High-Evidence Repurposed Drugs & GLP-1 Oncology (2026)
Series Hub · Living Review · September 2026
What already has trials, guidelines, or large human datasets — and what that does not authorize. Aspirin in PI3K-altered colorectal cancer is now guideline medicine. Metformin and statins are mixed. GLP-1 receptor agonists are the 2026 observational wave. This hub exists so “repurposed” stops meaning one bucket.
Door 1
Just resected
Ask whether the tumor was profiled for PI3K-pathway alterations. If stage II–III CRC and PI3K-altered, aspirin 100–162 mg × 3 years is the guideline conversation — not a forum protocol. Replete vitamin D. Start the exercise prescription before any supplement stack.
Door 2
Obesity, diabetes, or MASLD + cancer
If a GLP-1 or dual agonist is already indicated for the host (glycemic, weight, or cardiovascular), the 2026 oncology signal is a reason to discuss continuation, not a reason to start it as cancer therapy. Screen first for wasting, GI primary, and upcoming anaesthesia.
Door 3
On active systemic therapy
Do not add a metabolic agent during high-emesis chemo, mucositis, obstruction risk, or falling albumin/weight. Statins stay if cardio-indicated and interaction-checked. Metformin stays only if the diabetes indication is intact and eGFR allows.
Why this series exists
OneDayMD already publishes deeply on investigational antiparasitic combinations, metabolic stacking, and DMSO. Those remain Series 1–5. This hub is Series 8: the layer that should load first. The internet collapsed “repurposed” into whatever was trending on a podcast. Regulators and professional societies then collapsed the same word into a warning. Both simplifications are expensive for patients. Some cheap drugs now have randomized, molecularly selected outcome data. Some popular metabolic drugs have conference-level signals and hard contraindications in the same month. Ranking them is the work.
The editorial test for every page: can a newly diagnosed reader tell the difference between a guideline conversation, a reasonable off-label discussion, and a forum protocol? If not, the page is not ready.
What is not in this series
Ivermectin, fenbendazole, mebendazole, niclosamide, atovaquone, and DMSO are not ranked here. They live in the Repurposed Drug Series and the DMSO Series, with investigational/off-label labeling and the 2026 ASCO Clinical Notice on ivermectin and fenbendazole linked in full. Putting them on this hub would recreate the exact category error this series is designed to break.
Evidence badges
| Badge | Meaning | How we use it here |
|---|---|---|
| G | In a major guideline for a defined subgroup | Aspirin after resection in PI3K-pathway–altered stage II–III CRC (NCCN trajectory post-ALASCCA). |
| T1 | RCT or meta-analysis of RCTs on an oncology endpoint | ALASCCA aspirin; MA.32 metformin in early HER2− breast (negative). |
| T2 | Prospective data, Cochrane-level synthesis, or consistent meta-analysis | Cimetidine in CRC; statin mortality associations; exercise. |
| T3 | Large real-world / target-trial / ASCO abstract | GLP-1 incidence and stage I–III progression signals, 2025–2026. |
| T4 | Mechanism + small series only | Not featured as a lead agent in this series. |
| CONTRA | Do not use in this context | GLP-1 in progressive wasting, cachexia, obstruction, high-emesis chemo. |
A high tier in one tumor and one setting does not travel. Aspirin’s G/T1 status is for PI3K-pathway–altered resected CRC, not for pancreatic adenocarcinoma. GLP-1’s T3 status is for host-indicated use with an oncology signal — not for cachexia.
The ranked stack
Order is by strength of human oncology-relevant evidence plus how often the decision actually appears in clinic. Exercise is ranked 0 because it is still the intervention most likely to be skipped while patients bargain over pills.
