Human Studies vs Lab Research on Essential Oils: "Antimicrobial in a Dish"
Why an essential oil that kills germs in a dish is not proof it helps a person, with real examples of cell, animal and human studies and how to tell them apart.
Content last updated . This page has not been medically reviewed.
Safety first. Supplements can cause side effects and interact with medicines. Read mushroom safety and interactions before you use any mushroom product.
Short answer. A result in a dish or in animals is a clue, not a finding about people. Only studies in people can say whether an oil helps, and for most claims about oils they are few and small.
The three kinds of study
- Lab (in vitro) studies test an oil or one of its chemicals on cells, bacteria, viruses or enzymes in a dish.
- Animal studies test an oil in mice or rats. These are also called preclinical.
- Human studies test an oil in people. The best are randomized trials, and reviews that pool them.
Each answers a different question. This page uses the labels preclinical (cells or animals) and human trial so you can see which one a claim rests on.
Why a dish is not a body
In a dish, the oil can sit directly on the germs. In a person it is inhaled, or put on skin, and has to reach the right place without harming the person. Tea tree oil shows the practical problem even in the lab: its parts are only sparingly soluble in water, which makes it hard to test fairly. [1]Examples, side by side
| Oil and claim | What the source found | Kind of evidence |
|---|---|---|
| Tea tree oil (Melaleuca alternifolia) against scabies mites | A review reported promising effects against scabies mites in vitro and called for larger randomized trials. [2] | Preclinical, plus a few case reports |
| Tea tree oil as an antimicrobial | A major review said that where clinical work had been done, data were promising but thus far inadequate. [1] | Mostly lab, limited human |
| Tea tree oil gel for acne | One single-blind trial of 124 patients compared 5% tea tree oil gel with 5% benzoyl peroxide lotion and found both reduced acne lesions. [3] NCCIH summarizes the research as small and says more evidence is needed. [4] | Human trial, limited |
| Eucalyptus oils and inflammation | Leaf oils from several Eucalyptus species were tested on immune cells (macrophages) grown in a dish. The authors say clinical trials are still needed. [5] | Preclinical (cells) |
| Jasmine oil and pain (the study names Jasmine sambac L.) | The oil reduced pain and swelling responses in animals. The authors say more data are needed from human trials. [6] | Preclinical (animals) |
| Essential oils and the coronavirus that causes COVID-19 | A 2020 review said only computer docking and a few in vitro studies were available. [7] | Preclinical (lab and computer) |
"Antibacterial" and "antiviral" on a label
Words such as antibacterial or antiviral can describe a dish result. They do not tell you whether an oil treats an infection in a person. We found no human trial in the sources we read showing that an essential oil prevents or treats a viral infection. For this we looked at the COVID-19 review above and searched PubMed for essential oil antiviral reviews. Proper evidence not found. Do not use oils in place of care for an infection. See Essential oils vs medical treatments.When human studies exist
Aromatherapy trials are usually small and often measure how people feel. NCCIH says that for insomnia, little rigorous research exists and it is not known whether aromatherapy helps. [8] A review of reviews found the evidence was not convincing for any condition it covered. [9] See What trials measure and their limits and Human clinical studies.Frequently asked questions
If an oil kills bacteria in a dish, will it treat an infection?
Not necessarily. A dish result is a lead for research. Reviews of tea tree oil describe the human data as promising but inadequate. [1]
What does "preclinical" mean?
It means the study used cells or animals, not people. The jasmine oil pain study, for example, was done in animals. [6]
Are there human studies on essential oils at all?
Yes, mostly small ones on topics such as anxiety, sleep or skin. Reviews have judged the evidence not convincing for the conditions they covered. [9]
Is a lab result still useful?
Yes, as a reason to run careful human studies. The tea tree scabies authors, for example, call for larger randomized trials. [2]
Where can I learn what is known about a specific oil?
Start with the oil's own page, such as tea tree oil or lavender oil, and ask a pharmacist or clinician before relying on any oil for a health problem.
More details: evidence, limits and cautions
The details behind this page: cautions, limits of the evidence and where sources disagree. Open any note to read it.
Safety
References
- Carson CF, Hammer KA, Riley TV. Melaleuca alternifolia (Tea Tree) oil: a review of antimicrobial and other medicinal properties. Clinical Microbiology Reviews. 2006. Read 2026-10-23.
- Thomas J, et al. Therapeutic Potential of Tea Tree Oil for Scabies. American Journal of Tropical Medicine and Hygiene. 2016. Abstract read 2026-10-23.
- Bassett IB, Pannowitz DL, Barnetson RS. A comparative study of tea-tree oil versus benzoylperoxide in the treatment of acne. Medical Journal of Australia. 1990. Abstract read 2026-10-23.
- NCCIH. Tea Tree Oil. National Institutes of Health, US. Read 2026-10-23.
- Eucalyptus essential oils inhibit the lipopolysaccharide-induced inflammatory response in RAW264.7 macrophages. BMC Complementary Medicine and Therapies. 2020. Abstract read 2026-10-23.
- Jasmine sambac L. essential oil attenuates inflammation and pain via TRPV1. Inflammopharmacology. 2025. Abstract read 2026-10-23.
- COVID-19 and therapy with essential oils having antiviral, anti-inflammatory, and immunomodulatory properties. Inflammopharmacology. 2020. Abstract read 2026-10-23.
- NCCIH. Aromatherapy. National Institutes of Health, US. Last updated January 2020. Read 2026-10-23.
- Lee MS, Choi J, Posadzki P, Ernst E. Aromatherapy for health care: an overview of systematic reviews. Maturitas. 2012. Abstract read 2026-10-23.