them. He was quick to name this enemy who was threatening to kill him. Later on, several other people would appear to harass him and steal his possessions. Often he would believe he saw his wife surrounded by swindlers, thieves, and other shady characters who would torment him and order him to gamble on the Stock Market.

These multimodal hallucinations and delusions developed on the background of a cognitive impairment progressing to a “well characterized dementia.” In his comments, Parant made many references to Ball’s article but insisted on the fact that the psychiatric disturbances were in this case unremitting and associated with dementia.

In his textbook, Régis, a professor of psychiatry in Bordeaux (France), focused on nocturnal delirium with dreamlike qualities (“oniric” state).

However, a point not yet emphasized and one that I consider to be perhaps the most significant in regards to behavioral pathology is the mental confusion with dream-like delirium that occurs in paralysis agitans. Already in the clinical description by past authors . . . dreams, nightmares, visual hallucinations that are terrifying or linked to their work, and nighttime confusion have been noted in regards to parkinsonian patients. I have personally witnessed this very typical dream-like delirium in Parkinson’s disease that resurfaces episodically as the disease itself advances, and John Punton from Kansas City (1903) also published a very interesting and clear-cut case of Parkinson’s disease-related mental confusion with depressive symptoms and terrifying nighttime hallucinations.¹⁶

However, Régis was reluctant to ascribe these disturbances to PD itself.

It is possible that these confusional states and dream-like delirial episodes are due, not to Parkinson’s disease itself, but instead to a co-morbidity. In particular, arterial sclerosis occurred in all of my own patients and was also present in the case described by John Punton.¹⁶

Similar claims were made in Germany, but in an important article, the young psychiatrist König decidedly opposed the categorical view that mental disturbances in PD were “completely incidental.” Based on six personal cases from the psychiatric department of the University of Kiel and a thorough review of the literature, he reached the following conclusions (1913):

1. In many cases, an abnormal state of mood is part of the symptomatology. Frequently, a hypochondric depression can be noted, regularly with a tendency toward irritability and delusional ideas of injury.

2. In certain cases, the aforementioned mood disorders can develop into clear-cut psychosis, either of hypochondric and melancholic or of paranoiac nature, including sporadic hallucinations or suicide intent.

3. More often, however, a purely incidental combination with senile dementia, arteriosclerotic dementia or a prefinal delirium has to be taken into consideration.¹⁷

For König, hallucinations and other psychotic symptoms were direct manifestations of the implicit depressive state of PD and could not be considered as incidental comorbidities.

And that was my point: to demonstrate that in patients with Paralysis agitans there is practically no gap between their peculiar mental state and the mental disorders they occasionally show; that it is wrong to believe that the former is an integral part of the disease and the latter an incidental complication.¹⁷

In the 1920s, other authors discussing PD emphasized the possible occurrence of depression, sometimes with “psychotic” features. Jackson et al.¹⁸ reported on the “psychic manifestations” of paralysis agitans in five patients. Although published after the onset of the von Economo encephalitis pandemic, this work dealt with patients with PD, though in one of them (Case 4), an alternative diagnosis such as Lewy body dementia might today be considered. The authors insisted that “mental symptoms may be the first to make their appearance and may precede by months or years the classic neurologic signs” of PD. They described a characteristic “psychosis,” “the most marked involvement being found in the fields of perception, judgment, and affectivity.”

The mental trend is always depressive and the thought content is dominated by false percepts and concepts. Hallucinations are frequent; they are limited to the organic sensations and tactile sense, and are generally the result of sensory and vasomotor disturbances (. . .); on these false percepts are developed well defined delusions of somatic, self accusatory or paranoid nature.¹⁸

These five patients all experienced recurrent or chronic depression with psychotic features, apparently mood-congruent, developing in some cases years before the onset of parkinsonism and thus having a questionable relationship with PD. In only one patient (Case 3) were hallucinations reported, and they were exclusively auditory.

In this context, the work of Patrick and Levy is of value, because these authors carefully studied 140 patients with PD, excluding postencephalitic cases.¹⁹ The authors identified mental symptoms in 34% of cases, once again “largely in the form of depressive reaction” and “as frequently before as after the onset of the disease.” However, only three patients had “mental slowing-up, confusion, memory defects,” with no explicit mention of hallucinations.

