Herxheimer Reaction


Accepting New Patients Today, Call (202) 955-0003

Section 10: The Herxheimer Reaction

In a field in which clinical findings can be vague and imprecise, and where helpful monitoring laboratory tests are lacking, the Herxheimer reaction is an indispensable clinical tool in the treatment of persistent LD, or neuroborreliosis. In theory, Herxheimer reactions occur when an administered antimicrobial agent has successfully led to lyses of certain organisms (107). By definition this phenomenon is not unique to Borrelia burgdorferi. Fortunately, however, the Herxheimer reaction appears to provide a highly reliable barometer of therapy in Borrelia burgdorferi, so much so that a treatment course which lacks the Herxheimer response places the diagnosis of persistent Borrelia burgdorferi in serious doubt.

The reaction was first described in 1895 by an Austrian dermatologist Jarisch Adolf Herxheimer, who was practicing in Vienna, and later confirmed by his brother Karl Herxheimer, who was also a dermatologist, practicing in Frankfort (108,109). During these times, both physicians were responsible for treating syphilitic lesions and employed various preparations of mercury, arsenic and bismuth in these therapies. The key observation noted by both physicians was that, shortly after treatment of syphilitic skin lesions had been administered, many of their patients developed fever accompanied by rigors, drenching sweats, and nausea and vomiting. In addition, they found that the syphilitic skin lesions flared and became larger before healing; results were best in the patients that experienced this reaction, which typically lasted for 2 to 3 days.

The debate about the cause and nature of this predictable reaction has raged on for decades. Various theories have ranged from a vascular reflex mediated by the autonomic nervous system (110) to a direct toxic effect of the antimicrobial on tissues (111). In 1943 Mahoney described the first Jarisch-Herxheimer reaction in syphilitic patients treated with the relatively new antibiotic penicillin (112 ). Classically, the Jarisch-Herxheimer reaction occurred when treating the secondary stage of syphilis, at a time when a widespread rash may occur and the spirochetal burden is high. Even today, medical students are taught that a Jarisch-Herxheimer reaction occurs as a result of treatment of secondary syphilis. Most physicians are not aware that the Herxheimer reaction occurs in Borrelia burgdorferi infections and has been described in a variety of other diseases, many of them caused by spirochetal organisms such as Treponema pallidum (syphilis) and Borrelia burgdorferi in LD. A short list of treated spirochetal infections noted to manifest Herxheimer reactions includes Relapsing Fever (Borrelia recurrentis), Yaws (a subspecies of Treponema pallidum), Rat Bite Fever (spirullum minus), and perhaps Vincent’s Angina (spirochetal mouth forms) (113). Non-spirochetal infections manifesting the Herxheimer event after treatment include Brucellosis, Glanders, Anthrax, and even Leprosy (mycobacterium leprae) (113).

In 1972 Gudjonsson reported on a summary of experiments that entailed almost a decade of work (114). He concluded that the Herxheimer effect was not allergic in nature and was likely caused by a leukocyte pyrogen released at the time of phagocytosis. Most now believe that the pyrogens in question are exogenous pyrogens, or endotoxins, derived from components of the bacterial cell wall. In the case of Borrelia burgdorferi, the pyrogen is most likely the lipoprotein moiety which comprises the outer coat of the organism. These lipoproteins, specifically OspA and Osp B, have been shown to have potent B cell mitogenic and cytokine-stimulatory properties (78). It is widely recognized that antibiotic therapy may promote endotoxin release by virtue of its microbicidal effect which leads to the disintegration of the bacterial organism and exposure, or presentation, of endotoxin. On recognition of the endotoxin, polypeptides such as IL-1, interferons, or tumor necrosis factor (TNF), otherwise referred to as endogenous pyrogen (EP) or pyrogenic cytokines, are released by the monocyte/macrophage system (115). It should be pointed out that only minute quantities of EP are needed in order to generate fever and other systemic symptoms. In the extreme case, such as gram negative bacterial sepsis, high and persistent levels of endotoxin are present and lead to sepsis syndrome with capillary leak syndrome and vascular collapse. In the Borrelia burgdorferi model, with a relatively low number of organisms present and with limited and inconsistent die-off with each round of antibiotic therapy, one could conceive a model in which constitutional complaints, mediated by pyrogenic cytokines, are manifested in an ongoing and rather unpredictable manner. This then would represent the defining principle for the Herxheimer reaction in Borrelia burgdorferi infection. In predicting a pattern of response based on our knowledge of Borrelia burgdorferi infection, we would expect these symptoms to be worse initially, depending on die-off rates, and to not be life threatening, but likely to be life altering. Our clinical experience supports these precepts.

