sign in sign up
<<
Home , Alcian blue stain, Carbol fuchsin, Giemsa stain, Gram stain, Koilocyte, Papanicolaou stain

CLINICAL REVIEWS, July 1996, p. 382–404 Vol. 9, No. 3 0893-8512/96/$04.00ϩ0 Copyright ᭧ 1996, American Society for Microbiology

Detection of or Infectious Agents by Use of Cytologic and Histologic Stains

GAIL L. WOODS* AND DAVID H. WALKER Department of , University of Texas Medical Branch, Galveston, Texas 77555-0743

INTRODUCTION ...... 382 STAINS FOR DIAGNOSIS OF VIRAL ...... 383 Direct Detection in Smears ...... 383 Detection in Tissue Sections ...... 385 STAINS FOR DETECTION OF CHLAMYDIA TRACHOMATIS AND RICKETTSIAE ...... 387 Direct Detection in Smears ...... 387 Detection in Tissue Sections ...... 387 STAINS FOR DETECTION OF ...... 389 Direct Detection in Smears ...... 389 Detection in Tissue Sections ...... 392 STAINS FOR DETECTION OF MYCOBACTERIA ...... 393 Direct Detection in Smears ...... 394 Detection in Tissue Sections ...... 395 STAINS FOR DETECTION OF FUNGI...... 396 Direct Detection in Smears ...... 396 Detection in Tissue Sections ...... 397 STAINS FOR DETECTION OF PARASITES...... 399 Direct Detection in Smears ...... 399 Detection in Tissue Sections ...... 399 CONCLUSIONS ...... 400 REFERENCES ...... 401

INTRODUCTION inclusion-bearing cells are observed, the patient does not have CMV . For most patients with infectious , microbiological Frequently, the histopathologic damage, the host response, and identification techniques offer the most rapid and and the cultivated organism have a long-established associa- specific determination of the etiologic agent. On the other tion, lending strong support to the diagnosis. For example, hand, visualization of organisms in cytologic smears, tissue goes hand in hand with the pres- sections, or both often adds important information and in ence of caseous granulomas. Granulomas are identified micro- some circumstances is crucial to the establishment of a timely scopically by the pathologist by the criterion of clusters of diagnosis. Microbial culture alone cannot distinguish between activated ; caseous is identified by the colonization (the growth of organisms on the body surface) or, pathologist grossly by the presence of material resembling the in the case of some , asymptomatic shedding and tissue curds of cottage cheese or feta cheese. Pathologic lesions as- invasion. For example, recovery of , a com- sociated with an organism, however, often vary, depending on ponent of the normal flora of the oral cavity, from a sputum the competence of the host immune defenses; therefore, culture does not necessarily indicate that the organism is caus- knowledge of the immune status is useful in interpreting tissue ing pneumonia; this requires identification of pseudohyphae, lesions diagnostically. It is ideal for the pathologist to know the budding , or both invading the lung parenchyma. Isola- culture results from the microbiology and to con- tion of from a burn wound specimen sider them before finalizing a pathologic diagnosis; however, may represent colonization of the wound or a life-threatening this is not always possible. For example, mycobacterial and situation; histologic examination of tissue from the burn fungal cultures are held for several weeks, and if no organisms wound site reveals whether gram-negative bacilli are present are growing in these cultures at the time the pathologic diag- only on the dead tissue of the burn eschar or have invaded nosis must be made, which generally is within a few days, final viable tissue. Similarly, evaluation of tissue for the presence of culture results cannot be considered. compatible lesions can be an important factor in assigning an Other situations in which evaluation of smears and tissue etiologic role to an agent recovered in culture. For example, sections stained for organisms are useful include those when isolation of (CMV), a prevalent persistent multiple organisms are cultured and certain situations when agent, from a respiratory specimen such as bronchoalveolar none are recovered. When several organisms are cultivated, lavage fluid or lung tissue does not prove that it is the cause of the morphology of the predominant organism visualized in the illness; culture results must be considered in conjunction tissue sections can suggest the true causative agent. For exam- with clinical and pathologic data. If normal lung tissue and no ple, if a mixture of spp., Klebsiella pneumoniae, and aureus were recovered from sputum and if histologic examination of the lung specimen revealed * Corresponding author. Phone: (409) 772-4851. only large gram-negative bacilli, obviously the most likely eti-

382 VOL. 9, 1996 CYTOLOGIC AND HISTOLOGIC STAINS FOR DETECTION 383

TABLE 1. -induced changes in cellular morphology observed in cytologic preparations and tissue Virus Description of cytopathic changes HSV and VZV ...... Multinucleated giant cell (up to 100 ␮m) with molding of nuclei against each other, and/or infected cell with an intranuclear inclusion that fills the nucleus, pushing nuclear to the margin, or with a central dense irregular body surrounded by a clear zone and by clumped nuclear chromatin (inclusion scar) (Fig. 1 and 4)

CMV...... Enlarged cell with amphophilic intranuclear inclusion, often surrounded by a halo, and/or multiple granu- lar often cytoplasmic inclusions (Fig. 3); dying infected cells may appear shrunken and smudged with poorly defined inclusions (resembling inclusions of adenovirus)

Adenovirus...... Cell with an enlarged nucleus containing an amphophilic or basophilic inclusion and thin rim of (smudge cell) (Fig. 5); an earlier form of inclusion resembles a HSV inclusion

Human papillomavirus ...... “Balloon cell” (): enlarged intermediate cell containing one or more hyperchromatic irregular nuclei that are surrounded by a variably sized clear area (Fig. 2); smears stained by the Papanicolau method show isolated cells or small aggregates of cells showing dyskeratotic changes with orangeophilic cytoplasm and irregular, small, dense nuclei

Parvovirus B19...... Erythrocyte precursors with enlarged, ballooned, eosinophilic nuclei that have marginal inclusions (lantern cells) (Fig. 6)

JC virus...... Enlargement of oligodendrogliocyte nucleus with a basophilic, eosinophilic, or amphophilic intranuclear inclusion (Fig. 7)

BK virus ...... Large, dense, homogeneous, basophilic intranuclear inclusion that usually fills the entire nuclear envelope and has a smudgy appearance

Measles virus ...... Multinucleate giant cells with eosinophilic intracytoplasmic and sometimes intranuclear inclusions

RSV ...... Epithelial cell with rare pink intracytoplasmic inclusions, often in paranuclear location; particularly in pro- longed infections of immunocompromised hosts, giant cells may be present with more numerous and larger cytoplasmic inclusions (Fig. 8)

Variola and vaccinia viruses ...... Enlarged squamous epithelial cells with brick-shaped, eosinophilic intracytoplasmic inclusions (Guarnieri bodies) surrounded by a halo

Molluscum contagiosum virus ...... Squamous epithelial cell with a large, hyaline, acidophilic intracytoplasmic granular mass (molluscum body) that almost fills the cell (Fig. 9); inclusion stains brown with Lugol’s iodine

Rabies virus ...... Eosinophilic, round to oblong (2–10 ␮m) intracytoplasmic inclusions (Negri bodies) (Fig. 10) of variable size with basophilic stippling, most commonly in the hippocampus, horn of Ammon, and Purkinje cells of cerebellum; inclusions appear cherry-red with Seller stain

Hepatitis B virus ...... “Ground-glass” hepatocytes; intracytoplasmic deposits detected with special stains (modified trichrome, orcein, Victoria blue)

ologic agent would be K. pneumoniae. Histologic examination gist and pathologist to communicate effectively, hopefully lead- of tissue traditionally has been a method for diagnosis of in- ing to earlier, more precise diagnosis of patient . fection caused by noncultivable agents such as , Pneumocystis carinii, Treponema pallidum, the Whip- STAINS FOR DIAGNOSIS OF VIRAL INFECTIONS ple’s disease bacterium (Tropheryma whipplei), and Loboa loboi. Often, histologic or cytologic evaluation yields a more Diagnosis of a viral infection based on microscopic exami- timely diagnosis for diseases caused by organisms that are nation of stained smears or tissue sections is limited to viruses difficult or dangerous to culture (e.g., rickettsiae) or are de- that induce characteristic morphologic changes during their repli- tected late because they grow slowly (e.g., mycobacteria and cation in the cells they infect or for which specific antibodies to bartonellae). quantitatively significant and stable are commercially The goal of this review is to acquaint microbiologists with available. Morphologic features indicative of a viral infection the methods available for detecting infectious lesions and agents include the formation of inclusion bodies (masses of material by using cytologic and histologic stains. These approaches are consisting of viral particles or excess accumulation of products at times more powerful than those of the microbiology labo- of viral synthesis) in the host , cytoplasm, or both ratory, but they have limitations and pitfalls. For example, and, for some viruses, multinucleate giant cells (summarized in tissue occasionally is placed in fixative without considering the Table 1 and illustrated in Fig. 1 to 10) (37, 119, 157). potential need for cultures, in which case microbiologic studies are impossible unless additional tissue is obtained; and prior Direct Detection in Smears antimicrobial therapy may limit the ability to recover the re- sponsible pathogen in culture. Understanding cytopathologic Microscropic examination of cells in stained smears as a and histopathologic methods enables the clinical microbiolo- mechanism for diagnosis of viral infection was popularized in 384 WOODS AND WALKER CLIN.MICROBIOL.REV.

FIG. 1. Smear of cells from the uterine showing multinucleate giant FIG. 2. Smear of cells from the uterine cervix showing koilocyte (arrow) cell consistent with HSV infection. was used. Magnification, consistent with human papillomavirus infection. Papanicolaou stain was used. ϫ500. Courtesy of Vicki Schnadig, University of Texas Medical Branch, Magnification, ϫ500. Courtesy of Vicki Schnadig, University of Texas Medical Galveston. Branch, Galveston. the late 1940s by Tzanck, who used the technique to study cells that of for HSV, and identification of the specific scraped from the base of vesicular skin lesions as an aid in virus (HSV or VZV) is possible (68, 131, 155). dermatologic diagnosis. In particular, the Tzanck preparation In certain situations, examination of cytologic preparations allowed differentiation of (HSV) or varicella- stained with nonspecific stains is useful for diagnosis of several zoster virus (VZV) infection, based on specific cytopathic viral infections other than those caused by HSV and VZV changes (described in Table 1 and illustrated in Fig. 1), from (specific changes in cellular morphology are described in Table other dermatologic diseases with a similar clinical presentation 1). For example, examination of a Papanicolaou-stained smear (14, 69, 178). Cytologic preparations of material collected from of cells collected from the uterine cervix is a valuable screening a variety of sites have also proven useful for detection of tool for diagnosis of infection with genital human - cytopathic changes of several different viruses (5, 17, 202). viruses (Fig. 2) (6, 29, 97, 121, 122). Diagnosis of molluscum To perform the Tzanck test, the lesion selected for study is contagiosum, although usually based on the clinical appear- cleaned, the vesicle or pustule is opened or the crust is re- ance of the lesion, may be confirmed by examination of a moved with a scalpel blade, the base is scraped vigorously with smear of the cellular material collected from the central cavity the edge of the blade, and the cellular material collected is and stained with Lugol’s iodine (119). The -con- evenly spread on a glass microscope slide. As originally de- taining matrix takes up the stain, and the inclusions appear as scribed by Tzanck, the preparation is allowed to air dry and dark-brown masses. then stained with Giemsa or a Romanowsky polychrome dye Evaluation of cytologic preparations of bronchoalveolar la- such as the Wright stain; toluidine blue O is an acceptable vage fluid stained by the Papanicolaou technique or Giemsa alternative. These stains are simple to use, but air drying re- stain is a rapid method, used predominantly with specimens sults in poor definition of nuclear detail, which could make from immunocompromised hosts, for diagnosis of recognition of cytopathic changes difficult. Modifications of the caused by some viruses, including CMV (Fig. 3), adenovirus, technique that allow better preservation of cellular detail and respiratory syncytial virus (RSV), HSV or VZV, or thus may be more sensitive than air-dried preparations include virus (140, 202). For RSV, the changes often are subtle and are immediate fixation in 95% or 70% and the use of alternative stains, such as Papanicolaou, mod- ified Papanicolaou, Diff-Quik (another Romanowsky dye prep- aration), or Paragon Multiple Stain (9, 112, 155). Smears also may be stained with monoclonal antibodies, which increase diagnostic sensitivity and specificity and allow the diagnosis of infection with viruses that do not induce visible cytopathic changes. Although the Tzanck preparation is useful for rapid diagno- sis of cutaneous or mucocutaneous lesions caused by HSV or VZV, the test has limitations. It does not distinguish between HSV and VZV, and the sensitivity is less than 100%. Viral cytopathic changes are most likely to be detected in smears prepared from vesicles, followed by pustules, and then crusted lesions. When investigators have used clinical impression as the standard, the overall sensitivity of the Tzanck smear is about 50 to 60% for diagnosis of HSV infection and 64 to 100% for VZV, whereas the sensitivity of cell culture is ap- proximately 80% for HSV and only 26 to 60% for VZV (134, FIG. 3. Smear of cells collected by bronchoalveolar lavage from a patient 155, 162, 163). By with monoclonal antibodies, using with AIDS, showing a cell with intranuclear and intracytoplasmic inclusions (arrow) consistent with CMV infection. Papanicolaou stain was used. Magnifi- immunofluorescence or immunoperoxidase techniques, the cation, ϫ1,000. Courtesy of Vicki Schnadig, University of Texas Medical Branch, sensitivity of the direct smear increases but remains lower than Galveston. VOL. 9, 1996 CYTOLOGIC AND HISTOLOGIC STAINS FOR PATHOGEN DETECTION 385

FIG. 4. Paraffin-embedded section of skin from an immunocompromised FIG. 5. Paraffin-embedded section of from an immunocompromised patient with disseminated varicella showing cytopathic changes consistent with patient showing a “smudge cell” (arrow) consistent with adenovirus (culture or , i.e., multinucleate giant cell and confirmed the diagnosis). The section was stained with H&E. Magnification, many cells with intranuclear inclusions. H&E stain was used. Magnification, ϫ200. ϫ200.

of influenza and parainfluenza viruses and should be per- found only in a small proportion of epithelial cells, so that their formed if the immunofluorescence test is negative (47, 108, recognition requires careful search. Prolonged RSV infection 117, 147). Other uses of virus-specific antibodies include rapid in immunocompromised hosts, however, may result in produc- diagnosis of measles by detection of viral in smears of tion of multinucleate giant cells containing eosinophilic cyto- nasopharyngeal cells (124), diagnosis of CMV pneumonia by plasmic inclusions. To confirm infection with any of these vi- detection of virus-infected cells in smears of bronchoalveolar ruses, a smear may be stained with specific antibodies (165). lavage fluid (166), and differentiation of asymptomatic shed- Examination of a smear of Wright- or Giemsa-stained bone ding of CMV during reactivation from active disease. The last marrow aspirate for characteristic erythroblast inclusions (Ta- is based on detection and quantitation of antigenemia by stain- ble 1) often is necessary to confirm parvovirus B19 as the cause ing smears of peripheral leukocytes (buffy coat) with a mono- of chronic anemia in an immunocompromised patient, espe- clonal antibody against CMV pp65 structural antigen (15, 179). cially a person with AIDS, because serologic tests generally are not helpful in such cases. Infection with BK virus or CMV may Detection in Tissue Sections be diagnosed by detection of cells showing cytopathic changes characteristic of the respective virus in cytologic preparations The first step in the routine diagnosis of any infectious dis- of sediment stained by the Papanicolaou technique (5, ease from tissue specimens is examination of sections stained 17, 36). For CMV, however, shell vial centrifugation-enhanced with hematoxylin and (H&E). Although the individual cell culture with detection of CMV nuclear antigens is a more viruses cannot be seen, many induce easily recognized cyto- sensitive and more frequently used diagnostic test. pathic changes (described in Table 1 and illustrated in Fig. 4 to Microscopic examination of imprints of brain tissue stained 10) that in most cases are sufficient for diagnosis (37, 157). with Seller stain (a mixture of basic fuchsin and methylene Other features of viral infection are cell necrosis, which occurs blue) for Negri bodies is a useful method for rapid diagnosis of during release of viral particles after replication, and an in- rabies, especially in dogs (107). However, rabies in wildlife, flammatory infiltrate composed predominantly of ; particularly skunks and bats, may be caused by strains of the macrophages generally appear later in the infection. In certain virus that do not produce Negri bodies, and Negri-like cyto- plasmic inclusion bodies may occur in neural tissues in the absence of rabies (28, 45, 46). For these reasons, use of the Seller stain for diagnosis of rabies has for the most part been replaced by immunofluorescence staining with specific anti- bodies (107). For some viral infections, diagnosis based on visualization of characteristic morphologic changes in cells stained with a non- specific stain is not possible or is very difficult, but in many of these diseases a diagnosis can be made by immunofluorescence staining of smears with virus-specific antibodies. This approach to diagnosis is used most widely for detection of respiratory viruses (RSV, influenza A and B viruses, and parainfluenza viruses) in smears of nasopharyngeal cells collected by aspira- tion or swab. Advantages of immunofluorescence staining are rapid turnaround time, high specificity, and the ability to assess specimen adequacy. For RSV, the sensitivity of immunofluo- rescence staining is equal to or greater than that of other available diagnostic tests (enzyme immunoassay and cell cul- FIG. 6. Paraffin-embedded section of liver from a fatal case of erythroblas- tosis fetalis caused by paravovirus B19, showing extramedullary erythropoiesis ture) (47, 88, 117, 125, 169, 173). In contrast, cell culture is and lantern cells (erythrocyte precursors with eosinophilic nuclei with marginal more sensitive than immunofluorescence staining for detection inclusions) (arrow). The section was stained with H&E. Magnification, ϫ300. 386 WOODS AND WALKER CLIN.MICROBIOL.REV.

