Niacin and Glioblastoma: The Phase I–II Trial

Table of Contents

1. The Question the Trial Asked
2. How the Trial Was Designed
3. Phase I: Who Took Part and the Dose Escalation
4. Side Effects and Dose-Limiting Toxicities
5. Phase II Interim Result: 82% Versus About 54%
6. Reading the Numbers Carefully
7. How Niacin Might Act: The Immune Hypothesis in Plain Language
8. Earlier Research: Mice, Blood Markers and the Standard of Care
9. Niacin at Gram Doses: Safety Findings From Other Trials
10. The 2026 Press Release Compared With the Paper
11. Limitations and What Remains Unknown
12. Key Research Papers
13. Connections

1. The Question the Trial Asked

Glioblastoma grows quickly, spreads into surrounding brain tissue in thread-like extensions that surgery cannot fully remove, and almost always returns after treatment. The 2005 phase III trial by Stupp and colleagues, which enrolled 573 patients, showed that adding temozolomide to radiotherapy raised median survival from 12.1 to 14.6 months and two-year survival from 10.4% to 26.5%. That combination — maximal safe surgery, then radiotherapy with temozolomide, then further cycles of temozolomide — is the standard of care the Calgary trial builds on.

The Calgary group's question was narrow and specific: can controlled-release niacin be given safely on top of that standard treatment, at what dose, and does it raise the share of patients whose tumor has not progressed six months after diagnosis? The trial was described by its authors as first-in-human for this use. It does not test niacin as a replacement for surgery, radiotherapy or chemotherapy; every participant received the full standard treatment as well.

The idea came from the same laboratory's earlier work in mice and in cell culture (Sarkar 2020, described in section 8), which reported that niacin reactivated immune cells called monocytes and macrophages and slowed brain tumor growth. The trial is the first step in finding out whether anything similar happens in people.

Back to Table of Contents

2. How the Trial Was Designed

The trial is registered at ClinicalTrials.gov as NCT04677049. The registry lists the lead sponsor as a provincial cancer-control agency in Alberta, a start date of 18 March 2021, an estimated primary completion date of December 2027, an estimated total enrollment of 59 people across both phases, and a status of active, not recruiting. The abstract describes it as a single-arm, open-label phase I–II study.

• Single-arm. Every participant received niacin. There is no concurrent group receiving standard treatment alone.
• Open-label. Patients and doctors knew niacin was being given. There was no placebo.
• Eligibility. Adults aged 18 to 75 with newly diagnosed glioblastoma who were eligible for standard of care: surgery, radiotherapy and temozolomide.
• Intervention. A controlled-release niacin tablet added to standard treatment. In Canada the product was regulated as a natural health product.
• Phase I. Dose escalation within each patient, from 500 mg/day up to 3,000 mg/day, to find the dose-limiting toxicity (DLT), the maximum tolerated dose (MTD) and the recommended phase II dose (RP2D).
• Phase II. The target was an absolute increase of at least 20 percentage points in six-month progression-free survival (PFS-6M) over historical controls.
• Interim futility analysis. Planned for the point at which 24 patients became evaluable for PFS-6M. The trial would stop early if conditional power fell below 20% or the futility index rose above 80%.
• Historical benchmark. The abstract states standard-of-care figures of 15 months survival and a PFS-6M of 53.9%. The source of the 53.9% figure is not given in the abstract and does not appear in the abstract of the 2005 Stupp trial; it is the historical figure the authors used.

Progression-free survival at six months means the share of patients who are alive and whose tumor has not grown or returned on imaging six months into treatment. In this trial, progression was judged by central radiology review, meaning scans were assessed by reviewers separate from the treating team.

Back to Table of Contents

3. Phase I: Who Took Part and the Dose Escalation

Fifteen patients took part in phase I. The abstract describes them as follows:

• Median age 57 years, range 37 to 68.
• 40% were women (6 of 15).
• 47% had a methylated MGMT promoter (7 of 15).

