Peptides for Anxiety, Panic, and Stress: Where the Evidence Actually Lands

Anxiety is the most common mental-health diagnosis in the US, and the pharmacologic options — SSRIs, benzodiazepines, buspirone, hydroxyzine — have well-known limitations. SSRIs take weeks to work and produce emotional blunting. Benzodiazepines work but cause dependence, tolerance, and cognitive side effects. The search for alternatives has drawn attention to peptides with anxiolytic or stress-modulating mechanisms.

The peptide anxiety landscape in April 2026 is more interesting than it's been in twenty years, and easier to misread than people realize. Selank has real RCT data against medazepam. Oxytocin has dozens of trials with inconsistent results and a documented paradox where it can increase anxiety in some contexts. GLP-1 drugs have massive observational signals of reduced anxiety diagnoses paired with a pharmacovigilance counter-signal. Neuropeptide Y is in active Phase 2 trials for PTSD. BPC-157 has beautiful rodent data and zero human psychiatric evidence. CRH antagonists, the pharmacologically logical class, failed in Phase 3 despite confirmed target engagement.

For the cognitive-enhancement angle on Semax and Selank, see our companion neuropeptides insight. This piece focuses on the anxiety framing.

Before the individual peptides, here's the landscape sorted by how much evidence actually exists:

The top tier — Selank and oxytocin — still has important caveats. Selank's evidence is all Russian; oxytocin's is inconsistent. The middle tiers have promising signals but thin confirmatory data. The bottom tiers include the most community-hyped peptides (BPC-157) and the logically-designed class that didn't work (CRH antagonists).

If any peptide has "real" anxiolytic evidence, it's Selank — a synthetic tuftsin analog developed at the Russian Academy of Medical Sciences. The reason to take Selank seriously starts with a specific study: Zozulia et al. 2008, a randomized controlled trial of 62 patients with generalized anxiety disorder and neurasthenia comparing Selank against medazepam (a benzodiazepine).

The results:

• Comparable anxiolytic efficacy to the benzodiazepine
• No sedation, no muscle relaxation, no psychomotor slowing
• No tolerance or withdrawal in the study period
• Additional "antiasthenic and psychostimulant" effects — patients felt more energized, not blunted
• Two responder phenotypes: 40% rapid responders (HARS 20.3 → 7.0 by Day 3) and 60% gradual responders (HARS 16.1 → 6.2 by Day 14)

A 2015 Russian optimization study replicated and extended these findings. Rodent work shows Selank enhances diazepam's anxiolytic effect in unpredictable chronic mild stress models. The proposed mechanism is three-fold: enkephalinase inhibition (prolonging endogenous opioid activity), GABA-ergic modulation (similar downstream target to benzodiazepines but different receptor pharmacology), and BDNF effects (the neuroplasticity story shared with SSRIs).

What Selank doesn't have: a Western multicenter placebo-controlled RCT. Every positive study comes from Russia. This isn't automatically disqualifying — pharmacology doesn't know borders — but it means Selank has never survived the rigor of a modern FDA-registration-quality trial, and the possibility of systematic bias in the source literature cannot be ruled out.

Practical status: Selank is approved in Russia for generalized anxiety disorder and is sold as an intranasal spray (0.15%). In the US it's a research chemical sold by gray-market vendors. There are no registered US trials. The supply chain quality issues documented for research peptides apply with particular force here.

Intranasal oxytocin has been studied in social anxiety disorder, PTSD, and autism-related anxiety for over fifteen years. The results tell a more complicated story than oxytocin's popular reputation suggests.

The RCT landscape:

• A 2025 BMC systematic review of 16 RCTs (11 in PTSD, 5 in SAD) found that oxytocin improved amygdala-cortical functional connectivity and altered amygdala reactivity, with effects moderated by sex and symptom severity
• A 2024 meta-analysis of oxytocin as adjunct to psychotherapy (13 trials, 518 participants) showed reduced stress, improved therapeutic alliance, and reduced negative mental representations
• A 2019 systematic review of 15 RCTs across SAD, phobia, and depression concluded evidence was "inconclusive" with no significant effects on core symptomatology
• Guastella's 2009 RCT showed oxytocin + exposure therapy improved self-evaluations of appearance and speech in SAD patients, but did NOT improve overall symptom reduction

This is the definition of mixed evidence. Some trials hit, others don't, and the ones that hit often show effects on intermediate endpoints (brain imaging, therapeutic alliance) rather than core anxiety symptoms.

