Herxheimer’s Reaction: Symptoms, Causes, and Management

The Jarisch-Herxheimer reaction is a sudden inflammatory flare that occurs within hours of starting antibiotic treatment for certain infections, most commonly those caused by spiral-shaped bacteria called spirochetes. It can look and feel alarmingly like the infection getting worse, with fever, chills, and muscle pain, but it is actually the body’s immune system responding to a flood of debris released from dying bacteria. The reaction typically resolves on its own within a day, though in some settings it can be severe enough to require close medical monitoring.

What the Reaction Feels Like

The hallmark symptom is fever, which shows up in the vast majority of cases. In a study of neurosyphilis patients treated with penicillin, fever appeared in over 98% of those who developed the reaction.¹ Beyond fever, the most frequently reported symptoms are muscle aches, chills, weakness, and a feeling of general unwellness. A secondary analysis of a clinical trial in adults treated for early syphilis found that among those who reacted, about half reported muscle pain, close to half had chills, roughly four in ten felt weak, and about a third reported feverishness.² Sweating, nausea, and flushing round out the picture. If the patient already has a skin rash from the underlying infection, the rash often flares or becomes more vivid during the reaction.³

Symptoms typically begin within four to six hours of the first antibiotic dose, with the large majority of cases showing up within 12 hours. In the syphilis trial mentioned above, the median onset was about five hours, and the median duration was roughly 13 hours.² Most descriptions put the full window at 18 to 24 hours before everything settles.³ For the average person experiencing it, the reaction feels like a sudden, intense flu that arrives hours after treatment and is largely gone by the next day.

Why Antibiotics Trigger It

The reaction is fundamentally an immune overreaction to bacterial components that suddenly flood the bloodstream when the antibiotic starts killing the pathogen. The exact triggers have been debated for over a century, but research points to spirochetal lipoproteins, proteins embedded in the outer membrane of the bacteria, as the main culprit. When the bacteria die en masse, those lipoproteins are released and act as a kind of alarm signal to the immune system.⁴ Some studies have also pointed to endotoxin-like material, though the evidence on that front is more conflicted.⁵

What follows is a rapid spike in inflammatory signaling molecules. Studies in patients with louse-borne relapsing fever have documented sharp, transient rises in tumor necrosis factor alpha (TNF-alpha), interleukin-6, and interleukin-8, all chemicals the immune system uses to coordinate its attack response.⁶ This cytokine surge is essentially the same cascade that produces the miserable symptoms of sepsis, just compressed into a shorter, self-limiting window. Researchers have actually used the Jarisch-Herxheimer reaction as a human model for studying cytokine behavior in sepsis, because it follows a predictable, time-bound course that sepsis itself does not.⁷

Which Infections Carry the Risk

The reaction is classically associated with infections caused by spirochetal bacteria, and syphilis is the most well-known trigger. In the clinical trial of early syphilis treatment, roughly one in four patients experienced the reaction after receiving standard penicillin therapy.² The reaction tends to be more common in early-stage syphilis, when the bacterial load is higher and the immune system is more actively engaged with the infection.

Relapsing fever, caused by a different group of spirochetes transmitted by ticks or lice, carries a substantial risk. A systematic review of tick-borne relapsing fever found that about one in five treated patients developed the reaction, and in that group, the fatality rate among those who reacted was around 6.5%.⁸ That makes relapsing fever the setting where the Jarisch-Herxheimer reaction is arguably most dangerous.

Lyme disease, caused by yet another spirochete, can also trigger the reaction, though it is less commonly reported. Most documented Lyme-associated cases occur during early-stage disease, and the presentation differs somewhat from that seen in chronic Lyme arthritis, where isolated reports exist but are rare.⁹

Leptospirosis, caused by Leptospira species, rounds out the major spirochetal triggers. A systematic review of published leptospirosis cases found that the most common pattern was sudden-onset rigors followed by a temperature spike, with or without a drop in blood pressure, almost always after the first antibiotic dose.¹⁰ One study from Okinawa noted an especially high rate of 82% in treated leptospirosis patients.¹¹

Predicting Who Will React

Not everyone treated for these infections develops the reaction, and researchers have looked for ways to predict who is at higher risk. One promising marker is C-reactive protein (CRP), an inflammatory marker that can be measured with a routine blood test. In a study of active syphilis patients, those who went on to develop the reaction had substantially higher CRP levels before treatment than those who did not. The difference was large enough that a pre-treatment CRP cutoff was able to identify every patient who later reacted, though with some false positives among those who did not.¹² Earlier work had also found that patients with higher pre-treatment CRP tended to suffer the most intense systemic reactions, even though CRP levels did not correlate with the severity of the skin flare or the degree of fever individually.¹³

The type of antibiotic also appears to matter. Bactericidal drugs, which directly kill bacteria, may produce a more abrupt release of inflammatory debris compared to bacteriostatic drugs, which merely slow bacterial growth. In the Okinawa leptospirosis study, 90% of patients treated with bactericidal antibiotics developed the reaction versus 57% of those given bacteriostatic ones, though the sample was small enough that the difference did not reach conventional statistical thresholds.¹¹ In practice, the choice of antibiotic for these infections is dictated by what works against the pathogen, so switching to a bacteriostatic drug purely to avoid the reaction is rarely realistic. But knowing this distinction helps explain why the reaction can vary in intensity depending on the regimen used.

When the Reaction Gets Serious

For most people, the Jarisch-Herxheimer reaction is unpleasant but self-limiting. There are situations, however, where it demands real medical concern.

Pregnancy is the highest-stakes scenario in routine clinical practice. In a study of pregnant women treated for syphilis, 73% developed fever, but what made the situation dangerous was that two-thirds also experienced uterine contractions and decreased fetal movement, typically starting alongside the fever. Three of the monitored patients showed transient late decelerations in fetal heart rate. Among women who experienced the reaction, three delivered infants with congenital syphilis, including one stillbirth, while none of the women who did not react had fetal treatment failures.¹⁴ The implication is not that treatment should be withheld; untreated syphilis in pregnancy carries its own devastating risks. But pregnant patients with syphilis should ideally be monitored in a clinical setting when treatment begins.

Neurosyphilis presents another serious scenario. When the infection involves the brain and spinal cord, the reaction can extend beyond fever into neuropsychiatric territory. In a large cohort of neurosyphilis patients treated with penicillin, a quarter of those who reacted developed neuropsychiatric symptoms.¹ A case report documented one patient who developed altered consciousness, pupillary changes, and focal neurological signs ten hours after starting penicillin.¹⁵ These neurological manifestations can mimic a stroke or worsening meningitis, which creates diagnostic confusion if clinicians are not anticipating the reaction.

Beyond these specific populations, case reports across spirochetal infections have described the reaction producing acute respiratory distress, myocardial injury, dangerous drops in blood pressure, and seizures.⁴ These severe outcomes are uncommon in the syphilis setting but more of a concern in relapsing fever, where, as noted earlier, the fatality rate among those who react has been reported at around 6.5%.⁸

Managing and Preventing the Reaction

There is no universally accepted way to prevent the Jarisch-Herxheimer reaction. In clinical practice, most management is supportive: fluids, rest, and antipyretics like acetaminophen or ibuprofen for fever and body aches. Patients being treated for infections known to trigger the reaction should be warned that symptoms may worsen before they improve, so they do not panic or assume the antibiotic is failing.

The most scientifically interesting prevention strategy tested so far involves blocking TNF-alpha, the inflammatory signaling molecule that drives much of the reaction. In a landmark trial, patients with louse-borne relapsing fever were given sheep-derived antibody fragments against TNF-alpha before receiving penicillin. The treatment suppressed the reaction, reduced the associated rise in interleukin-6 and interleukin-8, and was suggested as potentially useful in other forms of sepsis.¹⁶ Despite these promising results, anti-TNF pretreatment has not become standard practice, partly because the reaction is self-limiting in most clinical settings, partly because TNF-blocking agents carry their own risks, and partly because later research into TNF blockade for human sepsis generally found that the harms outweighed the benefits.⁷

Corticosteroids are sometimes given before treatment in high-risk situations, particularly in neurosyphilis or ocular syphilis, where the inflammatory surge could cause irreversible damage. Evidence for this approach is largely observational and anecdotal, and no large randomized trial has established a definitive benefit. Still, many infectious disease specialists consider it a reasonable precaution when the stakes of an uncontrolled inflammatory flare are high.

For pregnant patients, the main management strategy is monitoring rather than prevention. Because withholding treatment would allow syphilis to devastate the fetus, the standard approach is to treat in a setting where fetal heart rate and uterine activity can be observed, with the obstetric team prepared to intervene if needed.

