Daytime Sleepiness, Agonist Therapy, and Driving in Parkinson Disease

Frucht and colleagues¹ proposed the unique association between treatment of Parkinson disease (PD) with 2 newer dopamine agonists and the occurrence of sudden onset of sleep, or sleep attacks. They described a series of 8 patients with PD treated with pramipexole or ropinirole who were involved in motor vehicle collisions because they fell asleep at the wheel. This small case series prompted grave concerns about the safety of prescribing the non-ergot dopamine agonists to patients with PD who drive. Recommended prescribing practices in many countries were altered.^{2 - 3} Multiple anecdotal and retrospective reports followed, all linking agonist therapy with excessive daytime sleepiness and sudden onset of sleep in patients with PD.^{4 - 7} Yet these reports did not systematically assess the frequency of daytime sleepiness in patients treated with other medications. This was a serious omission. Although levodopa initially was observed to cause transitory sleeplessness⁸ , within 5 years somnolence was also reported as an adverse effect.^{9 - 10} For example, in one series of 131 patients with PD treated with levodopa monotherapy, treatment was limited in 14% due to somnolence.¹¹

The controversy engendered by the report by Frucht et al is ongoing. The pivotal questions include whether sleepiness in PD is specific to treatment with dopamine agonists, whether PD patients can be screened for the likelihood of developing excessive daytime sleepiness and sleepiness while driving, and what drugs to use in the treatment of PD.

In this issue of THE JOURNAL, Hobson and colleagues¹² from the Canadian Movement Disorders Group address several of these questions. The study by Hobson et al differs from the prior case reports and retrospective observations by conducting a prospective evaluation of a large number of patients included from multiple movement disorders centers. This study systematically applied the validated Epworth Sleepiness Scale (ESS) and the Inappropriate Sleep Composite Score (ISCS) questionnaire, and assessed the contribution of other factors in addition to drug use in the development of daytime sleepiness in PD. A total of 638 patients with PD (420 of whom were actively driving) were studied using the ESS with 3 modifications, and the newly developed ISCS.

The major finding of the study was that daytime sleepiness and falling asleep at the wheel while driving were not associated with any specific antiparkinson drug or class of drugs. The authors also reported that excessive daytime sleepiness was frequent in PD, and correlated most strongly with severity and duration of PD, dosage of antiparkinson medication given as levodopa equivalents, and worse functional status. In predicting risk of falling asleep at the wheel, these investigators found that the scores on the ESS and the ISCS were the only factors independently predictive of events. The occurrence of sudden onset of sleep was infrequent, reported by only 3.8% of drivers. Only 3 of these patients did not have drowsiness prior to the sleep episodes. An ESS score cutoff of 7 had 75% sensitivity and 50% specificity for falling asleep at the wheel, and an ISCS score cutoff of 1 provided 52% sensitivity and 82% specificity. Thus, using both scales provided a high sensitivity and specificity for identifying patients with PD who fall asleep at the wheel.

This report addresses 3 major issues concerning excessive daytime sleepiness and sleep disturbances in PD. First, although the study by Hobson et al shows that excessive daytime sleepiness is common in PD, it does not address whether patients with PD are sleepier than other elderly groups. Identification of the frequency of and factors associated with sleep disturbances and daytime sleepiness in patients with PD is clinically important. A community-based study found that 60% of patients with PD have a sleep disturbance, compared with 33% of age-matched controls.¹³ The sleep disturbances can be broadly divided into those directly related to PD and its treatment and those that may occur with equal frequency in other elderly groups. The nocturnal sleep disturbances directly associated with PD include the reemergence of Parkinson symptoms during the night, causing sleep fragmentation and long periods of nocturnal insomnia; the occurrence of dopaminergic drug-related vivid dreams and hallucinations, sometimes causing nocturnal agitation and daytime sleepiness (day-night reversal); and the recently described rapid eye movement (REM) behavior disorder with a loss of descending motor inhibition during REM sleep, allowing patients to "act out their dreams" and cause sleep disruption and sleep-related injury to themselves and their caregivers.¹⁴ In addition to these disorders directly related to PD, other sleep disorders have been described frequently in patients with PD and in elderly individuals without PD.¹⁵ These include nocturia, periodic limb movement disorder, and sleep apnea. Although not explored in detail in the current study, the high frequency of nocturnal sleep disorders with consequent sleep deprivation may increase the susceptibility of a patient with PD to the sedative adverse effect of any centrally acting drug.¹⁴

