LL-37

LL-37 has been tested in people under controlled conditions, and the best of those tests came back negative. In a phase IIb trial that treated 148 patients with hard-to-heal venous leg ulcers, topical LL-37 closed no more ulcers than placebo: confirmed complete closure was 26.5 percent and 24.7 percent on the two strengths tested, against 25.3 percent on placebo.

That null result followed a smaller and more encouraging start. The 34-patient first-in-man trial reported healing rates about six times faster than placebo at its lowest strength, and a later 25-person trial in diabetic foot ulcers found better granulation tissue without a significant difference in wound area. All three trials put the peptide on the surface of a chronic wound under trial supervision, and in all three it was well tolerated.

None of them tested the reasons most people look LL-37 up. No controlled trial has given it by any route other than topically, and none has measured immunity, gut health, chronic infection or performance. It is also a peptide the body already makes, so research on a person's own LL-37 levels answers a different question from what happens when more of it is administered.

LL-37 is a 37-amino-acid peptide cut from a larger precursor, hCAP18, which is encoded by the CAMP gene. It is the only cathelicidin humans produce. Neutrophils release hCAP18 and enzymes such as neutrophil elastase and proteinase 3 cleave the active peptide from it, and a wide range of other cells make it too, including mucosal epithelial cells, monocytes, macrophages, mast cells and lymphocytes.

Because it is a working part of innate immunity, nobody disputes that LL-37 does something in the body. The open question is narrower: whether adding a synthetic copy from outside changes an outcome that matters. The trials that asked it used a synthetic version of the human sequence, and the company behind the phase IIb develops it for venous leg ulcers under the non-proprietary name ropocamptide.

What the Peptide Does

LL-37 is a cationic peptide, and its first job is breaking microbial membranes. In laboratory assays it inhibited the growth of E. coli, Pseudomonas aeruginosa, Staphylococcus aureus and the yeast Candida albicans. Structural work describes a hybrid mode of attack, partly forming pores and partly stripping lipids from the membrane surface like a detergent.

Its second job is signalling. LL-37 binds bacterial lipopolysaccharide and blocked its effect on vascular nitric oxide production with a potency similar to the antibiotic polymyxin B. It also draws neutrophils toward it through the formyl peptide receptor FPRL1. In wounds, the body's own LL-37 is strongly expressed at the edge of an acute injury early in healing and only weakly in chronic ulcers. That observation about the endogenous peptide is why topical supplementation of stalled ulcers became the clinical test.

Why the Route Matters

The membrane activity is not perfectly selective. LL-37 damages human cells in culture, and a 2025 review places that damage at concentrations similar to those needed to kill microbes. Plasma proteins also bind it, and serum markedly weakens its antibacterial effect.

Put together, those two properties close off the systemic route from both sides. In the bloodstream the peptide is either bound and blunted or, once binding capacity is exceeded, free to harm host cells. Applying it to a wound keeps it out of circulation, and every controlled human trial took that approach.

The overreach around LL-37 does not invent evidence. It selects the convenient trial and carries it somewhere it was never tested.

Quoting the 34-person trial and skipping the 148-person one. The sixfold healing-rate figure comes from the small first-in-man study. The phase IIb existed to confirm it and found closure rates within two percentage points of placebo. When a larger, better-designed trial disagrees with a smaller one, the larger trial is the better answer.

Promoting a subgroup to a result. The large-ulcer benefit came from an analysis run after the main comparison had failed, in arms of 21 to 24 people. In ulcers under 10 square centimetres the placebo group actually closed more often (44.2 percent) than either LL-37 group, though not significantly. That pattern is a hypothesis for the next trial, which is how its own investigators describe it.

Reading a tissue score as healing. In the diabetic foot ulcer trial the granulation index improved while wound area did not separate from placebo. Those are different endpoints, and only one moved.

Moving from the wound to the bloodstream. The case for injecting LL-37 borrows the credibility of topical trials and the biology of a peptide that serum blunts and that damages host cells in culture. The laboratory work argues against that route, and no human trial has tested it.