The “Die-Off” Claim and Its Limits

Online health forums are full of references to “Herxheimer reactions” or “Herx” in the context of conditions well beyond the reaction’s established scope. The most widespread extension involves Candida, the yeast that causes thrush and vaginal infections. Some practitioners and patients describe a worsening of symptoms after antifungal treatment as a “Candida die-off” and equate it with the Jarisch-Herxheimer reaction, attributing symptoms to a sudden flood of toxins released by dying yeast cells.¹⁷

The problem with this framing is that the Jarisch-Herxheimer reaction was characterized, studied, and mechanistically explained in the specific context of spirochetal infections. The cytokine cascade described in research involves spirochetal lipoproteins activating specific immune pathways. Whether a genuinely analogous process occurs during antifungal treatment of Candida has never been demonstrated in rigorous clinical research. Yeast cells do release cell-wall components when they die, and some of those components can provoke immune responses. But calling any post-treatment symptom flare a “Herxheimer reaction” stretches the term well past the evidence supporting it.

This matters because the label can be used to explain away side effects that actually warrant medical attention. If you start a new medication and feel dramatically worse, the answer might be a drug reaction, a worsening infection, or a dosing problem rather than a temporary healing crisis. The legitimate Jarisch-Herxheimer reaction has a specific profile: it occurs within hours of antibiotic treatment for a confirmed spirochetal infection, presents with a recognizable cluster of fever and inflammatory symptoms, and resolves within roughly a day. Anything that does not fit that pattern deserves investigation on its own terms rather than a reassuring label borrowed from a different context.

How It Differs From Antibiotic Allergy

One of the more practical points of confusion for patients is distinguishing the Jarisch-Herxheimer reaction from a true allergic reaction to the antibiotic itself. Both can produce fever, skin changes, and a general feeling of being unwell. The timing and character of the symptoms help separate them. The Jarisch-Herxheimer reaction begins within hours of the first dose, produces flu-like symptoms and an intensification of existing skin lesions, and resolves in under a day. An allergic drug reaction more often produces hives, itching, throat tightness, or new rashes that were not present before treatment, and it may worsen with continued dosing rather than improving.

The distinction has real consequences. A patient who develops the Jarisch-Herxheimer reaction should continue treatment, because the reaction means the antibiotic is working. A patient with a true drug allergy needs to stop the medication and switch to an alternative. Mislabeling a Herxheimer reaction as an allergy could lead to the patient being denied penicillin in the future, which is the single most effective treatment for syphilis and several other spirochetal infections. Conversely, dismissing a genuine allergic reaction as “just Herxheimer” could put the patient in danger. When there is any doubt, clinicians rely on the clinical context: Was a spirochetal infection confirmed? Did the symptoms start within the expected window? Do the symptoms match the established pattern? If the answers line up, the reaction is almost certainly Jarisch-Herxheimer, and treatment should continue under observation.

1
PubMed Central. Jarisch-Herxheimer reaction as a complication of penicillin G and ceftriaxone treatment in neurosyphilis
2
JAMA Network Open. Jarisch-Herxheimer Reaction After Benzathine Penicillin G Treatment in Adults With Early Syphilis: Secondary Analysis of a Randomized Clinical Trial
3
PubMed Central. Jarisch‒Herxheimer reaction
4
PubMed Central. The Jarisch-Herxheimer Reaction After Antibiotic Treatment of Spirochetal Infections: A Review of Recent Cases and Our Understanding of Pathogenesis
5
PubMed. Proposed mechanisms and preventative options of Jarisch-Herxheimer reactions
6
QJM: An International Journal of Medicine. The effect of antibody against TNF alpha on cytokine response in Jarisch-Herxheimer reactions of louse-borne relapsing fever
7
Journal of Antimicrobial Chemotherapy. Cytokines involved in human septic shock–the model of the Jarisch-Herxheimer reaction
8
PLOS Neglected Tropical Diseases. Tick borne relapsing fever – a systematic review and analysis of the literature
9
PubMed Central. The Jarisch-Herxheimer reaction associated with doxycycline in a patient with Lyme arthritis
10
PLoS ONE. The Jarisch-Herxheimer Reaction in Leptospirosis: A Systematic Review
11
Transactions of The Royal Society of Tropical Medicine and Hygiene. Clinical characteristics of laboratory-confirmed leptospirosis in Okinawa, Japan, 1974–2015: high incidence of Jarisch–Herxheimer reaction
12
PubMed Central. High Baseline C-Reactive Protein Level Can Predict the Occurrence of the Jarisch-Herxheimer Reaction in Patients with Active Syphilis
13
PubMed. C-reactive protein, C3, C4 and properdin during the Jarisch-Herxheimer reaction in early syphilis
14
PubMed. The Jarisch-Herxheimer reaction complicating syphilotherapy in pregnancy
15
PubMed Central. Jarisch-Herxheimer reaction in a patient with neurosyphilis
16
PubMed. Prevention of Jarisch-Herxheimer reactions by treatment with antibodies against tumor necrosis factor alpha
17
Toxicon. Candida die-off: Adverse effect and neutralization with phytotherapy approaches