PATHOGENESIS OF THE JARISCH-HERXHEIMER …receiving repository preparations of penicillin. The possibility that the Herxheimer reaction might be the result of a release of histamine

PATHOGENESIS OF THE JARISCH-HERXHEIMER REACTIONA REVIEW OF CLINICAL AND EXPERIMENTAL OBSERVATIONS*

BY

ALBERT HEYMAN, WALTER H. SHELDON, and LILIAN D. EVANSFrom the Departments ofMedicine and Pathology, Emory University School of Medicine, and

Grady Memorial Hospital, Atlanta, Georgia, U.S.A.

During the past few years we have been engaged ina clinical and experimental study of the patho-genesis of the Jarisch-Herxheimer reaction insyphilis. The results of this investigation havebeen published in a series of papers (Sheldon andHeyman, 1949; Sheldon, Heyman, and Evans,1951a, b, 1952). The present report summarizes ourobservations and presents our concepts of thepathogenesis of the Herxheimer reaction in syphilis,as well as in other infections. We shall use thisopportunity to speculate upon the significance of thereaction as a manifestation of a hypersensitivityphenomenon.The Herxheimer reaction has long been known to

be a common occurrence in the treatment of syphilis,but its pathogenesis has evoked little interest. Thereaction is generally attributed to the release ofendotoxins or spirochaetal breakdown substancesfollowing the initial administration of spirochaeti-cidal agents. These substances are thought toproduce systemic manifestations, such as fever,chills, and malaise, as well as an exacerbation ofsyphilitic lesions. The results of our studies seemto indicate that the Herxheimer reaction is a hyper-sensitivity phenomenon of the delayed type similarto the focal hypersensitivity type of tuberculinreaction.

Occurrence of Herxheimer-like Reactions in VariousDiseases

Herxheimer reactions occur not only in syphilis,but also in other spirochaetal infections, such asyaws (Dwindelle and others, 1946), Vincent'sinfection (Williams, 1941), and relapsing fever(Wolff, 1946). Recent reports indicate that thereaction may also appear in leptospiral diseases(Suchett-Kaye, 1951). It has been frequently

* Address delivered by Dr. Heyman to the Medical Society forthe Study of Venereal Diseases, September 28, 1951.

This study was aided by a grant from the National Institutes ofHealth United States Public Health Service.

observed in rat-bite fever caused by Spirillum minus(Brown and Nunemaker, 1942). The recognitionof Herxheimer reactions in this group of infectionscan probably be attributed to the use of organicarsenicals, which were the first effective chemo-therapeutic compounds. The development of newchemotherapeutic agents and antibiotics, however,has led to the recognition of Herxheimer reactionsin a variety of other infections. The methodscurrently employed in the treatment of brucellosisfrequently produce intensification of signs andsymptoms of the disease with transient episodes ofchills and fever. These reactions occur in 25 to50 per cent. of patients with brucellosis andfollow the use of aureomycin, chloramphenicol,terramycin, or combined sulphonamide-strepto-mycin therapy (Knight, 1950). The appearance ofthe reaction within a few hours after therapysuggests that the reaction is probably caused bydestruction of the causative organism. SimilarHerxheimer-like reactions have been described inpatients with glanders (Womack and Wells, 1949)and tularaemia (Foshay, 1947) treated with strepto-mycin. Mention has also been made recently of anexacerbation of symptoms in leprosy treated withdiasone (Faget and Erickson, 1948), and in anthraxtreated with aureomycin (Gold, 1950).The Herxheimer reaction in syphilis is thought to

be similar in many respects to the systemic type oftuberculin reaction, but there has been, to ourknowledge, no evidence that patients with tuber-culosis treated with either streptomycin or para-aminosalicylic acid show an intensification of signsor symptoms similar to the Herxheimer reaction.We have observed a few instances of transient feverduring the first few hours of streptomycin therapyin patients with tuberculosis, but no definite con-clusions could be drawn regarding its significance.Efforts to produce morphological or other evidenceof the Herxheimer reaction in tuberculous guinea-

50

on March 14, 2020 by guest. P

rotected by copyright.http://sti.bm

j.com/

Br J V

ener Dis: first published as 10.1136/sti.28.2.50 on 1 June 1952. D

ownloaded from

http://sti.bmj.com/

JARISCH-HERXHEIMER REACTION

pigs treated with streptomycin have been unsuccess-ful. Tissue removed from the draining sinus ofone patient with tuberculous lymphadenitis beforeand after the beginning of treatment with strepto-mycin showed morphological changes similar tothose observed in the lesions of syphilis during theHerxheimer reaction. We have had no opportunityof repeating this experiment.

