4,500 Children Gave a Blood Sample at Nine. What It Predicted at Eighteen

A long-running Bristol birth cohort found that inflammation levels measured in childhood tracked with depression nine years later. Here is what that finding means, what the genetics of depression has since added, and how much of it applies to any one person.

In Avon County, England, roughly 4,500 nine-year-olds gave a non-fasting blood sample. It was measured for two ordinary markers of inflammation: interleukin-6 and C-reactive protein — the same things a doctor might check for an infection.

Nine years later, at eighteen, the same young people sat structured psychiatric assessments.

The children who had been in the top third for IL-6 at age nine were more likely to be depressed at eighteen than those in the bottom third — an adjusted odds ratio of 1.55 (95% CI, 1.13–2.14), after accounting for sex, body mass index, ethnicity, social class, earlier psychological problems and maternal postnatal depression. The relationship was dose-dependent (Khandaker et al., JAMA Psychiatry, 2014, PMID: 25133871).

The direction matters. The inflammation came first.

Why that was surprising

Depression and inflammation had been seen together for years, but nobody could say which caused which. Being depressed is itself stressful, and stress raises inflammatory markers, so cross-sectional studies could not settle it.

A prospective birth cohort can. The Avon Longitudinal Study of Parents and Children followed these participants from birth, which is why a blood sample taken at nine can say something about a diagnosis at eighteen.

The broader case for inflammation as a causal contributor — rather than a consequence — is set out in detail by Miller and Raison (Nature Reviews Immunology, 2016, PMID: 26711676). It remains a hypothesis with good supporting evidence, not settled fact.

What genetics added

In 2018, a consortium led from the University of Queensland ran a genome-wide association meta-analysis of major depression across 135,458 cases and 344,901 controls and identified 44 independent significant loci (Wray et al., Nature Genetics, 2018, PMID: 29700475).

The single most important sentence in that paper is not about any gene:

All humans carry lesser or greater numbers of genetic risk factors for major depression.

There is no depression gene. There is a continuous distribution of risk that every person sits somewhere on, built from many variants of individually tiny effect. A follow-up meta-analysis raised the count to 102 independent variants (Howard et al., Nature Neuroscience, 2019, PMID: 30718901).

The Wray analysis also found that lower educational attainment and higher body mass appeared putatively causal for depression, and that depression and schizophrenia share part of their biological basis.

The limits, stated plainly

No single variant is worth naming. Each of the 44 loci shifts risk by a very small amount. Naming one would misrepresent how this works — which is why you will not find a marker of the week in this post.

A polygenic score explains only a small fraction of who becomes depressed. It captures a low single-digit percentage of variation in liability. Bereavement, poverty, isolation, chronic illness and childhood adversity are not in that score, and they matter enormously.

An odds ratio of 1.55 is a shift across a population, not a forecast for a person. Most children in that top IL-6 third were not depressed at eighteen.

Inflammation is not yet a treatment target. Trials of anti-inflammatory drugs in depression have produced mixed results, and nothing here is a reason to start or stop any medication.

This is genetic and biological predisposition, not diagnosis. Depression is diagnosed clinically, by a doctor, and treated effectively in most people. If you are struggling, that conversation belongs with a clinician — not with a genetic report.

What the app does, and what it does not

The Depression App applies the findings of the Wray et al. analysis to your uploaded genotype and places you on the distribution of polygenic risk described in that study, citing the source throughout.

It does not diagnose depression. It cannot tell you whether you will become depressed, and it does not measure inflammation. What it reports is where your genome sits on one measured axis among many — a number that is real, modest, and easy to over-read.

The children in Bristol are the better illustration of the science. Something measurable at nine shifted the odds at eighteen. It did not decide them.