Are NMN and NR Safe? What the Human Trials Recorded

The human NMN and NR trials recorded no serious side effects, but they ran for weeks and enrolled small groups. Here is what each trial reported about side effects, labs, and bedside measures, and what the records cannot show.

Jump to: - What the trials recorded - The highest dose we found - What the records cannot show - Who should ask a clinician first - Bottom line - Frequently asked questions - Next reads Are NMN and NR safe? In the human trials we checked, no trial that reported serious adverse events recorded one. That is a record of what participants reported, not a finding that either ingredient is safe for everyone. The trials are small and short, and several were written by people with ties to the companies that make the ingredients. Most enrolled fewer than 100 people, the largest enrolled 140, and the longest run with safety data lasted 24 weeks. Adverse events are any new symptom or problem a participant reported, whether or not the ingredient caused it. Labs are blood and urine values, such as blood counts, liver and kidney markers, and blood sugar. Bedside measures are blood pressure, heart rate, temperature, and oxygen saturation. ## What the trials recorded Our [NMN monograph](/ingredients/nmn-nicotinamide-mononucleotide) and [NR monograph](/ingredients/nr-nicotinamide-riboside) carry the wider evidence. The table covers the ten trials we checked that report safety data. | Trial | Who, and how many | Dose and duration | What the record reports | |---|---|---|---| | [Irie 2019](https://pubmed.ncbi.nlm.nih.gov/31685720/) | 10 healthy men, single-arm | One dose each of 100, 250, or 500 mg NMN, watched 5 hours | No significant symptoms; heart rate, blood pressure, oxygen saturation, and temperature unchanged; bilirubin rose, creatinine, chloride, and glucose fell, all within normal ranges | | [Okabe 2022](https://pubmed.ncbi.nlm.nih.gov/35479740/) | 30 healthy adults, 15 NMN, 15 placebo | 250 mg NMN a day, 12 weeks | No abnormal lab or physiological results; no obvious adverse effects | | [Igarashi 2022](https://pubmed.ncbi.nlm.nih.gov/35927255/) | 42 randomized men aged 65 or older | 250 mg NMN a day, 12 weeks | No serious adverse event; blood counts, blood chemistry, and liver and kidney markers did not differ from placebo | | [Yi 2022](https://pubmed.ncbi.nlm.nih.gov/36482258/) | 80 middle-aged healthy adults; placebo, 300, 600, or 900 mg NMN | Once daily, 60 days | No safety issues across adverse-event, lab, and clinical monitoring | | [Shiraki 2026](https://pubmed.ncbi.nlm.nih.gov/42082179/) | 7 men per group; type 2 diabetes, frailty, aged 65 or older | 250 mg NMN a day, 24 weeks | No systemic or eye adverse events in the NMN group | | [Martens 2018](https://pubmed.ncbi.nlm.nih.gov/29599478/) | 30 enrolled, 24 completed; healthy adults aged 55 to 79 | 1,000 mg NR a day, two 6-week crossover periods | 14 adverse events in 7 people, none serious; 2 dropped out for side effects, both on placebo; blood pressure change not significant | | [Conze 2019](https://pubmed.ncbi.nlm.nih.gov/31278280/) | 140 enrolled; healthy adults aged 40 to 60, BMI 25 to 30 | 100, 300, or 1,000 mg NR a day, 8 weeks | 95 adverse events in 61 people, none serious; no flushing; no difference in blood pressure or heart rate between groups; LDL cholesterol and homocysteine unchanged | | [Dellinger 2017](https://pubmed.ncbi.nlm.nih.gov/29184669/) | 120 healthy adults aged 60 to 80 | NR plus pterostilbene (a combination), standard or double dose, 8 weeks | No serious adverse events; NR milligrams not stated in the abstract | | [Berven 2023](https://pubmed.ncbi.nlm.nih.gov/38016950/) | 20 people with Parkinson's; 10 NR, 10 placebo | 1,500 mg NR twice a day (3,000 mg a day), 4 weeks | No moderate or severe adverse events; no significant difference in mild ones; slight early rise in serum homocysteine | | [Martens 2026](https://pubmed.ncbi.nlm.nih.gov/42478598/) | 