granulomatous infections: etiology and classification...granulomatous infections: etiology and...
TRANSCRIPT
146
Granulomatous Infections: Etiology and Classification
Alimuddin Zumla and D. Geraint James From the Academic Infectious Diseases Unit, Department ofMedicine,University College London Medical School; and the Royal Free
Hospital School ofMedicine, London, United Kingdom
Granulomatous disorders are frequently due to a wide variety of infections. Over the past decadeadvances in molecular diagnostic techniques have allowed identification of organisms involved ingranulomatous disorders that previously were of unknown etiol~gy. ~n the basis of cu~ently available information, granulomatous infections can now be classified m three categones. Group 1infections are due to a well-recognized organism. Group 2 comprises infections due to organismsthat have been recently identified in granulomas by molecular methods but are not readily isolatedby conventional microbiological techniques. Group 3 consists of disorders for which the causalorganisms have not yet been identified but are strongly suspected; further a~van~es in di~gnostic
techniques will lead to reclassification of some of these disorders as grou~ 2. ThIs r~vlew descnbe~ theetiology,histopathologic features, and classificationof granulomatous disorders, with an emphasis onthose of groups 2 and 3.
Granuloma Formation
A granuloma can be defined as a focal, compact collectionof inflammatory cells in which mononuclear cells predominate.Granulomas usually form as a result of the persistence of anondegradable product or as the result of hypersensitivity responses [1]. An overlap of the two mechanisms occurs in mostinfectious diseases since microorganisms can serve both asforeign bodies and as antigens for immunologic responses.Normally granulomas are the result of protective mechanismsand form when acute inflammatory processes cannot destroyinvading agents.
The granuloma forms by a stepwise series of events and isthe end result ofa complex interplay among invading organism,antigen, chemical, drug or other irritant, prolonged antigenemia, macrophage activity, T cell responses, B cell overactivity,circulating immune complexes, and a vast array of biologicalmediators [2]. Areas ofinflammation or immunologic reactivityattract monocyte-macrophages, which may fuse to form multinucleated giant cells [3]. Further cellular transformation ofmacrophages to epithelioid cells may occur.
The granuloma is an active site of numerous enzymes andcytokines and, with aging, fibronectin and progression factorssuch as platelet-derived growth factor, transforming growthfactor-,B, insulin-like growth factor, and TNF-a. There is aclose relationship between CD4+ T lymphocytes and activatedmacrophages showing increased expression ofmajor histocom-
Received 21 August 1995; revised 23 February 1996.Financial support: Camden and Islington Community Health Services NHS
Trust.Reprints or correspondence: Professor D. Geraint James, Visiting ~rofessor
of Medicine, Royal Free Hospital School of Medicine, Rowland HIll Street,London NW3 2PF, United Kingdom
Clinical Infectious Diseases 1996;23:146-58© 1996 by The University of Chicago. All rights reserved.1058-4838/96/2301-0020$02.00
patibility (MHC) class II molecules. The T helper cells recognize protein peptides presented to it by antigen-presenting cellsbearing MHC class II molecules.
The T cell induces IL-l on the macrophage, and thereaftera cavalcade of chemotactic factors promote granuloma genesis[2, 4]. IFN-'}' increases the expression of MHC class II molecules on macrophages, and activated macrophage receptorscarry a crystallizable fraction of IgG to potentiate their abilityto phagocytose [4]. The end result is the epithelioid granuloma,which progresses toward fibrosis under the impact of transforming growth factor and platelet-derived growth factor.
T cell subsets, macrophages, and other immune cells produceseveral cytokines, the amounts and types of which determinefurther subclassifications of granulomas. CD4+ T helper (Th)lymphocytes are necessary for the development of granulomasand are broadly classified into two functional types, Thl andTh2 [5]. Thl-type cells produce IL-2, IFN-'}', and TNF-,B uponstimulation with antigen and also participate in delayed-typehypersensitivity responses, while Th2-type cells make IL-4,IL-5, IL-6, IL-9, IL-I0, and IL-13, cytokines which are important for development of B cells and eosinophilia.
There appears to be a complex relationship between Thltype and Th2-type responses and infectious granulomas. Granulomas that synthesize predominantly Th2-type cytokines, suchas those that form in response to parasite ova, make only smallquantities of Thl-type cytokines chronically [6], whereas granulomas such as those of tuberculoid leprosy make large quantities of Thl-type cytokines and less of Th2-type lymphokines[7]. Granulomas of various infections may have different immunoregulatory mechanisms governing their formation andresolution [I, 2, 7].
Classification of Granulomatous Infections
Granulomatous disorders are frequently due to a wide varietyof infections. Over the past decade advances in molecular diag-
by guest on September 23, 2016
http://cid.oxfordjournals.org/D
ownloaded from
em 1996;23 (July) Granulomatous Infections 147
Table 1. Classification of granulomatous infections and associatedpathogens.
Group 1 Granulomatous Infections (Causal AgentsWell Defined)
nostic techniques have allowed identification of organismsinvolved in granulomatous disorders that previously were ofunknown etiology. On the basis of currently available information, granulomas associated with infections can be categorizedin three groups, as listed in table 1.
Group 1 granulomatous disorders are caused by well-recognized agents. In group 2, the causal agents have been suspectedover the years but have only recently been recognized by newerdiagnostic molecular techniques such as PCR. These agents arenot readily isolated by conventional bacteriologic techniques.Group 3 comprises granulomatous disorders that may be associated with particular infectious agents but for which the causalorganisms have not yet been determined. Future advances indiagnostic techniques may provide strong evidence of thecausal agents, in which case the disorders may be reclassifiedas group 2; any such advancements will allow better management of these idiopathic granulomatoses.
Group number,type of causal agents
Group IWell recognized
Mycobacteria
Bacteria
SpirochetesFungiProtozoaNematodesTrematodes
Chlamydia species
Rickettsia
Viruses
Group 2Recently recognized
Bacterium
Actinomyces
Group 3Suspected but not established
Measles virusMycobacterium? (see table 4)
Viral
?
Granulomatous disorders
Tuberculosis, leprosy, Buruli ulcer,
swimming pool (fish tank)
granuloma
Brucellosis, melioidosis, actinomycosis,
nocardiosis, granuloma inguinale,listeriosis, tularemia
Syphilis, pinta, yawsMycoses (see table 3)Leishmaniasis, toxoplasmosisVisceral larva migrans (toxocariasis)Schistosomiasis, paragonimiasis,
fascioliasis, clonorchiasisLymphogranuloma venereum, trachoma
Q fever (Coxiella burnetii infection)
Infectious mononucleosis,cytomegalovirus infection, measles,
mumps
Cat-scratch disease (Bartonella
henselae infection)Whipple's disease (Tropheryma
whippelii infection)
Crohn's diseasePrimary biliary cirrhosis
SarcoidosisKikuchi's disease
Chronic granulomatous disease of
childhood
Langerhans granulomatosis
A wide range of infectious organisms cause the granulomatous diseases that fall into this category (table 1).These diseasesshare similar histologic features and have identifiable and distinct etiologic agents.
Mycobacterial Infections
Mycobacteria are a large group comprising Mycobacteriumtuberculosis, Mycobacterium leprae, Mycobacterium aviumcomplex, Mycobacterium ulcerans, Mycobacterium marinum.several opportunistic mycobacteria, and the mycobacteria fromwhich BCG vaccine is produced. The nontuberculous mycobacteria produce a wide clinical spectrum of granulomatousdisorders, summarized in table 2. Mycobacteria have been implicated in the etiopathogenesis of sarcoidosis and Crohn' sdisease, and this is discussed in the section on group 3 disorders.
Tuberculosis. The etiologic agent of tuberculosis, M. tuberculosis, is clinically one of the most common microbialcauses of granulomas [2, 8]. The M. tuberculosis granuloma,classically characterized by the presence of central caseousnecrosis, is referred to as a tubercle. Typically, there is a centralarea of amorphous caseating granular debris and loss of cellulardetail, and acid-fast bacilli are present. This area is encircledby epithelioid cells, lymphocytes, histiocytes, fibroblasts, andoccasionally Langhans' giant cells. These histologic featuresof the granuloma are sufficiently characteristic to allow reasonably accurate diagnosis of tuberculosis [8].
While caseating granulomas are the classic finding in casesof tuberculosis, they are not always present; noncaseating granulomas may occur. When acid-fast bacilli are not identifiablein the granuloma, culture of the biopsy specimen may yieldM. tuberculosis. PCR with use of primers specific for M. tuberculosis is currently under evaluation for the identification ofthis species [9] and may develop into an important tool fordistinguishing it from other (nontuberculous) mycobacteria.
Immunosuppression may alter the classic histopathologicfeatures of tuberculosis. A spectrum of histopathologic granulomatous tissue reactions are seen in patients infected with HIV[10]. Classic well-formed granulomas have been observed inindividuals with early HIV disease whose CD4 cell counts haveremained adequate. In patients with advanced HIV disease, thegranulomas are less well formed, are more necrotic, and maycontain abundant bacilli, which may be demonstrated by appropriate staining techniques as well as by culture of biopsy specimens [9-11].
