Circadian Sleep Disorders Network - Advocating for people with misaligned body clocks
Printer-friendly page
ON THIS PAGE (links): Top
In a Nutshell
1. Scope
2. Prevalence
3. Causes
4. Diagnosis
5. Treatment
6. Dissemination
A General Comment

Circadian Sleep Disorders Network

June 1, 2019


  1. Researchers do not even agree on how to define circadian rhythm disorders.
  2. We do not know how many have these disorders.
  3. We do not know what the underlying causes are.
  4. We do not have a quick way to diagnose these disorders.
  5. We do not have good treatments.
  6. What we do know often isn't implemented by clinicians.
Each of these gaps in our knowledge corresponds with (and links to) a numbered major topic below.


Circadian rhythm disorders (CRDs) are hidden afflictions. Patients generally look well—if perhaps rather tired at times. We seem normal on a physical exam: you can't palpate our suprachiasmatic nucleus or listen to the beat of our internal clock with a stethoscope. There are some lab tests (e.g. DLMO), but they are not widely available and are hard to perform, needing to be done in the dark. No doctor or nurse is there to watch us lie in bed hoping to fall asleep: 2 a.m. . . . 4 a.m. . . . 6 a.m. . . . and sleep does not come. You have only our word for it, and maybe some sleep logs we've kept. It's easy to say we are doing something wrong. We aren't trying hard enough—or are trying too hard. We have no willpower. We aren't thinking the right thoughts. Our sleep hygiene isn't quite right.

Sleep deprivation is proven to be the most extreme form of torture known to humans. It is what totalitarian regimes use to break down resistance. But it's what many of us have to endure, week after week, year after year, if we wish to conform to a 9-to-5 society. Or we can sleep on our own schedule—getting up in the afternoon or later, or on a 25-hour cycle—but then what jobs can we get? What kind of social or family life can we have? (And even on our own cycle, many of us are foggy, lethargic, and tired, so it's not always a complete solution.)

Ours are invisible disorders, and so many doctors, family members, and friends don't believe us. Everyone has had to cope with a bad night of sleep from time to time or getting out of bed when they don't want to. Our conditions don't get a lot of sympathy. No one is going to host a telethon or a GoFundMe for us. But the suffering is real. Educations, careers, jobs, ambitions, social life, family life: all too often, they are destroyed. Health often suffers.

But in the midst of suffering and disbelief, we have one hope, one beacon in the 4 a.m. dark:


As the recent Nobel Prize showed the world, circadian rhythms are real. They are a fundamental aspect of all life on earth, from cyanobacteria to drosophila to humans. Circadian rhythms are literally written in our DNA. To think that a disruption of such a basic process could happen without causing a myriad of dysfunctions and symptoms is scientifically inconceivable. Science has finally caught up with what those of us with circadian disorders have known for years.

Researching CRDs benefits not only CRD patients, but everyone. Circadian health is increasingly recognized as foundational to physical and mental health, safety, productivity, and learning—to society as a whole. By studying disordered circadian rhythms, we can shed more light on the mysteries and complexities of circadian rhythms. We can better understand what makes our lives tick and how diseases take hold. Insights yielded from studying CRDs will have a direct effect on our understanding of diseases in many organ systems.

Science has told us why circadian rhythms are important. The next step, for medical scientists and clinicians, is to help alleviate the suffering circadian disorders cause: either by helping us change our rhythms or helping us live with them.

Much progress has been made, for which we are grateful. But more yet remains to be done. We hope that the cooperation of patient groups such as ours with the research community can help to bring about a new dawn in circadian research and treatment. It is in that spirit that we submit this document for your consideration.


Researchers should investigate the entire spectrum of circadian disorders, with regard to both diagnosis and degree of severity or impairment. There is a spectrum of disorders ranging from those which are relatively common but somewhat less impairing at one extreme, to disorders which are rare but severely impairing at the other. Mild cases of delayed sleep phase disorder (DSPD)—for example, someone falling asleep between midnight and 1 a.m.—do not appear to be rare (although precise statistics are lacking). However, such conditions can have a significant detrimental effect on a person's life and deserve medical attention due to the potentially large number of persons affected.

