Alpha-Lipoic Acid

What it is

Alpha-lipoic acid (ALA, also called thioctic acid) is an antioxidant widely used as a supplement [1][2]. Trials have used oral and intravenous forms [2][3].

How it might work

Proposed mechanism: ALA may help nerves by improving nerve cell metabolism and antioxidant defences [4]. In one small trial, ALA raised superoxide dismutase and lowered malondialdehyde, two markers of antioxidant activity [4].

What the research says

Diabetic nerve pain: a 2024 Cochrane analysis of three placebo-controlled trials (816 people) found ALA probably had little or no effect on symptom scores at six months, with high dropout risk in all three trials [1]. A 2023 analysis of ten trials (1,242 patients) found lower sensory symptom scores than placebo, with a dose-related trend, but no benefit on vibration sensing or nerve conduction [5]. A 2022 analysis reported lower symptom scores than placebo [2]. A 2026 analysis of 15 studies found 19 of 23 outcomes favoured ALA, while HbA1c and nerve conduction did not change [6]. An older 2012 analysis of 15 intravenous trials reported faster nerve conduction, but its authors said the trials were of poor quality [7]. A 2024 network analysis found ALA and gamma-linolenic acid both lowered symptom scores against placebo, with high heterogeneity between trials (I-squared 98%) [8]. Fibromyalgia pain: one small crossover trial found no significant difference from placebo [9].

Weight: several meta-analyses found about 1.3 kg [10], 0.7 kg [11] and 2.3 kg [12] more weight loss than placebo. A 360-person, 20-week trial found 1,800 mg a day led to a 2.1% greater loss than placebo [13]. A 24-week trial in 81 overweight adults with raised triglycerides found greater BMI loss with (R)-ALA, although triglycerides did not fall [14]. A crossover trial of 103 adults gave conflicting results, with mild weight loss reported alongside a statement that no real weight reduction was seen [15].

Blood sugar: results conflict. A 2018 analysis of 24 trials in people with metabolic disease found lower fasting glucose and HbA1c [16]. A 2021 analysis of 28 trials found no change in glucose or HbA1c, although insulin and insulin resistance fell [17]. A 2019 review found lower fasting blood sugar and HbA1c, with no change in insulin resistance [18]. A 2025 analysis in overweight adults found no significant change in intermediate disease markers, including lipids and glucose [19].

Doses used in studies

Oral doses in trials ranged from 400 to 1,800 mg a day [20][21][13]. The 2026 analysis found 600 mg a day was particularly effective for nerve symptoms [6]. Intravenous trials used 600 mg a day for about three weeks [3]. One nerve-pain trial ran for six months at 600 mg twice a day [22]. Weight trials ran from eight to 24 weeks, with 20 weeks in the largest [15][14][13].

Safety and interactions

Side effects: most were mild. In a 236-person trial, reported side effects were 25.4% with 1,800 mg a day versus 11.8% with placebo, mostly a burning feeling from the throat to the stomach [21]. In a 360-person trial, itching and hives were most common and generally mild [13]. Most side effects in pooled trials were gastrointestinal, with no difference in dropouts from placebo [8].

Upper limit: no official limit was found. The authors of one review describe doses up to 1,200 mg a day as safe [11].

Medicine interactions: none of the trials we reviewed tested interactions, so this is unknown. Because some trials lowered blood sugar [16], people on diabetes medicines should check with a doctor first.

Pregnancy and breastfeeding: no data were found in our searches.

Kidney and liver disease: no data were found in our searches on people with these conditions.

G6PD deficiency: a review found no evidence of harm from ALA in people with this condition [23].

Caveats and open questions

The 2024 Cochrane review rated all three trials at high risk of dropout bias [1]. Pooled results for nerve symptoms vary widely between studies (I-squared 98% in one analysis) [8]. Trials vary in dose, form and length, and results differ by dose [5].