In 2009, Zvosec and colleagues published a report in Sleep Medicine on three people who died after taking Xyrem, not illicit GHB. All three had taken other drugs, and two did not have a confirmed history of narcolepsy with cataplexy. The authors called attention to the drug’s respiratory depressant effects and risks for people who were obese or taking other medications.59 This prompted a letter to the editor of Sleep Medicine signed by twenty-four physicians, representing the European and North American sleep neurology establishment, defending Xyrem as “an effective and safe treatment for patients suffering from narcolepsy.” Zvosec also criticized a Jazz-sponsored study of Xyrem in people with sleep apnea, which concluded that the drug “might increase central apneas and cause oxygen desaturation in some individuals and should be used with caution.”60

  The FDA eventually came to a similar conclusion, but a full picture of Xyrem’s safety record was not available until several years after its approval. The close contact with patients built into Xyrem’s distribution system may have provided an opportunity for thorough monitoring, yet it appears that adverse events were underreported until 2011 (figure 15.3).

  At the time of an April 2011 visit by FDA inspectors, the company did not have adequate procedures for adverse event reporting, according to a warning letter released later that year.61 As a result, serious adverse events, including eighty-two deaths, had not been recorded by the central pharmacy and reported to the agency.62 Most reported deaths associated with Xyrem until 2011 had been missed.

  FIGURE 15.3. Xyrem adverse event reports by year, obtained via the FDA’s Adverse Event Reporting System Public Dashboard. Events appear in the database according to the date received, not when they occurred. Most Xyrem-associated deaths involved other drugs, and assigning causality is not possible based on this data alone.

  In 2012, the FDA found that many of the deaths came when Xyrem was combined with other drugs such as alcohol, benzodiazepines, or opioids.63 It appeared that a majority of adverse events and deaths occurred in patients who were prescribed Xyrem for off-label uses such as fibromyalgia, insomnia, or migraine. The agency recommended that Xyrem should be used cautiously, if at all, in patients with respiratory issues such as sleep apnea or COPD (chronic obstructive pulmonary disease) because of the risk of respiratory depression. The agency also added a contraindication against alcohol use to Xyrem’s label. A report from a nonprofit group, the Institute for Safe Medication Practices, concluded: “For almost a decade, the true adverse event profile of sodium oxybate was not known to the company, the medical community, or the FDA.”64

  STABILIZING THE DEFINITION OF NARCOLEPSY

  Jazz-sponsored studies have shown what many experts on narcolepsy often do not acknowledge. In the United States, diagnoses of narcolepsy without cataplexy have become several times more prevalent than diagnoses of narcolepsy with cataplexy, contradicting an assumption expressed in older publications on narcolepsy.65

  The Stanford-based authors of the Burden of Narcolepsy Disease study, which scanned a nationwide insurance database for narcolepsy-related claims from 2006 to 2010, expressed surprise at their findings. They wrote that it “invites suspicion regarding the accuracy of the diagnosis of narcolepsy without cataplexy by physicians in the United States.”66

  The numerical dominance of narcolepsy without cataplexy in the United States does not appear to have been the case before the twenty-first century. For example, earlier studies in Minnesota and the Seattle area found that narcolepsy with cataplexy was more common.67 It also contrasts with recent studies from other countries such as Switzerland and Spain.68 Thus, the increase in narcolepsy without cataplexy in the United States may be partially because of differences in diagnostic practices between countries, rather than differences between their populations. Put more simply, physicians in the United States may have been diagnosing narcolepsy less rigidly than in other countries. This could mean not checking for insufficient sleep before an MSLT or not investigating other potential reasons for excessive daytime sleepiness. In support of this point, U.S. military clinicians reexamined twenty-three service members diagnosed with narcolepsy at civilian sleep centers, but they were able to confirm the initial diagnosis in only two.69

  Stabilizing expert opinion on narcolepsy’s looser definition was in Jazz’s interest, even if the company was not responsible for the expansion in narcolepsy diagnoses. Initially Xyrem was approved specifically for the treatment of cataplexy but not for treating daytime sleepiness. The FDA had concluded that Xyrem was effective at reducing the frequency of cataplexy, but the data on daytime sleepiness was “very weak at best.”

