Alpha-Lipoic Acid: Glucose Disposal, Neuropathy and the Trials That Matter

Key Takeaways

Alpha-lipoic acid is two products wearing one name. In Germany it is thioctic acid, a licensed prescription treatment for diabetic polyneuropathy with a trial series stretching from 1995 to 2011 and a total enrolment in the thousands. In the English-speaking supplement market it is a "blood sugar support" antioxidant whose glucose evidence consists of a 74-patient pilot from 1999 and a pair of meta-analyses that do not agree with each other.

The neuropathy trials are worth reading in detail because they establish something most supplements never manage: a dose. Six hundred milligrams a day improves symptoms; twelve or eighteen hundred does not improve on it and roughly doubles the side effects. The four-year trial then shows what a symptom drug does and does not do to the underlying disease.

The glucose story is softer, and this article says so. And the safety story — an autoimmune hypoglycaemia syndrome with a known HLA association — is the part that belongs on every bottle and appears on none.

What Lipoic Acid Is, and What the Supplement Is Not

Inside the mitochondrion, lipoic acid is covalently bound to lysine residues on pyruvate dehydrogenase and alpha-ketoglutarate dehydrogenase, the enzyme complexes that gate entry to the citric acid cycle. Every aerobic cell synthesises it. There is no dietary deficiency state.

Supplemental alpha-lipoic acid is free, unbound lipoic acid, usually a racemic mix of the natural R-enantiomer and the synthetic S-enantiomer, at doses several hundred times the body's bound pool. It does not "top up" the enzyme cofactor. What it does is act as a dithiol: it reduces disulphides, regenerates glutathione, vitamin C and vitamin E, and chelates transition metals. The neuropathy rationale is that oxidative stress in the small vessels supplying peripheral nerves is part of the lesion, and a fast-acting antioxidant that reaches nerve tissue might relieve it.

That the compound also contains free sulphhydryl groups — the feature that makes it an antioxidant — is what makes it an immunological trigger in the wrong host. That comes later.

The Neuropathy Trials, in Order

ALADIN: the dose is set

The first trial is the one that established 600 mg. Three hundred and twenty-eight patients with type 2 diabetes and symptomatic neuropathy received three weeks of intravenous ALA at 100, 600 or 1,200 mg, or placebo. The Total Symptom Score fell by 5.0 points (63.5%) on 600 mg, 4.5 points (58.6%) on 1,200 mg, 3.3 points (43.2%) on 100 mg and 2.6 points (38.4%) on placebo. Six hundred milligrams beat placebo at P < 0.001 and had the best responder rate — 82.5% improving by at least 30%, against 57.6% on placebo.

The 1,200 mg arm was not better, and its adverse-event rate was 32.6% against 18.2% on 600 mg and 20.7% on placebo. Right from the first trial, the higher dose bought side effects and nothing else.

SYDNEY 2: it works by mouth

The intravenous data would be academic for a supplement user if oral ALA did not reproduce them. SYDNEY 2 tested that directly: 181 patients in Russia and Israel took 600, 1,200 or 1,800 mg once daily by mouth, or placebo, for five weeks after a placebo run-in.

Mean TSS fell by 4.9 points (51%) on 600 mg, 4.5 (48%) on 1,200 mg, 4.7 (52%) on 1,800 mg, and 2.9 (32%) on placebo — all three doses significant against placebo. Responder rates (at least a 50% reduction) were 62%, 50%, 56% and 26%. Stabbing pain, burning pain, the Neuropathy Symptoms and Change score and patients' global assessment all favoured ALA. The Neuropathy Impairment Score was "numerically reduced," which is the authors' way of saying it did not reach significance.

Safety showed a dose-dependent rise in nausea, vomiting and vertigo. The conclusion — that 600 mg once daily "appears to provide the optimum risk-to-benefit ratio" — is one of the more clearly earned sentences in the supplement literature.

The long trials: what a symptom drug cannot do

ALADIN III ran 509 patients through three weeks of IV followed by six months of oral ALA at 600 mg three times daily. At seven months symptom scores did not differ from placebo to a clinically meaningful degree; the authors attributed part of that to rising between-centre variability in scoring. Neuropathic deficit (NIS) did improve at day 19 (P = 0.02) but only trended at seven months (P = 0.09).

