RecoveryEvidence ReviewAug 2026

Best Peptides for Inflammation in 2026

Inflammation is a finding, not a diagnosis, and a diagnosed inflammatory condition is worked up and treated with a doctor. What follows is the published record study by study: who was enrolled, what endpoint was measured, what number came back.

10,061

Patients randomised in the canakinumab trial (PMID 28845751)

1,106

Adults with sepsis randomised in the thymosin alpha-1 phase 3 (PMID 39814420)

342

Adults with celiac disease randomised in the larazotide trial (PMID 25683116)

Start Here

Read this first: rheumatoid arthritis, inflammatory bowel disease, psoriasis, lupus and sarcoidosis are diagnosed and treated by a clinician. Nothing here is a protocol, a recommendation, or a substitute for that. A result in people with sepsis or celiac disease is a result in those people, and it reads differently once it is applied to someone without the diagnosis.

The word does two jobs. In a clinic it names a measurable immune process with numbers attached. Online it describes feeling puffy, sore and tired. A compound that shifts a blood marker may do nothing for the second.

What the Trials Measured

When a trial reports lower inflammation, one blood measurement usually moved. C-reactive protein, reported as hs-CRP in its high sensitivity form, rises with immune activity. Interleukin-6 drives CRP production. TNF-alpha is a separate signal. They do not have to move together.

A measured baseline makes any of it legible. The canakinumab trial enrolled people with hs-CRP of 2 mg per litre or more, treating that as residual inflammatory risk (PMID 28845751). A blood panel gives that baseline before anything else is decided.

Study by Study

One row per study. Milligram figures in the result column are trial arms, and the block underneath gives them with their source.

Compound	Population (n)	What was measured	Result	Design	PMID
Canakinumab (monoclonal antibody)	10,061 patients with a previous myocardial infarction, hs-CRP 2 mg per litre or more	Nonfatal myocardial infarction, nonfatal stroke or cardiovascular death	Hazard ratio 0.85 (95% CI 0.74 to 0.98, P = 0.021) at 150 mg; median hs-CRP fell 37 percentage points more than placebo at 48 months	Randomised, double-blind	28845751
GLP-1 receptor agonists	1,878 adults with type 2 diabetes, 25 studies (CRP from 18 of them)	C-reactive protein	SMD -0.39 (95% CI -0.72 to -0.06, P = 0.02, I2 = 88%)	Systematic review and meta-analysis, 2026	41660088
GLP-1 and dual GIP/GLP-1 agonists	1,991 adults with type 2 diabetes, 27 of 41 randomised trials	C-reactive protein or hs-CRP	SMD -0.37 (95% CI -0.59 to -0.14)	Meta-analysis of randomised trials, 2026	42449480
Semaglutide	17,604 adults aged 45 or over with cardiovascular disease and a body-mass index of 27 or more, no diabetes	Cardiovascular death, nonfatal myocardial infarction or nonfatal stroke	6.5% against 8.0%, hazard ratio 0.80 (95% CI 0.72 to 0.90, P less than 0.001), mean 39.8 months	Randomised, double-blind (SELECT)	37952131
Thymosin alpha-1	1,106 adults with sepsis, 22 centres in China	28 day all cause mortality	Hazard ratio 0.97 (95% CI 0.76 to 1.24) as corrected on 30 May 2025 (PMID 40447307)	Randomised, double-blind, phase 3 (TESTS)	39814420
Thymosin alpha-1	361 patients with severe sepsis, six teaching hospitals in China	28 day all cause mortality	26.0% against 35.0%, relative risk 0.74 (95% CI 0.54 to 1.02); nonstratified P = 0.062, log rank P = 0.049	Randomised, single-blind (ETASS)	23327199
Cibinetide (ARA 290)	64 subjects with sarcoidosis-associated small nerve fibre loss and neuropathic pain	Change in corneal nerve fibre area at day 28	Placebo-corrected 697 square micrometres (95% CI 159 to 1236, P = 0.012) at 4 mg; 109 and 431, both non-significant, at 1 and 8 mg	Randomised, phase 2b, 28 days	28475703
Thymosin beta-4 (RGN-259 eye drops)	72 subjects with moderate to severe dry eye, single centre	Ocular discomfort and inferior corneal staining at day 29 (primary)	Neither primary endpoint separated from placebo; day 28 discomfort in the environment challenge fell 27% (P = 0.0244)	Randomised, double-masked, phase 2	26056426
Thymosin beta-4 (RGN-259 eye drops)	9 patients with severe dry eye, two US sites	Ocular discomfort and total corneal fluorescein staining at day 56	35.1% lower discomfort (P = 0.0141) and 59.1% less staining (P = 0.0108) against vehicle control	Randomised, double-masked, phase 2	25826322
Icatibant	88 subjects in an acute hereditary angioedema attack, cutaneous or abdominal (icatibant 43, placebo 45)	Time to a 50% or greater reduction in symptom severity	Median 2.0 against 19.8 hours (P less than 0.001); in a separate 5 subject laryngeal group (icatibant 3, placebo 2), 2.5 against 3.2 hours	Randomised, double-blind, phase 3 (FAST-3)	22123383
Ziconotide	220 adults with severe chronic pain (112 against 108), mean baseline pain intensity 80.7	Visual analogue pain intensity, mean percentage improvement to week 3	14.7% against 7.2% (P = 0.036); dizziness, confusion, ataxia, abnormal gait and memory impairment reported as significant adverse events	Randomised, double-blind, intrathecal	16716870
Larazotide acetate	342 adults with celiac disease still symptomatic after 12 months or more gluten-free	Celiac Disease Gastrointestinal Symptom Rating Scale, average on treatment	Primary endpoint met at 0.5 mg (analysis of covariance P = 0.022); 1 mg and 2 mg arms did not differ from placebo	Randomised, double-blind, 12 weeks	25683116
Afamelanotide (alpha-MSH analogue)	94 patients (US study) and 74 (EU study) with erythropoietic protoporphyria	Hours of direct sunlight exposure without pain	US at 6 months: median 69.4 against 40.8 hours (P = 0.04). EU at 9 months: 6.0 against 0.8 hours (P = 0.005)	Two randomised, double-blind trials reported together	26132941
Repository corticotropin (porcine ACTH analogue)	18 healthy adults, 9 per treatment sequence	Granulocyte and lymphocyte counts, total cortisol-equivalent exposure	Granulocytes rose and lymphocytes fell on both arms, significantly less so on the ACTH analogue; 80 units equated to 30 mg methylprednisolone by cortisol-equivalent exposure	Randomised, crossover, open-label	26120075
BPC-157	Rats with a transected right Achilles tendon, group sizes not stated in the abstract	Load to failure, functional index, cell counts on histology	Higher load to failure and Young modulus, higher functional index, more mononuclear cells and fewer granulocytes than saline	Rat model, intraperitoneal, assessed to day 14	14554208
KPV (lysine-proline-valine)	Mice, dextran sodium sulfate colitis and CD45RB-high transfer colitis, group sizes not stated in the abstract	Body weight, histological infiltrate, colonic myeloperoxidase activity	Earlier recovery and stronger weight regain; infiltrates and myeloperoxidase activity significantly reduced	Two mouse models	18092346
KPV in hyaluronic acid nanoparticles	Mice with ulcerative colitis, group sizes not stated in the abstract	Mucosal damage and TNF-alpha	Oral nanoparticles in a hydrogel prevented mucosal damage and downregulated TNF-alpha more than the system without hyaluronic acid	Mouse model, oral delivery	28143741