One author who found a considerable rate of mental disturbances and hallucinations in PD was the Berlin neurologist and neuropathologist Lewy. According to him, psychopathologic syndromes in PD patients (n = 27) could be classified into three types: with delusion and hallucinations (n = 13), with derangement of the capacity to register and severe long-term memory failure (n = 2), and with a combination thereof (n = 12).²⁰ As to the characteristics of the hallucinations, Lewy added:

Only in 3 patients delusional ideas were not related to hallucinations. The other patients heard voices insulting them and telling them unpleasant things, on which the patients themselves based their delusional ideas. On the whole, hallucinations could be identified with certainty in 7 (out of 27) patients; in 5 cases, the hallucinations involved the acoustic field, in 2 cases the acoustic and the visual field, and in one perhaps the gustatory field. Acoustic and visual hallucinations were not very marked, and patients showed little emotional involvement with them.²⁰

As a group, these articles demonstrate that hallucinations in various modalities could occur in PD patients under specified instances, mainly depression, dementia, and delirium due to a variety of causes. However, given that articles specifically focused on behavioral aberrations and yet, except for Lewy’s report, infrequently reported hallucinations, it is likely that the problem, even in these circumstances, was not typically encountered.

From the von Economo encephalitis epidemic to the dopa era.

The interest in behavioral and psychotic symptoms in parkinsonism increased dramatically after the worldwide epidemic of von Economo encephalitis (1916 to 1927). The prevalence of hallucinations in postencephalitic syndromes has been considered to be much higher than in PD,²¹ but prospective comparative data are lacking. Hallucinations could occur in the acute phase of the disease or in the chronic phase, because patients surviving the encephalitis often presented with a wide range of sequelae, including parkinsonism and a variety of mental disturbances. In a series of 83 postencephalitic patients, Duncan, however, only recorded one case of hallucinations and delusions occurring during the early phase of the illness.²² In another study devoted to the “mental symptoms in epidemic encephalitis,” Folsom summarized his findings:

A consideration of the psychiatric clinical picture presented in this material permits a division into four groups: sixty-three cases characterized essentially by emotional instability; six cases of uncomplicated deliria; one case consisting of an acute hallucinatory paranoid episode; and one dementia praecox syndrome.²³

These retrospective studies may have underestimated the frequency of hallucinations, because delirium during the acute phase of the disease may have included these behaviors as part of the global mental derangement. von Economo stated that “delirium and other psychotic conditions occur commonly in the initial stage, and may usher in the other symptoms of the disease.”²⁴

During the chronic phase, hallucinations and delusions could be present, although they were not among the most frequent psychiatric features.²⁵ They were seen as a feature of several clinical syndromes but most predominantly as part of delirium and “hallucinatory-paranoid states with intact consciousness and without confusion.”²⁴ The studies specifically devoted to hallucinations and delusions in this setting developed into two main directions. Some authors, in France, insisted on the “oniric” or dreamlike states, elaborating on analogies with the recently described peduncular hallucinosis.²⁶ For German and Swiss authors, the main point was to differentiate these psychic disturbances from those of schizophrenia.^27–29 In this setting, Runge emphasized clinical characteristics helping to differentiate postencephalitis hallucinations from those occurring in other instances:

1. Patients (with encephalitis lethargica) do not look astonished or frightened because of their hallucinations. On the contrary, mostly they take them for granted. There are no doubts about the reality of the hallucinations.

2. In most of the cases visual and tactile hallucinations are prevailing. Acoustic hallucinations are less frequent.

3. Complex visual and multimodal hallucinations also occur, but are less distinct and less frequent than in patients with alcoholic delirium.

4. The content of the hallucinations changes rapidly over time.

5. Complex visual hallucinations relate to the patients’ situation in the hospital ward or to their professional life. There are also a variety of non-visual hallucinations that patients may occasionally perceive.

6. As a rule, the hallucinations do not terrify the patients; on the contrary, the hallucinations represent every-day situations, or consist of animal visions.

