Sleep problems worsened years before a psychiatric diagnosis in working-age Swedes and partly eased after it

Every two years, a questionnaire arrived in the mail for tens of thousands of working-age Swedes. Along with questions about work and health, it asked what time they went to bed and got up on workdays and days off. It also asked how often they struggled to fall asleep, woke too early or felt mentally exhausted. Every item was rated on a six-point scale that ran from never to five or more times a week.

Jie Guo, Anna Karin Hedström and colleagues at Karolinska Institutet, Stockholm University and China Agricultural University have now lined those answers up with Swedish national health registers. In the European Archives of Psychiatry and Clinical Neuroscience, they report that insomnia symptoms and fatigue progressively worsened in the years before a psychiatric diagnosis and partly eased after it.

Insomnia had already been identified as a predictor for depression. A 2019 pooled analysis of 13 studies found it also predicts the onset of anxiety, alcohol abuse and psychosis. Little research had previously tracked how sleep changes in the lead-up to psychiatric illness, and a survey repeated every two years is suited to exactly that. Across the whole group, a diagnosis was uncommon: of about 25,000 participants with none at the start, 716 were diagnosed with a psychiatric disorder during follow-up, fewer than 3 in 100.

How eight years of Swedish sleep questionnaires were matched to psychiatric diagnoses

The questionnaires belong to the Swedish Longitudinal Occupational Survey of Health, which since 2006 has followed people first sampled as working adults aged 16 to 64. Of the 29,651 who answered at least once between 2010 and 2018, the researchers removed anyone with an existing psychiatric diagnosis. New diagnoses came from the National Patient Register and the Cause of Death Register, so nobody had to report their own condition.

For each person with a new diagnosis, the researchers then ran the clock backward. The year of diagnosis became year zero, with every earlier and later questionnaire placed on either side of it. For everyone else, year zero was the end of follow-up. That let the team compare the two groups year by year, across the eight years prior to diagnosis and the six after.

Because the same people answered repeatedly and their diagnoses came from medical records, the study can date sleep changes relative to the timing of diagnosis with some confidence. The links between sleep and a later psychiatric diagnosis also held when the first two years of follow-up were left out. That guards against counting people who were already close to a diagnosis at the start. It would not catch an illness that builds over many years. Nobody was assigned to sleep better or worse, either. So the study cannot say whether sleep problems are risk factors that could be changed or early symptoms of an illness already under way. Despite adjustment for a long list of possible influences, from smoking to heart disease, those who sleep badly may differ in ways the analysis missed.

Mental fatigue and insomnia set people apart years before a psychiatric diagnosis

At their first questionnaire, every sleep complaint on the list was associated with increased risk of a later diagnosis. Mental fatigue had the strongest link. It carried about double the risk after accounting for age, sex, education, smoking, exercise, weight, sleeping pills and other conditions. Of the 4,161 participants who reported it, 231 were diagnosed with a psychiatric disorder, about 1 in 18, against 1 in 43 of those who did not.

The link between sleep duration and a later diagnosis was U-shaped, with risk higher at both ends. Compared with seven to eight hours of sleep per night, averaging under seven or over eight carried a 30% to 40% greater chance. Insomnia here meant trouble falling asleep, waking repeatedly or waking too early, at least three times per week. Participants with both nighttime insomnia and persistent fatigue had more than twice the risk of those with neither. When every sleep measure went into one analysis, mental fatigue and long sleep still stood out.

The repeated questionnaires show when those gaps opened. Trouble falling asleep, persistent fatigue, physical exhaustion and mental fatigue already differed eight years before diagnosis, the earliest point the analysis reached, and the differences widened from there. Frequent waking, difficulty getting up and unrefreshing sleep began to diverge around six to seven years out. Most complaints were at their worst near the time of diagnosis and then eased. Even so, insomnia symptoms and fatigue stayed above the levels seen in the unaffected group.

Three complaints followed a flatter line. Insufficient sleep, snoring and signs of sleep apnea were more common all along in those who went on to a diagnosis, with a gap that stayed about the same size. Two others kept worsening after diagnosis: long sleep duration and delayed wake-up times. Mornings had already been getting subtly later over the five years leading up to diagnosis.

