The One Peptide the FDA Panel Turned Down

When we covered the July advisory committee vote, the story was the six peptides that got through. There was a seventh on the agenda, and it is arguably the more interesting case, because it is the one that lost.

On July 24, the Pharmacy Compounding Advisory Committee voted against recommending emideltide, better known as DSIP or delta sleep-inducing peptide. The count was 6 in favor, 7 against, with one abstention. One vote the other way and the outcome flips.

So out of seven peptides reviewed across two days, six were recommended and one was not. That alone is worth noting. This was not a committee waving everything through.

Reason one: the evidence

FDA reviewers described the human trials of DSIP as small, poorly controlled, and inconsistent with each other.

That is a fair description. DSIP has been studied since the late 1970s, which sounds impressive until you look at the studies. The encouraging result came from six healthy volunteers. The trial in people with actual insomnia produced weak effects, and its own authors concluded that short term treatment was unlikely to be of major therapeutic benefit. Nearly fifty years of research has not settled whether it does anything.

Inconsistency is the specific problem. When trials of the same compound point in different directions, the usual explanation is that any real effect is small enough to be swamped by chance.

Reason two: the addiction question

This one is more unusual and did not come up for any of the other six.

DSIP is thought to work partly by stimulating endorphin release. Endorphins act on the same reward pathways that opioids do. Reviewers raised a theoretical concern that a compound working through that route could carry dependence potential.

Theoretical is the operative word. Nobody has demonstrated that DSIP is addictive. But the committee was being asked to approve unsupervised compounding of a sleep product, and sleep aids have a long history of dependence problems. In that context, an unresolved question about reward pathways is a reasonable thing to hesitate over.

Reason three: what is actually in the vial

The third objection is the one that deserves the widest attention, because it is not really about DSIP at all.

Reviewers said the substance is not adequately characterized and carries potential for peptide-related impurities from incomplete coupling reactions, truncations, and side reactions.

In plain terms: peptides are built by joining amino acids one at a time. Every join is a chemical reaction, and reactions do not go to completion every time. When a step fails you get a chain missing an amino acid, or a chain that stopped early, or one that reacted in the wrong place. These are similar enough to the intended molecule to be hard to separate out, and they can behave differently in the body.

This is a manufacturing argument rather than a biological one, and here is why it matters beyond this vote: it applies to essentially every synthetic peptide. It is the reason a certificate of analysis is worth something, a point we made when a lab-tested BPC-157 reference product launched. The committee raised it for DSIP, but nothing about the chemistry is unique to DSIP.

Five more in early 2027

The process continues. The same committee is expected to meet again by February 2027 to consider five further peptides:

• LL-37, the body's own antimicrobial peptide
• GHK-Cu, the copper peptide used in skin and wound healing
• Dihexa, an experimental cognitive compound
• Melanotan II, the tanning and libido peptide
• PEG-MGF, a longer-lasting version of MGF with polyethylene glycol attached to slow its breakdown

That is a noticeably different group from the July seven. Those were mostly healing, gut, and longevity compounds. This set reaches into skin, cognition, cosmetics, and muscle, which means the committee will be weighing evidence in areas where the human data is thinner still.

Melanotan II in particular will be an interesting test. It is widely used, its side effect profile is well known to anyone who has taken it, and it has essentially no controlled human trial evidence for the purposes people use it for.

What none of this is

The same caution from July applies. A recommendation is not an approval. Six peptides being recommended for a compounding list does not mean anyone has shown they work, and DSIP being rejected does not mean it has been shown to be harmful.

What the two outcomes together do show is that this committee reads the evidence file and sometimes says no. Given how the July vote was reported, that is worth knowing.

Sources: FDA: July 23-24, 2026 Pharmacy Compounding Advisory Committee meeting and RAPS on the committee backing two more peptides and rejecting one.