FIG. 7. Paraffin-embedded section of brain from a patient with AIDS show- FIG. 9. Paraffin-embedded section of skin showing squamous epithelial cells ing cytopathic changes consistent with JC virus infection, i.e., enlargement of with large intracytoplasmic inclusions diagnostic of molluscum contagiosum. The oligodendrogliocyte nucleus with intranuclear inclusion (arrows). H&E stain was section was stained with H&E. Magnification, ϫ200. used. Magnification, ϫ200.

cess is determining optimal dilutions of antibody and conju- viral infections, such as HSV bronchopneumonia or bronchitis, gate, which must be done for both manual and automated polymorphonuclear leukocytes (PMN) may be present, even in staining techniques. For liver tissue, an initial avidin-biotin the absence of secondary bacterial infection. The composition blocking step, which generally is not necessary with other types and degree of the inflammatory response to any infectious of tissue, must be performed. Obtaining adequate positive con- agent, however, depends in large part on the host immune trol tissue also is often a challenging task. A potential pitfall of status. is failure to detect viral (or other tar- Occasionally, the changes observed in H&E-stained sections get) antigen because of prolonged storage of the tissue in are insufficient for diagnosis and additional stains or other tests (i.e., a false-negative result). In these cases, ad- are necessary. Only a few nonspecific “special stains” have ditional steps for antigen retrieval must be performed, such as been useful for diagnosis of viral infections. Lendrum’s phlox- protease digestion or heating in a microwave oven or a 95 to ine-tartrazine method enhances detection of viral inclusion 99ЊC water bath with sodium citrate buffer. bodies, staining them red against a yellow background (106), Another consideration with regard to the use of immuno- and Parson’s stain and Schleifstein’s method have been used histochemical analysis is cost. Because antibodies are more for detection of Negri bodies. Orcein, modified trichrome, or expensive than chemicals and reagents used in nonspecific Victoria blue-nuclear fast red were once used for confirmation special stains, limiting their use to specific situations is reason- of chronic hepatitis B virus infection, apparently on the basis of able. For example, when H&E-stained sections of liver tissue the presence of disulfide bonds in hepatitis B virus surface show changes consistent with acute or chronic hepatitis, immu- antigen (44, 76, 171). Today, however, immunohistochemical nohistochemical analysis with commercial antibodies against assays with commercial virus-specific antibodies are preferred. the surface and core antigens of hepatitis B virus and against Immunohistochemistry has been perfected to a level such hepatitis C virus often identifies the responsible pathogen. that, in general, it is no more difficult to perform than many However, this is not necessary if the agent causing the infection nonspecific “special stains.” For institutions that perform large has been determined by serologic studies. With regard to di- volumes of immunohistochemical analyses, an automated agnosis of infection caused by HSV or CMV, immunohisto- stainer, which has the advantages of labor savings and consis- chemical studies are seldom convincingly positive in the ab- tency, is available. The most time-consuming step in the pro- sence of viral inclusions (168) and thus are most useful when

FIG. 8. Paraffin-embedded section of lung from an immunocompromised patient showing a multinucleate giant cell; viral cultures of lung tissue were FIG. 10. Paraffin-embedded section of cerebellum showing a Purkinje cell positive for RSV. The section was stained with H&E. Magnification, ϫ200. containing a Negri body (arrow). H&E stain was used. Magnification, ϫ200. VOL. 9, 1996 CYTOLOGIC AND HISTOLOGIC STAINS FOR PATHOGEN DETECTION 387 the cytopathic changes are not typical, as when dying cells during either the first or a return visit. Immunofluorescence infected with CMV appear shrunken and smudged with poorly staining is not recommended for use in cases of suspected defined inclusions, resembling inclusions of adenovirus. Simi- sexual abuse. larly, immunohistochemical studies with antibodies against hu- With regard to detection of rickettsiae, clinical samples that man papillomavirus, JC virus, or adenovirus may be useful for contain a sufficient quantity of organisms to be visualized di- confirmation of infection with the respective virus. rectly in smears have been limited to detection of human granulocytotropic ehrlichia in smears of peripheral blood STAINS FOR DETECTION OF CHLAMYDIA stained with a in which this as yet unculti- TRACHOMATIS AND RICKETTSIAE vated organism, closely related to Ehrlichia phagocytophila and Ehrlichia equi, can be observed growing as microcolonies in With regard to chlamydiae, stains are used predominantly cytoplasmic vacuoles of circulating (Fig. 11) (7). for diagnosis of infections caused by Chlamydia trachomatis; On rare occasions, Ehrlichia chaffeensis has been identified they have no proven value in diagnosis of infections caused by within cytoplasmic vacuoles, mainly in . For most Chlamydia psittaci or Chlamydia pneumoniae. For rickettsial rickettsioses, the obligately intracellular organisms are located diseases, because serologic test results usually are negative inside cells within the tissues and therefore are not available in during the acute state of the illness and cultivation is rarely sufficient quantities to be detected in smears. To address the performed, staining of organisms in biopsy specimens of a skin problem of collecting the target cells of rickettsial infection rash offers one of the few possibilities of providing a timely (i.e., the endothelial cells that line the blood vessels throughout laboratory diagnosis. the body), magnetic beads may be coated with a monoclonal antibody to a human endothelial antigen. By Direct Detection in Smears incubating these beads with anticoagulated blood from a pa- tient, circulating, detached endothelial cells are captured. Im- Microscopic examination of stained smears is most useful munofluorescence staining of endothelial cells captured in this for rapid diagnosis of sexually transmitted disease or inclusion manner has been used for identification of Rickettsia conorii in conjunctivitis caused by C. trachomatis. Intracytoplasmic inclu- French patients with boutonneuse fever (48). This technique sions of C. trachomatis may be seen in cytologic preparations of currently is limited to research ; the antibody is not conjunctival scrapings prepared with the . This commercially available. method of diagnosis, however, has been replaced for the most When clinical specimens are cultured in the yolk sac of an part by direct immunofluorescent staining of elementary bod- embryonated chicken egg or cell cultures or inoculated intra- ies with commercial monoclonal antibodies, which is both peritoneally into guinea pigs (e.g., Rickettsia rickettsii or Rick- more sensitive and more specific than the Giemsa stain. Cur- ettsia typhi) or mice (e.g., Orientia [Rickettsia] tsutsugamushi), rently, antibodies for diagnosis are available from several man- smears of the cultures are examined microscopically for the ufacturers, but not all are approved for all types of specimens. presence of organisms. The intracellular rickettsiae are visual- Antibodies against the species-specific major outer membrane ized when stained with a fuchsin stain such as the Gimenez or of C. trachomatis appear to be more specific and pro- Macchiavello method or with a Romanowsky stain such as duce more intense fluorescence than those against the chamy- Giemsa or Diff-Quick (67). dial (35, 200). Advantages of immunofluorescence staining for detection of Detection in Tissue Sections C. trachomatis are rapid turnaround time and the ability to assess specimen quality. Smears with columnar or metaplastic Tissue for histologic evaluation rarely is necessary for diag- squamous cells are acceptable, whereas those with few colum- nosis of chlamydial infections. One possible exception is diag- nar cells, excessive amounts of mucus, or a predominance of nosis of lymphogranuloma venereum, in which case lymph squamous cells are not. However, interpretation of the smear node biopsy may be performed. The presence of necrotizing is subjective, and microscopist fatigue can be a problem in granulomas in H&E-stained sections is consistent with lym- high-volume situations. The specificity of immunofluorescence phogranuloma venereum, but other organisms (Table 2) can staining, compared with culture as the “gold standard,” is 95% induce identical changes. The Giemsa stain has been used for or greater; the sensitivity varies from 50% to almost 100%, detection of chlamydiae in tissue sections, but the results are depending on the prevalence of infection in the population not consistent; therefore, other tests, most often , are being evaluated, the number of elementary bodies required for necessary for specific diagnosis. a positive result, and the technical ability of the microscopist Rocky Mountain spotted fever was demonstrated in 1906 to (8). In general, sensitivity is greater in populations with a high be an infectious disease that is transmissible experimentally to prevalence of disease and when a lower cutoff point for num- animals, and organisms suspected to be the etiologic agents ber of elementary bodies is required for a positive result. False- were observed in stained specimens from ticks. However, the positive results by immunofluorescence staining may occur criteria of Koch’s postulates were not met, because these ob- even with putative species-specific antibodies, perhaps as a ligately intracellular organisms could not be cultivated on agar result of nonspecific immunoglobulin binding or shared or broth medium. Visualization of the rickettsiae in endothe- epitopes. The problem of false positivity is most common with lial cells at the loci of vascular injury provided convincing rectal specimens; therefore, although some immunofluores- evidence for their etiologic role. The Giemsa method for stain- cence tests are approved for this source, culture is preferred. ing of rickettsial organisms in tissue sections was developed by Experts at the Centers for Disease Control and Prevention Wolbach between 1915 and 1922 and applied to and recommend that a positive result with a screening test (such as biopsy material of patients with Rocky Mountain spotted fever immunofluorescence staining) be verified with a supplemental and epidemic louse-borne fever (198). Wolbach’s Gi- test if a false-positive result is likely to have adverse medical, emsa preparation, however, is different from the routinely social, or psychological consequences (30). For laboratories available Giemsa tissue stain used to evaluate hematopoietic that use immunofluorescence staining, this recommendation cells, which does not stain rickettsiae. Subsequently, Pinker- currently means collection of a second specimen for culture, ton’s stain was introduced for detection of rickettsiae. Success- 388 WOODS AND WALKER CLIN.MICROBIOL.REV. VOL. 9, 1996 CYTOLOGIC AND HISTOLOGIC STAINS FOR PATHOGEN DETECTION 389 ful staining by these two methods requires close attention to those that have not yet been cultured on solid media, such as details such as fixation in Regaud’s solution for Wolbach’s spirochetes) are discussed in this section. Mycobacteria, which Giemsa stain or in Zenker’s solution for Pinkerton’s stain. have a much different composition (although some- Seldom are these freshly prepared fixatives available and used. what similar to actinomycetes), have quite different staining Observation of thin bacilli lying longitudinally, often parallel characteristics and therefore are discussed separately. with one another in the cytoplasm of endothelial cells, is diag- nostic of infection with an undetermined Rickettsia species. Direct Detection in Smears Identification of rickettsiae that have invaded vascular smooth muscle cells or the nuclei of infected cells is considered to be The most frequently used technique for direct detection of diagnostic of Rocky Mountain spotted fever. A modification of bacteria is the , introduced by the Danish patholo- the Brown-Hopps stain prepared with a substantially higher gist over a century ago and modified only concentration of basic fuchsin has been used for visualization slightly since then (10, 11, 19, 145). This procedure differen- of rickettsiae in tissue, but this method stains only a fraction of tially divides the bacteria into two groups, gram positive and the quantity of rickettsiae that are observed by specific immu- gram negative. Those that are gram positive retain the crystal nofluorescence or immunoenzyme staining (185, 186). violet-iodine complex and appear purple or deep blue, whereas Immunohistologic staining has been used to demonstrate R. gram-negative bacteria are decolorized by alcohol or rickettsii, R. conorii, R. akari, R. typhi, R. prowazekii, and Cox- but are visible because they red or pink with iella burnetii in human tissues (25, 90, 187–189). For patients in . Epithelial and inflammatory cells are also stained, whom a rash has developed, often not until day 3 to 5 of the thus providing information about the host immune response illness, rickettsiae can be stained in a 3-mm punch biopsy and the quality of the specimen. specimen of the cutaneous rash by immunofluorescence or In the hands of a skilled observer, the Gram-stained smear immunoenzyme methods applied to frozen sections or forma- is a powerful diagnostic tool, allowing rapid preliminary iden- lin-fixed, paraffin-embedded tissue. The sensitivity of this ap- tification of the organisms on the basis of their Gram reaction, proach is about 70%; the specificity in the hands of an expe- morphology, and arrangement and an assessment of their clin- rienced microscopist is 100%. ical significance, determined by the proportion of epithelial Although ehrlichiae can be suspected to be present on the and inflammatory cells (191). The most valuable information is basis of Giemsa or staining of tissue, E. chaffeensis can be gleaned from specimens that have a predominance of inflam- demonstrated by specific monoclonal antibody-based immu- matory cells and a single (or predominant) bacterial cell mor- noenzyme staining and the human granulocytic ehrlichia has photype. For example, the presence of numerous PMNs, few been detected by immunoenzyme staining of necropsy tissues epithelial cells, and many lancet-shaped gram-positive diplo- (7, 204). cocci in a sputum smear suggests pneumococcal pneumonia Cardiac valves removed from patients with chronic culture- (16, 55); a similar picture but substituting intracellular gram- negative endocarditis can be examined by immunohistologic negative diplococci for gram-positive diplococci, is suggestive methods for C. burnetii to document Q fever endocarditis. Q of Moraxella catarrhalis. A Gram-stained smear of urethral fever also may be suspected on the basis of the presence of exudate from a symptomatic male showing intracellular gram- fibrin ring granulomas (epithelioid granulomas with fibrinoid negative kidney-bean-shaped diplococci is diagnostic of gon- material and a central clear space) in H&E-stained sections of orrhea. Infection with an actinomycete ( or Nocar- a liver or biopsy specimen, but this finding is not dia spp.) is suspected on the basis of characteristic findings in pathognomonic. Practically, however, chronic Q fever is most a smear of material obtained from an appropriate source. A easily diagnosed serologically. smear of exudate from a cutaneous or subcutaneous lesion, material, or sputum typically contains many PMNs and delicate, beaded, gram-positive filaments, approximately 1 ␮m STAINS FOR DETECTION OF BACTERIA in diameter, that occasionally form granules embedded in a Bacteria that clinical microbiologists generally consider pink, amorphous matrix, or that fragment, appearing as chains usual (those detected in 24 to 48 h by routine culture methods or clusters of bacilli or coccobacilli (165). Differentiation of in the general bacteriology section of a laboratory) or unusual Actinomyces and Nocardia spp. is possible on the basis of re- (those that require a special medium and/or slightly longer sults of special stains (discussed later in this section). In con- incubation, such as actinomycetes and ,or trast to these examples, the clinical significance of a specimen

FIG. 11. Peripheral blood smear showing a polymorphonuclear leukocyte containing a morula (microcolony) (arrow) of human granulocytotrophic ehrlichiae. A Wright-Giemsa stain was used. Magnification, ϫ630. FIG. 15. Paraffin-embedded section of lung showing gram-negative bacteria invading a blood vessel wall (the culture grew Pseudomonas species). The section was stained with the Brown-Hopps stain. Magnification, ϫ400 (insert, ϫ1,000). FIG. 17. Paraffin-embedded section of lung showing acute inflammation. The section was stained with H&E. Fite stain (insert) shows thin, branching acid-fast bacilli. Culture of the lung tissue grew Nocardia asteroides. Magnification, ϫ630. FIG. 18. Smear of material aspirated from an osteolytic lesion involving the humerus of a patient with AIDS, demonstrating a long AFB resembling a shepherd’s crook. Cultures of the aspirated material grew Mycobacterium kansasii. The Ziehl-Neelsen stain was used. Magnification, ϫ1,250. Courtesy of Vicki Schnadig, University of Texas Medical Branch, Galveston. FIG. 19. Paraffin-embedded section of a caseous lesion in lung tissue showing granulomatous inflammation and a Langerhans giant cell. The section was stained with H&E. Magnification, ϫ400. Ziehl-Neelsen stain (insert) shows a few AFB (magnification, ϫ1,000); culture of the tissue grew Mycobacterium tuberculosis. FIG. 20. Paraffin-embedded section of a skin nodule from a person with lepromatous leprosy shows a large aggregate of macrophages. The section was stained with H&E. Magnification, ϫ100. Fite stain (insert) shows numerous AFB. Magnification, ϫ630. FIG. 21. (A) Smear of sediment of CSF from a patient with AIDS, showing poorly staining budding yeasts. Culture of the CSF grew Cryptococcus neoformans. The Gram stain was used. Magnification, ϫ1,500. (B and C) Paraffin-embedded section of spinal cord from an AIDS patient with fatal cryptococcal , showing budding yeasts. The section stained with mucicarmine (B) shows positively staining (red) capsules, and the section stained by the Fontana-Masson method (C) shows positively staining (brown) cells. Magnification, ϫ400. FIG. 27. Paraffin-embedded section of a nodular skin lesion showing a spherical, thick-walled brown sclerotic body diagnostic of chromoblastomycosis. The section was stained with H&E. Magnification, ϫ400. 390 WOODS AND WALKER CLIN.MICROBIOL.REV.

