ARA-290's largest trial met its primary endpoint. The endpoint was nerve fibers in the eye.

Why this note exists

Most of the compounds pepmg tracks fail an evidence audit in the same boring way: the human trials do not exist. ARA-290 fails differently, and more interestingly. The trials exist, they were properly randomized and masked, they were run with academic neurologists and anesthesiologists at Leiden University Medical Center, the Cleveland Clinic and Queen's University Belfast, and their results were published in indexed journals.[1][4][10] The question worth asking is not whether the evidence is real. It is what the real evidence actually measured.

In the index generated on August 15, 2026, 46 of 112 vendors carried an ARA-290 listing — 60 listings, twenty-eighth of 83 compounds by vendor coverage, 49 of them in stock.[19] Three of the 60 name cibinetide anywhere in the product title, one of those spelling it "Cibinitide."[19] Advertised sizes run from 5 mg to 160 mg, with 10 mg the most common at 28 listings and 16 mg next at seven — though the largest figures are multi-vial kit totals rather than single vials.[19] Every dose ever published in a human trial of this compound falls between 1 mg and 8 mg per day.[1][4][10]

A note on what kind of evidence this is: every participant count, route, endpoint and P value below comes from a human trial, from the results its sponsor posted to a public registry, or from a registry record — and each is reported with its design. This note contains no animal data at all, which is itself worth saying, because the great majority of the ARA-290 literature published since 2017 is in mice, rats and cell culture and none of it is used here to support a claim about people. Doses appear as their sources published them, with route and duration intact.

The anchor trial

Sixty-four people, four arms, twenty-eight days

Study NCT02039687 is the study. It ran from January 2014, a double-blind, placebo-controlled, parallel-group, dose-ranging phase 2 in subjects with sarcoidosis-associated small nerve fiber loss and neuropathic pain, with masking recorded in the registry as quadruple — participant, care provider, investigator and outcomes assessor.[2] Sixty-four people were randomized, 16 to each of 1 mg, 4 mg, 8 mg per day of subcutaneous cibinetide and 16 to placebo, for 28 consecutive days.[2][3] Sixty completed; one withdrew, one left for an adverse event, and two could not have their treatment assignment confirmed.[3] That is a well-designed small trial, and small trials are where most drugs that later work begin.

Its pre-specified primary endpoint was the change from baseline in corneal nerve fiber area at day 28, measured by corneal confocal microscopy — a non-invasive imaging technique that counts and sizes the small nerve fibers in the front of the eye.[2] The published result: placebo-corrected mean change of 109 µm² (95% CI −429 to 647) at 1 mg, 697 µm² (95% CI 159 to 1236; P = 0.012) at 4 mg, and 431 µm² (95% CI −130 to 992) at 8 mg.[1] Regenerating intraepidermal fibers, identified by GAP-43 staining, also increased in the 4 mg group (P = 0.035), and changes in corneal nerve fiber area correlated with changes in GAP-43 fibers (ρ = 0.575; P = 0.025) and with the six-minute walk test (ρ = 0.645; P = 0.009).[1]

Take that result at face value: at one dose out of three, over four weeks, a masked trial found more nerve fiber in the cornea than placebo did, and the effect tracked with two other measures. That is a real signal and this note does not discount it. What it is not is evidence that anyone felt better — and the same trial has something to say about that, because the sponsor posted its results.

What the sponsor posted

On the measures patients feel, placebo was not behind

ClinicalTrials.gov holds sponsor-posted results for this trial: unadjusted mean changes from baseline for each arm, with standard deviations and no P values.[3] They are worth reading next to the paper, because they cover the outcomes a person taking the drug would notice.

Unadjusted mean change from baseline to day 28 as posted by the sponsor. The registry publishes no P values for these comparisons, so they are raw numbers, not a statistical test — and with 13 to 16 people per arm the confidence intervals around any of them are wide.[3]

The pattern holds on the questionnaires too. On the Small Fiber Neuropathy Screening List the 4 mg arm improved by 8.7 points against placebo's 7.3; on the Neuropathic Pain Symptom Inventory, 14.50 against placebo's 12.88; on the Fatigue Assessment Scale, 2.1 against placebo's 2.2, with the 8 mg arm at 3.1.[3] On Brief Pain Inventory pain severity the 4 mg arm improved by 1.167 against placebo's 0.922.[3] None of these is a large separation, several run the wrong way, and the paper states the situation plainly in its own abstract: "Pain improved significantly in all groups."[1]

The pain result most often quoted for this compound needs its P value attached. The paper reports that subjects with moderate to severe pain at baseline showed "a clinically meaningful placebo-corrected decrease in pain intensity in the 4 mg group (P = 0.157)."[1] The developer's press release repeated the "clinically meaningful" half without the number.[13] A P value of 0.157 in a subgroup of a 64-person trial is a result that did not reach conventional significance, and describing it as clinically meaningful is an interpretation, not a finding.