Other disorders—severe DSPD, irregular sleep-wake disorder (ISWD), and non-24-hour sleep-wake disorder (Non-24), especially the sighted form—may be less common. However, they can be extremely disabling. Many patients are unable to work or have a normal family/social life. Although the disorders are rare, the extreme severity of the impairment means they desperately need attention. As one paper stated regarding Non-24, "This rare syndrome is extremely debilitating in that it is incompatible with most social and professional obligations." (Oren DA, Wehr TA. Hypernyctohemeral syndrome after chronotherapy for delayed sleep phase syndrome. N Engl J Med. 1992 Dec 10;327(24):1762. NOTE: The paper misspelled "hypernychthemeral.") Not only are the sleep-wake cycles of these patients incompatible with most occupational and societal demands, but many also suffer from other symptoms such as profound fatigue and/or mental fogginess, which accompany the circadian disruption, for reasons not fully understood at present.

Both milder manifestations and severe manifestations of circadian disorders need attention, the former because of the large number of people affected and the latter because of the extreme degree of disability involved.


a. Determine the Prevalence of DSPD and ASPD

There have been a few studies of prevalence of DSPD, but they differ in how they define the disorder and how they select the population.

One that is often quoted (Schrader H, Bovim G, and Sand T, The prevalence of delayed and advanced sleep phase syndromes (J Sleep Res. 1993 Mar;2(1):51-55)) is a Norwegian study that selected 10,000 adults at random and used a questionnaire followed (for likely cases) by a telephone interview. That provided a figure of 0.17% (about 1 in 600). It included people as young as 18, which might skew the result, and Norway is at very high latitude resulting in wide variation between summer and winter daylight. It didn't find any cases of advanced sleep phase disorder (ASPD). Other studies have included teenagers or were restricted to students, and resulted in much higher prevalence estimates.

We need additional studies that all use the same recognized definitions of DSPD and ASPD and select adults at random in the overall population. We also need studies on prevalence of CRDs in young children.

b. Determine the Prevalence of Sighted Non-24

With careful questioning we could also get a rough estimate of Non-24 prevalence in sighted people, which we suspect (from our survey—see next paragraph) is considerably higher than generally believed. Currently there is almost no evidence of the prevalence of sighted Non-24 and limited evidence for the prevalence of ISWD, mostly based on its occurrence in the context of other conditions such as head injury. (We do have good estimates of the number of Non-24 cases in the totally blind population.)

In our survey, 29% of respondents with CRDs believe they have Non-24. Some may have misunderstood the definition of Non-24. But of the people diagnosed with a CRD by a medical professional, 19% were diagnosed with Non-24. Our sample is biased by self-selection—people with Non-24 are more likely to have major problems with their lives and be more likely to participate in our survey. Still, the results suggest that the prevalence of Non-24 is much higher than is generally recognized. (For additional information about the Circadian Sleep Disorders Network (CSD-N) survey see Results quoted in this document reflect data from about 750 respondents.)


a. Study Multiple Causes

Circadian disorders may have different underlying causes in different patients. We need to investigate causes, tease out which causes are important for which patients, and tie the underlying cause to the recommended treatment.

While some occurrences of circadian disorders can be traced to very specific single causes—CRY1 mutation or another genetic mutation in some cases of DSPD, other genetic causes for ASPD, and total blindness in some cases of Non-24—more generally, circadian disorders appear to have multiple, sometimes overlapping etiologies. Causes may include:

as well as other possibilities yet unexplored, such as epigenetics, the microbiome, and endocrine disorders.

Researchers should also investigate causes of sighted Non-24, which is often overlooked. Too often, they focus on blindness, a recognized cause of Non-24—but not in sighted people.