Herxheimer Reaction in Various Stages of SyphilisEarly Syphilis.-The Herxheimer reaction is

thought to occur in approximately 40 per cent. ofpatients with early syphilis (Farmer, 1948). Theactual incidence, however, depends on the criteriadetermining the presence of the phenomenon, thetherapeutic agent used, and, perhaps, the type andseverity of the syphilitic lesions. The reactionusually consists of a single episode appearing withinthe first few hours of the institution of therapy.Multiple reactions have been observed, however, inpatients given initially small but subsequentlyincreasing doses of arsenicals or penicillin. Evidenceof Herxheimer reactions in cutaneous and mucosallesions may be observed for several days in patientsreceiving repository preparations of penicillin.The possibility that the Herxheimer reaction

might be the result of a release of histamine orsimilar substances suggested to us the use of anti-histaminic compounds, together with penicillin.The reaction occurred in patients with early syphilisreceiving either pyribenzamine or benadryl as oftenas in patients receiving antisyphilitic therapy alone(Stewart, 1949; Heyman and Sheldon, 1948).Attempts to find histamine-like substances in theserum of patients during the Herxheimer reactiongave negative results (Morrison, Heyman, and Shel-don, 1948).

Neurosyphilis.-The use of penicillin in the treat-ment of neurosyphilis produced more frequent andsevere Herxheimer reactions than were observedfollowing arsenical therapy. The frequency of thereaction in patients with symptomatic neurosyphilistreated with penicillin in this clinic is shown inthe Table. These patients were afebrile at thebeginning of therapy and received penicillin alone

TABLEHERXHEIMER REACTIONS IN NEUROSYPHILIS

FOLLOWING PENICILLIN THERAPY

No. Herxheimer ReactionDiagnosis Patients -__________Fm_r___________Symptomatic Fever Only Total

Paresis .. 44 6 17 23Tabes 8 1 1 2Meningomyelitis 19 2 2 4

in dosages varying from 1,000 to 50,000 units.During the first 24-36 hours after penicillin therapy,23 (52 per cent.) of the 44 patients with paresisshowed a definite febrile reaction with a rise oftemperature to 100° F. or more. Six of these 23patients also showed a transient intensification oftheir psychotic manifestations. Of eight patientswith tabes, two had evidence of a febrile reaction,one of whom also showed an increase in lightningpains. Two patients with meningomyelitis developedmore severe spinal cord manifestations which mayhave been the result of a Herxheimer reaction(Jones, Heyman, Smith, and Wilson, 1951).

It has been estimated that approximately 64 to79 per cent. of all patients with neurosyphilistreated with penicillin develop Herxheimer reactions(Hoekenga and Farmer, 1948). The incidence isparticularly high in paresis and seems to be morecommon in patients with markedly abnormal spinalfluid findings. The reactions in neurosyphilisusually consist of elevation of temperature, butoccasionally irreversible damage develops. Deathhas been reported after penicillin therapy in apatient with cerebral gumma (Scott, Maxwell, andSkinner, 1949), and in another with syphiliticpachymeningitis (Shaffer and Shenkin, 1950).Histological study of the nervous system of thesepatients revealed changes in the syphilitic lesionsresembling those occurring in early syphilis duringthe Herxheimer reaction. We had an opportunityof examining histological preparations of thegumma mentioned above.* The lesion showedmarked congestion, acute inflammation, and oedema,findings usually absent in patients with uncompli-cated gummatous lesions. There is little doubtthat the Herxheimer reaction was an importantfactor in the death of these patients.The changes in cerebral circulation and meta-

bolism caused by the Herxheimer reaction werestudied in a few patients with neurosyphilis (Heyman,Patterson, Nichols, and Jones, 1951). This investi-gation was carried out with the use of the nitrousoxide technique in six patients with paresis and inone with meningovascular syphilis. Only as lightincrease in the mean cerebral blood flow of thesepatients was observed. Two of the patients withclinical evidence of a Herxheimer reaction showed afall in cerebral metabolism. We are not certain ofthe significance of these findings because of thesmall number of patients studied. The expectedincrease in the cerebral circulation may have beenprevented by the marked oedema of the tissueswhich appears during the Herxheimer reaction.