One more asymmetry is worth naming, because it cuts against a simple story. Nerve fiber was measured two ways: in the cornea, where 4 mg beat placebo, and in skin biopsies, where placebo beat every dose on total intraepidermal nerve fiber density.[1][3] The paper's positive skin finding is for GAP-43-positive regenerating fibers specifically, a different and narrower measure than the total density in the registry table.[1] Both can be true. But a compound whose case rests on nerve regrowth showing regrowth on one imaging measure and not on the standard biopsy count is a reason for a phase 3, not a substitute for one.

The rest of the human record

Four more studies, none larger than 64

The published human literature on this compound is small enough to list in full.

2012, the pilot. Twenty-two patients with sarcoidosis and symptoms of small fiber neuropathy, randomized double-blind to intravenous ARA 290 at 2 mg three times weekly (n = 12) or placebo (n = 10) for four weeks. The Small Fiber Neuropathy Screening List improved more on drug (Δ −11.5 ± 3.04 versus Δ −2.9 ± 3.34 standard error, p < 0.05), as did the pain and physical functioning dimensions of the SF-36. And then the same sentence that recurs throughout this record: "The mean BPI and FAS scores improved significantly but equivalently in both patient groups." No safety concerns were raised.[4]

2013, the follow-up. A blinded, placebo-controlled trial of 28 days of daily subcutaneous ARA 290 in patients with documented small nerve fiber loss, reporting improved neuropathic symptoms, increased corneal small nerve fiber density, changed cutaneous temperature sensitivity and increased six-minute walk distance. The abstract does not state the number randomized; the paper carries a published erratum.[5]

2015, type 2 diabetes. A phase 2 of 4 mg per day subcutaneously for 28 days with a further month of follow-up. Among the 42 subjects with complete serial HbA1c measurements — 21 per group — mean HbA1c changed by −0.16% at day 28 and −0.21% at day 56 on ARA 290, against −0.01% and +0.21% on placebo (P = 0.002, repeated-measures ANOVA).[6] Two things belong beside that. Baseline HbA1c was not balanced — 7.3% on drug against 6.9% on placebo — and the corneal nerve fiber density result was a subgroup analysis in subjects more than one standard deviation from normal, tested by paired t-test within each arm (+2.6 ± 1.0 fibers/mm², n = 18, p = 0.02 on drug; +0.7 ± 1.3, n = 19, not significant on placebo) rather than by a comparison between arms.[6] Significant in one arm and not in the other is not the same claim as significantly different between arms. The corresponding registry record has carried the status "unknown" since its last update in September 2015.[7]

2015, healthy volunteers. Thirty-six healthy participants received a single 2 mg dose of ARA290 or placebo in a double-blind, randomized, parallel-group design, assessed one week later for the neural and cognitive processing of emotion. The authors' own conclusion: the effects observed "do not unequivocally support an antidepressant-like profile for ARA290," and no effects were seen on mood or affective symptoms.[8][9]

2020, diabetic macular edema. A phase 2 at the Belfast Health and Social Care Trust: patients self-administered cibinetide 4 mg per day subcutaneously for 12 weeks, with no control arm and no masking.[10][11] Nine were recruited and eight completed. There was no improvement in best corrected visual acuity (−2.9 ± 5.0), central retinal thickness (10 ± 94.6 µm), central retinal sensitivity (−0.53 ± 1.9 dB) or tear production (−0.13 ± 7.7 mm). The one measure that moved was a patient-reported vision questionnaire, in an unmasked single-arm study, by 2.7 ± 3.1 points. The authors' conclusion is a safety conclusion: "The cibinetide 12-week course was safe."[10]

Where it stopped

The last trial ended because the drug expired

On May 8, 2017, a week after the phase 2b appeared in print, the developer announced the result, stated that "Cibinetide has been granted US and EU Orphan Drug Designation for the treatment of sarcoidosis" and US orphan drug and fast track designations for neuropathic pain in sarcoidosis, referred to a "recently-completed successful end-of-phase 2 meeting with the FDA," and said it looked forward "to taking the next step in clinical development."[13] Those designations are real regulatory instruments and they are worth exactly what they are: orphan designation confers development incentives and fast track confers earlier and more frequent FDA contact. Neither is an approval, and neither is a finding that a drug works.

