Nicotinic acid × liver-injury adverse effects: safety by IR, SR and ER formulation

Useful facts when choosing a product

• Niacin-deficiency ascertainment and baseline nutritional biomarkers were not established in the core lipid trials/cases. Neither adequacy nor deficiency is assigned.
• Oral nicotinic acid is separated into IR/SR/ER. Nicotinamide, NR, NMN, inositol hexanicotinate and multi-ingredient energy drinks are not pooled as equivalent exposures. Product analysis/manufacturer is unreported in some cases.
• Total dietary intake and niacin equivalents (NE) are unreported in the core evidence. Reported drug mg/day is not total dietary intake or a nutritional-supplement recommendation.
• Concurrent vitamin intake was not comprehensively established. The ER label warns that niacin/niacinamide-containing vitamins may potentiate adverse effects, not that their hepatic risks are identical.
• Study-specific ALT/AST, bilirubin, INR, jaundice, encephalopathy, discontinuation and death remain separate. Lipid improvement is not this hepatic-safety endpoint.
• Label-listed catecholamine fluorometric and Benedict urine-glucose interference is separate and does not justify dismissing liver-enzyme elevations as artifacts.

What the research actually shows

30-second answer — High-dose oral nicotinic acid has a liver-injury hazard, including reported liver failure; the safety label is Warning. A small escalating-dose SR-versus-IR trial found more author-defined hepatotoxic effects with SR, but its rate cannot be generalized to all formulations. ER enzyme elevations, discontinuations, serious hepatobiliary events and severe case reports are distinct endpoints and do not quantify risk from ordinary nutritional intake or nicotinamide. [S01–S06]

The distinction is between **whether a hepatic hazard exists and how often it occurs**. The evidence supports a hazard, but does not support a universal rate pooled across formulations, exposures and outcome definitions. Warning is not an inverted efficacy score and does not mean that every exposed person is injured. [S01–S06; editorial synthesis]

1. Question and scope — The proposed question concerns liver injury in people exposed to oral nicotinic acid. Actual core evidence mainly involves pharmacological treatment of adults with lipid/vascular disorders and uncontrolled cases, not treatment of confirmed deficiency. Baseline nutritional status, total dietary intake and other vitamins are often unreported; these gaps were not filled with assumptions of adequacy or no co-exposure. [S01–S06]

IR, SR and ER retain the formulations named in their sources. The single translated label “sustained release” is not used to imply interchangeability or equivalent product release characteristics. HPS2-THRIVE studied **ER niacin plus laropiprant**, not NIASPAN alone. Nicotinamide, NR, NMN, inositol hexanicotinate and multi-ingredient energy drinks are outside the same-exposure quantitative synthesis. This safety endpoint is distinct from the trials’ registered efficacy primary endpoints; completed cardiovascular/stroke judgments were not repeated. [S01–S06; input molecular boundaries]

