Peptides for a herniated disc: BPC-157, TB-500 and ARA-290 measured against a 70% spontaneous resorption rate

Seven in ten herniations shrink without anyone doing anything

A 2024 meta-analysis pooled 31 studies and 2,233 patients treated conservatively for lumbar disc herniation. The overall rate of spontaneous resorption was 70.39%. Broken out by herniation type: 87.77% for sequestration, 66.91% for extrusion, 37.53% for protrusion, 13.33% for a bulge. Most of it happened inside six months.

An earlier meta-analysis of 11 cohort studies put the pooled figure at 66.66%.

Two independent pools, seven years apart, landed within four percentage points of each other. That is the number every claim below has to be read against. If a compound is taken for twelve weeks and the leg pain resolves, the base rate says that is the most likely outcome anyway — and the bigger the herniation, the more likely it resorbs, because extruded and sequestered fragments are the ones that disappear most reliably.

The trial record says the same thing from a different direction. In a randomised trial of 283 patients with six to twelve weeks of severe sciatica, 95% of patients in both arms — early microdiscectomy and prolonged conservative care — reported recovery at one year. Surgery got them there faster. It did not get more of them there.

Disc bulges are common in people who feel nothing at all

A systematic review of imaging in 3,110 asymptomatic people found disc protrusions in 29% of 20-year-olds and 43% of 80-year-olds, and disc degeneration in 37% and 96% respectively.

Two things follow. An MRI finding is not by itself the cause of a patient's pain. And a follow-up MRI showing a smaller herniation is not proof that anything was treated, because the herniation may never have been the pain generator and would probably have shrunk regardless.

A herniated disc is four problems wearing one name

Layer 1 — mechanical. A displaced fragment of nucleus pulposus occupies space it should not, and deforms a nerve root. A geometry problem.

Layer 2 — inflammatory. Nucleus pulposus is immunologically sequestered inside a healthy disc. Once it leaks out it behaves as foreign tissue: macrophages arrive, TNF-α and other cytokines rise, and the nerve root becomes chemically irritated. This layer is why some people with tiny herniations are in agony and some with enormous ones barely notice.

Layer 3 — neural. Sustained compression and sustained chemical exposure damage the nerve fibres themselves. Numbness, burning and weakness that persist after the mechanical problem resolves belong here.

Layer 4 — structural. The annular tear, the surrounding ligament and muscle, and the disc's own matrix are damaged tissue in a near-avascular structure that heals slowly.

The stack argument runs: layers 2, 3 and 4 have separate biology, no existing conservative treatment addresses layer 4 at all, and three peptides with three different mechanisms can therefore be combined without redundancy. Layer 1 is conceded to surgery or to time.

The layer-to-compound matrix

No peptide moves a displaced fragment, and nothing in the literature claims otherwise

Layer 1 is mechanical, and there is no proposed mechanism by which a fifteen-amino-acid peptide relocates extruded disc material. Resorption happens — at the rates above — through macrophage-mediated phagocytosis, neovascularisation of the fragment and dehydration, over months. Surgery removes the fragment in an hour.

That is the honest boundary. Everything below concerns pain, nerve function and soft-tissue repair, not disc geometry.

The chemistry hurts more than the squeeze, which is where the mechanism is strongest and the human record is worst

Layer 2 carries the best mechanistic case for a peptide stack and also the most instructive failures.

The mechanism is established. Leaked nucleus pulposus provokes a TNF-α-driven inflammatory response at the nerve root, and animal models show that TNF blockade delivered epidurally shortens recovery from radicular allodynia.

Then the human trials arrived. FIRST II randomised 40 patients with MRI-confirmed disc herniation and severe sciatica — all of them surgical candidates — to a single infusion of infliximab 5 mg/kg or placebo. Both groups improved substantially. Neither beat the other. Seven patients in each arm went on to discectomy anyway.

A three-arm trial of 84 adults compared epidural steroids, epidural etanercept and epidural saline in subacute lumbosacral radiculopathy. Steroids beat saline by 1.26 points on the leg pain measure, which did not reach significance. Etanercept fared worse than steroids on functional capacity, by a margin that did.

One transforaminal etanercept trial in 49 patients did find separation — but only at the lowest of three doses, 0.5 mg, with the 2.5 mg and 12.5 mg arms not separating, and at a significance threshold of P < 0.1.

Read together: targeted, potent, pharmaceutical-grade TNF-α blockade — the strongest available test of the inflammatory hypothesis as a treatment target — produced a null result in the best-controlled trial, an inverse dose-response in the one positive trial, and worse function than steroids head to head. That is the ceiling on the inflammatory argument. BPC-157 and TB-500 are proposed as weaker, non-specific, unmeasured modulators of a pathway that a monoclonal antibody could not reliably move.

