Detected, Listed, Reported: What the Kratom Death Numbers Actually Count

By Kemal Whyte, founder and CEO of Rebel X Brands.
Published 6 October 2026.

Ask how many people kratom has killed and somebody will give you a number. The more useful question is what that number counts. There is no single figure for kratom deaths in the United States. There are several, and each counts something specific: a substance detected on a toxicology screen, a substance a medical examiner listed as a cause, a death in a report to a poison center, a compound identified in a sample sent to a federal surveillance program.

Those are four different things. A number that travels without the thing it counts becomes a talking point, for people who want the plant gone and for people in my own industry who want the subject closed. The honest thing, for a seller more than anyone, is to say what each number counts and live with the answer.

I covered the poison center figures in August, in my column on gas station heroin, and I will not re-argue them. And nothing below says kratom is harmless. Several of these documents found deaths in which mitragynine was the only drug blamed. I quote those too.

Kratom deaths in the 2019 CDC report: detected, listed, the only one

Start with CDC's Morbidity and Mortality Weekly Report of April 2019. It drew on SUDORS, which abstracts death certificates, examiner and coroner reports and postmortem toxicology for unintentional and undetermined overdose deaths, across a patchwork: 11 states for July 2016 to December 2017, and 16 more for the last six months of 2017 only. It holds three numbers that mean three different things.

Of 27,338 overdose deaths, 152, or 0.56%, "tested positive for kratom on postmortem toxicology". That is detection. A medical examiner or coroner determined kratom to be a cause of death for 91 of the 152, which CDC calls kratom-involved. That is a listing, and a listing is not a sole cause: toxicology "detected multiple substances for almost all decedents", and any fentanyl was listed as a cause for 65.1% of the kratom-positive deaths. Among the 91 were seven "for whom kratom was the only substance to test positive on postmortem toxicology", with CDC's caveat in the same sentence: "although the presence of additional substances cannot be ruled out".

"152 kratom deaths" collapses detected into caused. "Only seven" drops every death in which kratom was listed alongside something else. Both take one row and present it as the table. CDC also warned that "Identification of kratom is method-dependent", so the data "might underestimate the number of kratom-positive deaths, although the extent cannot be determined." And they describe 2016 and 2017 in part of the country. Not the current market.

Fifteen deaths in Colorado, re-examined

The reference CDC cites for that caveat is a January 2019 letter in the New England Journal of Medicine from a Colorado public health official, Denver and Boulder County death investigators and university researchers. They found 15 Colorado death certificates from 1999 to 2017 mentioning kratom or mitragynine, the plant's primary alkaloid. Four had been reported as mitragynine only. Three of those had blood left, and comprehensive screening found other substances in all three, from etizolam, a benzodiazepine, to a synthetic amphetamine. Their summary: "we determined that 14 deaths clearly involved multiple drugs."

So "only substance" depends on how hard the laboratory looked. The letter cuts the other way too. It notes that "routine drug testing does not detect mitragynine", so kratom can be missed wherever nobody looks for it, and it concludes: "It is likely that mitragynine increases the risk of adverse events, especially when ingested with opioids or psychoactive drugs." Retesting changed the count. It did not clear the plant.

What FDA says now

FDA's page FDA and Kratom, current as of 2 December 2025, is the agency's live position:

FDA has warned consumers not to use kratom because of the risk of serious adverse events, including liver toxicity, seizures, and substance use disorder (SUD). In rare cases, deaths have been associated with kratom use, as confirmed by a medical examiner or toxicology reports. However, in these cases, kratom was usually used in combination with other drugs, and the contribution of kratom in the deaths is unclear.

Read the three phrases separately. "Rare" carries no number, and I will not pair it with anybody's figure as if FDA had. "Usually" is not always. And unclear is not none. A seller who quotes that paragraph as an acquittal has misread it, and so has a critic who stops before the word "However".

The older number people attribute to FDA comes from a February 2018 statement by the then Commissioner that is no longer on fda.gov. I quoted it from FDA's archive copy in my column of 28 September on why kratom is legal. I could not check it on a live FDA page for this piece, so I am not repeating it as a current count.

