Palmitoylethanolamide (PEA) for Pain: What the Trials Really Show

Palmitoylethanolamide, usually shortened to PEA, is a fatty molecule your body makes on its own, and it has quietly become a pain supplement that is popular online but little known to most shoppers. In Italy it has been sold since 2008 as a food for special medical purposes, and online it is pitched as a gentle, non-drowsy, "endocannabinoid" option for nerve and joint pain. Branded forms such as Levagen+ add claims for sleep, allergies, colds and memory.

The research looks impressive at first: several meta-analyses report large drops in pain scores. Look closer and the picture changes. The positive pooled results lean heavily on open-label studies (where patients knew they were taking PEA) and on trials connected to the companies that sell PEA, while most of the double-blind trials (where neither patients nor researchers knew who got PEA) run without a commercial stake found no benefit. This guide explains what PEA is, what each kind of evidence shows, why the branded consumer trials need careful reading, and what PEA costs.

What PEA is

PEA belongs to a family of fat-derived signaling molecules called N-acylethanolamines, which also includes anandamide, one of the body's endocannabinoids. It was first identified in egg yolk, soybean and peanut oil in 1957. In the late 1960s and 1970s it was sold in Czechoslovakia as Impulsin, a 300 mg tablet for flu and colds, and a later review describes several placebo-controlled trials from that era reporting fewer respiratory infections. We could not read those trials directly, so their details rest on that review.

How PEA works is still being worked out. Its best-supported target is PPAR-alpha, a receptor that dampens inflammation; in mice lacking that receptor, its anti-inflammatory effect disappeared. It may also raise levels of anandamide indirectly by competing for the same breakdown enzyme (FAAH), an idea called the "entourage effect," and in animal studies it reduced the release of inflammatory substances from mast cells (immune cells that release histamine). An independent 2020 pharmacology review urged journal reviewers to stop accepting descriptions of PEA as an endocannabinoid, because "it does not interact directly with CB receptors." That also fits with the absence of any reported "high" in trials.

Basic questions remain open. The same review noted that PEA's bioavailability and tissue distribution are not known. In one measurement cited by a 2016 review, a 300 mg dose of micronized PEA raised blood levels for only a few hours.

The evidence at a glance

These summaries describe research. They are not claims about what a supplement does. Persistent or severe pain, numbness or weakness should be evaluated by a clinician.

The meta-analyses: big numbers, shaky foundations

Most PEA articles cite one of four pooled analyses. Each looks strong in its abstract and weaker underneath.

The 2023 "double-blind only" meta-analysis. Austrian researchers, declaring no funding or conflicts, pooled 11 double-blind trials with 774 patients and reported a standardized mean difference of 1.68 (a way of comparing trials that used different pain scales; 0.8 or more is conventionally called large), an unusually large effect for any pain treatment. But the trials varied enormously (a heterogeneity score, I², of 93 percent, meaning most of the spread reflects real differences between trials rather than chance), they covered conditions as different as burning mouth, painful periods, carpal tunnel symptoms and knee pain, and 4 of the 11 tested PEA combined with another compound (polydatin), not PEA alone. Three of the 11 found no significant difference between groups (the paper's own text counts two). The paper is also internally inconsistent: its abstract and results text give different p-values for the main result, and it describes the effect size as points on a pain scale, which it is not. Its largest trial, 220 women with painful periods treated for 10 days, reports a confidence interval far narrower than standard methods give for that sample size: by our arithmetic, roughly 94 to 99.5 percent, against the published 97.6 to 98.6 percent. The meta-analysis's own quality table also records that trial as not described as double-blind. We found no correction or published comment on it. The authors themselves cautioned that the result "may not necessarily translate into a clinically significant benefit."

The 2016 pooled-data analysis combined raw data from 12 studies, but 9 of them were open-label, 7 had no comparison group at all, and two authors listed Epitech affiliations; Epitech is the Italian company that makes Normast.

The 2017 meta-analysis was dominated by a single sciatica trial of 636 patients published in a Spanish journal not indexed in PubMed, which we could not read directly. A later reanalysis of that trial reported a number needed to treat of 1.7 for the 600 mg dose (9 for 300 mg), meaning fewer than two people treated for one to benefit. That is far better than standard nerve pain medicines, whose numbers needed to treat in a major 2015 analysis were mostly about 6 to 8, and the reanalysis authors cautioned that an unknown number of patients in both groups may have recovered on their own during the three-week trial.

A 2022 meta-analysis included only two double-blind trials, and both were null; its positive pooled result came from open-label and observational studies, with signs of publication bias. A 2024 analysis drew 8 of its 9 studies from uncontrolled research. A 2025 analysis of 18 randomized trials reported a moderate-to-large effect and called PEA a promising alternative to opioids, a use no trial has tested; its reference list includes combination products, and we could not confirm exactly which trials it pooled.

