The discoveries did spur clinicians in Boston to begin treatment of one young woman’s early-onset narcolepsy with natalizumab (Tysabri), which is approved for multiple sclerosis and is thought to stall the advance of T cells into the nervous system.38 The patient herself wrote a blog post saying: “I still don’t know if Tysabri is helping me and whether my spinal fluid hypocretin levels might have stabilized or even increased with this treatment, but it feels good to know that I might have kick-started research into a new treatment approach focused on reversing this disease early in its course, rather than just managing a lifetime of symptoms.”39

  After the Swiss and Stanford papers were published, others jumped in with findings supporting the autoimmune mechanism.40 Precisely how immune destruction began was still unanswered. To attack hypocretin neurons, killer T cells would be needed in the hypothalamus—but in general, DQB1*0602 and other “class II” HLA proteins direct the activation of helper T cells, not killers.

  Some investigators argued that the autoimmune mechanism was not proven, because killer T cells that could be activated by hypocretin can be found in the blood of people without narcolepsy and autoreactive T cells have not been observed in the nervous system or CSF.41 Mignot’s most recent paper on the topic downplayed the concept of molecular mimicry, after an examination of autoreactive T cells did not find any activated by both influenza and hypocretin peptides.42

  For comparison, multiple sclerosis is well established as an autoimmune disorder, and the basics of the disease’s destruction of myelin—the insulating material surrounding neurons’ axons—have been known for years. But extensive research has not suggested that a single decisive autoantigen drives pathology, despite accumulating evidence for Epstein-Barr virus infection as a trigger.43 In narcolepsy, the available evidence suggests a scenario where someone has preexisting inflammation in the upper respiratory tract, followed by additional events—in some cases, vaccination—prompting T cells to enter the nervous system and begin hunting for hypocretin-producing cells.

  LIMITED EVIDENCE FOR PARTIAL HYPOCRETIN LOSS IN TYPE 2 NARCOLEPSY

  Neurologists have been attracted to the idea that an incomplete loss of hypocretin-producing cells could result in sleepiness without cataplexy. Perhaps a higher threshold of hypocretin loss is required for cataplexy to appear. With an intermediate amount of hypocretin production, someone might be sleepy but still have enough hypocretin to trickle out into the CSF. The trouble is that visualizing hypocretin neurons while someone is alive, without poking into the hypothalamus, isn’t possible at this point. For people with Parkinson’s disease, doctors can look for the loss of dopamine-generating neurons via a PET scan. Something analogous for narcolepsy doesn’t really exist.44

  Preserved brains from people with narcolepsy without cataplexy are scarce, so direct evidence supporting the “partial hypocretin loss” idea has rested on only a few examples. One was a man diagnosed with narcolepsy whose brain was studied by Siegel and colleagues at UCLA.45 They interviewed two of the man’s friends after his death from colon cancer to check that he had never experienced cataplexy. His MSLT results were consistent: in each of three tests, he fell asleep quickly and entered REM sleep twice. In the man’s brain, almost complete hypocretin loss was visible in the posterior part of the hypothalamus, but the anterior was relatively intact.

  More indirect evidence comes from a case report from the University of Bologna.46 A woman first became sleepy at age thirty-nine but fit MSLT criteria for IH. At age forty-three, she began to experience disrupted nighttime sleep and entered REM sleep three times on a second MSLT. Cataplexy appeared at age forty-five—and at that point the level of hypocretin in her CSF was lower than before. Thus, she gradually moved across all three diagnoses: IH, narcolepsy type 2, and, finally, narcolepsy type 1, presumably as hypocretin neurons were eliminated.

  Additional supportive evidence comes from some people with Parkinson’s disease, who have both excess daytime sleepiness and degeneration of hypocretin neurons, without low CSF hypocretin levels. Also, mice engineered with a partial loss of hypocretin-producing cells have mild narcolepsy-like symptoms.47

  UNFINISHED BUSINESS

  Most people diagnosed with narcolepsy or IH in the United States have reached that status by completing an MSLT and have never undergone a CSF hypocretin measurement.48 Without a specific test, people who fit a narcolepsy diagnosis and don’t have cataplexy may have hypocretin deficiency—or they might not.

