Is Kratom Safe? What the Evidence Actually Says

Natural Does Not Mean Harmless

The word natural does a lot of work in this market. Almost none of it is honest.

Foxglove is natural. So is hemlock. So is the poppy. Caffeine is natural, and it’s the most widely used psychoactive substance on earth. Nature does not sort plants into safe and unsafe for our convenience. It produces compounds that act on the human body. What they do depends on the compound. It also depends on the amount, the person, and what else is in that person at the time.

Kratom is a real plant with real chemistry. The DEA’s own fact sheet records at least 54 alkaloids isolated from it. It states that alkaloid content varies significantly with plant strain and maturity, which produces variation in the chemical profile of different kratom products.

Read that last part again. It is the sentence that governs everything else on this page. Two bags of kratom are not necessarily the same thing. That’s true across brands and across product types. To some degree it’s true across batches from the same field. It’s why we test. It’s also why nobody, including us, can tell you what a given amount will do inside you.

Respect for a plant starts with admitting what it actually is.

What the Best Evidence Actually Says

There is now real clinical evidence. It is worth knowing exactly what it found and exactly who it studied.

In 2026, FDA scientists and their collaborators published a pilot study of botanical kratom. Forty participants took part, in five groups of eight. Six got kratom and two got placebo in each group. The amount given rose in five steps. It rose only after a safety review cleared the step before it. The material was leaf. Its alkaloid levels matched botanical kratom products already described in the literature, and it carried only trace 7-hydroxymitragynine.

No deaths and no serious adverse events occurred. The most common adverse effects were sleepiness, vomiting and nausea. Pupillary constriction, an opioid-type effect, appeared from the second step upward. At the top step, participants reported increases on subjective measures associated with drugs of abuse, including drug liking and feeling high.

That is the FDA’s own study. It carries both halves of the answer in one set of results.

A second and larger trial was published in January 2026. It gave a characterized dried leaf powder to 116 healthy volunteers against placebo. Each participant received one administration, then fifteen daily ones at the same amount, with follow-up monitoring after that. It reported no serious adverse events and no deaths. The authors concluded the powder was safe and well tolerated at the amounts tested. That trial is covered from a different angle, its liver findings, on Long Term Kratom Use.

Now the part that any vendor quoting those headlines leaves out.

Look at who was in those rooms. The two trials studied different people, and neither group looks like a customer. The FDA pilot recruited people who use drugs recreationally and had prior experience with opioids. They were screened to be otherwise healthy. The larger trial recruited healthy non-smokers aged 18 to 55, inside a set weight range, who had either never used kratom or had not touched it for a year or more.

Both groups took a single well-characterized product. Both were monitored under controlled conditions.

Almost nobody buying kratom is in that room. Real customers take other medications. They smoke and drink. They have liver and kidney histories. They buy products of varying and undisclosed alkaloid content. Many take them every day over long stretches. The pilot study’s authors say it directly. Their results may not be representative of other kratom-containing products.

So here’s the honest reading. Under controlled conditions, with a known product, in screened participants, short-term kratom produced no deaths and no serious adverse events. At the top amount it produced the subjective effects that mark substances people return to. What that tells you about somebody taking an unknown product every day over several years is almost nothing. That person was not studied.

And the Other Half of the Record

The clinical trials sit alongside surveillance data, and the surveillance data is not reassuring.

In March 2026, CDC reported national survey findings alongside its own analysis of poison center data. Poison centers logged 14,449 kratom exposure reports between 2015 and 2025. There were 258 in 2015. There were 3,434 in 2025, a record, and an increase of about 1,200 percent. Reports climbed steadily through 2019, held level from 2020 to 2024, then surged. The sharpest rise was among adults aged 40 to 59, whose rates by 2025 nearly matched those of adults aged 20 to 39. Annual use prevalence was stable from 2019 to 2023 while lifetime use rose from 4 million to 5 million people, so more people are trying it and it is spreading across age groups rather than staying with the youngest.

CDC put that 2025 surge alongside the arrival of high-potency products marketed and sold under the name 7-OH. Those are built from isolated and lab-modified alkaloids rather than from leaf. DEA cited these same poison center figures in August 2026 when it placed three of them into Schedule I, mitragynine pseudoindoxyl, MGM-15 and MGM-16. All three are manufactured from purified isolates and do not occur in the plant. They were sold as flavored chewable tablets and liquid shots, marketed as kratom, and DEA’s order describes that branding as deceptive. They are not kratom, they are not what we sell, and they are not what this page is about.

