Does oral single-ingredient NR reduce intrahepatic fat in adults with metabolic dysfunction-related fatty liver?

Useful facts when choosing a product

• Trial doses are not personal dosing or treatment-change instructions.
• Oral single NR differs from NR plus pterostilbene or multiple metabolic activators.
• Trial-specific salt/active mass is recorded only where explicitly verified.
• NAD, liver enzymes, muscle mass and body weight are not intrahepatic-fat values.

What the research actually shows

# Does oral single-ingredient NR reduce intrahepatic fat in adults with metabolic dysfunction-related fatty liver?

TASK-1023 / R01-063 · NUT/R01 · Evidence cutoff 2026-09-17

## Thirty-second answer

**The current verdict is D, 28 points; safety is Caution. A reduction in intrahepatic fat from oral single-ingredient nicotinamide riboside (NR) has not been established for adults with metabolic dysfunction-related fatty liver.** The closest trial enrolled 40 obese men rather than a cohort selected using a fatty-liver diagnosis. Over 12 weeks, hepatic fat declined by 2.0 percentage points with NR and 0.2 points with placebo, but the between-group test was not significant (P=0.13). The reported 18% relative reduction describes change within the NR arm, not an 18% benefit over placebo. [S01]

A separate crossover trial had 13 completers with overweight/obesity and low mean intrahepatic fat. Both condition means rounded to 3.4% (P=0.85). Neither finding establishes **an exactly zero effect, equivalence or exclusion of benefit**. Fatty-liver trials combining NR with pterostilbene or several metabolic activators were not attributed to NR alone. [S02, S04, S05]

This is a completed initial content submission. Site ID, slug, URLs and first-publication date remain unassigned, not because the clinical decision is deferred. Document quality A refers to the declared same-author self-check, not independent external review, journal certification or a guarantee of no errors.

## The question and the actual evidence

The question is whether single-ingredient oral NR reduces intrahepatic fat compared with an appropriate control in adults with metabolic dysfunction-related fatty liver. Proposed diagnosis, formulation and comparison fields were not silently turned into study facts. Imaging/spectroscopy fat, liver enzymes, blood or muscle NAD, body weight, histology, inflammation/fibrosis and clinical events are different endpoints.

NR, nicotinic acid, nicotinamide, NMN, NAD+ and NRH are adjacent search terms, not interchangeable interventions. Habitual-lifestyle conditions in an NR-only trial differ from co-administration of pterostilbene or L-serine, NAC and L-carnitine tartrate. NIAGEN branding alone was not treated as verification of every trial batch's salt, purity or free-NR mass. The doses below are reported research doses, not personal dosing or treatment-change instructions. [S01, S02, S04, S05]

NAFLD, MAFLD and MASLD labels are retained according to actual study entry criteria. Diagnostic, alcohol and participant-level information was insufficient to retrospectively classify every Dollerup or Remie participant as having current MASLD. An elevated group mean liver-fat value or obesity alone does not document the same diagnosis in every participant. [S01, S02]

## Main human evidence

A percentage point (pp) is an absolute difference in fat percentage, not a relative percent change. Lower values are favorable. The errors below are SEM/SE as specified, not SD.

