Pyroluria — A Correction

What this page used to claim

It claimed that pyroluria is a metabolic condition in which kryptopyrroles are excreted in urine, dragging zinc and vitamin B6 out with them; that the resulting deficiency impairs GABA synthesis and leaves glutamate unopposed; and that this produces treatment-resistant anxiety, OCD, ADHD and poor stress tolerance. It put prevalence at 10–15% of the general population and up to 40% of people with anxiety disorders. It listed a symptom cluster — white spots on the nails, stretch marks without weight change, no dream recall, no morning appetite — as a recognisable clinical picture. It pointed readers at a urine test costing around £60 and said almost nobody was running it.

Some of that was presented with more certainty than the evidence carried, and some of it does not survive contact with the evidence at all.

Where the idea came from

The history matters, because it explains how something can be sincerely believed for sixty years without ever being established.

In the late 1950s Abram Hoffer and Humphry Osmond were monitoring urine by paper chromatography using Ehrlich’s reagent, and found an unidentified spot they called the mauve factor. Their original description associated it with psychosis generally rather than with any specific diagnosis. In 1969 the compound was identified as kryptopyrrole in Nature. In 1978 the same investigator reversed that identification, concluding it was hydroxyhemopyrrolin-2-one — HPL. The clinical apparatus built on top of it — the name pyroluria, the colorimetric assay, the zinc-and-B6 mechanism, the high-dose protocol — came from Carl Pfeiffer’s group in the 1970s.

Most of that foundational literature appeared in the Journal of Orthomolecular Medicine and its predecessors, which is not indexed in MEDLINE. That does not make its contents false. It does mean the work never passed the editorial screening that indexing requires, and it is why so much of it has gone unexamined for so long.

Why it is retired — four reasons, in order of severity

1. The assay probably cannot measure what it claims to measure

In 2024, Sherwin and Shaw at the University of Canterbury did something that had not been done in sixty years of argument: they applied Ehrlich’s reagent to authentic pure standards of both candidate compounds and worked through the reaction chemistry.

The practical consequence: a commercial test marketed as measuring HPL is reporting a number that is not a measurement of HPL. Whatever the assay detects, it is not the compound named on the report.

I want to be accurate about the strength of this, because overstating a correction is the same error in the opposite direction. It is a small study — ten volunteers — and the authors themselves say further work is needed. They are not hostile to the field; they go on to discuss where kryptopyrrole might come from, including gut microbiome metabolism of stercobilin, and a plausible interaction with zinc. The chemistry argument is the solid part. The identification of what the mauve factor actually is remains open.

2. The same study found no difference between anxious and non-anxious people

This is the finding that should have ended the conversation, and it gets less attention than the chemistry. Sherwin and Shaw measured pyrrole concentrations in volunteers screened with the GAD-7 anxiety scale. Concentrations in the anxious group and the control group were similar.

The authors are careful here too: GAD-7 does not distinguish state from trait anxiety, the numbers are small, and they call for further study rather than declaring the matter closed. But the direction is the opposite of what this page previously asserted.

3. A systematic review found no reliable association at all

In 2021, Warren, Sarris, Mulder and Rucklidge published a PRISMA systematic review specifically asking whether pyroluria holds up. They screened 73 articles. Three found significantly higher HPL in a psychiatric population than in controls. The remaining clinical studies either found no association or did not report data adequate to show one. There were no placebo-controlled treatment trials of the zinc-and-B6 protocol. None.

Their stated conclusion is that HPL testing is not recommended as a screening or treatment tool.

It is worth noting where this was published: the Journal of Alternative and Complementary Medicine. This is not mainstream psychiatry dismissing an outsider idea from a distance. It is the complementary and nutritional field examining one of its own claims and finding it wanting, which is a great deal harder to wave away — and a great deal more to that field’s credit.

4. The thresholds were never validated

The cut-offs in circulation — above 15 µg/dL as clinically significant, above 20 µg/dL as clearly positive — derive from Pfeiffer-era clinical practice. They were not established by any study measuring sensitivity or specificity against an outcome. There is no standardised assay across laboratories and no external quality assurance scheme. The analyte is unstable, degrading with light, heat and delay, so a result depends substantially on how the sample was collected and shipped.

An earlier revision of this page corrected the thresholds while leaving the rest of the argument standing. That was a half-measure. If the assay is not measuring the named compound and the compound has no established relationship to symptoms, the threshold question does not arise.

If you were tested, or treated

A few honest things.

• The number on your report does not mean what it said it meant. That applies whether it was high or low.
• Feeling better on zinc and B6 does not validate the diagnosis. It may mean you were low in zinc, which is measurable directly. It may be regression to the mean, or the ordinary course of a fluctuating condition, or the effect of everything else that changed at the same time. Individual improvement is real and worth having; it is not evidence that the mechanism was correct.
• High-dose B6 is not a free bet. Sensory peripheral neuropathy is a documented risk. The evidence sits in the region of roughly 117–200 mg per day rather than the far higher figure often quoted, and some case reports describe symptoms at lower intakes. If you are still taking a high-dose B6 preparation on the strength of a pyroluria result, that is worth raising with whoever is looking after you.
• Your symptoms are not retracted. Anxiety that does not respond to standard treatment is real, and so is the experience of being told there is nothing to find. What has been withdrawn is one explanation for it, not the problem itself.

Why this page still exists

It would be easier to delete it. The article performed well, the claim is popular, and nobody would have noticed for a while.

But a correction that only the corrector can see is not a correction. Anyone who found the original should find this, at the same address, which is why the URL has not changed. And there is a harder reason: I published it. The reasoning that produced it — a plausible mechanism, a clinical picture that seemed to recur, a test that produced numbers, a literature I did not check hard enough because it agreed with me — is exactly the reasoning that produces the next mistake. Writing that down is more useful than pretending it did not happen.

What would change my mind back

A validated assay, established against a known standard, with published sensitivity and specificity. Independent replication showing that whatever the assay measures differs reliably between people with and without the relevant symptoms. And at least one placebo-controlled trial of the zinc-and-B6 protocol against a defined outcome.

If those appear, they will be reported here, and this page will change again. That is the arrangement.