52 randomized with amnestic mild cognitive impairment; 42 completed | 500 mg NR twice a day, 12 weeks | No serious adverse events; nausea, diarrhea, and headache most common; 4 stopped for adverse events (3 placebo, 1 NR, for a low kidney filtration value) | ## The highest dose we found went to people with Parkinson's The highest dose we found was 1,500 mg of NR twice a day, or 3,000 mg a day, given for four weeks to 20 people with Parkinson's disease ([Berven and colleagues, 2023](https://pubmed.ncbi.nlm.nih.gov/38016950/)). The authors wrote that doses above 2,000 mg a day had not been tested in humans before. The abstract reports no moderate or severe adverse events and a slight early rise in serum homocysteine, which the authors said did not disturb methyl-group balance. A 30-day trial in newly diagnosed Parkinson's patients who had not started treatment called 1,000 mg of NR well tolerated and found lower inflammatory cytokines in blood and spinal fluid ([Brakedal and colleagues, 2022](https://pubmed.ncbi.nlm.nih.gov/35235774/)). Two small trials do not make a dose safe for anyone with the disease. ## What the records cannot show Time is the biggest gap. The longest NR trial we found ran six months in 90 people with peripheral artery disease ([McDermott and colleagues, 2024](https://pubmed.ncbi.nlm.nih.gov/38871717/)), but its abstract reports no adverse-event data. The longest run with safety data is the 24-week NMN trial in men with type 2 diabetes ([Shiraki and colleagues, 2026](https://pubmed.ncbi.nlm.nih.gov/42082179/)). The [longevity pillar](/health/nad-longevity-complete-guide) covers the wider evidence. Several abstracts say nothing about adverse events. They include a 40-person NR trial in stable COPD ([Norheim and colleagues, 2024](https://pubmed.ncbi.nlm.nih.gov/39548320/)), a 25-person kidney disease trial comparing NR with coenzyme Q10 ([Ahmadi and colleagues, 2025](https://pubmed.ncbi.nlm.nih.gov/39847432/)), and a 14-day trial of NR, NMN, and nicotinamide in 65 healthy adults ([Christen and colleagues, 2026](https://pubmed.ncbi.nlm.nih.gov/41540253/)). Several safety sections also come from company-affiliated authors. Conze and colleagues list ChromaDex affiliations, Dellinger and colleagues list Elysium Health authors, and Okabe and colleagues and Igarashi and colleagues list Mitsubishi Corporation Life Sciences authors, a company that supplied the NMN in the Igarashi trial. Read those sections as company-reported records, and check the paper before you repeat a number. The Igarashi trial also lost data to a supplier error, which led the authors to exclude 22 participants' 12-week results. No trial we checked was built to find the dose where side effects begin, so the records give tested doses from 100 mg to 3,000 mg but no ceiling. ## Who should ask a clinician first These are the trials' own exclusions, taken from the full texts. They are not a list of who should never take either ingredient. - Conze and colleagues (2019) excluded people with diabetes, people on blood-pressure or cholesterol medicines, and people with kidney or liver disease. - Igarashi and colleagues (2022) excluded men with a history of cancer treatment, heart failure, or heart attack, and men taking a prescription medicine that could affect the results. - Martens and colleagues (2026) excluded people with abnormal liver enzymes, an eGFR below 30, Parkinson's disease, epilepsy, multiple sclerosis, or medicines that affect cognition. - Martens and colleagues (2018) excluded people with abnormal kidney or liver blood chemistry, and people with alcohol dependence. If any of this describes you, ask a clinician before you start. That includes anyone on a prescription medicine, anyone with kidney, liver, or heart disease, and anyone who is pregnant or nursing. None of the abstracts we checked covers pregnancy or breastfeeding. Bring the milligrams per day you plan to