TABLE 2. Infectious agents commonly causing a mixed suppurative and granulomatous response and/or necrotizing granulomasa

Tissue stains for Organism or disease Organism appearanceb Comments organismsb

Nontuberculous mycobacteriac Auramine-rhodamine, Yellow fluorescent (with fluorochrome) or M. fortuitum-chelonae complex may not Kinyoun, Ziehl- red (with ) beaded bacilli stain by fluorochrome technique Neelsen

Blastomyces dermatitidis H&E, PAS, GMS Spherical multinucleate yeasts (8–15 ␮m) Cell wall of yeast may stain weakly with with thick (double contour) walls and sin- mucicarmine gle, broad-based buds (Fig. 25)

Sporothrix schenckii PAS, GMS Pleomorphic, spherical, oval, or cigar-shaped Yeast cells often are not detected in tissue yeasts (2–10 ␮m) with single buds

Paracoccidioides brasiliensis H&E, PAS, GMS Large spherical yeasts (5–60 ␮m) with multi- ple buds attached by narrow necks (“mari- ner’s wheel”)

Chromoblastomycosisd H&E, Fontana-Masson Large (6–12-␮m), spherical to polyhedral thick-walled, dark brown muriform cells (sclerotic bodies) with septations along one or two planes in subcutaneous tissue (Fig. 27); Ϯ pigmented hyphae

Systemic phaeohyphomycosise H&E, Fontana-Masson Brown pigmented hyphae (2–6 ␮m wide), branched or unbranched, often constricted at prominent septations

Acanthamoeba spp. H&E, PAS Trophozoites (22 ␮m in diameter) with large nucleolus and vacuolated cytoplasm; cysts have thick ectocyst containing endocyst, often shrunken (Fig. 32)

Yersinia enterocolitica, Yersinia Brown-Hopps Clumps of GNB Necrotizing granulomas with stellate ab- pseudotuberculosis scesses usual

Francisella tularensis NA NA Necrotizing granulomas with stellate ab- scesses usual

Bartonella henselae (?Afipia felis) Brown-Hopps, War- GNB (Warthin-Starry- and Dieterle-positive Spectrum of histologic changes is usual; thin-Starry, Dieterle bacilli) in cluster of necrotic cells (early early lesions show focal necrosis and lesion) small ; with disease progres- sion, PMNs become fragmented and macrophages surround the area; finally, areas of inflammation coalesce and are surrounded by epithelioid macrophages

Chlamydia trachomatis, serotypes NA NA Necrotizing granulomas with stellate ab-

L1–L3 scesses usual a An uncommon infectious cause of necrotizing granulomas is Prototheca spp. (olecranon bursitis). b Ϯ, may or may not be present; GNB, gram-negative bacilli; NA, not applicable because organisms typically are not detected in tissue, although tissue is stained with H&E. c Especially M. fortuitum-chelonae complex and early cutaneous lesions caused by M. marinum. d Most often Fonsecaea pedrosoi or Cladosporium carrionii. e Most often Exophila jeanselmei, Curvularia, Bipolaris, and Xylohypha spp., Phaeoannellomyces werneckii,orWangiella dermatitidis. showing multiple bacterial cell morphotypes representative of of well-mixed uncentrifuged urine is a reliable and economic, the normal flora and a predominance of squamous epithelial although labor-intensive, way to screen urine samples to de- cells is questionable. termine which are not clinically significant and thus do not In addition to providing preliminary information concerning require culture. Observation of one or more bacteria per oil identification, the Gram-stained smear allows assessment of immersion field correlates well with growth of 105 or greater the clinical relevance of a specimen. Various algorithms, based CFU/ml, but the sensitivity declines with lower colony counts on the numbers of epithelial cells and PMNs observed in a (143). Gram-stained smear, to determine whether a sputum specimen Despite its many uses in the clinical laboratory, the Gram represents saliva or deep respiratory have been pro- stain has limitations. , which lack a cell wall, and posed (132). Specimens judged to be saliva will not provide spirochetes do not stain by the Gram method. Certain gram- clinically useful information and should not be cultured. Sim- negative bacilli stain weakly and therefore are difficult to visu- ilar criteria also may be applied to endotracheal aspirate spec- alize unless the usual procedure is modified (described below). imens (130). Examination of a Gram-stained smear of a drop The integrity of the cell wall of gram-positive organisms that VOL. 9, 1996 CYTOLOGIC AND HISTOLOGIC STAINS FOR PATHOGEN DETECTION 391 are old, dead, or damaged by antimicrobial agents is often of organisms in CSF and other body fluids and in smears disrupted; in this case, the -iodine complex is not prepared from broth (103, 104, 126), and in at retained during decolorization and the organisms appear gram least one case it allowed detection of burgdorferi in negative. Moreover, some clostridia are regularly decolorized sediment of CSF (20). Actinomyces spp. may be detected in and stain gram negative. Technical pitfalls also may cause Papanicolaou-stained smears of cells from the uterine cervix, errors in analysis, but most mistakes can be avoided by careful predominantly of women who have had an intrauterine con- attention to technique. The most common problems are im- traceptive device in place (78). proper smear preparation, such as failure to select a portion of Wright and Giemsa stains, used routinely in the hematology the material most representative of the disease process or laboratory, demonstrate relapsing fever borrelia spirochetes in preparation of a smear that is too thick; excessive heat fixation stained smears of peripheral blood. May-Grunwald-Giemsa or with subsequent distortion of organism morphology; over- or a modification thereof also is useful for detection of Helico- underdecolorization; and inexperience. Artifacts and speci- bacter pylori in touch imprints prepared from endoscopic gas- men-related factors also may cause difficulty in smear inter- tric biopsy specimens, although these bacteria can also be pretation. In particular, bits of precipitated crystal violet stain detected by Gram stain (40–42, 129, 181, 205). Similarly, Le- may be misinterpreted as gram-positive cocci or fungal hyphae; gionella bacilli can be visualized in Gram-stained smears, but and background material, such as mucus or other protein- the cells often stain weakly unless carbol fuchsin is used as the aceous matter, may be misidentified as gram-negative bacilli or counterstain, and the organisms may be detected more easily make recognition of slender gram-negative bacilli difficult. by other methods (3). Greer et al. (73) found that for detection Techniques that enhance the reliability of Gram stain anal- of Legionella spp. in impression smears of fresh or formalin- ysis involve modifications in the components of the stain, the fixed lung tissue, the Gimenez technique, which uses carbol fixation process, and/or the method of smear preparation. Vi- fuchsin and stains the bacilli bright red against a blue-green sualization of certain gram-negative bacilli that stain weakly or background, is more sensitive than Gram stain, and a silver not at all with the dyes most commonly used today to stain stain has been shown to be useful for detection of the organism bacteria by the Gram technique (i.e., Bordetella, Campy- in body fluids (52). lobacter, Helicobacter, and Legionella spp. and some anaerobic Techniques used most commonly for detection of mycobac- gram-negative bacilli) may be improved by staining with safra- teria (described in the next section) or modifications thereof nin for 1 to 2 min rather than 30 s, by adding 0.05% basic often provide preliminary identification of select bacteria. The fuchsin to the safranin, or by substituting carbol fuchsin or modified Kinyoun procedure (involving mild-acid decoloriza- basic fuchsin for safranin. Alternatively, the enhancer reagent tion) differentiates Nocardia spp., which retain the carbol fuch- tartrazine with fast green may be added before the safranin sin and appear red, from Actinomyces spp., which do not. In counterstain. This step makes the background of inflammatory smears of clinical material or fixed tissue, Rhodococcus equi cells and mucus appear gray or green rather than red or pink, and Legionella micdadei bacilli often stain positively with the thus allowing easier visualization of gram-negative organisms modified acid-fast stain, and the L. micdadei bacilli stain with (91), but it causes a slight change in the color of organisms and weakly with Kinyoun (27, 79, 196). For L. compared with that in the routine Gram stain, which initially micdadei, the property of being acid fast is lost after the or- could cause confusion for the inexperienced observer. With ganism is cultivated on artificial culture media and when res- regard to fixation, results in less distortion of bacte- piratory specimens containing L. micdadei are processed for ria and tissue cells and less interference by background debris mycobacterial culture (discussed in the subsequent section). than does heat (115), and it provides for the most reliable Specific antibodies for immunofluorescence staining are staining of anaerobes. For optimal identification of anaerobes, commercially available for a few bacteria: Francisella tularensis, performing the entire staining procedure, including methanol , Treponema pallidum, Legionella pneumo- fixation, in an anaerobic chamber appears to lessen the ten- phila, and other Legionella species (50, 51, 60, 81, 85, 86). For dency of certain gram-positive anaerobes to stain gram nega- detection of B. pertussis, a smear of nasopharyngeal cells ob- tive (87). Finally, for examination of cerebrospinal fluid (CSF) tained by swab or aspirate is prepared. Although the results are and other body fluids, use of a cytocentrifuge to concentrate available within a few hours, the test has limitations. Its sensi- the sample increases the sensitivity of the method and allows tivity is much lower than that of culture, owing to the large better preservation of cell morphology and more uniform number of organisms necessary for microscopic visualization. staining (32, 158). Moreover, the test is associated with a high percentage of Nonspecific stains other than the Gram stain are useful for false-positive results, ranging from about 7 to 40%, a problem direct detection of some bacteria; none, however, allow differ- due to variability among the skill of technologists in interpre- entiation of gram-positive and gram-negative organisms. tation of results, low-quality smears, and nonspecific reagents enhances the staining of gram-negative bacilli (60). Direct immunofluorescence staining for detection of Le- and the oral spirochetes, increases the contrast between organ- gionella bacilli in a smear prepared from sputum or other isms and background debris, and allows detection of metachro- respiratory secretions or in touch imprints of lung tissue has a matic granules of Corynebacterium diphtheriae in smears pre- sensitivity of 20 to 70% and a specificity of 90 to 95%, depend- pared from Loeffler agar. The Wayson variation of the ing on the experience of the person interpreting the test and methylene blue stain allows detection of Francisella tularensis, the quality of the microscope (50). and which often is not visualized in a Gram-stained smear, and this pneumoniae and other streptococci may fluo- method appears to be more sensitive than the Gram stain for resce in tests for legionellae owing to natural antibodies in the detecting organisms in CSF in cases of bacterial meningitis conjugate or to the nonspecific reaction of immunoglobulin G (39). Acridine orange, a fluorochrome dye that at low pH (IgG) with components in the staphylococcal cell wall such as allows differentiation of bacteria, which show orange fluores- protein A. The coccal forms of these bacteria, however, gen- cence, from background material and mammalian cells, which erally are easily distinguished from the coccobacillary to rod- fluoresce green to yellow, is particularly useful for interpreta- shaped Legionella spp. and so should not cause a problem with tion of smears prepared from thick or purulent material (58, interpretation. Some Pseudomonas species, Stenotrophomonas 99). It also is more sensitive than the Gram stain for detection maltophilia, Bacteroides fragilis, and Bordetella pertussis cross- 392 WOODS AND WALKER CLIN.MICROBIOL.REV.

TABLE 3. Infectious causes of noncaseating “mature” granulomasa

Organism Tissue stainsb Organism appearanceb Commentsc

Mycobacterium tuberculosis Auramine-rhodamine, Yellow fluorescent (with fluorochrome) or Granulomas in many organs, often in conjunc- Kinyoun, Ziehl- red (with carbol fuchsin) beaded bacilli tion with caseating granulomas (Fig. 19) Neelsen

Nontuberculous mycobacteria As above As above Primarily M. marinum (skin lesions), M. szulgai (most commonly chronic pulmonary dis- ease), and M. leprae (tuberculoid leprosy)

Histoplasma capsulatum GMS, PAS Small yeast cells (2–5 ␮m in diameter) with Granulomas in many organs single buds, often in macrophages

Brucella spp. NA NA Granulomas typically in liver, bone marrow, occasionally in lung

Toxoplasma gondii H&E, PAS Spherical cysts (up to 30 ␮m); crescentic Granulomas in lymph node, eye (ocular toxo- tachyzoites (2–3 ␮mby4–8␮m) (Fig. 31) plasmosis), rarely liver, bone marrow. Tachyzoites are PAS positive; cyst wall is weakly PAS positive

Coxiella burnetti NA NA Epithelioid granuloma and fibrinoid material; Ϯ clear central space (fibrin-ring granuloma) in liver and bone marrow (not specific for C. burnetti, may be seen with other organisms)

Ehrlichia spp. NA NA Granulomas in bone marrow or liver

CMV NA NA Granulomas in liver

Schistosma spp. H&E Eggs in different stages of destruction and Early granulomas with intact, viable eggs con- calcification tain many PMNs, a high proportion of which are ; older granulomas have eggs in different stages of destruction and calcifi- cation and consist of epithelioid cells, giant cells, and eosinophils; eggs of S. haemato- bium and S. japonicum tend to calcify more often than do eggs of S. mansoni

Dirofilaria immitis H&E Worm in pulmonary artery, diameter 140– Worm may be well preserved (recent lesion) or 200 ␮m (males) to 300 ␮m (females), necrotic with various degrees of degenera- smooth cuticle tion and calcification (older lesion)

a Other organisms that may (but usually do not) cause noncaseating granulomas are Pneumocystis carinii, Candida spp., Treponema pallidum (secondary ; granulomas rarely in skin lesions, liver, lymph node), Trypanosoma cruzi (placenta), Epstein-Barr virus, hepatitis viruses, Strongyloides stercoralis, and Ascaris lumbricoides (ova). Organisms that cause noncaseating granulomas but are less commonly (or rarely) seen in the United States are Leishmania (cutaneous ) and Toxocara spp. and Loboa loboi. b NA, not applicable because organisms typically are not seen in tissue. c Ϯ, may or may not be present. react with polyclonal antibodies against Legionella spp., poten- seen in cases of overwhelming infection with one of the pyo- tially causing false-positive results (50). Direct immunofluores- genic bacteria (especially S. aureus). Bacterial vasculitis, char- cence staining for detection of Treponema pallidum spirochetes acteristic of infection with P. aeruginosa or Vibrio vulnificus,is in smears of material collected from lesions suggestive of pri- suggested by the presence of bluish-purple, granular staining of mary syphilis is especially useful in situations when immediate vessel walls. Similarly, eosinophilic to bluish-purple granular examination of the specimen is not feasible and the reliability microcolonies of Bartonella spp. are often visible in lesions of of the dark-field examination would therefore be low. bacillary angiomatosis or hepatic peliosis (Fig. 13). In H&E- stained sections of gastric biopsy specimens that show acute Detection in Tissue Sections gastritis (described in Table 5), curved bacteria consistent with H. pylori may be seen in the layer of mucus on the crypt Tissue diagnosis of a bacterial infection begins with the (64). Other stains, however, demonstrate H. pylori recognition of a consistent pattern of inflammation in H&E- more dramatically even though they are only marginally more stained sections (typical findings are summarized in Tables 2 to sensitive than H&E (Fig. 14). 6), although it is important to remember that the inflammatory A Gram stain is most commonly used for detection of bac- response varies depending on the immune status of the host. teria in tissue sections. The Brown-Brenn technique is pre- Individual bacteria generally are not detected in H&E-stained ferred for demonstration of gram-positive bacteria, whereas tissue sections; however, exceptions do exist. Granules often the Brown-Hopps method and the Goodpasture method are are present in cases of (Fig. 12), and purple- better for detection of gram-negative organisms (Fig. 15), in- staining clumps, representing small bacterial colonies, may be cluding H. pylori, Legionella spp., and Bartonella henselae (in VOL. 9, 1996 CYTOLOGIC AND HISTOLOGIC STAINS FOR PATHOGEN DETECTION 393

TABLE 4. Common infectious causes of acute inflammation, with or without abscess formation

Organism Tissue stains Organism appearancea Comments

Pyogenic bacteriab Brown-Brenn, GPC, GNDC, GNB Brown-Hopps

Actinomycetes Brown-Brenn, GMS, Beaded, gram-positive filaments (1 ␮mindi- In H&E-stained sections, granules bordered by Fite’s or Coates ameter), may fragment and form what ap- club-like eosinophilic material (Splendore- modified Fitec pear to be chains or clusters of gram-posi- Hoeppli phenomenon) may be seen (Fig. 12). tive bacilli or coccobacilli

Candida species H&E, PAS, GMS Budding yeasts (3–5 ␮m diameter), pseudo- hyphae, occasionally septate hyphae

Aspergillus speciesd H&E, PAS, GMS Septate hyphae (3–8 ␮m; diameter) with par- Hyphae often invade blood vessels, causing allel walls, often showing repeated, dichoto- thrombosis and occlusion with consequent in- mous branching at 45Њ angle farction and hemorrhage; oxalate crystals often are associated with A. niger; identification not possible unless fruiting heads present.