Published reviews have suggested a periodicity exists for Borrelia burgdorferi activity or replication, specifically that symptoms of fever and malaise, etc. occur at 4 week intervals (116). Others have created so called mathematical models to support this hypothesis. While we respect this author’s opinion, we find it difficult to understand how there could be synchronicity in any given polymorphic Borrelia burgdorferi population (often consisting of multiple strains) in any given host. The matter merits further study and validation. In females, increased symptoms and increased urinary shedding of Borrelia burgdorferi has been documented in the perimenstrual period, suggesting a hormonal influence (personal communication Dr. Nick Harris), and so it would seem possible that periodicity may exist in menstruating females (although our more seriously ill female LD patients routinely develop menstrual irregularity).

In our clinic, prior to starting any antimicrobial therapy, especially if our patient is naïve to treatment, we emphasize to the patient that they may notice certain significant clinical events while on therapy. In the occasional patient in whom we have no firm diagnosis but where we are suspicious enough to offer short-term empiric oral therapy, we are intentionally a bit vague about providing information to the patient about the Herxheimer effect, as we do not wish to influence a response by suggestion. We have found the Herxheimer response in LD to be as myriad as the course of persistent LD itself. It is naïve for one to expect to witness simply a flu-like syndrome, although this certainly happens. Instead, generally one sees an intensification of pre-existing symptoms, e.g. increased brain fog or muscle/joint pain, where these symptoms were reported prior to therapy. On the other hand, it is equally common to take reports of new symptoms, e.g. headache in a patient who previously reported symptoms other than headache. In general, the Herxheimer reaction is worse in our most seriously ill patients and most violent at the onset of therapy. The Herxheimer response typically occurs within 3 to 5 days, but may take up to 2 weeks to appear. These symptoms may persist for days or weeks and often become a major management concern as our patient may suffer considerably in the process of treatment. Eventually, as therapy progresses, we tend to witness a dampening of the intensity of the Herxheimer response, as well as some reports of positive clinical gains. Introduction of new therapy, as we cycle antibiotics through our treatment schedule (see treatment program to follow), invariably leads to intensification or new symptoms, all of which are unpleasant. In fact, if we do not observe a new response when therapy is added or substituted, we question the efficacy of our program. Later in this report, we refer to dermal or neurologic Herxheimer phenomena, which we feel reflect local manifestations of pyrogenic cytokines in response to treatment.

Regrettably, but not unexpectedly, we have treated a number of individuals whose Herxheimer experience is so intense and prolonged that continued treatment is virtually impossible. After exhausting all customary supportive and treatment measures, which incidentally never includes the use of systemic steroids, we have learned the art of using a balanced program of anti-oxidants, copius fluid administration, and appropriate wash out therapeutic periods during a treatment protocol. Much more of these concepts will be discussed in our next website edition of 2007.

The escalating headache symptoms experienced by some patients on intensive antibiotic treatment is a Herxheimer effect that merits special attention. This “Lyme” headache is thought to be linked to cerebral edema brought about by Borrelia burgdorferi die-off and the ensuing creation of inflammatory microfoci in the leptomeninges. In our experience, severe CNS symptoms, such as incapacitating headache, are more likely to occur when the patient has had prominent CNS symptoms or findings pre-therapy, e.g. encephalopathy, aseptic meningitis, optic neuritis, and so on. In 2-3% of our treated cases, we have had to resort to one or more therapeutic lumbar punctures to provide relief by lowering intracranial pressure.

Given the miserable experience which we may exact on our patients during therapy, we would perhaps be better off if we followed the advice of Russell McMillan, DDS, DPH, who wrote the Arthritis Trust of America in 1994 with his personal remedy for the Herxheimer reaction. “I take a saltz bath which consists of adding 1 cup salt, 1 cup soda, 1 cup Epsom salts, 1 cup aloe vera, to a hot bath which I remain in and keep hot for about 11/2 hours all the while consuming about 2 quarts of warm water. Evidently the perspiration and osmotic pressure removes the causative toxins. I find it quite helpful” (117). Hey, sounds ok to us.

Table of Contents

  • Contact Us

  • This field is for validation purposes and should be left unchanged.