Leprosy. Leprosy, or Hansen's disease, is a chronic granulomatous disorder caused by M. leprae, an obligate intracellularbacterium that predominantly affects the skin and nerves [12].First identified over a century ago by Armauer Hansen, thebacterium has defied all attempts to grow it in culture. Leprosy
by guest on September 23, 2016
http://cid.oxfordjournals.org/D
ownloaded from
148 Zumla and James
Table 2. Granulomatous diseases caused by nontuberculous mycobacteria.
em 1996;23 (July)
Site
Lung
Lymph node
Bones and joints
Skin and soft tissue
Multisystem
involvement
Clinical manifestation
or type of disease
Bronchopulmonary disease
Lymphadenitis
Osteomyelitis
Abscesses
Buruli ulcer
Swimming pool granuloma
Tuberculoid leprosy
Lepromatous leprosy
Meningitis, diarrhea, pneumonia,
lymphadenitis
Mycobacterial species
M. avium complex, M. kansasii, M. xenopi,
M. simiae, M scrofulaceum, M. chelonae,
M. malmoense
M. avium-M. intracellulare, M kansasii,
M fortuitum, M. scrofulaceum,
M. chelonae
M. avium-M. intracellulare, M. kansasii,
M fortuitum, M scrofulaceum,
M. marinum
M chelonae, M. fortuitum
M. ulcerans
M. marinum
M. leprae
M. leprae
M. avium-M. intracellulare, M. kansasii,
M. chelonae
manifests as a spectrum of well-described clinical, pathological, and immunologic features [12].
At one end of the spectrum is lepromatous leprosy, the lowresistance form with multiple lesions, extensive skin and visceral involvement, and diffuse infiltration of lesions with numerous leprotic baccilli. At the other end of the spectrum,tuberculoid leprosy, the highly resistant form, is characterizedby one or few skin lesions and involvement of the nerve at thesite of the lesion. Skin biopsies of patients with tuberculoidleprosy reveal a few acid-fast and alcohol-fast leprotic bacilliand a well-defined giant cell granuloma. Actual granulomatousinvasion and destruction of dermal nerves, occasionally withcaseous necrosis, sometimes occurs.
Within the dermis of patients with lepromatous leprosy, increased levels of Th2-type cytokines (lL-4, IL-5, and IL-IO)and decreased levels of IL-2 and IFN-y have been noted. Incontrast, ThI-type cytokines predominate in tuberculoid lesions[7]. The diagnosis of leprosy is based on a combination ofclinical and histologic findings. The paucity of acid-fast bacilliin the lesions of tuberculoid leprosy may sometimes make itdifficult to distinguish them from other infectious and noninfectious granulomas, especially those of sarcoidosis and syphilis[12]. Pt.R for the detection of M leprae in tissue lesions isbeing developed, and this may in the future allow more accuratediagnosis in such circumstances [13].
Buruli ulcer. M. ulcerans is the cause of chronic, relativelypainless, cutaneous Buruli ulcers. The disease is most prevalentin Africa and Australia. The organism causes extensive undermined ulcers on the extensor surface of the extremities. Thecenters ofthe ulcers are necrotic, and the edges are undermined;the organisms are usually found at the periphery, where granulation tissue is most extensive [14]. While it is relatively easyto diagnose Buruli skin ulcers on the basis of clinical features
and histologic findings, microbiological identification of thecausal mycobacteria may sometimes be quite difficult, requiring long periods of culture.
Newer techniques such as gas-phase chromatography arebecoming useful for identification of the acid-fast bacilli inlow-count subcultures [15]. A report from West Africa [16] ofa case of disseminated M. ulcerans infection that followed asnakebite and caused multifocal osteomyelitis illustrates thedifficulties that are associated with identifying the organismby conventional microbiological techniques. Although repeatedcultures were negative, the organism was identified as M. ulcerans by peR amplification of the genes coding for 168 rRNAfrom biopsy specimens.
Swimming pool (fish tank) granuloma. M marinum causesswimming pool (fish tank) granuloma [17]. Although the primary skin infection may be inconspicuous, the draining lymphnodes are extensively involved and caseous. A similar microscopic picture, with conspicuous plasma cell infiltration, is associated with granulomas due to other opportunistic mycobacteria. Fish tank granulomas develop in persons with minorabrasions who dip their hands in tropical fish tanks. Usually asolitary granuloma, nodule, or pustule forms, which may ulcerate or suppurate; however, multiple lesions may extend alongthe line of lymphatic vessels (in sporotrichotic infections).
Biopsy specimens that are cultured on Lowenstein-Jensenmedium at room temperature yield pigmented mycobacterialcolonies in 2~4 weeks. The response to treatment is variableand not dramatic. Antituberculous drugs, cotrimoxazole, andhigh doses of minocycline have been advocated.
A wide range of mycobacterial diseases occur in humans(table 2), and in some cases the clinical features and results ofstaining for acid-fast bacilli in granulomas may not reveal theidentity of the causal organisms, which may also be difficult
by guest on September 23, 2016
http://cid.oxfordjournals.org/D
ownloaded from
em 1996;23 (July) Granulomatous Infections 149
to isolate in culture. The development and application ofmolecular techniques such as PCR may in the future allow fairly
accurate diagnosis [18].
Bacterial Infections
Brucellosis. Brucellosis (Malta fever) is a multisystemgranulomatous disorder caused by small gram-negative, intracellular coccobacilli of the Brucella species [19]. The UnitedKingdom is one of 17 Brucella-free countries. It is more prevalent than its reported incidence in the United States, France,Spain, and Ireland [20]. Unpasteurized dairy products remaina common mode of transmission of this chronic granulomatousdisorder among travelers to areas of endemicity. Raw camel'sand goat's milk, raw sheep's and goat's liver, and reindeerbone marrow have all been associated with transmission.
Diagnosis, which may be difficult by means of culture, isusually based on clinical features and serological and histopathologic findings [20]. Delayed-type hypersensitivity is believedto be important in the formation of the tissue granuloma, whichlimits the spread of the organism [21]. Epithelioid-cell granulomas may be found in the reticuloendothelial system, lung, brain,bone, joints, and soft tissue and are difficult to distinguish from
granulomas due to other causes.Necrotizing hepatic granulomas due to brucellosis have been
described. In these cases, bacteriologic diagnosis may be difficult on the basis of blood or abscess pus cultures, althoughserology may be useful [22]. PCR tests for the identificationof Brucella species are being developed [23, 24], and thesemay be useful tools for the diagnosis of brucellar granulomas.
Melioidosis. Melioidosis is caused by a gram-negative bacillus, Burkholderia pseudomallei. Acute melioidosis presentswith protean clinical manifestations during which septicemia
may occur, causing septic shock and multiorgan involvement.This may progress to a chronic suppurative phase, in whichlung involvement, subcutaneous abscesses, lymphadenitis, andsuppurative parotitis are common [25, 26]. Granulomatous osteomyelitis has also been described [27].
The lesions vary histologically from acute, chronic inflammation with a focal granulomatous component to a predominantly granulomatous picture. Caseous or purulent central necrosis is surrounded by epithelioid and giant cells and byfibrosis in lungs, lymph nodes, and bone. The presence ofintracellular gram-negative bacteria within macrophages or giant cells and of similar extracellular bacilli may provide a clue
to the diagnosis [28].
More-specific enzyme immunoassays are being developedfor the detection of circulating antibody [29] or of B. pseudomallei antigen in the urine [30] of patients with melioidosis.PCR tests for the identification of B. pseudomallei have recently been described [31, 32], but these require field testingin areas of endemicity.
Treponemal Infections
The spirochetes include Treponema pallidum subspecies pal
lidum (the cause of syphilis), T. pallidum subspecies pertenue(responsible for yaws), and Treponema carateum (the causalagent of nonvenereal pinta in Latin America). All three arechronic granulomatous disorders that may cause diagnosticconfusion with other granulomatous disorders [33]. Mucocutaneous lesions of secondary syphilis show a spectrum ofhistopathologic changes, ranging from minimal infiltration togranulomatous infiltration throughout the dermis [34]; granulomatous infiltrates show endothelial proliferation, with mononuclear cell infiltration.
The syphilitic gumma may present as a microscopic togrossly visible lesion, with an enclosing wall of histiocytes,plasma cell infiltrate, and necrotic center cells with intact cellular outlines. peR to detect T. pallidum subspecies pallidumDNA in clinical specimens, including gumma biopsy material,appears to be highly sensitive and specific [35, 36].
Protozoan Infections
Leishmaniasis. Leishmaniasis refers to the spectrum ofclinical disease caused by the protozoan Leishmania [37]. Thisparasite is transmitted by phlebotomine sandflies, and the disease presents in three clinical forms: cutaneous, mucocutaneous, and visceral. Amastigotes of Leishmania species livewithin macrophages and are usually seen as single or focalcollections within macrophages in blood, aspirates of sternalmarrow, liver or spleen specimens, or special culture medium.
Leishmaniasis has been extensively studied as a model infection for analysis of cell-mediated immunity. Granulomatousresponses are a feature of cutaneous and mucocutaneous leishmaniasis. Cutaneous ulceration is characterized by a mononuclear cell infiltrate, with Th l-type responses emerging afterweeks of infection. Resolution of the infection depends on anincrease in the number of leishmania-specific CD4+ T cells ofthe Thl subset [38], following which a granulomatous responsewith epithelioid and giant cells emerges.