Topical LL-37 has a reasonably clean safety record across the three trials, and the phase IIb gives real rates. Of 148 treated patients, 64 (43.2 percent) reported 122 adverse events, mostly mild. At least one event was reported by 20 of 48 and 24 of 49 patients on the two LL-37 strengths, against 20 of 51 on placebo. Six events in four patients (2.7 percent) were judged possibly related to study treatment, one of those patients on placebo, and none was judged probably related. There were no deaths.

Total events and serious events were numerically higher on LL-37: 4 and 7 serious events in the active arms against 1 on placebo, 12 in 11 patients overall. They included erysipelas away from the treated ulcer, fractures, heart failure, a myocardial infarction and a stroke, and investigators judged none related to the drug. The authors attribute much of the gap in total events to reports of over- and under-application and to erysipelas. Infection of the treated ulcer was uncommon in every arm: 2.1 and 4.1 percent on LL-37, 5.9 percent on placebo. The first-in-man trial reported no local or systemic safety concerns, and in the diabetic foot ulcer trial one of 13 patients on LL-37 had mild irritant dermatitis attributed to wound exudate.

Those numbers describe a peptide applied to a wound for weeks under supervision. No human safety data exists for systemic use. FDA's published concerns about cathelicidin LL-37 are unusually specific: alongside immunogenicity and characterisation difficulties, it cites nonclinical findings suggesting detrimental effects on male reproduction and that the compound can be protumorigenic in some tissues. LL-37 is not FDA approved and sits on neither 503A list, because its nominators withdrew the nomination around April 2026, which left those concerns published rather than resolved. FDA has said it intends to consult its advisory committee on cathelicidin LL-37 before the end of February 2027. That is an agenda item, not a decision.

The evidence supports a short, specific list. LL-37 is a real human antimicrobial and immune-signalling peptide. Applied to chronic leg and foot ulcers under trial conditions, it was well tolerated over weeks of use. In one small trial it sped healing at its lower strengths; in the trial designed to confirm that, it did not close more ulcers than placebo; in a third, it improved granulation without shrinking wounds faster than placebo.

The evidence does not support any claim about systemic immunity, chronic infection, gut health, recovery or performance, because no trial has measured those, and the laboratory findings on serum binding and host-cell toxicity cut against the systemic route. The one live clinical question is whether LL-37 helps large venous leg ulcers specifically, and that waits on a trial powered to answer it.

LL-37 belongs in the small group of compounds with genuine controlled human data. What that data says, for the population it was designed around, is no.

Not a proven ulcer treatment. Its largest controlled trial missed the primary endpoint in venous leg ulcers, and the benefit seen in large ulcers came from a post hoc analysis still waiting for a trial of its own.

Not a systemic antimicrobial. Every controlled human trial applied it to the surface of a wound. Nothing in the human record covers injection, and the serum and cytotoxicity findings give specific reasons to doubt it.

Not interchangeable with the LL-37 your body makes. Low natural levels in chronic ulcers were the rationale for supplementing it, and the phase IIb showed that rationale did not translate into more healed ulcers.

Not approved by any regulator. There is no approved LL-37 product, and its development as ropocamptide has not produced one.

Not allowed in tested sport. LL-37 falls under WADA's S0 class, which bans at all times any pharmacological substance lacking current approval from a governmental health authority for human therapeutic use. The 2026 list never names it; the catch-all is the provision that applies.

LL-37 is my clearest example of a controlled trial giving a firm answer that happens to be no. Most compounds people ask me about have never been tested properly in humans. This one was, by a company that funded the work and had every reason to want it to succeed, and the phase IIb found no more healed ulcers than placebo across the people it enrolled.

That does not make LL-37 dull. The biology is real, the topical safety data is genuinely reassuring, and the large-ulcer signal deserves the dedicated trial its investigators asked for. What I will not do is let the small early trial stand in for the large one, or let wound-care data be borrowed for injections it never studied. Any argument for LL-37 that opens with the sixfold number and stops there has left out the trial that mattered most.

Taking evidence seriously includes taking it seriously when it says no. A negative result from a well-run trial is information, and here it is the most useful information available.