The second issue involves whether a specific drug (eg, pramipexole or ropinirole) or class of drugs (eg, dopamine agonists) should be avoided in patients with PD who drive. In 2 of the pivotal double-blind placebo-controlled trials designed to assess the efficacy of pramipexole and ropinirole, the rate of self-reported somnolence was 18.3% and 36.2%, respectively.^{16 - 17} Because these studies were not designed specifically to assess daytime sleepiness, there is no indication of the severity or duration of sleepiness, but somnolence was not a cause of drug discontinuation. Comparative studies of dopamine agonists and levodopa have shown that both classes of drug cause somnolence.^{18 - 20} Studies using the ESS confirm that any antiparkinson medication can cause daytime sleepiness in patients with PD.^{21 - 22} In one study of 55 patients taking levodopa, cabergoline, or pramipexole showed equivalent mean scores on the ESS for all 3 drugs, and no patient had sudden onset of sleep.²¹ In another study, in which 303 patients with PD were assessed using the ESS, there was no correlation of increased daytime sleepiness as measured by the ESS with the use of a particular dopaminergic medication.²² Consistent with the study by Hobson et al, this smaller study showed that daytime sleepiness correlated primarily with longer duration and severity of PD. In addition, male sex was found to be associated with higher ESS scores.

The third issue is whether there is a practical tool that clinicians can use to screen PD patients for the presence of daytime sleepiness and predict whether a patient is at risk for falling asleep while driving. The Multiple Sleep Latency Test (MSLT), which measures the mean sleep latency during 4 to 5 daytime nap opportunities using polysomnography, is widely accepted as the "gold standard" for quantifying daytime sleepiness.²³ An MSLT is both expensive and cumbersome, requiring a day in the sleep laboratory, making the MSLT impractical as a screening tool. In contrast, the ESS is a simple, 8-item questionnaire that is completed by the patient. The total score is obtained by summing the scores of these 8 items. The ESS was developed in 1991 by Johns²⁴ and was shown to have a high level of reliability and internal consistency.²⁵ However, the ability of the ESS to measure daytime sleepiness has been debated. In 2 recent studies, there has been no correlation between the MSLT and the ESS score in patients without PD.^{26 - 27} The ESS has been found to correlate with a patient's perception of problem sleepiness²⁶ and with other measures of functional status and well-being.²⁸ The ESS is thought to measure different aspects of daytime sleepiness not captured by the MSLT. The ESS has been previously used to assess sleepiness in PD.^{21 - 22} The MSLT has also been applied to PD, but associated or predictive factors for shortenened sleep latency could not be determined due to the small number of patients assessed.²⁹ The ISCS is promising as a separate clinically relevant measure for detecting pathological sleepiness, but it has not been evaluated for its psychometric properties. Although the ESS and ICSC require further testing in PD, the study by Hobson et al indicates that these scales used together are sensitive and specific for identifying the sleepy PD patient who is at risk of napping while driving.

The last, but perhaps most fundamental, question is whether patients with PD who drive should be treated with dopamine agonists. Patients with PD who drive generally are younger than those who do not and frequently are employed or required to drive for other reasons. Dopamine agonist therapy has been implicated with falling asleep at the wheel.^{1 ,4 - 7} However, dopamine agonist as initial monotherapy for PD delays the onset of dykinesias and disabling dyskinesias that affect up to 50% of patients receiving levodopa monotherapy.^{18 - 20} The dilemma in clinical decision making for treatment of younger, highly functional patients with PD is choosing whether to initiate agonist therapy that may prevent future disability.

The study by Hobson et al indicates that sleepiness behind the wheel can be detected by using simple, clinically practical questionnaires. Furthermore, the occurrence of "sleep attacks" in the absence of drowsiness is exceedingly rare. This report also shows that although sleepiness may be common in PD and in patients with PD who drive, the cause of the daytime sleepiness is not directly associated with a specific dopaminergic medication. The role of the underlying pathophysiological mechanisms related to PD, the presence of co-existing nocturnal sleep disturbance, and the contributions of all medications may play a role in the development of excessive daytime sleepiness. A clearer understanding of the role of each of these factors requires further exploration. In the meantime, the study by Hobson et al provides clinicians with 2 tools, the ESS and ISCS, to screen PD patients for the presence of daytime sleepiness. Patients considered to be at risk for falling asleep behind the wheel should be cautioned to avoid driving, regardless of drug treatment. The study by Hobson et al confirms the need for caution in the interpretation of anecdotal case reports and highlights the need for further rigorous investigations of the disorders of sleep and daytime alertness in patients with PD.