Characterisation of major depressive disorder subgroups defined by weight and sleep changes

Abstract

Major depressive disorder (MDD) is a psychiatric illness that is highly heterogeneous in its presentation. As a result, it is postulated that MDD may contain subgroups, potentially with different underlying aetiologies. One such subgroup is atypical depression, which is increasingly defined as MDD with increased weight and sleep during depressive episodes. This subgroup is associated with earlier age of onset, increased episode frequency and severity of MDD. The symptoms of increased weight and sleep (↑WS) are also postulated to define an immunometabolic subgroup of depression, as individuals with these symptoms have higher rates of obesity and metabolic syndrome and ↑WS MDD is associated with polygenic risk scores for body mass index (BMI), triglycerides, leptin and C-reactive protein. However, less is known about MDD with decreased weight and sleep (↓WS) during depressive episodes or individuals who do not meet the criteria for either ↑WS or ↓WS (uncategorised), as genetic and immunometabolic analyses often primarily compare ↑WS MDD to a ‘non-↑WS’ group containing both ↓WS and uncategorised individuals.

Further work is required to better understand MDD subgroups defined by weight and sleep changes. Including investigating the potential neurological and inflammatory co-morbidities, the underlying genetic aetiology and the relationship with obesity and inflammation of ↑WS, ↓WS and uncategorised MDD. The overarching aim of this thesis is to further characterise subgroups of depression defined using weight and sleep change to increase understanding of their aetiology, open new avenues for research, and identify biomarkers for subgroups.

Analyses were primarily conducted in the UK Biobank (UKB), but additional cohorts were utilised for out-of-sample analyses (Generation Scotland, GS) and meta-analysis (Australian Genetics of Depression Study, AGDS). Association analyses compared the depressive characteristics and symptoms, current mental health, neurological and inflammatory co-morbidities, measures of obesity and immunometabolic markers between weight-sleep subgroups. The genetic architecture and aetiology of weight-sleep subgroups was examined by performing case-control and case-only genome-wide association studies. First using genotype data and then imputed data, before results from UKB were meta-analysed with AGDS. Lastly, genetic correlation and Mendelian randomisation analyses were used to further investigate the relationship between weight-sleep subgroups, obesity, metabolism and inflammation. The ↑WS subgroup was observed to have more severe depressive characteristics and worse current mental health, and the ↓WS less severe depressive characteristics and better current mental health than uncategorised MDD. The ↑WS subgroup was also associated with an increased risk of co-morbid epilepsy, migraine and asthma. Both the ↑WS and uncategorised subgroups were associated with obesity, but sex-specific associations between ↓WS MDD and obesity observed as well. Furthermore, the ↑WS subgroup not only had a poorer immunometabolic profile, but also genetic overlap with obesity measures and leptin, c-reactive protein, and apolipoprotein A levels. and there was evidence to suggest that both body mass index and apolipoprotein A may have a causal effect on the ↑WS subgroup. Lastly, genetic analyses did not find a significant difference in the SNP heritability of weight-sleep subgroups, but between weight-sleep subgroup genetic correlation analyses also suggest there may be differences in the genetic aetiology of ↑WS. Overall, this thesis presents a series of investigations highlighting novel differences between subgroups of MDD defined using weight and sleep changes which are consistent with the hypothesis that ↑WS symptoms may identify an immunometabolic subgroup of MDD. This work emphasizes the importance of characterising subgroups of MDD, in order to understand the heterogeneity and underlying mechanisms in MDD as a whole.