A ReLACSing Blog #22: 15 Reasons Narcolepsy Type 1 is Over-diagnosed: Part I

Narcolepsy type 1, or the artist formerly known as narcolepsy with cataplexy, is the one central disorder of hypersomnolence (i.e. brain condition that causes excessive sleepiness) that has a well known biological cause and specific and objective clinical testing. Why is it then that the majority of patients are misdiagnosed with this condition? This lengthier-than-usual blog addresses 15 important reasons over two sections.

Early in my career, I somehow was able to put together a review paper entitled Disorders of Excessive Daytime Sleepiness Including Narcolepsy and Idiopathic Hypersomnia with the help of a colleague at the University of Michigan. Since that time, however, my concern about the field’s approach to these conditions has grown by magnitudes and my criticisms have become considerably sharper, as you will see here. Having encountered a swath of second opinions and transfers of management from more than six years of academic sleep medicine, the misdiagnosis of narcolepsy type 1 (NT1) is rampant. I have already addressed the sort-of made up conditions of narcolepsy type 2 (NT2, the artist formerly known as narcolepsy without cataplexy) and idiopathic hypersomnia (IH) that really have no biological basis currently. It was rather stunning that so many patients were misdiagnosed with NT1, which does have known biology and objective and accurate testing available. NT1 is a permanent neurological condition from one’s immune system destroying an area of the brain that regulates sleep. An individual loses their ability to regulate sleep-wake cycles, which results in excessive daytime sleepiness and host of other symptoms, most notably cataplexy, along with sleep paralysis, hallucinations when waking up or falling asleep, dream enactment behavior, and going directly into dream sleep (REM) upon falling asleep. It can be a debilitating condition.  

In my experience, more than 90% of patients seeing me from another sleep provider did not have NT1 but were carrying the diagnosis. To do my poor Jim Gaffigan audience member side comment impression: But Dr. Berkowski, how can you prove this? Isn’t this just your opinion? This guy doesn’t believe in anything. Are there any real sleep disorders? Maybe this guy needs some sleep himself! Yes, much of this is my clinical judgment and another doctor could see the same patient and disagree. However, there is OBJECTIVE testing. It is much harder to argue against an evidence-informed, objective test result than someone’s subjective assessment.

Part of my cockiness (your word, not mine)…err…confidence in this claim is that my record has been nearly undefeated when I have suspected a false diagnosis of NT1 and then the patient had the objective testing. Most, however, are ties because most of the time the objective testing never gets done (as I will explain in Part II #15), so maybe you could say my record is in the ratio of 4-0-10 (W-L-T; we’re not using the NHL terminology of overtime losses here because there are no shootouts in central disorders of hypersomnolence). I was blessed to have outstanding fellowship training at Stanford, one of the world’s first sleep centers, built by Guilleminault and Dement and their medical progeny, with the late Dr. Guilleminault providing me with clinical supervision several days per week. At the time I was there, Drs. Chad Ruoff and Emmanuel Mignot delivered outstanding teaching and mentorship in the area of narcolepsy. Mignot’s lab discovered that the cause of NT1 was the immune-mediated destruction of the hypocretin-(aka orexin) producing brain cells in the hypothalamus. Mignot could take a patient’s cerebrospinal fluid (CSF) sample, run it over to his lab, and see if there was any hypocretin present. In other words, from working with Mignot, you got to take the test a lot and see the answer key quite often: is it NT1 or not NT1? Again, you ask, why is NT1 so hard to diagnose correctly?

#1 An army of untrained sleep specialists: Prior to about 10–15 years ago, any physician could sit for the sleep medicine board examinations and get certification. Roughly 20—25 years ago, formal fellowship training in sleep medicine did not exist. Thus, you have a generation of physicians, many of whom were essentially self-taught with textbook knowledge. This has led to a wide discrepancy in practice abilities unique to this subspecialty. Clinical training is essential in medicine. Medical schools would not spend 2+ years with clinical rotations if hands-on, supervised learning were not important. Residencies in any specialty are a minimum of three years, most of which is clinical training. Thus physicians generally have 5+ years of experience in supervised patient care, sometimes double that amount. In sleep medicine, many just took an exam or maybe even just practiced under the umbrella of their primary specialty to expand to treat sleep patients. Additionally, sleep medicine became a pre-retirement option for many critical care/pulmonary physicians and neurologists who wanted a less demanding practice situation. Perhaps some treated it casually as a result. There was a huge boom in sleep testing more than two decades ago that also made it profitable, but the focus was more on sleep apnea. Some may have gone into this specialty with financial motivations (no longer the case!). For all these reasons, physicians treating sleep patients may not have the experience or knowledge to manage rare and complex conditions such as NT1, leading to incorrect diagnoses and management.

