palmitoylethanolamide.com.au
Menu
Home › Forms and Doses › Safety and interaction data

Safety and interaction data

This page collects what the studies on this site report about safety. It is a summary of published data, not a safety assessment, and it cannot tell anyone whether a product is safe for them. Every figure links to the study it came from.

What the independent reviews concluded

Gabrielsson 2016, written by academic pharmacologists with no company link, set the frame that still applies. For treatment up to 49 days, the trial data argued against serious adverse reactions at a rate of 1 in 200 or more. Beyond 60 days, too few patients had been studied to rule out a rate below 1 in 100. The reviewers called the six randomised trials available then of variable quality.

Rankin and Fowler 2020 found no published data on interactions beyond trials where the compound was added to other treatment, and no data on the route or rate of elimination.

The meta-analyses by Artukoglu 2017, Lang-Illievich 2023 and Vina 2025 each report that no major side effects were attributed to the compound in the trials they pooled. Artukoglu adds that side-effect reporting in those trials was often poor, which limits what “none reported” can mean.

Adverse events reported in trials

Across the trials here, the events attributed to the compound are mild and uncommon: gastrointestinal discomfort, dizziness and sleepiness. Examples with numbers:

  • Schifilliti 2014: 1 of 30 patients stopped for excessive sleepiness.
  • Rao 2025: 5 events on the product and 3 on placebo across 217 pain episodes, no difference between groups.
  • Germini 2017: the only adverse event, diarrhoea, occurred on placebo.
  • Noli 2019, in cats: 4 events on the product and 6 on placebo, none serious.
  • Briskey 2026, in dogs and cats: two cats on the product and one on placebo vomited.
  • Schweiger 2024 notes mild gastrointestinal effects in three of the studies it pooled.

Several trials report adverse events only in the placebo group, or none at all. A trial that reports none at all usually did not collect them systematically.

Blood tests

Some trials ran liver, kidney and blood-count panels before and after. None reported a change attributed to the compound: Schifilliti 2014 (60 days), Steels 2019 (8 weeks), Narayanaswam 2023 (12 weeks, with astaxanthin) and Guida 2010 (21 days). Most trials did not run blood tests.

Drug interactions

No interaction study exists. What exists is trials where the compound was added to another drug and the combined group was watched:

Absence of a reported interaction in a small add-on trial is not evidence of no interaction. Search add-on for the full list.

Duration

The longest controlled exposure in this database is about 6 months (Cantone 2024, 4 arms, no placebo; Passavanti 2017, observational). Most randomised trials ran 3 to 12 weeks. No trial followed participants for a year or more. This matches the gap Gabrielsson named in 2016.

Groups with little or no data

  • Pregnancy and breastfeeding: no trials.
  • Children: Khalaj 2018 (ages 4 to 12, with risperidone) and the eczema cream study Eberlein 2008 (ages 2 and up, observational). Nothing else.
  • Liver or kidney disease: excluded from most trials.
  • Horses: no safety trial. See horses.

Topical use

The patch-test study Maghfour 2021 found no irritation or sensitisation from PEA-containing products on normal skin. The eczema and itch trials (Rao 2024, Visse 2017, Yuan 2014) reported no adverse events attributed to the PEA cream, though Visse 2017 had six people in each group report worse skin symptoms.

Regulatory notes

In Australia the compound is a permitted ingredient in listed medicines and has a Schedule 6 poison entry with exemptions for therapeutic use and low-dose cosmetics. See status in Australia. That scheduling decision is a regulatory classification, not a finding from the trials above.