Your Depression and Your Heart Disease May Share the Same Genetic Roots

Walk the wards of any busy hospital and a pattern emerges, one that clinicians have noticed for decades without quite being able to explain. The patient with schizophrenia who keeps returning with gastrointestinal problems. The woman managing both PTSD and heart failure. The teenager with ADHD whose chart is thick with respiratory diagnoses, musculoskeletal complaints, a sleep disorder. For a long time, the working assumption was that these were coincidences, or perhaps downstream effects: mental illness leads to poor self-care, poor self-care leads to physical disease. A new genetic analysis of nearly 2 million people suggests the story is considerably stranger than that.

The study, published in Nature Communications, finds that the same chunks of DNA that predispose people to psychiatric disorders also predispose them to a wide range of physical illnesses, and that the overlap is not occasional or modest. It is, in the language of the paper, pervasive and substantial.

Andrew Grotzinger, an assistant professor of psychology and neuroscience at the University of Colorado Boulder who led the research, is careful about how he frames it. “The surprising finding here is not that psychiatric disorders and medical disorders are linked, but rather, how much they are linked,” he says. His team found that when you compare the genetic risk factors for mental and physical conditions, they overlap 42% of the time. “At the genetic level, we found that there is so much overlap they are really not two different classes of diseases at all.” That is a striking claim. And the data behind it are, by any measure, large.

The researchers used a statistical approach called Genomic Structural Equation Modelling, applied to GWAS summary statistics drawn from roughly 1.9 million individuals. They mapped 73 physical disorders across eight medical domains, from respiratory and circulatory conditions to digestive, musculoskeletal and genitourinary disease, and then examined how those genetic risk structures aligned with five clusters of psychiatric illness: compulsive disorders, psychotic and thought disorders, neurodevelopmental conditions, internalising disorders, and substance use disorders. The question was not whether any individual pair of conditions shared genetic risk. It was whether that sharing was transdiagnostic, cutting across diagnostic categories rather than being confined to tidy one-to-one pairings.

The answer was yes, though with some striking variation. Neurodevelopmental conditions turned out to be the most promiscuously entangled with physical disease. ADHD, in particular, showed a higher average genetic correlation with physical illnesses than it did with other psychiatric conditions, which is a genuinely odd finding if you still think of psychiatric disorders as something categorically different from diseases of the body. PTSD and major depression showed a similar pattern; their genetic signal appeared to be unusually broad, indexing risk pathways that sprawl across both psychiatric and physical domains in a way that most diseases, even severe ones, simply do not.

“In the clinic, you rarely see someone with just one condition walk into a room,” says Jeremy Lawrence, a PhD candidate in clinical psychology and the paper’s first author. “If we can better understand the cross-talk between these conditions, we can do a better job helping the whole patient.”

Not every psychiatric category behaved this way. Compulsive disorders, meaning OCD, Tourette syndrome and anorexia nervosa, showed relatively little genetic overlap with physical illness and, in one notable case, appeared to have a protective genetic relationship with a cluster of digestive conditions including acute pancreatitis and peptic ulcers. Psychotic and thought disorders, principally bipolar disorder and schizophrenia, sat somewhere in the middle, showing modest correlations with digestive, musculoskeletal and respiratory conditions, though with bipolar disorder driving much of the signal rather than schizophrenia. The picture that emerges is not a simple gradient from “mentally ill” to “physically ill” but something more like a partially overlapping web, where some psychiatric conditions are deeply embedded in the broader genome-wide architecture of physical disease and others seem to operate through quite different mechanisms.

To investigate this further, Grotzinger’s team identified a single common genetic factor that cut across 21 physical disorders spanning six of the eight medical domains, from heart failure and chronic airway obstruction to urinary tract infections and sleep disorders. Twenty-seven genome-wide significant risk loci mapped to this transdiagnostic physical illness factor. When they modelled that factor alongside the psychiatric clusters, the genetic correlations with internalising disorders, neurodevelopmental conditions, and substance use disorders were all substantial, ranging from about 0.61 to 0.68. Something in the shared genetic architecture of a depressive disorder is, in a real molecular sense, part of the same risk landscape as something in the shared genetic architecture of the body’s organs failing.

What that something is, nobody yet knows for certain. The correlations could reflect what geneticists call horizontal pleiotropy, where a genetic variant influences multiple traits simultaneously through separate pathways, possibly via some upstream process like genetically mediated personality or stress reactivity that cascades across systems. Or they could reflect vertical pleiotropy, where one condition causally influences another through a chain of effects: substance use damaging physical organs, depression affecting cardiovascular function through behaviour or inflammation. Probably both mechanisms are operating, in different proportions for different disorder pairs. Untangling them will require the kind of fine-mapping and multi-omic work that is only now becoming computationally feasible at scale.

There is also a practical constraint on how far to run with these findings: the study relied exclusively on GWAS summary statistics from people of European ancestry, which limits how far the results generalise. And while the genetic correlations are striking, they describe population-level patterns in common variants; they say relatively little about rare genetic mutations or about how these shared risk pathways play out in any individual patient. Grotzinger is also candid that the lack of diagnostic tools for psychiatric disorders has historically made the field seem less grounded than cardiology or oncology. “You can ask someone to spit in a tube or put a blood pressure cuff on to diagnose physical illness, but in many ways, we don’t have that for psychiatric disorders, so some have viewed them as more esoteric and less tangible,” he says. “Psychiatric disorders are just as real as any medical disease. Our findings help make that argument.”

Lawrence points to GLP-1 agonists as a rough sketch of where this kind of thinking might eventually lead: drugs developed for diabetes, later adopted for obesity, now showing early signals of usefulness in substance use disorders. If the genetic risk architecture of addiction and metabolic disease are genuinely overlapping, that pharmacological promiscuity starts to look less like a fluke and more like an inevitability. The next generation of drugs targeting these shared upstream pathways, whatever they turn out to be, might treat conditions we currently think of as wholly separate in one go. Whether that ambition can be translated into actual therapies depends on work that has not yet been done. But the old organisational logic of medicine, with its firm boundary between the psychological and the organic, is looking shakier by the year.

DOI: https://doi.org/10.1038/s41467-026-69218-1