* We are indebted to Dr. Virgil Scott, St. Louis, Mo, for sendingus the sections of this case.

51



j.com/

Br J V


ownloaded from

http://sti.bmj.com/

BRITISH JOURNAL OF VENEREAL DISEASES

Cardiovascular Syphilis.-The danger of a seriousHerxheimer reaction in cardiovascular syphilis afterpenicillin therapy is thought to have been greatlyexaggerated and is now believed to be of littlesignificance (Moore, 1949; Sinclaire and Webster,1951). Detailed information regarding the fre-quency of febrile Herxheimer reactions in patientswith cardiovascular syphilis and co-existent nervoussystem involvement is given in several reports(Tucker and Farmer, 1947; Sinclaire and Webster,1951 ; Wheeler and Curtis, 1951 ; Coale, Allen,and Delp, 1950). Only fourteen of 155 patientswith either aortic insufficiency or aneurysm developedfebrile reactions after penicillin-an incidence

*a...w< b :

:.0W~~ ~.AR t.......

X ~~~~~~~~A,, ., ,,J , ..6.~ ~ ~ ~ t

S~~~~~S

*~$ .1; :4*N ;;: ,,;

s; t < <S st4

(a) Before treatment. Connective tissue beneathepithelium (left upper corner) shows chronicinflammation with a few small blood vessels.

) ~~~~~~~0:s.Xt !:.4 .:i4 :..%' ,. . +.izg4*R.z t i ^ 4,

%* % x

(c) 18 hrs after treatment. Blood vesselsremain prominent. Leucocytes have disappeared.Stroma now infiltrated by large mononuclear cells.Epithelium (left upper corner) normal.

FIG. 1.-Condylomatous lesion.

considerably less than that occurring in earlysyphilis or neurosyphilis. Moreover, thirteen ofthe fourteen patients with Herxheimer reactions hadevidence of neurosyphilis. This suggests that thereaction might be related to the co-existing neuro-syphilis rather than to the cardiovascular syphilisalone. The development of angina pectoris,increased congestive failure, or rupture of anaortic aneurysm have been attributed to the effectsof penicillin therapy (Scott, Maxwell, and Skinner,1949; Dolkart and Schwemlein, 1945; Porter,1948; Callaway and others, 1946), but there is noclear-cut evidence that these manifestations wererelated to a Herxheimer reaction.

A.8w o W3 > e d o?S

r~~~~~~~t- * b a 9' g<-i. M

' t. '~~Ai *

W1 s zC.A.

', tr4 , ' C a*, ''

(b) 6 hrs. after treatment. Blood vessels nowprominent and crowded with leucocytes, whichinfiltrate vessel walls and oedematous connective

-- tissue. The epithelium (left) is also oedematous.

z.<A t >.°.:-.4~ ~ ~~4

4.r

(d) 72 hrs after treatment. Acute inflammatorychanges have subsided and the picture is similar tothat before treatment (Fig. la).

Phloxine-methylene blue x 530.

52



j.com/

Br J V


ownloaded from

http://sti.bmj.com/


Despite the lack of a well-authenticated instanceof a serious Herxheimer reaction in cardiovascularsyphilis, we believe that such reactions are possible.This opinion is based upon the study of a patient inthis hospital with gummatous aortitis who died afew hours after an initial administration of penicillin(Whorton and Denham, 1951). At autopsy, theaortic lesions showed the acute inflammatorychanges associated with the Herxheimer reaction.The death of this patient was not directly related tocardiovascular syphilis and was probably caused bypulmonary embolism. Gummatous aortitis is rela-tively uncommon and constitutes less than 3 percent. of patients with syphilitic aortitis (Gordon,Parker, and Weiss, 1942). As yet, there are noreports regarding histological evidence of a Herx-heimer reaction in patients with non-gummatoussyphilitic aortitis. Serious Herxheimer reactionsare therefore not of practical importance in themanagement of cardiovascular syphilis.