Reports to poison centers are not a death count

A report in CDC's MMWR on 26 March 2026, by University of Virginia, Baylor and VA researchers, used the National Poison Data System for 2015 to 2025. I covered it in the gas station heroin column: 233 kratom-associated deaths, 79% involving multiple substances. What matters here is the unit. An exposure report is "actual or suspected contact with a substance that prompted a consultation with a poison center, regardless of toxicity or clinical manifestations." The data are voluntary, and the authors say "determining which substance was most related to clinical effects or medical outcome, including death, was not possible."

So 233 counts deaths among the reports that were made, not the number kratom killed in eleven years. And a "single-substance" report means kratom was the only substance named in the report. Not the only one a laboratory found.

Florida separates cause from present, in writing

Florida's Medical Examiners Commission, publishing through the Florida Department of Law Enforcement (FDLE), prints its definitions on the first page of its drug reports:

A drug is indicated as the cause of death only when, after examining all evidence, the autopsy and toxicology results, the medical examiner determines the drug played a causal role in the death. It is not uncommon for a decedent to have multiple drugs listed as a cause of death. However, a drug may not have played a causal role in the death even when the medical examiner determines the drug is present or identifiable in the decedent. Therefore, a decedent often is found to have multiple drugs listed as present; these are drug occurrences and are not equivalent to deaths.

Cause is not sole cause. Present means found but not judged causal. And occurrences are not deaths, so rows for different drugs cannot be added.

In the 2024 annual report, published in October 2025, mitragynine was listed as a cause in 131 decedents and as present in 93, 224 occurrences in all, and FDLE reports it was listed as causing death in 58 percent of the deaths in which it was found. Occurrences fell 18 percent on 2023; deaths caused by mitragynine fell 28 percent. The 2025 interim report, published in May 2026 for January to June 2025, gives 52 cause, 45 present, 97 in total, and FDLE warns that the half-year data are "subject to change based on the submission of any pending cases". It means it: the January to June 2024 row read 134 occurrences in its own interim report and 136 when restated.

The reports give no reason for the falls, and I will not invent one. They do not break out 7-OH or the newer synthetic compounds. And a cause is still not a sole cause.

The other Florida figure in circulation is not a government one. The Tampa Bay Times ran a news investigation on 7 December 2023 headlined "Hundreds died using kratom in Florida. It was touted as safe." It counted accidental overdoses up to 30 June 2022 in autopsy reports citing kratom or mitragynine as a cause or contributing factor, which "left 587 kratom-related overdoses, 46 of which were determined to be caused by the herb alone." Kratom-only meant "mitragynine or kratom was the only substance listed as a cause of death or contributing factor", so other drugs could be detected if not blamed. A newspaper's count, with its own definitions. Do not add it to FDLE's.

The newer medical examiner studies

Three recent studies went back into examiner files. I report them as published, two from their abstracts alone. A team from Florida Atlantic University, the University of Florida and Harvard, in the Journal of Addiction Medicine in 2025, used Florida data for 2020 and 2021: of 551 decedents with confirmed mitragynine exposure, 36, or 6.5%, "used kratom as the sole substance", and the authors conclude that "dozens of decedents died of kratom (mitragynine) exposures alone". Everyone in that dataset had died, so its comparisons describe how deaths were classified, not the odds that a kratom user dies.

An Orlando group, in the Journal of Analytical Toxicology in May 2026, compared 38 Florida deaths with mitragynine listed as a cause against 111 where it was detected but not deemed causal, the second group from a single county. The first group "had no toxic levels of any other substance", which is not the same as no other substance. Their line: "Mitragynine toxicity accounts for a notable number of deaths."

In Oklahoma, the Chief Medical Examiner's office reported 28 cases from 2022 and 2023 in the Journal of Forensic Sciences. Mitragynine was in the cause of death in 22, "with only one case listing mitragynine as the sole drug contributing to death"; fentanyl was the most frequent companion. The authors judge "it is probable that mitragynine in many of these multi-drug exposures resulted in additive/synergistic effects contributing to the cause of death".