MGMT is a DNA-repair enzyme. When its gene's promoter is methylated, the tumor makes less of the enzyme and is less able to repair the damage temozolomide causes, so patients with a methylated MGMT promoter generally respond better to temozolomide. The share of methylated tumors in a small trial therefore matters when its results are compared with other groups of patients.

Each phase I patient started at 500 mg/day, and the dose was stepped up within that same patient toward a planned ceiling of 3,000 mg/day. This within-patient escalation differs from the more common design in which successive small groups of patients each receive one fixed, higher dose.

The abstract states the maximum tolerated dose as 2,000 mg/day. Its results line reads: “Niacin dose escalated up to 2000 mg/d is the ongoing RP2D”, meaning that escalation up to 2,000 mg/day is the regimen carried forward into phase II.

Back to Table of Contents

4. Side Effects and Dose-Limiting Toxicities

The most common side effect was flushing, reported in 10 of the 15 phase I patients; in 9 of those 10 it was grade 1, the mildest grade on the standard cancer-trial toxicity scale. Flushing — warmth, redness and tingling of the face and upper body — is the best-known effect of nicotinic acid. It is driven by the niacin receptor HCAR2 (also called GPR109A) on skin cells, which triggers release of prostaglandins that widen small blood vessels. Controlled-release forms were developed in part to blunt this effect by releasing the drug more slowly.

Two dose-limiting toxicities occurred at 2,500 mg/day:

• Grade 3 thrombocytopenia — a fall in blood platelets, the cells that help blood clot, to a level classed as severe.
• Hyperbilirubinemia — a rise in bilirubin, a breakdown product of red blood cells that the liver clears; a raised level can signal strain on the liver.

Both findings have context. Temozolomide itself commonly lowers platelet counts, so in a combined regimen it is not simple to attribute a platelet fall to one drug alone; the abstract records it as a DLT at that niacin dose. Liver effects, including raised liver enzymes and, rarely, serious liver injury, have long been documented with high-dose nicotinic acid, especially with some sustained-release formulations. These two events at 2,500 mg/day are why the maximum tolerated dose was set at 2,000 mg/day.

The abstract does not give a full table of all adverse events by grade; that is in the paywalled full text.

Back to Table of Contents

5. Phase II Interim Result: 82% Versus About 54%

At the interim analysis, six-month progression-free survival by central radiology review was 82.3%. The historical figure the authors compared it with was 53.9%. The authors' conclusion states: “The interim analysis already showed an absolute increase in PFS-6M of 28%”. The arithmetic is 82.3 minus 53.9, which is 28.4 percentage points.

That difference exceeds the 20-point target the trial was designed to look for. The registry lists the trial as active, not recruiting, with phase II still under way. Several things the abstract does not report:

• The exact number of patients included in the interim analysis. The analysis was planned at 24 evaluable patients; the abstract does not confirm the number analyzed.
• Overall survival — how long patients lived.
• How long patients had been followed at the time of analysis, beyond the six-month mark.
• Results split by MGMT methylation status, by extent of surgery, or by age in the phase II group.

The trial's primary completion is estimated for December 2027 in the registry. The final phase II result, with the full planned number of patients, has not been published.

Back to Table of Contents

6. Reading the Numbers Carefully

Percentage points, not a percentage

The gain the authors report is absolute: 28 percentage points, from about 54% to about 82%. Expressed as a relative change, 82.3 divided by 53.9 is about 1.53 — a relative increase of roughly 53%. A phrase such as “a 28% improvement” can be read either way, and the two readings describe very different sizes of effect. The paper's own wording is “absolute increase”.

A historical comparison, not a randomized one

The 53.9% figure comes from patients treated elsewhere, at another time, who were not selected by this trial's criteria. Patients who enter a trial tend to be fitter than patients in general: they must meet age limits, be well enough for full standard treatment and be willing to take part. Imaging techniques, surgery and supportive care also change over time. Each of these differences can make a new single-arm result look better than the historical benchmark without the added treatment doing anything. That is why single-arm phase II results in glioblastoma are treated as signals for a randomized trial rather than as proof.