  “I just want to know,” said Stacy Erickson Edwards, a woman living near Seattle who was diagnosed first with sleep apnea and then IH but suspected that she was developing cataplexy. “I’ve been trying to convince my doctors. We have good insurance, but they tell me the procedure is too invasive, and there’s no point. And I’m not crazy about my IH diagnosis, because to some people it sounds fake.”

  A commercial test for CSF hypocretin, performed by a federally certified lab, did not become available in the United States until recently. Testing for the DQB1*0602 risk factor was more accessible but could not confirm a diagnosis. Mignot tried to convince several companies to develop a commercial test but was repeatedly told the market for narcolepsy diagnosis was too small. His lab at Stanford provided CSF hypocretin measurements as a service to researchers and sleep specialists around the world but had to stop in 2017. Mayo Clinic Laboratories filled the gap in 2019.49 “It’s important for patients to be objectively tested, rather than for physicians just to rely on clinical history,” said the Mayo neurologist Michael Silber, in a statement explaining the value of the test. “Because it’s a lifelong disorder, we treat it with potentially harmful medications, and there’s a big differential diagnosis of other causes of sleepiness that could be confused with narcolepsy.”

  In some European countries, CSF hypocretin is measured more often as part of narcolepsy diagnosis than in the United States. But because lumbar puncture carries risks, CSF hypocretin is not like insulin in the blood, something that can be checked routinely.

  ANOTHER AUTOIMMUNE TARGET?

  The “incomplete hypocretin loss” idea is plausible for some people with narcolepsy type 2 or IH, but not enough sleepy human brains have been sliced and stained to fully evaluate it. Another mechanism might be reduced neuronal sensitivity to hypocretin, analogous to type 2 diabetes.50 Damage to the hypothalamus outside of hypocretin neurons may be a cause, but to produce something like narcolepsy type 2, the injury has to perturb the regulation of sleep while avoiding other vital functions. Also, for IH, unlike narcolepsy type 2, there is no known connection with HLA genes. That doesn’t mean an autoimmune mechanism is impossible, but it would have to happen in a different way.

  Speculatively, the immune system might be provoked to excise another group of neurons in the hypothalamus, in areas such as the paraventricular hypothalamic nucleus51 or the tuberomammilary nucleus (TMN), which produces the wake-promoting neurotransmitter histamine. Neurons in the TMN increase in number in people with narcolepsy type 1, possibly to compensate for the loss of hypocretin neurons.52 Adult mice with an engineered deficiency in histamine have trouble staying awake, spending more time asleep during both light and dark periods than wild-type mice.53 While these similarities are suggestive, more research is necessary to identify the neurobiological basis for both narcolepsy type 2 and IH.

  CHAPTER 14

  EVERYTHING OFF LABEL

  After all, we have all needed to write letters of medical necessity for patients with narcolepsy who were told that our prescribed treatment was not approved.… These issues are just as important as determining how many SOREMPs are on the patient’s MSLT.

  —Alon Avidan, 2012

  Until Balance Therapeutics started clinical trials testing PTZ in 2015, no company had ever pursued approval for the IH indication—but twenty years earlier, a similar vacuum had existed for narcolepsy. As a guide to the obstacles pharmaceutical companies face with IH, we can learn a lot by examining the past behavior of both industry and the FDA with respect to narcolepsy. Most of this chapter will explore the history of modafinil, the first medication specifically approved for narcolepsy.