Morphological Changes in Syphilitic Lesions duringthe Herxheimer Reaction

We have made numerous histological studies ofsyphilitic lesions in both patients and experimentalanimals during the Herxheimer reaction (Sheldonand Heyman, 1949). At the beginning of eachexperiment, lesions similar in size and appearancewere selected and one was excised before theinitiation of specific therapy. The other lesionswere removed at varying intervals after therapy.Histological changes consisting of transient acute in-flammation were observed in the syphilitic lesions 4 to6 hours after the beginning of treatment. At first, thecapillaries and small blood vessels (chiefly veins)became congested. They later became crowded withneutrophilic polymorphonuclear leucocytes, thevascular endothelium swelled, and the leucocytesmigrated through the vessel wall into the surroundingtissues which were oedematous. The acute processsubsided after 14 to 18 hours, at which time theoedema disappeared and large mononuclear cellsreplaced the polymorphonuclear leucocytes. Thechanges were no longer present 72 hours afterinitiation of therapy (Figs 1 and 2).These changes were encountered in practically

every patient with clinical evidence of a Herxheimerreaction. They do not occur in the natural courseof syphilitic infection, but are observed only afterspecific therapy. In our experience, the histologicalchanges were a more sensitive criterion of thepresence of the Herxheimer reaction than eitherelevation of temperature, leucocytosis, or thedevelopment of local or systemic manifestations.

FIG. 2.-Condylomatous lesion. Phloxine-methyleneblue x 690.

k4'~~~~~~~~

Aft*

(a) Before treatment. Connective tissue beneathepithelium (left) shows slight chronic inflammation.Capillaries not visible.

:. :

sSLr- .

i:: .:...

,<....... . :.:.0..... ..I

ii ....,,w.

(b) 4 hrs after treatment. Many congestedcapillaries containing leucocytes now occupy con-nective tissue beneath epithelium (top).

we....*? .t q!e.. , . j,;U -Aw

Sv:|: ..

*';'s*:_::*.

(c) 7 hrs after treatment. A small vein beneaththe epithelium (right upper corner) is crowded withleucocytes which have begun to infiltrate theoedematous connective tissue.

53

AsAL-

:. ..:...t 's

::.



j.com/

Br J V


ownloaded from

http://sti.bmj.com/


%~~~ ~ ~#.. $ *

%~~~~~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~~~~~ X-. B,, ,

*4~~~~~~4

-A

4L~~~~~~~~~~

4SWW.. Aw..

.4.r ,a£ ;R

(a) Lesion excised before injection of penicillin.Polymorphonuclear leucocytes are sparse amongthe inflammatory-cell infiltrate.

(a) Excised before penicillin treatment. No poly-morphonuclear leucocytes among inflammatory cells.

ment polmop :ouc ear ucct sno poi

nent part ofin*S' mmatory process.

44

(b) Excised 4 hrs after initial penicillin treat-

ment, polymorphonuclear leucocytes now a promi-nent part of inflammatory process.

; t x * ^

' ,.M ~ * 4 . X

(b) Excised 7 hrs after administration of penicillin.Many polymorphonuclear leucocytes.

FIG. 4.-Skin lesion of rabbit infected with S. minus.

Sr;Vr*t;,;,^,* ,, iX frUt , Phloxine-methylene blue x 480.

Herxheimer Reactions in Laboratory Animals

.. Histological changes similar to those observed in

..' human syphilitic lesions were produced in rabbits

t4O With experimental syphilis by the administration of

Aw.*0AJ*$) % i* penicillin or arsenoxide (Fig. 3).~~We made num6rous attempts to elicit these

;4r>., 4 changes by the injection of living and dead Tre-,<§tt,;,t 1* *A',:Wtt",,,w,,,> ;49*1i i "ponema pallidum. Before injection, the treponemata>^Se<J>:s9 t ; :*! were treated in various ways in an attempt to

* release antigenic substances. The procedures in-Excised '~%u cluded heating, repeated freezing and thawing,(c) Excised 8 hrs after treatment. Leucocytes numerous. '

FIG. 3.-Herxheimer reaction in skin syphiloma of rabbit after penicillin. Phloxine-methylene blue x 480.

54



j.com/

Br J V


ownloaded from

http://sti.bmj.com/


(a) Excised before injection of specific immune

serum. Polymorphonuclear leucocytes not part ofinflammatory reaction.

4 RO

A , 'J~~~~

(c) Excised before injection of normal serum.