The nodular stage of post-kala-azar dermal leishmaniasis ischaracterized by massive granulomas consisting of lymphocytes, plasma cells, and histiocytes, with numerous amastigotesof the species Leishmania donovani [39]. A preponderance ofCD8+ T cells occurs in the lesion, in contrast to normal levelsin the peripheral blood. A recent report described the use ofPCR for identifying Leishmania aethiopica DNA in formalinfixed and paraffin-embedded tissue specimens [40]: seven ofthe 40 skin biopsy specimens in which parasites were not foundhistopathologically were subsequently found to contain parasite DNA.
Toxoplasmosis. Toxoplasmosis refers to the clinical disease caused by the protozoan Toxoplasma gondii. Neonatalinfection occurs during acute maternal infection in pregnancy.Acquired toxoplasmosis presents as a glandular, fever-like syn-
by guest on September 23, 2016
http://cid.oxfordjournals.org/D
ownloaded from
150 Zumla and James
Table 3. Causes and immunopathologic features of granulomatous mycoses.
em 1996;23 (July)
Fungal species
Cryptococcus neoformans
Candida species
Sporothrix schenckii
Histoplasma capsulatum
varieties
Aspergillus speciesParacoccidioides brasiliensis
Coccidioides immitis
Blastomyces dermatitidis
Phialophora speciesPseudallescheria boydii,
Madurella species
Clinical condition
Cryptococcosis
Candidiasis
Sporotrichosis
Histoplasmosis
Aspergillosis
South American blastomycosis
Coccidioidomycosis
Blastomycosis
ChromoblastomycosisMycetoma
Immunopathologic feature(s)
Pneumonia, infarction, abscess,
meningitis, granuloma, fibrosisAbscess, necrosis, granulomas
(mucocutaneous or multisystem)
Granuloma (cutaneous, skeletal)Pneumonia, cavitation, granuloma
Necrotizing multisystem granulomasPneumonia, cavitation, granuloma
Pneumonia, cavitation, granuloma,
cutaneous plaques, nodulesGranuloma, microabscess, pneumonia,
chronic pulmonary or extrapulmonarydisease
Granuloma (cutaneous)Granuloma (skin and subcutaneous tissue)
drome that rapidly resolves. The T. gondii trophozoites encystand persist in nucleated cells as the cause of chronic, latentinfection that can reactivate with immunosuppression and produce multisystem disease.
The histopathologic changes characteristically occur with atriad of features [41]: reactive follicular hyperplasia, epithelioidhistiocytes, and monocytoid cells. Langhans' giant cells arenot typically seen. Accurate diagnosis may be difficult in manycases. PCR is increasingly being used to detect T. gondii DNAin various clinical samples [42], and its application to identification of T. gondii DNA in tissue biopsy specimens is underinvestigation.
Fungal Infections
Fungal infections are common causes of granulomas worldwide. Granulomatous fungal infections can present as localizedconditions or systemic illnesses (table 3). A wide range ofmedically important fungi can now be diagnosed by PCR [43].Histoplasmosis and coccidioidomycosis are caused by dimorphic fungi that produce clinical disease resembling tuberculosis[44-46].
Histoplasma infections in the lungs of immunocompetentpersons cause epithelioid-cell granulomas that undergo coagulative necrosis. Histologic differentiation from tuberculosis,sarcoidosis, and coccidioidomycosis may be possible by identification ofthin-walled yeast forms in Grocott-Gomori methenamine-silver nitrate stains. In immunodeficient persons, histoplasmosis may be fulminant and is not associated withformation of epithelioid-cell granulomas: focal collections ofmononuclear phagocytes containing yeasts are seen throughouttissues of the body, and the reticuloendothelial system becomesfull of macrophages containing yeasts [46, 47].
Coccidioides immitis spores inhaled by immunocompetentpersons induce a delayed-type hypersensitivity reaction. Mostprimary cases of coccidioidomycosis involving such patientsare asymptomatic. Lung lesions may develop in some patients,in association with fever, chest symptoms, and erythema nodosum. A minority of cases progress to disseminated infection.
The primary and secondary granulomatous lung lesions dueto C. immitis are similar to those due to Histoplasma species.C. immitis organisms are thick-walled, nonbudding spherulesoften filled with endopores. A neutrophil infiltrate may occuraround the area of granulomatous reaction when the spherulesrupture to release the endospores. In disseminated disease, theinflammatory response may be purely granulomatous, pyogenic, or mixed. Pyogenic lesions may dominate in immunosuppressed patients [46, 47].
Helminthic Infections
Several helminthic parasites can cause granulomas withpathological consequences. Ofparticular interest are schistosomiasis and visceral larva migrans.
Schistosomiasis. Schistosomiasis is caused by five mainhuman-blood flukes of the genus Schistosoma: S. haematobium,S. mansoni, S.japonicum, S. intercalatum, and S. mekongi.These currently infect more than 200 million people. Diagnosisis made by identification of the species-specific characteristicova in feces, urine, or tissues (usually rectal and liver biopsyspecimens). During the chronic phase of the infection, the mature worms produce large numbers of ova, which becometrapped in tissues and induce formation of granulomas [48],with chronic fibro-obstructive sequelae.
Granuloma formation is the result of a combination of factors: a foreign-body reaction to the deposited ova (which may
by guest on September 23, 2016
http://cid.oxfordjournals.org/D
ownloaded from
em 1996;23 (July) Granulomatous Infections 151
or may not be identifiable within the granuloma) and a cellmediated delayed-type hypersensitivity reaction to antigenicdeterminants of the parasite [49]. The giant cells are of themultinucleate type, and the eosinophils are conspicuous. Theova contain living larvae that release a variety of toxic andantigenic substances through the egg wall, which itself is antigenic.
These substances induce and maintain the granulomatousresponse. The granulomatous response appears to be tightlyregulated and involves development of Th l-type and Th2-typecytokine responses [50, 51].
Toxocariasis (visceral larva migrans). Dog and cat larvalnematodes, especially Toxocara canis and Toxocara cati, mayinfect children and lead to the syndrome of visceral larva migrans [52]. Swallowed infected eggs emerge as larvae in theintestine, whence they pass to the liver, lungs, brain, and eye,producing eosinophilic granulomas in these different organs.This infection should be particularly suspected in children withhepatomegaly or miliary granulomas in liver biopsy specimens,miliary lung mottling, eosinophilia, and focal posterior uveitisor endophthalmitis. Liver granulomas may reveal part of thelarvae with surrounding eosinophilia.
Viral Infections
Measles. The paramyxovirus family is a group of RNAviruses that includes the measles virus, which has been implicated in the etiology of idiopathic granulomatous disorders,sarcoidosis, and Crohn's disease. The measles (rubeola) virusis spread by respiratory droplets and it multiplies within theupper respiratory tract epithelial cells, mononuclear cells, Band T lymphocytes, and macrophages. Viremia follows, whichdisseminates the virus throughout the body. T cell-mediatedimmunity that controls the infection develops and produces themeasles rash.
The blotchy rash of measles consists of dilated skin vessels,edema, and a nonspecific mononuclear perivascular infiltrate[47,53]. Lymphoid organs have marked follicular hyperplasia,large germinal centers, and randomly distributed multinucleategiant cells (Warthin-Finkeldey cells), which have eosinophilicnuclear and cytoplasmic inclusion bodies. In the lung, peribronchial and interstitial mononuclear infiltration occurs. The measles virus has been implicated in the etiopathogenesis of sarcoidosis and Crohn's disease and is discussed in the sectionon group 3 disorders.
Herpesviruses. The herpes group of viruses are doublestranded DNA viruses that cause a spectrum of human disease.The Epstein-Barr virus is the etiologic agent of the clinicalcondition infectious mononucleosis and is transmitted by closecontact, especially via saliva during kissing. It has also beenimplicated in the pathogenesis of several disorders such asBurkitt's lymphoma, nasopharyngeal carcinoma, B celllymphomas, and sarcoidosis.
Infectious mononucleosis in its benign form is characterizedby fever, sore throat, lymphadenopathy, splenomegaly, and theappearance of atypical activated T cells in the peripheral blood.Further multisystem involvement may occur, leading to hepatitis, meningoencephalitis, and pneumonitis. Atypical lymphocytes are seen in the portal areas and sinusoids of liver biopsyspecimens, and scattered isolated areas of focal parenchymalnecrosis filled with lymphocytes may occur [47].
This histologic picture is difficult to distinguish from that ofother types of viral hepatitis, such as hepatitis B, and from thatof cytomegalovirus [54] infections. Specific diagnosis can bemade by culture of the virus or identification of virus-specificDNA by PCR [55].
Group 2 Granulomatous Disorders (Recently RecognizedCausal Agents)
Cat-Scratch Disease
Cat-scratch disease or fever is also known as benign lymphoreticulosis or regional granulomatous lymphadenitis [56]. Itonly occurs in humans, especially those who are scratchedor bitten by kittens and then develop regional lymphadenitisproximal to the site of injury. Primary involvement is that ofthe lymph nodes, which first show lymphoid hyperplasia. Later,scattered granulomas with central areas of necrosis coalesce toform abscesses.
The histopathologic features of cat-scratch disease are notdiagnostic and may be mistaken for tularemia, lymphogranuloma venereum, syphilis, brucellosis, atypical mycobacterialinfections, fungal infections, and toxoplasmosis [47]. WarthinStarry silver staining is used to detect Bartonella henselae,which may be present in the early phase of the disease. A skintest antigen has been made from lymph node pus. It is inoculated intradermally, and the degree of induration and erythemais measured at 48 hours.
