NADH

Chronic fatigue syndrome (ME/CFS)

The Vall d'Hebron hospital group in Barcelona (Jesus Castro-Marrero and Jose Alegre) ran most of these trials (PMID: 20447621, PMID: 34444817); the first came from Georgetown University working with the Birkmayer group (PMID: 10071523).

• Forsyth 1999, randomized double-blind crossover, 26 patients, NADH 10 mg a day or placebo for 4 weeks each with a 4-week washout. Eight of 26 patients (31 percent) responded favorably to NADH, against 2 of 26 (8 percent) on placebo (PMID: 10071523).
• Santaella 2004, randomized, 31 patients, NADH versus nutritional supplements plus psychological therapy for 24 months, no placebo. The 12 patients given NADH had a large drop in symptom scores in the first three months (p<0.001), after which scores were similar in both groups (PMID: 15377055).
• Alegre 2010, randomized double-blind, 86 enrolled (77 completed), NADH 20 mg a day or placebo for 2 months. Most measures did not differ between groups. Anxiety scores fell by 1.0 point with NADH (p<0.05), and peak heart rate after an exercise test fell by 8.1 beats per minute within the NADH group, while patients and investigators rated the two treatments as equally effective (PMID: 20447621).
• Castro-Marrero 2015 and 2016, randomized double-blind, NADH 20 mg with coenzyme Q10 200 mg a day or placebo for 8 weeks. Both reports cite one registration (NCT02063126), with 73 and 80 patients. The registered primary outcome was peak heart rate during a cycle test; the 2016 report describes a fall from baseline within the active group (p=0.022) rather than a difference from placebo, while fatigue, a secondary outcome, fell more than with placebo (p=0.03) and pain and sleep did not improve (PMID: 26212172). The 2015 report found a lower fatigue score than placebo (p<0.05) and, in blood mononuclear cells, a higher NAD+/NADH ratio, ATP, coenzyme Q10 and citrate synthase and lower lipid peroxides (PMID: 25386668).
• Castro-Marrero 2021, randomized double-blind, 207 women, NADH 20 mg with coenzyme Q10 200 mg a day or placebo. The primary endpoint was the change in the FIS-40 fatigue score, and the authors report improvements within the active group but no difference between groups, so the primary endpoint was not met (PMID: 34444817). The tablets were made and donated by a supplement company (PMID: 34444817).
• Systematic review 2025. Fourteen ME/CFS supplement studies with 809 participants carried a high risk of bias; NADH and NADH with coenzyme Q10 were among the supplements reporting lower fatigue, but small samples and missing data prevented firm conclusions (PMID: 39940333).

Parkinson's disease

• Birkmayer 1993, open-label, 885 patients, NADH by intravenous infusion or oral capsules. The Birkmayer Institute for Parkinson Therapy in Vienna reported improvement in disability in about 80 percent of patients, with oral NADH comparable to intravenous NADH. There was no control group (PMID: 8101414).
• Dizdar 1994, randomized double-blind, 5 patients given NADH and 4 given saline, intravenous for 4 days and then intramuscular at 2 and 4 weeks. Disability scores did not change significantly, and cerebrospinal fluid markers did not change (PMID: 7887134).
• Kuhn 1996, open-label, 15 patients, intravenous NADH daily for 7 days. Parkinson's rating scale scores improved (p=0.025), and plasma levodopa levels rose (PMID: 9013405).
• Review 1998. A University of Virginia review listed the support for NADH as claims from cell culture and human studies plus case series, noted that a placebo-controlled trial failed to show clear benefit, and concluded that recommendations for NADH in Parkinson's disease were premature (PMID: 9805207).

Dementia

• Birkmayer 1996, open-label, 17 people with Alzheimer-type dementia, 8 to 12 weeks. Mini-Mental State scores improved by a mean of 8.35 points; the author described the study as a pilot from which no definitive conclusion could be drawn (PMID: 8834355).
• Rainer 2000, open-label, 25 people with Alzheimer, vascular or frontotemporal dementia, NADH 10 mg a day for 3 months added to cholinergic drugs. In the 19 who completed the study, no cognitive effect was found (PMID: 11459000).
• Demarin 2004, randomized double-blind, 26 people with probable Alzheimer's disease, NADH 10 mg a day or placebo for 6 months. In 12 matched pairs, total Mattis Dementia Rating Scale scores were higher with NADH (p<0.05), with better verbal fluency and visual-constructional scores, but attention, memory and clinician-rated dementia severity did not differ (PMID: 15134388).

Jet lag

• Birkmayer 2002, randomized double-blind, 35 healthy employed adults after an overnight west-to-east flight with a 3-hour time difference, sublingual NADH 20 mg or placebo. The NADH group did better on 4 cognitive test measures and reported less sleepiness than the placebo group (PMID: 12385067). The study came from the Birkmayer group in Vienna.

Animal and cell work

• Cells. NADH raised dopamine production up to threefold in rat PC12 cells (PMID: 9247090).
• Animals. Swallowed NADH did not raise urinary nicotinamide metabolites in mice (PMID: 16802695), while oral or injected NADH raised cortical NADH fluorescence in rats (PMID: 12076318). A 26-week rat study found no treatment-related toxicity (PMID: 12635493).

What the trials recorded

In a 26-patient chronic fatigue syndrome crossover trial, no severe adverse effects related to NADH were observed (PMID: 10071523). In a 35-person jet lag trial of sublingual NADH, no adverse effects were observed (PMID: 12385067). In an open-label study of 17 people with Alzheimer-type dementia followed for 8 to 12 weeks, no side effects were reported (PMID: 8834355).

For NADH combined with coenzyme Q10, an 8-week trial in 80 patients described the combination as generally safe and well tolerated (PMID: 26212172). In the 207-patient trial, 8 people in the active group left the study because of adverse effects (5 with upper abdominal pain and 3 with dizziness) against 11 in the placebo group, and no serious adverse events were reported (PMID: 34444817).

A 2023 systematic review of 10 randomized trials with 489 participants, covering NADH and NAD precursors together, listed muscle pain, nervous system complaints, fatigue, sleep disturbance and headache among reported side effects and judged none a serious risk (PMID: 37971292). A 2025 systematic review of dietary supplements in ME/CFS, which included NADH trials, noted that six studies reported adverse effects such as nausea and insomnia across the supplements reviewed (PMID: 39940333).

Animal toxicology

In a 26-week study from the Birkmayer group in Vienna, which makes a stabilized NADH product, 80 rats were split between daily NADH tablets and placebo (PMID: 12635493). NADH caused no deaths or treatment-related clinical signs and no differences in blood tests or organ weights, and tissue findings were similar to controls; treated females gained more weight and ate up to 10 percent more food (PMID: 12635493).

Gaps in the safety record

The controlled NADH trials summarized here lasted from one day to 6 months, apart from one 24-month comparison without placebo (PMID: 15377055). No human safety data in pregnancy or breastfeeding were found; the 207-patient trial excluded pregnant and breastfeeding women (PMID: 34444817).