  First, we should review why the lack of drugs approved by the FDA for idiopathic hypersomnia has been a major concern of the U.S. hypersomnia community. Conventional stimulants are available and familiar to many IHers, but they have their limitations and drawbacks. One IHer wrote on a Facebook thread: “I have never actually liked Adderall because it really does make my head feel like a jumbled mess with the unpleasant physical sensation of too much caffeine, but it has stayed my drug of choice because as long as I stick to my schedule, it’s the only thing that keeps me out of bed and off the couch.”1

  For reference, a survey of the Hypersomnia Foundation’s registry found that a majority of IHers on medication were taking conventional stimulants such as amphetamines or methylphenidate. The second-most-popular options were modafinil or armodafinil, used by 38 percent of IHers.2

  If someone has an IH diagnosis, any prescription written by their doctor was considered “off label.” It is illegal for pharmaceutical companies to promote their products for off-label purposes: indications for which they did not have FDA approval. Doctors do have the professional capacity to prescribe a medicine for a patient if they think it is appropriate, and off-label prescriptions are common in oncology, psychiatry, and pediatric medicine.3

  While physicians can practice medicine independently, insurance companies in the United States generally take their cues from the FDA. In 2007, the AASM (American Academy of Sleep Medicine) provided advice for the treatment of narcolepsy and “other hypersomnias of central origin,” but it took years for insurers’ policies to reflect that guidance.4 An AASM panel later recommended modafinil as a first-line treatment for IH because clinicians had experience with it and there was much evidence for its efficacy and safety—but that’s not the same as an official FDA stamp of approval.

  Until recently, IH was considered too rare and possibly too poorly defined for a pharmaceutical company to bother seeking approval for one of its medications. To use a legal analogy, IH lacked standing with the FDA. IH’s lack of standing played out in the bureaucratic battles IHers waged against their insurers. In 2018, a member of the Atlanta hypersomnia support group invited me to review what they and their spouse dealt with. It typically began when they tried to get a prescription filled for modafinil: FDA approved for excessive daytime sleepiness associated with narcolepsy, sleep apnea, and shift work—but not for IH. The first obstacle the couple would routinely face was a requirement for prior authorization: a filter for expensive, risky, or possibly unnecessary medications. Their doctor, Lynn Marie Trotti, would write a support letter, but the request would still be denied. The IHer’s spouse showed me one of the insurance company’s letters, which cited the FDA-approved indications for modafinil, with IH not on the list. Usually at least one additional level of appeal was necessary. “Most people would just stop at the first appeal denial,” they said.

  The IHer had unfulfilled education and career plans that IH and other conditions had eroded. The spouse had worked at a large company, where the couple had been able to eventually get coverage for both modafinil and flumazenil after a few months. But when the spouse joined a startup and changed health insurance, they had to start all over again. Sometimes a paperwork delay would knock them back down a rung on the ladder of approvals. The Sisyphean process forced the couple to pay for both modafinil and flumazenil out of pocket, although modafinil was relatively affordable at a discount pharmacy. Financially, they were just surviving and treading water, since a significant amount of the spouse’s salary went to support the IHer’s medications.

  In their letters to insurance companies, they’d tried citing research on the fuzzy distinction between IH and narcolepsy type 2. That didn’t seem to work; for effective was persistence—and making a personal connection with someone working for the insurer. Getting coverage for flumazenil, which required a special order from a compounding pharmacy, was even more of a challenge. Citing past adverse reactions to other medications such as methylphenidate and clarithromycin was helpful.

  BUREAUCRATIC BARRIERS

  Doctors in the United States have reported that prior authorization causes delays that can hurt patient care. In a 2021 American Medical Association survey, one-third of physicians reported that such delays have caused at least one of their patients to experience a serious problem, such as hospitalization, the development of a birth defect, disability, or death.5 In that same survey, more than 80 percent of surveyed doctors said their patients sometimes abandon recommended treatment because of prior authorization obstacles.

  People with narcolepsy have prior authorization challenges too, but the extra difficulty associated with IH has been a reason why IH has been undercounted. Physicians sometimes assign patients with IH a narcolepsy diagnosis instead, to make it easier for their patients to obtain prescribed medications.