An independent 2016 review by Swedish pharmacologists put it fairly: the data "point to efficacy of PEA over placebo (assuming no publication bias), but more information is needed to be able to gauge this efficacy vs. current treatment regimes." We found no Cochrane review of PEA for pain and no pain-society guideline that recommends it; the UK's NICE neuropathic pain guidance does not mention it.

What the independent trials found

The trials least likely to be shaped by commercial interest tell a consistent story.

• Spinal cord injury nerve pain (Denmark, 2016). In a double-blind trial of 73 people published in the journal Pain, 1,200 mg a day of ultramicronized PEA for 12 weeks, added to usual care, made no difference to pain intensity, spasticity, sleep or mood, although participants on PEA used less rescue paracetamol. Epitech supplied the product, but the trial was run independently.
• Carpal tunnel symptoms (Spain, 2017). In a double-blind trial of 61 people with no commercial funding, 600 mg a day for 60 days did not improve any clinical or nerve-conduction measure, though the PEA group improved on one questionnaire scale and the authors called for trials of higher doses.
• Nerve pain after chemotherapy (US, 2024). In a Mayo Clinic-led double-blind trial of 88 people, neither 400 nor 800 mg a day for 8 weeks improved established chemotherapy-related nerve symptoms, quality of life or thinking. The authors concluded that "PEA failed to improve" the condition.
• Exercise recovery (Belgium, 2024). In a small crossover trial of 11 men at KU Leuven, with no company funding stated, Levagen+ made no difference to soreness, strength or muscle-damage markers.

Some registered trials have no publication we could find, including a double-blind fibromyalgia trial registered by Epitech and completed in 2022 (a paper may still appear, or may exist without the registry number). A trial in bipolar depression posted results on ClinicalTrials.gov showing no advantage over placebo (depression scores fell 12.4 points on PEA and 13.0 on placebo) without a journal paper.

The company-funded pain trials

The positive double-blind trials mostly come from groups linked to PEA's makers. In a trial of 70 adults with diabetes-related nerve pain, sponsored by Gencor Pacific (which sells Levagen+), 600 mg a day for 8 weeks lowered pain scores compared with placebo. In a single-site trial of 111 adults with knee pain from osteoarthritis, 300 or 600 mg a day for 8 weeks improved joint scores; the 2023 meta-analysis lists it as company-funded. Separately, the conflict-of-interest statement of a 2021 review co-written by R. Venkatesh, one of the knee trial's authors, says he "has personal relations to one of the members of Gencor Pacific." A second 12-week knee trial, by researchers at the Indian company BioAxion Innovations, missed its main goal, with the placebo group improving more at week 4.

Other positive trials, in burning mouth, jaw joint pain and multiple sclerosis injection-site pain, used Epitech's product; the painful-periods trial did not report its manufacturer; several tested PEA combined with polydatin, and many were small. A 2026 trial of 120 adults with nerve-related low back pain, run with Indena, which makes the phospholipid PEA it tested, also reported less pain than placebo. None of this proves the results are wrong, but it means the evidence has not been independently confirmed.

Sleep, allergies, colds and memory: the Levagen+ trials

Levagen+ is marketed for everyday uses, and each claim rests on a single trial funded by Gencor Pacific or Pharmako Biotechnologies, the companies behind the ingredient. Several, including the sleep, cold and allergy trials, were run by one Australian contract research firm, RDC Clinical.

• Sleep. 125 adults were randomized to 350 mg a day or placebo for 8 weeks. Overall sleep quantity and quality did not differ; time to fall asleep improved only in a subgroup who took more than 10 minutes at the start.
• Colds. In 426 adults over 12 weeks, the PEA group reported fewer upper respiratory infection episodes (39 vs 64), meeting the main goal, but more people dropped out on PEA (19) than on placebo (9).
• Allergies. In 108 people over 2 weeks, the main nasal symptom score did not differ from placebo; a subgroup with higher starting scores improved.
• Memory. In a 39-person crossover trial (each person took PEA and placebo in turn), PEA reduced errors on one memory test and raised a blood marker (BDNF), while another memory measure did not differ.
• Muscle. In 52 people doing 8 weeks of resistance training, lean mass gains did not differ; jump height improved more on PEA, while bench-press strength improved more on placebo.
• COVID-19 inflammation. In 60 adults, 5 of 6 main inflammation markers did not differ.

In several of these papers, the conflict-of-interest statement shown on PubMed says "none," while the funding section of the full text names Gencor Pacific or Pharmako. Journals often treat funding and conflicts as separate declarations, but a reader who sees only the PubMed record would miss the sponsor. Single, unreplicated, sponsor-funded trials are a starting point, not a basis for buying a supplement.

Micronized, ultramicronized, Levagen+: does the form matter?

PEA does not dissolve well in water, so makers grind it into very fine particles (micronized or ultramicronized PEA, Epitech's approach) or mix it with a dispersing system (LipiSperse, used in Levagen+). The marketed figure that Levagen+ is absorbed 1.75 times better comes from a single-dose study of 28 people over 4 hours, comparing it with what the paper called "standard" PEA (a later review by Gencor-linked authors describes it as non-micronized). Peak blood levels did not differ significantly, the paper appeared in a journal not indexed in PubMed, and it was funded by Gencor Pacific and Pharmako, the two companies behind the ingredient.

