Apart from damage or dysfunction of the CNS visual hallucinations may also arise from a pure peripheral pathology caused by lesions of the optical nerves or an ocular pathology as in macular degeneration, retinopathy, or cataract. This association of impairment of peripheral vision and complex visual hallucinations in aged psychologically normal people is called Charles Bonnet's syndrome. Typically there are no concomitant psychotic symptoms and the patient is usually aware of the unreality of his experiences. However, despite a widespread agreement about hallmarks of the phenomenology, a universally accepted definition has not been found yet.1

Little is known about the underlying pathophysiology. A widely accepted hypothesis postulates a reduced afferent input causing a “release” with disinhibition of engrams in the visual association cortex that are experienced as hallucinations. Indeed a recent fMRI study has shown an increased activation of the ventral extrastriate cortex in Charles Bonnet's syndrome.2 Some authors suggest that hallucinations in the syndrome may share a common mechanism that also evokes hallucinations in some central disorders (for example, after infarction of the visual cortex) involving cholinergic and serotonergic pathways.3

Although often neglected or misdiagnosed in clinical practice4 a peripheral visual pathology seems to be an important differential diagnosis of complex visual hallucinations. In a large study of 500 visually handicapped patients Teunisseet al 5 found a prevalence of Charles Bonnet's syndrome of 11%. The occurrence of the syndrome was significantly associated with older age (>64) and a severe impairment of visual acuity (<0.3 in the best eye).

Therapeutic options for Charles Bonnet's syndrome still remain poor and of uncertain benefit for the individual patient. Even without any intervention in some patients the hallucinations can fade away within a few weeks or months. However, there are also many reports of a continuous course with ongoing tenacious hallucinations for up to 8 years.6 An improvement of visual acuity—for example, after cataract extraction—or also a deterioration can eliminate the hallucinations.7 Many of the widely used psychotropic drugs such as benzodiazepines, antidepressant drugs, or classic neuroleptic drugs have not been effective. Only a few reports exist of successful pharmacotherapy, with carbamazepin, valproate, melperone, or cisaprid.8-11 Also, non-pharmacological strategies based on reassurance and self education can be helpful.1 For instance, some patients reported an influence of intensive thoughts. Even admission to hospital interrupted the hallucinations in some cases; however they recurred after discharge.

We describe a patient with a 2 year history of Charles Bonnet's syndrome with macular degeneration, with persistent and frequent hallucinations that have disappeared after treatment with gabapentin.

The 86 year old woman had a 2 year history of complex visual hallucinations on being admitted to our hospital. A senile macular degeneration had been diagnosed by her ophthalmologist 10 years previously. She complained of a daily and repetitive occurrence of images predominantly showing human beings such as medieval women and knights in bright colours, but also torsos or isolated heads. None of the faces were familiar to her. They were of realistic size, coincided with normal perception of the external space, and mainly emerged when looking at a wall or lying supine facing the ceiling. The hallucinations were exclusively of a visual nature and static, but moved when she moved her eyes. They never occurred when her eyes were closed. She also sometimes experienced hallucinations of tiny homunculi strolling on the floor and climbing on her legs when she tried to step on them. Rarely, the content of the hallucinations changed while being watched—for example, from a female to a male head. The duration of the phenomena ranged from seconds to a few minutes. The patient recognised an increase in hallucinations during exhaustion or inflammatory diseases with raised temperature. A condition which regularly evoked hallucinations was using a mobile phone. A complex pattern of rhomboid shapes emerged with a short latency and faded away soon after having switched off the phone. The patient had full insight into the non-realistic nature of her experiences and she did not feel distressed by them. None the less, she argued that the hallucinations sometimes interfered with perception when she was driving a car; therefore convincing her to seek therapy.

We asked the patient to document the time and content of the hallucinations throughout the day in a pretreatment diary. It showed that they were most likely to occur in the morning and in the evening. The most prevalent features were parts of human beings, predominantly heads. Sometimes, objects such as old fashioned clocks or tombstones were described.

There was no psychiatric history. Medical history showed no diseases apart from hypertension and a severe polyarthrosis. She only irregularly took an antihypertensive medication (an angiotensin II antagonist and a diuretic) and pain killers (tilidine). Her general practitioner also prescribed Ginkgo biloba extract and pentoxifylline to treat the hallucinations, but without any effect. On admission to our hospital she was only taking homeopathic medication.

