IH less rare than we thought?
A rare sleep disorder — idiopathic hypersomnia or IH, which I wrote an entire book about — may be not as rare as previously thought.
That was the headline from a press release put out last week by the American Academy of Neurology, about a study published in the journal Neurology. The press release got a smidge of attention from media outlets. Healthline went as far as asking unaffiliated experts about it, which was nice to see.
The number the Neurology paper arrives at, based on data from a long-running study from Wisconsin, is quite high: 1.5%. It would make IH prevalence comparable to that of better-known neurologic/psychiatric conditions such as epilepsy, schizophrenia or bipolar disorder.
That would be amazing, but the 1.5% number is probably too large, for two reasons. First, in my opinion, the threshold of subjective sleepiness is too low. It’s possible to get an Epworth Sleepiness Scale score of 11 out of 24, the minimum for inclusion, without saying that you have a “high chance of dozing”, as opposed to moderate or slight, in any of the eight situations in the ESS questionnaire. (These investigators agreed with my assessment of ESS thresholds.) Also, according to Table 1, only half of the 12 people flagged as having IH said they had an “uncontrollable urge to fall asleep during a typical day.”
Second, the identified study participants all fell asleep quickly in the laboratory, but there was no requirement for them to have decided that their sleepiness bothers them enough to seek help for it. None had been prescribed stimulants, although some were taking antidepressants or thyroid supplements.
The Wisconsin study should be seen as asking: How many people are objectively sleepy enough, and *could* be diagnosed with IH if they were coaxed into a sleep specialist’s office? As the Neurology authors conclude, “there is likely a sizeable discrepancy between the number of persons with IH relative to those who seek treatment.”
I agree with the stance taken by Johns Hopkins sleep specialist Mark Wu, who was quoted by Healthline: “Hey, 1% of people out there are really sleepy. And then probably a subset of them have idiopathic hypersomnia.” I do think IH is under-recognized and needs more attention. And there’s a lot of room between previous estimates of IH prevalence and the big numbers from this study and from Maurice Ohayon’s work, which used phone surveys.
(I actually talked about different ways of measuring IH prevalence this summer at the Hypersomnia Foundation’s conference — see below. The Wisconsin study was not published at that point, but I mention a presentation by one of the authors, Paul Peppard, near the end.)
The most realistic estimates of IH prevalence in the United States have relied on looking at insurance claims. This gives a much smaller number of around 1 in 10,000 people. But the insurance approach is asking a different question; it’s more like asking: “How many people have actually been diagnosed with IH?” This approach has some shortcomings because some people who could get diagnosed with IH lack access to healthcare, or they get bogged down within the healthcare system and are told they have depression or some other condition.
The Wisconsin study used the MSLT (Multiple Sleep Latency Test) to measure physiological/objective sleepiness, yet experts in the field — including authors of the 2023 Neurology paper — have published studies highlighting the limitations of the MSLT. Many patients subjectively experience excessive daytime sleepiness, but have a difficult time demonstrating it in the laboratory through the MSLT. One of the authors, David Plante, has been testing alternative methods of measuring excessive sleepiness and has shown that using those methods could be more inclusive.
As another author, Emmanuel Mignot, remarked to me: “It’s hard to know how to slice the pie.” For example, according to studies from France, people with IH who display long sleep duration often don’t fall asleep quickly in the laboratory. None of the people flagged as having IH in the Wisconsin study had long sleep duration, but full-blown IH with long sleep is going to be rarer.
Some limitations of the Wisconsin study: to be part of it, someone needed to be employed by state government. Out in the wider world, IH often interferes with the ability to work. Also, most of the people in the Wisconsin study are middle-aged, while IH is typically diagnosed at a younger age, when someone has trouble staying awake in school. It would be wonderful to see a study of IH from another country’s healthcare system (like this one examining narcolepsy prevalence in Germany).
The other interesting finding from the Wisconsin study is that a large number (about 40%) of the people flagged as having IH experienced remission, in that they later reported being not as symptomatically sleepy. My friend Michelle Chadwick of Hypersomnolence Australia has argued that apparent “remission” means that an individual was misdiagnosed. Now, the business of deciding who is sleepy enough and who has genuine IH versus something else is tricky, and so is second-guessing two respected experts in the field! But my reading of the trajectory graph in Figure 2 is that the people who reported higher ESS scores were less likely to display symptomatic remission.
Looking ahead, a host of experts have proposed to revise the classifications of narcolepsy and IH in the ICSD (International Classification of Sleep Disorders) to better reflect clinical reality. At the APSS meeting this summer, when Plante unveiled ICSD-3 revisions for hypersomnolence disorders, he said that their scope was limited. Big changes to narcolepsy and IH were not on the table and more data are needed for any future realignments, he said. After this study, it’s worth asking — what else do we need?