Zygomycetes H&E, PAS, Ϯ GMSe Broad (8–25 ␮m in diameter), nonseptate (or Hyphae often invade blood vessels, causing sparsely septate) ribbon-like, often distorted thrombosis with consequent infarction, hemor- hyphae with nonparallel walls (Fig. 26) rhage, and necrosis; sporangia rarely seen

a GPC, gram-positive cocci; GNDC, gram-negative diplococci; GNB, gram-negative bacilli. b Includes many bacteria, such as staphylococci, streptococci, species, most members of the Enterobacteriaceae, Pseudomonas aeruginosa, Haemophilus influenzae, and many other gram-negative bacilli. c Nocardia positive; Actinomyces negative (may be weakly positive with Putt’s modification of Fite stain). d Hyphae of other hyaline moulds, such as and Fusarium species, have a similar appearance in tissue. e Ϯ, organisms stain weakly with the stain. lesions of cat scratch disease) (26, 83, 123, 176). The bartonel- and alcian blue (pH 2.5) stains, a technique as sensitive as the lae, however, tend to stain faintly with the Brown-Hopps Warthin-Starry method for detecting the bacteria (65), or per- method and therefore may be difficult to find. Spirochetes are forming an H&E counterstain on the section already stained not detected with a tissue Gram stain; therefore, other tech- with the Warthin-Starry or Steiner silver stain, a difficult task niques often are used for diagnosis of infections caused by that histologists generally prefer to avoid. these organisms. Both Actinomyces and Nocardia spp. stain with Grocott- Silver impregnation, with various modifications of the War- Gomori’s methenamine silver stain, which most commonly is thin-Starry, Dieterle, or Steiner and Steiner stain, is among the used to enhance visualization of fungi (discussed in a later most sensitive methods for detection of bacteria, including section) (137). As with smears from specimens, these two gen- those that stain weakly with a tissue Gram stain, such as Le- era usually are distinguished in tissue with a modified acid-fast gionella spp. (Fig. 16), and is the only one that allows visual- stain (Fig. 17): Fite’s and Coates’ modified Fite stain perform ization of spirochetes (1, 22, 49, 53, 92, 96, 164, 182, 196, 203). equally well, but Actinomyces spp. often stain weakly with Silver impregnation, however, is not specific, staining virtually Putt’s modification of the Fite stain, so that this variation all bacteria, spirochetes, and fungi, and it has the pitfalls as- probably should not be used for this purpose (82). Nocardia sociated with any silver-based stain, most of which relate to the spp., however, are not invariably acid fast, and the filaments of intensity of development. With overexposure, bacteria may some Streptomyces spp. are weakly acid fast, so that culture is appear too dark and there may be a black precipitate on the critical for diagnosis. Legionella micdadei also appears acid fast tissue, making visualization of the bacteria difficult; alterna- in tissue, a characteristic most frequently demonstrated with tively, with underexposure, bacteria often appear light brown Kinyoun and Fite stains and less often with Ziehl-Neelsen stain or yellow. Of the silver impregnation methods, the Warthin- (133, 152, 170, 196). For detection of all species of Legionella, Starry method appears to be the most sensitive, but it is also the Wolbach modification of Giemsa stain appears to provide the most difficult to duplicate from day to day. Only acid-clean consistently good staining of the bacterial cells (59). Specific glassware can be used, and strict attention to pH is critical. The antibodies for immunofluorescence staining are useful for tis- Dieterle stain has many steps and requires the use of pyridine, sue diagnosis of legionella pneumonia and for detection of T. which has an unpleasant odor, but the results generally are pallidum (84, 85). consistent. The Steiner and Steiner method has fewer steps than the Dieterle method but appears to be less sensitive than STAINS FOR DETECTION OF MYCOBACTERIA the Warthin-Starry method for detection of certain bacteria. Various other special stains have proven useful for detection Over the past few decades, the number of mycobacterial of selected bacteria in tissue sections. H. pylori may be dem- infections in the United States has risen, in part because of the onstrated with modified Giemsa, acridine orange, Wright-Gi- resurgence of tuberculosis and in part because of the increas- emsa, or toluidine O, but none appear to be as sensitive as the ing numbers of cases of disseminated Mycobacterium avium Warthin-Starry stain (77, 110, 114, 144, 159, 190). Moreover, complex disease in patients with AIDS. Given the public health none of these stains adequately allows simultaneous detection importance of control of tuberculosis, much emphasis recently of H. pylori bacilli and histopathologic evaluation of salient has been placed on use of the most sensitive stain for detection features of the gastric mucosa. The latter problem can be of acid-fast bacilli (AFB) and rapid reporting of smear results overcome by staining a single tissue section with Steiner, H&E, (172). 394 WOODS AND WALKER CLIN.MICROBIOL.REV.

TABLE 5. Common infectious causes of nonorganizing mixed acute and chronic or chronic inflammation

Organism Tissue stainsa Organism appearance Commentsa Legionella species Warthin-Starry, Dieterle, or Steiner Pleomorphic, tapered and centrally Infiltrate usually is mixture of PMNs (pneumonia) and Steiner; Ϯ Brown-Hopps; constricted bacilli (0.3–0.7 by and macrophages, fibrin almost Wolbach modification of Giemsa; 2–4 ␮m), predominantly intra- always is present; frequent and Kinyoun or Fiteb; specific anti- cellular, in PMNs and macro- characteristic feature is leukocyto- bodies for IFA or DFA phages; rare filamentous forms; clasis of cellular exudate appear as short, blunt rods with Dieterle stain (or other silver impregnation method) (Fig. 16)

Helicobacter pylori Warthin-Starry, Dieterle, or Steiner Curved bacteria in mucus layer Gastric biopsy specimens show acute and Steiner; Giemsa; toluidine (Fig. 14) gastritis consisting of glandular O; acridine orange; Ϯ H&E; , epithelial cell degenera- “triple stain” (Steiner, H&E, and tion, and intestinal ; alcian blue) PMNs, eosinophils, lymphocytes, and PC in lamina propria

Cryptosporidium spp. H&E, Fite Rows or clusters of basophilic Small intestine is main target (may spherical structures (2–4 ␮min infect biliary tree, pancreas, gall- diameter) attached to microvil- bladder, and respiratory tract); lous border of epithelial cells tissue architecture is often normal (Fig. 30) with light infection but is dis- torted with heavy infection (muco- sal flattening with an infiltrate of lymphocytes, PC, and occasionally PMNs in the lamina propria)

Microsporidia H&E, Brown-Brenn or Brown- Organisms and round to oval Sites of involvement same as for Hopps, Warthin-Starry, modified spores (2–5 ␮m in diameter) cryptosporidia; tissue architecture trichrome located in cytoplasm of entero- often normal with light infection; cytes (or other ciliated cells) be- with severe infection, villi are tween microvillous border and shortened and blunted, epithelial nucleus (Fig. 29) cells are necrotic, and the lamina propria contains lymphocytes and occasional PMNs

Treponema pallidum Warthin-Starry, Dieterle, or Steiner Thin, spiral bacilli Characteristic features include oblit- and Steiner erative endarteritis (concentric endothelial and fibroblastic prolif- erative thickening) and PC infiltrate

a PC, plasma cells; IFA, indirect immunofluorescence assay; DFA, direct immunofluorescence assay; Ϯ, may or may not stain with this stain. b Legionella micdadei positive.

Direct Detection in Smears stain; therefore, bacilli often appear “gram neutral” or as Fluid specimens submitted for detection of mycobacteria “gram ghosts” (56, 80, 175), although they may stain gram should be concentrated (after decontamination if nonsterile, positively in smears prepared from clinical specimens contain- such as sputum) before portions are used to prepare a smear ing large numbers of mycobacteria. for staining and to inoculate media for culture (31). Tradition- Three staining procedures are most commonly used for vi- ally, specimens are concentrated by centrifugation (with aero- sualization of mycobacteria in smears. Two of these, Ziehl- sol-free sealed centrifuge cups) for at least 15 min at Ն3,000 ϫ Neelsen and Kinyoun, use carbol fuchsin as the primary stain g (149). Recently, Saceanu et al. reported that for sputum (13). The Ziehl-Neelsen stain requires heat; the , specimens from a selected group of patients, the sensitivity of in which the concentrations of carbol fuchsin and phenol are AFB with the Kinyoun stain was greatly improved increased, does not. The third method uses the fluorochrome by using cytocentrifugation for smear preparation (154). This dye auramine O with or without rhodamine (12, 116, 177). finding, however, was not confirmed in a more extensive evalua- Smears stained with carbol fuchsin are examined by light mi- tion in which auramine O was used for visualization of AFB (201). croscopy under oil immersion, whereas those stained with au- The Gram stain typically is not used for visualization of ramine O are examined by fluorescence microscopy at lower mycobacteria in smear preparations, because the large amount magnifications (ϫ150 and ϫ450). AFB are red in smears of in their cell walls renders them impermeable to the stained with carbol fuchsin; they fluoresce yellow-green against dyes used in the procedure, giving the cells a variable appear- a black background when stained with auramine O and orange- ance. To the astute microbiologist, however, this feature actu- yellow when auramine and rhodamine are used. Cells typically ally may provide a clue to their presence. Decades ago, when appear as slightly curved bacilli (1 to 10 ␮m long and 0.2 to 0.6 gentian violet was used in the Gram stain, mycobacteria often ␮m wide) that often are beaded or banded, but coccoid or appeared as slender, poorly stained, beaded gram-positive ba- filamentous forms, which could potentially cause confusion cilli (80). In general, mycobacteria do not take up crystal violet with Nocardia spp., are occasionally seen. In general, the ap- or safranin, which today are customarily used in the Gram pearance of AFB does not provide a species identification; VOL. 9, 1996 CYTOLOGIC AND HISTOLOGIC STAINS FOR PATHOGEN DETECTION 395

TABLE 6. Infectious agents causing histiocyte aggregates/diffuse infiltrates

Organism Tissue stains Organism appearancea Comments Mycobacterium avium Auramine-rhodamine, Yellow fluorescent (with fluorochrome) Bacilli stain positive with PAS and complexb Kinyoun, Ziehl- or red (with carbol fuchsin) beaded GMS Neelsen, PAS, GMS bacilli

Mycobacterium genavense As above As above Bacilli are PAS negative

Mycobacterium leprae Fite Red, beaded bacilli Lepromatous leprosy (Fig. 20)

Leishmania spp. H&E, Giemsa Ovoid amastigotes (2–4 ␮m in diame- Associated with disseminated cutaneous ter) within macrophages and visceral leishmaniasis (poorly formed granulomas in liver); may re- semble histoplasmosis, but amastig- otes have a kinetoplast and are PAS and GMS negative

Listeria monocytogenes Brown-Brenn, Brown- Short GPB Most commonly seen in fatal neonatal Hopps listeriosis

a GPB, gram-positive bacilli. b In patients with AIDS. however, cells of certain species have features that may be Detection in Tissue Sections useful diagnostically. For example, cells of Mycobacterium kan- sasii often appear as cross-barred bacilli, larger than M. tuber- As with all infectious agents, tissue diagnosis of a mycobac- culosis and resembling a “shepherd’s crook” (Fig. 18), and cells terial infection begins by examination of H&E-stained tissue of M. avium complex may be more coccobacillary than those of sections. Although organisms cannot be seen, the pattern of M. tuberculosis. inflammation provides the first indication that mycobacteria Advantages of auramine O over carbol fuchsin are the ability should be considered among the differential diagnoses (typical to read more fields in less time and the increased sensitivity, findings include a variation of granulomatous inflammation, as which may be even further enhanced by staining at 37ЊC rather described in Tables 2, 3, 6, and 7 and illustrated in Fig. 19 and than conventional staining at room temperature (118). Given 20), thus triggering further evaluation with special stains for these features, experts at the Centers for Disease Control and AFB. However, in certain patient groups, particularly those Prevention now recommend a fluorochrome stain for detection with AIDS, the typical response may not develop, and the of AFB in smears, and they suggest reporting results of smears diagnosis must be suspected on the basis of clinical presentation. from respiratory specimens within 24 h of receipt in the labo- Stains used most commonly for detection of AFB in tissue ratory (172). A potential disadvantage of auramine O, how- sections are Ziehl-Neelsen (Fig. 19), Kinyoun, and auramine O ever, is that cells of the Mycobacterium fortuitum-chelonae com- (with or without rhodamine) (100, 116). As with smears, fluo- plex occasionally stain poorly. Therefore, when the latter rochrome stains are more sensitive (21, 95) and positively organisms are suspected (for example, in post-surgical wound fluorescing bacilli are much easier to detect. Demonstration of infections) but no organisms are seen, staining over the smear Mycobacterium leprae bacilli in tissue sections stained with with carbol fuchsin is recommended. Moreover, to the inexpe- Ziehl-Neelsen stain, however, is difficult; therefore, various rienced observer, fluorescent debris may be misinterpreted as modifications of this stain have been developed for diagnosis of bacilli. Positive fluorochrome smears may be overstained to leprosy. Many years ago, the Putt modification was used, but save the slides for future reference (31). currently the Fite stain (Fig. 20), which includes peanut or

FIG. 12. Paraffin-embedded section from a draining sinus tract of a patient FIG. 13. Paraffin-embedded section of a lesion of peliosis showing a granular with abdominal actinomycosis, showing a characteristic granule with peripheral microcolony of bacteria. The section was stained with H&E. Magnification, clubs. The section was stained with H&E. Magnification, ϫ200. ϫ200. 396 WOODS AND WALKER CLIN.MICROBIOL.REV.