This range of possible causes may explain why treatments don't work the same for everyone. One would expect, for example, that for people with DSPD who are not sensitive enough to light, very bright light in the morning is what works. For those oversensitive to light, light restriction in the evening may be key.

b. Investigate Amplitude and Other Factors

Research and diagnosis seem mainly concerned with circadian phase. Circadian rhythm also has amplitude: some evidence suggests that people with a high amplitude are very alert during their day (whenever that is) and sleep soundly at night; those with low amplitude are tired all day and don't sleep as well at night. Low amplitude, for example, occurs in ISWD and often as a result of aging. There is also a "waveform": some people (square wave) fall asleep quickly and wake up quite rapidly; others (more sinusoidal) fall asleep more gradually and waking up fully takes time. In addition, some folks can more easily shift their sleep hours, while others are quite inflexible. We would like to see more attention paid to these factors.

c. Determine Why CRD Patients Are So Tired

Many CRD patients feel tired, jet-lagged, and mentally foggy even when sleeping on their body's preferred schedule. In our survey 56% feel "somewhat tired" or worse, and 19% "very tired" or worse—even when sleeping on their body's preferred schedule. Is this related to internal desynchronization or phase difference of some rhythms, severe sleep inertia, or other factors? We greatly need better understanding of this phenomenon and better recognition of its occurrence.

Patients who force themselves to fight their natural sleep-wake cycle are, of course, fatigued. Is this because they don't get enough hours of sleep or because they sleep at the wrong time, or both?

d. Study Effects of Age

Many of our members report that their DSPD has gotten worse with age, both in progressing to later sleep times and in the inflexibility they experience when needing to get up early for work, family, etc. In our survey, 61% of those responding say their sleep time has become less flexible; 40% say it has become less regular as they have gotten older. Some members with DSPD and Non-24 report increasing problems with sleep consolidation and irregularity of timing, sometimes resembling ISWD or an exaggerated effect of aging on the sleep cycle.

There is conflicting research on the effects of aging on circadian rhythms, with some studies suggesting a phase advance and others finding no difference in phase or tau, suggesting that early morning awakening in the aged is a problem of sleep continuity. The idea of a phase advance leads some clinicians to expect DSPD or Non-24 to ameliorate with age, but this has not been the experience of our members. More research on circadian disorders and aging is needed.

e. Investigate Comorbidities—Chicken or Egg?

Often, the presumption is that medical and mental health conditions cause sleep disruption. Our survey data suggest it is often the other way around. Of those with a circadian rhythm disorder and depression, 63% said the CRD preceded the depression.

f. Investigate Comorbidities—Prevalence

Anecdotally and per our survey results, a concerning number of CRD patients report hypersomnia, ADHD, pain, migraine, fibromyalgia, ME/CFS, depression, anxiety, and endocrine disorders. We'd like to know whether prevalence is higher among CRD patients than in people without CRDs. Do CRDs and/or resulting sleep deprivation cause these conditions, or vice-versa? Or does an underlying mechanism give rise to CRDs and these conditions? Investigating the answers can point the way to treatments for CRDs and these other medical conditions. Researchers should also examine the presence and significance of CRDs in various developmental and neurological conditions, such as autism, ADHD, Smith-Magenis syndrome, and others.


a. Develop Biomarkers

We need biomarkers for circadian phase that don't require difficult or expensive laboratory conditions—certainly ones that don't require imposed sleep patterns. We also need biomarkers for tiredness/alertness that correlate with subjective experience as well as with cognitive testing. Markers for amplitude and other factors (see Section 3b above) would be helpful too.

b. Determine the Usefulness of DLMO in Diagnosis

DLMO is taken as a marker of circadian phase. We believe this to be valid for normal sleepers. But do we have evidence that it also is valid for DSPD patients? Or is it possible that different circadian processes become desynchronized (thus also helping to explain why many of us take so long to fully wake up), meaning melatonin production is shifted relative to the sleep-wake cycle? Same question for ASPD. Is DLMO useful in diagnosing non-24? ISWD?