Worsening sleep before a psychiatric diagnosis may be both warning sign and cause

Guo, Hedström and their co-authors read the build-up as a disorder announcing itself. The steady widening of the gap, they write, strengthens the case that disturbed sleep belongs to a prodromal phase, “a subtle clinical state that precedes overt psychiatric illness”. Several of the complaints double as symptoms. The US National Institute of Mental Health lists fatigue, difficulty sleeping, waking too early in the morning and oversleeping among common signs of depression.

Mental fatigue fits the same reading. It outperformed several traditional insomnia symptoms as a predictor. The researchers suggest that difficulty getting through the day may reflect emerging psychiatric vulnerability more closely than poor sleep itself. Yet many complaints did not return to the level of participants who stayed well, and two kept getting worse. The authors concede that this persistence complicates a purely prodromal interpretation.

The biology behind these patterns went untested. Chronic sleep disruption has been tied to inflammation, the stress-hormone system, the brain’s chemical signaling and its capacity to rewire. All of them have also been implicated in depression and anxiety. Long sleep that lingers after diagnosis could reflect an unsettled body clock, the effects of medication or a response to physiological or emotional strain, the authors propose.

An Oxford-led team has made the case for cause. In a 2020 review in The Lancet Psychiatry, they argued that insomnia and other mental health conditions share causes and feed each other. The stronger pathway, they concluded, is typically disrupted sleep as a causal factor in other psychiatric problems. Treating insomnia, the Oxford group added, lessens other mental health difficulties.

A trial at the University of California, Los Angeles tested the preventive side of that claim. It enrolled 291 adults aged 60 or older with insomnia disorder but no recent depression. They received two months of either sleep education or cognitive behavioral therapy for insomnia, which retrains sleep habits and the thoughts that keep people awake. Over up to three years, major depression occurred in 12.2% of the therapy group and 25.9% of the comparison group, about half the rate. It was a single site, with older adults and one condition, but treating sleep changed who became depressed.

The Swedish team’s own previous work shows how general that warning is. Using the same survey, they reported in Sleep Medicine earlier this year that similar changes preceded the diagnosis of neurological disease by several years. They included later sleep timing and increasing insomnia symptoms, fatigue and breathing problems in sleep. A slide in sleep, in other words, does not point to one kind of illness.

The Swedish registers never said which psychiatric diagnosis people received

The linked data recorded that a psychiatric diagnosis had been made, with no detail of which one. That meant the team could not check whether the pattern differed across psychiatric diagnostic categories, such as depression and anxiety. The count took in psychiatric disorders (ICD codes F00-F99) from the National Patient Register and the Cause of Death Register. That span of the World Health Organization’s disease classification also contains F51, nonorganic sleep disorders, a category that lists nonorganic insomnia. Some later diagnoses may therefore have been sleep problems in their own right, and the paper does not report setting them apart. Nor did the records reach all of health care. Sweden’s national patient register does not contain data on primary care, so a diagnosis made only by a family doctor would not have counted. The authors add that subthreshold or undiagnosed psychiatric conditions, meaning milder or unrecognized problems, may be missed.

Everything about sleep was self-reported. Sleep length came from the bedtimes and wake-up times people gave, and by that yardstick four in ten participants averaged more than eight hours a night. The method counts any time spent lying awake in bed. The survey held no clinical data on sleep apnea diagnosis, severity or treatment. Compared with people who did not reply, participants were more often women, older, born in Sweden and university educated. The response rate also slipped, from 57% in 2010 to 48% in 2018. The authors say this limits how far the results reach beyond the Swedish working-age population.

When a long slide in sleep is worth raising with a doctor

Where repeated monitoring is feasible, the researchers write, these complaints “could serve as early warning signs to guide preventive efforts”. For any one person, the chance of a diagnosis stays small. Even among participants with both insomnia and persistent fatigue, who had more than twice the risk, 132 of 2,156 received a diagnosis during follow-up. About 94 in 100 did not.

Anyone whose sleep has been sliding can act on existing advice. The NHS advises seeing a GP, or family doctor, if changing sleeping habits has not helped. The same goes for trouble sleeping for months, or insomnia that makes daily life hard to cope with. A GP may offer cognitive behavioral therapy, in person or through an online self-help program.

Poor sleep is worth treating for its own sake, and in this study it was rarely followed by a psychiatric diagnosis.