FIG. 5.-Skin lesion of rabbit infected with

exposure to spirochaeticidal drugs and incubationwith immune serum, but the results of these experi-ments were inconclusive. Attempts to demonstratetreponemal toxins by intravenous and intracerebralinjection of living and dead spirochaetes into youngmice also produced no results.

Since Herxheimer reactions are known to occurin patients with Spirillum minus infections followingspirillicidal therapy, we began a study of thisinfection in laboratory animals (Sheldon, Heyman,and Evans, 1951a). Biopsies of the cutaneous

AW ... *'

* kg,*7* *e EEv

(r. V4A C EU.

immune ~serum. Noedsintiflrainwtpoyorhnula leucrcytes

*T, .t. _*e ,.

* r-s.^.._'¶¶u& 4 *

Af7 .* 4 . .I- .:

46.~~~~~~~~~~~~~~~U

(b) Excised 8 hrs after injection ofnra specific

immune serum Note-hdistinctral infiltreationewth

rato wihiflrtoofpolymorphonuclearlecyts

* -~~~~~~W%~~~~~~~

(a)~~~~~~~~~~~~~~bExcisedbeorbsatrinjectionofspecificimn

withny6mor8phonucers. uocts

45L A

~~~#4e~~~~4

(d) Excised 8 hrs after injection of normal serum.No appreciable difference between (c) and (d).

S. minus. Phloxine-methylene blue x 480.lesions of rabbits infected with S. minus were takenbefore and at 2-hour intervals more treatmentwith penicillin (Fig. 4). Changes very similar tothose observed in syphilitic lesions of rabbitsfollowing penicillin therapy were found in the skinlesions of the Spirillum-infected animals. Thesechanges consisted of a transient, acute inflammatoryreaction with infiltration of polymorphonuclearleucocytes, oedema and congestion. The reactionappeared within 4 hours and reached its peakwithin 6 to 8 hours.

55



j.com/

Br J V


ownloaded from

http://sti.bmj.com/

BRITISH JOURNAL OF VENEREAL DISEASES..... }.;.

- a::* ~~~~~~~~~~~~. ..........

40

:,.9t $s'w' .*

(a) Lesion excised before injection of serum.

(b) Excised 4 hrs after injection.

Ap"2...ii!.......:

*.:: *|a-t:::. .

4~~~~~~4'IDli Mv

(c) Excised 8 hrs after injection. Note marked

diffuse infiltration with polymorphonuclear leuco-

cytes, which in (b) pack a small vessel (right centre).

FIG. 6.-Herxheimer reaction in skin syphiloma of

rabbit after injection of syphilitic serum. Phloxine-methylene blue x 480.

Lysing and immobilizing antibodies against S.minus are known to be present in the serum ofanimals infected with this organism (McDermott,1928). We found that dilutions of serum(1: 4 or less) of rabbits infected with S. minus werecapable of rapidly immobilizing and lysing thisorganism. A single large dose (70 ml.) of serumpooled from rabbits with untreated S. minus infectionwas injected intravenously into rabbits with skinlesions of this disease. The skin lesions excisedfrom the recipient animals following injection ofimmune serum showed histological changes similarto those observed after administration of penicillin.The lesions of rabbits receiving an equal amount ofserum pooled from uninfected animals showed nodiscernible changes (Fig. 5, previous page).

These observations led us to investigate theeffects of administration of syphilitic immuneserum in rabbits with experimental syphilitic infec-tion (Sheldon, Heyman, and Evans, 1951b). Serumfrom rabbits with untreated syphilis of 8 to 10weeks duration is known to contain immobilizingantibodies for T. pallidum (Nelson and Mayer, 1949).Immune serum (70 ml.) pooled from infected animalswas injected intravenously into each of six rabbitswith skin syphilomata. Individual syphilomata wereexcised from the recipient animals before and at2-hour intervals following injection of the serum.The lesions of 5 of the 6 animals showed histologicalevidence of a transient acute inflammation similarto that observed in Herxheimer reactions producedby penicillin (Fig. 6). The skin lesions of an equalnumber of control rabbits, receiving serum pooledfrom uninfected rabbits, showed no changes (Fig.7, opposite). The serum used was bacteriologicallysterile and care was taken to avoid contaminationby pyrogenic substances.These experiments suggested that the Herxheimer