| Study and actual population | Intervention, control and common conditions | Liver-fat method, timing and denominator | Verified finding and interpretation | |---|---|---|---| | Dollerup 2018: randomized double-blind parallel trial; 40 healthy sedentary obese men aged40–70, BMI>30, no prescription medicines; not enrolled by confirmed fatty-liver diagnosis | Oral NR1000mg twice daily, total2000mg/day, for12weeks versus matching placebo;20 per arm. Habitual diet/activity maintained; no assigned weight-loss/exercise program | 1.5T proton MRS; LWR/(1+LWR) expressed as fat percentage; baseline and12weeks. Baseline liver n20 NR/n19 placebo; Figure4 time-specific n19–20/n18–19. Exact complete-pair liver n unverified | Baseline NR11.3±1.8%, placebo14.1±1.9% (mean±SEM). Within-arm change NR−2.0pp (relative−18%), placebo−0.2pp (relative−1.4%); treatment-by-visit mixed-model P=0.13. The −1.8pp difference of rounded changes is our arithmetic, not a reported exact model coefficient or CI. [S01] | | Remie 2020: randomized double-blind crossover;15 recruited,13 completed,7 women/6 men;45–65years, mean59±5years and BMI30.2±2.6 (SD); no active disease | Oral NR1000mg/day versus placebo,6weeks per condition,4–7week washout. Habitual diet/activity maintained. Thirteen people crossed over, not26 independent participants | 3.0T proton MRS; T2-corrected CH2/unsuppressed water peak×100; day36 of each condition at17:00 after ≥3h fast. Default paper n13, with no IHL-specific exception stated | After-condition NR3.4±1.2%, placebo3.4±1.3% (mean±SE); printed paired difference0.0±0.0pp, P=0.85. These are rounded printed numbers; exact paired difference/SE/two-sided CI unverified. Indirect evidence from a low-liver-fat population. [S02] |

The MRS percentages differ even in denominator: lipid/(lipid+water) in Dollerup versus the CH2 lipid-peak/water-peak ratio in Remie. They were not relabeled as identical MRI-PDFF measurements, mechanically transformed using group means, or combined into a single percentage of fatty-liver treatment success. [S01, S02]

### Dollerup: diagnosis, analysis and the favorable signal

Baseline BMI was32.4±0.5 with NR and33.3±0.6kg/m² with placebo; HbA1c was5.6±0.1% and5.8±0.1%, respectively (SEM). Participants had metabolic risk, but this was not a treated-diabetes or diagnostically defined MASLD cohort. A quantitative long-term alcohol history and full diagnostic exclusion of alternative fatty-liver causes were not verified. Avoiding alcohol for72hours before clamp visits is not proof of a long-term nonalcoholic diagnosis. [S01, Table1/Methods]

All40 participants completed the study. Two technically incomplete final clamp examinations were not automatically counted as liver-fat dropouts. Hospital pharmacy staff handled block-of-four randomization, packaging and blinding. Participants/data collectors were masked; codes were released after study completion. The analysis used all available measurements in a repeated-measurement mixed model and also checked complete data by repeated-measurement ANOVA. The primary/sample-size endpoint was the clamp M-value, not hepatic fat; the paper acknowledges possible underpowering for hepatic fat. The NCT02303483 identifier and prerecruitment registration statement were verified in the paper, but the original registry outcome list/history could not be separately verified. [S01, S06]

Investigators described apparent reductions among9/13 NR and6/15 placebo participants whose baseline HLC exceeded5%. This was exploratory post hoc inspection, not a prespecified clinically meaningful responder threshold or a confirmatory between-arm test. No RR, ARR, NNT or MASLD treatment-response rate was created from it. The whole-cohort P=0.13, multiple metabolic measurements and missing between-arm CI remain part of the interpretation. [S01, HLC results/Figure4]

**The counterargument is preserved:** the point estimate favors NR, and the NR group already had2.8pp lower hepatic fat at baseline. A small trial not powered for hepatic fat cannot rule out clinically important benefit solely because its test is nonsignificant. Conversely, the favorable within-arm change alone cannot establish efficacy. [S01, Discussion]

### Remie: low liver fat and a rounded zero

Entry criteria included BMI27–35kg/m², exercise below3hours/week, weight stability for at least6months and alcohol no more than2servings/day. That alcohol limit was not converted to invented grams/day. Participants used no medications or supplements interfering with study outcomes and were asked to retain their usual diet/activity. The body-weight comparison P=0.55 is a different endpoint from liver-fat P=0.85. [S02, Methods/Results]

