The largest randomized trial of injected NAD+ we found used 10 mg.

Why this note exists

NAD+ is one of the most widely stocked compounds in the research market, and the case for it is usually made from biology that is not in dispute: it is an essential molecule in cellular metabolism.[10] The question this note asks is narrower. What happens when people are actually given NAD+ itself, by vein or needle, and what has been measured?

In the index generated on September 17, 2026, 102 of 108 vendors carried a NAD+ listing, 189 listings in all, the third-widest vendor coverage of any compound pepmg tracks after BPC-157 and GHK-Cu.[12] Where a size is stated, 500 mg is the most common (79 listings), followed by 1,000 mg (48); three listings are nasal sprays.[12]

A note on what kind of evidence this is: everything below is human data unless labeled otherwise, and each study carries its design, size and funding in the same paragraph as its result. Two reports discussed here are preprints and have not been peer-reviewed. Doses are reported as the studies published them, and pepmg does not convert a trial dose into a protocol for a research vial.

The new evidence

A 180-patient heart-failure trial, and what it did and did not show

The largest randomized trial of directly administered NAD+ pepmg found was published online in September 2025 in the American Journal of Cardiovascular Drugs. It enrolled 180 adults at a single hospital in Hefei, China, with heart failure caused by ischemic cardiomyopathy, a left ventricular ejection fraction of 45% or less and NYHA class II–III symptoms. Patients were randomized 1:1 to "intravenous NAD⁺ (10 mg/day) or an equivalent placebo (5% glucose/normal saline) for a duration of 7 days, alongside guideline-directed medical therapy."[1]

The primary endpoint was the change in ejection fraction at one month. In the full text, ejection fraction rose from 40.96% to 45.44% in the NAD+ group and from 40.24% to 42.44% on placebo, a change of +4.49 against +2.25 points (p = 0.023).[2] The secondary outcomes all moved in the same direction without reaching significance: NT-proBNP at day 7 (p = 0.102), the six-month composite of major cardiac and cerebrovascular events, 14.6% against 24.7% (p = 0.089), and improvement in NYHA class at one month (p = 0.088) and six months (p = 0.115).[1] No differences were seen in heart structure.[1]

Four details belong next to that number. First, the trial was sized to detect a 3.7-point difference in ejection fraction; the observed difference between groups was about 2.2 points.[2] Second, ejection fraction was read by "a single accredited sonographer," and while the paper describes a "blinded randomization scheme developed by an independent statistician," it does not state whether patients, treating physicians or that sonographer knew who received NAD+.[2] Third, the study was supported in part by "a research grant and study drug supply from Kaifeng Knature Pharmaceutical Co. Ltd.," and the methods name the NAD+ product's supplier as "Kaifeng Kangnuo Pharmaceutical Co., Ltd," citing a Chinese drug approval number (H41024721); the authors' conflict statement declares no conflicts.[2] Fourth, the authors' own conclusion, that NAD+ "can enhance cardiac function," is stronger than a single-center trial with non-significant clinical outcomes can carry; they themselves call for "more extensive multicenter trials focused on clinical endpoints."[1]

Safety at this dose was unremarkable: one patient on NAD+ reported minor dizziness that resolved in about an hour, and one on placebo reported chest discomfort.[2]

An earlier and smaller trial points the same way, more weakly. In 2024, a Beijing group randomized 60 older heart-failure patients to "saline and 50 mg NAD+ with 50 mL of normal saline for 7 days." Improvement in NT-proBNP and ejection fraction was better on NAD+ "although not statistically significant"; the trial was registered retrospectively, in August 2023.[3] Its conclusion describes "a positive effect on improving cardiac function," which its own primary measures did not establish.[3]

The doses that are sold

At 500 mg and up, the human record is three small studies

The heart-failure trials used 10 mg and 50 mg a day. The 2026 clinic study describes four daily 500 mg infusions as "the typical commercial 'loading dose' recommended by the compounding pharmacy," and 500 mg is the most common research-vial size.[6][12] The human studies at those doses are small, and two of them were run by or with the maker of a competing product.

Enrollment, dose and design as published. A fifth, the 2019 pharmacology study, enrolled 11 men (8 NAD+, 3 saline) and is described below.[1][3][4][5][6]

The 2019 pharmacology study. Eight healthy men aged 30–55 received 750 mg of NAD+ intravenously over six hours, with three given saline. Plasma NAD+ did not change "until after 2 h" and was about 398% above baseline at six hours; the authors "observed no adverse events" during the infusion.[4] It measured blood and urine metabolites, not any clinical outcome. One author is a director of NAD+ Research Inc., and another received consulting fees from it.[4]

The 2024 preprint. This pilot analyzed 37 participants in its first part and 16 in its second, comparing 500 mg intravenous NAD+ with nicotinamide riboside (NR) and, in the first part, oral NR and saline. The IV groups were blinded; the oral group could not be. The report has not been peer-reviewed. It reported more infusion-related adverse experiences with NAD+ than with NR, increased white-cell and neutrophil counts in the NAD+ group, and no attributable adverse events during the 14-day follow-up. This was a small tolerability and biomarker study, not a trial of longevity or disease prevention.[5]