| Rank | Agent | Best-defined oncology lane | Tier | Do not flatten into |
|---|---|---|---|---|
| 0 | Exercise (dose it) | Adjuvant and survivorship outcomes across solid tumors; the cleanest host intervention. | T1–T2 | A lifestyle footnote under supplements. |
| 1 | Aspirin | Adjuvant stage II–III CRC with somatic PI3K-pathway alteration (PIK3CA exon 9/20, other PIK3CA, PIK3R1, PTEN). Lynch prevention is a separate lane. | G / T1 | “Take aspirin for cancer.” |
| 2 | Cimetidine (vs default PPI) | CRC perioperative / adjuvant survival signal. Contrast with PPI associations of worse outcomes in breast trial pools and GBM. | T1–T2 | Stop every PPI without a GI-bleed indication check. |
| 3 | Statins | Continue when cardiovascular indication exists. Mortality associations are supportive, not an oncology license. Check TKI pairs. | T2 | Starting atorvastatin “for the tumor.” |
| 4 | Metformin | Keep when diabetes / insulin-resistance indication is intact. MA.32: no IDFS benefit in early HER2− breast. | T1 negative in that setting; T2 elsewhere | Metformin as universal anti-cancer. |
| 5 | Propranolol bundle | Peri-biopsy / adrenergic-stress hypothesis; selected perioperative settings. | T2 | A substitute for beta-blockade indicated for heart rate / BP. |
| 6 | Vit D / melatonin / ω-3 | Repletion and supportive-care adjuncts. | T2 | Disease-modifying headlines. |
| 7 | GLP-1 / dual agonists | Host-indicated metabolic therapy with 2026 observational signals on incidence and metastatic progression. | T3 · CONTRA if wasting | Starting semaglutide as cancer treatment. |
Part A — High-evidence agents, in brief
A1. Aspirin after PI3K-altered CRC resection
ALASCCA (NEJM 2025) randomized patients with resected localized CRC and PI3K-pathway alterations to aspirin 160 mg daily or placebo for three years. Among PIK3CA exon 9/20 hotspot mutations, 3-year recurrence was 7.7% vs 14.1% (HR 0.49). Among other PIK3CA / PIK3R1 / PTEN alterations, 7.7% vs 16.8% (HR 0.42). Severe adverse events were more common with aspirin (16.8% vs 11.6%). Subsequent NCCN-facing guidance: profile stage II–III colon tumors for somatic PI3K-pathway alterations and consider aspirin 100–162 mg daily for three years after surgical recovery, concurrent with adjuvant chemotherapy when used. A 2026 translational meta-analysis pooling ASCOLT, SAKK 41/13, and ALASCCA estimated a 39% reduction in DFS events in exon 9/20 mutants (HR 0.61).
Separate lane — Lynch syndrome: CAPP2 established high-dose aspirin for prevention; CaPP3 (2026) found 100 mg had a similar cancer-risk profile to 600 mg with less bleeding. Do not paste Lynch prevention doses onto adjuvant CRC, or the reverse. Contraindications: active bleeding, ulcer disease, concurrent anticoagulation without a plan, clinically important thrombocytopenia, aspirin-exacerbated respiratory disease.
A2. Cimetidine versus the default PPI
Cimetidine is the most under-discussed agent in this stack: older CRC survival data and a Cochrane-level signal, with an immune-adhesion rationale. The live clinical fork is usually not “add cimetidine.” It is “why is this patient on a potent PPI by default?” Pooled breast-cancer trial data have associated PPIs with worse OS/PFS and more grade ≥3 adverse events. In newly diagnosed glioblastoma, potent ALDH1A1-activating PPIs (omeprazole, pantoprazole) have been associated with inferior survival versus other acid suppression. None of that is a reason to stop ulcer prophylaxis in a high-risk GI bleeder. It is a reason to ask whether an H2 blocker or a local antacid is enough.
A3. Statins — cardio-oncology first
Meta-analyses associate statin use with lower cancer-specific mortality. That is not an indication to start a statin for the tumor. The page that earns its place is the interaction table: roughly one in four recent oncology approvals interacts with a statin; simvastatin is the noisiest, pravastatin the quietest; selected combinations (adagrasib, tucatinib, asciminib) are contraindicated. Pharmacovigilance has flagged ICI-plus-statin irAE patterns (colitis, pneumonitis, myocarditis). Monitor. Do not perform a ritual discontinuation of a drug that is protecting the heart of a patient who may now survive long enough to need one.