The gold standard for determining intrinsic circadian rhythm in the laboratory is a "constant routine" used in an attempt to eliminate the effects of eating, sleeping, and activity on melatonin production. This is a difficult and expensive procedure. Some physicians have used home melatonin testing as a more practical alternative. Is this useful as a diagnostic tool?

Melatonin testing is not covered by health insurance; recreating dark laboratory conditions at home is challenging, potentially undermining test results; and hospital-based melatonin testing is limited to study settings. We support efforts to make DLMO testing widely available in clinical (non-research) settings and to secure insurance reimbursement.

c. Research Possible Overlap Between Sleep Disorders

While differential diagnosis is important, more attention should be paid to the overlap of different sleep and circadian disorders than is currently the case.

For example, we have been approached by a number of narcolepsy patients, and many of them go to bed late and have trouble getting up in time. Do they really have a CRD, but were misdiagnosed with narcolepsy, since the MSLT can give false positives when people are not sleeping on their body's preferred time (this is well-known for people doing shift work)? Or do a lot of narcolepsy patients also have circadian disorders?

We note that 40% of our survey respondents have been correctly diagnosed with other sleep disorders in addition to their CRD.

We noted earlier (Section 3c) that many CRD people are tired even when sleeping on their body's preferred schedule. Is this an inherent part of the CRD? Or another sleep disorder? Or perhaps the disorders are not as distinct as we usually think of them, but overlap?

d. Do Not Reclassify "Non-Circadian" DSPD

Some recent studies of DSPD have made a distinction between a subgroup with delayed DLMO and another with normal DLMO but a large lag between DLMO and sleep onset. Researchers have suggested that these subgroups be classified as "circadian" and "non-circadian" DSPD. While we support this research and any research into the causes and diagnosis of circadian disorders, we urge caution in applying the "circadian" vs "non-circadian" distinction. Our concerns fall into two categories:

1. We have reason to believe that this distinction is based on studies of patients with mild DSPD and will not hold up in more severe cases. The reasons for this are detailed in Appendix A.

2. The suggestion has been made that only circadian DSPD should be legitimately categorized as DSPD. That may have unfortunate consequences. For example, if the lag between DLMO and sleep onset is due to decreased homeostatic sleep pressure, we don't have a diagnostic category called "homeostatic sleep disorder" and so such patients may lose a DSPD diagnosis without being given an appropriate alternate diagnosis. Essentially, they would become diagnostic orphans. Insomnia, as usually understood, may not apply to such patients and may lead the incorrect treatment. For example, CBT and/or sleeping medications may be of benefit for insomnia associated with anxiety or psychophysiological overarousal, but it's not clear if they would help someone whose root problem has to do with disturbances of sleep homeostasis or other factors that may lead to a lag between DLMO and sleep onset. Extreme eveningness would be another factor that might cause a lag between DLMO and sleep onset, but is also distinct from psychophysiological insomnia. We would suggest that any persistent sleep delay continue to be classified as DSPD (as in most current diagnostic criteria) unless or until research clarifies other options.


a. Base Treatments on Studies of CRD People

CRD treatments are based on the physiology of people with normal circadian rhythms: their phase-response curves (PRCs), relative timing of dim light melatonin onset (DLMO), and relative timing of core body temperature nadir. Perhaps this is one reason (among many) that CRD treatment often fails. We need clinical studies of people with CRDs to determine their PRCs, relative timing of DLMO, and relative timing of temperature nadir (among many other objectives).

Researchers should make use of the online communities of CRD patients, particularly as a source of volunteers for clinical trials. This is particularly important for the more severe or more rare disorders, which may be otherwise harder to study due to lack of subjects. Such communities would include the Circadian Sleep Disorders Network (CSD-N), the Niteowl mailing list, and Facebook and Reddit groups. CSD-N in particular has developed a patient registry and survey which can be both a source of information and a resource for seeking patient volunteers for studies. Recruiting from sleep specialist practices might be another, if the doctors are notified that there is a particular interest in severe cases.