The nerve layer has real randomised evidence, in a disease nobody reading this has

ARA-290, generic name cibinetide, is an eleven-amino-acid peptide derived from erythropoietin that activates the innate repair receptor without stimulating red cell production. Its human record is the strongest of the three, and none of it is in spines.

A phase 2 pilot randomised 22 sarcoidosis patients with small-fibre neuropathy to intravenous ARA-290 or placebo for four weeks. The treated group improved significantly more on the small-fibre neuropathy screening list. Pain intensity on the Brief Pain Inventory improved in both groups, equivalently.

A phase 2b trial of 64 patients tested three doses against placebo over 28 days, with corneal nerve fibre area as the primary endpoint. The 4 mg arm gained 697 μm² over placebo (95% CI 159 to 1,236; P = 0.012). Regenerating intraepidermal GAP-43+ fibres increased in the same arm. Pain improved in every arm including placebo, and the placebo-corrected pain difference in the 4 mg group did not reach significance (P = 0.157).

What that record supports: ARA-290 measurably regrows small nerve fibres in a metabolic and inflammatory neuropathy. What it does not support: any claim about a nerve root under mechanical compression, which is a different injury with a different cause and an established fix — removing the compression. The pain endpoints were the weakest part of both trials.

The structural claim is animal data, and no animal has ever had one of these peptides tested on a disc

BPC-157's most cited spine result is a rat spinal cord compression model. A single intraperitoneal injection after a 60-second compression of the sacrocaudal cord produced better motor recovery, fewer lost motoneurons and less demyelination, followed out to 360 days.

TB-500's disc evidence is one in-vitro study: exogenous thymosin β4 reduced apoptosis in cultured human intervertebral disc cells.

A 2026 review mapping the thymosin β4 and TB-500 literature found spine and intervertebral disc work limited to in-vitro studies.

Neither compound has been given to an animal with an induced disc herniation and measured against a control for resorption rate, annular integrity, disc height or radicular pain behaviour. Not once. The full mechanistic case, dosing record, pharmacokinetics and safety data for each are set out separately.

The only published test of the two together found the combination beat neither one alone

Thirty-two rats underwent standardised Achilles tendon transection and repair, then four weeks of intraperitoneal treatment across four arms: control, BPC-157 at 10 µg/kg/day, TB-500 at 60 µg/kg/day, and both together.

TB-500 alone reached significance on maximum load to failure. TB-500 alone and the combination both lowered total Movin scores against control. BPC-157 alone was numerically better than control without reaching significance on total scores. The combination did not beat either single agent.

The authors' own reading is that the two peptides may converge on shared downstream pathways. That is a direct hit on the stack rationale, delivered by the only experiment that has ever tested the pair in any tissue.

No trial has tested these three compounds together, in a disc, in anything

Searched: PubMed and ClinicalTrials.gov. Result: zero completed or ongoing trials of BPC-157, TB-500 or ARA-290 in disc herniation, radiculopathy or intervertebral disc repair, in humans or in animals — and zero trials of any two of them in combination outside the rat Achilles study above.

The registry record for BPC-157 in humans is one phase 2 trial, recruiting, in acute hamstring strain.

clinicaltrials.gov

ClinicalTrials.gov NCT07437547 — BPC 157 for Acute Hamstring Muscle Strain Repair (Phase 2, recruiting)

BPC 157 for Acute Hamstring Muscle Strain Repair — Phase 2 — Recruiting.

↗ clinicaltrials.gov

2026-08-04 08:28s33 · #56aaa08a0137

ARA-290's completed phase 2 registrations sit in sarcoidosis neuropathy, type 2 diabetes and diabetic macular oedema, the last of which was terminated.

clinicaltrials.gov

ClinicalTrials.gov NCT02039687 — Efficacy of ARA 290 on Corneal Nerve Fiber Density and Neuropathic Symptoms in Sarcoidosis (Phase 2, completed)

Study of Efficacy of ARA 290 on Corneal Nerve Fiber Density and Neuropathic Symptoms of Subjects With Sarcoidosis — Phase 2 — Completed.

↗ clinicaltrials.gov

2026-08-04 08:28s34 · #4acab3ad5e43

The three-compound stack for a herniated disc is a mechanistic hypothesis assembled out of adjacent literatures. It has never been the subject of an experiment. A protocol presented with expected outcomes is extrapolating across a species boundary, a tissue boundary and a disease boundary simultaneously.

What the alternatives deliver, and what each one costs

cost.sidecarhealth.com

Sidecar Health — cost of an epidural steroid injection in the United States

Published US price range for a single epidural steroid injection, with the Medicare allowable near $161 and cash prices commonly running from several hundred to several thousand dollars depending on facility, imaging guidance and region.

