Severe mental disorders like major depressive disorder, bipolar disorder, and schizophrenia share a genetic overlap with markers of metabolic health, but the nature of this relationship differs depending on the specific condition. New evidence indicates that the genetic pathways linking depression to metabolic markers resemble those of physical conditions like heart disease, whereas schizophrenia and bipolar disorder display a reversed genetic pattern. These findings were published in the American Journal of Psychiatry.
People diagnosed with severe mental disorders frequently experience higher rates of physical health problems, such as type 2 diabetes and coronary artery disease. These physical conditions often involve disruptions in metabolism, the process by which the body converts food into energy. Metabolic processes leave behind small molecules called metabolites in the bloodstream. Examples include cholesterol, fatty acids, and glucose.
“People with severe mental disorders die 10 to 20 years earlier than the general population, mostly from cardiometabolic disease,” said Dennis van der Meer, a research group leader at the Centre for Precision Psychiatry at the University of Oslo and an adjunct at Maastricht University. “It is hard to tell from observational data how much of this is shared biology versus medication and lifestyle. Genetics gives us a way to look at the shared biology without those confounders, and the required large-scale data for both the disorders and blood metabolic markers has only recently become available.”
Scientists wanted to understand if the shared occurrence of mental and physical health conditions stems from a common genetic foundation. Both severe mental health conditions and metabolic traits are known to be influenced by genetics. By examining the DNA sequences of large groups of people, researchers can estimate how much of the genetic influence on one trait overlaps with another.
Medications used to treat schizophrenia and bipolar disorder tend to cause weight gain and metabolic disruption as side effects. This makes it difficult to tell if poor metabolic health is a consequence of treatment, lifestyle factors, or an underlying biological predisposition. Mapping the genetic relationships between these disorders and metabolic markers provides a way to look past the effects of medication to examine the foundational biology.
To conduct the research, the scientists analyzed genetic data from several massive international databases. They looked at summary statistics from over 300,000 individuals of European descent, combining information from the UK Biobank and the Estonian Biobank. The researchers focused on 249 specific metabolic markers measured from blood samples. These included various lipids, amino acids, and indicators of inflammation.
The authors compared the genetic profiles associated with these markers to the genetic profiles of major depressive disorder, bipolar disorder, and schizophrenia. They also included genetic data for body mass index, type 2 diabetes, and coronary artery disease for comparison.
For a portion of the analysis looking at real-world diagnoses, the researchers examined medical records from 207,836 participants in the UK Biobank. This group included 13,887 people with major depressive disorder, 769 with bipolar disorder, 447 with schizophrenia, 17,299 with type 2 diabetes, and 25,638 with coronary artery disease. The researchers adjusted their statistical models for age and sex. They also controlled for the influence of body mass index in a secondary genetic analysis to see if body weight alone explained the genetic associations.
The analysis indicated an extensive genetic overlap between the metabolic markers and all three mental health conditions. However, the pattern of this overlap was surprisingly different among the disorders. Major depressive disorder showed a genetic correlation pattern that almost perfectly matched the patterns seen for type 2 diabetes, coronary artery disease, and high body mass index.
In contrast, schizophrenia and bipolar disorder showed an opposing genetic correlation pattern. The genetic variants associated with an increased risk for schizophrenia and bipolar disorder were linked to lower levels of the cardiometabolic risk markers. This divergence in genetics contradicted what the researchers saw in the real-world medical records, where all three disorders were associated with poor metabolic health in practice.
“How sharp the divergence was” came as a surprise, van der Meer told PsyPost. “These disorders are genetically correlated with each other and look similar in patients, yet their genetic relationships with the metabolic markers went in opposite directions. We expected some differences in point estimates but definitely not opposites across the board.”
This divergence provides evidence for the role that environmental factors and treatment play in certain mental health conditions.
“Mental and metabolic health are connected at a biological level, but differently per disorder,” van der Meer explained. “Depression genetically resembles type 2 diabetes and heart disease. Schizophrenia and bipolar disorder are the opposite: genetically they align with more favorable levels of many metabolic markers, even though patients have poor metabolic health in practice. For those disorders the metabolic burden appears largely driven by non-genetic factors, medication and lifestyle in particular, meaning it should be preventable.”