The oxytocin paradox. More importantly for patients considering oxytocin for anxiety: Grillon et al. demonstrated that intranasal oxytocin significantly increased defensive responses to unpredictable threat compared to both placebo and vasopressin. A Frontiers 2014 review documented that oxytocin can potentiate ACTH release, intensify negative memories, and worsen anxiety toward unpredictable stressors. Oxytocin's effects appear deeply context-dependent — it strengthens social bonding in safe contexts and amplifies threat response in uncertain ones.

The practical implication: oxytocin is not a universal anxiolytic. For socially-mediated anxiety in structured therapeutic contexts (exposure therapy for phobia, psychotherapy adjunct), the evidence is cautiously supportive. For generalized anxiety or unpredictable-threat anxiety, there's a real possibility oxytocin makes things worse. This is unusual among anxiolytics and deserves to be understood before use.

The most surprising anxiety-peptide story of the past two years didn't come from a peptide designed for anxiety. It came from the GLP-1 weight-loss boom.

The observational signal is massive. An Epic Research / Cosmos analysis of 3.08 million diabetic and 929,000 non-diabetic patients on GLP-1 drugs found:

• Non-diabetic semaglutide users: 37% lower depression diagnoses, 31% lower anxiety diagnoses vs matched controls
• Diabetic users: 45% lower depression, 44% lower anxiety

A March 2026 analysis replicated the pattern with 44% lower depression risk and 38% lower anxiety risk during active semaglutide treatment. A 2025 Frontiers in Psychiatry real-world study linked tirzepatide to lower suicidality risk compared to other weight-loss drugs. The SURMOUNT post-hoc analysis (Obesity, 2026) found no elevated depression or suicidality signal for tirzepatide in obesity patients without major psychopathology.

The counter-signal is also real. A EudraVigilance pharmacovigilance analysis (PMC10960895) showed elevated reporting odds ratios for semaglutide — anxiety 1.26, depressed mood 1.70, suicidality 1.45 — after June 2021. An April 2025 paper proposed that GLP-1s may specifically drive depression and suicidal ideation in patients with genetic predisposition toward low dopamine function, which would mean the effect is bidirectional depending on underlying neurobiology.

The methodological caveat matters. None of the anxiolytic signal comes from RCTs designed to measure anxiety as a primary endpoint. Observational data from 3 million patients is powerful but confounded by selection: people who lose weight on GLP-1s feel better for many reasons, and people who start GLP-1s may differ systematically from those who don't.

Proposed mechanisms include central GLP-1 receptor effects on amygdala reactivity, HPA axis modulation, dopamine-serotonin crosstalk, and reduced chronic inflammation. The science is plausible. What's missing is a dedicated RCT to separate anxiolytic effects from the broader mood benefits of weight loss and metabolic improvement.

Clinically, patients starting semaglutide or tirzepatide for weight loss can reasonably expect that most will see anxiety reduction as a secondary benefit. A small subset, possibly with low-dopamine genotype, may see the opposite. This is increasingly part of the routine GLP-1 informed-consent conversation.

If any peptide becomes the first FDA-approved peptide anxiolytic, it's likely to be neuropeptide Y (NPY).

The biological rationale is clean. Low CSF NPY correlates with PTSD symptom severity across multiple cohorts. NPY Y1 receptor activation in the amygdala reduces fear and stress responses in every preclinical model tested. Intranasal NPY reverses anxiety and depressive-like behavior in the Single Prolonged Stress PTSD model in rats.

Phase 1b results are encouraging. Sayed et al. (2018, International Journal of Neuropsychopharmacology) dosed 26 PTSD patients across five dose cohorts (1.4–9.6 mg intranasal NPY). It was well tolerated at all doses, and high-dose (but not low-dose) NPY was associated with reduced anxiety on some measures.

NCT04071600 is an active intranasal NPY trial in PTSD, currently running. It's the most promising Western peptide anxiolytic trial in active enrollment.

What's not yet known: efficacy in non-PTSD anxiety disorders, long-term tolerance, whether the amygdala engagement translates to robust symptom improvement at scale. NPY has been a promising target for two decades — the shift from preclinical promise to active Phase 2 is genuine progress, but Phase 2 doesn't equal approval.

NPY is the peptide to watch for the next 3–5 years. If the Phase 2 results hit, it would be the first designed-for-anxiety peptide with Western RCT evidence.