2. Expert evidence table — | Evidence | Actual population, exposure, comparator and time | Hepatic endpoint and reported finding | Scope limit | |---|---|---|---| | [S01](https://jamanetwork.com/journals/jama/article-abstract/366299) | 46 adults with hypercholesterolemia, 23 per SR/IR arm; 500→1000→1500→2000→3000 mg/day, each step 6 weeks. Not all completed the maximum dose. | Author-defined hepatotoxic effects; exact operational criteria not available in the abstract. **SR 12/23 (52%) versus IR 0/23. All-cause discontinuation before completing the maximum dose: 18/23 versus 9/23.** | Signal from one small escalation trial. The 12 events are not relabeled as adjudicated DILI or liver failure, nor generalized to all SR products. Zero IR events do not prove absence of risk. | | [S02](https://dailymed.nlm.nih.gov/dailymed/drugInfo.cfm?setid=dd60fe81-3d54-4a29-9502-601be89c575e) | Official ER label summary of 3 placebo-controlled trials: 245 ER recipients, final 500–3000 mg/day, mean 17 weeks. Distinct from the labeled 500–2000 mg/day range. | >3×ULN elevations in patients with normal baseline transaminases; discontinuation for >2×ULN. **No >3×ULN elevation in the normal-baseline subgroup (subgroup denominator not reported). 2/245 ER recipients (<1%; recalculated 0.82%) discontinued for >2×ULN.** | Do not record 0/245 DILI. Liver-specific placebo counts, trial overlap and severe-DILI incidence cannot be established from this summary. | | [S03](https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4156937/) | AIM-HIGH: ER 1500–2000 mg/day in 1718 participants versus masked control containing IR 100–150 mg/day in 1696; statin-based care in both groups, mean follow-up 36 months. | Serious hepatobiliary adverse events (MedDRA system organ class). **24/1718 (1.4%) versus 17/1696 (1.0%), reported p=0.29.** | Nonsignificance does not establish equivalence or safety. This broad category is not adjudicated DILI; participants passed a tolerability run-in. | | [S04](https://pmc.ncbi.nlm.nih.gov/articles/PMC3640201/) | HPS2-THRIVE: ER nicotinic acid 2000 mg+laropiprant 40 mg/day in 12838 participants versus placebo in 12835; simvastatin±ezetimibe in both arms, median follow-up 3.9 years. | Consecutive ALT >3×ULN without muscle damage; separately ALT >3×ULN plus bilirubin ≥2×ULN. **48 versus 30 participants, annual rates 0.10% versus 0.06%, p=0.04. The ALT/bilirubin criterion occurred in 14 versus 18, p=0.48. Hepatitis with no alternative cause identified: 4 versus 2; all recovered.** | Do not attribute the whole combination-product effect to niacin alone. Annual and cumulative rates, and all-results versus routine-only columns, differ. Run-in selection, sparse hepatitis and multiple comparisons limit inference. | | [S05](https://pmc.ncbi.nlm.nih.gov/articles/PMC6211325/) | 74-year-old woman without known liver disease. In hospital, IR 500 mg three times daily was changed to ER 500 mg three times daily; total 1500 mg/day unchanged. Immunosuppressants, steroid escalation and intercurrent infection were present. | Author-attributed acute liver failure, with coagulopathy and encephalopathy despite relatively low bilirubin. **Encephalopathy on day 5; day 6 ALT 523 U/L, AST 1091 U/L, bilirubin 1.8 mg/dL, INR 10.1. Fatal outcome.** | A case report based on temporal association and assessment of alternatives. No formal causality score or incidence denominator. ER three times daily is not the current labeled once-at-bedtime regimen. | | [S06](https://journals.sagepub.com/doi/10.1177/23247096231224349) | 79-year-old man with pre-existing compensated cirrhosis, diabetes and metabolic syndrome. Reported self-initiated ER 2500 mg/day for 15 years; release characteristics were not independently verified. | Author-attributed cholestatic injury with decompensation of pre-existing cirrhosis. **Jaundice and encephalopathy, followed by death on hospital day 16.** | Metabolic steatohepatitis and other confounding prevent assigning the entire cirrhosis to niacin alone. Fifteen years of exposure is not an established latency; no denominator or formal causality score is available. |

3. Definitions, causality and contrary findings — Enzyme elevations are biochemical signals. Magnitude, repeat measurements, symptoms, bilirubin, coagulation, muscle injury and alternative causes matter. Author-defined hepatotoxic effects, regulatory laboratory findings, serious hepatobiliary events, clinical hepatitis and acute liver failure are not interchangeable outcomes. Historical events were not retroactively relabeled as independently adjudicated modern DILI. No individual RUCAM/RECAM score was fabricated. The AASLD guidance landing page and partial primary excerpts were available, but its full text was not comprehensively reviewed. [S01–S07]

The randomized SR–IR comparison strengthens causal interpretation within that small escalation experiment. It does not establish a general SR liver-failure frequency, a fixed 3-g risk or safety of IR outside 23 studied recipients. Exact operational event definitions, event doses, baseline liver exclusions and co-medications remain unavailable from the accessed abstract. [S01]

Contrary or less adverse findings are preserved: no >3×ULN elevation in the ER label’s normal-baseline subgroup; no statistically significant excess of broad serious hepatobiliary events in AIM-HIGH; and no significant excess of the combined ALT/bilirubin criterion in HPS2-THRIVE. Missing subgroup denominators, broad outcome categories and nonsignificance do not prove safety. There is no direct SR-versus-ER risk ratio from these heterogeneous sources. [S02–S04]

In HPS2-THRIVE, repeated ALT elevations without muscle injury were more frequent, but the product also contained laropiprant. Statin background treatment and active tolerability selection remain part of the comparison. ALT/bilirubin thresholds do not by themselves prove causal Hy’s-law DILI; hepatitis without another identified cause was sparse. [S04]

The 2018 case adds a temporal formulation-switch signal and fatal clinical severity, but includes an ER three-times-daily schedule, immunosuppressants, steroid escalation and infection. It did not demonstrate recovery after withdrawal. The 2024 cirrhosis case includes substantial pre-existing metabolic liver disease, so niacin cannot be declared the sole cause of the entire cirrhosis. Neither case provides an incidence denominator, comparative risk or NNH. Histological descriptions are the authors’ text, not a new interpretation of biopsy images. [S05–S06]

4. Numerical checks and time scales — All calculations are explicitly separate from source-reported values. Author-defined SR events were 12/23=52.17%, versus IR 0/23; no pseudocount was inserted to create an RR. The ER label’s 2/245=0.82% describes discontinuation for >2×ULN, not adjudicated injury. [S01–S02; CALC01–02]

