Not the oxygen drops: broken sleep tracked with depression, insomnia explained it
In 1,238 people from two community cohorts, apnea severity showed no link to depressive symptoms. Sleep fragmentation did, and that association disappeared once insomnia symptoms entered the model.
There are two standard stories about why disturbed sleep and depression travel together. One locates the damage in intermittent hypoxia — the repeated oxygen desaturations of sleep apnea acting on the brain. The other locates it in fragmentation and in the subjective experience of not sleeping. A cross-sectional analysis published in the Journal of Sleep Research on 28 August tests both against each other in community samples, and only one survives [s1].
The design
The analysis harmonised data from two population-based cohorts: the Sleep and HEalth in women (SHE) study and the Men Androgen Inflammation Lifestyle Environment and Stress (MAILES) study, giving a combined sample of 1,238 participants — 68% male, mean age 57 (SD 12), mean BMI 28.0 kg/m² (SD 4.2) [s1].
Using population cohorts rather than sleep-clinic patients is the methodological point. The authors explicitly frame the question as one that has been hard to answer in samples carrying clinical referral bias [s1]: people who reach a sleep laboratory are there because something already brought them, and that something is often mood.
Objective measures of obstructive sleep apnea — the apnea-hypopnea index and the oxygen desaturation index — were recorded alongside objective fragmentation measures, sleep efficiency and wake after sleep onset, plus subjective reports of insomnia symptoms and depressive symptoms [s1]. Logistic regression models adjusted for age, sex, BMI, alcohol and physical activity, and pathway analyses explored mediation [s1].
What it found
Depressive symptoms were present in 17% of the sample [s1].
Neither of the intermittent-hypoxia variables was linked to depressive symptoms: no association was found for the apnea-hypopnea index or the oxygen desaturation index [s1].
The fragmentation variables were. Reduced sleep efficiency was associated with depressive symptoms (odds ratio 0.84; 95% CI, 0.72–0.98) and increased wake after sleep onset with higher odds (1.26; 95% CI, 1.06–1.51) [s1].
Then the models added insomnia symptoms, and both associations lost significance [s1]. The authors' reading is that while sleep fragmentation was associated with depressive symptoms, insomnia symptoms appear to be the key factor [s1].
Why that ordering matters
The finding is not that fragmentation is irrelevant. It is that the measurable, objective fragmentation in this sample did not carry information about depressive symptoms beyond what the subjective insomnia report already carried.
That is consistent with a large prospective literature. A meta-analysis of 34 cohort studies covering 172,077 participants, with an average follow-up of 60.4 months, found a pooled relative risk of 2.27 (95% CI, 1.89–2.71) for depression among people with insomnia [s2]. That analysis came with substantial heterogeneity (I² = 92.6%; P < 0.001) and evidence of publication bias on Egger's test, though correcting with trim-and-fill did not alter the combined estimate [s2].
Insomnia, in other words, has long been the variable with prospective predictive power. What the new analysis adds is a direct comparison, in the same participants, against objective polygraphic measures that are often assumed to be closer to the biology.
The limits, which are real
This is cross-sectional. Nothing in it establishes direction. Depression fragments sleep and generates insomnia complaints at least as readily as the reverse, and a single-timepoint analysis cannot separate those. The prospective evidence for insomnia preceding depression comes from elsewhere [s2].
The sample is also 68% male [s1], which matters given that both insomnia and depression are more commonly reported in women, and the two harmonised cohorts were each single-sex by design [s1]. Combining them produces a larger sample at the cost of a confounded sex distribution.
And a null on the apnea variables in a community sample is not a null in a clinical one. The severity distribution in an unselected population sits well below that of patients referred for suspected apnea, and an effect that requires severe hypoxic burden to appear would not be visible here.
The practical implication the authors draw is a screening one: assessing insomnia in people with depression and depression in people with insomnia, and treating both, may reduce symptom burden [s1]. That is a modest claim, and it is the one the data support.
Sources
- [s1] "Effect of Sleep, Sleep Apnoea and Insomnia Symptoms on Depressive Symptoms in Combined harmonised Community-Based Cohorts—A Cross-Sectional Analysis," Journal of Sleep Research, 28 August 2026. https://doi.org/10.1111/jsr.70431
- [s2] "Insomnia and the risk of depression: a meta-analysis of prospective cohort studies," BMC Psychiatry, 5 November 2016. https://doi.org/10.1186/s12888-016-1075-3
Sources
- Effect of Sleep, Sleep Apnoea and Insomnia Symptoms on Depressive Symptoms in Combined harmonised Community-Based Cohorts—A Cross-Sectional Analysis — Journal of Sleep Research , August 28, 2026
- Insomnia and the risk of depression: a meta-analysis of prospective cohort studies — BMC Psychiatry , November 5, 2016
More on
Sleep apnea plus insomnia tripled mortality risk in a six-year cohort of 2,401 patients
Moderate-to-severe sleep apnea alone roughly doubled all-cause mortality risk. Adding comorbid insomnia to the same patients raised it further — evidence for treating the combination as its own risk category.
Waking in the night is the norm. Struggling to get back to sleep is not.
About a third of adults on two continents wake at least three nights a week. The 7.7% who cannot resume sleep carry almost all of the daytime damage — and the cause is often not what the sleeper thinks.
Patients with insomnia and apnea saw a bigger blood-pressure drop from CPAP
After four weeks, the comorbid-insomnia group's blood pressure fell 20 points more than the apnea-only group. The finding flips the usual framing of COMISA as simply a worse-outcomes phenotype.
Sleep apnea plus insomnia raised hypertension risk — but only in one insomnia subtype
In a Pennsylvania cohort followed 7.5 years, the combination carried the highest risk when insomnia came with objectively short sleep. Insomnia with normal sleep duration showed no significant association.