take. ## Bottom line Across the trials that reported serious events, none recorded one. The longest run with safety data was 24 weeks, and none of the trials we checked reports an interaction test with a prescription drug. If you take a prescription medicine, have kidney, liver, or heart disease, or are pregnant or nursing, ask a clinician before the first dose. Most trials dosed in milligrams per day, so start with the milligram figure on the label. ## Frequently asked questions ### Are NMN and NR safe? No one can call either ingredient safe for everyone from these records. In the human trials we checked, none that reported serious adverse events recorded one. The longest run with safety data was 24 weeks, and most trials enrolled fewer than 100 people. Ask a clinician before you start, especially if you take a prescription medicine. ### Do NMN and NR have side effects? Some participants reported side effects, and placebo groups reported them too. Conze and colleagues (2019) recorded 95 adverse events in 61 of the 140 people enrolled, across doses of 100 to 1,000 mg of NR a day for eight weeks. None were serious, and event types and rates were similar across groups ([PubMed](https://pubmed.ncbi.nlm.nih.gov/31278280/)). Okabe and colleagues (2022) reported no obvious adverse effects with 250 mg of NMN a day for 12 weeks ([PubMed](https://pubmed.ncbi.nlm.nih.gov/35479740/)). ### What side effects did NR trials record? The most common were nausea, diarrhea, and headache, in a 12-week trial of people with amnestic mild cognitive impairment ([Martens and colleagues, 2026](https://pubmed.ncbi.nlm.nih.gov/42478598/)). One participant stopped after a low estimated kidney filtration value on a lab test. In the Parkinson's trial, NR produced a slight early rise in serum homocysteine and no moderate or severe adverse events ([Berven and colleagues, 2023](https://pubmed.ncbi.nlm.nih.gov/38016950/)). ### Is NMN safe over the long term? The records cannot answer that. The longest NMN run with safety data lasted 24 weeks, in seven men per group with type 2 diabetes and frailty, with no systemic or eye adverse events in the NMN group ([Shiraki and colleagues, 2026](https://pubmed.ncbi.nlm.nih.gov/42082179/)). A 12-week trial of 42 randomized older men found no serious events and no lab differences from placebo ([Igarashi and colleagues, 2022](https://pubmed.ncbi.nlm.nih.gov/35927255/)). No trial we checked tracked safety for a year or more. ### Who should not take NMN or NR? The trials do not say who should never take either ingredient. They do show who was kept out. Conze and colleagues (2019) excluded people with diabetes, people on blood-pressure or cholesterol medicines, and people with kidney or liver disease. Igarashi and colleagues (2022) excluded men with a history of cancer treatment, heart failure, or heart attack. Ask a clinician if any of that describes you, or if you are pregnant or nursing, since none of the abstracts we checked covers either. ### Do NMN and NR interact with medicines? The trials we checked did not test a drug interaction. Conze and colleagues (2019) excluded people taking blood-pressure or cholesterol medicines, which keeps them out of the record rather than showing what happens when the drugs combine. The NR-SAFE Parkinson's trial tied one result to the time since the last levodopa dose but reported no interaction test ([Berven and colleagues, 2023](https://pubmed.ncbi.nlm.nih.gov/38016950/)). Tell your prescriber before you start. ## Next reads - [The NR monograph](/ingredients/nr-nicotinamide-riboside): the dosing and safety sections for NR, in full. - [NMN and NR doses against the trials](/health/nad-longevity-complete-guide/nmn-nr-doses-vs-the-trials): every human dose we checked, set against label servings. - [NMN vs NR: which reigns supreme](/blog/aging-gracefully/nad-precursors-nmn-vs-nr-which-reigns-supreme): how the two precursors differ in effect and mechanism.