  In 2005, the FDA approved Xyrem for excessive daytime sleepiness in the context of narcolepsy, based on two studies. One showed its subjective benefits in combination with stimulants.70 The other study did not require participants to have narcolepsy with cataplexy and showed that Xyrem could partly substitute for modafinil.71 The approval meant that the overall diagnosis of narcolepsy was relevant for insurance coverage, rather than the specific symptom of cataplexy.

  Later, Jazz supported a June 2012 conference that sought to establish a consensus on diagnosis of narcolepsy type 2. A report from the conference noted that some participants thought narcolepsy type 2 might lie in the borderland between narcolepsy type 1 and IH, while others perceived it as a distinct entity.72 For his part, David Rye came back from the conference grumbling about his peers’ willingness to rely on the MSLT.73 The conference report emphasized the limitations of the MSLT yet said that it remained the most important test for narcolepsy type 2 diagnosis.

  THE FLOOD OR THE TRICKLE?

  Despite Xyrem’s demonstrated efficacy in managing narcolepsy symptoms, sleep researchers remain uncertain about the drug’s mechanism of action. The history of its active ingredient, GHB, stretches back decades before scientists knew much about the neurological basis of narcolepsy.

  Biochemically, the drug resembles GABA. For its acute knockout effects, GHB is thought to act through GABA-B receptors, which are separate from the GABA-A receptors for benzodiazepines. GHB binds to GABA-B receptors with less affinity than GABA itself. Like GABA-A receptors, GABA-B receptors are inhibitory, but they trigger a different set of signals inside neurons, and their distribution in the brain is also different.74

  As far back as the 1960s, neuroscientists detected GHB as a natural metabolite in the mammalian brain, leading to speculation about its possible physiological role as a neurotransmitter.75 However, when someone takes the drug as a fast-acting sedative, the brain is flooded with GHB at a concentration hundreds of times higher than the amounts naturally found there. Do the drug’s benefits come primarily from the flood or the trickle in its aftermath? The drug could also be affecting brain cells’ metabolism, because GHB is quickly converted into an energy source.

  One 2016 review offered this explanation: “The mechanism of the majority of GHB pharmacological effects are thought to interfere primitively with the brain GABA system,” but “it seems difficult to consider GHB pharmacological effects mediated only and purely via GABA-B receptor stimulation.”76 In addition to GABA-B receptors, several research groups have identified “high affinity GHB receptors” bound by GHB in the brain—tight enough to hold on even after the flood washes away. However, the biochemical identity of these receptors has been elusive. In 2021, Danish scientists found an enzyme that appears to be the long-sought target, but investigating that enzyme’s relevance for narcolepsy remains to be done.77

  When used for management of narcolepsy, GHB’s effects are not immediate. With respect to daytime sleepiness, it can take days or weeks for patients to feel the benefits.78 The drug appears to be gradually altering how the brain’s sleep and wake circuits function, even after it has washed out of the body. “Chronic treatment is doing something to reset the circuits,” said Tom Kilduff, a neuroscientist in California whose laboratory has studied GHB extensively in animal models. “We don’t understand the pathway that makes GHB effective therapeutically,” Kilduff added. “That makes it a very attractive scientific question.”

  GABA-B receptors are present on several groups of neurons in the brain that regulate sleep. Because the drug’s therapeutic benefits are clearest in narcolepsy, it is reasonable to ask whether it bolsters or substitutes for the function of hypocretin neurons so central to narcolepsy type 1. Which neurons are important for GHB’s effects remains an unanswered question, Kilduff said. “Hypocretin neurons are part of the story, but not the whole story,” he said.

  When other researchers gave rats GHB, they were surprised to find that it didn’t activate brain regions connected with reward and addiction, such as the nucleus accumbens.79 Some experiments indicate that GHB acts on other neurons that connect the thalamus and cortex, which drive slow wave oscillations during non-REM sleep.