A4. Metformin — the honest autopsy
Diabetic cohorts have shown lower cancer incidence on metformin for years. Residual confounding is the standard critique; the ADA has said cancer rumors should not drive diabetes-drug choice. MA.32 randomized 3,649 patients with early HER2-negative breast cancer to metformin 850 mg twice daily or placebo and found no invasive disease-free survival benefit (HR 0.96). That result belongs in paragraph one of any metformin-for-cancer essay. Remaining open lanes: insulin-resistant hosts, selected gynecologic and prostate settings, AMPK–mTOR biology. Safety: eGFR thresholds, hold around iodinated contrast and hypoxia, B12 depletion, and no silent stack with berberine plus a GLP-1 without glucose logs.
Part B — GLP-1 / incretin oncology
Treat incretin mimetics as a new organ system in cancer care, not as a weight-loss sidebar. Three questions must never be collapsed: prevention in people who already qualify for the drug; continuation after a diagnosis in someone who is not wasting; and initiation during active systemic therapy or cachexia. The answers diverge.
| Question | 2025–2026 human signal | What it is not |
|---|---|---|
| Prevention in diabetes + obesity | Target-trial emulation vs DPP-4 inhibitors: composite obesity-related cancer aHR ~0.93; colon/rectal strongest. | A 50% prevention claim. Effect size is modest and residual confounding remains. |
| After diagnosis, stage I–III, no wasting | ASCO 2026 matched analyses: less progression to stage IV vs DPP-4i in NSCLC (~10% vs 22%), breast (~10% vs 20%), CRC (~13% vs 22%), HCC (~19% vs 28%). High tumor GLP-1R associated with better OS (HR ~0.67). | A registrational oncology indication. Still observational. |
| With checkpoint inhibitors | Matched real-world cohort: ~31% lower 5-year all-cause mortality; some irAE rates lower. Hypothesis-generating. | Proof that GLP-1s “boost immunotherapy.” |
| High-risk liver disease → HCC | Very large real-world reductions in incident HCC across viral, alcohol, MASLD, cirrhosis strata. | An effect size to quote as fact until a dedicated RCT exists. |
| NMIBC on BCG | Lower progression to muscle-invasive disease and lower mortality; recurrence not clearly different. | A reason to start GLP-1 solely for bladder cancer. |
Mechanism remains unsettled. Weight, insulin, and inflammation explain part of the obesity-related cancer signal. Movement in NSCLC and the association of tumor GLP-1 receptor expression with survival keep a possible direct or immune-mediated effect on the table. Neither story is mature enough to drive a start/stop decision by itself.
The only GLP-1 rule that belongs above the fold
Lean mass is the missing vital sign
In obesity trials, roughly 20–30% of GLP-1-associated weight loss can be lean tissue. In oncology, low skeletal muscle predicts more severe toxicity, more dose reductions, and worse survival. A falling scale during immunotherapy is not automatically “the drug is working.” Pair any continuation decision with protein targets, resistance training, and a muscle proxy (grip, thigh/calf, CT L3 muscle if scans already exist). Supportive-care commentary in 2026 is explicit: dietetics and body composition should govern initiation during chemo-radiotherapy.
Peri-operative and oral-drug reality
GLP-1s delay gastric emptying. Aspiration under anaesthesia is the documented risk; society guidance still disagrees on whether to hold the day of surgery, hold a week for weekly agents, or individualize with a 24-hour liquid diet and rapid-sequence induction. This hub does not adjudicate the anaesthesia literature. It requires that someone owns the plan before a resection date. Oral TKIs, capecitabine, and endocrine agents can see delayed Tmax. Flag it. Do not invent dose changes on a blog.
GLP-1 / dual agonist — hold or continue
Oncology decision aid · September 2026 · Not a protocol · Shared decision with oncology and the prescriber. Print this section and take it to clinic.
1. Does this person already have a HOST indication?
T2D · obesity + CV risk · selected MASLD · already on a GLP-1 / dual agonist
No host indication
Do not start a GLP-1 as cancer therapy. Observational signals are not an oncology license. STOP.
Yes — run all four screens before any continuation
A positive answer on any screen moves the case to HOLD. Screens are about safety and nutrition, not about how good the ASCO abstract looked.