Nine years later, a registry search on August 30, 2026 returns four interventional studies of this compound and no phase 3.[12] The most recently registered of the four is the Belfast macular edema study, and its record is the sharpest single fact in this note. It ran from April 2016, it is marked TERMINATED at nine participants, it was first posted to ClinicalTrials.gov in October 2024 — seven years after it stopped — and the reason it gives for stopping is a single line:[11]

Read that carefully, because it is easy to over-read. A trial ending because its drug supply ran out is a supply and funding event. It is not a safety signal, it is not a regulatory rejection, and the registry does not say why no replacement was available. What it does establish is that by 2017 the investigational supply of this compound had run down and was not restocked — and that no trial of it has begun since. A recent narrative review from Brigham and Women's Hospital, published on August 24, 2026, includes cibinetide among regenerative peptides of interest in chronic pain and reaches the same general conclusion this note does about the class: most remain unapproved by FDA and clinical evidence in humans is limited.[18]

The absence of a phase 3 is a fact about the program, not a verdict on the molecule. Drugs stall for money, for corporate reasons, for competing priorities. But a research market selling a compound in 2026 is selling it on the strength of a program that has not run a new trial in a decade, and buyers deserve to know that the sentence "it has clinical trials" and the sentence "it is in clinical development" are no longer the same sentence.

Who ran the studies

The record is small and it is largely one group's

This is stated as fact, not as an accusation: nearly every human study of ARA-290 lists employees or officers of the developer among its authors. The phase 2b paper carries four Araim Pharmaceuticals affiliations alongside the Cleveland Clinic, Leiden, Manchester and Weill Cornell Medicine-Qatar.[1] The 2013 and 2015 papers carry Araim affiliations.[5][6] The Belfast paper states its conflicts in the plainest available terms: four named authors "are officers and own stock in Araim Pharmaceuticals, which is developing cibinetide for clinical use," and a fifth is a consultant to the company.[10]

Sponsor involvement in trials of a sponsor's drug is normal and disclosed involvement is the system working. It matters here only because of scale: with four registered trials and no independent replication, there is no second group's result to check the first against. The one study in the set with a genuinely independent lead sponsor is the type 2 diabetes record, and it has never reported.[7] The measurement technique itself has been examined separately — a 2018 paper argues that corneal nerve fiber size adds diagnostic utility in small fiber neuropathy — but that is a paper about the endpoint, not an independent test of the drug.[17]

Sport

WADA named this compound, then named its mechanism

ARA-290 has an anti-doping history that most compounds in this index do not, and it is short and clear. The World Anti-Doping Agency's 2016 Prohibited List named it outright, at S2.1.2: "Non-erythropoietic EPO-Receptor agonists, e.g. ARA-290; asialo EPO; carbamylated EPO."[15]

The 2026 list no longer prints the name. Its equivalent entry, S2.1.5, reads "Innate repair receptor agonists, e.g. asialo EPO; carbamylated EPO (CEPO)" — and "innate repair receptor" is the term the compound's own developers coined for its target.[16][1] The S2 section opens by prohibiting the substances listed "and other substances with similar chemical structure or similar biological effect(s)," is in force in and out of competition, and contains only non-specified substances.[16] pepmg is not an anti-doping authority and does not issue eligibility rulings; an athlete's status question goes to their own anti-doping organization. The point here is narrower and is simply what the documents say: this compound was named on a prohibited list, and the category that replaced the naming is defined by its mechanism.

Three things this note is not saying

First, it is not saying the trials were bad. The phase 2b was randomized, quadruple-masked, placebo-controlled, dose-ranging and pre-specified its primary endpoint, and its sponsor posted results to a public registry when it was under no great pressure to do so.[2][3] That is better conduct than most of what this index touches.

Second, it is not saying the compound does nothing. A masked trial found more corneal nerve fiber at 4 mg than at placebo, with a confidence interval excluding zero, and all three doses walked farther on the six-minute walk test than placebo did.[1][3] Those are the findings a phase 3 exists to confirm or refute, and no phase 3 was run.[12]

Third, it is not saying the compound is unsafe. Across every published study — 22, 64, 42, 36 and 9 participants, at 2 to 8 mg, for four to twelve weeks — the reports describe no safety concerns, no serious adverse reactions and, in the Belfast study, no anti-cibinetide antibodies.[4][1][6][10] What that record cannot speak to is exposure beyond twelve weeks, because no published study has gone there.

What it is saying is narrow. ARA-290's evidence is a single positive surrogate endpoint, at one of three doses, in 16 people per arm, over 28 days, alongside patient-reported outcomes on which placebo did as well or better — and a development program that has not started a trial since April 2016. That is a reasonable basis for a phase 3 trial. It is a thin basis for anything else.

Questions people are asking