  GHB appears to be an exception to the general rule that artificial sedation does not substitute for natural sleep.80 GHB is known to increase slow waves, which are thought to be important for memory consolidation and possibly for the other aspects of restorative sleep. However, slow wave oscillations provoked by GHB don’t look the same as physiological slow wave sleep.81

  Remember the Michigan woman whose cataplexy was made worse by the blood pressure medication prazosin? GHB might be accomplishing the reverse, by stabilizing the locus coeruleus. The locus coeruleus is a group of neurons in the brainstem that produce norepinephrine, whose lack of activity is critical during cataplexy. Beyond cataplexy, it is possible that the neurons needed for GHB’s alleviation of daytime sleepiness are separate. The puzzle highlights how sleep and sleepiness are orchestrated by networks of neurons, not by discrete centers in the brain.

  BACLOFEN: POOR MAN’S XYREM

  If acting through GABA-B receptors was the only way GHB was exerting its effects, then the muscle relaxant baclofen, less expensive and less prone to abuse, might be able to do the job. Sleep researchers have wondered since the 1970s whether baclofen—a canonical GABA-B agonist—might be helpful for people with narcolepsy. A few members of the Atlanta hypersomnia community have tried baclofen for themselves, describing it as “the poor man’s Xyrem.”

  Because of Xyrem’s side effects, cost, and restrictions around its use, clinicians keep proposing baclofen as a substitute for GHB, despite the negative experiences of influential figures such as Stanford’s Christian Guilleminault.82 Some clinicians have reported that their narcolepsy patients have had a positive response to baclofen.83 Kilduff’s work on mouse models has also suggested that R-baclofen could be effective for narcolepsy.84 (R-baclofen is one of two mirror-image forms of the molecule; the R-form has higher affinity for GABA-B receptors.) Baclofen’s lack of interaction with the shadowy high-affinity GHB receptor suggests that baclofen’s and GHB’s shared mechanism must be related to GABA-B receptors. However, since narcolepsy type 1 and IH are different disorders, the factors that determine efficacy for GHB or baclofen may also differ between the two groups of patients.

  CHAPTER 16

  BIOMARKERS OF SLEEPINESS—AND IH

  In our busy world, natural long sleepers attract less interest, although their capacity of spending many hours asleep could be regarded as no mean accomplishment.

  —Alexander Borbely and colleagues, 1996

  For more than a century, scientists have been looking for the stuff in the brain that makes someone feel sleepy. It’s becoming clear that sleepiness is better thought of as a pattern embedded within our brain cells, rather than a substance that can be extracted from the body.

  Practically speaking, it’s much easier to detect physiological signs of sleepiness, like droopy eyelids or impaired reaction time, than it is to isolate sleepiness itself. The signs don’t cause sleepiness but are tightly associated with it. Scientists who study sleep have long considered it a top goal to establish “biomarkers of sleepiness,” and millions of dollars of investment have gone into developing them.

  We can easily see the need when considering public safety. A driver who has not slept in twenty-four hours is as impaired as someone who is legally drunk, according to measurements of their driving performance.1 The federal government estimates that thousands of crashes and hundreds of fatalities occur every year nationwide as a result of drowsy drivers. Also, a sizable percentage of commercial truck drivers have obstructive sleep apnea, putting them at higher risk of a crash.

  Clinicians know how to screen for and detect sleep apnea, but someone can be drowsy for other reasons, including simply not getting enough sleep. But some people don’t recognize that they are pathologically sleepy until it overwhelms them. A broader test for drowsiness, not only one identifying breathing disruptions, could catch people who present potential risks. Professional groups whose performance depends on their ability to stay awake for long periods, such as truck drivers, pilots, or surgeons, could be asked to undergo some kind of alertness test or, alternatively, spit into a cup or even give a blood sample.