When recruiting and studying patients with CRDs, distinguish between a) those who sleep according to the dictates of their body clock; b) those who sleep according to society's (school, work, etc) dictates; c) those who've successfully treated their CRD; and d) those for whom treatment has failed.

b. Link Treatment With Cause

CRD treatments have failed for many patients. We need treatments that are more effective—targeting underlying pathology, which may vary from patient to patient (see Section 3a). When multiple causes are present, treatment may need to address all relevant causes to be successful. Severity also interacts with treatment: a therapy that advances sleep by 1 hour may be of great help to someone who typically falls asleep at 1 a.m., but of minimal benefit to someone who falls asleep at 6 a.m.

c. Improve Evaluation of Treatment Effectiveness

Some patients shift their sleep-wake times with a treatment under study but still feel awful. Shifted schedule, DLMO, and temperature nadir are not necessarily reliable indicators of treatment success—not if the patient is still tired, jet-lagged, and mentally foggy, with impaired cognition and functioning. (Why is there sometimes a mismatch between objective and subjective findings?)

Also, clinical studies tend to follow patients over the short term only.

Research must consider the patient's subjective experience—do they feel rested and alert, with healthy cognition and functioning?—and follow patients over the long term.

d. Improve Parameters for Light and Dark Therapy

Light therapy is often prescribed as a treatment for CRDs. But only 27% of all CSD-N survey respondents who tried light therapy said it helped somewhat or more, and only 14% said they achieved their desired sleep schedule.

Members of our community are given wildly varying prescriptions for administering light and dark therapy. We need more research based on the circadian rhythms and PRCs of disordered, not normal, patients. We need evidence-based answers to these questions, as they pertain to people with DSPD, ASPD, Non-24, and ISWD—not people with normal CRs: When should light therapy begin, relative to the patient's circadian rhythm? How rapidly or slowly should we advance the timing? How many lux? Blue, green, and/or white light? For how many minutes or hours? Continuous or pulsed light?

For light restriction: How few lux? Should we block green as well as blue light? Or all light? Starting how many hours before bed? How rapidly or slowly should we advance?

We note (see Section 3a) that there may be different underlying causes for DSPD in different patients. This suggests that PRCs and treatment parameters should be investigated separately for each underlying cause.

For light treatment of Non-24, we have noted that the best results seem to occur when the individual waits until they have rotated around almost to their desired sleep time and then starts light therapy (as opposed to simply starting it on a random day). This seems to follow from the light PRC, but we need research to confirm this clinically.

e. Improve Melatonin Recommendations

Members of our community are given wildly varying prescriptions for administering melatonin therapy.

There is some research on the best timing of melatonin for DSPD, but more is needed. Is the best time 6 hours before DLMO? 12 hours before spontaneous wake time? Other? How rapidly or slowly should the patient advance their schedule?

We note that the same dose of melatonin can result in blood levels that differ as much as 35-fold from one individual to another (Aldhous M, Franey C, Wright J, and Arendt J, Plasma concentrations of melatonin in man following oral absorption of different preparations (1985); Waldhauser F, Waldhauser M, Lieberman HR, et al, Bioavailability of oral melatonin in humans (1984)). We need research to optimize melatonin dose for each individual patient instead of one-size-fits-all.

We need recommendations on how to begin titrating to find the optimal dose. Clarify the use of immediate vs delayed release. Can melatonin be taken safely and effectively for an indefinite period of time? Is it safe for children, and at what dose?