↗ cost.sidecarhealth.com

2026-08-04 08:33s42 · #80acf6e58858

The epidural steroid number deserves a second look, because it is the closest analogue to what the peptide stack proposes — an anti-inflammatory delivered at the nerve root. A 6.2-point improvement on a 0–100 leg pain scale falls below every proposed threshold for a clinically important change, which run from 10 to 30 points. The most established anti-inflammatory intervention for this exact condition produces a short-term effect smaller than patients can reliably notice, and no durable one at twelve months.

Why one person's twelve-week result cannot settle anything

Take a patient with an extruded L5/S1 herniation and severe radicular leg pain. The published resorption rate for extrusions is 66.91%, most of it inside six months. The one-year perceived-recovery rate under conservative care is 95%.

Run a twelve-week peptide protocol. Suppose the pain resolves and a repeat MRI shows the fragment has shrunk.

The arithmetic: roughly two thirds of such patients show fragment shrinkage in that window with no treatment at all, and the overwhelming majority report recovery within a year. A single positive case therefore carries close to zero information. Separating a real effect from that base rate needs randomisation, imaging read blind to allocation, and enough patients to detect a difference against a control arm already recovering at 66% to 70%.

That is why the reports below are labelled anecdotal and left as anecdotes, including the persuasive ones.

What people who have tried it report, positive and negative

Two patterns are worth naming. The reports describing relief describe pain relief on a two-to-six-week timescale, which is also the timescale of natural inflammatory resolution, and none of them include a blinded or controlled comparison; one of them says so explicitly. The reports describing failure come from people who used the same compounds at similar doses for tendon and joint problems and noticed nothing across repeated attempts. Both sets are single-person observations without controls, and neither is evidence of anything beyond what those people experienced.

In US law none of this is a medicine

BPC-157 and thymosin beta-4 fragment (LKKTETQ) — the sequence sold as TB-500 — both appear in the FDA's Category 2 bulk drug substances record, the list of substances that may present significant safety risks when compounded. The stated reasons are immunogenicity risk, peptide-related impurities and characterisation difficulty. For TB-500 the agency states it has identified no human exposure data at all.

Statement

Official disclosure

FDA — Certain Bulk Drug Substances for Use in Compounding that May Present Significant Safety Risks (current as of 22 April 2026)

“Compounded drugs containing BPC-157 may pose risk for immunogenicity for certain routes of administration and may have complexities with regard to peptide-related impurities and active pharmaceutical ingredient (API) characterization. FDA has identified no, or only limited, safety-related information for the proposed routes of administration.”

Aug 4, 2026 · fda.gov

2026-08-04 08:28s32 · #6f5783547545

The practical consequences:

• Neither compound is on the 503A or 503B bulks lists, so a US compounding pharmacy cannot lawfully prepare either one.
• Every remaining US supply channel is a research-chemical vendor selling under a research-use-only label, outside pharmaceutical manufacturing controls and outside any sterility or identity guarantee a patient could rely on.
• ARA-290 is a clinical-stage investigational drug that was never approved and is not commercially available; material sold under that name is unverifiable.
• BPC-157 and TB-500 are both prohibited under the World Anti-Doping Code, which makes any competing athlete a strict-liability problem independent of whether the compounds work.

None of that is a claim about efficacy. It is the supply and legal reality sitting underneath any decision about them.

Twelve weeks, one patient, with the clock and the meter running

A 41-year-old with an MRI-confirmed L5/S1 extrusion and left-sided radicular leg pain at 7/10, four weeks after onset. Dollar figures are US cash prices or common allowed amounts; hours are the patient's own time.

Week 0 — what exists. Leg pain 7/10, back pain 4/10, Oswestry Disability Index 44. No motor deficit, no bowel or bladder involvement, no saddle anaesthesia. MRI already obtained: $400–$1,800 cash, $250–$400 typical Medicare allowable, 1 hour. Four baseline measures recorded: leg pain numeric rating, ODI score, straight leg raise angle in degrees, single-leg heel raise count on the affected side. Cost of measuring: $0 and 20 minutes.

Week 0 decision point. No red flags, so the fork is conservative care against early surgery. The randomised evidence says both reach 95% recovery at one year and surgery gets there faster. Choosing conservative care buys a slower course, not a worse one.

Weeks 1–2 — load management. Two physical therapy visits a week: 4 visits at $75–$150 = $300–$600, 4 hours in clinic plus 3.5 hours of home exercise. NSAIDs as tolerated: about $12 for the fortnight. Expected trajectory from the base rate: leg pain unchanged or slightly worse in week 1, beginning to ease in week 2.

