Oil pulling: what the trials actually measured

Oil pulling has been tested more often than most home remedies — twenty-five randomised trials and 1,184 people in the largest meta-analysis. What almost nobody reports is what those trials measured: bacteria in a saliva sample, plaque scored after disclosing, and gum inflammation scored by eye, over periods between seven and forty-five days, in participants who mostly had nothing wrong with their mouths.

That is not nothing. It is also not an answer to “does this keep my teeth?”, and the gap between those two things is the whole story here.

What people claim, and what was tested

Claims range from the plausible to the impossible: less plaque and gingivitis, fresher breath, whiter teeth, “pulling toxins” from the body, clearer skin, treating systemic disease. Only the first was ever measured.

The toxin claim deserves dismissing specifically rather than politely. Swishing oil in your mouth does not draw toxins out of your bloodstream. There is no mechanism by which it could: your oral mucosa is not a filtration membrane, and substances do not migrate out of your organs into a mouthful of coconut oil because it happens to be there. Detoxification is what your liver and kidneys are for.

The mechanism that is not nonsense

There is a real, if modest, rationale — two of them.

• Bacterial cell membranes are lipid-based. Many oral bacteria have fatty outer membranes, and fat dissolves in fat. Extended contact with oil can plausibly disrupt and carry away some of the bacteria and biofilm on a tooth surface.
• Coconut oil contains lauric acid, which has antimicrobial activity in laboratory conditions. Saliva may also partly saponify the oil during swishing, producing a mild soap-like effect.

Neither is fantasy. Both are a long way from “replaces your dental routine”, and both are the kind of mechanism that produces exactly the short-term surrogate results the trials found.

What the reviews found, in order

Four systematic reviews have pooled this literature. They are worth taking in sequence, because they disagree in an instructive way.

1. 2016, five trials, 160 participants Three trials reported no significant difference in plaque index between oil pulling and control (p = 0.28, 0.94 and 0.38); two reported no significant difference in modified gingival index against chlorhexidine (p = 0.32 and 0.64). Trials lasted 10 to 45 days. The review’s own conclusion was that oil pulling “may have beneficial effects” over the short period investigated, and that future trials should be more rigorous.
2. 2020, coconut oil only, four trials, 182 participants Seven to fourteen days. Significant reductions in salivary bacterial colony count and plaque index score were reported, but no meta-analysis was performed because the studies were too clinically heterogeneous, quality was mixed and risk of bias was high.
3. 2022, nine trials Salivary bacterial colony counts were significantly reduced (mean difference 17.55, 95% CI 2.56 to 32.55). Plaque index was not (−0.10, 95% CI −0.33 to 0.14). Gingival index was not (−0.05, 95% CI −0.12 to 0.02).
4. 2024, twenty-five trials, 1,184 participants The largest. Against non-chlorhexidine comparators, oil pulling improved modified gingival index scores substantially (standardised mean difference −1.14, 95% CI −1.31 to −0.97). Against chlorhexidine, chlorhexidine was better at reducing plaque (SMD 0.33, 95% CI 0.17 to 0.49). And then the sentence that governs the rest: “The overall quality of the body of evidence was very low.”

So two reviews found no plaque effect, one found chlorhexidine better on plaque, and the largest found a genuinely large gingival-index effect on evidence its own authors graded very low. Those are not contradictions so much as different comparators: oil pulling looks decent against “nothing much” and worse against a proper antiseptic.

Two problems that run through all of it

The participants mostly did not have a problem. In the 2024 review, more than half the trials — seventeen of twenty-five — involved people with no reported oral health issues. You cannot demonstrate much improvement in gum inflammation in people whose gums are not inflamed, and a trial that tries will produce small, noisy numbers whichever way the intervention works. This is the same defect that runs through the flossing evidence, and it is worth learning to spot: check the baseline before you read the result.

Nobody could be blinded. You cannot disguise twenty minutes of swishing oil, so every participant knew which group they were in, and both examiner-scored and self-reported outcomes are vulnerable to that. Add the attention effect — someone committing twenty minutes a day to their mouth is generally paying more attention to it in every other respect — and a chunk of any measured improvement has an ordinary explanation.

