Suicide is a major public health problem; in the UK the most recent data suggests a standardised rate of 10.1 per 100,000 (ONS 2024), with 6650 suspected suicides occurring from mid 2023 to mid 2024.

Numerous risk factors for suicide have been identified including particular age groups, male sex, a history of self-harm, substance and alcohol misuse, having a mental health problem and current social stresses (Franklin et al 2017). However, risk factors are often static, and not amenable to any kind of therapeutic modification (we cannot help aging, for example). It might be useful to identify factors that are modifiable to allow interventions to be developed.

Sleep is a fundamental biological process; problems with sleep can have profound impacts on our subjective health, including our mental health. Our sleep and wakefulness are governed by multiple factors including an intrinsic clock-like rhythm that our brains generate, the ‘circadian rhythm’ (Meyer et al 2022).

We all differ in level of activity over the 24 hour cycle e.g. whether we have stable patterns of waking or going to sleep, and whether we have a morning preference – “larks” or evening preference – “owls”. Social and environmental factors like light levels, social and occupational activities, exercise and meal times also influence these rhythms.

In this study, Walsh and colleagues set out to explore whether particular patterns of social and circadian rhythms were associated with suicidal thoughts and behaviours (STB).

Methods

The authors took a broad approach in this meta-analysis and systematic review of epidemiological studies. They used a robust search strategy to identify candidate papers, and the analysis presented was pre-registered in the PROSPERO database.

For exposures, the authors included both subjective (i.e. self-reported) and objective (i.e. using some technological product like actigraphy, which tracks physical movements) measures of social and circadian rhythms. The outcomes were suicidal thoughts and behaviours. Studies that measured exposures and outcomes using cross-sectional and prospective designs were included. Studies that only focused on other constructs, such as deliberate self-harm were excluded.

The investigators also extracted several additional study characteristics in order to carry out moderation analyses to explore heterogeneity in their results, including the age and sex of participants, clinical setting, and mental health diagnoses. They did not assess study quality in a systematic way.

Results

The authors identified a total of 2687 studies of which 52 were included in the review. After meta-analysis, subjective evening preference (being a “night owl”, compared to being a “morning lark” or an intermediate chronotype) was associated with both increased suicidal thoughts and behaviours (STB) and suicide attempts, with small effect sizes. It appeared that these associations were stronger in participants with depression, and in older participants.

The authors detected evidence for publication bias among these studies and used a technique that allowed them to estimate what the association between ‘eveningness’ and STB would be if possibly missing publications were available. They found that this would have reduced the size of the association, but it still appeared to remain above the null.

Looking at other measures of social and circadian rhythms, the study found a moderate-size association between “delayed sleep-wake phase disorder” and STB. Delayed sleep-wake phase disorder is a specific, clinically diagnosed sleep problem where a person’s circadian rhythm is not aligned with the rest of their life, so they have problems getting to sleep, waking up on time, or feeling excessively tired in the day.

Multiple other measures of circadian rhythms including measurements of melatonin and genes associated with circadian function were not associated with any measures of suicidal thoughts or behaviours. However, in a sensitivity analysis they found a large association between actigraphy-measured “amplitude” – the difference in activity between the most and least active hour of the day – and STB.

In addition to the cross-sectional analyses above, the authors found some evidence that there were prospective associations between self-reported measures of circadian disruption and STB. However, only one study (Maruani et al 2023) had a follow-up period of more than two weeks; that study found that within individuals with “treatment resistant depression”, at a year follow-up, daytime sleepiness and daytime dysfunction were associated with suicidal thoughts.

Conclusions

The authors concluded that:

evening chronotype is significantly, yet modestly, associated with STB [suicidal thoughts and behaviours], SI [suicidal ideation], and SA [suicide attempts], although the strength of these associations is attenuated when accounting for publication bias.

They also observed relationships between suicidal thoughts and behaviours (STB) and social rhythm disruption, overall circadian rhythm disorders, mean daily activity and actigraphy-assessed amplitude.

There were no associations between other measures including melatonin levels and circadian genotypes.

Strengths and limitations

Strengths

This study addresses a relevant question and synthesises a large amount of evidence. The authors have attempted to be comprehensive and cover a breadth of the different possible exposures and outcomes, and considered both subjective and objective methods of measuring circadian rhythms. The authors carried out meta-analyses where possible to give an overall synthesis of associations, and examined for presence of publication bias and the impact this might have on their results. Their discussion is comprehensive and thoughtful, linking this study to a range of other evidence and relevant theoretical models.

Limitations

This meta-analysis of epidemiological studies, mostly cross-sectional, is unable to assess whether the observed relationships are causal. For example, does having an evening preference cause increased suicidal thoughts or behaviours, or does a pre-disposition towards these types of thoughts cause delayed sleep?

Another key challenge for this study is the combination of quite a disparate set of studies with very different characteristics. For example, the three prospective studies they combined were one study with a year of follow-up and two studies with only 2 weeks of follow-up. It is hard to imagine that these follow-up periods are clinically comparable.

The authors could have further improved their meta-analysis by using a formal risk of bias tool to assess the quality of the included studies such as the ROBINS-E tool (Higgins et al 2024). Assessing risk of bias could allow further sensitivity analysis comparing higher and lower-quality studies.

Although the authors provide a link to the data they used, the link leads to an OSF repository to which access has to be requested, so the data is not open access. It is also difficult to identify all of the included papers as the reference list does not include all the reviewed studies. These problems limit the ability of readers to scrutinise the results.