The sample-size calculation concerned insulin-stimulated skeletal-muscle glucose disposal. The hepatic-fat finding was not a successful fatty-liver primary endpoint. The paper generally reports n13 and mean±SE, with paired t tests for parametric and Wilcoxon tests for nonparametric data. The specific liver-test branch and registry endpoint history were not independently established. The Discussion explicitly states that many outcomes were tested without multiple-comparison adjustment. The hepatic-fat P=0.85 is not an adjusted significant finding. [S02, S07]

The trial did not select people with high baseline liver fat and therefore cannot substitute directly for a trial in confirmed fatty liver. Without the unrounded paired coefficient and SE, the printed0.0±0.0 is not an exactly zero effect, CI[0,0] or equivalence result. Marginal condition SEs were not used as though the crossover comprised independent groups. [S02]

## Why other evidence was not attributed to single NR

| Source | Verified scope | Handling in this verdict | |---|---|---| | Lapatto 2023 twin study | Reused study design/source extraction from3135: main16pairs/32people all receiving NR; separate4pairs/8people placebo comparison;250→1000mg/day escalation and5months. The final paper's liver table was inaccessible in current attempts | The main before/after change is not an NR–placebo contrast. An indexed “unchanged” lead or preprint snippet was not promoted to verified final numerical data. No new quantitative hepatic-fat estimate or grading weight was added. [S03] | | Dellinger 2023 NAFLD | MRI-PDFF HFF≥15%, age18–70, BMI25–39.9; NR250+PT50mg/day or NR500+PT100mg/day versus placebo for26weeks. Randomized111, mITT94 (placebo31/low36/high27); change comparisons with placebo P=0.4119/0.6396 | Combination treatment without an NR-only arm. Whole-cohort HFF change was not significantly better than placebo. Within-arm enzyme reductions and post hoc HFF strata are not NR-only liver-fat efficacy. [S04] | | Zeybel 2021 NAFLD | BMI>27, liver fat>5.5%;31 randomized (CMA20/placebo11),30 completed; NR1g+L-serine12.35g+NAC2.55g+L-carnitine tartrate3.73g orally once daily for14days, then twice daily for56days, versus placebo; MRI-PDFF outcome | Four-component metabolic activators, with no arm isolating NR. Reported within-CMA liver-fat reduction was not assigned to NR. Within-arm significance versus nonsignificance in placebo is not a between-arm test. [S05] | | Lapatto 2026,49 BMI-discordant twin pairs | Observational phenotyping, transcriptomics and epigenetics | Not an assigned NR hepatic-fat intervention. [S09] | | NCT06044935 | Official indexed title concerns ketosis, fat oxidation and metabolic rate; only a registry shell was readable | Current recruitment, dose, diet detail and liver-fat results unverified; third-party plans were not filled in as trial results. [S08] |

Some NRPT Table2 change rows do not equal arithmetic differences of rounded timepoint means; the original change rows were retained, without inventing an explanation or correction.

The NRPT study permitted diabetes/metformin and required stable statin dosing, while excluding NASH, cirrhosis and other major liver causes. Its red-wine restriction controlled overlapping pterostilbene/resveratrol exposure and was not treated as the general diagnostic alcohol threshold. The CMA trial's antidiabetic-drug exclusion and body-weight adjustment were different conditions. An actual NAFLD cohort exposed to a different combination does not establish a causal effect of single NR. Study-specific details remain separate in study_extraction.json. [S04, S05]

## Nutrition, formulation and safety

| Field | Verified value or reason for uncertainty | |---|---| | Baseline nutrition/deficiency | Neither quantitative single-NR trial was established as treatment of a confirmed B3/NR-deficiency cohort. Missing deficiency assays/thresholds do not prove nutritional sufficiency in every participant. [S01, S02] | | Total diet and added intake | Dollerup repeated a3-day pre-examination food record and restricted other vitamins/supplements; Remie maintained habitual diet. Numerical total dietary B3 intake was not verified for these hepatic-fat analyses. Added reported NR was2000/1000mg/day, not simply summed with other molecular B3 forms. [S01, S02] | | Salt, active amount and coingredients | Oral capsules/single NR were verified in the two trials; study-specific explicit salt and free-active equivalents were not. NRPT/CMA coingredients were separated. [S01, S02, S04, S05] | | Biochemical and clinical endpoints | NAD-related changes, ALT/AST, weight and muscle mass are not intrahepatic fat. Earlier independent NR verdicts were not re-graded. [S01, S02; supplied candidate originals] | | Adverse events/testing interference | Dollerup12weeks: minor-AE participants4/20 NR versus2/20 placebo. NR: pruritus, sweating, bloating and transient stool changes; placebo: reflux/loose stools. No serious AE reported. Remie reported no AEs/side effects in the13-completer study,6weeks per condition. These were not established as specific laboratory-interference studies. [S01, S02] |