The twenty-minute problem

Here is the practical objection, and for most people it decides the matter.

Oil pulling asks for fifteen to twenty minutes a day. Cleaning between your teeth takes two. Brushing takes four. If somebody has twenty spare minutes for their mouth and is not yet cleaning between their teeth, spending it on oil pulling is a bad trade: swishing does not remove plaque from between two contacting teeth, and that is exactly where gum disease starts.

As an addition to a routine that is already complete, the calculation changes and the answer becomes “if you enjoy it, fine”. The problem is only substitution.

Risks, such as they are

• Inhaling the oil. Lipoid pneumonia from accidentally breathing oil into the lungs is rare but documented. Do not do it lying down, while distracted, or with young children.
• Upset stomach if swallowed. Spit it out, always.
• Jaw ache. Twenty minutes of continuous swishing is genuinely tiring, and a poor idea if you have jaw joint problems.
• Blocked drains. Coconut oil solidifies below about 24 °C. Spit into a bin, not a sink.
• Displacement, which is the real one — doing this instead of brushing, cleaning between teeth, or seeing a dentist.

What a better trial would look like

It is worth saying, because it is the same short list every time this pattern turns up.

• Participants who actually have gingivitis at baseline, so there is room to measure a change.
• A comparator that is a real alternative — not just “carry on as normal”, which tells you about attention as much as about oil.
• Long enough to reach an outcome that matters: bleeding sites at six months, pocket depths at a year, not plaque index at three weeks.
• Examiners who do not know the allocation, even if participants must.

None of that is exotic. It is what the reviewers themselves keep asking for, in the last paragraph of every one of these papers.

The verdict

Oil pulling is safe, cheap, and probably does something modest to gum inflammation over a few weeks. It is not a detox, it does not meaningfully whiten teeth — see natural teeth whitening for why — and it treats no disease. No trial has followed anyone long enough to know whether it prevents a single cavity. If you enjoy the ritual and you are already brushing twice daily with fluoride toothpaste and cleaning between your teeth, it is a harmless addition. If you are doing it instead of any of those, it is costing you.

That framing goes well beyond oil pulling. Most oral health additions — rinses, supplements, probiotics — are worth considering only once the mechanical basics are actually in place, because nothing you swish or swallow removes plaque from a tooth. The same question is worth asking of any of them: what, exactly, did the trials measure, and for how long?

Do oral probiotics work? →

If you want to try it anyway

A tablespoon of coconut or sesame oil, swished gently for ten minutes — twenty is tradition, not a requirement — spat into a bin, mouth rinsed with water afterwards. Do it before brushing, not instead of it, and stop if your jaw complains. Half the trials used sesame rather than coconut, for what that is worth, which is not much.

See a dentist if

• Your gums bleed when you brush, and have done for more than two weeks.
• You have taken up oil pulling because something hurt. Pain is a reason to be examined, not a reason to swish.
• You are using it as a substitute for treatment you have already been advised to have.

Fig. 1 is free to reuse under CC BY 4.0 — see the diagrams page.

References

1. Gbinigie O, Onakpoya I, Spencer E, et al. Effect of oil pulling in promoting oro dental hygiene: a systematic review of randomized clinical trials. Complement Ther Med 2016;26:47–54. DOI · PubMed · Funding: see ledger
2. Woolley J, Gibbons T, Patel K, Sacco R. The effect of oil pulling with coconut oil to improve dental hygiene and oral health: a systematic review. Heliyon 2020;6(8):e04789. DOI · PubMed · Funding: see ledger
3. Peng TR, Cheng HY, Wu TW, Ng BK. Effectiveness of oil pulling for improving oral health: a meta-analysis. Healthcare (Basel) 2022;10(10):1991. DOI · PubMed · Funding: see ledger
4. Jong FJX, Ooi J, Teoh SL, et al. The effect of oil pulling in comparison with chlorhexidine and other mouthwash interventions in promoting oral health: a systematic review and meta-analysis. Int J Dent Hyg 2024;22(1):78–94. DOI · PubMed · Funding: see ledger