Vial quality is a separate question, and the certificate of analysis guide covers how to check what is in a vial.

Trial Doses as Published

Figures as published by each trial:

Canakinumab trial (Ridker et al., PMID 28845751): 50 mg, 150 mg or 300 mg subcutaneously every three months, or placebo.
SELECT (Lincoff et al., PMID 37952131): semaglutide 2.4 mg subcutaneously once weekly, or placebo.
Cibinetide phase 2b (Culver et al., PMID 28475703): 1, 4 or 8 mg per day for 28 days, or placebo.
Larazotide acetate trial (Leffler et al., PMID 25683116): 0.5, 1 or 2 mg three times daily for 12 weeks, or placebo.
Afamelanotide trials (Langendonk et al., PMID 26132941): a subcutaneous implant containing 16 mg every 60 days, or placebo.
Ziconotide 301 (Rauck et al., PMID 16716870): intrathecal infusion from 0.1 microgram per hour, raised over three weeks, mean dose at termination 0.29 microgram per hour.
RGN-259 phase 2 (Sosne et al., PMID 26056426): thymosin beta-4 ophthalmic solution 0.1 percent as eye drops for 28 days, or placebo.
Repository corticotropin study (Lal et al., PMID 26120075): 80 units subcutaneously daily for five days, compared with 1 g intravenous methylprednisolone.

Disclaimer: these are figures recorded from published trials, several given under specialist supervision, one infused into spinal fluid through an implanted pump, one an eye drop. They are not doses for anyone to take, not a schedule, and not advice. Any dosing decision belongs with a prescriber.

Trials in a Named Diagnosis

The diagnosis is part of the result. A trial in hereditary angioedema, one in celiac disease and one in sarcoidosis-associated nerve fibre loss each measured something their own condition made measurable, and a result in one is not a result in another. Larazotide acetate is one to sit with, because its 0.5 mg arm met the primary endpoint while its higher arms did not (PMID 25683116), which the intestinal permeability guide gives more context for.