AIM-HIGH’s crude serious-hepatobiliary proportions were 1.40% versus 1.00%, an approximately 0.39-percentage-point difference. The exploratory crude RR was 1.39 (log-Wald 95% CI 0.75–2.58); the source-reported p=0.29 remains separate. This is not a causal-DILI effect estimate. [S03; CALC03]

For HPS2-THRIVE, 48/12838=0.37% and 30/12835=0.23% are crude participant proportions, whereas the published annual rates of 0.10% and 0.06% use 46239 and 46359 person-years. Median follow-up was not multiplied by randomized n to manufacture person-time. ALT-related active-run-in discontinuation, 472/38369=1.23% over a mean 7.4 weeks, is not a post-randomization comparative risk. [S04; CALC04–06]

5. Safety, liver disease and co-exposures — Do not substitute release formulations solely on equal milligrams. Review IR-to-SR/ER switches, pharmacological doses, baseline liver disease, alcohol and co-medications. US ER labeling contraindicates active liver disease or unexplained transaminase elevations and requires laboratory monitoring. Suspected symptoms such as jaundice, dark urine, persistent nausea, marked weakness or confusion warrant prompt medical assessment; stopping, restarting or changing formulation requires clinician-directed decisions. This is not a personal dosing or treatment-change instruction. [S02, S05–S06]

The US ER label specifies baseline tests, every 6–12 weeks during the first year and periodic testing thereafter, approximately every 6 months. This is product-label information, not a personal schedule for every product or patient. Historical trial doses up to 3000 mg/day were not turned into the current labeled maximum or a recommended dose. Selection against active liver disease and intolerance limits extrapolation to cirrhosis, substantial alcohol use, children and pregnancy. [S02–S04]

AIM-HIGH and HPS2-THRIVE are add-on strategies on shared statin-based treatment. Historical statin intolerance in the Nawaz case is not current statin co-exposure. Unreported co-medications or vitamins are not coded as absent. [S03–S06]

6. Mandatory NUT/R01 extraction — - **baseline_nutrient_status** (`not_reported`): Niacin-deficiency ascertainment and baseline nutritional biomarkers were not established in the core lipid trials/cases. Neither adequacy nor deficiency is assigned. - **molecule_and_form** (`confirmed_with_limits`): Oral nicotinic acid is separated into IR/SR/ER. Nicotinamide, NR, NMN, inositol hexanicotinate and multi-ingredient energy drinks are not pooled as equivalent exposures. Product analysis/manufacturer is unreported in some cases. - **diet_total_and_supplement** (`not_reported`): Total dietary intake and niacin equivalents (NE) are unreported in the core evidence. Reported drug mg/day is not total dietary intake or a nutritional-supplement recommendation. - **other_vitamins** (`not_reported`): Concurrent vitamin intake was not comprehensively established. The ER label warns that niacin/niacinamide-containing vitamins may potentiate adverse effects, not that their hepatic risks are identical. - **blood_markers_and_symptoms** (`confirmed`): Study-specific ALT/AST, bilirubin, INR, jaundice, encephalopathy, discontinuation and death remain separate. Lipid improvement is not this hepatic-safety endpoint. - **interference** (`confirmed_with_limits`): Label-listed catecholamine fluorometric and Benedict urine-glucose interference is separate and does not justify dismissing liver-enzyme elevations as artifacts.

7. Assessment and publication state — Human comparative data, clinical reports and labeling support a liver-injury hazard; the precision of formulation-specific clinical-DILI rates is much weaker. Both Warning and the uncertainty are retained. The efficacy A–F formula, score anchors and axes are inapplicable; `grading_status=not_applicable_or_policy_missing`. This is not a formal GRADE assessment. [S01–S06; supplied safety policy]

The principal studied intervention is medicinal pharmacological oral exposure, so the provisional S classification is changed to **M** for the core evidence, while self-supplementation cases remain distinct strata. The existing **liver** category is retained. No new site ID, URL or semantic code was issued. All 3130 index records and 8 candidate originals were compared. IDs **128, 2346, 3082, 3083, 3091, 3092, 3129 and 3130** supply reusable formulation/source/safety context; their completed content and scores remain untouched. [Input classification and duplicate audit]

Precise formulation-specific absolute rates of consistently adjudicated DILI or liver failure remain unestablished. The small SR trial’s detailed definition, the ER normal-baseline subgroup denominator, and long-term, pediatric, pregnancy and baseline-liver-disease risks across formulations are unresolved. Full-text/supplement access limitations do not prove safety or absence of human studies.