For melatonin treatment of Non-24, we have noted that the best results seem to occur when the individual waits until they have rotated around almost to their desired sleep time and then starts melatonin (as opposed to simply starting it on a random day). This seems to follow from the melatonin PRC, but we need research to confirm this clinically.

f. Stop Recommending Phase-Delay Chronotherapy

Phase-delay chronotherapy (progressively delaying sleep time 1 – 3 hours each day until sleeping at the desired time) is not a long-term solution for DSPD, and appears to involve unacceptable risk of the DSPD becoming Non-24. Our survey results indicate that, while this may work in the short term, nearly all patients (99% in our survey) revert to their previous schedule after a while. More seriously, this may cause the DSPD to become Non-24, which is far more difficult to live with. (See e.g. Oren DA and Wehr, TA, Hypernyctohemeral Syndrome after Chronotherapy for Delayed Sleep Phase Syndrome, 1992. NOTE: The paper misspelled "hypernychthemeral.") In our survey, some 24% of patients who tried phase-delay chronotherapy said they became Non-24 as a result. We believe utmost caution should be used before prescribing this treatment until we have a better understanding of how to avoid the precipitation of Non-24.

g. Consider a Different Endpoint for Non-24

Some of us have speculated that Non-24 patients might find it easier to entrain to a delayed 24-hour schedule rather than to a normal one. This is based on the expectation that Non-24 might be due to some of the same underlying causes as DSPD, and to the observation that some patients had DSPD before it became Non-24.

h. Investigate Timing of Meals

Some patients are told to eat earlier in the day in order to shift their circadian rhythms. Is eating a zeitgeber? Investigate whether, and to what extent, meal timing affects circadian rhythms in people with CRDs.

i. Consider New Treatment Paradigms

We need treatments that are less draconian and burdensome, so people needn't be tethered to perfectly timed and executed light exposure and light restriction, every day, for the rest of their lives. Also, many patients report adverse effects from melatonin.

Most important, we need treatments that work.

Perhaps other treatment approaches, based on entirely new paradigms, are called for. Therefore, we encourage researchers to explore new avenues. One example: investigate the possible connection between gut microbiome and CRs. Another: study drugs affecting the cellular clock that cause phase shifts or change the intrinsic period. Other examples are awaiting discovery.


A lot of circadian research knowledge has not been taught in medical school or filtered down to clinicians, including primary care doctors, medical specialists, mental health professionals, and even some sleep medicine physicians. As a result of some clinicians' ignorance, patients have been misdiagnosed, shamed, prescribed medications they don't need (or denied those they do need), prescribed ultimately harmful treatments, and denied referrals to specialists who could potentially help them.

a. What Clinicians Need to Know (and Often Don't)

b. Treatment Concerns—Light

Doctors often prescribe therapy at a specific time based on desired sleep schedule, for example light therapy at 7 a.m.. This generally means that the patient is trying to move their schedule by many hours at once. In many cases of severe DSPD, it may mean that the light is provided during the phase delaying part of the patient's PRC. We believe that light therapy should initially begin when the patient would wake up on their body's preferred schedule, and be gradually advanced earlier as the patient gradually advances their sleep time, until arriving at their desired schedule.

c. Melatonin

Many community doctors seem to prescribe melatonin doses in the range of 3 – 10 mg. That may be too high for many patients (Lewy AJ, Emens JS, Sack RL, Hasler BP, Bernert RA, Low, but not high, doses of melatonin entrained a free-running blind person with a long circadian period (2002); Fadden, J, Melatonin: Less Is (Sometimes) More (2010) - a blog post discussing the previous article). Some of us have had good effects from as little as ¼ mg. We would like to see more guidance for doctors.

Many doctors are not aware of the distinction between hypnotic and chronobiotic use of melatonin. Hypnotic use involves a larger dose taken shortly before bedtime to produce rapid somnolence and aid in insomnia. Chronobiotic use involves taking melatonin with reference to the PRC to maximize phase shift. Typically, small doses a number of hours before DLMO are recommended by researchers, but community clinicians do not often know that.