Afamelanotide, an alpha-melanocyte-stimulating hormone analogue, was measured in hours of sunlight without pain (PMID 26132941), an endpoint built around its condition. Background sits on the thymosin alpha-1 profile, the ARA 290 profile and the TB-500 profile, which keeps the dry eye drops separate from injected use.

BPC-157 and KPV

BPC-157. A 2026 narrative review in Pharmaceutics, searching PubMed, Embase and the Cochrane Library to April 2026, reports the available clinical data as fewer than 30 subjects across three uncontrolled pilot studies, none using standardised pharmaceutical preparations, and a plasma half-life under 30 minutes (PMID 42198317). The BPC-157 profile and the Wolverine stack guide cover how it is discussed in practice.

KPV. The tripeptide lysine-proline-valine is a fragment of alpha-melanocyte-stimulating hormone, and its colitis results were measured in mice (PMID 18092346, PMID 28143741). The search log records what those queries returned. See the KPV profile.

Pain as an Endpoint

A search for a peptide for inflammation and a search for a peptide for pain can land on different trials, because the compounds were tested against different endpoints. A 2022 review of venom peptide calcium channel blockers describes ziconotide in clinical use for chronic intractable pain and notes that its adverse effects and its intrathecal route limit how many patients can receive it (PMID 34259147).

The cibinetide trial measured both kinds of endpoint and they moved apart: nerve fibre area separated from placebo at 4 mg, P = 0.012, while pain improved in every group including placebo and the placebo-corrected decrease in that arm was reported at P = 0.157 (PMID 28475703). For everyday joint pain the trialled options sit in the joint pain supplements guide and the omega-3 guide.

When to See a Doctor

Joint swelling, morning stiffness lasting over an hour, or symmetrical joint pain.
Unexplained weight loss, fever, night sweats, or a rash alongside joint pain.
Blood in stool, persistent diarrhoea, or recurring abdominal pain.
Pain that wakes you at night or worsens steadily rather than fluctuating.
An hs-CRP that stays high on repeat testing with no explanation.

Each item is a reason to be examined. The cost of self-managing a treatable autoimmune disease is the damage accumulating while it goes untreated. Wider framing sits in the therapeutic peptides guide and the arthritis evidence guide.

Questions

Why does one trial report heart attacks and another report a blood marker?+

They chose different endpoints, and the two are not interchangeable. An event endpoint counts things that happened to people, such as a heart attack, a stroke or a death, which takes thousands of participants and years of follow-up to accumulate (PMID 28845751, PMID 37952131). A marker endpoint reads a value out of a blood sample, which a much smaller and shorter study can do. A marker moving is a reason to run the larger trial, not a substitute for having run it.

Does a result in a disease population apply to a healthy adult?+

That is a separate question from the one the trial answered. A trial measures what happened in the people it enrolled, on the dose, route and duration it used, and someone without that diagnosis differs on several of those at once. Investigators say as much about the reverse direction: the repository corticotropin crossover, run in healthy volunteers, concluded that the clinical relevance of its findings in autoimmune disease populations is unknown and requires further evaluation (PMID 26120075). The gap does not close by running the inference the other way.

Why do trials of the same kind of compound use such different routes?+

The route is chosen for the tissue the trial is aiming at, and it becomes part of the result. Thymosin beta-4 was given as an eye drop to reach the ocular surface (PMID 26056426). Ziconotide was delivered intrathecally, to reach the spinal cord directly (PMID 16716870). KPV was packaged into hyaluronic acid nanoparticles inside a hydrogel so that an oral dose could survive the gut and reach colon tissue in mice (PMID 28143741). A result obtained by one route says little about another.

How much does a rodent result tell you about a human one?+

It tells you what happened in that model. The injury or the colitis was induced deliberately, and the dose, route and timing were set by the researchers. A person arriving in a clinic matches little of that. Rodent work is where a hypothesis is generated and where a range to test in people is first guessed at, which is a different job from showing that something works in people.

Searches Run for This Page

Each line records a query, the index, the date and the count returned. A count is a fact about a search on a date, not a statement about the world: a compound-name search cannot see synonyms, MeSH entry terms or work held outside PubMed.

("BPC 157"[All Fields] OR "BPC-157"[All Fields] OR "body protection compound 157"[All Fields] OR "pentadecapeptide BPC 157"[All Fields]) AND randomized controlled trial[pt] in PubMed on 9 August 2026 returned 0 records.
("BPC 157"[All Fields] OR "BPC-157"[All Fields]) in PubMed on 9 August 2026 returned 225 records.
(KPV[ti] OR "lysine-proline-valine"[ti] OR "lysine proline valine"[ti] OR "alpha-MSH(11-13)"[All Fields]) AND (randomized[tiab] OR randomised[tiab] OR clinical trial[pt]) in PubMed on 9 August 2026 returned 0 records.
thymosin beta 4[All Fields] AND randomized controlled trial[pt] in PubMed on 9 August 2026 returned 7 records.