Skip to content

Peptide research

LL-37 wound-healing trials and the limits of immune-support claims

What LL-37 human trials show about wound healing, why topical and injected evidence differ, and what current FDA safety and compounding records say.

By Jabful · · 7 min read

This guide is educational, not a recommendation to use LL-37. It contains no dosing or treatment instructions. Sources checked November 11, 2026. No clinician review is claimed.

LL-37 has been tested in people. The important question is what those experiments can actually establish. A study of a peptide placed on a chronic wound does not automatically support claims about injections for whole-body immunity, infection prevention or recovery.

The human evidence includes encouraging early wound findings, a larger trial without an overall healing benefit, and a small diabetic-foot study with a positive tissue measure but several negative outcomes. There is also research outside wound care. Keeping those studies separate produces a more useful answer than either “clinically proven” or “never studied in humans.”

Start with the product and the route

LL-37 is a peptide derived from the human protein hCAP18. In the first venous-ulcer trial, researchers used a synthetic peptide with the same amino-acid sequence as the naturally occurring LL-37. The study product was applied to wounds alongside compression care, rather than injected for general wellness. Grönberg and colleagues, 2014

That distinction is fundamental. A molecule's role in the body is not itself proof that adding it as a drug improves health. Nor can the word “human” tell a reader whether a paper concerns human cells, measurements of a naturally present peptide, or people actually receiving an intervention.

For any LL-37 claim, identify three things before interpreting the result: the preparation, how it was administered, and the outcome being measured. Changing any of them changes the question that the evidence needs to answer.

An encouraging small trial was not the final answer

The 2014 randomized, double-blind study included 34 people with hard-to-heal venous leg ulcers. After a placebo run-in, participants entered a four-week treatment comparison. The lowest-strength LL-37 group had a significantly faster modeled healing rate than placebo. The middle-strength group's numerical advantage did not meet the conventional statistical threshold; the highest-strength group showed no benefit on that comparison. This was a small, early study whose main purpose was safety assessment. Original trial

A faster modeled reduction in wound area is not the same claim as reliable complete closure. Likewise, improvement relative to someone's own pretreatment period is a different comparison from improvement relative to placebo. Those distinctions matter when an impressive number moves from a paper into a benefit list.

An early positive finding can justify a larger trial. It does not make the larger trial optional, or mean that an unfavorable later result can be ignored.

The larger trial found no overall healing advantage

HEAL LL-37 was a randomized, double-blind, placebo-controlled study of topical treatment alongside compression care. It randomized 149 people; 148 received treatment and 144 entered the full analysis set. Across the overall study population, the investigators did not find a significant healing advantage over placebo. A later, post hoc analysis suggested benefit among participants with larger ulcers. Mahlapuu and colleagues, 2021

The trial registry identifies confirmed complete wound closure during the treatment period as the primary outcome. That is a more concrete endpoint than a broad promise of “repair.” EU trial results, 2018-000536-10

The larger-wound finding is worth investigating, but it does not replace the overall result. A subgroup examined after the main analysis is better treated as a hypothesis for another study than as confirmation that a treatment works generally. Reporting only that favorable subgroup would give readers a different impression from reporting the trial as a whole.

Conversely, an overall null result does not prove that every possible formulation or patient subgroup can never benefit. It sets a limit on what this particular study demonstrated. Both the negative main result and the exploratory signal belong in a fair account.

A diabetic-foot study measured more than one kind of healing

A separate randomized, double-blind trial studied 25 people, with 13 assigned LL-37 cream and 12 placebo, over four weeks. The ulcers were uninfected or mildly infected. Both groups also received standard wound care. The LL-37 group improved more on the granulation index, the proportion of wound area covered by granulation tissue. However, the between-group difference in wound-area reduction was not statistically significant. Neither the measured inflammatory markers nor aerobic bacterial counts showed the hoped-for reduction. Miranda and colleagues, 2023

This is a genuine positive finding on one endpoint, not proof of all the proposed effects. Granulation tissue, wound size, complete closure, bacterial burden and inflammation are different measurements. A favorable change in one should not be rewritten as a favorable result in all five.

The small sample also limits confidence about who might benefit and how consistently. The result should not be generalized to severe infection or every kind of wound. A description of a research endpoint is not advice about managing a diabetic foot problem.

Other human studies do not establish general immune support

It would be inaccurate to say that LL-37 has never been injected in human research. A completed melanoma study investigated injections directly into tumors. Its official registry lists four enrolled participants, while the posted participant-flow and outcome tables cover three. It was a small, uncontrolled study in a very specific cancer setting. ClinicalTrials.gov, NCT02225366

That record establishes that this route has been investigated. It cannot validate subcutaneous wellness injections, infection prevention or a general claim of stronger immunity. “Injected” is not a single interchangeable treatment category: placement into a tumor and administration elsewhere ask different clinical questions.

There is also a 2023 report involving an oral preparation made with recombinant LL-37-producing Lactococcus lactis. The single-center, open-label randomized study involved 238 adults hospitalized with COVID-19. It reported earlier viral-RNA test conversion in an early-treatment subgroup, but the overall comparison of conversion time in Table 2 was not statistically significant. This was a different product, with a specific infectious-disease endpoint, not a trial of injected synthetic LL-37. Zhao and colleagues, 2023

Neither a faster laboratory-test change nor a wound-tissue finding is a universal measure of immune health. For a broad “immune support” claim, the missing step is evidence showing the claimed benefit with the actual product and route in the relevant population.

What the safety findings do and do not show

The HEAL trial described topical treatment as well tolerated, but it did not report an absence of adverse events. Some events were considered possibly treatment-related. Serious events occurred, although investigators judged none related to the study drug. HEAL safety results, section 3.6

Those observations are relevant to the preparation and setting studied. They do not establish long-term safety, exclude uncommon harms or answer the safety question for other routes. “The body makes it” is also not a substitute for a safety assessment of an administered product.

FDA currently lists cathelicidin LL-37 among bulk substances whose compounding nominations were withdrawn. The agency flags potential immune reactions, peptide impurities and difficulties characterizing the active ingredient, and says available safety information is insufficient. It also cites nonclinical concerns about male reproduction and tumor-promoting effects in some tissues. These are safety signals and uncertainties, not proof that LL-37 causes those outcomes in people. Withdrawal of a nomination is not a safety clearance. FDA safety-risk page

Does the October 2026 peptide policy cover LL-37?

LL-37 is not among the six substances in Appendix A of the HHS interim compounding policy reviewed here. The document also explicitly distinguishes compounded products from FDA-approved drugs. A policy concerning other named peptides should not be presented as approval, or as blanket authorization for every peptide. HHS policy, opening section and Appendix A

Regulatory status, product quality and clinical effectiveness are separate questions. A claim about one cannot answer the other two. The practical reading rule is to check the named substance and the actual document rather than rely on a headline about “peptides” as a group.

The evidence supports continued investigation of LL-37, especially carefully defined wound-care questions. It does not support converting those studies into a dependable promise of whole-body immune improvement. Preserve the route, patient group, endpoint and uncertainty whenever a research finding becomes a public claim.

Published by Jabful · MultiMode Labs Inc.. This evidence overview is educational, not personalized medical advice or a recommendation to use a peptide. It distinguishes human research from preclinical findings and does not provide doses or treatment schedules.

How we review evidence and handle corrections

Product details: current App Store listing · Privacy · Contact and corrections