Safety is **Caution**. This is not a warning borrowed automatically from nicotinic acid. Small, selected samples and short exposure do not establish safety of long-term high doses, pregnancy, pediatric use, advanced liver or severe renal disease, or specific drug combinations. Unreported adverse events or a within-arm ALT reduction cannot provide that assurance. The evidence is not a reason to independently stop existing care, diet or exercise or turn trial doses into a personal prescription. [S01, S02; applicability judgment]

## Final values under the supplied rules

**S / R1 / I1 / E0 / B1 / CX → D → zero strong axes → fixed 28 points.** The supplied grade calculator was executed, with no override. The score is a Chamgap rule-based anchor, not treatment success probability or official GRADE. Exact rule files/hashes and execution are in grading_audit.json and calculations.json.

S: intrahepatic fat is a surrogate, not evidence for cirrhosis, cancer, mortality or histological improvement.

R1: the closest randomized evidence is the obese-men cohort with elevated mean liver fat. The lower-fat Remie cohort differs in dose, duration and MRS definition and does not establish comparable replication or repeated refutation. R1 does not certify adequate liver-primary power.

I1: public/foundation support is mixed with manufacturer product provision and a Dollerup coauthor's advisory, equity and grant interests.

E0 codes nonsignificant verified single-NR liver-fat contrasts under the supplied rule, not exact zero, equivalence or exclusion of benefit. The favorable -2.0-pp/18%-relative within-NR signal is retained.

B1: the decisive closer Dollerup cohort has one explicitly listed defect, total n40<200. Twelve weeks is not <12 weeks; all completed and all available data entered a mixed model. Liver underpowering is not counted again as a separate small-sample defect. Remie's small/6-week/completer limitations are recorded as indirect evidence. Inaccessible registration, commercial support and missing CI are not invented extra bias defects.

CX: exact relevant two-sided between-arm CIs could not be verified or reconstructed. Available mean/SE/n, model tests and a baseline-SD sensitivity were checked, but change covariance/model detail and unrounded Remie paired SE are missing. No personal30%-MRI-PDFF-response rule or invented MCID is used to assert C1 or E~.

Conventional magnitude sensitivity was also examined. Under the supplied rubric's descriptive d=0.2/0.5/0.8 conventions, baseline SEMs and n from Dollerup yield pooled baseline SD about8.16pp; dividing the rounded change contrast−1.8pp by that SD gives about−0.22. **This is a baseline-SD-scaled descriptive sensitivity, not the paper's SMD, change-score SD, exact model effect or CI.** It only contextualizes the favorable point estimate; it does not establish or exclude clinical benefit. Missing change covariance and exact model coefficient/SE prevent valid reconstruction of the relevant two-sided CI. [S01; calculations.json CALC01–04; supplied rubric]

The content is complete with the narrow conclusion that directly applicable evidence for the selected metabolic-fatty-liver population is limited. Human liver-fat studies exist, so no_human_study=false. Indirect low-fat cohorts, mixtures and mechanisms were not accumulated into repeated refutation or established benefit.