Implications for practice

This study provides an interesting addition to our understanding of risk factors for suicidal thoughts and behaviours. However, circadian traits are complex and at least partially genetically determined (e.g. chronotype is estimated to have a heritability of 13.7%, Jones et al 2019). It is unclear whether knowledge of a pattern of circadian rhythms would produce a useful avenue for intervention.

More broadly, prediction tools for suicide and self-harm have not yet proved clinically useful. For example, a recent model by Fazel and colleagues (Fazel et al 2023) had a sensitivity of 82%, and specificity of 54%. This means that in people who presented to healthcare services with deliberate self-harm, the model failed to identify 1 in 5 people who went on to die by suicide, and around half of people identified as at most risk did not go on to die. The challenge of predicting suicide is extremely difficult because on a population level, suicide is fortunately a rare event.

Statement of interests

I have no personal or professional link to this study or its authors.

Links

Primary paper

Walsh RFL, Maddox MA, Smith LT, Liu RT, Alloy LB. Social and circadian rhythm dysregulation and suicide: A systematic review and meta-analysis. Neurosci Biobehav Rev. 2024 Mar;158:105560. doi: 10.1016/j.neubiorev.2024.105560. Epub 2024 Jan 23. PMID: 38272337; PMCID: PMC10982958.

Other referenced studies

Fazel S, Vazquez-Montes MDLA, Molero Y, Runeson B, D’Onofrio BM, Larsson H, Lichtenstein P, Walker J, Sharpe M, Fanshawe TR. Risk of death by suicide following self-harm presentations to healthcare: development and validation of a multivariable clinical prediction rule (OxSATS). BMJ Ment Health. 2023 May;26(1):e300673. doi: 10.1136/bmjment-2023-300673. PMID: 37385664; PMCID: PMC10335583.

Franklin JC, Ribeiro JD, Fox KR, Bentley KH, Kleiman EM, Huang X, Musacchio KM, Jaroszewski AC, Chang BP, Nock MK. Risk factors for suicidal thoughts and behaviors: A meta-analysis of 50 years of research. Psychol Bull. 2017 Feb;143(2):187-232. doi: 10.1037/bul0000084. Epub 2016 Nov 14. PMID: 27841450.

Higgins JPT, Morgan RL, Rooney AA, Taylor KW, Thayer KA, Silva RA, Lemeris C, Akl EA, Bateson TF, Berkman ND, Glenn BS, Hróbjartsson A, LaKind JS, McAleenan A, Meerpohl JJ, Nachman RM, Obbagy JE, O’Connor A, Radke EG, Savović J, Schünemann HJ, Shea B, Tilling K, Verbeek J, Viswanathan M, Sterne JAC. A tool to assess risk of bias in non-randomized follow-up studies of exposure effects (ROBINS-E). Environ Int. 2024 Apr;186:108602. doi: 10.1016/j.envint.2024.108602. Epub 2024 Mar 24. PMID: 38555664; PMCID: PMC11098530.

Jones SE, Lane JM, Wood AR, van Hees VT, Tyrrell J, Beaumont RN, Jeffries AR, Dashti HS, Hillsdon M, Ruth KS, Tuke MA, Yaghootkar H, Sharp SA, Jie Y, Thompson WD, Harrison JW, Dawes A, Byrne EM, Tiemeier H, Allebrandt KV, Bowden J, Ray DW, Freathy RM, Murray A, Mazzotti DR, Gehrman PR, Lawlor DA, Frayling TM, Rutter MK, Hinds DA, Saxena R, Weedon MN. Genome-wide association analyses of chronotype in 697,828 individuals provides insights into circadian rhythms. Nat Commun. 2019 Jan 29;10(1):343. doi: 10.1038/s41467-018-08259-7. PMID: 30696823; PMCID: PMC6351539.

Maruani J, Molière F, Godin O, Yrondi A, Bennabi D, Richieri R, El-Hage W, Allauze E, Anguill L, Bouvard A, Camus V, Dorey JM, Etain B, Fond G, Genty JB, Haffen E, Holtzmann J, Horn M, Kazour F, Nguon AS, Petrucci J, Rey R, Stephan F, Vaiva G, Walter M; FondaMental Advanced Centers of Expertise in Resistant Depression (FACE-DR) Collaborators; Lejoyeux M, Leboyer M, Llorca PM, Courtet P, Aouizerate B, Geoffroy PA. Diurnal symptoms of sleepiness and dysfunction predict future suicidal ideation in a French cohort of outpatients (FACE-DR) with treatment resistant depression: A 1-year prospective study about sleep markers. J Affect Disord. 2023 May 15;329:369-378. doi: 10.1016/j.jad.2023.02.106. Epub 2023 Feb 24. PMID: 36842655.

Meyer N, Harvey AG, Lockley SW, Dijk DJ. Circadian rhythms and disorders of the timing of sleep. Lancet. 2022 Sep 24;400(10357):1061-1078. doi: 10.1016/S0140-6736(22)00877-7. Epub 2022 Sep 14. Erratum in: Lancet. 2023 May 13;401(10388):1570. doi: 10.1016/S0140-6736(23)00908-X. PMID: 36115370.

Photo credits

• Photo by Brandon on Unsplash
• Photo by insung yoon on Unsplash
• Photo by Luis Argaiz on Unsplash
• Photo by Adrian Swancar on Unsplash
• Photo by Susan Q Yin on Unsplash
• Photo by Sangharsh Lohakare on Unsplash