Week 2 — first measurement gate. Re-record all four measures. The failure condition here is not "still in pain". It is new motor weakness, loss of bladder or bowel control, or progressive numbness in a saddle distribution. Any of those ends the conservative track that day and becomes a surgical referral.

Weeks 3–6 — where a peptide protocol would sit if one is run. At research-vendor pricing, BPC-157 5 mg runs $45–$60 a vial and TB-500 starts near $25 a vial; a twelve-week two-compound course lands between $150 and $400 in materials, plus bacteriostatic water, syringes and swabs at roughly $30. Time cost is about 10 minutes a day, or 4.7 hours across twelve weeks. There is no established dose for this indication, because no dose has ever been tested for this indication, and neither compound is approved for human use in the United States. What is run here is unlicensed self-experimentation, and the expected value of the spend cannot be estimated from the published record, because the published record for this condition is empty.

thepeptidecatalog.com

Where to Buy BPC-157: Price and Vendor Guide (2026)

Legitimate vendors with proper documentation typically charge $5-$8/mg. Since the FDA moved BPC-157 to the 503A Category 2 bulks list, compounding pharmacies can no longer produce it, and every legitimate source is now a research peptide vendor.

↗ thepeptidecatalog.com

2026-08-04 08:33s43 · #7bb9093ec80f

Week 6 — second measurement gate. Re-record the four measures. The published inflection points are six months for resorption and two to eight weeks for the inflammatory peak to settle. A usable decision rule: leg pain down at least 30% from baseline continues the current course; less than 30% moves to escalation.

Weeks 6–8, escalation if that gate fails — epidural steroid injection. One transforaminal injection: $800–$5,000 cash, $161 Medicare allowable, 3 hours including travel and recovery. Expected effect from the pooled data: about 6 points on a 0–100 leg pain scale short term, below the 10-to-30-point range proposed as clinically important, and not durable at twelve months. It is a real option with a small honest effect size, and it is the cleanest available test of whether the pain is inflammatory rather than mechanical.

Weeks 8–12 — continued rehabilitation. 8 further physical therapy visits at $75–$150 = $600–$1,200, 8 clinic hours plus 7 hours of home work.

Week 12 — the state at the end. Conservative-track running total: $912–$1,812 in physical therapy, about $70 in medication, $0–$5,000 if an injection was used, plus $180–$430 if a peptide protocol was run. Patient time: roughly 28 hours. Base-rate expectation: substantially improved leg pain in the majority, an ongoing but declining course in a large minority, and a repeat MRI showing fragment reduction in about two thirds if one is taken.

Week 12 decision point. Persistent disabling radicular pain past twelve weeks, with imaging that matches the symptoms, is where microdiscectomy stops being premature. At $15,000–$50,000 list — $14,137 above non-operative care over two years in the trial data — it buys faster relief, not a higher chance of eventual recovery.

What this walk-through cannot establish. Whether the peptides did anything. One patient, no control, a 66.91% base rate for extrusion resorption: the twelve-week outcome is equally compatible with a large effect, no effect, and a harmful effect. That is a property of the design, not of the compounds.

Who the decomposition is useful to

Chiropractic offices, physical therapy clinics and sports medicine practices field this question directly, because patients arrive having already read a vendor page and want a verdict. The layer breakdown gives a practice something specific to say instead of an endorsement or a brush-off: the mechanical layer is time or surgery, the inflammatory layer has a real mechanism and a discouraging human trial record, the neural layer has genuine randomised data in a different disease, and the structural layer is animal data only.

It also fixes the referral boundary. Progressive motor weakness, bowel or bladder change, and saddle anaesthesia are surgical emergencies, and no conversation about peptides belongs near them.

What would change the answer

Four experiments would move the matrix, in ascending order of cost:

1. A rodent lumbar disc herniation model — autologous nucleus pulposus implantation is the standard preparation, and it is the model already used in the radicular pain literature — with BPC-157, TB-500, both, and vehicle, measuring mechanical allodynia thresholds and herniation volume. This is the missing first experiment and it is cheap.
2. A dose-response study for either peptide in any spinal tissue. Every dose in circulation was chosen by analogy to the tendon and gut literature.
3. A randomised, imaging-blinded human trial in radiculopathy with a conservative-care control arm, powered against a 66% to 70% control resorption rate, with leg pain and ODI as primary endpoints and fragment volume as a secondary.
4. A trial of ARA-290 in compressive radiculopathy — the only one of the three with a human safety database large enough to justify starting one.

Until at least the first of those exists, the layer-to-compound matrix has three empty cells in its evidence column, and 70.39% is the number any future result will be measured against.

The sibling objects for this page, each one inspectable on its own terms: