Herxheimer reaction, die-off and toxins
When feeling worse at the start of treatment is real, and when "detox" is used to explain away harm.
How solid is the research? Trust score 3 of 5:Promising
The Jarisch-Herxheimer reaction is well documented, but almost only in the treatment of spirochaetal infections such as syphilis, Lyme disease, leptospirosis and relapsing fever. That treatments against fungi, gut parasites or "gut cleanses" trigger a comparable reaction is not proven.
Healthy state
When antibiotics kill certain bacteria very quickly, components of the bacteria are released. The immune system responds with a brief burst of inflammation: chills, fever, headache and muscle pain, sometimes a stronger skin rash. This is the Jarisch-Herxheimer reaction.
It typically begins within 24 hours of the first dose and fades after a few hours. It is triggered mainly by lipoproteins from spirochaetes, and messengers such as TNF, IL-6 and IL-8 mediate it. Spirochaetes are spiral-shaped bacteria, which include the causes of syphilis, Lyme disease, leptospirosis and relapsing fever. In syphilis, it occurred in just under a third (31.5 %) of 355 people treated in one study, most often in the early stage. That it occurs with fungi, gut parasites or gut bacteria during a cure has not been shown.
Similar reactions to dying pathogens also exist beyond spirochaetes, but where the pathogens sit in tissue. In river blindness (onchocerciasis), worm larvae live in the skin and eyes. When treatment kills them, the immune system responds with itching, rash, eye inflammation and sometimes a drop in blood pressure: the Mazzotti reaction. It is stronger the more larvae were present beforehand. With tapeworm cysts in the brain (neurocysticercosis), treatment can increase the inflammation around the dying cysts and trigger seizures, which is why the guideline recommends starting corticosteroids beforehand. What these cases share is that the immune system is in direct contact with the pathogens. Microbes inside the gut are largely kept apart from it by the Mucus layer A gel layer of mucins on the gut wall that keeps bacteria at a distance. In the colon, the inner layer is almost free of microbes. More in the chapter “The mucus layer and additives” and the gut wall; such a reaction has not been shown for them.
How the body really detoxifies: no product does this, but a team of organs. The liver converts fat-soluble substances in two steps so that they become water-soluble. The kidney excretes the water-soluble part in urine, the liver sends the rest with bile into the gut, and the gut carries it out with the stool. Some of it is split off again by gut bacteria and reabsorbed (enterohepatic circulation), which is why fibre and regular digestion help. Skin and lungs excrete substances too. This system works constantly and without a cure; it works as long as the organs are healthy.
Toxins from the gut do exist, just differently from how they are often described. Some bacteria make poisons, such as Clostridioides difficile A bacterium that can spread after antibiotics and cause severe diarrhoea with its toxins. More in the chapter “After antibiotics: C. difficile and faecal transplants”. And Endotoxin (LPS) A component of the outer coat of certain bacteria. When it reaches the blood, it triggers inflammatory responses. More in the chapter “The gut barrier” (lipopolysaccharide, LPS) from the envelope of certain bacteria constantly reaches the blood in small amounts. In mice, a high-fat diet raised this level two- to threefold and promoted inflammation and insulin resistance. What helps against this is a tight gut barrier and a healthy liver, not detox cures. There are exceptions in which “binding” is a real treatment, such as Bile acids Substances made by the liver that make fat digestible. About 95 percent are reabsorbed in the last part of the small intestine; gut bacteria transform the rest. More in the chapter “Bile acids: when bile causes diarrhoea” binders for bile acid diarrhoea, but that is a medicine prescribed by a doctor for a proven cause.
Disrupted state: causes and effects
On the internet almost every worsening during a cure is called “Herx” or “die-off”: with anti-Candida diets, herbal cures for small-intestine overgrowth or parasites, Probiotics Live microorganisms that are meant to provide a benefit in sufficient amounts. The effect depends on the individual strain. More in the chapter “Using fermented foods well” or “detoxes”. There is no evidence for this. The interpretation is the dangerous part: anyone who understands worsening as a sign of healing carries on, even though a side effect, an intolerance or another illness may be behind it.
What can actually get worse during such cures usually has ordinary reasons. New fibre, Prebiotics Food components, usually certain fibres, that specifically feed beneficial gut bacteria, such as inulin or GOS. More in the chapter “Prebiotics that build up gently” or probiotics increase fermentation and with it gas, bloating and cramps, especially at the start. Herbal products such as essential oils irritate the stomach and gut. Antifungals and antibiotics have side effects of their own such as nausea, diarrhoea or headache. Anyone who suddenly drops sugar, caffeine, alcohol or carbohydrates can get withdrawal headache, tiredness and irritability. Then there is the nocebo effect: if you expect a worsening, you notice symptoms more. That is real, but not a sign that pathogens are dying.
The word “die-off” comes up especially often with strong probiotics and ferments, such as milk kefir made from grains. Grain kefir is something different from most shop-bought kefir: a laboratory analysis found a completely different community in it, led by Lactobacillus kefiranofaciens, while Lactococcus lactis dominated in commercial products. Various yeasts add to this, and which microbes end up in the glass, and how many, depends on the grains, the fermentation time and the temperature. Anyone who starts drinking it suddenly takes in many unknown live microbes, plus residual lactose, acid, carbon dioxide, a little alcohol and, depending on the fermentation, biogenic amines such as Histamine A messenger substance of the body that is also found in aged and fermented foods. Too much can cause headache, flushing or diarrhoea. More in the chapter “Histamine intolerance: much searched, little proven”. Bloating, cramps, loose stools, headache or flushing in the first days can be explained by this without anything “dying off”. With Small intestinal bacterial overgrowth (SIBO) Too many bacteria in the small intestine, where normally few live. Bloating and diarrhoea are typical; it is usually diagnosed with a breath test. More in the chapter “When the small intestine is too densely colonised” there may be more to it: in a small study without a proper control group, all 30 people with brain fog and bloating had been taking probiotics, many had D-lactic acid in their urine, and after stopping the probiotics and taking an antibiotic the fog improved in most of them. No study has tested reactions to kefir as a sign that it is working.
Some people also feel worse at first after a faecal microbiota transplant. In a review of 50 publications, 28.5 % of those treated had side effects, most often abdominal complaints such as cramps, bloating and more frequent stools, followed by brief fever (in around 3 %). This resembles a Herxheimer reaction, but counts as a side effect of the procedure and should be reported to the treating practice. The same applies to herbal cures with oregano oil or berberine and to antifungal drugs: whatever gets worse afterwards is first of all a possible side effect of the remedy.
One example of the harm: people diagnosed with “chronic Lyme disease” were sometimes given intravenous antibiotics for months. The US Centers for Disease Control (CDC) described blood poisoning, bone infections, abscesses and C. difficile infections as a result. A long-term benefit of such treatments has not been shown.
Detox programmes as a whole do not deliver what they promise. A review found not a single controlled study showing a detoxifying effect of commercial programmes in humans; the few small studies have major methodological weaknesses. According to a systematic review, colonic cleansing for general health is not supported, whereas there are numerous reports of harm. Coffee enemas have led to inflammation of the colon and rectum, and there are also injuries and electrolyte disturbances. In Inflammatory bowel disease Umbrella term for Crohn’s disease and ulcerative colitis: long-term, relapsing inflammation of the gut that needs medical treatment. More in the chapter “Immune system and inflammation”, enemas are dangerous.
“Binders” are not harmless either. Activated charcoal is an emergency measure in acute poisoning shortly after ingestion, but it also binds medicines and nutrients, and there is no evidence for long-term use. In 2016 the US Food and Drug Administration (FDA) warned against edible clay because products contained raised lead levels.
A true Jarisch-Herxheimer reaction can rarely become dangerous: with a fall in blood pressure, liver or kidney problems and, in pregnancy, contractions. It should therefore be medically supervised.
What often helps
Lifestyle
Diet, exercise, sleep, social contact and dealing with stress.