Neurological examination was normal. A dry atrophic macular degeneration was confirmed by our ophthalmologist. Visual acuity was 0.4 in the left eye and 0.6 in the right eye, without perimetric signs of scotomas. No cognitive dysfunction (mini mental state examination 29/30, above average performance in testing alertness, and selective attention) or additional psychotic symptoms could be found. Laboratory tests were normal. Her EEG and brain SPECT disclosed no pathology. Cranial MRI only showed an age related circumscribed frontal atrophy but no abnormalities in the brain.

We started pharmacotherapy with gabapentin (300 mg/day). The patient reported only one hallucinatory event on each of the next 2 days. After that, the hallucinations disappeared; confirmed in a follow up examination 3 months later. The medication has been well tolerated without any side effects. There was no visual deterioration, confirmed by an examination by her own ophthalmologist at this time.

In our case report the complex visual hallucinations, both normal sized and “Liliputian”, accompanied by full insight as well as preserved cognitive skills without a specific brain pathology in morphological and physiological studies, fit in well with the typical clinical picture of Charles Bonnet's syndrome. Also the increase during the morning and evening has been described reminiscent of the emergence of hypnagogic hallucinations in normal subjects. It is noteworthy that visual acuity was less impaired than usually reported in patients with Charles Bonnet's syndrome, which emphasises that a severe loss of vision is not a necessary condition.

As already mentioned, the natural course of the hallucinations differs greatly between patients with Charles Bonnet's syndrome. It sometimes only covers a short period, with a spontaneous remission. Therefore, a therapeutic approach seems not to be necessary for all patients. None the less, an effective pharmacotherapy is needed for those with highly frequent and chronically ongoing hallucinations that are not responsive to non-pharmacological interventions and cause a considerable impairment of daily life. This reflects the situation in the patient presented, who came to our hospital to get an efficient therapy as she had a 2 year history of chronic hallucinations which had not responded to previous interventions. Facing the paucity of data in the field of therapeutic options we eventually considered gabapentin to be a likely favourable drug for treatment of Charles Bonnet's syndrome for the following reasons:

(1) Anticonvulsant drugs in general may influence abnormal neuronal excitations caused by release mechanisms. This is supported by two reports of an effective treatment of Charles Bonnet's syndrome with carbamazepin and valproate.9 10

(2) There is a wide non-epileptic use of gabapentin as well as of conventional anticonvulsant drugs including therapy of peripherally caused “phantom pains”, possibly having similar pathophysiology in another modality.

(3) Compared with the above mentioned drugs gabapentin seems to have fewer side effects (for example, compared with carbamazepin or neuroleptic drugs, which often cause marked sedation or cognitive impairments) and fewer interactions with comedication, providing a safer application, especially in the predominantly elderly Charles Bonnet's syndrome population. The properties of gabapentin require less time for increasing the dosage as in many other antiepileptic drugs and therefore can possibly shorten the period of stay in hospital. However, there are conjectures that GABA related anticonvulsant drugs may cause visual field defects, which might be of interest especially in the already visually impaired patients with Charles Bonnet's syndrome. The exact action of gabapentin on neuronal systems has not been worked out but probably involves multiple mechanisms, apart from GABA.12 By contrast with many reports on vigabatrin, there has not to our knowledge been substantial evidence for a causal association between visual field defects and gabapentin, although transient tritanopia and critical flicker fusion paradigms might be slightly influenced by the drug.13 14Also in the patient reported here an ophthalmological follow up examination did not show a visual deterioration. None the less, further studies on treatment with gabapentin should consider this topic with special concern.

In our patient, a well tolerated low dosage application of gabapentin coincided with a full remission of the hallucinations within 2 days after having started the medication and no relapses were reported in a follow up examination 3 months later. Considering the 2 year history of continuous daily repeated hallucinations this strongly points to a causal correlation, suggesting gabapentin to be an efficient and safe treatment for Charles Bonnet's syndrome. This remains to be proved in a larger group of patients.

In view of the current data on Charles Bonnet's syndrome, therapeutic approaches should be adjusted for each patient as there are possibly interindividual inconsistencies in responsiviness to treatment.15 To that end, a broader range of potentially effective drugs would increase the options.