  A 2019 survey conducted by the Hypersomnia Foundation makes the burden clear. It found that more than 70 percent of people with IH reported having prescribed medications initially denied by their insurance company. Appeals were successful about half the time, and a third never started the appeals process.6

  The foundation’s survey also found that 30 percent of people with IH have chosen not to fill a prescription because of out-of-pocket costs. Half were facing costs of more than $250 per month, substantial for a group of people with a chronic disorder that can interfere with employment. The survey indicated IHers’ out-of-pocket costs were comparable to those of people with multiple sclerosis, the most expensive in a 2016 survey of common neurological conditions.7

  For IH and prior authorization, health insurers’ policies varied. Those who obtained coverage through Medicare and Medicaid were subject to state-by-state variations. Openings have come incrementally as a result of pressure from physicians and patients. In 2020, the largest, United Healthcare, stated that coverage for IH could be allowed for modafinil, armodafinil, and conventional stimulants, while others’ policies did not mention IH.8 (This picture did change after 2021.)

  GATEWAY TO BILLIONS

  Before the 1990s, no medication was specifically FDA approved for narcolepsy. Narcolepsy was generally treated with amphetamines and REM-suppressing tricyclic antidepressants, which both have a long list of unpleasant side effects. In 1982, the American Narcolepsy Association complained to Congress that “almost nothing is being done to find better medications for treating the symptoms of narcolepsy.”9 A rare disease such as narcolepsy offered little incentive for a company to conduct clinical trials and develop alternatives.

  The 1983 Orphan Drug Act was designed to address this problem, which applies to rare diseases in general. It set out incentives such as seven years of market exclusivity after approval, grants for exploratory studies, and tax credits for up to half the cost of clinical trials. Decades later, the Orphan Drug Act is widely regarded as a success. In fact, critics have argued that it has gone too far, pushing companies to develop expensive therapies for rare diseases. Companies could then turn around and repurpose their medications for more widespread conditions.10 As of 2019, almost a third of drugs in the industry pipeline targeted rare diseases, and rare diseases represented the majority of new drug approvals.11

  Modafinil, previously known by its brand name of Provigil, is an example of the “orphan-to-blockbuster” phenomenon. It was the first medication explicitly approved for excessive daytime sleepiness associated with narcolepsy in 1998. But modafinil’s use by people who didn’t have narcolepsy quickly grew dominant commercially. It demonstrates how billions of dollars in revenue lie in the space between caffeine and amphetamines and how an aggressive pharmaceutical company used narcolepsy as a gateway to more lucrative opportunities.

  Surgeons and pilots took the drug in order to stay awake longer and complete their duties safely. So did college students studying for exams and computer programmers looking to maximize their productivity.12 In 2013, New York magazine proclaimed modafinil “Wall Street’s New Drug of Choice.” In 2018, White House physician Ronny Jackson was reported to distribute modafinil freely to presidential aides on overseas flights, a practice considered mildly scandalous.13

  Modafinil has represented a kind of forbidden fruit for many people with chronic illnesses. The novelist M. J. Hyland, who has multiple sclerosis, has written about difficulties convincing the United Kingdom’s National Health Service to allow her a prescription: “The problem isn’t the drug, but the curse of its image: the stigma of its ‘recreational’ use; the idea that the benefit of cognitive endurance is inherently greedy and frivolous, in the same dirty class as steroids, Botox and Viagra.”14

  ON ITS FEET AND FIGHTING

  Modafinil’s early development was driven by the French sleep research giant Michel Jouvet. Starting in the early 1980s, Jouvet and his colleague Helene Bastuji in Lyon prescribed modafinil on a limited scale to both narcolepsy and IH patients.15 They reported that it was safe and effective at combating sleepiness, although it did not suppress cataplexy.

  In the 1970s, modafinil’s manufacturer, Lafon Laboratories, had discovered a related drug called adrafinil and marketed it as a cognitive enhancer for the elderly. Still, Lafon was reluctant to develop modafinil for narcolepsy because of the limited market size. Jouvet convinced the company to start clinical trials and promoted potential military applications for the drug, leading to its use by the French military in the 1991 Gulf War. He told a Paris defense conference that modafinil “could keep an army on its feet and fighting for three days and nights with no major side-effects.”