CDC also said plainly that its data do not record whether an exposure involved leaf or one of those products, so the figures cannot be split by product type. That’s not a defense and we’re not going to use it as one. It cuts both ways, and nothing in this record clears leaf either. What it does mean is that the eleven-year line above is not eleven years of the same product, and the steepest year in it is the year those products arrived. The next several years of poison center data will be the first capable of separating the two, and we will be reading them.

The FDA’s Adverse Event Reporting System is a separate record. As of the DEA’s October 2025 fact sheet, kratom was co-involved in 1,486 cases in that system between 2008 and 2025. Of those, 1,387 were classified as serious and 715 involved a death.

Co-involved is the operative word, and we are not going to soften it. Those reports do not establish that kratom caused those outcomes. Adverse event databases collect what gets reported rather than what happened. But the number is not small. It is not decreasing. A company telling you kratom is safe has to explain it away. We would rather you just see it.

Two findings matter more than the totals. Poison centers recorded 233 kratom-associated deaths across those eleven years, and 184 of them, 79 percent, involved more than one substance. Opioids appeared in 62 percent of the deaths, benzodiazepines and stimulants in 20 percent each, alcohol in 19. The authors state their own limit on all of it: when several substances are reported in one case, determining which one was most related to the outcome, death included, was not possible. Kratom-associated is not kratom-caused, and this record does not let anyone say otherwise in either direction. Why that distinction keeps getting lost is the subject of the next section. The authors name kratom taken with alcohol, opioids, benzodiazepines, stimulants and antidepressants as the combination that concerns them. That pattern is the single most actionable thing in the entire safety literature. It is covered on Kratom and Alcohol and on Kratom Drug Interactions.

Why the Case Reports Are Weaker Than Either Side Admits

Most of what gets quoted about kratom harm comes from case reports. A case report is one clinician writing up one patient. A review published in Frontiers in Pharmacology in August 2025 went through 59 of those papers, covering 95 patients. What it found about the quality of that literature matters more than any single case in it.

Toxicology confirmed mitragynine in 55 of the 95 patients. For 40 of them there was no toxicological confirmation that kratom was present at all. Some of those write-ups attributed the harm to kratom anyway.

Among the 55 confirmed cases, 35 died and 20 lived. Other substances turned up in 32 of the 35 who died, against 7 of the 20 who survived. The authors found that difference statistically significant. Blood concentrations did not sort the two groups either. The lowest concentration measured among those who died was below the lowest measured among those who lived. The authors describe the concentrations as less informative than they had expected. There is no number marking a line between the two groups, which is exactly why we will not print one.

Their conclusion was that the literature does not establish that kratom alone raises the risk of severe acute harm. They called for work that separates kratom from everything else these patients took.

Now the part we are obliged to tell you, because that conclusion helps us. The authors work for a consulting firm. By their own disclosure in the paper, that firm has been engaged by kratom processors and distributors in litigation, and one author has served as a paid expert in kratom litigation. That does not make the analysis wrong. Their methodology is transparent, and their point about missing toxicology stands on its own. But we would flag that disclosure without hesitation if the finding had gone the other way. So we are flagging it now. Read it as one input with an interest attached, not as the verdict.

One Thing Among Many, Not the Whole Thing

Nothing works as the only thing you do.

Water is essential, and you can drink too much of it. Exercise is good for you, and people injure themselves with it every day. Sleep, food, sunlight, work, company: every one of them is part of a life, and none of them is a life. A plant is not exempt from that just because it grew in a field rather than a factory.

The fastest way to lose whatever you value about kratom is to take it every day without thinking about it. That sentence is not a warning about morality. It’s what the tolerance literature describes, and it’s the most consistent finding in the whole subject.

Both Things Are True at Once

Millions of Americans have taken kratom and never called a poison center, never went to an emergency room, and never had a problem worth reporting. That’s real.

Thousands did. That’s also real.

Any page giving you only one of those two facts is selling you something. That includes pages selling prohibition as much as pages selling powder. The useful question is not whether kratom is safe. It is what separates the people who never had a problem from the people who did.

The literature points at the same handful of things every time. How much. How often, and for how long. What else was in the body. Whether the product was what the label said. Whether the person had a condition or a medication that made them a bad candidate to begin with. And whether anyone was watching for the early signs.

Every one of those is a decision, not a dice roll. That is the whole reason the Responsible Use guide exists, and every one of those decisions has a page of its own.