reaction might be a hypersensitivity phenomenon inwhich the hosts' tissues reacted with substancesreleased by destruction of T. pallidum. The effectof ACTH on the Herxheimer reaction was thenstudied (Sheldon and others, 1952), as it was thenthought to inhibit or suppress the delayed type ofhypersensitivity reaction (Long and Favour, 1950).ACTH (in daily doses of 5 mg./kg.) was givenintramuscularly to rabbits with skin syphilomata.Lesions excised on the third day of administrationof ACTH were compared with those removed beforethe onset of hormone therapy. A marked decreasein the intensity of the inflammatory and reparativeprocesses was evident following the administrationof ACTH. The polymorphonuclear leucocyteswere reduced in number and the collections oflymphocytes at the periphery of the lesions were

56

:k ...:..:.

i.....W.F.1.



j.com/

Br J V


ownloaded from

http://sti.bmj.com/


less extensive. There was no decrease in thenumber of large mononuclear cells. New for-mation of fibroblasts and capillaries was diminished,and the central oedematous core of the lesions hadlargely disappeared. The animals were then givenpenicillin therapy and other lesions were excised atregular intervals, while ACTH therapy was con-

tinued. Histological changes of the Herxheimerreaction were readily visible in these lesions andresembled in every respect the changes developingafter penicillin alone (Fig. 8, overleaf). ClearlyACTH does not inhibit the development of themorphological changes of the Herxheimer reaction.

DiscussionDefinite histological changes have been observed

in the syphilitic lesions of both patients and experi-mental animals during the Jarisch-Herxheimerreaction. These changes consist of a nonspecific,transient, acute inflammation and are believed to bethe morphological basis for the clinical mani-festations of the Herxheimer reaction. The histo-logical picture and the time of appearance of thereaction resemble those occurring in tuberculouslesions following the introduction of large amountsof tuberculoprotein in hypersensitive individuals(Rich, 1944). In experimental syphilis, however,the reaction could not be produced by the admini-stration of spirochaetal antigens.At present, it appears that the Herxheimer

phenomenon in syphilis can be elicited only whenspirochaetes are destroyed by administration of eitherappropriate antibiotics, chemotherapeutic agents orspecific immune serum. Such serum has beenshown to contain immobilizing antibodies for T.pallidum (Nelson and Mayer, 1949), which probablydestroy sufficient numbers of treponemata to evoke aHerxheimer reaction. Destruction of the causativeorganism seems to be an essential factor in thepathogenesis of the reaction. This is furtherindicated by the demonstration of Herxheimer-likereactions in a variety of nonspirochaetal infectionstreated with new antibiotic agents.

Destruction of the aetiological agent, however, isnot sufficient in itself to explain many of the knownfactors regarding the Herxheimer phenomenon. If,as Moore, Farmer, and Hoekenga (1948) havepointed out, the reaction were due solely to therelease of spirochaetal breakdown substances, itsincidence and severity should vary with (1) thenumber of treponemata present in a given individualat the time of treatment, and (2) the number oforganisms destroyed by specific therapy. Theseworkers, however, have observed that the incidenceof Herxheimer reactions in patients with sero-

KC~~~~~~--4;

*s

4~~~~~I

*~~~~t#>+q5%sw +*

.>~~ ~ 9*~~~~~~

(a) Lesion excised before injection of serum.

x ** s t

1*.

F -f c nW s

w S*(b) Excised 4 hrs after injection.

i,*4-gi* )s 54 *+qW 2 ;

.4"

~~~~~~~4'~~ ~ ~ ~ ~ ~ ~ ~~~4t£o*(c) Excised 8 hrs after injection.

FIG. 7.-Injection of normal serum producing nochanges in skin syphiloma of rabbit. Phloxine-methylene blue x 480.

57



j.com/

Br J V


ownloaded from

http://sti.bmj.com/


~ ~ ~ ~~~~~~~~~~~~~

'_. ....... 6.

..

-o.XAz y ;~~~~~~~~~~~~~~~g

.........*:C .........(a) Before administration of ACTH or penicillin.

b) Syphiloma excised on 3rd day of ACTHtreatment immediately before injection of penicillin.Decreased cellularity indicates diminished in-flammatory-cell infiltration.

4~~~~~~~~:.;','-.

jection. Acute inflammatory celloedema.

