| JOURNAL HOME | CME HOME | THIS MONTH | PAST ISSUES | ETOC | COLLECTIONS |
| AUTHORS | REVIEWERS | EDITORIAL BOARD | FEEDBACK | RSS | HELP |
| ||||||||||||||
| ||||||||||||||
|
|
|
|
||||||||||||
|
||||||||||||||
From the Departments of *Anesthesiology, Neurological Surgery and Neurology, University of Washington School of Medicine, Seattle, Washington.
Address correspondence and reprint requests to Lucinda A. Grande, MD, Department of Anesthesiology, Box 356540, University of Washington School of Medicine, Seattle, WA 98195. Address e-mail to cgrande{at}u.washington.edu.
Abstract
A middle-aged woman began experiencing spells of profound anterograde amnesia several months after beginning intrathecal baclofen treatment for generalized dystonia. Her spells met criteria for transient global amnesia, but were unusual because of their frequent recurrence and because their frequency was somewhat dose-dependent on baclofen. Fludrocortisone decreased the frequency of these episodes, and sublingual nitroglycerin both prevented and terminated them. Baclofen-induced amnesia in rodents is a reliable model of memory impairment. In contrast, baclofen-induced memory impairment in humans is uncommon. Baclofen- associated transient global amnesia has not previously been reported.
Intrathecal baclofen is approved for patients with spasticity of spinal or cerebral origin who are intolerant of or unresponsive to oral baclofen.1 It is also used for other painful chronic neurological conditions, including dystonia.2 Common central nervous system side effects are hypotonia, somnolence, and dizziness. Rare adverse effects include seizures, euphoria, dysphoria, hallucinations, fatigue, tremor, dystonia, and ataxia.1
Baclofen has also been associated with memory problems. Amnesia has been noted as a rare adverse effect occurring during clinical trials of intrathecal baclofen but is not further described.1 Memory impairment was described as an occasional feature of a chronic baclofen intoxication syndrome.3 One report of three patients describes oral baclofen-induced memory impairment involving inability to remember or recognize names of familiar persons and places.4
We report an unusual case of memory impairment associated with intrathecal baclofen in which discrete episodes of profound amnesia occurred repeatedly. Each spell met criteria for a rare neurological entity known as transient global amnesia (TGA).5–9
CASE DESCRIPTION
Background
A female patient developed symptoms of dystonia in the neck at age 28 yr, 2 mo after jaw surgery and a fall resulting in a minor head and neck injury. Symptoms progressed to include the face, esophagus, upper and lower extremities and trunk. A tracheostomy was placed at age 37 after vocal cord spasms led to oxygen desaturation.
Treatment for her dystonia included oral gabapentin and tizanidine and periodic botulinum toxin injections. The most effective treatment was oral baclofen, but it caused symptomatic hypotension. Therefore, at age 39, she underwent placement of a catheter and pump for intrathecal baclofen.
The patient’s family history was positive for nonspecific neuromuscular disorders, but genetic testing was negative for the most common dystonia mutation, DYT1. She had a stable family life, was employed part-time, and did not smoke or use alcohol. She had mild depression controlled with citalopram.
Amnesia History
Several months after beginning treatment with intrathecal baclofen, she began experiencing dissociative spells. During these episodes, she was able to conduct routine activities such as paying for groceries or driving a car, but she experienced profound anterograde memory loss such that she had no memory of those periods and did not recognize a time lapse. Family and coworkers reported that during spells she asked questions repeatedly about the current situation and immediate future. Occasional severe spells included retrograde amnesia; on emergence from those she was unable to recall newly acquired memories of people and places from the previous several weeks.
Spell frequency varied somewhat with baclofen administration. For example, spells which occurred 2–3 times per day on an infusion rate of 1330 µg/d stopped after downward titration to 555 µg/d, but resumed after a few months. Repositioning of the catheter tip from the thoracic to the cervical region to improve dystonia was followed by an increase in spell frequency, which returned to baseline after withdrawal to its original position.
The episodes lasted from a few minutes to several hours and were often preceded by fatigue and a mild burning chest pain. They sometimes occurred at times of emotional stress, sometimes within about 15 min of standing after prolonged sitting, and sometimes with no apparent stimulus. There was no shaking, tongue biting, or bowel incontinence, but urinary incontinence occurred during several severe spells. Two spells were captured on inpatient electro-encephalogram monitoring; both were negative for epileptiform activity. Psychiatric counseling, biofeedback, acupuncture, and hypnosis had little beneficial effect.
Amnesia Treatment
Five years after baclofen pump placement, episodes of hypotension as low as 70/45 were discovered during prespell fatigue, and treatment with oral fludrocortisone 0.1 mg daily was begun. A decrease in spell frequency occurred, from every 3 or 4 days to every 14 or 15 days. The remaining episodes were not preceded by hypotension.
A trial of sublingual nitroglycerin 0.3 mg as needed was begun for symptom relief. When administered by family shortly after spell onset, her awareness sometimes sharpened over several minutes, concluding with a return to normal consciousness. She began using nitroglycerin regularly to avert spells, when warned by an aura of fatigue or chest pain.
A trial of isosorbide dinitrate was begun; after initially promising results, the trial was terminated during the second month due to increasing frequency and severity of spells. Symptoms returned to baseline after discontinuation. Topiramate reduced spell frequency somewhat and was continued. During the subsequent 3 yr, there was a trend toward a slow increase in spell frequency.