Although it is important to recognize the utility of the Gram stain in detection of fungi, other techniques generally are pre- ferred for this purpose. The KOH preparation is used primar- ily for visualizing unstained fungal elements in skin, hair, and nails and in specimens, such as sputum and vaginal secretions, that contain large amounts of cellular material, because KOH dissolves and much of the other interfering background protein-rich debris. Background artifacts, however, may not be completely eliminated by KOH, sometimes making interpreta- tion of the preparation difficult. To overcome this problem, many laboratories have substituted calcofluor white for KOH or use a combination of the two reagents. Calcofluor white is a fluorescent brightener or whitening agent that binds to cellu- lose and chitin in fungal cell walls and fluoresces with a blue- white color when exposed to UV radiation. Aqueous solutions of calcofluor white have an absorption spectrum of 300 to 412 FIG. 14. Paraffin-embedded section of a gastric biopsy specimen showing bacteria, consistent with Helicobacter pylori, in the layer of mucus on the crypt nm; therefore, microscopes fitted with selective filters for ex- epithelium. The Steiner-hematoxylin and eosin- was used (65). citation of fluorescein cannot be used. Moreover, a mercury Magnification, ϫ200. vapor lamp rather than a quartz halogen bulb is recommended, because the low energy output of the latter is not suitable for calcofluor white fluorescence. Fungal elements may be de- mineral oil in the deparaffinizing solution (the Ziehl-Neelsen tected in Papanicolaou-stained smears (89), but the combina- stain does not), is preferred (146). In addition to stains specif- tion of this stain with calcofluor white is more effective than is ically for detection of AFB, M. avium complex bacilli stain the Papanicolaou stain alone for detection of fungi (127). positively with periodic acid-Schiff (PAS; a stain typically used Fungi also will be detected in smears stained with acridine for detection of fungi, discussed in the next section), which is orange (33, 34), but this technique has not been widely unique among the mycobacteria and, therefore, a useful diag- adopted in the mycology laboratory. nostic feature. Mycobacteria occasionally are visualized as Certain stains have a specific niche for diagnosis of fungal gram-positive bacilli in sections stained by the Brown-Hopps infections. For example, examination of smears of bone mar- or Brown-Brenn method. Commercial polyclonal antibodies row or peripheral blood stained by the Wright or Giemsa against Mycobacterium bovis ( Calmette-Gue´rin strain) technique frequently provides the first indication of dissemi- and Mycobacterium duvalii may be used to detect mycobacteria nated histoplasmosis. India ink or nigrosin is used for detection in tissue by immunohistochemical techniques. Immunohisto- of cells of Cryptococcus neoformans in CSF. Although the use chemistry appears to be more sensitive than the Kinyoun and of India ink for diagnosis of cryptococcal meningitis has been Fite stains overall, and for M. tuberculosis the rate of detection replaced by the more sensitive antigen detection tests (latex of organisms in tissue is similar to that of combined Kinyoun, agglutination or enzyme-immunoassay) in most laboratories, it Fite, and fluorochrome stains (195). Available antibodies, how- still has value for rapid diagnosis of disease in patients with ever, are not specific, staining several fungi as well as myco- AIDS and for use in emergency situations. bacteria (194). Moreover, antibodies are much more expensive For diagnosis of P. carinii pneumonia, smears of induced than Kinyoun, Fite, and fluorochrome stains; therefore, limit- sputum or bronchoalveolar lavage fluid or imprints of lung ing immunohistochemistry studies to cases in which the acid- tissue may be stained by several different techniques, some fast stains are negative and the clinical history is very sugges- useful for detection of the organisms and others useful for tive of mycobacterial disease is reasonable. detection of the cyst walls (93, 161). The Giemsa technique (or the Diff-Quik modification) stains free trophozoites, which STAINS FOR DETECTION OF FUNGI have a central red nucleus and pale blue cytoplasm, and the organisms within cysts, but not the cyst walls. This is a rapid Direct Detection in Smears Several stains or reagents allow direct detection of fungi in clinical specimens. With the Gram stain, yeast cells typically appear gram positive. A notable exception is Cryptococcus neo- formans, whose cells often appear gram negative, occasionally as pale lavender to red, round globules resembling fat droplets, or show a symmetrical arrangement of gram-positive granular inclusions (Fig. 21A) (18). Moreover, cryptococcal yeast cells often are surrounded by an amorphous orange-staining mate- rial. These staining characteristics actually are useful diagnos- tically, especially when seen in smears of CSF. In addition to C. neoformans, fungal hyphae may stain poorly with the Gram stain. The technique is not useful for the diagnosis of histoplas- mosis, and it is considerably less sensitive than other stains for detection of Pneumocystis carinii (discussed later in this sec- tion). Unstained cysts of P. carinii, however, may be recognized in impression smears of lung tissue stained by the Gram FIG. 16. Paraffin-embedded section of lung from a patient who died from method, providing a clue to the diagnosis more rapidly than Legionnaires’ disease, showing numerous silver-staining bacilli scattered the more commonly used techniques (54). throughout. The Dieterle stain was used. Magnification, ϫ315. VOL. 9, 1996 CYTOLOGIC AND HISTOLOGIC STAINS FOR PATHOGEN DETECTION 397

TABLE 7. Common infectious causes of caseous granulomasa

Organism Tissue stains Organism appearance Comments

Mycobacterium tubercu- Auramine-rhodamine, Yellow fluorescent (with fluorochrome) or Polyclonal antibodies available; granulomas may losis Kinyoun, Ziehl-Neelsen red (with carbol fuchsin) beaded bacilli not develop in immunocompromised hosts, espe- cially those with AIDS

Nontuberculous myco- As above As above M. kansasii usually larger than M. tuberculosis (re- bacteriab sembles “shepherd’s crook”), with broad bands (Fig. 18)

Histoplasma capsulatum GMS, PAS Round to oval yeast cells (2–5 ␮m) with Resembles Torulopsis glabrata, Penicillium marnef- single buds fei, and Leishmania spp. (Table 6)

Coccidioides immitis H&E, GMS, PAS Spherules (20–200 ␮m) with or without Developing spherules may resemble Blastomyces internal endospores (2–5 ␮m) (Fig. 24) dermatitidis (Table 2)

Cryptococcus neofor- H&E, GMS, PAS, mucicar- Variably sized yeast cells (2–15 ␮m) with Reaction typical of pulmonary infection caused by mans mine, Fontana-Masson single buds (Fig. 21) capsule-deficient strains in immunocompetent hosts; immunodeficient hosts infected with an encapsulated strain typically show minimal or no inflammatory response

a Other organisms that may cause caseating granulomas are Blastomyces dermatitidis, Sporothrix schenckii (systemic disease), , Aspergillus, and rarely Nocardia spp. b Most commonly M. kansasii, M. avium complex (immune competent host), and M. scrofulaceum. and inexpensive method for diagnosis but appears to be less were detected as darker areas of nonfluorescence, with the sensitive than nonspecific cyst walls stains and immunofluores- mycelial walls appearing black and the central areas of the cence staining with monoclonal antibodies; therefore, it may be hyphae appearing dark green to brown. most valuable as a screening test, with confirmation of negative Because not all fungi are easily recognized in H&E-stained smears by a more sensitive assay (38, 199). The Papanicolaou sections, especially when present in small numbers within the stain is useful for detection of the diagnostic foamy alveolar lesion, special stains are subsequently performed to enhance casts that accompany cysts (156). Nonspecific stains that dem- visualization of the organisms and, occasionally, allow identi- onstrate the cyst walls include Gram-Weigert, cresyl echt vio- fication based on a characteristic morphology (reviewed for let, toluidine blue O (70, 197), calcofluor white (167), and commonly encountered in Tables 2, 4, and 7 and Grocott-Gomori methenamine silver (GMS; including rapid illustrated in Fig. 21 to 27). Frequently, however, identification microwave modifications) (2, 70, 161, 167, 197). Immunofluo- based on morphology is not possible, and so culture is neces- rescence with specific monoclonal antibodies appears to be the sary. For example, dichotomously branching, thin, septate hy- most sensitive diagnostic technique (2, 38, 136, 167, 199) and phae do not always indicate Aspergillus spp.; other fungi (for requires the shortest time for microscopic examination. More- example, Pseudallescheria boydii or Fusarium species) have a over, because antibodies are specific, immunofluorescence similar appearance in tissue (Fig. 22). Assigning the definitive staining is especially useful in situations in which differentia- diagnosis of aspergillosis requires the presence of fruiting heads, tion of P. carinii cysts from nonbudding yeasts is difficult. which, except in oxygen-rich cavities, is an infrequent finding. Some special stains allow detection of virtually all fungi. Of Detection in Tissue Sections these, GMS and the Hotchkiss-McManus PAS stains, each of which was designed as a histochemical test for glycogen and As with direct smears, several stains are useful in the tissue diagnosis of fungal infections (120). The H&E stain demon- strates the pattern of inflammation (summarized in Tables 2, 4, and 7), which, in conjunction with the clinical information, is key in formulating a differential diagnosis, and it allows detec- tion of some fungi, especially aspergilli, zygomycetes, and many yeasts. In tissue sections stained with H&E, fungal hy- phae are best visualized by lowering the substage condenser so that hyphal walls refract the unfocused light. Moreover, for fungi that demonstrate autofluorescence when the H&E- stained tissue is examined under UV illumination, rapid diag- nosis often is possible without the need for additional special stains. In one study, autofluorescence was most valuable for identification of Coccidioides immitis, Candida species, and Aspergillus species in sections from solid parenchymatous or- gans, loose , granulomas, or necrotic tissue debris (71). Fungal elements fluoresced bright green to yellow- green against the olive-green to yellow-brown of the tissue background. Neither Histoplasma capsulatum (one specimen) FIG. 22. Paraffin-embedded section of skin lesion showing a “granule” com- nor the zygomycetes (five specimens) included in that study posed of septate hyphae. Cultures of the tissue grew Pseudallescheria boydii. were autofluorescent, but in the cases of zygomycosis, hyphae GMS stain was used. Magnification, ϫ200. 398 WOODS AND WALKER CLIN.MICROBIOL.REV.

FIG. 23. Paraffin-embedded section of lung from a patient with AIDS. Intra- FIG. 25. Paraffin-embedded section of a verrucous skin lesion showing spher- alveolar cysts of Pneumocystis carinii (arrow) are evident. GMS stain was used. ical yeast cells with thick walls, and one demonstrates a single broad-based bud Magnification, ϫ200. consistent with Blastomyces dermatitidis (culture confirmed the diagnosis). The section was stained with GMS. Magnification, ϫ200. mucin, probably are the most widely used (75, 120). The first chemical process in each method is oxidation of adjacent hy- the yeast’s capsule), i.e., Mayer’s mucicarmine (Fig. 21B) droxyl groups of cell wall to aldehydes. With (probably the most popular), alcian blue, colloidal , and GMS, aldehyde oxidation products reduce methenamine silver the Fontana-Masson stain (Fig. 21C), which apparently oxi- nitrate to metallic silver, and fungal cell walls appear brown to dizes the -like pigment of the yeast as it reduces silver black against the pale-green background of the counterstain. (101). The Fontana-Masson stain is particularly useful for dis- With the PAS stain, aldehydes combine with the Schiff reagent; tinguishing capsule-deficient C. neoformans from H. capsula- fungi appear red-purple against a green counterstained back- tum and Blastomyces dermatitidis (151). Neither type of stain, ground. GMS provides the best contrast of the special fungal however, is absolutely specific for C. neoformans. Rhinosporid- stains, but it has pitfalls: the natural color of cell walls is ium seeberi stains positively with capsular techniques, but its masked, silver deposited on erythrocytes and naked nuclei can morphology (endospores within large spherules) should not be mimic yeasts, and the potential for overstaining of background confused with that of C. neoformans. Other issues are poten- material exists. In many laboratories, the microwave tially more problematic. Cell walls of the yeast form of B. modification of the GMS is preferred (23). Advantages include dermatitidis may stain weakly with mucicarmine (and the other speed and lower risk of the section washing off the slide; capsular stains). With Fontana-Masson, Cryptococcus laurentii however, the microwave must have a digital pad, allowing a also stains positively and yeast cells of Sporothrix schenckii cutoff time of seconds, and its performance with the stain must show faint brown pigment; however, because C. laurentii is a be thoroughly evaluated before the procedure is instituted. rare human pathogen and because yeast cells of S. schenckii Stains that may be substituted for GMS or PAS are Gridley’s are uncommonly seen in tissue and typically are cigar-shaped stain, the principle of which is similar to PAS, and several rather than round like C. neoformans, in practice, neither fluorescent stains, including calcofluor white, , Uvi- should present a diagnostic problem. However, when individ- tex 2B, and perhaps Tinopal CBS-X (72, 74, 94, 128, 160, 183). ual stains do not provide a definitive diagnosis, combining the A few special stains are useful in identification of selected Fontana-Masson stain with a capsular stain provides distinctive fungi. Techniques that are relatively specific for C. neoformans staining of C. neoformans not observed in other fungi (105). include those that stain acid mucopolysaccharides (present in The Fontana-Masson stain also allows detection of dematia- ceous fungi when the dematiaceous nature of their cell walls is not evident. The Ziehl-Neelsen stain, although typically used for detec- tion of mycobacteria (discussed above), is helpful for identifi- cation of B. dermatitidis when only small forms of the yeast, which are difficult to distinguish from H. capsulatum, are present. Persistent acid fastness of the organisms after treat- ment of sections with 1 N hydrochloric acid at 60ЊC for 15 min identifies the yeast as B. dermatitidis (184). Antibodies against P. carinii are commercially available for use in immunohistochemical assays (2, 102). Given that this method of diagnosis is more costly than nonspecific special stains, which in most cases will detect organisms or cysts (Fig. 23), limiting the use of immunohistology or immunocytology to specific situations is reasonable. Antibodies are most helpful in the diagnosis of P. carinii pneumonia when there is an atypical host response, such as granulomas or hyaline membranes, and when small numbers of organisms are present and in exami- FIG. 24. Paraffin-embedded section of a bone marrow biopsy specimen shows spherules of Coccidioides immitis (culture confirmed the diagnosis). The nation of tissues from extrapulmonary sites in cases of sus- section was stained with PAS. Magnification, ϫ200. pected disseminated disease. VOL. 9, 1996 CYTOLOGIC AND HISTOLOGIC STAINS FOR PATHOGEN DETECTION 399

mercial reagents also includes an antibody against Giardia lam- blia (62). Cysts of C. cayetanensis are visible in direct wet mounts viewed under UV illumination, appearing as autofluo- rescent neon-blue circles (111). Stains for detection of micro- sporidia are the Weber modification of trichrome (involves a 10-fold-higher content of chromotrope 2R than that of the trichrome generally used for stools and a much longer staining time), a further modified trichrome-blue stain, calcofluor white, and Uvitex 2B (43, 153, 180, 192). Specimens other than stool submitted for detection of par- asites generally are processed differently, but there are excep- tions. The direct wet mount is a rapid technique for detection of in vaginal discharge; its major prob- lem is low sensitivity. Other techniques for detection of T. vaginalis include the KOH preparation (described in the pre- vious section) and various stains. For example, trichomonads FIG. 26. Paraffin-embedded section of lung showing broad, nonseptate rib- bon-like fungal hyphae, consistent with a zygomycete. GMS stain was used. may be visualized in Papanicolaou-stained smears of cells from Magnification, ϫ200. the uterine cervix, but this method for diagnosis of trichomo- niasis is associated with both false-positive and false-negative results (142). Diff-Quik is more sensitive than the wet mount STAINS FOR DETECTION OF PARASITES for detection of trichomonads, but acridine orange appears to be the most reliable technique (61, 109). Direct Detection in Smears The direct wet mount also is a useful first step in the exam- Various types of direct preparations, depending on the spec- ination of duodenal aspirates, sputum samples, skin snips, and imen source, are useful for diagnosis of a parasitic infection. small pieces of muscle; and if parasites are detected, the spec- The specimen most frequently submitted for detection of par- imen is fixed and stained to confirm the diagnosis. For duode- asites is stool, and the usual request is examination for ova and nal aspirates and sputum, the sample is stained with trichrome parasites (O&P). The first step in this process, although not a or iron hematoxylin; skin snips are fixed in absolute methyl stain, is a direct wet mount, used primarily for visualization of alcohol and stained with a Giemsa stain. Giemsa staining is motile protozoan trophozoites. Next, Lugol’s or D’Antoni’s recommended for detection of many other parasites, including iodine or lactophenol cotton blue is added to enhance the blood pathogens ( species, species, Trypan- visibility of internal structures of trophozoites, cysts, and ova osoma species, Leishmania donovani, and microfilariae) in that might be present (141). Preparation of a permanently thick and thin blood films; Leishmania donovani in bone mar- stained smear is the essential final step in the typical O&P row aspirates and Leishmania species in aspirates or imprints evaluation. The stains used most frequently are the trichrome of cutaneous or mucocutaneous lesions; trypanosomes in technique of Wheatly or iron hematoxylin (174, 193). Accept- lymph node imprints; in imprints of brain, able alternatives to these methods are a rapid modification of lung, or other tissues; and G. lamblia (Fig. 28), microsporidia, the trichrome stain (57), a trichrome technique in which and Entamoeba histolytica in imprints of gastrointestinal biopsy Hemo-De is substituted for xylene (135), and a modification of specimens (40, 42). iron hematoxylin that incorporates a carbol fuchsin step, thus allowing detection of acid-fast organisms (discussed later in this section) (139). Most problems encountered with these Detection in Tissue Sections staining techniques occur because the stool specimen is too old, the smears are too dense, or fixation is inadequate. Im- The H&E-stained section has two purposes in the tissue mature cysts fix more easily than do old ones, and Entamoeba diagnosis of parasitic infections. First, it allows detection of coli cysts require a longer fixation time than do those of other most parasites or their eggs (examples are shown in Fig. 29 to species. Not all parasites are readily detected by the steps included in the usual O&P examination of stool. For optimal detection of Cryptosporidium parvum, Cyclospora cayetanensis, and micro- sporidia, other techniques must be used. In most clinical mi- crobiology laboratories, it is not cost effective to look for these organisms in all specimens submitted for “O&P examination”; each must be requested specifically. Alternatively, a laboratory director, in conjunction with Infectious Disease clinicians, may choose to evaluate all stool specimens from patients with AIDS who have chronic diarrhea for these organisms, because the likelihood of such infections is high in this population. Fluo- rescent stains such as auramine-rhodamine or acridine orange are useful screening techniques for detection of C. parvum, Isospora belli (which also will be detected by trichrome or iron hematoxylin), and C. cayetanensis, but modified Kinyoun or Ziehl-Neelsen stain is recommended for confirmation (24, 111, 113). For detection of cryptosporidia, immunofluorescence staining with monoclonal antibodies is more sensitive than the FIG. 28. Imprint of a duodenal biopsy specimen showing Giardia lamblia nonspecific staining techniques (4, 62), and one of the com- trophozoites. Giemsa stain was used. Magnification, ϫ200. 400 WOODS AND WALKER CLIN.MICROBIOL.REV.