._.~........^.,t~~~~~~~~~~~~~~~~~~~~~~~~~~~....(e) Syphiloma excised 8 hrs after penicillin in-jection. Marked acute inflammatory-cell infiltration.

*~~~~~~~~~~~~~~~~~~~~~~~.........

f)Syphiloma excised 24 hrs after penicillin in-

jection Subsidence of inflammationFIG. 8.-Skin syphiloma of rabbit. Phloxine-methylene blue x 660.

58



j.com/

Br J V


ownloaded from

http://sti.bmj.com/


negative primary syphilis is approximately the sameas in patients with secondary syphilis, even thoughthe latter group presumably harbours a greaternumber of treponemata. Moreover, the reactionoccurs in patients with early syphilis who receivepenicillin in doses too small to produce any apparentdecrease in the number of spirochaetes in cutaneousand mucosal lesions. It is further known thatHerxheimer reactions occur in lesions of latesyphilis in which spirochaetes are notoriously few.These apparently conflicting observations can be

understood if it can be shown that the Herxheimerreaction is a hypersensitivity phenomenon. In ahypersensitive host the liberation of even minuteamounts of antigenic spirochaetal substances couldproduce a severe systemic reaction. The incidenceand severity of the reaction would then be deter-mined primarily by the degree of hypersensitivityrather than the number of organisms. A Herx-heimer reaction, therefore, would be produced byany dose of penicillin destroying enough treponematato release a sufficient quantity of spirochaetalproducts to act as an antigen.The failure of ACTH to inhibit the Herxheimer

reaction does not conflict with the concept that thisphenomenon is a hypersensitivity reaction. It hasrecently been shown that this hormone does notinterfere with the basic immune mechanisms of thedelayed type of hypersensitivity (Sheldon, Cummings,and Evans, 1950; Derbes and others, 1950). Thefailure of antihistaminic agents, such as pyriben-zamine, to inhibit the Herxheimer reaction suggeststhat the release of histamine is not a factor in theproduction of the reaction.Regarding the role of hypersensitivity in the

pathogenesis of the Herxheimer reaction, it isimportant to point out that hypersensitivity of thetuberculin type is present in all but one (anthrax) ofthe nonspirochaetal infections in which Herxheimer-like reactions are now being observed. Positiveskin reactions of the tuberculin type develop inbrucellosis, leprosy, tularemia, and glanders followingintradermal inoculation of specific antigens. Itappears, therefore, that the existence of the delayedtype of hypersensitivity may be an essential factorin the occurrence of the Herxheimer reaction. Thelack of suitable antigens, rather than the absenceof hypersensitivity, may explain the absence of skinreactions in patients with spirochaetal diseases. Itis probable that accurate assessment of the role ofhypersensitivity in the pathogenesis of the Herx-heimer reaction will be determined only aftercultivation of the Treponema pallidum has beenachieved and antigenic fractions of the organismhave been isolated.

The delayed type of hypersensitivity, manifestedby a positive skin reaction to specific antigens,exists in such virus and fungus infections as lympho-granuloma venereum, mumps, coccidioidomycosis,and histoplasmosis. We believe that, as newchemotherapeutic agents are developed for thesedisorders, Herxheimer reactions will no longer beconfined to the spirochaetal diseases, but will beencountered in a variety of infections.The prevention of the Herxheimer reaction in

neurosyphilis is of practical importance. It hasbeen shown that the use of small doses of eitherpenicillin or arsenicals in initiating therapy will notprevent the reaction. Although initial treatmentwith bismuth over a period of several weeks mayprevent a Herxheimer reaction in early syphilis,there is no evidence that such treatment is effectivein neurosyphilis. In fact, Sinclaire and Webster(1951) have reported that the use of heavy metals upuntil the time of penicillin therapy did not preventHerxheimer reactions in five patients having bothcardiovascular and neurosyphilis. Fever therapyhas been suggested as a means of preventing theHerxheimer reaction in patients with paresis. Thismay not be practical, however, since fever therapyitself often produces untoward complications andis thought by many workers to be unnecessary in thetreatment of neurosyphilis. It must be concludedthat at present there is no practical nmethod forpreventing the Herxheimer reaction in patients withneurosyphilis.

SummaryClinical and experimental observations on the

pathogenesis of the Jarisch-Herxheimer reaction arereviewed. Available evidence indicates that theHerxheimer reaction should be regarded as ahypersensitivity phenomenon of the delayed type.