- Take worsening seriously and note it down When did the symptoms begin, how long do they last, what was changed? A true Herxheimer reaction begins in the first 24 hours and lasts hours. What increases over days is more likely a side effect or intolerance. A daily scale from 0 to 10 (for example in this platform’s symptom diary) makes trends visible. Promising How we rate the evidence Established Recommended in medical guidelines or supported by several consistent meta-analyses of randomised human studies. Well supported Several randomised studies in humans with largely consistent results. Promising Single randomised or controlled studies in humans, not yet independently confirmed — or many consistent observational studies. Plausible Two cases: first observational or pilot studies with small samples — or a settled basis that is not in dispute, such as movement, sleep or meal pauses. But not tested in humans for symptoms. Experimental Barely studied in humans. Mainly lab and animal studies or theoretical reasoning; whether it helps is open. [1]
- Start new products one at a time If you start several things at once, you cannot tell what helps and what harms. So start them one after another, a few days apart, and stick to the dose on the pack. Never dose prescription drugs yourself, and have anything new checked for interactions at a pharmacy or practice first. Plausible How we rate the evidence Established Recommended in medical guidelines or supported by several consistent meta-analyses of randomised human studies. Well supported Several randomised studies in humans with largely consistent results. Promising Single randomised or controlled studies in humans, not yet independently confirmed — or many consistent observational studies. Plausible Two cases: first observational or pilot studies with small samples — or a settled basis that is not in dispute, such as movement, sleep or meal pauses. But not tested in humans for symptoms. Experimental Barely studied in humans. Mainly lab and animal studies or theoretical reasoning; whether it helps is open.
- Support the detox organs, do not replace them What really relieves liver, kidney and gut is unspectacular: little alcohol, enough to drink, plenty of fibre, regular bowel movements, enough sleep and exercise. This is demonstrably healthier than any cure, even if it feels less like “cleansing”. Promising How we rate the evidence Established Recommended in medical guidelines or supported by several consistent meta-analyses of randomised human studies. Well supported Several randomised studies in humans with largely consistent results. Promising Single randomised or controlled studies in humans, not yet independently confirmed — or many consistent observational studies. Plausible Two cases: first observational or pilot studies with small samples — or a settled basis that is not in dispute, such as movement, sleep or meal pauses. But not tested in humans for symptoms. Experimental Barely studied in humans. Mainly lab and animal studies or theoretical reasoning; whether it helps is open. [2]
- Start strong ferments and probiotics slowly With grain kefir, home-made ferments or high-dose probiotics, start with small amounts and only increase when things go well. If symptoms grow over days, pause instead of “pushing through”. Headache, palpitations or flushing shortly after drinking point more to histamine than to die-off. Anyone with small intestinal bacterial overgrowth or who develops brain fog should discuss strong probiotics with a doctor. Plausible How we rate the evidence Established Recommended in medical guidelines or supported by several consistent meta-analyses of randomised human studies. Well supported Several randomised studies in humans with largely consistent results. Promising Single randomised or controlled studies in humans, not yet independently confirmed — or many consistent observational studies. Plausible Two cases: first observational or pilot studies with small samples — or a settled basis that is not in dispute, such as movement, sleep or meal pauses. But not tested in humans for symptoms. Experimental Barely studied in humans. Mainly lab and animal studies or theoretical reasoning; whether it helps is open. [3][4]
Natural supplements
Supplements do not replace a medical check-up. Discuss dosage and interactions with your doctor or pharmacist.
- Be careful with "detox" binders Clay, zeolite or activated charcoal are often recommended against "die-off". There is no evidence, products can be contaminated, and activated charcoal can stop medicines from working. Promising How we rate the evidence Established Recommended in medical guidelines or supported by several consistent meta-analyses of randomised human studies. Well supported Several randomised studies in humans with largely consistent results. Promising Single randomised or controlled studies in humans, not yet independently confirmed — or many consistent observational studies. Plausible Two cases: first observational or pilot studies with small samples — or a settled basis that is not in dispute, such as movement, sleep or meal pauses. But not tested in humans for symptoms. Experimental Barely studied in humans. Mainly lab and animal studies or theoretical reasoning; whether it helps is open. [5][6]
- Chlorella, coriander and “heavy-metal chelation” Algae and herbs said to “draw out” heavy metals rest on a few, mostly animal or very small studies. For people without a proven poisoning, this is not supported. Anyone who suspects heavy-metal exposure should have it measured by a doctor. Experimental How we rate the evidence Established Recommended in medical guidelines or supported by several consistent meta-analyses of randomised human studies. Well supported Several randomised studies in humans with largely consistent results. Promising Single randomised or controlled studies in humans, not yet independently confirmed — or many consistent observational studies. Plausible Two cases: first observational or pilot studies with small samples — or a settled basis that is not in dispute, such as movement, sleep or meal pauses. But not tested in humans for symptoms. Experimental Barely studied in humans. Mainly lab and animal studies or theoretical reasoning; whether it helps is open. [2]
Medical treatments
Information, not a treatment recommendation. Decide together with a doctor. Approval and availability differ from country to country.
- Spirochaete treatment: know the reaction, do not stop on your own If you are treated for syphilis, Lyme disease or leptospirosis, know that fever and chills are possible on the first day. This is not a sign of allergy. Still tell your doctor, and immediately if you have breathing problems, circulation problems or are pregnant. Well supported How we rate the evidence Established Recommended in medical guidelines or supported by several consistent meta-analyses of randomised human studies. Well supported Several randomised studies in humans with largely consistent results. Promising Single randomised or controlled studies in humans, not yet independently confirmed — or many consistent observational studies. Plausible Two cases: first observational or pilot studies with small samples — or a settled basis that is not in dispute, such as movement, sleep or meal pauses. But not tested in humans for symptoms. Experimental Barely studied in humans. Mainly lab and animal studies or theoretical reasoning; whether it helps is open. [1][7]
- No long-term antibiotics without a clear diagnosis Months of antibiotics for unclear symptoms have no proven benefit and can cause serious infections and C. difficile. Well supported How we rate the evidence Established Recommended in medical guidelines or supported by several consistent meta-analyses of randomised human studies. Well supported Several randomised studies in humans with largely consistent results. Promising Single randomised or controlled studies in humans, not yet independently confirmed — or many consistent observational studies. Plausible Two cases: first observational or pilot studies with small samples — or a settled basis that is not in dispute, such as movement, sleep or meal pauses. But not tested in humans for symptoms. Experimental Barely studied in humans. Mainly lab and animal studies or theoretical reasoning; whether it helps is open. [8]
- Never dismiss warning signs as die-off High fever, blood in the stool, severe abdominal pain, yellowing of the skin or eyes, breathlessness or confusion need medical attention right away, whatever protocol you are on. Well supported How we rate the evidence Established Recommended in medical guidelines or supported by several consistent meta-analyses of randomised human studies. Well supported Several randomised studies in humans with largely consistent results. Promising Single randomised or controlled studies in humans, not yet independently confirmed — or many consistent observational studies. Plausible Two cases: first observational or pilot studies with small samples — or a settled basis that is not in dispute, such as movement, sleep or meal pauses. But not tested in humans for symptoms. Experimental Barely studied in humans. Mainly lab and animal studies or theoretical reasoning; whether it helps is open.
- No colon cleanses or enemas for “cleansing” Neither coffee enemas nor colon irrigation in a practice are supported for general health, and they can injure, transmit infections and shift the salt balance. Bowel preparation before a colonoscopy is something different and belongs in medical hands. Well supported How we rate the evidence Established Recommended in medical guidelines or supported by several consistent meta-analyses of randomised human studies. Well supported Several randomised studies in humans with largely consistent results. Promising Single randomised or controlled studies in humans, not yet independently confirmed — or many consistent observational studies. Plausible Two cases: first observational or pilot studies with small samples — or a settled basis that is not in dispute, such as movement, sleep or meal pauses. But not tested in humans for symptoms. Experimental Barely studied in humans. Mainly lab and animal studies or theoretical reasoning; whether it helps is open. [9][10]
- If symptoms persist, look for the cause If symptoms persist despite several cures, a medical work-up (blood count, inflammation markers, coeliac test, stool examination, endoscopy if needed) helps more than the next cure. Often irritable bowel, a sugar intolerance or small-intestine bacterial overgrowth is behind it. Promising How we rate the evidence Established Recommended in medical guidelines or supported by several consistent meta-analyses of randomised human studies. Well supported Several randomised studies in humans with largely consistent results. Promising Single randomised or controlled studies in humans, not yet independently confirmed — or many consistent observational studies. Plausible Two cases: first observational or pilot studies with small samples — or a settled basis that is not in dispute, such as movement, sleep or meal pauses. But not tested in humans for symptoms. Experimental Barely studied in humans. Mainly lab and animal studies or theoretical reasoning; whether it helps is open.
From the community: what people report
Experiences and tips from people with the same problems that come up again and again, but for which there are no meaningful studies yet. They are explicitly not evidence.
“Die-off” as an explanation for every worsening
As in the Candida chapter, very widespread: a worsening after starting treatment is read as “die-off”, usually combined with the advice to increase more slowly instead of enduring a strong reaction. Some stop the product on a strong reaction, a few warn explicitly against calling every worsening “die-off” too quickly. For gut microbes a genuine Jarisch-Herxheimer reaction is not documented.
Binders against “released toxins”
It is widespread to take binders during a treatment – activated charcoal, bentonite, zeolite, chlorella – to catch “toxins” that are said to be released when germs die off, often spaced apart from meals and medicines. In the forums it is known that binders can also bind medicines.
Enemas and coffee enemas
A recurring topic is enemas, above all coffee enemas, often as a multi-part series over weeks, as well as colon irrigation at a clinic. Those affected report that this helps against “toxins”, constipation and mucus. Medical evidence is lacking; coffee enemas and colon irrigation can shift the electrolyte balance, cause injury, and with unclean practice cause infections, and in inflammatory bowel disease they are dangerous.
Personal reports can mislead: what helps many people may be a placebo, and what helps one may harm another. Treat them as ideas to raise with your doctor or pharmacist – not as instructions.
In the mood for something completely different? Another one, at random
How we rate the evidence
- Trust score 5 of 5:Established
- Recommended in medical guidelines or supported by several consistent meta-analyses of randomised human studies.
- Trust score 4 of 5:Well supported
- Several randomised studies in humans with largely consistent results.
- Trust score 3 of 5:Promising
- Single randomised or controlled studies in humans, not yet independently confirmed — or many consistent observational studies.
- Trust score 2 of 5:Plausible
- Two cases: first observational or pilot studies with small samples — or a settled basis that is not in dispute, such as movement, sleep or meal pauses. But not tested in humans for symptoms.
- Trust score 1 of 5:Experimental
- Barely studied in humans. Mainly lab and animal studies or theoretical reasoning; whether it helps is open.
Sources
- Review 2017: the Jarisch-Herxheimer reaction in spirochaetal infections, course and causes
- Klein and Kiat 2015: Detox diets for toxin elimination and weight management, a critical review of the evidence (Journal of Human Nutrition and Dietetics)
- Study 2018: brain fog, bloating, probiotics and D-lactic acidosis in 38 patients
- Review 2026: safety and efficacy of probiotic preparations
- STAT 2016: the US FDA warns against bentonite clay for ingestion because of high lead levels
- Juurlink 2016: Activated charcoal for acute overdose, a reappraisal (British Journal of Clinical Pharmacology)
- Study 2010: Herxheimer reaction in 355 people with syphilis treated with penicillin, frequency 31.5 %
- US CDC 2017: serious infections caused by treatments for "chronic Lyme disease"
- Acosta and Cash 2009: Colonic cleansing for general health promotion, a systematic review (American Journal of Gastroenterology)
- Lee et al. 2020: Proctocolitis from coffee enema, case report (ACG Case Reports Journal)
- Cani et al. 2007: metabolic endotoxaemia, a high-fat diet raises endotoxin in the blood of mice
- Kanza et al. 2024: Ocular Mazzotti reactions after moxidectin or ivermectin in river blindness (Parasites & Vectors)
- IDSA/ASTMH guideline 2017: Diagnosis and treatment of neurocysticercosis, corticosteroids before antiparasitic treatment
- Nejati et al. 2022: Grain kefir and commercial kefir compared by microbes and metabolites (Applied Sciences)
- Wang et al. 2016: Systematic review of adverse events of faecal microbiota transplantation in 50 publications (PLoS One)
Your progress
Log in to mark chapters as read, rate them or report them for review. An account is optional and free.
Is something wrong here?
If you think something is wrong, outdated or misleading, report it for review. Every report is read and answered.
Log in to mark chapters as read, rate them or report them for review. An account is optional and free.