The cat-scratch antigen skin test is positive for about 90%of patients who are clinically suspected of having the disease.This test will become redundant when techniques for amplifying specific nucleotide sequences with peR come into general use. There is no well-recognized response to antibiotics,and recovery usually occurs without treatment.
Until recently, the causal agent remained controversial. Twoorganisms, Afipia felis and B. henselae, have been identifiedas potential candidates [57]. Substantial evidence is now accumulating that B. henselae, a small pleomorphic gram-negativebacillus, is the etiologic agent [58-60].
A PCR hybridization assay designed to amplify DNA fromB. henselae and A. felis was used on pus aspirates from 89skin test-positive patients with cat-scratch disease. B. henselaeDNA was found in 96% of samples, while no samples containedA.felis DNA [58]. An indirect fluorescent antibody assayfor B. henselae-specific antibody has been described thatappears to be sensitive for the diagnosis of cat-scratch disease [61].
by guest on September 23, 2016
http://cid.oxfordjournals.org/D
ownloaded from
152 Zumla and James em 1996;23 (July)
Whipple's Disease
George Hoyt Whipple of Johns Hopkins Hospital initiallydescribed a 37-year-old missionary who presented with fever,polyarthritis, and steatorrhea. His report was entitled "A Hith
erto Undescribed Disease Characterized Anatomically by Deposits of Fat and Fatty Acids in the Intestinal and Mesenteric
Lymphatic Tissues" [62]. Whipple's disease is a chronicmultisystem granulomatous disorder that affects middle-agedwhite males. It presents, as did Whipple's patient, with fever,polyarthritis, weight loss, and diarrhea progressing to malabsorption.
Hepatosplenomegaly and generalized lymphadenopathy maybe present, and biopsy of the lymph node, liver, or small intestine will reveal sarcoid granulomas, of which PAS stains will
reveal foamy macrophages. The PAS-positive material corre
sponds with lysosomes containing bacilliform bodies. Electron
microscopy reveals rod-shaped bacilli, termed Whipple bacilli[63], Whipple's associated bacterial organisms, or (more correctly) Tropheryma whippelii [64-66].
The nucleic acids extracted from an endoscopic biopsy specimen of the proximal small bowel of a patient with Whipple'sdisease were subjected to amplification and nucleotide sequencing by PCR. The resulting PCR product from the bacterial 16SrRNA was then the subject of a computer database search forthe rRNA sequences most similar to it.
T. whippelii is an actinomycete that behaves like an intracel
lular pathogen and responds particularly well to antibiotics thatenter cells easily. It is an erythrocyte-associated organism (likeBartonella bacilliformis, Babesia species, and Plasmodiumspecies, which are also hemotropic). T. whippelii DNA has
been detected by PCR examination of peripheral blood, pleuraleffusion cells [66, 67], biopsy tissue from the intestine [68],heart [69], or vitrectomy specimen of the eye [70].
The PCR primers that have been developed are specific forT. whippelii, and this technique is now one of the standardmethods for establishing the diagnosis of Whipple's disease[66, 67, 70]. The technique ensures a more specific diagnosissince the organism cannot be cultivated. Other data show thatthere may be a closely related second causative agent of Whip
ple's disease [71, 72].
Group 3 Granulomatous Disorders (Infective AgentStrongly Suspected)
Primary Biliary Cirrhosis
This condition of progressive, nonsuppurative, granulomatous destruction of intrahepatic bile ducts has also been termedxanthomatous biliary cirrhosis, but its most accurate descriptionis chronic nonsuppurative destructive cholangitis [73]. Livercell necrosis may become superadded, later progressing to cirrhosis. Epithelioid granulomas associated with the bile ductsare found in about one-third of the cases of primary biliarycirrhosis (PBC). The condition has been likened to chronic
graft-versus-host rejection, with similar structural changes in
the bile, lacrimal, and pancreatic ducts, which have a highconcentration of human leukocyte class II antigens on the epithelial surface.
PBC is distinguished by the fact that it predominates in
women in the reproductive years of age and by the presenceof serum mitochondrial antibodies. It is classified as an autoim
mune disorder and is associated with other autoimmune conditions, including rheumatoid arthritis, systemic lupus erythematosus, scleroderma, Sjogren's syndrome, and the CRESTsyndrome.
The etiology of PBC remains unknown. Two infectiousagents have been postulated as possible trigger cofactors [74,75]. The antigen specific for PBC serum is M2, a component
of the pyruvate dehydrogenase complex of mitochondrial en
zymes. There are similarities between mitochondrial components and Escherichia coli, and cross-reactivity between bile
duct mitochondria and bacteria may be conceivable [73]. Mitochondrial antibodies cross-react with subcellular constituentsor gram-negative intestinal or urinary tract organisms.
An increased incidence of gram-negative urinary tract infections in patients with PBC has been reported. Another infectiveassociation was suggested in a report from Barcelona that incriminated Mycobacterium gordonae. This ubiquitous organism causes granuloma formation, and serum antibodies are evident in PBC serum. Furthermore, there is cross-reactivity
between this organism and the important PBC-defining M2
antigen. PCR has disclosed the presence of this mycobacteriumin the tissues and bile of patients with PBC [75].
Sarcoidosis
Sarcoidosis is a multisystem disorder of unknown etiology[76]. An antigen (chemical or infectious organism)-driven,cell-mediated immune response leads to a cytokine cascade, tosarcoid granuloma formation, and eventually to sarcoid fibrosis.An intense search for an infectious cause has yielded manytheories, yet none is conclusive (table 4).
In 1961, Edith Mankiewicz [77], of Montreal, reported thatbacteriophages, lytic for mycobacteria, can be isolated withhigher frequency from stool and resection specimens from pa
tients with tuberculosis and sarcoidosis. Raised titers ofphageneutralizing antibodies were found in tuberculous patients infected with mycobacteria harboring mycobacteriophages butnot in cases of sarcoidosis.
While mycobacteriophages are common to both diseases
the persistence ofphages without antibody was the explanationgiven for the absence of M. tuberculosis in sarcoid tissue [78].This observation led to further experiments in which guineapigs were infected with tubercle bacilli alone or together withmycobacteriophages, and the histological responses (analyzedblindly) were of two dissimilar types. The first was caseousnecrosis with the presence of acid-fast bacilli, and the secondwas a different response modified by the mycobacteriophages:
by guest on September 23, 2016
http://cid.oxfordjournals.org/D
ownloaded from
cm 1996; 23 (July) Granulomatous Infections 153
Table 4. Causal agents implicated in cases of sarcoidosis.
the phage-infected animals showed a granulomatous sarcoidlike picture with very few atypical lysogenic mycobacteria.
The conclusion reached was that phages of different antigenic composition determined a shift in the antigen-antibodyresponse toward either caseation or noncaseating granulomaformation. It seemed at that time that sarcoidosis resulted fromunrestrained lytic phage activity against mycobacteria [77, 78].
At about the same time, Chapman and Speight [79] reporteda high incidence of serum antibodies to mycobacteria in patients with sarcoidosis. Serum antibody levels were particularlyhigh against photochromogen and scotochromogen antigens.The patients' antibody status changed from negative to positiveas sarcoidosis manifested itself, and in many cases antibodiesappeared 2 months following the onset of erythema nodosum.
Ghosts of mycobacteria have been reported to be detectedby fluorescent microscopy of scalene lymph node biopsy specimens from patients with sarcoidosis and from controls. Auramine-rhodamine stains disclosed shining golden rods resembling mycobacteria in specimens from patients with sarcoidosisbut not in those from controls [80]. Recent advances in molecular biology have resurrected these observations; PCR has detected mycobacterial DNA in clinical samples from patientswith sarcoidosis [81, 82].
Bronchoalveolar lavage samples, bronchial washings, andtissue specimens have been assayed by PCR for mycobacterialDNA. M. tuberculosis DNA has been found in 50% of patientswith sarcoidosis, and nontuberculous mycobacterial DNA hasbeen found in 70% of them. Another study extracted RNAfrom five sarcoid and five normal spleens [81]. Extracts wereassayed by liquid-phase DNA/RNA hybridization with a DNAprobe specific for the rRNA of the M. tuberculosis complex.
Causal agents implicated
Bacteria
Mycobacteria
Streptococci
Propionibacterium acnes
Borrelia burgdorferi
Mycoplasma species
Nocardia species
Viruses
Epstein-Barr
Herpesvirus
Rubella
Measles
Cytomegalovirus
Coxsackievirus B
Retrovirus
Others
Beryllium
Zirconium
Pine pollen
Peanut dust
Clay (ingestion)
Reference(s)
[76-92]
[94]
[88,93]
[95,96]
[97]
[98]
[99]
[100, 101]
[102]
[102]
[102]
[102,103]
[104]
[lOS]
[105][106,107]
[107]
[108]
The number of hybrids obtained with the sarcoid spleens (fromwhich mycobacteria were neither seen on microscopy nor cultured) was 4.8 times higher than that obtained with normalspleens.
However, the results obtained by PCR are being fiercelycontested by German [83], Scottish [84], and French [85] investigators. The findings of the German group contradict those ofSaboor et a1. [81], and the group concludes that mycobacterialDNA cannot be detected in tissues or bronchoalveolar lavagecells in cases of sarcoidosis. They use the technique to distinguish tuberculosis (which produces a positive result) from sarcoidosis. The Glasgow investigators [84] also failed to detectmycobacterial DNA in sarcoid lymph nodes. The French investigators [85], using PCR, rarely found DNA from M. tuberculo
sis in cases of sarcoidosis.The literature for and against a mycobacterial cause of sar
coidosis is summarized in table 5. The existence ofmycobacterial sarcoidosis would validate several claims of infective orchemical causes of sarcoidosis: Scadding's long-held views ontuberculous causation [86], a Swedish suggestion [87] of aninteraction between mycobacteria and a virus, a Japanese theory [88] regarding Propionibacterium as the causal agent, andthat of investigators in Houston [89] who isolated cell walldefective spheroplasts of acid-fast mycobacteria from sarcoidskin biopsy specimens.
The controversy continues [90-93] and many different antigens have come under suspicion, including other bacteria [9497], fungi [98], viruses [99-104], and chemicals [105-- 108].Many attempts have been made but tissue culture has failed touncover a virus. Elevated antibody responses to several viruseshave been reported, but such responses may only reflect B-cellhyperreactivity rather than a causal relationship.
Crohn's Disease
Crohn's disease is a chronic granulomatous disorder affecting predominantly the gut, but it has multisystem involvementand variable clinical manifestations. The pathological changesmay involve any parts and any layer of the alimentary tract.Transmural inflammation consists of chronic inflammatorycells with lymphoid aggregates scattered throughout. Noncaseating sarcoid-like granulomas may be present in all layers ofdiseased and unaffected tissue throughout the alimentarytract [109].
As with sarcoidosis, controversy prevails concerning a causalagent. Clinicopathologic studies at the Royal Free Hospital(London) demonstrated the presence of granulomatous vasculitis and vessel thrombosis, giving rise to focal microulcers. Theinvestigations have shown early superficial mucosal changesand disruption of the capillary basement membrane precedingthe formation of the microulcer [I 10]. These inflammatorychanges may extend beyond the alimentary tract to the lung,causing pulmonary vasculitis, granulomatous interstitial lymphocyte infiltration, and interstitial fibrosis [II I, 112].
by guest on September 23, 2016
http://cid.oxfordjournals.org/D
ownloaded from
154
Table 5.
Reference( s)
[77,78]
[79]
[80]
[87]
[93]
[89]
[81]
[82]
[83]
[84]
[85]
[90]
[91]
[92]
Zumla and James
Summary of the literature for and against a mycobacterial cause of sarcoidosis.
Laboratory method used or featureYear(s) of study Location of study analyzed (finding)
1961-1964 Montreal Reaction to mycobacteriophage infection1964 Dallas Presence of mycobacterial antibodies in
serum (positive)
1971 Prague Auramine-rhodamine staining,
fluorescent microscopy (positive)
1972 Stockholm Mycobacteria-virus interaction
1984 Tokyo Culture of lymph nodes (positive for
Propionibacterium acnes)1988 Houston Skin biopsy (cell-wall-defective, acid-
fast bacilli noted)1992 London PCR (positive for mycobacterial DNA)
1992 London Liquid-phase hybridization (positive for
mycobacterial rRNA)1992 Borstel, Germany PCR (negative for mycobacterial DNA)
1992 Glasgow PCR (negative for mycobacterial DNA)
1992 Paris PCR (negative for mycobacterial DNA)
1993 London PCR (positive for mycobacterial DNA)
1993 Ohio ELISA (positive for antibodies to
Mycobacterium paratuberculosis)1993 Copenhagen Western blotting (positive for antibodies
to Mycobacterium tuberculosis)
ern 1996;23 (July)
Against this background is the conflict about whether thecausal agent is Mycobacterium paratuberculosis [113], othermycobacteria [114], or the measles virus [115, 116]. Measlesvirus nucleocapsids have been detected in foci of granulomatous inflammation of the intestine in cases of Crohn's disease[117]. It has been postulated that Crohn' s disease may be aresult of chronic granulomatous vasculitis that develops in reaction to a persistent infection with measles virus within thevascular endothelium [118].
Langerhans' Cell Granulomatosis
The term Langerhans' cell granulomatosis refers to proliferative disorders of histiocytes, previously referred to as histiocytosis X [119]. It encompasses a group of disorders of unknown etiology characterized by granulomatous infiltration ofthe lungs, bone, skin, lymph nodes, and brain. The clinicalconditions have been known by several names, based on thetype of presentation, sites of involvement, rate of progression,and degree of associated immune dysfunction. They includeeosinophilic granuloma, Letterer-Siwe disease, and HandSchuller-Christian disease. These are now considered to bedifferent expressions for the same basic disorder, in which theproliferation of Langerhans' cells results from disturbances inimmunoregu1ation.
Langerhans' histiocytes are bone marrow-derived monocyte-macrophage cells; they include Langerhans' epidermalcells, Kupffer's cells in the liver, osteoclasts, and alveolar macrophages. They are human leukocyte antigen-DR-positive
functioning macrophages that present antigen to T cells andplay a role in cell-mediated immunity. Unlike histiocytes,Langerhans' cells can be stained immunohistochemically forS-100 protein and OKT-6.
Lung biopsy reveals a mixed celullar exudate, foam cells,eosinophils, and characteristic X bodies (Birbeck granules) inmacrophages. Langerhans' or X bodies are an ultrastructuralfeature in 90% of patients [120]. They are identical to thegranules in Langerhans' epidermal cells and consist of intracytoplasmic rod-, plate-, or cup-like pentalaminar structures.
The presence of these tennis racket-shaped ultrastructuralBirbeck granules is diagnostic of the disorder. They have surface adenosine triphosphate activity identifiable by gold fluorescence. These diagnostic cells are readily found by bronchoalveolar lavage, and this technique may make lung biopsyunnecessary. It may also be a likely means of detecting apossible causal agent.
Chronic Granulomatous Disease of Childhood (CGD)
CGD comprises a genetically heterogeneous group of diseases that have in common the functional disorder of a multicomponent enzyme system, NADPH oxidase, found in phagocytic cells [121-124]. This defect leads to severe infections,particularly with Staphyloccocus aureus, Serratia marcescens,Burkholderia cepacia, and Aspergillus species. Organisms thatlack catalase supply the neutrophil with the hydrogen peroxidefor their own destruction. Thus, catalase-negative organismssuch as pneumococci or streptococci present no problem because they are normally destroyed.
by guest on September 23, 2016
http://cid.oxfordjournals.org/D
ownloaded from
em 1996;23 (July) Granulomatous Infections 155
There are three autosomal recessive types, corresponding tomutations in p22p h0X, p47ph0X, and p67phox
. The classic X-linkeddisorder (found in ,....,60% of cases), which affects the geneencoding gp91phox, occurs in boys and is diagnosed in infancy,usually on the basis of an infection with a catalase-producingbacterium or fungus. The diagnosis is confirmed on the basisof a nitroblue tetrazolium test or other test showing defectivesuperoxide production.
Hepatosplenomegaly is common and does not reflect activeinfection per se. Lymphadenopathy is often undetected, butwhen it is significant it usually reflects an active infection.Severe granulomatous problems can occur in association withCGD, such as obstruction of the gut and the genitourinarysystem. Weeping granulomatous skin lesions occur in the setting of an active wound, typically post-surgically. The histological appearance is marked by lipid-laden macrophages. Whilethe etiology of CGD is genetic, there may be undetected pathogens responsible for some of the granulomatous complicationsof the disease.
Orofacial Granulomatosis (Melkersson-Rosenthal Syndrome)
This is a rare granulomatous disorder of the mouth andadjacent tissues, involving the oral mucosa, gum, lips, tongue,pharynx, eyelids, and skin of the face. Melkersson [125] described an association between facial edema and facial paralysis. Rosenthal [126] added the features of lingua plicata orscrotal tongue. Other clinical features include granulomatouscheilitis, edema of the gums and scalp, salivary gland dysfunction, granulomatous blepharitis, trigeminal neuralgia, Raynaud's phenomenon, and chronic hypertrophic granulomatousvulvitis [127-129].
Biopsy of an involved area reveals a noncaseating granuloma. Unlike the findings in cases of sarcoidosis, there are nochest radiographic changes or uveitis, and the Kveim-Siltzbachskin test is negative. The etiology remains unknown.
Kikuchi's Disease
This disorder was described in 1972 by a Japanese pathologist and is characterized by lymphadenitis with focal reticulumcell hyperplasia, nuclear debris, and phagocytosis [130]. Thelymphadenitis may be proliferative, necrotizing, or xanthomatous. Immunohistologic analysis shows that the predominantcells involved in the lymphadenitis are various types of histiocytes, plasmacytoid monocytes, and T cells; B cells are absent[131]. Clinically there is localized, tender, cervicallymphadenopathy with an upper respiratory tract prodrome.
Kikuchi's disease was initially thought to occur mostly inwomen under the age of 30 years, although a recent study of79 cases showed that the preponderance of cases involvingfemales is not striking [131]. The disease runs a self-limitingcourse and is associated with a recurrence rate of3%. Kikuchi'sdisease has been recognized as a presenting feature of mixed
connective-tissue disease and has been noted in associationwith adult Still's disease and systemic lupus erythematosus, afinding leading to the hypothesis that Kikuchi's disease andautoimmune rheumatic disorders may share a common etiology [132].
