Patients who have sustained clinically significant traumatic brain injury (TBI) often complain of sleep disturbances. Immediately following TBI, difficulty in falling asleep and frequent awakenings are commonly reported, whereas after several years excessive somnolence is more typical.1 However, despite the ubiquity of sleep abnormalities following TBI, objective documentation is lacking for the occurrence of persisting severe insomnia specifically associated with brain injury.2 The present case report describes a patient who, as a result of cerebral concussion, developed an enduring inability to sleep that has been virtually intractable to behavioral and pharmacological therapies.
In 1991, at the age of 35, the patient was dragged from his car and repeatedly beaten about the right side of the head with a blunt object. Although his recollection of the details of the attack is not clear, he believes that he may have momentarily lost consciousness. He was able, however, to drive his car from the scene of the assault. Prior to his injury the patient was employed as a paralegal and his medical history was unremarkable. Premorbid intellectual functioning was estimated with the National Adult Reading Test, which indicated a Verbal IQ of 103 (58th percentile) and Full Scale IQ of 106 (61st percentile). The patient indicated that before being injured, he experienced no sleep disturbance and usually slept 6 to 8 hours per night.
Difficulties in sleeping began immediately after the injury, in association with severe memory and concentration problems. The patient reported that he typically goes to bed between 10:30 p.m. and 12:00 midnight and takes up to 4 hours to fall asleep. When he does sleep he experiences a variable number of awakenings, only occasionally due to extraneous noise, and once awake he has extreme difficulty falling asleep again. According to his sleep log the patient averages about 15 hours' total sleep per week. He does not snore and does not awaken with a choking sensation or with gasping. Although the patient occasionally exhibits hypnic jerks while in bed, his symptoms are not consistent with restless legs syndrome. It should be noted that the patient's reduced sleep time is not due to a decreased need to sleep. He described feeling exceptionally tired, and to even a casual observer he appeared extremely fatigued.
Ambulatory 24-hour electroencephalograms with conventional electrode montages, performed 1 and 4 years after the injury, were described as normal; no epileptiform activity was noted. An all-night polysomnogram with 12 additional EEG recording channels was performed 18 months after the injury and had results consistent with the patient's descriptions of his sleeping problems. Sleep latency was 283 minutes; total sleep time was 62 minutes, with stage 1 sleep 10.3 minutes and stage 2 sleep 51.7 minutes. There was a complete absence of slow-wave and REM sleep. The patient awakened spontaneously without apparent cause and did not fall asleep again during the monitoring period.
An MRI examination of the brain was unremarkable except for the presence of a left posterior parietovenous angioma. However, 4 years after the injury a positron emission tomography scan using [18F]deoxyglucose, performed at the hospital of the University of Pennsylvania, revealed abnormally increased metabolic activity in the posterior cingulate gyrus as well as in several small regions of the left prefrontal cortex. Abnormally decreased metabolism was observed in the medial temporal lobe bilaterally, right visual cortex, left caudate, left putamen, and anterior dorsal aspect of the thalamus. Increases and decreases were evaluated with respect to other areas of the brain.
In addition to his complaints of sleeping difficulties, the patient described feeling somewhat sad and has also noted increasing difficulty with concentration and memory. Evaluation with the Beck Depression Inventory 4 years postinjury confirmed moderate to severe depression.
Neuropsychological assessment was performed to obtain objective evaluation of the patient's cognitive difficulties. Memory functioning was assessed with Wechsler Memory Scale—Revised subtests (Logical Memory I and II, Verbal Paired Associates I and II, Digit Span) and the Cambridge Automated Neuropsychological Test Assessment Battery (CANTAB)3 computerized tests of visual pattern recognition, spatial recognition, delayed matching to sample, and visuospatial short-term memory capacity. The norms used for each of the tests were those provided by the test developers.
The patient's memory functioning, even under conditions of low load (i.e., wherein what is to be remembered is very simple and recall is immediate), was moderately to severely impaired. Under conditions of high load, the patient's memory functioning was profoundly impaired. The one exception was the very simple Forward Digit Span test, wherein a sequence of numbers must be repeated in the order in which they were heard; the patient scored at the 50th percentile for men of his age. The more difficult task of hearing a sequence of numbers and repeating them in the reverse order from which they were presented (Backward Digit Span) was only in the 9th percentile. His ability to immediately recall a short paragraph was in the 18th percentile; if he had to wait about 30 minutes before being asked to recall the story, his performance deteriorated to the 6th percentile.
Performance on a simple visual pattern recognition test was moderately subpar (approximately 8th percentile). However, on the relatively demanding delayed matching-to-sample test, in which complex visual patterns have to be identified after delays varying from 0 seconds up to 12 seconds, the patient's difficulties were severe (<1st percentile). Similarly, when he was asked to recall the spatial location of test stimuli, his performance was at chance levels (i.e., similar to what would result from pure guessing: <1st percentile). Taken together, all of these findings represent a major impairment of working memory with a relatively rapid loss of new information.