## Search scope, information state and revision conditions

Cutoff 2026-09-17. The complete supplied 3138-record index, nine NR originals, most recent research/search/source map and deployment receipt were reviewed first. Public web-index searches used Korean/English NR, liver fat/hepatic lipid, NAFLD/MAFLD/MASLD, trial/registry,2025–2026 and correction/retraction terms, followed by primary-paper and registry opens. The 34 actual queries and access limits are in search_log.json. No subscription Embase/CENTRAL census or author contact was conducted; result totals or complete inclusion fractions were not invented.

| Information state | Exact unresolved item | Current handling and revision condition | |---|---|---| | confirmed | Hepatic-fat measurements/reported values and nonsignificant comparisons in two distinct single-NR cohorts; compound boundaries | Supports the present conclusion; paper and registry of one study are not counted as different trials | | not_reported / inaccessible | Dollerup exact model coefficient/two-sided CI and paired liver n; Remie unrounded paired SE | Revise the same page when original model tables/individual data/supplements change the available values; no current interpolation | | inaccessible | Registry outcome/change histories, Lapatto2023 final liver table and some supplements | Explicitly bounded access; gaining access prompts source comparison, not automatic efficacy confirmation | | not_reported / unverified | Participant-level MASLD eligibility, quantitative long-term alcohol, total B3 intake, deficiency, salt/free-active amount and actual medication changes | No extrapolation to all MASLD, deficiency treatment or a specific chemical form without corresponding data | | not_applicable | Isolated NR effect from mixtures; converting NAD/liver enzymes into hepatic fat | Different interventions/endpoints without an isolating contrast; no conversion |

Revision conditions include an adequately powered, preregistered single-NR imaging trial in diagnostically defined target patients with appropriate control, between-arm CI and diet/weight/medication information; original-paper corrections/retractions; or verified numerical or classification-boundary errors. These are future evidence conditions, **not unfinished clinical TODOs**. The present author has finalized clinical decisions; technical handoff is limited to format, ID, collision, build and display checks.

## Reuse, publication history and references

314,929,1068,3093,3094,3135,3136,3137 and3138 concern different independent claims. No exact completed single-NR intrahepatic-fat claim was identified in the complete supplied index and nine originals, so this is new content. Existing NR searches/source extractions were reused, without changing old grades or writing. Supplied completion records/receipts establish TASK1019→3135,1020→3136,1021→3137 and1022→3138 as actually published. An original manifest's earlier unassigned state remains historical and is separate from this new TASK1023's unassigned site metadata. [input_access.json; duplicate_audit.json; provenance/input_14.json, input_20.json, input_21.json]

**[S01]** A randomized placebo-controlled clinical trial of nicotinamide riboside in obese men: safety, insulin-sensitivity, and lipid-mobilizing effects. DOI: 10.1093/ajcn/nqy132. https://www.brennerlab.net/files/dollerup18.pdf

Access: full_text_primary_author_hosted_pdf_and_table_figure_screenshots. Locations: PDF p2 Table1 and Methods; PDF p4 study preparations and primary M-value; PDF p6 MR method/statistics; PDF p7 Figure4 and all-available-data analysis; PDF p8 HLC results, subgroup and safety/Table3; PDF p10 discussion: absolute and relative reductions, liver power limitation and conflicts.

**[S02]** Nicotinamide riboside supplementation alters body composition and skeletal muscle acetylcarnitine concentrations in healthy obese humans. DOI: 10.1093/ajcn/nqaa072. https://pure.amsterdamumc.nl/ws/files/107460914/nqaa072.pdf

Access: full_text_primary_repository_pdf_and_screenshot. Locations: Journal pp414-415 population/design (repository PDF pp4-5); Journal p416 MRS methods, day36 and lipid/water denominator (PDF p6); Journal p417 sample size, statistics, n13, adverse events (PDF p7); Journal p420 Ectopic lipid storage (PDF p10); Discussion and acknowledgments for funding/product support.

**[S03]** Nicotinamide riboside improves muscle mitochondrial biogenesis, satellite cell differentiation, and gut microbiota in a twin study. DOI: 10.1126/sciadv.add5163. https://pmc.ncbi.nlm.nih.gov/articles/PMC9839336/

Access: provided_previous_design_extraction_and_current_indexed_lead; current_primary_liver_table_inaccessible. Locations: Provided candidate3135 sources/research extraction; Current indexed PMC lead mentions Table3; final table not directly readable.