The 2026 clinic review. Published in Frontiers in Aging, this is a retrospective medical-record review at a commercial wellness clinic: six clients given 500 mg NAD+ IV daily for four days, eight given NR IV.[6] The paper reports that "[a]ll clients (n = 6) in the NAD+ group reported moderate to severe abdominal cramping, diarrhea, nausea, vomiting, increased heart rates, pain in throat, congestion, and chest pressure during infusion," which "ceased immediately upon infusion completion," and that NAD+ infusions averaged 97 minutes against 37 for NR because clients slowed them.[6] There was no placebo group, the authors acknowledge the study "was designed and executed by a commercial entity that employed the researchers," and the NR was donated by its manufacturer.[6]

These studies assess pharmacology and short-term tolerability, not clinical benefit. Infusion-related symptoms in small studies should not be generalized to every dose or route.[4][5][6]

A newer injection preprint changes the route picture. An April 2026 report includes a 45-person randomized pilot of NR, NAD+ and placebo given intravenously, intramuscularly or subcutaneously. Only 15 participants received NAD+: four intramuscularly, five intravenously and six subcutaneously. It reports no attributable unexpected adverse events or experiences. The study was funded by ChromaDex, is not peer-reviewed, and is too small to establish long-term safety or clinical benefit. Its second, 39-person study evaluated NR, not NAD+. Thus, limited subcutaneous NAD+ data do exist; an absence-of-studies claim would be inaccurate.[13]

Regulation

FDA said no in 2019. It has not finished saying it.

No FDA-approved drug contains NAD+, which is why it came before FDA as a candidate for the section 503A "bulks list": the list covers substances that are "neither the subject of an applicable United States Pharmacopeia (USP) or National Formulary (NF) monograph nor components of FDA-approved drugs."[7] FDA evaluated NAD for one use, fatigue in multiple sclerosis, and FDA's evaluation in a September 2019 proposed rule is blunt.[7]

On stability, NAD "degrades substantially when exposed to light, moisture, alkaline pH, or standard room temperatures and would not be stable under ordinary storage conditions absent multiple compensatory measures." On safety, "[n]onclinical data found in the literature are inadequate to characterize the potential toxicity profile for NAD," and FDA "did not find sufficient clinical data about NAD to evaluate whether it is safe for use in compounded drug products." On effectiveness, "[w]e identified no published studies that support the use of NAD for the treatment of fatigue in patients with multiple sclerosis."[7] The Pharmacy Compounding Advisory Committee had voted on May 8, 2017 not to include it, and FDA concluded that "[t]he proposed rule would not place NAD on the 503A Bulks List."[7] It reached the same conclusion for NADH.[7]

Seven years on, that is still a proposal. FDA's compounding page, current as of May 14, 2026, says of the September 2019 rule that "[a]fter considering public comments, the agency will issue a final regulation."[8] Until it does, NAD sits where it has sat: FDA's categories list, updated the same day, names "Nicotinamide Adenine Dinucleotide (NAD)" in Category 1.[9] For Category 1 substances, FDA says it "does not intend to take action against a compounder" provided the conditions in its interim guidance are met, and that they may stay there "until the agency decides on inclusion."[8]

That is the whole of the tolerance. Category 1 is an enforcement posture toward state-licensed pharmacies and physicians compounding for patients; it is not an approval, it is not a finding that NAD is safe or effective, and it says nothing about vials sold for research use.[8] FDA's 2019 stability language concerns the substance itself; pepmg makes no claim about the condition of any vendor's product.[7]

The precursors

Most NAD+ research is not about NAD+

The large body of human NAD+ work tests oral precursors, mainly nicotinamide riboside and nicotinamide mononucleotide, not NAD+ itself. A 2025 review in Nature Metabolism summarized it this way: "although preclinical studies support the idea that supplementation with NAD+ precursors is a promising therapeutic strategy to promote healthy ageing, human clinical trials have shown limited efficacy," and evidence for an age-related decline in human NAD+ "has been consistently observed only in a limited number of studies."[10] Animal results are a large part of the case for NAD+ biology; the review warns that "the extrapolation of rodent-based data is not straightforward."[10]

The precursor trial most likely to move the field has finished but not reported. NOPARK, a phase 3 randomized trial of oral nicotinamide riboside in early Parkinson's disease, enrolled 410 people and completed in June 2025; as of September 18, 2026 its registry record lists no results.[11] Whatever it shows will be a result for an oral precursor in Parkinson's disease. It will not be evidence about injected NAD+.

Three things this note is not saying

First, it is not saying NAD+ does nothing. A 180-patient randomized trial reported a statistically significant improvement in its primary endpoint, and a smaller one pointed the same direction; that is a real signal in heart failure that deserves the multicenter trials its authors call for.[1][3]

Second, it is not saying NAD+ infusions are dangerous. The adverse experiences reported at 500 mg were unpleasant and common in two small studies, and they stopped when the infusions stopped; beyond those infusion-day symptoms, the preprint reports "no attributable adverse events" through 14 days of follow-up and the clinic review "no documented adverse events."[5][6] What is missing is a controlled study large enough, and long enough, to say more.

Third, it is not saying FDA has banned NAD+. It proposed not to allow NAD in pharmacy compounding and has not finalized that proposal; in the meantime it has said it does not intend to act against compounders who meet its interim conditions.[7][8][9]

What it is saying is narrower: the best human trial of injected NAD+ is one single-center study in heart failure at 10 mg a day, and the doses that are actually sold have been studied in a few dozen people, for tolerability, not benefit.[1][4][5][6]

Questions people are asking