Targeted searches include material through 2026-09-16. The partially accessible Duffy 2024 case supplies no formulation-specific quantitative estimate, and a 2026 energy-drink report concerns a different multi-ingredient exposure. Neither exclusion means there are no human studies. Access limitations and the non-exhaustive search scope are recorded. Content is complete with declared uncertainty; self-check is not independent peer review. **Only technical safety-format mapping, ID linkage, build and deployment remain for Codex; no new clinical values are delegated.** [S07–S09; search log]

Why the safety label is Warning

Human comparative data, clinical reports and labeling support a liver-injury hazard; the precision of formulation-specific clinical-DILI rates is much weaker. Both Warning and the uncertainty are retained. The efficacy A–F formula, score anchors and axes are inapplicable; `grading_status=not_applicable_or_policy_missing`. This is not a formal GRADE assessment. [S01–S06; supplied safety policy]

The principal studied intervention is medicinal pharmacological oral exposure, so the provisional S classification is changed to **M** for the core evidence, while self-supplementation cases remain distinct strata. The existing **liver** category is retained. No new site ID, URL or semantic code was issued. All 3130 index records and 8 candidate originals were compared. IDs **128, 2346, 3082, 3083, 3091, 3092, 3129 and 3130** supply reusable formulation/source/safety context; their completed content and scores remain untouched. [Input classification and duplicate audit]

Precise formulation-specific absolute rates of consistently adjudicated DILI or liver failure remain unestablished. The small SR trial’s detailed definition, the ER normal-baseline subgroup denominator, and long-term, pediatric, pregnancy and baseline-liver-disease risks across formulations are unresolved. Full-text/supplement access limitations do not prove safety or absence of human studies.

Targeted searches include material through 2026-09-16. The partially accessible Duffy 2024 case supplies no formulation-specific quantitative estimate, and a 2026 energy-drink report concerns a different multi-ingredient exposure. Neither exclusion means there are no human studies. Access limitations and the non-exhaustive search scope are recorded. Content is complete with declared uncertainty; self-check is not independent peer review. **Only technical safety-format mapping, ID linkage, build and deployment remain for Codex; no new clinical values are delegated.** [S07–S09; search log]

Counterevidence and limits. Enzyme elevations are biochemical signals. Magnitude, repeat measurements, symptoms, bilirubin, coagulation, muscle injury and alternative causes matter. Author-defined hepatotoxic effects, regulatory laboratory findings, serious hepatobiliary events, clinical hepatitis and acute liver failure are not interchangeable outcomes. Historical events were not retroactively relabeled as independently adjudicated modern DILI. No individual RUCAM/RECAM score was fabricated. The AASLD guidance landing page and partial primary excerpts were available, but its full text was not comprehensively reviewed. [S01–S07] The randomized SR–IR comparison strengthens causal interpretation within that small escalation experiment. It does not establish a general SR liver-failure frequency, a fixed 3-g risk or safety of IR outside 23 studied recipients. Exact operational event definitions, event doses, baseline liver exclusions and co-medications remain unavailable from the accessed abstract. [S01] Contrary or less adverse findings are preserved: no >3×ULN elevation in the ER label’s normal-baseline subgroup; no statistically significant excess of broad serious hepatobiliary events in AIM-HIGH; and no significant excess of the combined ALT/bilirubin criterion in HPS2-THRIVE. Missing subgroup denominators, broad outcome categories and nonsignificance do not prove safety. There is no direct SR-versus-ER risk ratio from these heterogeneous sources. [S02–S04] In HPS2-THRIVE, repeated ALT elevations without muscle injury were more frequent, but the product also contained laropiprant. Statin background treatment and active tolerability selection remain part of the comparison. ALT/bilirubin thresholds do not by themselves prove causal Hy’s-law DILI; hepatitis without another identified cause was sparse. [S04] The 2018 case adds a temporal formulation-switch signal and fatal clinical severity, but includes an ER three-times-daily schedule, immunosuppressants, steroid escalation and infection. It did not demonstrate recovery after withdrawal. The 2024 cirrhosis case includes substantial pre-existing metabolic liver disease, so niacin cannot be declared the sole cause of the entire cirrhosis. Neither case provides an incidence denominator, comparative risk or NNH. Histological descriptions are the authors’ text, not a new interpretation of biopsy images. [S05–S06]

Review performed and remaining limitations

Precise formulation-specific absolute rates of consistently adjudicated DILI or liver failure remain unestablished. The small SR trial’s detailed definition, the ER normal-baseline subgroup denominator, and long-term, pediatric, pregnancy and baseline-liver-disease risks across formulations are unresolved. Full-text/supplement access limitations do not prove safety or absence of human studies. The AASLD full text, some trial supplements and some recent case full texts were inaccessible. Current conclusions disclose the accessed scope; Codex is not asked to fill new clinical values.

Evidence Table

Receipt — 9 References

Evidence access cutoff: 2026-09-16. 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.