The disease occurs worldwide and has often been confusedwith toxoplasmosis, cat-scratch disease, tuberculosis, infectiousmononucleosis, and non-Hodgkin's lymphoma. The cause ofthe disease is not known. A viral etiology is strongly suspectedon the basis of clinical features, although serological andultrastructural studies have not yet identified an infectiousagent [133].
Conclusions
Classification schemes are most useful when they provideinsight into the purpose of events and the etiologic mechanismsthat govern them. Granulomas are remarkably complex inflammatory foci that sequester organisms or other substancesresistant to degradation. While a large number of granulomatous disorders are recognized, infections are clearly the mostcommon underlying causes of granulomas in general.
It is apparent that granulomatous conditions ofdiverse etiologies share common histologic features, although the etiologicagent is not always identifiable. Although the histopathologicpatterns in various infectious granulomas may be sufficientlydifferent to prevent an accurate diagnosis, atypical presentations may necessitate identification of the specific etiologicagent by direct microscopic examination, culture, serology, ormolecular detection.
In several granulomatous diseases the etiologic agent is difficult to identify by microscopic examination, culture, or serological means. The availability of PCR for the diagnosis ofseveral infectious conditions has been of major importance indetecting the etiologic agents of granulomatous disease conditions that previously were considered to be of unknown etiology. In particular, PCR has been used successfully for detectingthe bacteriologic causes of Whipple's disease and cat-scratchdisease.
Further applications of molecular techniques may pin downthe microbiological etiology (if any) of several granulomatousdisorders of group 3 (table I), of which the etiology remainselusive.
References
1. Warren KS. Granulomatous inflammation. In: Lepow IH, Ward PW, eds.
Inflammation: mechanisms and control. New York: Academic Press,
1972:203 -18.2. Williams GT, Williams WJ. Granulomatous inflammation-a review. J
Clin Pathol 1983; 36:723 - 33.
3. Silverman L, Shorter RG. Histogenesis of the multinucleated giant cell.
Lab Invest 1963; 12:985-90.4. James DG. What makes granulomas tick'? Thorax 1991;46:734-6.
by guest on September 23, 2016
http://cid.oxfordjournals.org/D
ownloaded from
156 Zumla and James ern 1996;23 (July)
5. Heinzel FP. Th I and Th2 cells in the cure and pathogenesis of infectiousdiseases. Curr Opin Infect Dis 1995;8:151-5.
6. Weinstock JV. Function and regulation of granulomatous responses (experimental granulomatosis). In: James DG, ed. Sarcoidosis and othergranulomatous disorders. New York: Marcel-Dekker, 1994:99-133.
7. Yamamura M, Uyemura K, Deans RJ, et al. Defining protective responsesto pathogens: cytokine profiles in leprosy lesions. Science 1991;254:277-9.
8. Cucin RL, Coleman M, Eckardt JJ, Silver RT. The diagnosis of miliarytuberculosis: utility of peripheral blood abnormalities, bone marrowand liver needle biopsy. J Chronic Dis 1973;26:355-61.
9. Schluger NW, Kinney D, Harkin TJ, Rom WN. Clinical utility of thepolymerase chain reaction in the diagnosis of infections due to Myco
bacterium tuberculosis. Chest 1994; 105:1116-21.10. Zumla A, James DG. Lung immunology in the tropics. In: Sharma OP,
ed. Lung disease in the tropics. New York: Marcel-Dekker, 1991:164.
11. Lucas S, Nelson AM. Pathogenesis of tuberculosis in human immunodeficiency virus-infected people. In: Bloom BR, ed. Tuberculosis: pathogenesis, protection and control. Washington, DC: American Societyfor Microbiology, 1994:503-16.
12. Zumla A. Sarcoidosis and leprosy: a comparative perspective. In: JamesDG, ed. Sarcoidosis and other granulomatous disorders. New York:Marcel-Dekker, 1994:681-705.
13. Santos AR, Goes Filho JT, Nery JA, et al. Evaluation of PCR mediatedDNA amplification in non-invasive biological specimens for subclinical detection of Mycobacterium leprae. FEMS Immunol Med Microbiol 1995;11:113-20.
14. Dodge OG. Mycobacterial skin ulcers in Uganda: histopathological andexperimental aspects. J Pathol Bacteriol 1964;88:167- 74.
15. Delaporte E, Savage C, Alfandari S, Piette F, Leclerc H, Bergoend H.Ulcere de Buruli chez une malade Zairoise infectee par le virus del'immunodeficience humaine. Ann Dermatol Venereol 1994; 121:557-60.
16. Hofer M, Hirschel B, Kirschner P, et al. Brief report: disseminated osteomyelitis from Mycobacterium ulcerans after a snakebite. N Engl JMed 1993;328:1007-9.
17. Gray SF, Smith RS, Reynolds NJ, Williams EW. Fish tank granuloma.Brit Med J 1990;300:1069-1070.
18. Shinnick TM, Jonas V. Molecular approaches to the diagnosis oftuberculosis. In: Bloom BR, ed. Tuberculosis: pathogenesis, protection andcontrol. Washington, DC: American Society for Microbiology, 1994:511-31.
19. Bruce D. Malta fever. BMJ 1889; 1:1101-7.20. Trujillo IZ, Zavala AN, Caceres JG, Miranda CQ. Brucellosis. Infect Dis
Clin North Am 1994;8:225-41.21. Adams DO. The biology of the granuloma. In: Ioachim HL, ed. Pathology
of granulomas. New York: Raven Press, 1983:1-20.22. Debat-Zoguereh D, Badiaga S, Uzan E, Le Treut YP, Lebreuil G, Bour
geade A. Granulome necrosant hepatique d'origine brucellienne. Apropos d'un cas. Rev Med Interne 1995; 16:63-6.
23. Romero C, Gamazo C, Pardo M, Lopez-Goff I. Specific detection ofBrucella DNA by PCR. J Clin Microbiol 1995;33:615-7.
24. Bricker BJ, Halling SM. Differentiation of Brucella abortus bv. 1, 2, and4, Brucella melitensis, Brucella ovis, and Brucella suis bv. 1 by PCR.J Clin Microbiol 1994;32:2660-6.
25. Leelarasamee A, Bovomkitti S. Melioidosis: review and update. RevInfect Dis 1989; 11:413- 25.
26. Lumbiganon P, Viengnondha S. Clinical manifestations of melioidosisin children. Pediatr Infect Dis J 1995; 14:136-40.
27. SirikuIchayanonta V, Subhadrabandhu 1. Melioidosis: another etiologyof granulomatous osteomyelitis. Report of 2 cases. Clin Orthop 1994;308:183-6.
28. Wong KT, Puthucheary SD, Vadivelu J. The histopathology of humanmelioidosis. Histopathology 1995;26:51-5.
29. Rugdech P, Anuntagool N, Sirisinha S. Monoclonal antibodies to Pseu
domonas pseudomallei and their potential for diagnosis of melioidosis.Am J Trop Med Hyg 1995;52:231-5.
30. Desakom V, Smith MD, Wuthiekanun V, et al. Detection of Pseudomo
nas pseudomallei antigen in urine for the diagnosis of meiloidosis.Am J Trop Med Hyg 1994;51:627-33.
31. Kunakom M, Markham RB. Clinically practical seminested PCR forBurkholderia pseudo mallei quantitated by enzyme immunoassay withand without solution hybridization. J Clin Microbiol 1995;33:2131-5.
32. Lew AE, Desmarchelier PM. Detection of Pseudomonas pseudomallei
by PCR and hybridization. J Clin MicrobioI1994;32:1326-32.33. Tramont EC. Treponema pallidium (syphilis). In: Mandell GL, Bennett
JE, Dolin R, eds. Mandell, Douglas and Bennett's principles and practice of infectious diseases. 4th ed. Vol. 2. New York: Churchill Livingstone, 1995:2117-33.
34. Pandhi RK, Singh N, Ramam M. Secondary syphilis: a clinicopathologicstudy. Int J Dermatol 1995;34:240-3.
35. Burstain JM, Grimprel E, Lukehart SA, Norgard MV, Radolf JD. Sensitive detection of Treponema pallidum by using the polymerase chainreaction. J Clin Microbiol 1991;29:62-9.
36. Horowitz HW, Valsamis MP, Wicher V, et al. Brief report: cerebralsyphilitic gumma confirmed by the polymerase chain reaction in aman with human immunodeficiency virus infection. N Engl J Med1994;331: 1488-91.
37. Peters W, Killick-Kendrick R, eds. The leishmaniases in biology andmedicine. London: Academic Press, 1987.
38. Locksley RM, Louis JA. Immunology ofleishmaniasis. Curr Opin Immunol 1992;4:413-8.
39. Ghosh MK, Nandy A, Addy M, Maitra TK, Ghose AC. Subpopulationsof T lymphocytes in the peripheral blood, dermal lesions and lymphnodes of post kala-azar dermal leishmaniasis patients. Scand J ImmunoI1995;41:1l-7.
40. Laskay T, Miko TL, Negesse Y, Solbach W, Rollinghoff M, FrommelD. Detection of cutaneous Leishmania infection in paraffin-embeddedskin biopsies using the polymerase chain reaction. Trans Roy Soc TropMed Hyg 1995;89:273-5.
41. Dorfman RF, Remington JS. Value oflymph-node biopsy in the diagnosisof acute acquired toxoplasmosis. N Engl J Med 1973;289:878-81.
42. Novati R, Castagna A, Morsica G, et al. Polymerase chain reaction forToxoplasma gondii DNA in the cerebrospinal fluid of AIDS patientswith focal brain lesions. AIDS 1994;8:1691-4.
43. Makimura K, Murayama SY, Yamaguchi H. Detection of a wide rangeof medically important fungi by the polymerase chain reaction. J MedMicrobiol 1994;40:358-64.
44. Davies SF. Fungal pneumonia. Med Clin North Am 1994;78:1049-65.45. Wong JS, Herman SJ, de Hoyos A, Weisbrod GL. Pulmonary manifesta
tions of coccidioidomycosis. Can Assoc Radiol J 1994;45:87-92.46. Wheat J. Histoplasmosis and coccidioidomycosis in individuals with
AIDS: a clinical review. Infect Dis Clin North Am 1994; 8:467-82.47. Samuelson J, Von Lichtenberg F. Infectious diseases. In: Cotran RS,
Kumar V, Robbins SL, eds. Robbins pathologic basis of disease. Philadelphia: WB Saunders, 1994:305-77.
48. Warren KS. The pathology, pathobiology and pathogenesis of schistosomiasis. Nature 1978;273:609-12.
49. Warren KS, Domingo EO, Cowan RBT. Granuloma formation aroundschistosome eggs as a manifestation of delayed hypersensitivity. AmJ PathoI1967;51:735-56.
50. Chikunguwo SM, Quinn JJ, Ham DA, Stadecker MJ. The cell-mediatedresponse to schistosomal antigens at the clonal level. III. Identificationof soluble egg antigens recognized by cloned specific granulomagenicmurine CD4+ Thl-type lymphocytes. J Immunoll993; 150:1413-21.
51. Wynn TA, Cheever AW, Jankovic D, et al. An IL-12 based vaccinationmethod for preventing fibrosis induced by schistosome infection. Nature 1995;376:594-6.
by guest on September 23, 2016
http://cid.oxfordjournals.org/D
ownloaded from
em 1996;23 (July) Granulomatous Infections 157
52. Huntley CC, Costas MC, Lyerly A. Visceral larva migrans syndrome:
clinical characteristics and immunologic studies in 51 patients. Pediat
rics 1965; 36:523-36.53. Halsey NA, Job JS. Measles. In: Warren KS, Mahmoud AAF, eds. Tropi
cal geographic medicine. New York: McGraw-Hill, 1994.
54. Heyries L, Perreard M, Monges 0, Chrestian MA, Gerolami A. Hepatic
granulomatosis associated with mononucleosis syndrome secondary to
cytomegalovirus infection: apropos of 2 cases in healthy adults. Rev
Med Interne 1994; 15:744-6.
55. Nahass GT, Penneys NS, Leonardi CL. Analysis of the polymerase chain
reaction in the detection of herpesvirus DNA from fixed and stained
tissue sections. Arch Dermatol 1995; 131:805-8.
56. Daniels WB, MacMurray FG. Cat scratch disease: report of one hundred
sixty cases. JAMA 1954; 154:1247-51.
57. Drancourt M, Raoult D. La maladie des griffes du chat et la pathologie
aBartonella (Rochalimaea). Presse Med 1995;24:183-8.
58. Bergmans AMC, Groothedde J-W, Schellekens JFP, van Embden IDA,
Ossewaarde JM, Schouls LM. Etiology of cat scratch disease: compari
son of polymerase chain reaction detection of Bartonella (formerly
Rochalimaea) and Afipia felis DNA with serology and skin tests. J
Infect Dis 1995; 171:916-23.
59. Anderson B, Sims K, Regnery T, et a1. Detection of Rochalimaea
henselae DNA in specimens from cat scratch disease patients by PCR.
J Clin Microbiol 1994; 32:942-8.
60. Goral S, Anderson B, Hager C, Edwards KM. Detection of Rochalimaea
henselae DNA by polymerase chain reaction from suppurative nodes
of children with cat-scratch disease. Pediatr Infect Dis J 1994; 13:
994-7.
61. Dalton MJ, Robinson LE, Cooper J, Regnery RL, Olson JG, Childs JE.
Use of Bartonella antigens for serologic diagnosis of cat-scratch dis
ease at a national referral center. Arch Intern Med 1995; 155:1670-6.
62. Whipple GH. A hitherto undescribed disease characterized anatomically
by deposits of fat and fatty acids in the intestinal and mesenteric
lymphatic tissues. Johns Hopkins Hospital Bulletin 1907; 18:382-91.
63. Spapen HOM, Segers 0, De Wit N. Electron microscopic detection of
Whipple's bacillus in sarcoidlike periodic acid-Schiff-negative granu
lomas. Digest Dis Sci 1989;34:640-3.
64. Reiman DA, Schmidt TM, Mac Dermott RP, Falkow S. Identification of
the uncultured bacillus of Whipple's disease. N Engl J Med 1992;327:
293-301.
65. Donaldson RM Jr. Whipple's disease-s-rare malady with uncommon
potential [editorial]. N Engl J Med 1992;327:346-8.
66. Dobbins WO III. The diagnosis of Whipple's disease [editorial]. N EnglJ Med 1995;332:390-2.
67. Lowsky R, Archer GL, Gyles G, et al. Brief report: diagnosis of Whipple's disease by molecular analysis of peripheral blood. New Engl J
Med 1994;331:1343-6.
68. Maiwald M, Meier-Willersen HJ, Hartmann M, von Herbay A. Detection
of Tropheryma whippelii DNA in a patient with AIDS. J Clin Microbiol
1995;33:1354-6.
69. Wendler 0, Mendoza E, Schleiffer T, Zander M, Maier M. Tropheryma
whippelii endocarditis confirmed by polymerase chain reaction. Eur
Heart J 1995; 16:424-5.
70. Rickman LS, Freeman WR, Green WR, et al. Brief report: uveitis caused
by Tropheryma whippelii (Whipple's bacillus). N Engl J Med 1995;
332:363-6.
71. Wilson KH, Blitchington R, Frothingham R, Wilson JAP. Phylogeny of
the Whipple's disease-associated bacterium. Lancet 1991;338:474
475.
72. Harmsen 0, Heesemann J, Brabletz T, Kirchner T, Muller-Hermelink
HK. Heterogeneity among Whipple's-disease-associated bacteria [let
ter]. Lancet 1994;343:1288.
73. Sherlock SO. Diseases of the liver and biliary system. 9th ed. Oxford:
Blackwell, 1993:461.
74. Hopf U, Moller B, Stemerowicz R, et al. Relation between Escherichia
coli R (rough)-forms in gut, lipid A in liver, and primary biliary cirrho
sis. Lancet 1989;2:1419-22.
75. Vilagut L, Vila J, Vinas 0, et al. Cross-reactivity of anti -Mycobacterium
gordonae antibodies with the major mitochondrial autoantigens in pri
mary biliary cirrhosis. J HepatoI1994;21:673- 7.
76. James DG, ed. Sarcoidosis and other granulomatous disorders. New York:
Marcel-Dekker, 1994.
77. Mankiewicz E. Mycobacteriophages isolated from persons with tubercu
lous and nontuberculous conditions. Nature, London, 1961: 191:1416
1417.
78. Mankiewicz E. The relationship of sarcoidosis to anonymous bacteria.
Acta Med Scand Suppl 1964;425:68- 73.
79. Chapman JS, Speight M. Further studies of mycobacterial antibodies in
the sera of sarcoidosis patients. Acta Med Scand Suppl 1964;425:
61-7.
80. Richter J, Bartak F, Halova R. Detection of mycobacteria by fluorescent
microscopy in sarcoidosis. In: Proceedings of the 5th International
Conference on Sarcoidosis. Prague: Universita Karlova, 1971:83-4.
81. Saboor SA, Johnson NM, McFadden J. Detection of mycobacterial DNA
in sarcoidosis and tuberculosis with polymerase chain reaction. Lancet
1992; 339: 1012-5.
82. Mitchell IC, Turk JL, Mitchell DN. Detection of mycobacterial rRNA in
sarcoidosis with liquid-phase hybridisation. Lancet 1992: 339:
1015-7.
83. Gerdes J, Richter E, Rusch-Gerdes S, et al. Mycobacterial nucleic acids
in sarcoid lesions [letter]. Lancet 1992; 339: 1536-7.
84. Thakker B, Black M, Foulis AK. Mycobacterial nucleic acids in sarcoid
lesions [letter]. Lancet 1992;339:1537.
85. Bocart D, Lecossier 0, De Lassence A. Valeyre D, Battesti J-P, Hance
AJ. A search for mycobacterial DNA in granulomatous tissues from
patients with sarcoidosis using the polymerase chain reaction. Am Rev
Respir Dis 1992; 145:1142--8.
86. Scadding JG. Further observations on sarcoidosis associated with M.
tuberculosis infection. In: Proceedings of the 5th International Confer
ence on Sarcoidosis. Prague: University Karlova, 1971:89-92.
87. Hanngren A, Biberfeldt G, Carlens E, et al. Is sarcoidosis due to an
infectious interaction between virus and mycobacterium? In: Proceed
ings of the 6th International Conference on Sarcoidosis. Tokyo: Uni
versity of Tokyo Press, 1974:8-11.
