Fine-needle aspiration of histoplasmosis in the era of endoscopic ultrasound and endobronchial ultrasound

Cytomorphologic features and correlation with clinical laboratory testing




Histoplasmosis has a textbook cytologic description with numerous intracellular organisms that are readily apparent on routine stains. This is based on series and reports describing histoplasmosis in immunosuppressed patients with disseminated disease. With the advent of ultrasound-guided (US) fine-needle aspiration (FNA) techniques, a marked increase in the cytologic diagnosis of histoplasmosis in immunocompetent patients is noted.


A search identified all cytology cases diagnosed with Histoplasma within the past 10 years. Cases were reviewed, along with patient demographic, clinical, and laboratory data.


A total of 40 FNA cases of histoplasmosis were identified. Patients ranged in age from 15 years to 86 years. There were 23 female patients and 17 male patients; 37 were immunocompetent and 3 were immunosuppressed. Sixteen patients were being staged for primary tumors of other sites; others presented with primary pulmonary symptoms or histoplasmosis was noted incidentally. Specimens were composed of bland acellular necrosis, most commonly with granulomas (77.5%); only rare intracellular organisms were present on routine stains, and variable extracellular organisms were noted on Grocott methenamine silver stain (GMS) stain. GMS stain on direct smears was found to be more sensitive than cell block. Laboratory studies for urine antigen, yeast, and mycelial antibody (by compliment fixation), serum antibody (by immunodiffusion), and culture were positive in 11.8%, 59.1%, 4.5%, 47.6%, and 3.4% of cases, respectively.


In an endemic region, histoplasmosis presents more commonly in immunocompetent patients as localized fibrocaseous disease on FNA and is often identified by high-resolution imaging. FNA is increasingly used in the diagnosis because of endoscopic ultrasound and endobronchial ultrasound. GMS stain on direct smears is more sensitive than cell block. In general, laboratory tests have low sensitivity in this patient population. Cancer (Cancer Cytopathol) 2013;121:508-17. © 2013 American Cancer Society.


Histoplasma capsulatum is a dimorphic fungus predominately found in soils enriched with bird and bat excreta. Although worldwide in distribution, it is endemic in a broad area of the central and eastern United States, concentrated along the Mississippi and Ohio River valleys.[1-3] Infection occurs when aerosolized microconidia of the mold are inhaled from the environment.[1, 4] Once in vivo, the microconidia transform to the yeast phase and are found within histiocytes and tissues of the reticuloendothelial system.[3] It is estimated that greater than 80% of young adults residing in endemic regions of the United States have been previously infected with H. capsulatum.[1]

The vast majority of infected immunocompetent persons are either asymptomatic or have a mild illness that is never fully recognized as a fungal infection.[1, 3] However, histoplasmosis may take other forms. Acute pulmonary histoplasmosis is generally a self-limited illness most commonly presenting in young persons and mistaken for acute bronchopneumonia. Patients present with fever, malaise, and headache, and if a chest radiograph is obtained it may show hilar lymphadenopathy.[1, 2] Biopsy is unusual for this form of the disease and unless the inoculum is massive, these patients need no further treatment.[1] Chronic cavitary histoplasmosis typically occurs in older persons and likely represents reactivation of a latent infection.[1] These patients present with constitutional symptoms and productive cough, possibly with hemoptysis, dyspnea, and fibrocavitary lesions noted on radiologic imaging, overlapping with the signs and symptoms noted in chronic obstructive pulmonary disease.[1] It is important to note that these lesions of histoplasmosis can present as positron emission tomography (PET)-positive nodules and can mimic malignancy.[5-7] Histologic examination shows granulomatous inflammation with a fibrocaseous core of necrotic debris in which fungal organisms may be identified, but generally the inflammatory milieu is out of proportion to the organism burden.[2, 8] The disseminated form of histoplasmosis is generally observed in individuals with significant cell-mediated immune deficiency, specifically those with a deficient or less-developed T helper cell 1 (Th1) lymphocyte response, which is partially responsible for granuloma formation.[1] The disseminated form is prototypically described in conjunction with human immunodeficiency virus (HIV) and can represent either primary infection or disease reactivation.[1, 8-11] Hematologic malignancy, solid organ transplant, high-dose corticosteroid administration, and the extremes of age have also all been described as predisposing conditions leading to dissemination.[1, 11] More recently, the advent of biologic drugs, especially tumor necrosis factor-α (TNF-α) antagonists, that modulate important cell-mediated immune factors in the treatment of rheumatic conditions has lead others to report cases of disseminated histoplasmosis in a variety of clinical settings.[12-14] The underlying immune defect in these patients leads to the inability of macrophages to transform and form granulomas, thereby isolating and containing the organism from the rest of the body. As a result, these patients experience asymptomatic hematogenous dissemination throughout the reticuloendothelial system by parasitized macrophages.[1] The histology for this form of the disease typically demonstrates abundant intracellular yeast filling macrophages with little granuloma formation and it is this form that is most commonly depicted in textbooks and articles.