DISCUSSION
The patient’s spells met diagnostic criteria for TGA,5 including witnessed attack of <24 h duration, anterograde amnesia, no clouding of consciousness or loss of personal identity, cognitive impairment limited to amnesia (no apraxia or aphasia), and no recent history of head trauma or seizures. The repetitive questioning behavior was typical of TGA, but the frequent spell recurrence was very unusual.6–9 The association of TGA with baclofen has not previously been reported.
The annual incidence of TGA is estimated at about 3–30/100,000.6 Since its first description in 1958,7 its etiology has not been explained satisfactorily by an ischemic, migrainous, or epileptic mechanism.6 Recurrence of our patient’s TGA and the discovery of exacerbating and alleviating factors provide a unique opportunity to examine this phenomenon. Association with baclofen suggests dysfunction of 
-aminobutyric acid (GABA)-mediated synaptic transmission. Symptomatic relief by nitroglycerin suggests a role for nitric oxide (NO) metabolism. A brief review of relevant animal studies provides a framework for interpreting her condition.
Baclofen, a GABA-B receptor agonist, induces anterograde amnesia in rodents. This phenomenon is so reliable that baclofen-induced amnesia is a model used to study the role of GABA and the GABA-B receptor in learning and memory.10–14 GABA-B receptor activation by baclofen inhibits release of GABA and other neurotransmitters.13 Baclofen administered to rats impairs working memory in a dose-dependent fashion.14 It selectively disrupts acquisition of short-term memory (stimulus exposure 
15 s), an effect associated with disruption of firing in specific cell types in the hippocampus.11 The rarity of baclofen-induced memory impairment in patients suggests there is a protective feature of human memory potentially lacking in this patient.
NO is a highly diffusible free radical with a broad range of physiological functions including smooth muscle relaxation and neuronal signaling.15–17 It is an important mediator of memory formation in rat hippocampus.18,19 Amnesia caused by administration of an inhibitor of NO synthesis (7-nitroindazole) is reversed by administration of a NO donor (molsidomine).20 Baclofen-induced amnesia in rats is also reversed, and prevented, by administration of the NO donor molsidomine.12 Baclofen-induced amnesia in our patient was likewise reversed and prevented by administration of a NO donor (nitroglycerin). The ultimate failure of isosorbide dinitrate, a long-acting NO donor, may be explained by drug tolerance, which can occur during its use for stable angina.21
NO modulates the release of GABA and other neurotransmitters.22,23 Its effect on GABA release is biphasic and concentration-dependent; at high concentration, NO stimulates GABA release, whereas below a threshold concentration, it inhibits GABA release.12,24–26 Because of this biphasic property, an abrupt decrease in hippocampal NO concentration theoretically could act like the flipping of a switch which turns off GABAergic synaptic transmission and consequently interrupts memory acquisition.
Under what conditions might an abrupt decrease in hippocampal NO concentration occur? The patient’s TGA spells sometimes followed sympathetic stimuli such as the baroreflex (when standing after prolonged sitting) and emotional stress. Sympathetic stimulation interrupts NO synthesis in the endothelium of the peripheral vasculature via autonomic transmission, resulting in vasoconstriction.16 Although there is no major neural pathway by which it might likewise interrupt hippocampal NO synthesis,16 sympathetic stimulation does result in production of stress hormones, some of which modulate hippocampal memory processes27,28 and are also implicated in brain NO metabolism.29 We hypothesize that a stress hormone-mediated reduction in hippocampal NO synthesis caused spells in our patient. Fludrocortisone reduced spell frequency, perhaps by preventing orthostatic hypotension and an exaggerated baroreflex.
TGA occurred in this patient with intrathecal but not oral baclofen. Normally, intrathecal administration results in fewer central nervous system side effects. Since the plasma concentration is about 100 times lower with intrathecal versus oral administration,1 drug exposure is limited primarily to tissue accessible by diffusion from cerebrospinal fluid. Exposed tissue includes the dorsal horn, baclofen’s putative site of action in spasticity and dystonia.1 However, the hippocampus, which is adjacent to the lateral ventricle, is also exposed. Biologically relevant diffusion of intrathecally delivered drugs into periventricular brain regions such as hippocampus has been demonstrated in gene therapy studies in mice.30 The periventricular hippocampal layers (about 200 µm in rats) are particularly dense with GABA-B receptors.31 This memory-related region may therefore be targeted with high specificity by baclofen when delivered intrathecally.
In summary, we propose that TGA episodes were caused by an abrupt reduction in hippocampal NO synthesis and a prolonged decrease in NO concentration. The decrease in NO concentration flipped a physiological switch, interrupting hippocampal GABA- ergic neurotransmission and memory acquisition. These episodes were sometimes triggered by the release of stress hormones after sympathetic stimuli. Intrathecal baclofen contributed by reducing baseline GABA-ergic synaptic transmission in the hippocampus. Our patient may have inherited a vulnerability to TGA within the NO-GABA signaling pathway.
The proposed mechanism of our patient’s TGA, acute dysfunction in NO-mediated hippocampal GABA-ergic transmission, is consistent with the TGA literature. For example, precipitating events are usually situations causing powerful sympathetic stimulation, such as immersion in cold water, sexual intercourse, severe pain, and emotional stress.6,9 Our patient’s unique condition, in which recurrent baclofen-induced TGA was exacerbated by sympathetic stimuli and alleviated by nitroglycerin and fludrocortisone, may provide insights into the etiology of TGA.
ACKNOWLEDGMENTS
The authors thank Jon Hillyer, MD, Buck Hartford, PA-C, Gary Gretch, MD, Enayat Niakan, MD, and Donald Bright, MD, for participating in patient care and providing helpful history.
Footnotes
Accepted for publication November 30, 2007.
REFERENCES
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