#2 Poor understanding of cataplexy: As I briefly discuss in this video, cataplexy is a highly unusual phenomenon in which the signals from the brain to certain muscles are interrupted and the tone of the muscles decreases for a split-second to seconds. Think of a brown out with the lights in your home flickering. Also unusual is that cataplexy may be triggered by strong positive emotions of all things, such as anticipation of humor, telling a joke, laughing, or even other emotions like surprise. Cataplexy can occur in some of the facial muscles, eye muscles, shoulders, and even legs if severe enough. Someone with cataplexy can find something humorous and the head can bob or the eyelids start drooping for a few seconds until the individual can have the emotions pass. It is not clear why certain emotions trigger cataplexy, but the movements are due to the sleep-stage disruption caused by the loss of the hypocretin-producing brain cells in the hypothalamus. Normally during rapid eye movement sleep (REM) the body is essentially paralyzed with the exception of very few muscles (eye muscles for example) in part preventing a person from acting out dreams. This is why we don’t sleep standing up like giraffes, elephants, zebras or whatever animal I cannot remember sleeps upright (horses? what’s the difference between a zebra and horse? I’ll tell you that narcolepsy is a zebra, not getting enough sleep is a horse). In NT1, the brain may send the signal that decreases the tone of the muscle, just like it would in REM sleep, but it may do so while the individual is awake! Thus a person with cataplexy is awake and conscious but going into the REM-associated muscle atonia, with shoulders starting to slouch, all the while surprised that their 14 year-old did the dishes out of the blue while they were out shopping for groceries. If they had cataplexy in the arms or shoulders, they may have even dropped the groceries with the eggs cracking and spilling on the floor. I’m confident the 14 year-old would volunteer to help with the cleanup however.

There are so many mimics of cataplexy. Passing out is not cataplexy; patients do not get symptoms of lightheadedness or other symptoms of almost passing out (vasovagal), like flushing, increased heart rate, nausea, etc. This would be the 14 year-old that was dehydrated and overheated (having already done the dishes and cleaned up spilled eggs that morning), becoming lightheaded in the church choir on Sunday morning and nearly fainting. The boy almost passed out, but he does not have cataplexy like mom. Seizures are not cataplexy. Falling asleep suddenly is not cataplexy. There is no muscle weakness that lasts any substantial length of time. The muscles are not weak, they just have lost their input for a few seconds. All of these and more can be confused with cataplexy. If cataplexy is exclusively linked to NT1, and if someone is very sleepy and has symptoms similar to these cataplexy mimics, then you can see how someone can be misguided down the path of NT1 diagnosis.

Henry Fuseli, Public domain, via Wikimedia Commons

#3 Assumption that only those with NT1 can have sleep paralysis and other sleep-related symptoms: Though true cataplexy can really only be due to NT1, other symptoms associated with narcolepsy: excessive daytime sleepiness, frequent awakenings at night, hallucinations when falling asleep or waking up, sleep paralysis, and vivid dreams can be seen in healthy individuals under the right circumstances. I had an episode of sleep paralysis a few weeks ago, and it is quite uncomfortable, but I couldn't help but think how many providers only associate it with narcolepsy. Sleep paralysis is a corollary of cataplexy in that your brain is coming out of REM sleep into wakefulness but the muscle atonia/paralysis is still present so you are essentially waking up paralyzed. This is depicted in Fuseli’s painting “The Nightmare” as an incubus standing on top of a woman preventing her from getting up from bed. Many patients with shift work, irregular sleep-wake patterns, highly disrupted sleep, or using the snooze button on the alarm excessively can have more frequent sleep paralysis. 

#4 It is a logical flaw to screen for exceedingly rare conditions: If you look at most screening questionnaires at a sleep clinic or prior to doing a sleep study, you’ll see questions related to obstructive sleep apnea (OSA), insomnia, restless legs syndrome (RLS), parasomnias, and maybe circadian rhythm (biological clock) disorders. Sidling up to these common disorders is good old narcolepsy as if you don’t screen for it, you may miss it. The prevalence of the following is roughly OSA 20%, insomnia 10%, RLS 5%, sleepwalking 2%, circadian rhythm 3%, etc. Narcolepsy type 1? 0.014% according to this study. Does it make sense to ask about symptoms of all these conditions as if they were in the same stratosphere? OK, you’re saying, this is maybe a waste of time but what’s the harm? There can be significant harm. There is a scientific way to approach screening. False positives can be very harmful. There is a risk and benefit to doing any test. Currently, it is more beneficial to do a colonoscopy at age 45 but not ages 4–5. This is because there are rare risks of doing the procedure, but more importantly, there are more significant risks of false positives–finding things that were not meant to be found and the cascade of effects as a result of addressing what was found. In the case of NT1, false positives on inappropriate screening questionnaires can lead to further testing potentially and leads us to the next number.