NATHAN 1 is the definitive long trial: 460 patients with mild-to-moderate neuropathy, 600 mg oral ALA or placebo, four years. The primary composite endpoint — impairment score plus seven neurophysiological tests — did not differ between groups (P = 0.105). Secondary analyses were kinder: the NIS improved on ALA (P = 0.028), as did the lower-limb NIS (P = 0.05), and more patients showed clinically meaningful improvement while fewer progressed (P = 0.013). Nerve conduction did not worsen on placebo either, which meant the trial could not test whether ALA prevented deterioration — there was little to prevent.

Serious adverse events over four years were 38.1% on ALA versus 28.0% on placebo, without a clear pattern. Tolerability and discontinuation rates were equivalent.

The meta-analysis of the four intravenous trials (ALADIN, ALADIN III, SYDNEY and NATHAN II; 1,258 patients) gives the summary figure: a 24.1% relative symptom advantage for ALA over placebo after three weeks (95% CI 13.5–33.4), a 16.0% advantage on lower-limb impairment, and responder rates of 52.7% versus 36.9%. Benefit was first detectable at day 8.

What this series shows, taken together: oral 600 mg relieves the symptoms of diabetic neuropathy within weeks, slows the accumulation of clinical impairment over years, and does not restore nerve conduction or reverse the disease. That is a B for symptoms and an honest null for cure.

Glucose Disposal: What the Data Actually Are

The mechanistic claim — that ALA improves insulin-stimulated glucose uptake — comes from a specific study, and it is worth knowing how small it was.

In 1999, 74 patients with type 2 diabetes were randomised to placebo or oral ALA at 600 mg once, twice or three times daily for four weeks, with an isoglycaemic clamp before and after. Significantly more ALA-treated patients showed an increase in insulin-stimulated glucose disposal, and because there was no dose–response the three ALA arms were pooled: glucose disposal rose 27% relative to placebo (P < 0.01). The authors called it explorative and asked for confirmation. Twenty-five years later, the confirmation at scale has not arrived.

Two meta-analyses disagree, and the disagreement is instructive. The 2018 pool reported an HbA1c standardised mean difference of −1.22 — an effect larger than metformin's — from trials in "patients with metabolic diseases," a category loose enough to include studies that should not be pooled. The 2021 analysis, restricted to glycaemic markers and modelled for dose and duration, found a modest improvement in insulin and HOMA-IR and nothing on glucose or HbA1c. The insulin effect was duration-dependent and non-linear.

The number to plan around is the second one. ALA lowers a fasting-insulin-derived index by about half a point and does not move the marker a clinician uses to diagnose or manage diabetes. That is a C, and it is why ALA sits on the weaker arm of the fork in The Insulin Sensitivity Protocol, while berberine — which does move HbA1c — carries the load.

R-ALA, Racemic ALA and the Empty Stomach

The R-enantiomer is the form the body makes and binds; the S-enantiomer is a by-product of synthesis. The obvious question is whether R-ALA alone is a better supplement.

Pharmacokinetically, yes. In 12 healthy adults given 600 mg of R-lipoic acid as the sodium salt, plasma peak concentration and area under the curve were significantly higher, and time to peak shorter, than for plain R-ALA or racemic ALA — the salt is more soluble and less prone to polymerising in the gut. Three 600 mg doses at 15-minute intervals approached the exposure of a 20-minute intravenous infusion.

Clinically, there is no answer. Every efficacy trial above used racemic ALA at 600 mg. No trial shows that 300 mg of R-ALA reproduces SYDNEY 2. Buying R-ALA is buying better absorption and zero outcome data; the reasonable position is that 300 mg of R-ALA is a pharmacokinetic approximation of the trialled dose, not a proven equivalent.

Timing is settled: food reduces the bioavailability of both enantiomers, so take it 30 minutes before a meal.

The Insulin Autoimmune Syndrome Story

Insulin autoimmune syndrome — Hirata disease — is spontaneous hypoglycaemia in someone never given insulin, with extremely high measured insulin and high-titre antibodies against their own insulin. The antibodies bind insulin after a meal and release it unpredictably hours later, producing fasting and post-absorptive lows. It was described in Japan in 1970, is strongly associated with HLA-DRB1*04:06 in Japanese patients, and is typically triggered by sulphhydryl-containing drugs — classically methimazole.

ALA is a sulphhydryl compound. Japanese cases on ALA appeared from 2006. Then, in 2014, an Italian group reported six European patients — three men, three women, median age 63 — who developed episodes of fasting and post-absorptive hypoglycaemia with mainly neuroglycopenic symptoms 30 to 120 days after starting ALA at 600 mg/day. None had an insulinoma or insulin exposure. Stopping ALA reduced the episodes; all were treated with glucose and prednisone at 12.5–25 mg/day.