Two peptides deserve careful honest treatment because they're heavily marketed for anxiety without commensurate human evidence.

Semax — the ACTH(4-7) analog covered in our cognitive enhancement insight — has specific anxiety-relevant preclinical data. In a chronic unpredictable stress rat model, Semax produced antidepressant-like effects on anhedonia and restored hippocampal BDNF. In the CCK-4 panic model, Semax blocked CCK-4-induced anxiety and depression-like behavior. The mechanism is rapid elevation of hippocampal BDNF and TrkB receptor activation, plus serotonergic and dopaminergic modulation. The human story is Russian-only, and the Russian studies are mostly about cognitive and stroke-recovery endpoints, not anxiety disorders. A Western anxiety RCT of Semax doesn't exist.

BPC-157 has the most elaborate preclinical anxiolytic dataset of any peptide. Tohyama et al. 2004 documented anxiolytic effects in shock probe/burying and light/dark tests. Sikiric 2000 found BPC-157 at 10 ng/kg and 10 µg/kg equivalent to imipramine (15, 30 mg) and nialamide (30, 40 mg) in rat forced swim. The mechanism includes modulation of serotonin synthesis, normalization of dopamine (counteracting both amphetamine overstimulation and haloperidol blockade), and putative GABA-ergic effects.

Zero human psychiatric trials have been conducted. All anxiolytic claims for BPC-157 in humans are extrapolation from rodent data.

Real-world reports are mixed. Some users report calmness they attribute to gut-brain axis modulation. Others report paradoxical anxiety spikes, anhedonia, or brain fog. ExcelMale and Reddit threads document both patterns. Whether the paradoxical response reflects serotonergic modulation in vulnerable individuals, mast cell activation in MCAS-prone users, or something else, isn't clear. Either way, the human anxiety evidence for BPC-157 is anecdote, not trial data.

A complete picture has to walk past the failures. Each one teaches something.

CRH antagonists — corticotropin-releasing hormone receptor antagonists — were supposed to be the next generation of anxiolytics. The biological rationale was textbook-clean: a hyperactive HPA axis is a hallmark of anxiety disorders, and blocking the CRH-1 receptor should dampen the cascade.

It didn't translate. Coric 2010 ran a multicenter RCT of pexacerfont in GAD against escitalopram and placebo — pexacerfont failed to separate from placebo, escitalopram worked. Kwako 2015 tested verucerfont in anxious alcohol-dependent women: reduced HPA axis response and amygdala reactivity on imaging, zero behavioral improvement in distress, anxiety, or craving.

The imaging and biomarker data confirmed target engagement. The behavior didn't move. The lesson: neuroendocrine modulation doesn't automatically produce subjective anxiety improvement. Cortisol and amygdala reactivity are related to anxiety but aren't identical to it. The preclinical CRH model of anxiety didn't translate to chronic human GAD.

TRH (thyrotropin-releasing hormone) / protirelin had the opposite problem. In the 1990s, multiple studies showed rapid, robust anxiolytic and antidepressant effects from IV or intrathecal TRH. Callahan 1993: 26 depressed women, 0.2 mg IV protirelin, significant improvement on anxiety measures at 2 hours. Marangell's intrathecal TRH work in refractory depression: 5/8 patients achieved ≥50% depression score reduction and significant suicidality reduction. The drug worked. The delivery didn't — effects were short-lived, and IV/intrathecal isn't an outpatient anxiolytic. By the time ketamine arrived as a practical rapid-acting antidepressant with IV/intranasal formulations, TRH had effectively been abandoned.

DSIP (Delta Sleep-Inducing Peptide) is often marketed as an anxiolytic and HPA axis regulator. The controlled clinical evidence is thin. A 1995 study showed no inhibitory effect of DSIP on ACTH or cortisol response to physiological or stressor stimuli in humans. There's essentially no modern RCT program. Wellness-space claims substantially outrun the evidence.

The clinical question most patients considering peptide anxiolytics actually have is: how does Selank (or others) compare to the standard options? The honest comparison:

On paper, Selank looks favorable across most dimensions. In practice, the limiting factor is evidence quality and supply chain. A patient with severe anxiety has access to well-characterized medications with clear efficacy and well-understood risks. A patient using Selank from a research-peptide vendor has a drug with real pharmacology, suggestive evidence, and significant uncertainty about what's actually in the vial.