FIG. 29. Paraffin-embedded section of a small bowel biopsy specimen from a FIG. 31. Paraffin-embedded section of brain from a patient with AIDS show- patient with AIDS showing clusters of round structures (spores) in the enterocyte ing a cyst (arrow) and nearby cluster of tachyzoites (arrowhead) of Toxoplasma cytoplasm (arrows), consistent with the diagnosis of microsporidiosis. The gondii. The section was stained with H&E. Magnification, ϫ200. Brown-Brenn Gram stain was used. Magnification, ϫ200.

which often resemble host macrophages and ganglion cells in 32). Second, for those that are not readily visualized, the pat- H&E-stained sections. By combining PAS with diastase diges- tern of inflammation, in conjunction with the clinical history, tion on a contiguous section, identification is even more secure may provide a clue to their presence and thus initiate a more because glycogen is the reactant in the amoebae. PAS also careful search and further evaluation using special stains. For helps distinguish intracellular yeasts of H. capsulatum (PAS example, microsporidia and cryptosporidia are extremely small positive) from Leishmania spp. (PAS negative). GMS, PAS, and therefore easily overlooked in H&E-stained sections (Ta- and calcofluor white stain Acanthamoeba cysts, thus facilitating ble 5) (63, 98, 138, 148, 150, 192). Tissue Gram stains (Fig. their differentiation from tissue macrophages, which they fre- 29) (Brown-Brenn produces much better enterocyte cytoplas- quently resemble in H&E-stained sections. Immunohistology, mic background contrast for microsporidia spores than does using commercial antibodies against T. gondii, may be useful Brown-Hopps) probably are the most useful for rapid and for diagnosis in situations in which organisms are difficult to reliable diagnosis of microsporidiosis; spores appear dark blue recognize, such as in H&E-stained sections of necrotic lesions or red (they often are gram variable), in sharp contrast to the in brain tissue (Fig. 31 and 32). Acid-fast techniques stain faint yellow-brown background, but nonspore stages cannot be hooklets of Echinococcus granulosus and schistosome eggs. In seen (192). Microsporidia also can be detected by silver im- the case of helminths, which generally are readily detected in pregnation techniques (Warthin-Starry stain), which allow vi- H&E-stained sections, special stains that demonstrate the ex- sualization of spores and intermediate stages, the Chromo- oskeleton, such as Weigert’s iron hematoxylin or Russell-Mo- trope-2R modified trichrome stain, and Giemsa stain, which vat pentachrome, aid in identification by enhancing visualiza- stains sporozoites (66, 192). With regard to cryptosporidia, the tion of the characteristic morphologic features. Fite stain may be useful; organisms, which lose their acid-fast quality in histologic preparations appear bright blue (Fig. 30). CONCLUSIONS Special stains also provide a means to differentiate parasites from inflammatory cells and serve as an aid in identification. In the current climate of requirements for greater efficiency PAS is useful for identifying trophozoites of E. histolytica, and decreased costs of patient care, new paradigms for infec- tious disease diagnosis are needed. Too often in the past, clinician, pathologist, and microbiologist have worked inde-

FIG. 30. Paraffin-embedded section of a small bowel biopsy specimen from a patient with AIDS showing clusters of spherical structures attached to the mi- crovillus border of the epithelial cells and free in the adjacent lumen, consistent FIG. 32. Paraffin-embedded section of brain from a patient with AIDS show- with the diagnosis of . The section was stained with the Fite ing numerous Acanthamoeba cysts (arrow) and trophozoites (arrowhead) around stain. Magnification, ϫ200. a blood vessel. H&E stain was used. Magnification, ϫ200. VOL. 9, 1996 CYTOLOGIC AND HISTOLOGIC STAINS FOR PATHOGEN DETECTION 401 pendently toward contributing their skills to the diagnosis of bacilli in histologic sections. Am. J. Clin. Pathol. 36:37–40. infectious diseases. Lack of communication, incomplete clini- 22. Bridges, C. H., and L. Luna. 1957. Kerr’s improved Warthin-Starry technic. Study of the permissible variations. Lab. Invest. 6:357–367. cal information accompanying patient samples submitted to 23. Brinn, N. T. 1983. Rapid metallic histological staining using the microwave the clinical microbiology and laboratories, oven. J. Histotechnol. 6:125–129. and failure of the microbiologist and anatomic pathologist to 24. Bronsdon, M. A. 1984. Rapid -modified acid-fast stain of pursue further information have led to delayed or missed di- Cryptosporidium oocysts in stool specimens. J. Clin. Microbiol. 19:952–953. 25. Brouqui, P., J. S. Dumler, and D. Raoult. 1994. Immunohistologic demon- agnoses and prolonged illness. Optimally, each diagnostic pro- stration of Coxiella burnetii in the valves of patients with Q fever endocar- fessional would know the patient’s clinical signs, symptoms, ditis. Am. J. Med. 97:451–458. course of illness, and working differential diagnosis, and the 26. Brown, R. C., and H. C. Hopps. 1973. Staining of bacteria in tissue sections: most cost-effective tools would be applied in a logical manner. a reliable gram stain method. Am. J. Clin. Pathol. 60:234–240. 27. Bryce, E. A., R. L. Barteluk, and M. A. Noble. 1991. Community-acquired For some situations, immediate molecular diagnosis by PCR or Legionella micdadei pneumonia presenting as acid fast bacilli in sputum. other contemporary method will provide the best cost-benefit Clin. Microbiol. Newsl. 13:95–96. outcome. In other situations, a particular set of cultures, cyto- 28. Butts, J. D., T. W. Bouldin, and D. H. Walker. 1984. Morphological char- logic and histologic examinations, or a combination of these acteristics of a unique intracytoplasmic neuronal inclusion body. Acta Neu- ropathol. 62:345–347. tests will achieve the diagnosis, successful treatment, and ear- 29. Casas-Cordero, M., C. Morin, M. Roy, M. Fortier, and A. Meisels. 1981. liest recovery, discharge from the hospital, and return to work. Origin of the koilocyte in condylomata of the human cervix. Ultrastructural Presently, we do not know how to achieve the goal of cost- study. Acta Cytol. 25:383–392. effective diagnosis under many conditions. The clinical micro- 30. Centers for Disease Control and Prevention. 1993. Recommendations for the prevention and management of Chlamydia trachomatis infections. Mor- biologist and surgical pathologist working together and under- bid. Mortal. Weekly Rep. 42(RR-12):1–39. standing the capabilities, strengths, and weaknesses of the two 31. Cernoch, P. L., R. K. Enns, M. A. Saubolle, and R. J. Wallace, Jr. 1994. approaches could establish guidelines and test them to deter- Cumitech 16A, Laboratory diagnosis of mycobacterioses. Coordinating ed., mine their effect on patient outcome as well as on the true A. S. Weissfeld. American Society for Microbiology, Washington, D.C. 32. Chapin-Robertson, K., S. E. Dahlberg, and S. E. Edberg. 1992. Clinical and costs, including professional and technical manpower effort, laboratory analyses of cytospin-prepared gram stains for recovery and di- supplies, and reduced or prolonged hospitalization. agnosis of bacteria from sterile body fluids. J. Clin. Microbiol. 30:377–380. 33. Chick, E. W. 1961. Acridine orange fluorescent stain for fungi. Arch. Der- REFERENCES matol. 83:305–309. 1. Aberer, E., and P. H. Duray. 1991. Morphology of Borrelia burgdorferi: 34. Clark, R. F., and M. E. Hench. 1962. Practical application of acridine structural patterns of cultured borreliae in relation to staining methods. orange stain in the demonstration of fungi. Am. J. Clin. Pathol. 37:237–238. J. Clin. Microbiol. 29:764–772. 35. Cles, L. D., K. Bruch, and W. E. Stamm. 1988. Staining characteristics of six 2. Amin, M. B., E. Mezger, and R. J. Zarbo. 1992. Detection of Pneumocystis commercially available monoclonal immunofluorescence reagents for direct carinii. Comparative study of monoclonal antibody and . Am. diagnosis of Chlamydia trachomatis infections. J. Clin. Microbiol. 26:1735– J. Clin. Pathol. 98:13–18. 1737. 3. Arnow, P. M., and P. Gardner. 1983. Usefulness of the gram stain in 36. Coleman, D. V. 1975. The cytodiagnosis of human polyomavirus infection. Legionnaires’ disease. Clin. Microbiol. Newsl. 5:125–127. Acta. Cytol. 19:93–96. 4. Arrowood, M. J., and C. R. Sterling. 1989. Comparison of conventional 37. Cowdry, E. V. 1934. The problem of intranuclear inclusions in virus dis- staining methods and monoclonal antibody-based methods for Cryptospo- eases. Arch. Pathol. 18:527–542. ridium oocyst detection. J. Clin. Microbiol. 27:1490–1495. 38. Cregan, P., A. Yamamoto, A. Lum, T. VanDerHeide, M. MacDonald, and L. 5. Ashton, P. R. 1983. Infectious organisms in cytologic material. Lab. Med. Pulliam. 1990. Comparison of four methods for rapid detection of Pneu- 14:227–233. mocystis carinii in respiratory specimens. J. Clin. Microbiol. 28:2432–2436. 6. Ayre, J. E. 1960. Role of the halo cell in cervical cancerigenesis. A virus 39. Daly, J. A., W. M. Gooch III, and J. M. Matsen. 1985. Evaluation of the manifestation in premalignancy? Obstet. Gynecol. 15:481–491. Wayson variation of a methylene blue staining procedure for the detection 7. Bakken, J. S., J. S. Dumler, S. M. Chen, M. R. Eckman, L. L. Van Etta, and of in cerebrospinal fluid. J. Clin. Microbiol. 21:919–921. D. H. Walker. 1994. Human granulocytic ehrlichiosis in the upper Midwest 40. Debongnie, J. C., C. Beyaert, and G. Legros. 1989. Touch cytology, a useful United States: a new species emerging? JAMA 272:212–218. diagnostic method for diagnosis of upper gastrointestinal tract infections. 8. Barnes, R. C. 1989. Laboratory diagnosis of chlamydial infections. Clin. Dig. Dis. Sci. 34:1025–1027. Microbiol. Rev. 2:119–136. 41. Debongnie, J. C., M. Delmee, P. Mainguet, C. Beyaert, J. Haot, and G. 9. Barr, R. J., J. Herten, and J. H. Graham. 1977. Rapid method for Tzanck Legros. 1992. Cytology: a simple, rapid, sensitive method in the diagnosis of preparations. JAMA 237:1119–1120. Helicobacter pylori. Am. J. Gastroenterol. 87:20–23. 10. Bartholomew, J. W. 1962. Variables influencing results, and the precise 42. Debongnie, J. C., J. Mairesse, M. Donnay, and X. Dekoninck. 1994. Touch definition of steps in gram staining as a means of standardizing the results cytology. A quick, simple, sensitive screening test in the diagnosis of infec- obtained. Stain Technol. 37:139–155. tions of the gastrointestinal mucosa. Arch. Pathol. Lab. Med. 118:1115– 11. Bartholomew, J. W., and T. Mittwer. 1952. The Gram stain. Bacteriol. Rev. 1118. 16:1–29. 43. DeGirolami, P. C., C. R. Ezratty, G. Desal, A. McCullough, D. Asmuth, C. 12. Bennedsen, J., and S. O. Larsen. 1966. Examination for tubercle bacilli by Wanke, and M. Federman. 1995. Diagnosis of intestinal microsporidiosis by fluorescence microscopy. Scand. J. Respir. Dis. 47:114–120. examination of stool and duodenal aspirate with Weber’s modified 13. Bishop, P. J., and G. Neuman. 1970. The history of the Ziehl-Neelsen stain. trichrome and Uvitex 2B stains. J. Clin. Microbiol. 33:805–810. Tubercle (London) 51:196–206. 44. Deodhar, K. P., E. Tapp, and P. J. Scheuer. 1975. Orcein staining of 14. Blank, H., C. F. Burgoon, G. D. Baldridge, P. L. McCarthy, and F. Urbach. hepatitis B antigen in paraffin sections of liver . J. Clin. Pathol. 1951. Cytologic smears in diagnosis of herpes simplex, herpes zoster, and 28:66–70. varicella. JAMA 146:1410–1412. 45. Derakhshan, I. 1975. Is the Negri body specific for rabies? A light and 15. Boeckh, M., P. M. Woogerd, T. Stevens-Ayers, C. G. Ray, and R. A. Bowden. electron microscopical study. Arch. Neurol. 32:75–79. 1994. Factors influencing detection of quantitative cytomegalovirus anti- 46. Derakhshan, I., M. Bahmanyar, A. Fayaz, S. Noorsalehi, M. Mohammad, genemia. J. Clin. Microbiol. 32:832–834. and P. Ahouraii. 1978. Light-microscopical diagnosis of rabies: a reap- 16. Boerner, D. F., and P. Zwadyk. 1982. The value of the sputum gram’s stain praisal. Lancet i:302–303. in community-acquired pneumonia. JAMA 247:642–645. 47. Dominguez, E. A., L. H. Taber, and R. B. Couch. 1993. Comparison of rapid 17. Bossen, E. H., W. W. Johnston, J. Amatulli, and D. T. Rowlands, Jr. 1969. diagnostic techniques for respiratory syncytial and influenza A virus respi- Exfoliative cytopathologic studies in organ transplantation. Am. J. Clin. ratory infections in young children. J. Clin. Microbiol. 31:2286–2290. Pathol. 52:340–344. 48. Drancourt, M., F. George, P. Brouqui, J. Sampol, and D. Raoult. 1992. 18. Bottone, E. J. 1980. Cryptococcus neoformans: pitfalls in diagnosis through Diagnosis of mediterranean spotted fever by indirect immunofluorescence evaluation of Gram-stained smears of purulent exudates. J. Clin. Microbiol. of Rickettsia conorii in circulating endothelial cells isolated with monoclonal 12:790–791. antibody-coated immunomagnetic beads. J. Infect. Dis. 166:660–663. 19. Bottone, E. J. 1988. The gram stain: the century-old quintessential rapid 49. Duray, P. H., A. Kusnitz, and J. Ryan. 1985. Demonstration of the Lyme diagnostic test. Lab. Med. 19:288–291. disease spirochete by a modified Dieterle stain method. Lab. Med. 16:685– 20. Bozsik, B. P. 1985. Borrelia burgdorferi stained with acidic acridine orange. 687. Ann. Immunol. 25:335–344. 50. Edelstein, P. H. 1993. Legionnaires’ disease. Clin. Infect. Dis. 16:741–749. 21. Braunstein, H., and S. M. Adriano. 1961. Fluorescent stain for tubercle 51. Edelstein, P. H., and M. A. C. Edelstein. 1989. Evaluation of the Merifluor- 402 WOODS AND WALKER CLIN.MICROBIOL.REV.