Case counts, none a rate. Together they retire two lazy sentences. "Kratom deaths are really fentanyl deaths" is wrong: examiners have found deaths with mitragynine as the only drug blamed. "Kratom killed all of them" is wrong too. Most involved other drugs too, and there the honest reading is Oklahoma's: in many of them, mitragynine probably added to the harm.

The DEA's laboratory figures count identifications, and they cannot be added together

DEA TOX is, in the DEA's words, "a surveillance program that aims to detect novel psychoactive substances in fatal and nonfatal overdose cases", using samples that hospitals, examiners, poison centers and police choose to send. It was built to spot new drugs, not to count deaths.

The DEA's temporary scheduling order of 26 August 2026 says that between February 2025 and May 2026 "mitragynine pseudoindoxyl has been identified in at least 56 overdose cases, of which 48 were fatal events", and that MGM-15 turned up in 17 overdose cases from February through April 2026, 16 of them fatal. Query date: 15 May 2026. The order does not say what else was in those samples, or whether these compounds caused the deaths.

The figures for 7-OH sit in a different document, the DEA's notice of intent of 6 July 2026. It proposes scheduling 7-OH above a threshold, and as of 6 October 2026 I have found no order following it. DEA TOX "has identified 7-hydroxymitragynine in 85 cases since 2019", it says. "Of these, 55 were fatal and 30 were non-fatal cases." Query date: 5 March 2026. And "Samples often contained other related alkaloids (mitragynine and mitragynine pseudoindoxyl) or other drug classes", fentanyl among them.

Do not add these up. The DEA's July notice on mitragynine pseudoindoxyl gave 31 cases, 25 fatal, for February 2025 to February 2026, query date 27 February 2026; the August order's 56 covers a longer window that includes it. Two snapshots of one tally. And since the DEA lists mitragynine pseudoindoxyl among what 7-OH samples often contained, one case could sit in more than one count. FDA's July 2025 report put it plainly: "it generally cannot be discerned whether deaths are related to mitragynine, 7-OH, or mitragynine pseudoindoxyl, or some combination thereof."

That report also found the number of fatal cases with any of these detected approximately three-fold higher in 2023 to 2025 than in 2019 to 2022 in that voluntary sample, coinciding, in FDA's words, with "more-concerning kratom-related products in the marketplace, such as 7-OH." I said in August that the concentrated end of this shelf is my industry's responsibility. Coinciding is not causing: these numbers cannot show what caused the rise, or rule those products out.

When the laboratory only looks for mitragynine

In its July 2025 assessment of 7-OH, FDA wrote that "Many forensic laboratories test for mitragynine as a marker of kratom use", and that in those cases "7-OH overdose cases and fatalities may incorrectly be classified as kratom and/or mitragynine-related". A laboratory that does find it has the opposite problem, because "7-OH is a known metabolite of mitragynine in humans"; the DEA's July notice says the same. And a 2025 study in the Journal of Analytical Toxicology found that after 60 days of refrigerated storage, the compound's "concentrations dropped over 50% from initial concentrations."

So nobody can split the counts cleanly between leaf and concentrated products. A leaf seller who says those were 7-OH deaths, not ours, claims a precision the data do not have. A regulator who folds them all into "kratom" makes the same mistake from the other side.

What a death certificate is asked to say

Underneath most of these numbers is a form. The National Center for Health Statistics' guide to completing the death certificate in drug toxicity deaths, dated May 2019, explains why detected and listed differ on purpose:

The drugs mentioned should not necessarily be an exhaustive list of all the substances detected by toxicology testing. Toxicology testing may identify the presence of multiple drugs; however, only those drugs present in concentrations considered sufficient to have caused or contributed to the death, given the health of the decedent and the circumstances surrounding the death, should be listed on the death certificate.

It expects certificates to be amended when toxicology arrives late, which is one reason counts can move after publication. And its conclusion says where everyone quoting these numbers should start: "After completing a thorough medical and death scene investigation, the medical examiner or coroner is in the best position to determine the cause of death."

