Circadian Depression's 23% Rests on a Questionnaire, Not a Clock Test

Three versions of one number have travelled this week. Headlines say 1 in 4 Patients. The University of Sydney's release says nearly 25 percent. The paper behind both says 23%, which is 2,604 of 11,380 people. All three describe the paper in the British Journal of Psychiatry, published online Tuesday, Oct. 20, in which Sydney researchers propose a "circadian" subtype of depression tied to a disrupted body clock. They sorted people using six self-reported criteria from an online survey, not a clock test, and they describe the result as exploratory.

Circadian depression is not a diagnosis. ABC News reports it is not currently listed as a standalone diagnosis in the DSM, the manual clinicians use to diagnose mental illness. Nor is the idea new: Live Science says it was First conceptualized in 2021. What the paper offers is a proposed subtype, a way of grouping people whose depression looks alike so treatment might one day be matched. Nature reports that researchers in the field say more evidence is needed before circadian depression can be formally recognized as a disorder or treatment guidelines are changed.

So what is the 23% a share of? Not everyone with depression, and certainly not everyone who sleeps badly. First author Dr. Emiliana Tonini, co-senior author Dr. Jacob Crouse and Prof. Ian Hickie drew on the Australian Genetics of Depression Study, an online survey. Participants were recruited through government invitations based on prescription records, or a media campaign for adults who had experienced clinical depression. The paper describes the cohort as predominantly adults with severe, recurrent depressive disorders: only 4% reported a single episode, and the median was six. Some 75% were female, and the study kept only people of genetically inferred European ancestry. The data date from 2016 to 2018.

The arithmetic needs a second look as well. The paper rounds 22.88% to 23%. A true quarter of 11,380 would be 2,845 people, 241 more. Rounding up to "one in four" is a small stretch. The bigger one is the denominator: mostly severe, repeatedly depressed volunteers, sorted by a questionnaire. A quarter of that group says little about depression in general.

Six self-reported criteria built the subtype, and genes were checked afterward

Who counts as circadian? Anyone who answered yes to at least three of six self-reported items:

• a seasonal pattern to mood, sleep and behaviour;
• a delayed sleep midpoint, the halfway point between falling asleep and waking on free days, past an age-adjusted cut-off;
• social jetlag, a big gap between free-day and work-day sleep timing (explained here; two hours or more for ages 35 to 64);
• answering "Definitely an evening type", a natural lean toward late nights and late mornings, known as an evening chronotype;
• sleeping "much more than usual" during the worst depressive period;
• feeling "very tired" in the first half-hour after waking.

Genes came second. Once the groups existed, the team compared polygenic scores, estimates of inherited susceptibility added up across thousands of small genetic variations, between them. No wearable device and no physiological clock measure, such as the timing of melatonin release, was used. The authors say it themselves: the criteria did not directly capture internal physiological circadian markers (e.g. phase, amplitude, misalignment). The label describes how people say they sleep and feel, not what their clock is doing.

Why three? The paper says the cut-off was selected as an optimal balance of criteria matching the team's earlier clock studies, and chosen to give appropriate subgroup sizes that would support statistical power. It also calls the criteria exploratory rather than definitive. Andrew Coogan, a behavioural neuroscientist at Maynooth University who was not involved, cited to the Guardian a lack of detail as to how the cluster of features of, and the criteria for, the proposed circadian subtype of depression was arrived at.

Across all 11,380 people, 51.2% said they slept much more than usual when at their worst. Another 38.6% were very tired in the morning, and 24.8% called themselves definite evening types. The most common combination inside the subtype was exactly that trio: 481 people, 18% of the group. If half of everyone in a depression study reports oversleeping, then "three of six" is closer than it sounds.

The second weakness is the proxy itself. Brant Hasler, a psychologist and sleep researcher at the University of Pittsburgh who was not involved, told Live Science:

"One can identify as an evening-type, and report later sleep timing, and still have physiological circadian timing that is not particularly delayed,"

Brant Hasler, psychologist and sleep researcher, University of Pittsburgh

Hasler added that such self-reports correlate only modestly with direct measures like melatonin in saliva.

The genetic results are small tilts, not tests. People in the subtype carried higher scores for ADHD (odds ratio 1.11) and for a later sleep midpoint (1.15). An odds ratio of 1 means no difference, so these sit just above it. The largest, a score for evening preference (odds ratio 0.68, meaning higher genetic liability to eveningness), partly echoes a criterion used to define the group. A schizophrenia score (1.07) did not survive correction for multiple testing. About 1,400 people meeting criteria for atypical depression, and about 700 with bipolar disorder, were placed in the non-circadian group; the authors call circadian and atypical depression partly overlapping but separable constructs.