The phase shifting properties of exogenous melatonin are determined by the timing relative to current DLMO (or current bedtime as an approximation if DLMO is unknown). Many clinicians prescribe melatonin timed with reference to desired rather than current bedtime. This is unproven, seems to go against the science, and does not seem to work well for most of our members.

d. Phase-Delay Chronotherapy

Clinicians should not recommend phase-delay chronotherapy. Please see Section 5f.

e. Sleep Studies

Sleep studies (polysomnography) only make sense if the patient is sleeping on their body's natural schedule, yet most sleep labs require the study be done within a fairly narrow time window during the night.

MSLT in particular has been shown to give false positives for narcolepsy if the patient has been working night shifts. It would be expected to give similarly false results for someone with a CRD who has been forcing themselves to work day shifts, or who is being tested outside their body's natural awake time.

f. Medical Appointments and Lab Tests

CRD patients sometimes report that they can't get late-day appointments with their physicians—even sleep-medicine physicians. This is a barrier to treatment. If patients accept earlier appointments and drive, their safety—and the safety of others—is at risk. For some, forcing themselves up early even once can affect their productivity at work for several days, as some of our members report (and see Section 6h below).

Rarely are doctors aware that they should adjust the timing of certain blood tests (for example, cortisol and adrenal testing) in accordance with the patient's circadian rhythm. As a result, test results may not be accurate and treatment may not be appropriate—and could even be harmful.

g. Sleep Hygiene

Most advice regarding sleep hygiene is, at best, based on studies of normal subjects or patients with insomnia or, at worst, is clinical lore without an evidence base. It is rarely evidence-based with regard to CRDs. Traditional sleep hygiene may or may not work for patients with CRDs. "Keep regular sleep hours" is hardly useful advice for someone with Non-24. Sleep hygiene advice can be helpful in some cases, but it can also be irrelevant or even harmful. Too often, it slides into blaming of the patient for their disorder.

Most of us have had circadian disorders for years, if not decades. We have experience in what works for us and what does not. We've tried reading in bed or not reading in bed, getting out of bed when we can't sleep or staying in bed, exercising before bed, or avoiding exercise. What works for one may not work for another and may or may not match up with traditional sleep hygiene advice.

We suggest that sleep hygiene advice be based primarily on either (a) actual controlled studies of what works for CRDs; or (b) listening to the experience of individual patients with CRDs.

h. Inflexibility

For many of us, DSPD consists of TWO issues: our circadian phase is shifted later, and our sleep hours are resistant to change. Inflexibility of sleep timing is one major difference between being an "evening type" and having DSPD.

Many doctors do not recognize how inflexible our sleep hours are. Some insist that the prescribed treatment would work if we followed the timing rigorously, including forcing ourselves awake at the specified time. In many cases it still doesn't work, and patients have been pushed into serious sleep deprivation by their doctors.

It is worth noting that to become a doctor, to get through internship and residency, one has to be able to be flexible with sleep. So doctors are a subpopulation that may have little personal experience with the rigidity of sleep time that many CRD people experience.

i. Impact

Many clinicians are unaware of the impact of circadian disorders on the lives of patients.

Some children with CRDs fall asleep in class; others miss part or all of the school day, and the school takes legal action against their parents. Many of our members have failed classes and have had to drop out of school or college. Others whose disorder developed somewhat later are able to finish college or obtain advanced degrees, but find they have very limited career options due to their disorder. Those with severe DSPD are usually unable to work 9 to 5 jobs, while those with Non-24 or ISWD are unable to work on any fixed schedule and often cannot find employment of any kind.

Family life and social life are also adversely impacted. Partnering, parenting, and caring for elderly parents may be compromised, sometimes severely. Often, spouses, family, and friends do not understand our disorders.

