The evidence base

What The Neuropathy Trials Actually Used

Four things have controlled human trials in peripheral neuropathy, and every one of them comes with a dose on record. Those four figures are the most portable piece of knowledge anyone shopping this shelf can carry, because they work on any bottle in any shop.

By Nervelin Team · Last updated October 29, 2026 · 9 min read · The evidence base · Reviewed by Dr. Stephen Barrett, DPM

Why the dose column is the portable part

An ingredient name on a bottle tells you almost nothing. The same plant appears at 10 mg and at 1,000 mg, in extracts standardised to wildly different strengths, in products that cost the same. What separates them is the amount, and the only amount worth measuring against is the one a trial actually administered.

So the table below is built around the dose column rather than the findings column. Whatever you conclude about whether any of these works, the figures are fixed and they are the yardstick.

Five rows, and the dose column is the one worth carrying into a shop. Every figure is the amount a named trial administered.

Three of the five rows ran for months or years. That is worth noticing before anything else: this is not a field where a fortnight settles anything, and a bottle bought as a two-week trial has not tested the question the research asked.

Alpha-lipoic acid, and the four-year trial

SYDNEY 2 is the trial most often quoted in this category. It gave alpha-lipoic acid orally at 600 mg once daily and reported on symptomatic diabetic polyneuropathy over five weeks. The dose is the part to remember: one capsule-sized amount, once a day.

NATHAN 1 is the harder test and the more interesting one. Same 600 mg once daily, run for four years, which is an unusual commitment for a supplement ingredient and puts it in different company from most of this shelf.

What the long trial changes about the short one

A five-week result tells you something can shift symptoms quickly. A four-year result asks whether anything is still true later, which is the question anybody buying six bottles is implicitly asking.

Reading both together is the honest way to use this ingredient: a specific amount, tested over both a short and a long horizon, in a defined population. That is a great deal more than most of this aisle can say.

Benfotiamine, and why the form is the point

Benfotiamine is a fat-soluble derivative of thiamine, vitamin B1. The form is the whole reason it is studied separately: it is absorbed differently from ordinary thiamine, which is why a trial of one is not a trial of the other.

BENDIP is the randomised, double-blind, placebo-controlled study in diabetic polyneuropathy, and it ran two arms at 300 mg and 600 mg a day over six weeks. Two doses in one trial is unusually useful, because it puts a range on the shelf rather than a single number.

Two things to carry out of this section:

• A label saying “thiamine” is not a label saying “benfotiamine”, and the research is on the second.
• The studied range is 300 to 600 mg a day, which is a large amount by supplement standards and hard to hide inside a blend.

Acetyl-L-carnitine, and a Cochrane verdict

The pooled analysis of two 52-week trials is the substantial piece of work here. It ran 500 or 1,000 mg three times daily — so 1,500 to 3,000 mg a day — and reported on pain, nerve regeneration and vibratory perception in chronic diabetic neuropathy.

Three times daily at a gram a time is a demanding routine, and it is worth stating plainly because it is the kind of detail that vanishes when an ingredient is reduced to a name on a front label.

The Cochrane review assessed the field as a whole, which is the right document to read before deciding what the pooled analysis means. A systematic review exists to weigh trials against each other rather than to repeat the most favourable one.

Vitamin B12, which is a different kind of question

B12 belongs in this list for a reason that makes it unlike the other three. It is not a supplement with a trial in neuropathy so much as a deficiency that causes neuropathy, and replacing it is treatment of a cause rather than support of a symptom.

The Diabetes Prevention Program follow-up established the scale of the problem: long-term metformin use is associated with B12 deficiency, and metformin is among the most widely prescribed medicines in the world.

A systematic review of supplementation in metformin users looks at what happens when the deficiency is corrected, which is the practical question for anyone in that position.

The reason this matters to a shopper is blunt: if a medicine has depleted your B12, a botanical blend is not the answer to that problem and a blood test is. This is the clearest example in the whole category of a treatable cause being identifiable rather than guessed at.

What the guidelines do with all of this

The 2022 American Academy of Neurology guideline update covers oral and topical treatment of painful diabetic polyneuropathy. It is a guideline about medicines, and the honest reading of it is that supplements are not what it recommends.

That is not a contradiction of everything above. A guideline weighs the whole evidence base against the demands of clinical practice, and 'a trial exists' is a lower bar than 'this is what to do first'. Both statements can be true.

What it means practically is an order of operations. Find the cause, address what is addressable, and treat a supplement as something taken alongside that rather than instead of it.

Reading a botanical panel against these four

Now the comparison this article exists to make. Put the four trial rows beside the label this website sells and the result is unambiguous.

Four rows, and the right-hand column is the same in all four. The nine names on this panel are a different set of ingredients with a different research base.

The nine names on this panel — L-lysine, apple cider vinegar, Garcinia cambogia, Tongkat Ali, horny goat weed, wild yam, cinnamon, bitter melon and raspberry ketones — are a different set of ingredients with research in different conditions. The ingredients page sets each one beside what it was actually studied in and the dose that study used.

There is a second obstacle underneath the first. Even if one of the four were on this panel, the 1000mg blend total means no amount could be read for it, so the comparison would still be impossible. Two separate reasons, and either one on its own would be enough.

References