Is artemisinin really behind this case of hepatic toxicity?

Artemisinin is a natural product isolated from the leaves of the annual wormwood, Artemisia annua. Used originally in Chinese herbal medicine, the pure compound is employed in Africa as an inexpensive antimalarial drug. In April, 2009, the multinational pharmaceutical company Novartis received FDA approval for a combination drug called Coartem®, comprised of the semi-synthetic artemisinin analog, artemether, and another novel antimalarial, lumefantrine.
An herbal preparation of artemesinin has recently been associated with a single case of hepatic injury as reported in this week’s issue of the CDC publication, Morbidity and Mortality Weekly Report (MMWR).

On August 21, 2008, a man aged 52 years in Seattle, Washington, went to his primary-care physician with symptoms of severe fatigue and dark urine. His medical history included lactose intolerance and irritable bowel syndrome but no known hepatic dysfunction or alcohol abuse. His only medication was a multivitamin. Two weeks earlier, the patient had visited a naturopathic provider for long-standing abdominal discomfort that the provider attributed to a parasitic infection after stool studies reportedly showed an “unidentifiable protozoan.” The naturopathic provider had started him on a 6-week course of an herbal supplement containing 100 mg of artemisinin, two capsules orally three times a day, resulting in a dose of 7.5 mg/kg/day of artemisinin. The supplement was manufactured and sold through a company in the United States. Approximately 1 week into therapy, the patient developed worsening abdominal pain and dark urine. Three days later, on August 18, he stopped taking the supplement when his symptoms did not abate, and 3 days after that, he went to his primary-care physician.

I’ll leave it to my clinical colleagues to evaluate whether this workup by the naturopath was appropriate.

Physical examination by the primary-care physician revealed mild scleral icterus and upper abdominal tenderness. The patient reported no fever, cough, diarrhea, or other symptoms. He reported no significant alcohol use, additional use of over-the-counter medications (e.g., acetaminophen), ill contacts, recent international travel, or exposure to unsafe food or water. Laboratory findings were consistent with hepatitis: a serum alanine aminotransferase of 898 IU/L (normal: 10–55 IU/L), aspartate aminotransferase of 280 IU/L (normal: 10–40 IU/L), bilirubin of 3.1 mg/dL (normal: 0.2–1.2 mg/dL), and alkaline phosphatase of 258 IU/L (normal: 40–150 IU/L). Five months earlier, on March 12, as part of an evaluation for inflammatory bowel disease, all laboratory values had been found within normal ranges.
Among laboratory findings on August 21, the following were within normal ranges: white blood cell count, hemoglobin, hematocrit, platelets, sodium chloride, serum creatinine, glucose, and calcium. The patient’s potassium (3.4 mmol/L [normal: 3.4–5.2 mmol/L]) and carbon dioxide content (22 mmol/L [normal: 22–31 mmol/L]) were borderline normal, and blood urea nitrogen was just below the normal range (8 mg/dL [normal: 9–25 mg/dL]). Laboratory analysis for hepatitis A antibody total and antibody IgM; hepatitis B core antibody, core antibody IgM, surface antigen, and surface antibody; and hepatitis C antibody all were negative. Laboratory testing detected no acetaminophen. Examination of the patient’s stool for ova and parasites was negative.
The patient was admitted to the hospital on August 21, for continued monitoring and supportive care and discharged home on hospital day 3. During the next 2 weeks, the patient’s liver function test results and symptoms gradually improved and had returned to normal by September 4.