  Like people with sleep apnea, people with narcolepsy or IH are assumed to have a higher risk of car crashes, but just a few studies have looked directly at their risk.2 The most recent one from France found that patients with narcolepsy or IH had double the risk of having a car crash in the last five years, compared to controls.3 Most studies say that the MWT can predict driving performance. But one paper by Dutch sleep researchers compared the MWT, along with the psychomotor vigilance test and a sustained attention test, to an actual driving performance test for evaluating someone’s fitness to drive. The authors concluded that “none of the tests had adequate ability to predict impaired driving, questioning their use for clinical driving fitness evaluation in narcolepsy and IH.”4

  Both researchers and policy makers have lamented the lack of a broadly applicable “breathalyzer for sleepiness.”5 It has been a challenge to get to the point of having good biological markers. Scientists are close to being able to spot acute sleep loss, but chronic sleep insufficiency—getting fewer hours of sleep than ideal for a few weeks—may be harder to discern. A 2015 workshop on the topic concluded: “Laboratory measurements such as EEG-based assessment of sleepiness or objective performance measures are expensive and complex, not amenable to general use and there is no conclusive evidence that they provide gold standard measurements by which assessment of the state of chronic sleep loss can be determined.”6

  More recently, there have been signs that the field is getting closer to its long-standing goal. The Australian company Optalert currently produces infrared goggles that measure how fast and how far the eyelids open after a blink, generating a drowsiness score. The goggles are a scientifically rigorous way of measuring droopy eyelids.7 In a different mode, researchers at the University of Surrey in the United Kingdom have been closing in on a blood test that could reliably identify otherwise healthy people who have skipped a night of sleep. Their still-experimental test looks for changes in activity in a panel of genes, tracking cellular stress.8

  The extent of the effort to identify biomarkers of sleepiness is beyond the scope of this book, but some of this research might eventually be applied to people with narcolepsy or idiopathic hypersomnia or those suspected of having these conditions. Currently, these types of measures are not used in the sleep clinic. The standard ways to gauge someone’s sleepiness are subjective, like the Epworth Sleepiness Scale, or operational, like the MSLT or MWT, which serve as imperfect measures of daytime sleepiness.

  Looking ahead, it could be revealing to apply newer biological tests of sleepiness to people with IH. IHers often say that they feel like they’ve stayed up all night, despite spending an amount of time sleeping that most people would consider more than enough. Having an easily measurable biomarker of sleepiness might provide an explanation for how they feel.

  We don’t know how IHers would perform on the newer biological tests. Perhaps some IHers will have droopy eyelids or a slower reaction time for more of the day, or perhaps patterns of gene activity in their blood might match the effect of sleep deprivation. Others may not display the same similarities. Would that mean their feelings are less valid? This points to how sleepiness can be viewed both objectively or subjectively.

  In this chapter, we will turn to two types of studies that do reveal something about IH and how it affects the brain. The researchers were looking for what IHers have in common, as well as what distinguishes them from other groups of sleepy patients, such as people with narcolepsy or obstructive sleep apnea. These studies are examples of insights that the increased attention to IH has made possible.

  PLANTE’S PSYCHIATRIC PERSPECTIVE

  The first study we will examine was conducted at the University of Wisconsin’s sleep clinic. Its director, David Plante, comes to IH research with the perspective of a psychiatrist—he told me that his training background in psychiatry has given him a “pragmatic and symptom-based lens” through which to view IH. A major theme of his published work has been on the overlap between IH and depression. “Depression and hypersomnolence are intertwined,” he said at a talk for the Hypersomnia Foundation in 2019. “Depression is a common presumptive diagnosis in IH, yet many studies seek to exclude depression in cohorts of people with central nervous system hypersomnia.”

  While the majority of people experiencing depression report disturbed sleep or insomnia, about a quarter display hypersomnolence, Plante said. Most research on sleep in depression focuses on insomnia, but hypersomnolence has been well documented as a feature of atypical depression. In the context of depression, hypersomnolence has also been associated with impairment in daily life, resistance to treatment, and suicide risk.9

  Plante noted that in the DSM-5 diagnostic manual used by psychiatrists, the category that is the equivalent of IH—hypersomnolence disorder—is more flexible than in the International Classification of Sleep Disorders. In particular, it is more driven by subjective symptoms than IH (box 16.1). The MSLT is sometimes used in sleep clinics to differentiate IH from a psychiatric disorder such as depression, but this distinction may not be meaningful, he said.10