88. Nakata Y, Kataoka M, Kimura 1. Sarcoidosis and Proprionibacterium
acnes [in Japanese]. Nippon Rinsho 1994;52:1492-7.
89. Graham DY, Markesich DC, Kalter DC. Yoshimura HH. Isolation ofcell-wall defective acid-fast bacteria from skin lesions of patients withsarcoidosis. In: Proceedings of the 11th World Congress on Sarcoidosisand Other Granulomatous Disorders. Amsterdam: Excerpta Medica.Elsevier, 1988:1614.
90. Fidler HM, Rook GA, Johnson NM, McFadden J. Mvcobacterium tuber
culosis DNA in tissue affected by sarcoidosis. BM.I 1993; 306:546- 9.91. Reid .ID, Chiodini RJ. Serologic reactivity against Mvcobacterium para
tuberculosis antigens in patients with sarcoidosis. Sarcoidosis 1993:10:32-5.
92. Milman N, Andersen AB. Detection of antibodies in serum against M
tuberculosis using western blot technique. Comparison between sar
coidosis patients and healthy subjects. Sarcoidosis 1993; 10:29 3 I.
93. Abe C, Iwai K, Mikami R. Hosoda Y. Frequent isolation of Propionibac
terium acnes from sarcoidosis lymph nodes. Zentralbl Bakteriol Mik
robiol Hyg (B) 1984;256:541-7.
94. Miyagawa Y, Asoh H, Nakanishi M, Shigematsu N. Epithelioid cell
granuloma formation in Lewis rats induced by streptococcal cell wall(SCW). In: Abstracts of the 12th World Congress on Sarcoidosis and
Other Granulomatous Disorders. Kyoto, Japan: University of TokyoPress, 1989: I07.
95. Montemurro L, Rizzato G. Is sarcoidosis a borreliosis? Sarcoidosis 1991;8:134-5.
by guest on September 23, 2016
http://cid.oxfordjournals.org/D
ownloaded from
158 Zumla and James ern 1996;23 (July)
96. Bing H, Quing L, Wang F, Cheng A, Wei 1. Borrelia burgdorferi infection may be the cause of sarcoidosis. Chin Med J 1992;195:560-3.
97. HannukselaM, JannsonE. Isolationof a mycoplasmafromsarcoidtissue.In: IwaiK, HosodaY, eds. Proceedingsof the 6th InternationalConference on Sarcoidosis. Tokyo: Tokyo University Press, 1974:4.
98. Uesake E, Izumi T, Tsuki S. Nocardia-like organisms isolated from lesions of sarcoidosis. In: Iwai K, Hosoda Y, eds. Proceedings of the6th InternationalConferenceon Sarcoidosis. Tokyo:TokyoUniversityPress, 1974:3.
99. RottoliP, BianchiBandinelliML, RottoliL, ZazziM, PanzardiG, Valensin PE. Sarcoidosis and infections by human lymphotropic viruses.Sarcoidosis 1990;7:31-3.
100. Hirshaut Y, Glade P, Vieira LO, Ainbender E, Dvorak B, SiltzbachLE.Sarcoidosis, another disease associated with serologic evidence forherpes-likevirus infection. N Engl J Med 1970;283:502-6.
101. AblashiD, SalahuddinZ, Imam P, et al. Serologicalassociationof a newherpes virus (HBLV) with sarcoidosis. In: Proceedings of the II thCongresson Sarcoidosisand Other Granulomatous Disorders.Amsterdam: Excerpta Medica, 1988:68.
102. RottoliP, ValensinPE, BianchiBandinelliM, BracialeP, FortelioniGM,Vagliasindi M. Antibody responses to viral antigens in sarcoidosis[letter]. Sarcoidosis1993;10:73-4.
103. HillerdalG, Frisk G, Nettelbadt0, DiderholmH. High frequencyofIgMantibodies to coxsackie B virus in sarcoidosis patients and patientswith asbestos-related lesions. Sarcoidosis 1992;9:39-42.
104. Petterson HB, Iversen OJ, Eklung A. Retrovirus: a possible etiologicalagent of sarcoidosis. In: Abstracts of the 12th World Congress onSarcoidosis and Other Granulomatous Disorders. Kyoto, Japan: University of Tokyo Press, 1989:120.
105. Shelley WB, Hurley HJ. The allergic origin of zirconium deodorantgranulomas. Br J Dermatoll958; 70:75.
106. CummingsMM,DunnerE, SchmidtBH, BarnwellJB.Conceptsof epidemiology of sarcoidosis. Postgrad Med J 1956;19:437.
107. Cummings MM, Dunner E, Williams JH Jr. Epidemiologic and clinicalobservations in sarcoidosis. Ann Intern Med 1959;50:879-90.
108. Comstock GW, Keltz H, Sencer DJ. Clay eating and sarcoidosis: a controlled study in the state of Georgia. Am Rev Respir Dis 1961;84(5part 2)(suppl):130-4.
109. Rubin E, Farber JL. The gastrointestinal tract. In: Rubin E, Farber JL,eds. Essential pathology. Philadelphia: Lippincott, 1988.
110. WakefieldAJ, Sawyerr AM, Dhillon AP, et al. Pathogenesisof Crohn'sdisease: multifocal gastrointestinal infarction. Lancet 1989;2:105762.
Ill. Kayser K, Probst F, Gabius HJ, MOller KM. Are there characteristicalterations of lung tissue associated with Crohn's disease? Pathol ResPract 1990;186:485-90.
112. Wakefield AJ, Pittilo RM, Sim R, Cosby SL, Stephenson JR, DhillonAP, Pounder RE. Evidence of persistent measles virus infection inCrohn's disease. J Med ViroI1993;39:345-53.
113. SandersonJD. Mycobacteria in Crohn's disease [letter]. BMJ 1993;306:1131.
114. CiclitiraPI. Does Crohn's diseasehave a mycobacterialbasis? [editorial].BMJ 1993;306:733-4.
115. Smith MSH, WakefieldAJ. Viral associationwith Crohn's disease. AnnMed 1993;25:557-61.
116. EkbomA, WakefieldAJ, ZackM, AdamiHO. Perinatalmeaslesinfectionand subsequent Crohn's disease. Lancet 1994;344:508-10.
117. Lewin J, Dhillon AP, Sim R, Mazure G, Pounder RE, Wakefield AI.
Persistent measles virus infection of the intestine: confirmation byimmunogoldelectron microscopy. Gut 1995;36:564-9.
118. WakefieldAJ, Ekbom A, Dhillon AP, Pittilo RM, Pounder RE. Crohn'sdisease: pathogenesis and persistent measles virus infection. Gasteroenterology 1995; 108:911-6.
119. James DG. Definitionand classification. In: James DG, ed. Sarcoidosisand other granulomatous disorders. New York: Marcel-Dekker, 1994:33-5.
120. HanceAJ, CandranelJ, SolerP, BassetF. Pulmonaryand extrapulmonaryLangerhans' cell granulomatosis (Histiocytosis X). SeminRespir Med1988;9:349-68.
121. Smith RM, Curnutte JT. Molecular basis of chronic granulomatousdisease. Blood 1991;77:673-86.
122. Curnutte JT. Chronic granulomatous disease: the solving of a clinicalriddle at the molecular level. Clin Immunol Immunopathol 1993;67(3)(suppl):S2-15.
123. Thrasher AJ, Keep NH, Wientjes F, Segal AW. Chronic granulomatousdisease. Biochim Biophys Acta 1994;1227:1-24.
124. SchapiroBL,NewburgerPE, KlempnerMS, DinauerMe. Chronicgranulomatous disease presenting in a 69-year-old man. New Engl J Med1991;325:1786-90.
125. Melkersson E. Case of recurrent facial paralysis with angioneuroticedema. Hygiea 1928;90:737-41.
126. RosenthalC. Facialislahmungund lingua plicata.Z Neurol Psycholl990;131:475-501.
127. Larsson E, Westermark P. Chronic hypertrophic vulvitis-a conditionwith similarities to cheilitis granulomatosa (Melkersson-Rosenthalsyndrome). Acta Dermatovenereol (Stockholm) 1978;58:92-3.
128. Klaus SN, BrunstingLA. Melkersson's syndrome(persistentswellingofthe face, recurrent facial paralysis and lingua plicata): report of case.Proc Mayo Clin 1959;34:365- 70.
129. WiesenfeldD, FergusonMM,MitchellDN,et al. Oro-facialgranulomatosis: a clinical and pathological analysis. Q J Med 1985;54:101-113.
130. KikuchiM. Lymphadenitis showingfocal reticulumcell hyperplasiawithnuclear debris and phagocytosis. Nipon Ketsueki Gakkai Zassi 1972;35:379-80.
131. Kuo T-t. Kikuchi's disease (histiocyticnecrotizinglymphadenitis): a clinicopathologic studyof79 caseswith an analysisof histologicsubtypes,immunohistology, and DNA ploidy. Am J Surg Patho11995;19:798809.
132. Gourley I, Bell AL, Biggart D. Kikuchi's disease as a presenting featureof mixed connective tissue disease. Clin Rheumatoll995; 14:104-7.
133. Fred HL. Kikuchi-Fujimoto disease. Houston Med 1991;7:115-7.
by guest on September 23, 2016
http://cid.oxfordjournals.org/D
ownloaded from