REFERENCESBrown, T. McP., and Nunemaker, J. C. (1942). Bull.

Johns Hopk. Hosp., 70, 201.Callaway, J. L., Noojin, R. O., Flower, A. H., Kuhn,

B. H., and Riley, K. A. (1946). Amer. J. Syph.,30, 110.

Coale, L. H., Allen, M. S., and Delp, M. H. (1950).J. Kans. med. Soc., 51, 102.

Derbes, V. J., Dent, J. H., Weaver, N. K., and Vaughan,D. D. (1950). Proc. Soc. exp. Biol., N.Y., 75, 423.

Dolkart, R. E., and Schwemlein, G. X. (1945). J.Amer. med. Ass., 129, 515.

Dwindelle, J. H., Rein, C. R., Steinberg, T. H., andSheldon, A. J. (1946). Amer. J. trop. Med., 26, 311.

Faget, G. H., and Erickson, P. T. (1948). J. Amer. med.Ass., 136, 451.

Farmer, T. W. (1948). Ibid., 138, 480.Foshay, L. (1947). Amer. J. Med., 2, 467.Gold, H. (1950). Ibid., 8, 31.

59



j.com/

Br J V


ownloaded from

http://sti.bmj.com/


Gordon, W. H., Parker, F., and Weiss, S. (1942).Arch. intern. Med., 70, 396.

Heyman, A., Patterson, J. L., Nichols, F. T., and Jones,R. W. (1951). Amer. J. Syph., 35, 301.

- , and Sheldon, W. H. (1948). Unpublished data.Hoekenga, M. T., and Farmer, T. W. (1948). Arch.

intern. Med., 82, 611.Jones, R. W., Heyman, A., Smith, W. A., and Wilson, R.

(1951). Amer. J. Syph., 35, 72.Knight, V. (1950). Ann. N. Y. Acad. Sci., 53, Art. 2,

p. 332.Long, J. B., and Favour, C. B. (1950). Bull. JohnsHopk. Hosp., 87, 186.

McDermott, E. N. (1928). Quart. J. Med., 21, 433.Moore, J. E. (1949). Amer. J. Syph., 33, 43.-, Farmer, T. W., and Hoekenga, M. T. (1948).Trans. Ass. Amer. Phys., 61, 176.

Morrison, J. L., Heyman, A., and Sheldon, W. H.(1948). Unpublished data.

Nelson, R. A., and Mayer, M. M. (1949). J. exp.Med., 89, 369.

Porter, R. R. (1948). Va. med. Mon., 75, 357.Rich, A. R. (1944). "Pathogenesis of Tuberculosis".Thomas, Springfield, Illinois.

Scott, V., Maxwell, R. W., and Skinner, J. S. (1949).J. Amer. med. Ass., 139, 217.

Shaffer, B., and Shenkin, H. A. (1950). Amer. J. Syph.,34, 78.

Sheldon, W. H., Cummings, M. M., and Evans, L. D.(1950). Proc. Soc. exp. Biol., N. Y., 75, 616.

and Heyman, A. (1949). Amer. J. Syph., 33, 213.and Evans, L. D. (1951a). Ibid., 35, 411.

-, (195lb). Ibid., 35, 405.(1952). Ibid., 36, 77.

Sinclaire, H. A., and Webster, B. (1951). Ibid., 35, 312.Stewart, J. J. (1949). Arch. Derm. Syph., Chicago,

60, 427.Suchett-Kaye, A. I. (1951). Lancet, 1, 90.Tucker, H. A., and Farmer, T. W. (1947). Arch.

intern. Med., 80, 322.Wheeler, C. E., and Curtis, A. C. (1951). Amer. J.

Syph., 35, 319.Whorton, C. M., and Denham, S. W. (1951). Ibid.,

35, 255.Williams, R. H. (1941). Arch. intern. Med., 68, 80.Wolff, B. P. (1946). Ann. intern. Med., 24, 203.Womack, C. R., and Wells, E. B. (1949). Amer. J.

Med., 6, 267.

60



j.com/

Br J V


ownloaded from

http://sti.bmj.com/

Documents

PATHOGENESIS OF THE JARISCH-HERXHEIMER …receiving repository preparations of penicillin. The possibility that the Herxheimer reaction might be the result of a release of histamine