The ledger, and what each line counts

Figure	Source	What it counts	Not this
152 of 27,338	CDC, April 2019. Deaths 2016 to 2017, part of the US.	Kratom detected on postmortem toxicology.	Cause. A national total.
91	Same report.	Kratom listed as a cause by an examiner or coroner.	Sole cause.
7	Same report.	Kratom the only substance to test positive.	That nothing else was there.
15, of which 14 multi-drug	NEJM, January 2019. Colorado, 1999 to 2017.	Certificates naming kratom or mitragynine, reviewed and partly retested.	A rate.
233	MMWR, March 2026. Reports 2015 to 2025.	Deaths among poison center exposure reports.	How many died. Which substance caused a death.
131 cause, 93 present	Florida 2024 annual report.	Mitragynine listed as a cause, or present only.	Sole cause. Any separate 7-OH figure.
52 cause, 45 present	Florida 2025 interim report, January to June 2025.	The same, for half a year.	A final count. Subject to change.
56, 48 fatal; 17, 16 fatal	DEA order, August 2026. Query 15 May 2026.	Mitragynine pseudoindoxyl, then MGM-15, identified in submitted overdose samples.	Cause. What else was present.
85, 55 fatal	DEA notice, July 2026. Query 5 March 2026.	7-OH identified in submitted overdose samples since 2019.	Cause. Anything you can add to the row above.

Opinion: what follows for sellers, me included

First, stop quoting the friendly half of a sentence. "Usually used in combination" sits in the same FDA paragraph as "confirmed by a medical examiner", and unclear means unclear. Florida's examiners listed mitragynine as a cause in 131 decedents in 2024. Any page telling a customer that kratom deaths are really somebody else's deaths is wrong, and I would rather my own brands said less than said that.

Second, when we cite a number, we should say what it counts in the same sentence. If a page under one of my brands does not, I want it fixed. Find one and send it to me.

Third, the concentrated products. The data cannot separate their deaths from the leaf's. That is not an ambiguity to hide behind. It is the reason my industry should never have sold those products under the plant's name, and the reason the leaf may carry some of their numbers until somebody tests well enough to tell the two apart.

And report. FDA "encourages health care professionals and consumers to report any adverse reactions to the FDA's MedWatch program or the Safety Reporting Portal." If a customer describes a serious reaction, point them there and to a clinician; if it is still happening, 911 or Poison Control comes first.

Opinion: what follows for regulators

Test beyond mitragynine, and publish what you tested for. CDC's own authors asked for that in 2019: "Documentation of postmortem toxicology testing protocols is needed to further clarify the extent to which kratom contributes to fatal overdoses." The Colorado authors put it as a rule: "Careful examination of deaths apparently due to kratom only must include comprehensive toxicology screening."

Count product types separately, as the 2026 poison center authors urged: "surveillance should distinguish product types to assess risks." And print the definition beside the number, the way Florida already does. A count without its definition is a headline.

Then publish the split again. CDC's SUDORS page, dated February 2025, says 49 states and the District of Columbia now contribute, recording "All drugs detected" and "Drugs contributing to death". I could not find a published CDC count of kratom-involved deaths for any year after 2017. If one exists, tell me. If not, it should.

One caution my own side will not enjoy. A 2026 analysis of SUDORS data from 45 jurisdictions found that, compared with Schedule I states, "kratom detection rates were significantly higher in regulated states", and in states with no statewide policy. It is an ecological study of detection, not of deaths caused, and it does not show that a ban saves lives. I would say the same if it had pointed the other way.

Every row is a person

Behind every row in that table is a person and a family, and arguing over definitions changes none of that. I am not saying the numbers are too high, or too low. I am saying each one answers a specific question, and quoting it as the answer to a different one is a way of not looking at the people inside it.

My other columns are collected on my Unfiltered page. If a new report changes any figure above, or shows I have described one wrongly, tell me and I will correct this column at the top.

Disclaimer: This article summarizes publicly available material as of 6 October 2026 and links to or names primary sources so you can check them. Death counts are revised as pending cases close. It is general commentary, not legal or medical advice, and no substitute for a clinician. Sections marked as opinion are the author's own, and the author holds commercial interests in botanical brands. Nothing here is a health claim, and statements about botanical products have not been evaluated by the Food and Drug Administration.

Written by Kemal Whyte, founder and CEO of Rebel X Brands, Austin, Texas.