In addition to memory deficits, standard testing4 demonstrated psychomotor slowing, microsmia, and severe impairment (≤1st percentile) of attention and executive functioning. In addition, neurological examination indicated slowing of repetitive movements, rigidity upon passive manipulation, reduced right arm swing during walking, and postural instability, consonant with the indications of basal ganglia dysfunction shown on PET.
As noted previously, the patient had significant depression, a state often associated with memory disorders. Moreover, as noted by Alexander et al., "severe depression is commonly accompanied by psychomotor slowing, impaired attention, decreased cognitive flexibility and poor retrieval memory."5 However, the neuropsychological deficits we observed were apparently present immediately after the patient's head injury, whereas depression developed only later.
Although the patient disclaimed lingering fears from the assault, the initial psychiatric consultation and treatment mistakenly focused on a diagnosis of posttraumatic stress disorder (PTSD) with associated anxiety and insomnia. Symptoms corresponding to diagnostic criteria for PTSD (e.g., reexperiencing of the trauma, attenuated responsiveness or involvement with the external world) were not present. The patient was repeatedly reassured that with therapy his symptoms would disappear, and fluoxetine was initiated with doses up to 40 mg per day. Augmentation of drug treatment with concurrent administration of either doxepin (50 mg twice a day) or trazodone (150 mg per day) met with failure, as did bedtime administration of either amitriptyline (50 mg) or pentobarbital (3 grains). The patient was placed in individual and group therapy but found no relief from his symptoms. Changes in behavior based on recommendations from sleep disorder specialists to improve "sleep hygiene" were also ineffective.
Follow-up psychiatric evaluation diagnosed TBI, and attempts were made to ameliorate symptoms with pharmacotherapy. Over the past several years the patient has cooperated with therapeutic attempts involving the use of many drugs both individually and in combination. Unfortunately, there is no consensus of medical opinion offering guidance in the treatment of insomnia of this type, and therefore a trial-and-error approach using reasoned judgment is necessary. Although hypnotics were tried for obvious reasons, other medications were chosen in light of specific symptoms such as anxiety (buspirone). The PET findings of basal ganglia disorder suggested the possible use of drugs with action in this area. Although results actually confirmed the validity of this approach, other considerations were countervailing, as explained below. For various periods over several years, the following were administered: desipramine (25—300 mg), desipramine plus T3, pemoline (18.75—37.5 mg), divalproex (250—500 mg), carbamazepine (200 mg), buspirone (5—20 mg), zolpidem (10 mg), chloral hydrate (2 gm), haloperidol (2—4 mg), thioridazine (25—200 mg), risperidone (2—12 mg), risperidone plus ergoloid mesylate (1 mg), risperidone plus haloperidol and diphenhydramine, primidone (250 mg) plus risperidone and diphenhydramine, carisoprodol (1,400 mg) plus risperidone, diazepam (10 mg) plus temazepam (7.5—15 mg), melatonin, and carisoprodol plus diazepam and temazepam.
At the time of the current neuropsychological testing the patient was receiving diazepam (Valium), carisoprodol (Soma), and diphenhydramine (Benadryl). Most relevant are diphenydramine and diazepam, since these can produce drowsiness and thereby impair attention and slow information processing. However, the entire range of the patient's difficulties cannot be explained by his drug regimen.
Efficacy of treatment was based on patient report via a sleep log. Drug dosages were titrated upward over a period of 6 to 8 weeks unless adverse reactions (e.g., nightmares with pemoline, nausea and vomiting with carbamazepine) caused termination sooner. Risperidone in the above-noted combinations ameliorated symptoms such that sleep for up to 33 hours per week was claimed. At times the patient would sleep 3 hours consecutively. More commonly, the patient slept approximately 20 hours per week with the combination of risperidone, haloperidol, and diphenhydramine. Unfortunately, growing concerns about the risk of permanent motor side effects (e.g., tardive dyskinesia), compounded by the emergence of extrapyramidal motor signs in association with the PET findings of basal ganglia disruption mentioned above, prompted termination of neuroleptic administration.
The sleep disturbance described in this patient differs from insomnia associated with delayed sleep phase syndrome, a disorder frequently caused by TBI. In the latter disorder, insomnia is caused by a delay in the preinjury sleep-wake pattern so that the patient's circadian rhythm conflicts with the activities of daily living.6 Delayed sleep syndrome, unlike the present case, is not associated with loss of slow wave and REM sleep7 and is ameliorated by various drug treatments such as melatonin.8
The neuronal mechanisms of this patient's difficulties are problematic. Certainly, his neuropsychological deficits may be secondary to unremitting sleep deprivation.9 In addition, many of the abnormalities of cerebral metabolism described in the PET scan are also seen after sleep deprivation.10 However, a similar clinical picture, including loss of slow wave and absence of normal REM, has been described in patients with a familial disease characterized by degeneration of the anterior and dorsomedial thalamus. The present case, wherein PET abnormalities were found in the same area, accords with the possibility that traumatic injury of this zone may also produce persisting loss of normal sleep patterns.11 Unfortunately, the presence of abnormalities in other areas precludes a definitive conclusion.