The herbal supplement was tested independently by a laboratory at Georgia Tech and found to contain very close to the 100mg of artemisinin as indicated per dose with no impurities or adulterants detected. The MMWR case report confirms my own searches that artemesinin has been used by millions of patients without a single case of hepatotoxicity.
I’m hard-pressed to think of how artemisinin could cause liver toxicity other than considering its endoperoxide group (the O-O group in the structure shown above can generates hydrogen peroxide). Artemisinin is also metabolized by CYP2B6, perhaps competing for metabolism with another drug the patient may have been taken that could predispose to hepatotoxicity.
It would have been nice for the CDC to disclose the identity of the supplement in question so that other researchers could investigate it and clinicians could be aware of counseling against its use.
So was this case caused by some other agent the patient was taking without disclosure.or was it due to the artemisinin? This is a critical issue since the drug is used so widely for malaria as advised by the World Health Organization.
I’m concerned about this story getting blown out of proportion. The case is a single case and causation has not been demonstrated. Idiopathic hepatotoxicity (liver injury without an identifiable cause) is estimated to occur in 1 out of 10,000 patients taking certain drugs but it accounts for about 10% of all cases of drug-induced liver failure.
Any and all discussion is welcome below. My gut (and my liver) tells me that something else was going on in this case besides artemisinin.
h/t via Twitter from Jonathan Eisen et al.


6 thoughts on “Is artemisinin really behind this case of hepatic toxicity?

  1. I’d also be leery of attributing causation to artemisinin. Among other things, what about the “longstanding abdominal discomfort” that was the patient’s initial reason for visiting the naturopath?

  2. If this patient had some kind of iron overload syndrome, that might explain it. Artemisinin has a peroxide bond, and if that interacted with iron, either free or as heme, that would generate hydroxyl via the Fenton reaction. He was taking a multi-vitamin, which likely gave him plenty of iron. If he had enough nitric oxide, the iron and heme gets coordinated by NO and the Fenton reaction is blocked.
    The cytochrome P450 enzymes are all pretty uncoupled, and so make a lot of superoxide when they are activated. This superoxide destroys NO local to the microsome or mitochondrion and disinhibits the P450 accelerating its activity. If there was not enough catalase inside the microsomes, and the H2O2 got out into the cytosol, it could act upon the artemisinin and form hydroxyl in the cytosol.
    I think the CDC should do a more thorough work-up on this person and find out “for sure” if it was artemisinin (because artemisinin is an important generic antimalarial so this stuff needs to be known). I would suggest looking at his iron status and physiology. If he does have a hereditary iron overload syndrome he needs to know about it.

  3. could the patient’s I.B.S. have increased the absorption of artemisinin to this extent? What were his other risk factors for liver disease, for instance his work exposures, past drinking, etc etc. Whole thing sounds like an intentionally created action against artemisinin use.

  4. Probably the most reliable way to decide attribution in suspected Drug Induced Liver Injury is the CIOMS/RUCAM scale. There’s not really enough info in the above post to tell his RUCAM score for sure, but I make it +7 or +8. I disregarded the hospital discharge at day 3 – if he was known to improve within 8 days instead of the quoted “two weeks” he’d be at +9.
    A RUCAM score of +7 is considered “probable DILI” and a score of +8 is considered “highly probable.”
    Note that RUCAM score includes the lack of prior published reports as part of the scale.
    My “4A’s” of DILI: Any drug, Any dose, Any patient, Any time.

  5. I am suspicious of this “The herbal supplement was tested independently by a laboratory at Georgia Tech and found to contain very close to the 100mg of artemisinin as indicated per dose with no impurities or adulterants detected.”
    One can only detect impurities/adulterants one is looking for. An effective herb is a mass of impurities with a drug mixed-in. If you give me a tablet that should be lactose, magnesium stearate and artemisinin I can quickly tell you if it is contaminated; not so an herb. I agree that the identity of the product should be made public for surveillance. However, natural (and unnatural) batch-to-batch variation means this may never be seen again.
    By “unnatural variation” I mean contamination with another herb. According to Dan Hurley in “Natural Causes” (Broadway Books, 2006) there was a situation in Europe in which stefania was contaminated with (deadly) aristolochia. A few years later, the same occurred in the US. He offered no explanation for how this happened, apparently twice. The most innocent explanation is that people collecting herbs may mistake their identity and collect similar-looking plants.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s