#5 The abysmally non-specific, inaccurate, imprecise, and poorly interpreted “objective” study known as the MSLT: The multiple sleep latency test (MSLT) is a series of 4–5 naps spaced out by two hours throughout the day, following an overnight sleep study (PSG/polysomnogram). It determines how fast an individual falls asleep and whether they can enter REM sleep within 15 minutes of falling asleep, which is highly abnormal. During a fellowship lecture, Dr. Bill Dement told us that he and Dr. Mary Carskadon developed this test about 40 years ago to illustrate how uniquely abnormal the sleep architecture was of those with NT1 compared to healthy individuals. This was to be the signature test to diagnose narcolepsy. He said he did not know at the time how many individuals without NT1 would have similar results. This is why NT2 and IH have come about, as the MSLT showed abnormal results in people who did not have NT1, and the clinicians could not figure out why, so they needed new diagnostic categories. 

Later in the development of the cutoffs and numbers associated with the test, according to Emmanuel Mignot, a group of clinicians got in a room to decide what was normal. He said they debated what was normal versus abnormal in terms of falling asleep. Some thought 3 minutes, others 10 minutes. They finally settled on 8 minutes as the magic number. Unfortunately, many people can fall asleep in these five naps in less than 8 minutes on average (including medical residents) and have absolutely no medical problem other than a poor schedule the few days before the test. The test simply tells how fast a person could fall asleep on that day and under those circumstances. It is non-specific. This means that the results can be consistent with a condition like NT1 but other people without the condition can have the same results. Many papers have come out showing that this test is highly imprecise, even in one from my mentor Dr. Ruoff. Though the precision is better when the test is repeated for NT1, largely because this is a known permanent condition that remains more constant, it is utterly useless to diagnose NT2 or IH, even though these conditions were created from the results of this test! Over-screening for NT1, leads to over-testing with the MSLT, which leads to false positives, even for NT1. 

Can you tell that this person fell asleep?

#6 MSLT is prone to human error: We are getting closer to AI algorithms interpreting sleep studies, but there are still humans who have to interpret an MSLT. A sleep technologist first scores the MSLT by identifying the 30 second epoch of squiggly EEG lines visually on the study in which they feel the patient went into stage N1 (light) sleep for the majority of the 30 second period. This can be difficult to eyeball sometimes. They also have to identify if an epoch meets criteria for REM sleep. Is that a rapid eye movement? Or did the patient pull out their phone and is scrolling up and down their Facebook news feed? After the tech scores the study, a board-certified sleep physician has to over-read the scoring and provide interpretation. This then hits on point #1, which relies on training. Another important point is how much time does the physician have to look at this stuff in this medical environment? I have seen many misinterpreted MSLTs that went from narcolepsy-like to normal in a matter of a few oversights.

#7 Has Anyone Heard of the HLA DQB1*06:02 allele?: That’s the last time I am going to write DQB1*06:02. Henceforth we shall refer to the HLA DQB1*06:02 allele as the HLA for narcolepsy. Not to get into the weeds on this, but an allele is like a flavor of a gene, and in close to 99% of cases of hypocretin deficiency (objective NT1, see Part II of this blog), they have this narcolepsy HLA. HLA stands for Human Leucocyte Antigen and it is a powerful component of the immune system and thought to be directly involved in the immune system’s attack on hypothalamus, leading to the destruction of the aforementioned hypocretin-producing neurons. 

The narcolepsy HLA is virtually 100% sensitive but not specific at all. In other words, having this allele is necessary to have NT1 but not sufficient, not even close to sufficient. About 25% of the general population can have the HLA allele for narcolepsy but clearly 25% of the population does not have the condition, maybe 0.0025%. Almost 100% of the other 75% of people cannot have NT1 because they lack the HLA for narcolepsy. (Please cue up the clip from Anchorman: “60% of the time it works every time”). The upshot is that if a patient does this genetic test and does not have this HLA, they cannot have NT1. Find another diagnosis. The book is closed. 

Unfortunately, most clinicians are not aware of this test. Again, in 6+ years in academic sleep medicine, I encountered only a few who had this test performed but had a diagnosis of NT1 before I saw them. It’s so unknown, that a doctor reviewing the case for insurance coverage had never heard of it and he called me to send him literature to educate him. He approved the test. The patient never took the test (see #15 in Part II). This should be done on everyone with suspected NT1. Why play around with the risk of labeling a person with this condition who does not have it? I have seen some testing kits online for $195. I cannot verify the quality of some of these versus at a medical lab, but the cost will not break the bank regardless of when the insurance industry wakes up (pun intended) and pays for the simple test. If this can definitively exclude the diagnosis of a permanent brain condition leading to inability to stay awake or regulate sleep cycles, wouldn’t most people pay less than $200 to know this??? The answer turns out to be “no” and not because of the cost…

Well, I have gone on too long! Let’s pause here and finish the remaining eight reasons next week for Part II.

-Andy Berkowski, MD of ReLACS Health, who has heard that though Google search algorithms often prefer long-form 5000+ word blog articles as more authoritative on a subject, most readers cannot stomach writings this long in this day and age with social media and attention spans an’ all. Then again, how many great double-albums were there? (My favorites are The Beatles’ White Album, Pink Floyd’s The Wall, and a distant third and fourth to Smashing Pumpkins’ Mellon Collie and the Infinite Sadness and Red Hot Chili Peppers’ Stadium Arcadium)