The HLA typing was the finding. Five carried HLA-DRB1*04:03, and one carried *04:06 — a different allele from the Japanese association, and one that is more common in Europeans. Two further European cases in 2019, both women, both *04:03, confirmed it. A 2024 review of 68 genotyped cases worldwide summarised the split: DR4 predominates everywhere; *04:06 with methimazole in East Asians, *04:03 with lipoic acid in Europeans.

Three practical conclusions:

• The syndrome is rare, and its true incidence is unknown because it is underdiagnosed — a supplement user with hypoglycaemia is usually told to eat more frequently.
• The onset window is one to four months after starting, not the first week. A clean first month is not reassurance.
• Any unexplained hypoglycaemia on ALA means stop it and measure insulin, C-peptide and insulin antibodies. Not a snack.

Anyone with a personal history of insulin autoimmune syndrome, or known HLA-DRB1*04:03 or *04:06 carriage, should not take ALA. Anyone on insulin or a sulfonylurea already has a hypoglycaemia risk that ALA compounds, and needs their prescriber involved before adding it.

Dosing and Interactions

600 mg racemic ALA once daily, 30 minutes before breakfast. That is the trialled dose for neuropathy and the highest dose with a favourable side-effect profile. There is no evidence that more helps for symptoms or for glucose, and both ALADIN and SYDNEY 2 show more nausea, vomiting and vertigo above it.

The thyroid interaction deserves a note because it is repeated widely. Searching for a human study finds a pig feeding trial and nothing else. It is not listed as an interaction on this site for that reason.

Who Should Take It

Someone with diabetes and symptomatic distal neuropathy — burning, stabbing, numbness in the feet — has the best case in the supplement literature for a five-week trial at 600 mg. They should also have an HbA1c that is being managed, because ALA treats the symptom and not the cause, and NATHAN 1 shows it does not reverse the disease.

Someone with prediabetes and no neuropathy is buying a half-point HOMA-IR improvement with no demonstrated HbA1c effect. That is honestly labelled as the C-grade fork of The Insulin Sensitivity Stack, for people who are not candidates for the inositol arm.

Anyone with prior unexplained hypoglycaemia, anyone on insulin or a sulfonylurea without prescriber oversight, and anyone pregnant should leave it alone.

Frequently Asked Questions

What dose of alpha-lipoic acid works for neuropathy?

Six hundred milligrams once daily by mouth. In SYDNEY 2 it cut symptom scores by 51% against 32% on placebo in five weeks. Doses of 1,200 and 1,800 mg were not better and produced more nausea, vomiting and vertigo.

Does alpha-lipoic acid lower blood sugar?

Modestly, on a fasting-insulin index. The 2021 meta-analysis of 28 trials found HOMA-IR fell by 0.48 and insulin fell slightly, with no change in glucose or HbA1c. The 1999 clamp pilot found a 27% rise in glucose disposal but has not been confirmed at scale.

Is R-lipoic acid better than regular ALA?

It is absorbed better — the sodium salt of R-ALA gave higher and faster plasma peaks than racemic ALA in healthy volunteers. But every efficacy trial used racemic ALA at 600 mg, and no trial shows R-ALA at a lower dose matches those results.

Can alpha-lipoic acid cause low blood sugar?

Yes, two ways. Additively, with insulin or sulfonylureas. And rarely, through insulin autoimmune syndrome — antibodies against your own insulin — which appeared one to four months after starting 600 mg/day in European patients carrying HLA-DRB1*04:03. Unexplained hypoglycaemia on ALA means stop and test.

Does it reverse diabetic neuropathy?

No. Over four years in NATHAN 1, 600 mg did not change the primary composite of impairment and nerve tests. It improved clinical impairment scores and fewer patients progressed. It slows and relieves; it does not restore.

Should I take it with food?

No — 30 minutes before a meal. Food reduces the bioavailability of both enantiomers.

Related Research

• Alpha-Lipoic Acid (ALA)
• The Insulin Sensitivity Protocol
• The Insulin Sensitivity Stack: Berberine + Inositol + ALA + Chromium
• Myo-Inositol: The Insulin Sensitiser for PCOS and Beyond
• NAC: The Glutathione Master Antioxidant