7. Moving hallucinations are very frequent.

8. In a great number of cases, it can be observed that hallucinations may be induced or influenced by suggestive actions of the physician.¹¹

From the 1920s to the early 1970s, the scarce literature on mental symptoms in the course of PD is difficult to interpret, because several authors pooled the patients with PD and those with postencephalitic parkinsonism.^21,30–32 However, a few studies explicitly focused on patients with PD. Mjönes, in his investigation of 194 patients with PD, observed that approximately 40% of the patients had mental symptoms, which he subclassified into “reactive” (mainly depressive symptoms) and “organic” types (cognitive impairment).³³ Mjönes only mentioned hallucinations in two patients, and both had development of haptic hallucinations. One of them had a genetic form of parkinsonism. In another study of the natural history of PD, hallucinations were not recorded, although the possible occurrence of dementia was acknowledged.³⁴ However, in a study focusing on psychiatric symptoms in parkinsonism, Mindham found a high rate of hallucinations in patients with paralysis agitans (n = 36): 19% had auditory hallucinations, and 11% had visual hallucinations.³⁵ These high figures likely represent a preselected patient population because 89% of the sample was depressed and more than a quarter were designated as confused.

Further historical issues.

Mental disorders, including hallucinations, seemed to occur more frequently in PD after l-dopa became available,³⁶ and there is no doubt that dopaminergic treatment is associated with a high prevalence of hallucinations. Estimates in current studies still vary widely depending on patient population and methodology, but an overall estimate of approximately 33% is reasonable.¹ However, pharmacologic treatments were available during the prelevodopa era, and some of them may have been occasionally responsible for hallucinations. As early as the 19th century, opium, ergot products from rye, and belladonna were commonly prescribed, in particular atropine and scopolamine, two natural belladonna alkaloids.³⁷ Charcot also advocated the use of hyoscyamine, another alkaloid extracted from various plants of the Solanaceae family.³⁸ Most of these products had atropinic properties and could therefore be responsible for delirium and hallucinations. Synthetic anticholinergics were introduced in the 1950s and soon became another—if not the main—cause of hallucinations, especially in patients with coincident cognitive impairment.³⁹ Contemporary prospective studies of hallucinations in PD have included patients who, in almost every case, were receiving antiparkinsonian drugs. However, in a recent study devoted to ophthalmologic features in early, untreated PD patients with normal visual acuity, one fourth of the 30 patients had formed hallucinations.⁴⁰ None of these patients had development of dementia after 2 years of follow-up. This surprisingly high rate supports the view that hallucinations may occur spontaneously in the course of patients with PD.

Another limitation in the interpretation of early studies is the absence of dementia with Lewy bodies (DLB) in the neurologic nosology. First described in isolated case reports in the 1960s and the 1970s, DLB later took its place as the second most common form of dementia after AD.⁴¹ The clinical diagnosis relies on the presence of dementia, cognitive fluctuations, visual hallucinations, and parkinsonism. Visual hallucinations occur in 60% to 70% of patients. It is likely that some patients reported in the early literature as having PD and dementia would currently receive a diagnosis of DLB. However, dementias developing in the course of PD and DLB share many clinical, neuropsychological, and neurobiological characteristics, and the clinical distinction between the conditions is based on an arbitrary temporal criterion: a diagnosis of DLB is made if parkinsonism developed less than 1 year before (or after) the onset of dementia.⁴² It has been suggested that both conditions belong to a same spectrum of Lewy body disease. Interestingly, parkinsonism may develop late or remain absent in the course of DLB, and not all patients receive antiparkinsonian agents. It has been shown that hallucinations commonly develop spontaneously in these patients, demonstrating that taking antiparkinsonian drugs is not a necessary condition for the hallucinations to develop.⁴³

Although direct comparison of the frequency of hallucinations before and after the development of dopaminergic treatment is not available, historical documents demonstrate that hallucinations, mainly of a visual or auditory nature, did occur in patients with presumed PD before the levodopa era. Hallucinations developed primarily in association with depression, dementia, and confusional states. The lack of prospective studies (cause of underestimation) and the wide early use of nondopaminergic drugs, mainly anticholinergics (cause of overestimation), preclude an accurate estimate of the prevalence of hallucinations in the natural history of untreated PD. However, the typical hallucinatory syndrome occurring within a clear sensorium and in the absence of a major depressive disorder, as encountered nowadays in patients surviving for many years with PD and receiving chronic dopaminergic therapy, cannot be clearly identified in any of the early reports. However, contemporary reports of spontaneous occurrence of hallucinations in the course of untreated parkinsonism and, above all, DLB demonstrate that structural and functional changes in Lewy body diseases may, in some cases, be sufficient to trigger hallucinations. These data fit contemporary models of hallucinations in PD in which dopamine exerts a facilitating effect that acts in conjunction with disease-related factors.⁴⁴