A small interim group

With roughly two dozen patients, a handful of patients moving from progressed to not-progressed changes the percentage by several points. In a group of 24, each patient is worth about 4 percentage points.

A confidence interval that cannot be right

The abstract prints the 95% confidence interval for the 82.3% figure as 82.14–82.46%. An interval that narrow, about plus or minus 0.16 percentage points, could only come from many thousands of patients. For a group of about 24, a 95% interval around 82% would span roughly 20 or more percentage points in total. The printed interval is very likely a typographical or reporting error in the abstract, and it is not a measure of how precise the result is. The full text may give the correct figure; it was not available for this page.

The MGMT mix matters

Nearly half the phase I patients had methylated MGMT tumors, which respond better to temozolomide. Whether the phase II group had a similar mix, and whether the historical benchmark had the same mix, is not stated in the abstract. A difference in that mix alone can move six-month progression-free survival considerably.

Back to Table of Contents

7. How Niacin Might Act: The Immune Hypothesis in Plain Language

The researchers' hypothesis is that niacin works not by attacking the tumor directly but by waking up the immune system's “clean-up” cells. In plain language:

• Myeloid cells. Monocytes circulate in the blood; when they move into tissue they become macrophages. The brain has its own resident version, microglia. Together these are called myeloid cells.
• Tumors disarm them. Glioblastomas are full of macrophages and microglia, but the tumor reprograms many of them into a state that supports tumor growth and suppresses other immune cells.
• The proposal. In the group's 2020 laboratory study, monocytes exposed to niacin produced a signaling protein, interferon-α14, which slowed the growth of brain tumor-initiating cells taken from patients. In mice, niacin increased the number of macrophages and microglia in the tumor, and those cells appeared to restrain it.
• Niacin's known targets. Niacin is a precursor of NAD+, the coenzyme at the center of cellular energy metabolism, and it activates the HCAR2 receptor found on immune cells as well as skin and fat cells. The group's 2023 review discusses both routes.

In people, the evidence for this mechanism so far is indirect. The phase I immune study (Poon 2026) found changes in blood immune markers in patients taking niacin. Neither that study nor the clinical trial abstract shows immune cells attacking or killing tumor cells inside patients' brains. The immune-reactivation explanation is the researchers' working hypothesis, supported by mouse and cell-culture experiments and by shifts in blood markers.

Back to Table of Contents

8. Earlier Research: Mice, Blood Markers and the Standard of Care

Sarkar et al., 2020 — mice and cell culture

Published in Science Translational Medicine, this study screened existing drugs for the ability to reactivate monocytes against brain tumor-initiating cells, and niacin emerged as a candidate. Its findings:

• Monocytes exposed to niacin reduced the growth of brain tumor-initiating cells derived from patients, through production of interferon-α14.
• In mice with brain tumors, niacin increased macrophages and microglia in the tumor, reduced tumor size and prolonged survival.
• Adding temozolomide to niacin extended survival in mice further.
• The benefit disappeared in mice whose circulating monocytes had been depleted, pointing to those cells as necessary for the effect.

Poon et al., 2026 — immune markers in the phase I patients

Published in Neurology: Neuroimmunology & Neuroinflammation, this is the immune analysis of the same trial's phase I patients. With niacin, the authors report:

• more circulating memory T cells and natural killer (NK) cells;
• fewer nonclassical monocytes;
• higher blood levels of the signaling proteins IL-12p70 and G-CSF;
• lower levels of GRO-α.

The authors classify their own findings as Class IV evidence, the lowest tier, because the trial is open-label with no blinding and no comparison group. Some of these shifts could also reflect the effects of surgery, radiotherapy, temozolomide or steroids given over the same weeks.

