Back

Association Between Circadian Rhythms, Sleep, And Depression: A Bidirectional Mendelian Randomisation Study In The UK Biobank

Paz, V.; Williams, D. M.; Richards, M.; Tassino, B.; Silva, A.; Garfield, V.

2024-12-10 epidemiology
10.1101/2024.12.08.24318680 medRxiv
Show abstract

Increasing evidence links circadian and sleep disruptions with depression, though whether they are causally related remains unclear. This study used Mendelian randomization (MR) with exposures instrumented using genome-wide genetic variants and data from the UK Biobank (n=408 480; mean age=56.9 years) to explore causal associations between circadian and sleep traits, and depression. We conducted bidirectional MR analyses between morningness, sleep duration, insomnia symptoms, and depression, and secondary unidirectional analyses with circadian and sleep traits as exposures and mixed anxiety-depressive symptoms and anxiety disorders as outcomes. Using inverse-variance weighting and sensitivity analyses for horizontal pleiotropy (MR-Egger and Weighted Median Estimator), we found suggestive associations between morningness and lower odds of depression (diagnosis and treatment: OR=0.89, 95%CI=0.80, 0.98), and depression and shorter sleep duration ({beta}=-4.67 minutes, 95%CI=-7.44; -1.89). Depression showed an effect on insomnia (any symptoms: OR=1.22, 95%CI=1.11; 1.34; frequent symptoms: OR=1.30, 95%CI=1.19; 1.42), while frequent insomnia was associated with increased odds of depression (diagnosis: OR=1.14, 95%CI=1.10; 1.19; diagnosis and treatment: OR=1.10, 95%CI=1.06; 1.14). Insomnia also had an effect on mixed anxiety-depressive symptoms (any symptoms: OR=1.19, 95%CI=1.08, 1.32; frequent symptoms: OR=1.20, 95%CI=1.17, 1.23), but no associations were found between the exposures and anxiety disorders. These findings shed light on the causal links between the circadian system and depression.

Matching journals

The top 7 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.