**[S04]** Nicotinamide riboside and pterostilbene reduces markers of hepatic inflammation in NAFLD: A double-blind, placebo-controlled clinical trial. DOI: 10.1002/hep.32778. https://onlinelibrary.wiley.com/doi/full/10.1002/hep.32778

Access: full_text_primary_html. Locations: Participants and Methods; Table1 mITT n94; Table2 primary HFF change; Table3 explicitly post hoc subgroup; Table4 liver enzymes.

**[S05]** Combined metabolic activators therapy ameliorates liver fat in nonalcoholic fatty liver disease patients. DOI: 10.15252/msb.202110459. https://link.springer.com/article/10.15252/msb.202110459

Access: full_text_primary_html. Locations: Methods: study design, subjects and intervention; Methods: primary MRI-PDFF and statistical analysis; Results: clinical trial.

**[S06]** NCT02303483 registry record. https://clinicaltrials.gov/study/NCT02303483

Access: registry_shell_only; full outcome/history payload inaccessible. Locations: Page shell and attempted API access; primary endpoint taken from S01 paper, not claimed independently registry-verified.

**[S07]** NCT02835664 registry record. https://clinicaltrials.gov/study/NCT02835664

Access: registry_shell_only; full outcome/history payload inaccessible. Locations: Page shell; paper S02 supplies registry identifier and power endpoint.

**[S08]** Effect of Nicotinamide Riboside on Ketosis, Fat Oxidation & Metabolic Rate. https://clinicaltrials.gov/study/NCT06044935

Access: official_indexed_title_and_page_shell; current_full_protocol_and_results_unverified. Locations: Indexed registry title and low-carbohydrate context; no verified liver-fat result.

**[S09]** The effect of obesity and aging on NAD+/Sirtuin metabolism transcription and DNA methylation in subcutaneous adipose tissue of monozygotic twin pairs discordant for BMI. DOI: 10.1038/s41366-025-02007-w. https://www.nature.com/articles/s41366-025-02007-w

Access: full_text_primary_html (cookie-redirect URL). Locations: Abstract and Methods: 49 BMI-discordant monozygotic twin pairs, phenotyping and transcriptomics.

**[S10]** Underpowered or negative? A crucial distinction. DOI: 10.1007/s00125-019-4853-x. https://link.springer.com/article/10.1007/s00125-019-4853-x

Access: publisher_metadata_preview_only. Locations: Title/publication metadata; full letter not read.

Why this is classified as D (28)

The supplied rules map S/R1/I1/E0/B1/CX with zero strong axes to D28. E0 codes the verified nonsignificant comparisons, not a mathematically zero effect, equivalence or exclusion of benefit.

Counterpoint. Adequately powered single-NR trials in a clearly diagnosed fatty-liver population with exact between-arm CIs could revise this conclusion. The present clinical decision is complete.

Rejudgment record. Actual supplied calculator and fixed anchor applied without override

Review performed and remaining limitations

Exact model-based between-arm CIs and some liver denominators, the unrounded Remie paired SE, registry detail and some primary/supplementary material, diagnostic MASLD applicability, long-term alcohol, nutrition/deficiency and salt/active-equivalent detail remain unverified. This does not establish benefit exclusion, equivalence or an exactly zero effect.

Search scope and limitations. search_log.json

Evidence Table

Receipt — 10 References

Evidence access cutoff: 2026-09-17. Each citation states its actual access level (full-text transcription, abstract, index, or other) and remaining limitations. Bibliographic checking does not certify the study results or treatment efficacy.

What this document does and does not do

Chamgap is an information source. It reports what research has and has not confirmed; it does not tell readers what to take or buy. That decision belongs to readers and, when needed, medical or legal professionals. This verdict reflects literature available up to the search date and may change as new research appears. Nothing here is medical advice.