To the best of our knowledge, relatively few recent reports (the majority of which contained very limited numbers of patients) have highlighted the cytomorphological findings of histoplasmosis.[15-20] Many patients described in these studies were immunosuppressed with HIV. These reports paint a portrait of the cytologic features consistent with dissemination of Histoplasma with abundant intracellular organisms evident on routine stains in most if not all cases. However, in our recent practice in an endemic region, immunocompetent patients were found to present more commonly, frequently as a result of increased diagnostic imaging by computed tomography (CT) or PET, including for cancer staging, and they present much more often with localized fibrocaseous disease. Fine-needle aspiration (FNA) is an increasingly important diagnostic test for histoplasmosis. With the advent of ultrasound-guided FNA techniques, including endoscopic ultrasound (EUS) and endobronchial ultrasound modalities, once difficult-to-reach areas of the body are examined and biopsied. In the current study, we share our 10-year experience with FNA as the primary diagnostic test for histoplasmosis in a predominately immunocompetent patient population. We discuss the cytological findings that are significantly different from the characteristic description in significantly immunosuppressed patients. Because many of these patients were suspected to have malignancy, we describe clues that should raise the possibility of a Histoplasma infection, discuss staining methods that may increase the diagnostic yield from FNA material, and review the salient clinical factors and laboratory data that bring these patients to clinical attention and to the presumptive diagnosis.


We searched the cytology files at the University of Iowa Hospitals and Clinics for all cases in which yeast morphologically consistent with Histoplasma was identified by cytology over a 10-year period (2002-2012). All cases for which cytology material was available (smear and/or cell block) were included in the study. We reviewed all available slides on these cases documenting the presence or absence of bland acellular necrosis, granulomas, giant cells, conspicuous neutrophilic infiltrate, intracellular fungal forms, extracellular fungal forms, and identifiable fungi on routine stains (Romanowsky and Papanicolaou [Pap]). We defined bland acellular necrosis as granular, acellular material lacking inflammatory cells. Cell outlines, commonly found in coagulative necrosis associated with malignancy, were also lacking. Histoplasma was presumptively diagnosed when yeast forms, measuring 2 μm to 4 μm in diameter, were identified on Romanowsky stain, Pap stain, or Grocott methenamine silver stain (GMS). We also documented the quantity of yeast on GMS-stained direct smears as rare (0-2 yeast/high-power field [HPF]), few (3-5 yeast/HPF), moderate (5-10 yeast/HPF), many (> 10 yeast/HPF), and too numerous to count. All cases for which a cell block preparation was available underwent GMS staining to identify organisms and compare it with the direct smear method.

Patient demographic data, clinical history, and laboratory data were retrieved from the electronic medical record and the following items were recorded: patient age and sex; immune status; significant comorbidities; clinical presentation; treatment; clinical follow-up; and clinical laboratory testing, including specimen and blood cultures, urine antigen, and serological evaluation for antibodies. We defined the patients' immune status as either immunocompetent or immunosuppressed based on the information obtained from the electronic medical record. Although we agree that several definitions and forms of immunosuppression exist, for practical purposes, we defined immunosuppression in the current study as having a hematologic malignancy; being infected with HIV; receiving concurrent treatment with alkylating agents, antimetabolites, immunomodulators, radiation, or large doses of corticosteroids; and as occurring at the extremes of age. Furthermore, we did not classify patients with a history of malignancy who were not currently receiving treatment or those with diabetes mellitus, renal failure, or alcoholism as being immunosuppressed.