Legionella immunofluorescent reagent for identifying and detecting 21 Le- 80. Hinson, J. M., R. W. Bradsher, and S. J. Bodner. 1981. Gram-stain neu- gionella species. J. Clin. Microbiol. 27:2455–2458. trality of Mycobacterium tuberculosis. Am. Rev. Respir. Dis. 123:365–366. 52. Faine, S., P. Edelstein, B. D. Kirby, and S. M. Finegold. 1979. Rapid 81. Hook, E. W., III, R. E. Roddy, S. A. Lukehart, J. Hom, and K. K. Holmes. presumptive bacteriological diagnosis of Legionnaires disease. J. Clin. Mi- 1985. Detection of Treponema pallidum in lesion exudate with a pathogen- crobiol. 10:104–105. specific monoclonal antibody. J. Clin. Microbiol. 22:241–244. 53. Farrier, R., and A. S. Warthin. 1930. A study of the effect of PH upon the 82. Hotchi, M., and J. Schwarz. 1972. Characterization of actinomycotic gran- third improved Warthin-Starry method for demonstrating Spirocheta pal- ules by architecture and staining methods. Arch. Pathol. 93:392–400. lida in single sections. Am. J. Syph. 14:394–401. 83. Humberstone, F. D. 1963. A gram-type stain for the simultaneous demon- 54. Felegie, T. P., A. W. Pasculle, and A. Dekker. 1984. Recognition of Pneu- stration of gram-positive and gram-negative microorganisms in paraffin mocystis carinii by Gram stain in impression smears of lung tissue. J. Clin. sections. J. Med. Lab. Technol. 20:153–156. Microbiol. 20:1190–1191. 84. Hunter, E. F., P. W. Greer, B. L. Swisher, A. R. Simons, C. E. Farshy, J. A. 55. Fine, M. J., J. J. Orloff, J. D. Rihs, R. M. Vickers, S. Kominos, W. N. Crawford, and K. R. Sulzer. 1984. Immunofluorescent staining of Trepo- Kapoor, V. C. Arena, and V. L. Yu. 1991. Evaluation of housestaff physi- nema in tissues fixed with formalin. Arch. Pathol. Lab. Med. 108:878–880. cians’ preparation and interpretation of sputum gram stains for community 85. Ito, F., E. F. Hunter, R. W. George, V. Pope, and S. A. Larsen. 1992. Specific acquired pneumonia. J. Gen. Intern. Med. 6:189–198. immunofluorescent staining of pathogenic treponemes with a monoclonal 56. Fisher, J. F., M. Ganapathy, B. H. Edwards, and C. L. Newman. 1990. antibody. J. Clin. Microbiol. 30:831–838. Utility of Gram’s and Giemsa stains in the diagnosis of pulmonary tuber- 86. Jethwa, H. S., J. L. Schmitz, G. Dallabetta, F. Behets, I. Hoffman, H. culosis. Am. Rev. Respir. Dis. 141:511–513. Hamilton, G. Lule, M. Cohen, and J. D. Folds. 1995. Comparison of mo- 57. Flournoy, D. J., S. J. N. McNabb, E. D. Dodd, and M. H. Shaffer. 1982. lecular and microscopic techniques for detection of Treponema pallidum in Rapid trichrome stain. J. Clin. Microbiol. 16:573–574. genital ulcers. J. Clin. Microbiol. 33:180–183. 58. Forsum, U., and A. Hallen. 1978. Acridine orange staining of urethral and 87. Johnson, M. J., E. Thatcher, and M. E. Cox. 1995. Techniques for control- cervical smears for the diagnosis of . Acta Dermato-venereol. ling variability in gram staining of obligate anaerobes. J. Clin. Microbiol. (Stockholm) 59:281–282. 33:755–758. 59. Frenkel, J. K., L. H. Baker, and A. M. Chonko. 1979. Autopsy diagnosis of 88. Johnston, S. L. G., and C. S. Siegel. 1990. Evaluation of direct immuno- Legionnaires’ disease in immunosuppressed patients. A paleodiagnosis us- fluorescence, enzyme immunoassay, centrifugation culture, and conven- ing Giemsa stain (Wohlbach modification). Ann. Intern. Med. 90:559–562. tional culture for the detection of respiratory syncytial virus. J. Clin. Mi- 60. Friedman, R. L. 1988. Pertussis: the disease and new diagnostic methods. crobiol. 28:2394–2397. Clin. Microbiol. Rev. 1:365–376. 89. Johnston, W. W. 1981. in opportunistic respiratory infection. 61. Fripp, P. J., P. R. Mason, and H. Super. 1975. A method for the diagnosis Lab. Manage. 19:43–49. of Trichomonas vaginalis using acridine orange. J. Parasitol. 61:966–967. 90. Kass, E. M., W. K. Szaniawski, H. Levy, J. Leach, K. Srinivasan, and C. 62. Garcia, L. S., A. C. Shum, and D. A. Bruckner. 1992. Evaluation of a new Rives. 1994. Rickettsialpox in a New York City hospital, 1980 to 1989. monoclonal antibody combination reagent for direct fluorescence detection N. Engl. J. Med. 331:1612–1617. of Giardia cysts and Cryptosporidium oocysts in human fecal specimens. 91. Kee, J. C., A. Hill, and K. C. Chapin. 1994. Use of an enhanced Gram stain J. Clin. Microbiol. 30:3255–3257. method for detection of organisms from body fluids and blood culture 63. Genta, R. M., C. L. Chappell, A. C. White Jr., K. T. Kimball, and R. W. broths, abstr. C-242, p. 533. In Abstracts of the 94th General Meeting of the Goodgame. 1993. Duodenal morphology and intensity of infection in AIDS- American Society for Microbiology 1994. American Society for Microbiol- related intestinal cryptosporidiosis. Gastroenterology 105:1769–1775. ogy, Washington, D.C. 64. Genta, R. M., G. M. Lew, and D. Y. Graham. 1993. Changes in the gastric 92. Kerr, D. A. 1938. Improved Warthin-Starry method of staining spirochetes mucosa following eradication of Helicobacter pylori. Mod. Pathol. 6:281– in tissue sections. Am. J. Clin. Pathol. 8:63–67. 289. 93. Kim, H., and W. T. Hughes. 1972. Comparison of methods for identification 65. Genta, R. M., G. O. Robason, and D. Y. Graham. 1994. Simultaneous of Pneumocystis carinii in pulmonary aspirates. Am. J. Clin. Pathol. 60:462–466. visualization of Helicobacter pylori and gastric morphology: a new stain. 94. Koch, H. H., and M. Pimsler. 1987. Evaluation of Uvitex 2B: a nonspecific Hum. Pathol. 25:221–226. fluorescent stain for detecting and identifying fungi and algae in tissue. Lab. 66. Giang, T. T., D. P. Kotler, M. L. Garro, and J. M. Orenstein. 1993. Tissue Med. 18:603–606. diagnosis of intestinal microsporidiosis using the chromotrope-2R modified 95. Kommaredii, S., C. R. Abramowsky, G. L. Swinehart, and L. Hrabak. 1984. trichrome stain. Arch. Pathol. Lab. Med. 117:1249–1251. Nontuberculous mycobacterial infections: comparison of the fluorescent 67. Gimenez, D. F. 1964. Staining rickettsiae in yolk-sac cultures. Stain Technol. auramine-O and Ziehl-Neelsen techniques in tissue diagnosis. Hum. 39:135–140. Pathol. 15:1085–1089. 68. Gleaves, C. A., C. F. Lee, C. I. Bustamante, and J. D. Meyers. 1988. Use of 96. Korbi, S., M.-F. Toccanier, G. Leyvraz, J. Stalder, and Y. Kapanci. 1986. murine monoclonal antibodies for laboratory diagnosis of varicella-zoster Use of silver staining (Dieterle’s stain) in the diagnosis of cat scratch virus infection. J. Clin. Microbiol. 26:1623–1625. disease. 10:1015–1021. 69. Goldman, L., R. M. McCabe, and F. Sawyer. 1960. The importance of 97. Koss, L. G., and G. R. Durfee. 1956. Unusual patterns of squamous epi- cytology technic for the dermatologist in office practice. Arch. Dermatol. thelium of the uterine cervix: cytologic and pathologic study of koilocytotic 81:359–368. atypia. Ann. N. Y. Acad. Sci. 63:1245–1261. 70. Gosey, L. L., R. M. Howard, F. G. Witebsky, F. P. Ognibene, T. C. Wu, V. J. 98. Kotler, D. P., T. T. Giang, M. L. Garro, and J. M. Orenstein. 1994. Light Gill, and J. D. MacLowry. 1985. Advantages of a modified toluidine blue O microscopic diagnosis of microsporidiosis in patients with AIDS. Am. J. stain and bronchoalveolar lavage for the diagnosis of Pneumocystis carinii Gastroenterol. 89:540–544. pneumonia. J. Clin. Microbiol. 22:803–807. 99. Kronvall, G., and E. Myhre. 1977. of bacteria in clinical 71. Graham, A. R. 1983. Fungal autofluorescence with ultraviolet illumination. specimens using acridine orange buffered at low pH. Acta Pathol. Micro- Am. J. Clin. Pathol. 79:231–234. biol. Scand. 85:249–254. 72. Green, L. K., and D. G. Moore. 1987. Fluorescent compounds that nonspe- 100. Kuper, S. W. A., and J. R. May. 1960. Detection of acid-fast organisms in cifically stain fungi. Lab. Med. 18:456–458. tissue sections by fluorescence microscopy. J. Pathol. Bacteriol. 79:59–68. 73. Greer, P. W., F. W. Chandler, and M. D. Hicklin. 1980. Rapid demonstra- 101. Kwon-Chung, K. J., W. B. Hill, and J. E. Bennett. 1981. New, special stain tion of Legionella pneumophila in unembedded tissue. An adaptation of the for histopathological diagnosis of cryptococcosis. J. Clin. Microbiol. 13:383– Gimenez stain. Am. J. Clin. Pathol. 73:788–790. 387. 74. Gridley, M. F. 1953. A stain for fungi in tissue sections. Am. J. Clin. Pathol. 102. Lachman, M. F., R. W. Cartun, C. A. Pedersen, and S. R. Cole. 1990. 23:303–307. Immunocytochemical identification of Pneumocystis carinii in formalin- 75. Grocott, R. G. 1955. A stain for fungi in tissue sections and smears using fixed, paraffin-embedded tissues. Lab. Med. 21:808–810. Gomori’s methenamine- technic. Am. J. Clin. Pathol. 25:975– 103. Larson, A. M., M. J. Dougherty, D. J. Nowowiejski, D. F. Welch, G. M. 979. Matar, B. Swaminathan, and M. B. Coyle. 1994. Detection of Bartonella 76. Gubetta, L., M. Rizzetto, O. Crivelli, G. Verme, and S. Arico. 1977. A (Rochalimaea) quintana by routine acridine orange staining of broth blood trichrome stain for the intrahepatic localization of the hepatitis B surface cultures. J. Clin. Microbiol. 32:1492–1496. antigen (HBsAg). Histopathology 1:277–288. 104. Lauer, B. A., L. B. Reller, and S. Mirrett. 1981. Comparison of acridine 77. Guglielmetti, P., N. Figura, A. Rossolini, S. Quaranta, E. Fanteria, R. orange and Gram stains for detection of microorganisms in cerebrospinal Signori, and E. Camarri. 1991. The usefulness of the acridine-orange stain fluid and other clinical specimens. J. Clin. Microbiol. 14:201–205. in identifying Helicobacter pylori in gastric biopsies. Microbiologica 14:131– 105. Lazcano, O., V. O. Speights, Jr., J. G. Strickler, J. E. Bilbao, J. Becker, and 134. J. Diaz. 1993. Combined histochemical stains in the differential diagnosis of 78. Gupta, P. K. 1982. Intrauterine contraceptive devices. , Cryptococcus neoformans. Mod. Pathol. 6:80–84. pathologic changes and clinical implications. Acta Cytol. 26:571–613. 106. Lendrum, A. C. 1947. The phloxine-tartrazine method as a general histo- 79. Hilton, E., R. A. Freedman, F. Cintron, H. D. Isenberg, and C. Singer. 1986. logical stain for demonstration of inclusion bodies. J. Pathol. Bacteriol. Acid-fast bacilli in sputum: a case of Legionella micdadei pneumonia. 59:399–404. J. Clin. Microbiol. 24:1102–1103. 107. Lennette, E. H., and N. J. Schmidt. 1979. Diagnostic procedures for viral, VOL. 9, 1996 CYTOLOGIC AND HISTOLOGIC STAINS FOR PATHOGEN DETECTION 403