This is the peptide-medicine paradox in microcosm: the molecule may be genuinely useful, but the regulatory and supply chain infrastructure around it is not medicine-grade. For anxiety specifically — where placebo effects are large, expectations matter, and the patient population is often sensitive to subtle changes — this matters more than in most peptide applications.

A few safety angles specific to anxiety peptide use deserve explicit mention.

The mast cell / histamine axis matters here. Substance P, VIP, and somatostatin are known direct human mast cell secretagogues (Church et al. 1989). Reports of 'anxiety' or 'panic' from peptides in MCAS-prone patients may be mast cell degranulation masquerading as a panic attack — the symptoms overlap (palpitations, flushing, chest tightness, sense of impending doom). BPC-157 users who report paradoxical anxiety may be experiencing mast cell activation rather than a direct psychiatric effect. Interestingly, oxytocin inhibits mast cell degranulation in some models (Scientific Reports, 2016), which may explain part of its anxiolytic effect in inflammatory-anxiety presentations.

The oxytocin paradox is real and not widely understood. Oxytocin can increase anxiety to unpredictable threats, potentiate ACTH release, and intensify negative memories. Not a universal benefit.

GLP-1 subgroup risk also deserves attention. The 2025 work suggesting low-dopamine genotypes may be at increased psychiatric risk on GLP-1s is provocative and not yet fully characterized. Patients with personal or family history of depression, suicidality, or dopamine-spectrum conditions should have explicit informed-consent conversations before starting.

The research-peptide supply chain matters. Selank, Semax, BPC-157, DSIP, and the GH secretagogues are commonly sold as research chemicals with limited quality verification. FDA sampling suggests roughly 40% of online and compounded peptides are mislabeled or contaminated. For an anxiety indication — where expectation effects are substantial and the population may be particularly reaction-prone — quality matters more than in many other indications.

Benzodiazepine interactions are an open question. If you're considering Selank specifically for benzodiazepine discontinuation, the enkephalinase-inhibition mechanism may interact with opioid systems in ways that aren't well-studied. Selank isn't a like-for-like benzodiazepine substitute, and withdrawal protocols should be clinician-supervised regardless of whether a peptide is in the mix.

Informed consent on evidence quality is the bigger point. Peptide anxiety therapy sits mostly outside the evidence-based medicine framework. The tradeoff isn't 'peptide vs SSRI' on equivalent grounds. It's 'well-characterized drug with known risks vs less-characterized drug with theoretical advantages and unknown risks.' That's a defensible choice for some patients, but the framing should be explicit.

The peptide anxiety landscape in April 2026 has more substance than the wellness marketing suggests, and more uncertainty than most clinic protocols acknowledge.

Selank has the most robust anxiety RCT data of any peptide. The trials are Russian, and a Western multicenter RCT doesn't exist. The comparison to benzodiazepines is genuinely favorable on paper, but supply chain quality is the practical limiting factor.

Intranasal oxytocin has real pharmacology for socially-mediated anxiety in structured therapeutic contexts. The results are inconsistent, and the oxytocin paradox means it isn't a universal anxiolytic. It can worsen anxiety to unpredictable threats.

GLP-1 drugs show large observational anxiolytic signals in millions of patients, with some counter-signals in pharmacovigilance data. The net effect is likely beneficial for most, potentially harmful for a subgroup with dopamine-sensitivity genotypes. No dedicated anxiety RCT exists yet.

Neuropeptide Y is the furthest-along Western peptide anxiolytic pipeline. Intranasal NPY has Phase 1b data in PTSD and active Phase 2 trials. The space to watch over the next 3–5 years.

BPC-157's anxiolytic claims are entirely rodent extrapolation. Community reports are mixed, including paradoxical anxiety in some users that may represent mast cell activation rather than a direct psychiatric effect.

CRH antagonists, TRH, and DSIP are important failures or abandonments. Each teaches something about why neuroendocrine modulation doesn't automatically produce subjective anxiolysis.

For most patients with anxiety disorders, evidence-based standard care (CBT, SSRIs, benzodiazepines for acute use, buspirone) remains the defensible first line. Peptide therapy is most reasonably considered as an adjunct, an option for treatment-refractory cases, or a choice made with explicit understanding of the evidence gap.

The anxiety peptide conversation is more evidence-based than many of the other 'peptides for X' conversations on this site. The evidence it does have is thin, Russian, observational, or contradictory often enough that honest framing requires acknowledging the uncertainty out loud. If a clinician is pitching Selank or oxytocin or BPC-157 with the confidence of established therapy, the confidence is ahead of the data.