To explore this further, the researchers used a specialized statistical model to estimate the actual fraction of shared causal genetic variants, regardless of whether the variants increased or decreased risk. This model indicated that major depressive disorder shares about 77 percent of causal genetic variants with the metabolic markers. Bipolar disorder shared 85 percent, and schizophrenia shared 52 percent.
The key difference lay in the direction of the genetic effects. For major depressive disorder, most of the shared genetic variants affected the disorder and the metabolic markers in the same direction. For bipolar disorder and schizophrenia, roughly half of the shared genetic variants had opposing effects on the mental disorder and the metabolic marker.
The authors also used a statistical technique called Mendelian randomization to estimate causal pathways between the traits. This method uses genetic information to mimic the effects of a randomized trial. The results suggested that metabolic markers have a robust causal effect on the development of all three mental disorders.
“The individual genetic correlations are modest, which is expected for complex traits,” van der Meer said. “What matters is the consistency of the pattern across all 249 markers, and that many markers show causal effects on the disorders, what makes them potentially modifiable targets.”
Major depressive disorder was the only condition that also showed a broad causal influence back onto the metabolic markers. Interestingly, the analysis did not find reliable evidence that high body mass index or type 2 diabetes directly cause severe mental disorders. Instead, the metabolic markers appear to act as an intermediate step, bridging the physical and mental health conditions.
Finally, the researchers mapped the shared genetic variants to specific genes and biological pathways. They identified 1,056 genes shared between the mental disorders and the metabolites. These genes are known to be involved in energy metabolism, immune system function, and the structural development of brain synapses. Tests showed these genes are highly active not just in the brain, but also in the heart and liver.
The exact mechanisms linking schizophrenia and bipolar disorder to poor metabolic health in real life require further exploration. The genetic data suggests these two disorders are associated with healthier metabolic profiles, yet patients frequently suffer from metabolic diseases. This discrepancy points to environmental factors, such as the side effects of antipsychotic medications, diet, and sedentary lifestyles, as major drivers of metabolic illness in these populations.
“This is not an argument against medication, antipsychotics remain essential treatment,” van der Meer emphasized. “The implication is that metabolic side effects need proactive monitoring and management. Also, our analyses were limited to European-ancestry samples, so the findings need validation in other populations.”
Another limitation involves the use of electronic health records to identify people with mental health conditions. Medical records only capture individuals who seek treatment and receive a formal diagnosis, which likely misses people with milder symptoms who never interact with the healthcare system. The metabolic markers measured in this study are also highly correlated with one another, making it difficult to isolate exactly which specific biological molecule drives the observed effects.
Looking forward, the researchers plan to continue exploring how these biological pathways might guide treatment.
“We want to dissect this at a finer level, e.g. whether diagnostic subtypes have distinct metabolic genetic profiles, ideally in deeply phenotyped clinical cohorts,” van der Meer said. “We are also looking into the effects of these metabolites on other disorders and brain measures (see our preprints). Ultimately, we’d like to know whether metabolic pathways can be treatment targets, which the recent GLP-1 findings in psychiatry make all the more relevant.”
Medications like GLP-1 receptor agonists, which are commonly used for weight loss and type 2 diabetes management, have recently drawn interest from scientists investigating whether they might also alleviate psychiatric symptoms.
The study, “Divergent Patterns of Genetic Overlap Between Severe Mental Disorders and Metabolic Markers,” was authored by Dennis van der Meer, Alexey A. Shadrin, Sara E. Stinson, Elise Koch, Jaroslav Rokicki, Zillur Rahman, Jacob Bergstedt, Aigar Ottas, Ida E. Sønderby, Linn Rødevand, Julian Fuhrer, Daniel S. Quintana, Anders M. Dale, Kevin S. O’Connell, Srdjan Djurovic, Kelli Lehto, Lili Milani, Maris Alver, and Ole A. Andreassen.