The treatment signal is a self-rating, not a trial

The treatment claim is where the story shortened most. On Oct. 21, New Scientist posted:

Compare that with the paper, which says we propose that individuals with the circadian subtype may benefit preferentially from chronotherapies, meaning treatments that target the body clock, and cites growing evidence from other work. The university release goes further, calling the subtype more responsive to behavioural and medical therapies targeting the body clock's unique brain circuits, such as melatonin, melatonin-targeted antidepressants and light therapy. Crouse said clinicians should reach for these much earlier.

Here is what the study measured. Participants rated how well each of 10 commonly prescribed antidepressants had worked for them, from 0 (not at all) to 2 (very well). The subtype was associated with lower ratings for SSRIs (odds ratio 0.88, 95% confidence interval 0.82 to 0.94) and SNRIs (0.89), two widely used antidepressant families. An odds ratio below 1 here means somewhat lower odds of a higher rating. It is a modest shift, with no difference for mirtazapine (0.96) or amitriptyline (0.97), and the subtype reported a broader range of side effects.

And here is what it did not measure: light therapy, melatonin or any behavioural treatment. The "more responsive" line comes from the proposal, not from a result. Light therapy does have established evidence, but elsewhere: the National Institute of Mental Health calls it a mainstay for treating winter-pattern SAD, a seasonal-pattern form of depression with its own diagnostic criteria. (For melatonin, see the site's piece on how dosing differs between the U.S. and Europe.) Tonini told ABC News the antidepressant finding came from self-reported data, so more research was needed to test those drugs in this group.

Hickie's line in the release is the boldest: Most young people will be offered SSRIs first up, but for at least a quarter there are better circadian-focused medical and behavioural options to help treat and manage their condition. His disclosed interests appear below. "At least a quarter" is more than the paper's 23%, and "better" is a hypothesis the paper frames as "we propose".

Independent experts split in tone, not direction. Philip Mitchell, a psychiatry professor at the University of New South Wales, told ABC:

"At the moment it's not reaching the bar where this could enter clinical practice, but I think this is very thought-provoking for the field,"

Philip Mitchell, scientia professor of psychiatry, University of New South Wales

Jayashri Kulkarni, who directs the Monash Alfred Psychiatry Research Centre, called it a really well-conducted study, though the Guardian reports she noted it did not account for menopause's hormonal effects on sleep. Ruth Benca of Wake Forest University offered a reading that cuts against a clean clock story: People with more pronounced circadian disturbances often have more severe illness and other psychiatric conditions, which can contribute to treatment resistance, she told Live Science. Severity could be doing some of the work. The authors' own check, restricted to people meeting major-depression criteria, left three genetic findings and one treatment finding no longer significant under a strict statistical correction.

A bad night's sleep is not a diagnosis

If you sleep badly, are a night owl or sleep too much when low, this study does not say you have circadian depression. It looked only at people already diagnosed with depression, and the features it counted are common in that group. Shanthakumar Wilson Rajaratnam, a professor of sleep and circadian medicine at Monash University, told ABC that being a night owl or waking very tired might not be driven solely by circadian dysfunction. Still: It's important to address the sleep and circadian complaints in patients that are presenting with psychiatric disorders, including depression, as well as treating the depression symptoms, he said.

The typical subtype member in the data was a woman of about 38 whose depression began around 19. Mean age was 37.87 versus 45.56 outside the subtype. The subtype also scored higher on psychological distress and on suicidal ideation. In the U.S., anyone struggling or having thoughts of suicide can call or text the 988 Suicide & Crisis Lifeline at 988, as the National Institute of Mental Health notes.

On treatment, nothing has changed today: it is not in guidelines, and the researchers and outside experts all say more evidence is needed. Crouse's own advice: For those who find that antidepressants are not working for them, it is important to speak to your clinician for advice about your individual circumstances. And the NHS is explicit: Do not stop taking an antidepressant suddenly or without talking to your doctor first. ABC reports that Mitchell and Rajaratnam agreed that asking patients about circadian patterns was warranted, in addition to, not instead of, traditional treatment options.

What would settle it? The paper's own closing line says: Pending independent replication, investigation of its biology and predictive utility are warranted. Disclosure belongs here too. The paper says Hickie is Chief Scientific Advisor to and an equity shareholder in InnoWell Pty Ltd, a digital mental-health company, and has led projects supported by pharmaceutical companies including Servier. The release adds that he researched agomelatine, a melatonin-targeted antidepressant, with Servier funding.

One real test is under way. The MELODY trial, with Hickie as principal investigator, is a 12-week randomised trial in 660 people aged 18 to 30 with major depression or bipolar II and sleep complaints, comparing melatonin with placebo and with digital cognitive behavioural therapy for insomnia, with depressive symptoms at week 12 as the main outcome. The trial's published protocol notes earlier melatonin trials were small, from 6 to 142 participants, and Medical Republic says it is still recruiting. But it enrols people on sleep complaints, not on the six-feature label, so even MELODY cannot tell us whether this subtype is real.