Alpha-lipoic acid vs Benfotiamine
The two supplements with real randomised programmes in diabetic peripheral neuropathy, different mechanisms, and the same ceiling: symptom scores move, the nerve does not.
Alpha-lipoic acid
Thioctic acid; NOT alpha-linolenic acid
Genuinely reduces diabetic neuropathy symptoms (licensed for it in Germany), but the 4-year NATHAN 1 trial missed its primary endpoint, weight loss is under a kilogram, and it can trigger autoimmune hypoglycaemia in HLA-DRB1*04:06 carriers.
Full Alpha-lipoic acid evidence reviewBenfotiamine
S-benzoylthiamine O-monophosphate, milgamma
Raises thiamine far better than thiamine HCl and activates the transketolase shunt exactly as advertised — but the 12-month BOND trial found no effect on nerve structure or function.
Full Benfotiamine evidence reviewSide by side
| Alpha-lipoic acid | Benfotiamine | |
|---|---|---|
| Marketed for | Nerve pain, blood sugar, weight loss, antioxidant | Diabetic neuropathy, nerve health, anti-glycation / AGE defence |
| Evidence grade | Clinical data | Clinical data |
| Family | NAD⁺ & mitochondrial | Metabolic / AMPK |
| Key human source | Ziegler 2004 meta-analysis, Diabet Med | BOND trial, BMJ Open Diab Res Care 2026 |
marks a row where the two differ. Evidence grades come from our evidence matrix, which grades each molecule on the human data for the use it is marketed for.
Our verdict
Alpha-lipoic acid has the deeper file, and it is still bounded. A meta-analysis of four placebo-controlled trials in 1,258 patients found 600 mg intravenously daily for three weeks improved the Total Symptom Score by 24.1% over placebo, and SYDNEY 2 showed oral 600 mg daily improved symptoms with higher doses buying nothing extra — but a 2023 meta-analysis of ten oral trials found no improvement in the visual analogue pain scale, vibration perception threshold or nerve conduction, and the four-year NATHAN 1 trial in 460 patients missed its primary composite endpoint (P = 0.105) with more serious adverse events on drug than on placebo. Benfotiamine's pharmacology is the best-proven part of it, delivering roughly elevenfold the plasma thiamine exposure of thiamine hydrochloride, and the 12-month BOND trial confirmed that end of the chain by raising all six thiamine analytes it measured — yet corneal nerve fibre length did not differ from placebo, and neither did any morphometric, neurophysiological, clinical or quality-of-life outcome, with only the Neuropathy Symptom Score even trending (p = 0.098). Both are licensed medicines in Germany and unapproved supplements in the United States, they have never been compared with each other, and two different mechanisms have now hit the same wall.
Alpha-lipoic acid fits if
Alpha-lipoic acid is the better-supported of the two for symptomatic diabetic polyneuropathy at 600 mg a day, provided the insulin-autoimmune-syndrome signal in HLA-DRB1*04:06 carriers is treated as a real contraindication rather than a footnote.
Benfotiamine fits if
Benfotiamine makes sense mainly for correcting a documented thiamine-pathway deficit in someone with diabetes, since raising thiamine is the one thing it has unambiguously been shown to do.