Clinical Features

Forty patients (23 females and 17 males) had a presumptive FNA diagnosis of histoplasmosis during the study period. The clinical features of the patient cohort are summarized in Table 1. Patient age ranged from 15 years to 86 years (mean, 42 years). Thirty-seven patients (92.5%) were considered to be immunocompetent by our study criteria. Only 3 patients (7.5%) were immunosuppressed: 1 patient who was receiving treatment with a TNF-α antagonist for rheumatoid arthritis, 1 who was treated with chemotherapy for lymphoma, and 1 elderly male with a presumed immune deficiency due to age. None were known to be HIV positive. Sixteen patients (40%) were being evaluated or staged for known or suspected primary tumors at other sites, but were not under medical treatment at the time of biopsy. Biopsy sites included the lung (16 patients), lymph nodes (22 patients), adrenal gland (1 patient), and thyroid bed (1 patient). (Table 2) Thirty-nine patients (97.5%) presented with lymphadenopathy: 38 (95%) of the hilar or mediastinal lymph nodes, 1 (2.5%) of cervical lymph nodes, 1 (2.5%) of perihepatic/peripancreatic lymph nodes, 1 (2.5%) of perigastric lymph nodes, and 1 (2.5%) of retroperitoneal lymph nodes. Other sites included 14 patients (35%) with lung nodules or masses, 1 patient (2.5%) with a mass of the thyroid bed, 2 patients (5%) with adrenal masses, and 1 patient (2.5%) with a pancreatic lesion. Ten patients (25%) presented with constitutional symptoms (fever, weight loss, and/or night sweats). Available clinical follow-up ranged from < 1 month to 99 months, with a mean follow-up of 23 months.

Table 1. Clinical and Laboratory Information of Patients Presenting With Histoplasmosis
Patient No.Age, Years/SexSiteImmune statusClinical PresentationTreatmentaHistoplasma Yeast AntibodybHistoplasma Mycelial AntibodybHistoplasma Urine AntigencFollow-Upd
  1. Abbreviations: CT, computed tomography; CXR, chest x-ray; DLBCL, diffuse large B-cell lymphoma; DM, diabetes mellitus; F, female; GCT, germ cell tumor; IC, immunocompetent; IS, immunosuppressed; LAD, lymphadenopathy; M, male; MM, malignant melanoma; NA, not available; PET, positron emission tomography; PTC, papillary thyroid carcinoma; RA, rheumatoid arthritis; SOB, shortness of breath; TNF-α, tumor necrosis factor- α.

  2. a

    Treatment with itraconazole in all “yes” cases.

  3. b

    Histoplasma yeast and mycelial serum antibody by complement fixation (titer).

  4. c

    Histoplasma urine antigen by enzyme immunoassay.

  5. d

    Patient was asymptomatic unless otherwise noted.

150/FLymph nodeICMass on CTNo1:16<1:8NA4 mo, resolving symptoms
236/FLungICPET-positive nodule, history of breast carcinomaYes1:8<1:8Negative10 mo
327/FLymph nodeICMass on CTYes<1:8<1:8NA16 mo, decreased adenopathy
446/MLymph nodeICMass on CTYes<1:8<1:8Negative48 mo, recurrent hemoptysis
540/FLungICMass on CTNoNANANA13 mo
643/FLungICLung nodules, history of uterine carcinomaNoNANANA<1 mo
720/FLymph nodeICMass on CXRNoNANANA24 mo
842/MLungICLung nodule, history of testicular carcinomaNoNANANA22 mo
922/MLungICHilar LAD, history of PTCNoNANANA41 mo
1052/FLymph nodeICFlu-like symptoms, mass on CTNoNANANA89 mo
1182/FLymph nodeICLAD, history of MMYesNANANA92 mo, died of MM
1215/MLymph nodeICRight neck massNo<1:8<1:8Negative38 mo
1337/MLymph nodeICMass on CTNoNANANA3 mo
1438/FLymph nodeICMass on CT, history of pancreatic lesionYes1:16<1:8Negative5 mo, intermittent abdominal pain
1562/FLymph nodeICIncidental nodule on CTNo<1:8<1:8NA38 mo
1622/MLymph nodeICDysphagia, LAD on CTYes1:8<1:8Negative12 mo
1745/FLungICIncidental right hilar mass on CTNoNANANegative52 mo
1834/FLymph nodeICLAD, history of PTCNo1:32<1:8Negative47 mo
1955/FLungICChest pain, SOB, mass on CTNoNANANA17 mo
2061/FLymph nodeICLAD, history of breast carcinomaYesNANANA3 mo
2159/FLungICIncidental lung nodule on CTNoNANANA2 mo
2253/MLymph nodeICPET-positive nodule, history of MMNoNANANA5 mo
2350/MLymph nodeICMass on CTNoNANANA3 mo
2429/MLungICFatigue, LAD on CTYesNANANegative43 mo, continued fatigue
2532/MLungICHiccups, fatigue, mass on CTNo1:1281:16NA3 mo
2650/FLymph nodeICFever, lung nodulesYes1:64<1:8Negative3 mo, clinical improvement
2756/FLungICCough, SOB, mass on CTNo1:16<1:8Negative5 mo
2845/FLungICPulmonary nodules, history of uterine carcinomaNo<1:8<1:8NA10 mo
2925/FLymph nodeICChest discomfort, mass on CXRYes1:16<1:8Negative30 mo
3041/FLymph nodeICSOB, chest pain, LAD on CTNo<1:8<1:8NA29 mo
3115/MLungICLAD, history of mixed GCTNo1:32<1:8NA4 mo
3231/MLungICCough, mass on CTYes<1:8<1:8Negative10 mo, fibrosing mediastinitis
3329/MLymph nodeICPersistent cough and LAD on CTYes1:1024<1:8Negative22 mo, interval increase
3420/MLymph nodeICDysphagia, LAD on CTYes<1:8<1:8NA7 mo, noncompliant, progressive symptoms
3529/MLymph nodeICConstitutional symptoms, mediastinal massYes1:16<1:8NA25 mo, night sweats
3650/FLungICMass on CT/CXR, uncontrolled DMNo1:512<1:8Negative4 mo
3735/FLungICPET-positive lung nodule, history of breast carcinomaNo<1:8<1:8Negative1 mo
3886/MAdrenal glandISAdrenal masses on CT, age-related immune suppressionYesNANAPositive1 mo, improving
3958/MThyroid bedISRA treated with TNF-α antagonist, history of PTCYesNANAPositive37 mo
4037/FLymph nodeISMediastinal LAD, under treatment for DLBCLYesNANANegative99 mo
Table 2. Anatomic Distribution of 40 Cases of Histoplasmosis
Anatomic LocationNo. of Cases
Lymph nodes22
Hilar (15) 
Mediastinal (4) 
Extrapulmonary (3) 
Adrenal gland1
Thyroid bed1