Many tasks that most people take for granted—making a doctor's appointment, going to the bank, or going shopping—can become difficult challenges for those not awake during daylight hours. It's not unusual for us to hear questions like this: "I would normally expect to sleep from 8 a.m. to 4 p.m. tomorrow, but I have an 11 a.m. appointment. Is it better to try to get 2 hours of sleep and get up at 10, or should I just try to stay awake for 4 or 5 hours longer?" There is no easy answer. Often, trying to make such an appointment further deranges already abnormal circadian rhythms, and it can take many days for the person to recover from the sleep loss and disrupted circadian cycle.

j. Lack of Support from Clinicians

It is difficult to find a primary care doctor who recognizes CRDs and is willing to refer to a sleep specialist. Even sleep specialists are often unaware of the basic medical facts regarding circadian disorders and, in some cases, even question if these are legitimate disorders. All too often, they give advice based on the idea that if patients simply force themselves to wake at a set time, that will cure them. Some prescribe other unproven treatment regimens.

As a result, clinicians are often unwilling to certify that a patient is unable to adjust their sleep hours. This presents a significant problem for parents requesting accommodation from their child's school, students requesting accommodation from their college, and employees requesting accommodation from their employer. Often such accommodations can spell the difference between staying in school or dropping out, or between keeping a job or losing it. The consequences can be life-altering.


All medical research—including non-circadian studies—and all medical tests should evaluate patients for circadian rhythm. They should note (even for normal CRs) the time of day of all physiological measurements (blood tests, gene expression), because many of these vary by time of day.

Appendix A. Interaction of Severity and Causation

With regard to DSPD, the original Montefiore study specified patients' sleep onset as between 2 a.m. and 6 a.m. A number of studies in recent years have focused on patients who typically fall asleep around midnight to 1 a.m. In one respect, this is a positive development, as it gives insight into a possibly common condition which is often misunderstood and can be applicable to a wide population. At the same time, we urge caution in generalizing the results from such studies to more severely affected DSPD patients. Severely affected patients may differ in both the causes and the treatments of their condition.

One specific issue is a manifestation of Berkson's paradox, a finding in the science of causality, with particular applications to epidemiology. The paradox states that apparent correlations between two variables may differ depending on the subset of the population examined. One classic example is a finding that the correlation between the presence of respiratory disease and cardiovascular disease is much higher if the sampling group is hospitalized patients rather than outpatients. In this case, the reason is easy to see: someone who has two serious conditions is far more likely to end up in hospital than someone who has one.

Recent studies on DSPD have attempted to distinguish what are termed "circadian" and "non-circadian" causes of the condition (see, for example, Murray JM, Sletten TL, et al, Sleep, 2017 Jan 1;40(1)). Specifically, "circadian" DSPD is defined as a delay in the circadian rhythm, as indicated typically by DLMO, relative to normal. "Non-circadian" DSPD is characterized by an abnormal lag between DLMO and sleep onset. In mild cases, it has been possible to divide patients into one or the other category. Berkson's paradox suggests—and clinical evidence confirms—that patients with severe DSPD may have both circadian and non-circadian causes. For example, someone may have a DLMO at 1 a.m. (circadian delay) and also a longer-than-normal lag of 4 hours between DLMO and sleep onset (non-circadian), meaning they fall asleep at 5 a.m.

In addition, these two factors are not causally independent. Someone whose lag between DLMO and sleep onset is very long, will also be prone to sleep through the phase advance portion of the light PRC and end up with a circadian delay, as well. The combination of the two may result in a severe case of DSPD. Past research confirms that severe DSPD often exhibits both a delay in the temperature/melatonin cycle and a lag between that cycle and sleep propensity/sleep onset.

A similar situation occurs in sighted Non-24, where several studies have shown there is typically not just a greater-than-24-hour temperature rhythm, but also a very great phase delay of sleep propensity/onset relative to the temperature cycle.

Understanding the role of both types of causation is important for properly classifying and treating DSPD and Non-24. For example, treating only one component would be unsuccessful if both factors are present. In a more general sense, conclusions about causation based on studies of milder cases may not apply to more severe cases, and vice-versa.

© 2023 Circadian Sleep Disorders Network
Last modified Jan 5, 2022
Problems: contact