Does Parkinson’s start in the gut? What the evidence actually shows

One of the most provocative ideas in neuroscience is that Parkinson’s disease — a disorder defined by tremor, rigidity and the loss of dopamine neurons in the brain — may not begin in the brain at all. A growing body of evidence places its earliest events in the gut, years before a neurologist would ever see the patient. The idea is genuinely supported, genuinely incomplete, and a useful case study in how to read microbiome science without overreading it.

The clue that came first: constipation

Long before the motor symptoms, many people who go on to develop Parkinson’s have gut trouble — most consistently, constipation. This is not folklore. In the Honolulu Heart Program, a prospective cohort of nearly 7,000 men followed for up to 24 years, Abbott et al. (2001) found that men with fewer than one bowel movement a day had roughly three times the risk of later being diagnosed with Parkinson’s, compared with men who were more regular — with the diagnosis often arriving more than a decade later. A symptom in the gut was foreshadowing a disease of the brain.

Braak’s ascending hypothesis

The anatomist Heiko Braak gave the observation a mechanism. In two 2003 papers, he proposed that the protein aggregates that define Parkinson’s — clumps of misfolded alpha-synuclein, the Lewy pathology — appear first in the lowest, most peripheral parts of the nervous system and climb upward over years. His staging scheme put the earliest damage in the dorsal motor nucleus of the vagus nerve and the olfactory system, reaching the dopamine-producing substantia nigra only at later stages. A companion paper went further, proposing that some unknown trigger crosses the gut lining and travels up the vagus nerve into the brain. Gut first, brain later.

The protein really is in the gut — early

That model made a testable prediction: the pathological protein should be findable in the gut wall before the disease is diagnosed. It is. Examining archived gastrointestinal biopsies taken from people for unrelated reasons, Stokholm et al. (2016) found aggregated, phosphorylated alpha-synuclein in the gut tissue of individuals who were only later diagnosed with Parkinson’s — in some cases up to 20 years before. The misfolded protein was sitting in the enteric nervous system — the gut’s own dense web of neurons — long before the brain showed its hand.

Cut the wire, lower the risk

If the trigger really ascends the vagus nerve, then severing that nerve should interrupt the journey. A natural experiment exists: decades ago, truncal vagotomy was a common surgery for peptic ulcers. Mining Danish national registries, Svensson et al. (2015) found that people who had a full truncal vagotomy had a lower subsequent risk of Parkinson’s, an effect most visible 20 years out. A later Swedish cohort pointed the same way. The honest caveat: in the Danish study the overall truncal-vagotomy result did not reach statistical significance — it is a striking, supportive signal, not a closed case.

Where the microbes come in

The disease also comes with a changed gut community. The landmark human study, Scheperjans et al. (2015), found that people with Parkinson’s had markedly less Prevotellaceae in their stool than matched controls, with other shifts tracking the severity of gait and balance problems. Far larger and higher-resolution, Wallen et al. (2022) used deep metagenomic sequencing in roughly 490 patients and 230 controls and found a broad, “disease-permissive” reshaping — over a third of measured species, genes and pathways altered. Whether that altered microbiome is a cause, a consequence, or a bystander is exactly the question these studies cannot, by design, answer.

The strongest causal evidence is in mice — and that matters

The closest thing to a causal demonstration comes from the lab, not the clinic. Sampson et al. (2016), working in the Mazmanian lab, used mice engineered to over-produce alpha-synuclein. Raised germ-free, those mice were largely protected from motor deficits and brain inflammation; give them a microbiome — especially one transplanted from Parkinson’s patients rather than healthy donors — and the symptoms worsened. The microbes were not incidental; in this model they were required. The unavoidable caveat: this is a genetically engineered mouse. Rodent guts, immune systems and disease timelines are not human ones, and “required in mice” is not “proven in people.”

Constipation precedes the tremor by years; the rogue protein appears in the gut wall first; severing the vagus seems to lower the risk; and in mice, the microbiome is necessary for the disease. That is a remarkably consistent story — built almost entirely from association and animal models.

Not one disease, and not always the gut

The most useful refinement is that Parkinson’s may not be a single road. The “body-first versus brain-first” model — laid out with multimodal imaging by Horsager et al. (2020) — proposes two routes: a body-first form in which pathology really does begin in the gut and autonomic nerves and ascends (often heralded by REM-sleep behaviour disorder), and a brain-first form that starts centrally. If that is right, the gut-origin story is true for a subset of patients, not all of them.

The honest bottom line

The gut-origin hypothesis for Parkinson’s is one of the best-supported ideas in microbiome neuroscience — and it is still a hypothesis. The human evidence (constipation cohorts, biopsies, vagotomy registries, microbiome surveys) is consistent but observational, vulnerable to confounding and reverse causation: an early gut that moves sluggishly could be the first symptom of brewing brain pathology rather than its cause. The clean causal evidence lives in mice. What you can say with confidence is that Parkinson’s has a real, early gut dimension worth taking seriously — and that the leap from “involved” to “originates there” is the part the field is still working to earn. For the signalling routes underneath all this, see our gut–brain axis primer.