Wuerch, Roldan Urgoiti and Yong, 2023 — review

In Neurotherapeutics, the same group reviewed niacin's roles in supplying NAD+ and in signaling through the HCAR2 receptor, and its study in multiple sclerosis, Alzheimer's disease, Parkinson's disease, glioblastoma and ALS. The review places the glioblastoma trial within a wider research program on niacin and the nervous system.

Stupp et al., 2005 — the standard of care

The 573-patient randomized trial in the New England Journal of Medicine that established radiotherapy plus temozolomide as first treatment: median survival 14.6 months versus 12.1 months with radiotherapy alone, and two-year survival 26.5% versus 10.4%. The Calgary trial adds niacin to this regimen.

Back to Table of Contents

9. Niacin at Gram Doses: Safety Findings From Other Trials

The doses in the glioblastoma trial, up to 2,000 mg/day, are far above the amounts of niacin found in whole foods such as liver, fish, poultry, peanuts, brown rice and mushrooms, and in the range once used as a cholesterol-lowering drug. The largest long-term trial of that dose comes from heart disease rather than cancer.

HPS2-THRIVE (2014) randomized 25,673 adults with vascular disease to 2 g/day of extended-release niacin combined with laropiprant (a drug added to reduce flushing) or to placebo, on top of statin therapy, and followed them for a median of 3.9 years. Its results:

• No reduction in major vascular events: 13.2% with niacin versus 13.7% with placebo.
• An absolute excess of serious adverse events with niacin, including disturbances of diabetes control (+3.7 percentage points), infections (+1.4), gastrointestinal events (+1.0) and bleeding (+0.7).

That population and setting differ from the glioblastoma trial: older adults with heart and vessel disease, on statins, treated for years rather than months, and with laropiprant in the combination. The findings are context for what gram-dose niacin can do over long periods, not a prediction for patients with brain tumors. The bleeding and infection signals are of particular interest given that chemoradiation also lowers platelets and white cells, and the glioblastoma trial's own dose-limiting toxicities included a severe fall in platelets.

Back to Table of Contents

10. The 2026 Press Release Compared With the Paper

A university press release about this trial was published in October 2026, about ten months after the paper appeared online. Comparing its main points with the paper, abstract and registry:

• “82%” and “a 28% improvement compared with historical data.” The paper reports an absolute increase of 28 percentage points (82.3% versus 53.9%). Read as a relative gain, the figure would be about 53%. The paper's own term is “absolute increase”.
• Immune cells “rejuvenated” to attack and kill cancer cells. The trial abstract does not test this. The supporting evidence is the 2020 mouse and cell-culture study and the 2026 analysis of blood immune markers; neither shows immune cells killing tumor cells in patients. It is the researchers' hypothesis.
• A planned total of 48 participants. The registry lists an estimated enrollment of 59 across phases I and II. The figure of 48 may refer to phase II alone; that was not confirmed.
• 24 patients in the interim analysis. The abstract states that the interim analysis was planned at 24 evaluable patients; it does not state the number analyzed.
• Comparison group. The release refers to historical data, but the paper's design — single arm, no concurrent control, interim result — is central to how much weight the 82% figure carries.
• Timing. The paper was published on 28 November 2025; the October 2026 release describes research already in print.

The funding sources named in the release, a national health-research funding agency and a provincial cancer charity, were not checked against the paper, whose funding section is behind the paywall. The PubMed declarations record that the authors declared no competing interests.

Back to Table of Contents

11. Limitations and What Remains Unknown

The authors' own limitations section is in the paywalled full text and was not read. Limitations that follow directly from the design:

• No concurrent control group. Every patient received niacin, so the trial cannot by itself separate niacin's effect from patient selection, chance or changes in care.
• Historical comparison. The 53.9% benchmark's source is not stated in the abstract, and historical groups differ from trial groups in fitness, tumor biology and care.
• Small interim cohort. Around two dozen patients, with each patient worth several percentage points.
• Interim, not final. The phase II portion continues, with primary completion estimated for December 2027.
• Open-label. No blinding of patients or investigators, though progression was judged by central radiology review.
• A misprinted statistic. The confidence interval in the abstract cannot be correct for a group of this size.