rickettsial and chlamydial infections, 5th ed. American Public Health As- 136. Ng, V. L., N. A. Virani, R. E. Chaisson, D. M. Yajko, H. T. Sphar, K. sociation, Washington, D.C. Cabrian, N. Rollins, P. Charache, M. Krieger, W. K. Hadley, and P. C. 108. Leonardi, G. P., H. Leib, G. S. Birkhead, C. Smith, P. Costello, and W. Hopewell. 1990. Rapid detection of Pneumocystis carinii using a direct Conron. 1994. Comparison of rapid detection methods for influenza A virus fluorescent monoclonal antibody stain. J. Clin. Microbiol. 28:2228–2233. and their value in health care management of institutionalized geriatric 137. Oddo, D., and S. Gonzalez. 1986. Actinomycosis and nocardiosis: a mor- patients. J. Clin. Microbiol. 32:70–74. phologic study of 17 cases. Pathol. Res. Pract. 181:320–326. 109. Levett, P. N. 1980. A comparison of five methods for the detection of 138. Orenstein, J. M., J. Chiang, W. Steinberg, P. D. Smith, H. Rotterdam, and Trichomonas vaginalis in clinical specimens. Med. Lab. Sci. 37:85–88. D. P. Kotler. 1990. Intestinal microsporidiosis as a cause of diarrhea in 110. Loffeld, R. J. L. F., E. Stobberingh, J. A. Flendrig, and J. W. Arends. 1991. human immunodeficiency virus-infected patients: a report of 20 cases. Helicobacter pylori in gastric biopsy specimens. Comparison of culture, Hum. Pathol. 21:475–481. modified Giemsa stain, and immunohistochemistry. A retrospective study. 139. Palmer, J. 1991. Modified iron hematoxylin/Kinyoun stain. Clin. Microbiol. J. Clin. Pathol. 165:69–73. Newsl. 13:39–40. 111. Long, E. G., E. H. White, W. W. Carmichael, P. M. Quinlisk, R. Raja, B. L. 140. Parham, D. M., P. Bozeman, C. Killian, G. Murti, M. Brenner, and I. Swisher, H. Daugharty, and M. T. Cohen. 1991. Morphologic and staining Hanif. 1993. Cytologic diagnosis of respiratory syncytial virus infection in a characteristics of a cyanobacterium-like organism associated with diarrhea. bronchoalveolar lavage specimen from a bone marrow transplant recipient. J. Infect. Dis. 164:199–202. Am. J. Clin. Pathol. 99:588–592. 112. Lynch, P. J., O. Welsh, and B. Naylor. 1970. Cytodiagnosis of vesicular viral 141. Parija, S. C., and P. K. Prabhakar. 1995. Evaluation of lacto-phenol cotton lesions—modified Papanicolaou stain for herpesvirus infections. Cutis blue for wet mount preparation of feces. J. Clin. Microbiol. 33:1019–1021. 6:79–84. 142. Perl, G. 1972. Errors in the diagnosis of Trichomonas vaginalis infection as 113. MacPherson, D. W., and R. McQueen. 1993. Cryptosporidiosis: multiat- observed among 1199 patients. Obstet. Gynecol. 39:7–9. tribute evaluation of six diagnostic methods. J. Clin. Microbiol. 31:198–202. 143. Pezzlo, M. 1988. Detection of urinary tract infections by rapid methods. 114. Madan, E., J. Kemp, T. U. Westblom, M. Subik, S. Sexton, and J. Cook. Clin. Microbiol. Rev. 1:268–280. 1988. Evaluation of staining methods for identifying Campylobacter pylori. 144. Potters, H. V. P. J., R. J. L. F. Loffeld, E. Stobberingh, J. P. van Spreeuwel, Am. J. Clin. Pathol. 90:450–453. and J. W. Arends. 1987. Rapid staining of Campylobacter pyloridis. Histo- 115. Mangels, J. I., M. E. Cox, and L. H. Lindberg. 1984. Methanol fixation. An pathology 11:1223–1226. alternative to heat fixation of smears before staining. Diagn. Microbiol. 145. Provine, H., and P. Gardner. 1974. The Gram-stained smear and its inter- Infect. Dis. 2:129–137. pretation. Hosp. Pract. 9:85–91. 116. Mansfield, R. E. 1970. An improved method for the fluorochrome staining 146. Putt, F. A. 1951. A modified Ziehl-Neelsen method for demonstration of of mycobacteria in tissues and smears. Am. J. Clin. Pathol. 53:394–406. leprosy bacilli and other acid-fast organisms. Am. J. Clin. Pathol. 21:92–95. 117. Matthey, S., D. Nicholson, S. Ruhs, B. Alden, M. Knock, K. Schultz, and A. 147. Rabalais, G., G. Stout, and S. Waldeyer. 1992. Rapid detection of influen- Schmuecker. 1992. Rapid detection of respiratory viruses by shell vial za-B virus in respiratory secretions by immunofluorescence during an epi- culture and direct staining by using pooled and individual monoclonal demic. Diagn. Microbiol. Infect. Dis. 15:35–37. antibodies. J. Clin. Microbiol. 30:540–544. 148. Rabeneck, L., R. M. Genta, F. Gyorkey, J. E. Clarridge, P. Gyorkey, and 118. McCarter, Y. S., and A. Robinson. 1994. Detection of acid-fast bacilli in L. W. Foote. 1995. Observations on the pathological spectrum and clinical concentrated primary specimen smears stained with rhodamine-auramine course of microsporidiosis in men infected with the human immunodefi- at room temperature and at 37ЊC. J. Clin. Microbiol. 32:2487–2489. ciency virus: follow-up study. Clin. Infect. Dis. 20:1229–1235. 119. McCracken, A. W. 1980. Microscopic clues to viral and chlamydial diseases. 149. Rickman, T. W., and N. P. Moyer. 1980. Increased sensitivity of acid-fast Lab. Med. 11:152–158. smears. J. Clin. Microbiol. 11:618–620. 120. McGinnis, M. R. 1992. Diagnosing fungal infections using histopathology. 150. Rijpstra, A. C., E. U. Canning, R. J. Van Ketel, J. K. M. E. Schattenker, and Clin. Adv. Treat. Fungal Infect. 3:11–16. J. J. Laarman. 1988. Use of light microscopy to diagnose small-intestinal 121. Meisels, A., and R. Fortin. 1976. Condylomatous lesions of the cervix and microsporidiosis in patients with AIDS. J. Infect. Dis. 157:827–831. vagina. I. Cytologic patterns. Acta Cytol. 20:505–509. 151. Ro, J. A., S. S. Lee, and A. G. Ayala. 1987. Advantage of Fontana-Masson 122. Meisels, A., R. Fortin, and M. Roy. 1977. Condylomatous lesions of the stain in capsule-deficient cryptococcal infection. Arch. Pathol. Lab. Med. cervix. II. Cytologic, colposcopic and histopathologic study. Acta Cytol. 111:53–57. 21:379–390. 152. Rogers, B. H., G. R. Donowitz, G. K. Walker, S. A. Harding, and M. A. 123. Miller-Catchpole, R., D. Variakojis, J. W. Vardiman, J. M. Loew, and J. Sande. 1979. Opportunistic pneumonia. A clinicopathological study of five Carter. 1986. Cat scratch disease. Identification of bacteria in seven cases of cases caused by an unidentified acid-fast bacterium. N. Engl. J. Med. 301: lymphadenitis. Am. J. Surg. Pathol. 10:276–281. 959–961. 124. Minnich, L. L., F. Goodenough, and C. G. Ray. 1991. Use of immunoflu- 153. Ryan, N. J., G. Sutherland, K. Coughlan, M. Globan, J. Doultree, J. Mar- orescence to identify measles virus infections. J. Clin. Microbiol. 29:1148– shall, R. W. Baird, J. Pedersen, and B. Dwyer. 1993. A new trichrome-blue 1150. stain for detection of microsporidial species in urine, stool, and nasopha- 125. Minnich, L., and C. G. Ray. 1980. Comparison of direct immunofluorescent ryngeal specimens. J. Clin. Microbiol. 31:3264–3269. staining of clinical specimens for respiratory virus antigens with conven- 154. Saceanu, C. A., N. C. Pfeiffer, and T. McLean. 1993. Evaluation of sputum tional isolation techniques. J. Clin. Microbiol. 12:391–394. smears concentrated by cytocentrifugation for detection of acid-fast bacilli. 126. Mirrett, S., B. A. Lauer, G. A. Miller, and L. B. Reller. 1982. Comparison J. Clin. Microbiol. 31:2371–2374. of acridine orange, methylene blue, and Gram stains for blood cultures. 155. Sadick, N. S., P. D. Swenson, R. L. Kaufman, and M. H. Kaplan. 1987. J. Clin. Microbiol. 15:562–566. Comparison of detection of varicella-zoster virus by the Tzanck smear, 127. Monheit, J. E., G. Brown, M. M. Kott, W. A. Schmidt, and D. G. Moore. direct immunofluorescence with a monoclonal antibody, and virus isolation. 1986. Calcofluor white detection of fungi in cytopathology. Am. J. Clin. J. Am. Acad. Dermatol. 17:64–69. Pathol. 85:222–225. 156. Schumann, G. B., and J. J. Swensen. 1991. Comparison of Papanicolaou’s 128. Monheit, J. E., D. F. Cowan, and D. G. Moore. 1984. Rapid detection of stain with the Gomori methenamine silver (GMS) stain for the cytodiag- fungi in tissues using calcofluor white and fluorescence microscopy. Arch. nosis of Pneumocystis carinii in bronchoalveolar lavage (BAL) fluid. Am. Pathol. Lab. Med. 108:616–618. J. Clin. Pathol. 95:583–586. 129. Montgomery, E. A., D. F. Martin, and D. A. Peura. 1988. Rapid diagnosis 157. Schwarz, T. F., A. Nerlich, B. Hottentrager, G. Jager, I. Wiest, S. Kantimm, of Campylobacter pylori by Gram’s stain. Am. J. Clin. Pathol. 90:606–609. H. Roggendorf, M. Schultz, K.-P. Gloning, T. Schramm, W. Holzgreve, and 130. Morris, A. J., D. C. Tanner, and L. B. Reller. 1993. Rejection criteria for M. Roggendorf. 1991. Parvovirus B19 infection of the fetus. Histology and endotracheal aspirates from adults. J. Clin. Microbiol. 31:1027–1029. in situ hybridization. Am. J. Clin. Pathol. 96:121–126. 131. Moseley, R. C., L. Corey, D. Benjamin, C. Winter, and M. L. Remington. 158. Shanholtzer, C. J., P. J. Schaper, and L. R. Peterson. 1982. Concentrated 1981. Comparison of viral isolation, direct immunofluorescence, and indi- Gram stain smears prepared with a cytospin centrifuge. J. Clin. Microbiol. rect immunoperoxidase techniques for detection of simplex 16:1052–1056. virus infection. J. Clin. Microbiol. 13:913–918. 159. Slater, B. 1990. Superior stain for Helicobacter pylori using toluidine 132. Murray, P. R., and J. A. Washington II. 1975. Microscopic and bacterio- O. J. Clin. Pathol. 43:961. logic analysis of expectorated sputum. Mayo Clin. Proc. 50:339–344. 160. Slifkin, M., and R. Cumbie. 1988. Congo red as a fluorochrome for the 133. Myerowitz, R. L., A. W. Pasculle, J. N. Dowling, G. J. Pazin, M. Puerzer, rapid detection of fungi. J. Clin. Microbiol. 26:827–830. R. B. Yee, C. R. Rinaldo, and T. R. Hakala. 1979. Opportunistic lung 161. Smith, J. W., and M. S. Bartlett. 1982. Laboratory diagnosis of Pneumo- infection due to “Pittsburgh pneumonia agent.” N. Engl. J. Med. 301:953–958. cystis carinii infection. Clin. Lab. Med. 2:393–406. 134. Nahass, G. T., B. A. Goldstein, W. Y. Zhu, U. Serfling, N. S. Penneys, and 162. Solomon, A. R., J. E. Rasmussen, J. Varani, and C. L. Pierson. 1984. The C. L. Leonardi. 1992. Comparison of Tzanck smear, , and DNA Tzanck smear in the diagnosis of cutaneous herpes simplex. JAMA 251: diagnostic methods in detection of herpes simplex and varicella-zoster 633–635. infection. JAMA 268:2541–2544. 163. Solomon, A. R., J. E. Rasmussen, and J. S. Weiss. 1986. A comparison of 135. Neimeister, R., A. L. Logan, and J. H. Egleton. 1985. Modified trichrome the Tzanck smear and viral isolation in varicella and herpes zoster. Arch. staining technique with a xylene substitute. J. Clin. Microbiol. 22:306–307. Dermatol. 122:282–285. 404 WOODS AND WALKER CLIN.MICROBIOL.REV.

164. Steiner, G., and G. Steiner. 1944. New simple silver stain for demonstration 185. Walker, D. H., M. S. Burday, and J. D. Folds. 1980. Laboratory diagnosis of of bacteria, spirochetes, and fungi in sections from paraffin-embedded tis- Rocky Mountain spotted fever. South. Med. J. 73:1443–1447. sue blocks. J. Clin. Lab. Med. 29:868–871. 186. Walker, D. H., and B. G. Cain. 1978. A method for specific diagnosis of 165. Stevens, D. A. 1983. Clinical and clinical laboratory aspects of nocardial Rocky Mountain spotted fever on fixed, paraffin-embedded tissue by im- infection. J. Hyg. 91:377–384. munofluorescence. J. Infect. Dis. 137:206–209. 166. Stirk, P. R., and P. D. Griffiths. 1988. Comparative sensitivity of three 187. Walker, D. H., and J. H. S. Gear. 1985. Correlation of the distribution of methods for the diagnosis of cytomegalovirus lung infection. J. Virol. Meth- Rickettsia conorii, microscopic lesions, and clinical features in South African ods 20:133–142. tick bite fever. Am. J. Trop. Med. Hyg. 34:361–371. 167. Stratton, N., J. Hryniewicki, S. L. Aarnaes, G. Tan, L. M. DeLa Maza, and 188. Walker, D. H., C. Occhino, G. R. Tringali, S. DiRosa, and S. Mansucto. E. M. Peterson. 1991. Comparison of monoclonal antibody and calcofluor 1988. of rickettsial eschars: the tache noire of Boutonneuse white stains for the detection of Pneumocystis carinii from respiratory spec- fever. Hum. Pathol. 19:1449–1454. imens. J. Clin. Microbiol. 29:645–647. 189. Walker, D. H., F. M. Parks, T. G. Betz, J. P. Taylor, and J. W. Muehlberger. 168. Strickler, J. G., J. C. Manivel, C. M. Copenhaver, and V. L. Kubic. 1990. 1989. Histopathology and immunohistologic demonstration of the distribu- Comparison of in situ hybridization and immunohistochemistry for detec- tion of Rickettsia typhi in fatal murine typhus. Am. J. Clin. Pathol. 91:720–724. tion of cytomegalovirus and herpes simplex virus. Hum. Pathol. 21:443–448. 190. Walters, L. L., R. E. Budin, and G. Paull. 1986. Acridine-orange to identify 169. Takimoto, S., M. Grandien, M. A. Ishida, M. S. Pereira, T. M. Paiva, T. Campylobacter pyloridis in formalin fixed, paraffin-embedded gastric biop- Ishimaru, E. M. Makita, and C. H. O. Martinez. 1991. Comparison of sies. Lancet i:42. enzyme-linked immunosorbent assay, indirect immunofluorescence assay, 191. Washington, J. A. 1979. Use and abuse of the gram-stained smear. Clin. and virus isolation for detection of respiratory viruses in nasopharyngeal Microbiol. Newsl. 1:4–5. secretions. J. Clin. Microbiol. 29:470–474. 192. Weber, R., R. T. Bryan, D. A. Schwartz, and R. L. Owen. 1994. Human 170. Talamo, T. S., A. M. Norbut, and G. F. Kessler. 1980. Opportunistic pneu- microsporidial infections. Clin. Microbiol. Rev. 7:426–461. monia caused by a new acid-fast bacterium. Am. J. Clin. Pathol. 74:842–845. 193. Wheatley, W. B. 1951. A rapid staining procedure for intestinal amoebae 171. Tanaka, K., W. Mori, and K. Suwa. 1981. Victoria blue-nuclear fast red and flagellates. Am. J. Clin. Pathol. 21:990–991. stain for HBs antigen detection in paraffin section. Acta. Pathol. Jpn. 194. Wiley, E. L., B. Beck, and R. G. Freeman. 1991. Reactivity of fungal 31:93–98. organisms in tissue sections using anti-mycobacteria antibodies. J. Cutane- 172. Tenover, F. C., J. T. Crawford, R. E. Huebner, L. J. Geiter, C. R. Hors- ous Pathol. 18:204–209. burgh, Jr., and R. C. Good. 1993. The resurgence of tuberculosis: is your 195. Wiley, E. L., T. J. Mulhollan, B. Beck, J. Tyndall, and R. G. Freeman. 1990. laboratory ready? J. Clin. Microbiol. 31:767–770. Polyclonal antibodies raised against Bacillus Calmette-Guerin, Mycobacte- 173. Thomas, E. E., and L. E. Book. 1991. Comparison of two rapid methods for rium duvalii, and Mycobacterium paratuberculosis used to detect mycobac- detection of respiratory syncytial virus (RSV) (TestPack RSV and Ortho teria in tissue with the use of immunohistochemical techniques. Am. J. Clin. RSV ELISA) with direct immunofluorescence and virus isolation for the Pathol. 94:307–312. diagnosis of pediatric RSV infection. J. Clin. Microbiol. 29:632–635. 196. Winn, W. C., and R. L. Myerowitz. 1981. The pathology of the Legionella 174. Tompkins, V. N., and J. K. Miller. 1947. Staining intestinal with pneumonias. A review of 74 cases and the literature. Hum. Pathol. 12:401– iron-hematoxylin-. Am. J. Clin. Pathol. 17:755–758. 422. 175. Trifiro, S., A. M. Bourgault, F. Lebel, and P. Rene. 1990. Ghost mycobac- 197. Witebsky, F. G., J. W. B. Andrews, V. J. Gill, and J. D. MacLowry. 1988. teria on Gram stain. J. Clin. Microbiol. 28:146–147. Modified toluidine blue O stain for Pneumocystis carinii: further evaluation 176. Trowell, J. E., A. Young, K. J. Saul, P. W. Gant, and G. D. Bell. 1987. Simple of some technical factors. J. Clin. Microbiol. 26:774–775. half-Gram stain for showing presence of Campylobacter pyloridis in sections. 198. Wolbach, S. B. 1916. The etiology of Rocky Mountain spotted fever. (A J. Clin. Pathol. 40:702. preliminary report.) J. Med. Res. 34:121–128. 177. Truant, J. P., W. A. Brett, and W. Thomas, Jr. 1962. Fluorescence micros- 199. Wolfson, J. S., M. A. Waldron, and L. S. Sierra. 1990. Blinded comparison copy of tubercle bacilli stained with auramine and rhodamine. Henry Ford of a direct immunofluorescent monoclonal antibody staining method and a Hosp. Med. Bull. 10:287–296. Giemsa staining method for identification of Pneumocystis carinii in in- 178. Tzanck, A., and G. R. Melki. 1954. Cyto-diagnosis in dermatology, p. 87– duced sputum and bronchoalveolar lavage specimens of patients infected 102. In R. M. B. Mackenna (ed.), Modern trends in dermatology. Hoeber, with human immunodeficiency virus. J. Clin. Microbiol. 28:2136–2138. New York. 200. Woods, G. L., and J. A. Bryan. 1994. Detection of Chamydia trachomatis by 179. Van der Bij, W., R. Torensma, W. J. vanSon, J. Anema, J. Schirm, A. M. direct fluorescent antibody staining. Results of the College of American Tegzess, and T. H. The. 1988. Rapid immunodiagnosis of active cytomeg- Pathologists Proficiency testing program, 1986–1992. Arch. Pathol. Lab. alovirus infection by monoclonal antibody staining of blood. J. Med. Virol. Med. 118:483–488. 25:179–188. 201. Woods, G. L., E. Pentony, M. J. Boxley, and A. M. Gatson. 1995. Concen- 180. van Gool, T., F. Snijders, P. Reiss, J. K. M. E. Schattenkerk, M. A. van den tration of sputum by cytocentrifugation for preparation of smears for de- Bergh Weerman, J. F. W. M. Bartelsman, J. J. M. Bruins, E. U. Canning, tection of acid-fast bacilli does not increase sensitivity of the fluorochrome and J. Dankert. 1993. Diagnosis of intestinal and disseminated microspo- stain. J. Clin. Microbiol. 33:1915–1916. ridial infections in patients with HIV by a new rapid fluorescence tech- 202. Woods, G. L., A. B. Thompson, S. L. Rennard, and J. Linder. 1990. Detec- nique. J. Clin. Pathol. 46:694–699. tion of cytomegalovirus in bronchoalveolar lavage specimens. Spin ampli- 181. Van Horn, K. G., and B. M. Dworkin. 1990. Direct Gram stain and cation and staining with a monoclonal antibody to the early nuclear antigen test to detect Helicobacter pylori. Diagn. Microbiol. Infect. Dis. 13:449–452. for diagnosis of cytomegalovirus pneumonia. Chest 98:568–575. 182. Van Orden, A. E., and P. W. Greer. 1977. Modification of the Dieterle 203. Young, B. J. 1969. A reliable method for demonstrating spirochaetes in spirochete stain. J. Histotechnol. 1:51–53. tissue sections. J. Med. Lab. Technol. 26:248–252. 183. Wachsmuth, E. D. 1988. A comparison of the highly selective fluorescence 204. Yu, X., P. Brouqui, J. S. Dumler, and D. Raoult. 1993. Detection of Ehrli- staining of fungi in tissue sections with Uvitex 2B and calcofluor white chia chaffeensis in human tissue by using a species-specific monoclonal M2R. Histochem. J. 20:215–221. antibody. J. Clin. Microbiol. 31:3284–3288. 184. Wages, D. S., and D. J. Wear. 1982. Acid-fastness of fungi in blastomycosis 205. Zaitoun, A. M. 1992. Use of Romanowsky type (Diff-3) stain for detecting and histoplasmosis. Arch. Pathol. Lab. Med. 106:440–441. Helicobacter pylori in smears and tissue sections. J. Clin. Pathol. 45:448–449.