2. Is ANY of the four screens positive?
HOLD · DO NOT START
Incretin is a nutrition or safety problem, not a metabolic opportunity. Use another glycemic agent if diabetes must be treated. Reassess only when weight is stable, gut is quiet, and treatment intensity drops.
MAY CONTINUE
Host indication intact. No wasting. No high-risk gut primary in the acute phase. Oncology is informed. This is continuation of indicated therapy plus a monitoring plan — not a new cancer protocol.
- Protein target and 2–3× weekly resistance work assigned. Scale is not the only vital sign. Track grip or mid-thigh; use CT L3 muscle if scans already exist.
- Review at week 4 and week 12: weight trajectory, strength, nausea, stools, glucose, and whether oral cancer drugs still look absorbed (delayed Tmax possible).
- Anaesthesia plan documented before any procedure. Hold-duration vs 24-hour liquid diet + RSI is unsettled in society guidance — someone must own it.
- Hold immediately for obstruction symptoms, intractable vomiting, suspected aspiration, or a new cachexia diagnosis. Do not restart from a blog.
- Compounded or “research” incretins are not an oncology product. Brand or regulated pharmacy only, if therapy continues.
Sources informing this tool: JACC CardioOncology incretin state-of-the-art 2026; ASCO 2026 GLP-1 abstracts on incidence and stage I–III progression; Supportive Care in Cancer 2026 nutrition-centred monitoring; multi-society peri-operative GLP-1 guidance (hold windows still unsettled).
Clinic ask-sheet · two conversations
Bring this page. Leave with a named owner for each decision.
Conversation A · Aspirin / PI3K
- Tumor type / stage:
- Resection date:
- Was the tumor sequenced? Y / N / pending
- PIK3CA exon 9 or 20: Y / N / unknown
- Other PIK3CA / PIK3R1 / PTEN: Y / N / unk.
- MSI / MMR / POLE (CRC):
- If stage II–III CRC and PI3K-pathway altered: has adjuvant aspirin 100–162 mg × 3 years been considered? Y / N / contraindicated
- Bleed history / ulcer / anticoagulant:
- Platelets last value:
- Start date if agreed:
- Review date:
- Lynch syndrome / CaPP context (separate):
Conversation B · GLP-1 / dual
- Current agent / dose / start date:
- Host indication: T2D / obesity+CV / MASLD
- Compounded product? Y / N (if Y, flag)
- Weight 8 weeks ago: now:
- Unintentional loss? Y / N
- Albumin / CRP:
- Grip or thigh trend:
- Primary site GI / pancreas / GB? Y / N
- Vomiting / obstruction symptoms? Y / N
- Next anaesthesia date:
- Aspiration plan owner:
- Flowchart result: HOLD / PAUSE / CONTINUE
- Oncology aware? Y / N
- Week-4 review date:
- Week-12 review date:
- Hold trigger written in notes? Y / N
Also on the table this visit · do not let these get crowded out by the trendy molecule
- Exercise prescription written (minutes, resistance days), not just “stay active.”
- PPI: is there a bleed-level indication, or can an H2 blocker / local antacid be used? (GBM and some breast data make this non-trivial.)
- Statin: cardiovascular indication reviewed; interaction check against current TKI / ICI / targeted agent.
- Metformin: only if the diabetes indication is intact. Do not restart because of a cancer blog. eGFR / B12 noted.
- No investigational antiparasitic is discussed on this sheet. That is a different series, with a different evidence grade and an ASCO notice attached.
How this hub will be maintained
Living review. The evidence file is updated of an NCCN, ASCO, NEJM, Lancet, or JACC CardioOncology change that touches aspirin, metformin, statins, cimetidine, or incretin mimetics. Each update logs what moved tiers. Failed hypotheses stay on the page; they are how this brand stays early without becoming reckless.
Related
- Metabolic Cancer Series — 7-layer stack (this hub becomes Layer 1; investigational antiparasitics drop to Layer 3).
- Repurposed Drug Series — ivermectin / benzimidazoles, labeled investigational, ASCO notice attached.
Medical review: OneDayMD Editorial Board · Last updated 7 September 2026 · Version 1.0 hub · Next scheduled review: after any 2026 Q4 guideline drop.
OneDayMD Series 8 · onedaymd.com
.png)
Comments
Post a Comment