Questions the published record does not yet answer:

• Does niacin added to standard treatment lengthen overall survival, not just delay progression at six months?
• Does the effect hold in a randomized, controlled comparison?
• Does the result differ between MGMT-methylated and unmethylated tumors?
• Do the immune changes seen in blood also occur inside the tumor, and do they relate to which patients do better?
• What are the full side-effect rates at 2,000 mg/day alongside temozolomide over the whole course of treatment?
• Would other forms of vitamin B3, such as nicotinamide, which does not activate the HCAR2 receptor and does not cause flushing, behave the same way? This trial tested nicotinic acid only.

As of October 2026, controlled-release niacin is an investigational add-on in glioblastoma, studied in one single-arm trial whose final results are pending.

Back to Table of Contents

Key Research Papers

1. Roldan Urgoiti G, de Robles P, Tsang RY, Willson M, Ghosh S, Faruqi M, Lim G, Loewen S, Nordal R, Cairncross G, Leckie C, Poon CC, Yong VW (2025). A phase I-II study of niacin in patients with newly diagnosed glioblastoma: safety and interim phase II analysis. Journal of Neuro-Oncology 176(1):101 — PubMed PMID: 41313494
2. Poon CC, Hagen KM, Sarkar S, Mirzaei R, Silva C, Ueno A, de Robles P, Roldan-Urgoiti G, Yong VW (2026). Niacin modulates immune responses in a phase I dose-escalation clinical trial of newly diagnosed glioblastoma. Neurology: Neuroimmunology & Neuroinflammation 13(2):e200530 — PubMed PMID: 41632924
3. Sarkar S, Yang R, Mirzaei R, Rawji K, Poon C, Mishra MK, et al. (2020). Control of brain tumor growth by reactivating myeloid cells with niacin. Science Translational Medicine 12(537):eaay9924 — PubMed PMID: 32238578
4. Wuerch E, Urgoiti GR, Yong VW (2023). The promise of niacin in neurology. Neurotherapeutics 20(4):1037-1054 — PubMed PMID: 37084148
5. Stupp R, Mason WP, van den Bent MJ, Weller M, Fisher B, Taphoorn MJ, et al. (2005). Radiotherapy plus concomitant and adjuvant temozolomide for glioblastoma. New England Journal of Medicine 352(10):987-996 — PubMed PMID: 15758009
6. HPS2-THRIVE Collaborative Group; Landray MJ, Haynes R, Hopewell JC, et al. (2014). Effects of extended-release niacin with laropiprant in high-risk patients. New England Journal of Medicine 371(3):203-212 — PubMed PMID: 25014686

Trial registry: ClinicalTrials.gov NCT04677049. Full text of the primary paper: DOI: 10.1007/s11060-025-05351-z (subscription).

PubMed Topic Searches

1. PubMed: niacin glioblastoma
2. PubMed: niacin myeloid cells brain tumor
3. PubMed: glioblastoma PFS-6M with temozolomide
4. PubMed: HCAR2 receptor and microglia
5. PubMed: single-arm glioblastoma trials and historical controls

Back to Table of Contents

Connections

• Vitamin B3 Benefits Hub
• Vitamin B3 (Niacin) Overview
• B3 as an NAD+ Precursor
• B3 for Cholesterol & Cardiovascular
• B3 for Schizophrenia & Mental Health
• News, 11 October 2026
• Brain Cancer
• Oncology
• Vitamin B3 Toxicity
• Niacin Skin Flushing
• Niacin and Liver Damage
• NAD+ / NMN
• Complete Blood Count
• Liver Function Tests

Back to Table of Contents