We found no human trial comparing one PEA form with another on pain or any other health outcome; the only head-to-head human comparisons we found measured blood levels. A 2016 independent review reached the same conclusion about micronized versus unmicronized PEA: "evidence for superiority of one formulation over the other is currently lacking." Higher blood levels after one dose do not show that a product works better.

How PEA is sold

In Italy and Spain, PEA is sold as a food for special medical purposes (for example Normast, made by Epitech), a category used under medical supervision. That is not a medicine approval and involves no pre-market review of whether it works. In the US, PEA is sold as a dietary supplement. We found no FDA GRAS notice or new dietary ingredient notification for PEA; in 2018 Gencor told a trade publication it relies on its own self-affirmed GRAS ("generally recognized as safe") assessment, a route the law allows in which experts chosen by the company, not the FDA, judge the safety data.

Cost

Here is what two products cost on November 1, 2026. Several retailers blocked our price checks, so this is a small sample. Prices change.

The Emerald Labs price is from Vitacost and the Nootropics Depot price from the brand's store and a SuppCo listing, read November 1, 2026. Cost per day is our arithmetic; the Levagen+ figure assumes 350 mg of Levagen+ supplies about 300 mg of PEA, as in the trials.

Per milligram, the branded Levagen+ product cost about five times as much as plain micronized PEA, and no trial shows the pricier form works better. Talk with a clinician before trying PEA for ongoing pain.

Is it worth trying?

Start with a diagnosis. Chronic pain, numbness or tingling has many causes, some serious, and the treatments with the best evidence are the ones a clinician can offer. Do not replace or delay prescribed treatment with PEA.

If you and your clinician decide to try it, the short-term safety record is reassuring, and trials typically used 600 mg a day, sometimes up to 1,200 mg, usually for 2 to 12 weeks. Plain micronized PEA costs far less than branded forms, with no evidence the branded forms work better. Set a time limit, such as 8 weeks, and stop if you notice no difference.

Be skeptical of everyday claims. The sleep, allergy, memory and immune claims for Levagen+ each rest on a single company-funded trial, and several missed their main goal. For other ingredients studied for nerve symptoms, see our nerve pain supplements guide and our benfotiamine review.

Safety

Frequently asked questions

What is palmitoylethanolamide (PEA)?

PEA is a fatty molecule the body makes on its own, related to the endocannabinoid anandamide. It was identified in 1957, sold in Czechoslovakia as a flu and cold tablet in the late 1960s and 1970s, and is now sold in Italy and Spain as a food for special medical purposes and in the US as a dietary supplement.

Does PEA work for pain?

The evidence is mixed. Meta-analyses report large drops in pain scores, but they rely heavily on open-label or company-linked studies and show very high variation. Independent double-blind trials in spinal cord injury nerve pain, carpal tunnel symptoms and nerve pain after chemotherapy found no significant benefit. Persistent pain should be evaluated by a clinician.

Is PEA a cannabinoid? Will it make me high?

No. PEA does not act directly on cannabinoid receptors, and an independent pharmacology review urged journal reviewers to stop accepting descriptions of it as an endocannabinoid. No trial has reported intoxicating effects, though we found no dedicated study of abuse potential.

Is Levagen+ better than regular PEA?

Not shown. The claim that Levagen+ is absorbed 1.75 times better comes from one single-dose study of 28 people, compared with "standard" PEA, which the makers describe as non-micronized, with no significant difference in peak blood levels. It was funded by the ingredient's makers. No trial has compared PEA forms on pain or any other health outcome; head-to-head studies measured only blood levels. In our price check, Levagen+ cost about five times as much per milligram.

How much PEA do studies use?

Most pain trials used 600 mg a day, some up to 1,200 mg, usually for 2 to 12 weeks. Levagen+ consumer trials mostly used 350 to 700 mg a day of Levagen+ (about 300 to 600 mg of PEA); a COVID-19 trial reported a higher dose.

Is PEA safe?

In short trials, side effects were mostly mild and similar to placebo, and no serious side effects have been attributed to PEA. But too few people have taken it for longer than about 60 days to rule out rarer problems, interaction data with medicines are lacking, and there are no human pregnancy data.

The bottom line

PEA is a molecule your body makes, with a plausible anti-inflammatory mechanism and a reassuring short-term safety record. The headline evidence for pain, though, comes mostly from pooled analyses built on open-label and company-linked studies, and from single-site trials connected to the companies that sell it. The independent double-blind trials, in spinal cord injury nerve pain, carpal tunnel symptoms and nerve pain after chemotherapy, found no benefit, and the branded consumer claims rest on single sponsor-funded trials.

If you want to try it, do so with a clinician, keep in mind that no trial shows a pricier branded form works better than plain micronized PEA, and give it a fixed trial period. Keep expectations modest, and keep ongoing pain under medical care.