Seventeen patients (42.5%) were treated with itraconazole. Seven patients (17.5%) experienced continued or progressive symptoms, 4 (10%) had improving clinical symptoms, and the remaining 29 patients (72.5%) were asymptomatic at the time of last follow-up; 1 patient died of malignant melanoma during the follow-up period and 1 patient developed sclerosing mediastinitis, a rare complication that has been associated with mediastinal histoplasmosis.[21, 22]

Cytologic Findings

The cytologic material for all cases consisted of Romanowsky-stained and Pap-stained direct smears and GMS-stained direct smears and/or cell blocks. Cell blocks were available for 18 cases. The smear specimens were noted to have white to yellow, creamy to granular necrotic material admixed with various elements of blood. Microscopically, all cases were characterized by bland acellular necrosis (Fig. 1 Top) that did not contain cell outlines or inflammatory cell infiltrates. Thirty-one cases (77.5%) contained granulomas (Fig. 2). Giant cells and a conspicuous neutrophilic inflammatory infiltrate were present in 4 cases (10%) and 3 cases (7.5%), respectively. Only 3 cases (2 from immunocompetent patients and 1 from an immunosuppressed patient) demonstrated intracellular organisms that were recognizable on routine cytologic stains (Figs. 3 and 4). All cases had extracellular fungi within a background of bland acellular necrosis (Fig. 1 Bottom) and the organism burden ranged widely: rare (1 case; 2.5%), few (6 cases; 15%), moderate (11 cases; 27.5%), many (21 cases; 52.5%), and too numerous to count (1; 2.5%). The most common scenario included abundant, bland acellular necrosis with no identifiable organisms on routine stains and variable numbers of organisms on GMS-stained direct smears. The cytologic findings of these cases are summarized in Table 3.

Figure 1.

(Top) Bland acellular necrosis without a conspicuous neutrophilic infiltrate, characteristic of the necrosis observed with Histoplasma, is shown (Romanowsky stain, × 400). (Bottom) Grocott methenamine silver-stained direct smear with several yeast is shown in a background of bland acellular necrosis (× 1000).

Figure 2.

Epithelioid granulomas without identifiable organisms are shown on Romanowsky stain (× 400).

Figure 3.

Rare intracellular yeast within a histiocyte (arrow) in an immunocompetent patient is shown. The concurrent Grocott methenamine silver-stained direct smear for this patient demonstrated many extracellular yeast in a background of necrosis. We note that the identification of Histoplasma is difficult when they occur singly within histiocytes and these were only noted on review (Romanowsky stain, × 1000).

Figure 4.

Numerous intracellular yeast within histiocytes are shown in an immunosuppressed patient (Romanowsky stain, × 1000).

Table 3. Cytomorphology of 40 Cases of Histoplasmosis
  1. Abbreviations: GMS, Grocott methenamine silver stain; TNTC, too numerous to count.

Bland necrosis40/40 (100%)
Granulomas31/40 (77.5%)
Neutrophilic infiltrate3/40 (7.5%)
Giant cells4/40 (10%)
Intracellular fungi3/40 (7.5%)
Quantity: Rare2/40 (5%)
TNTC1/40 (2.5%)
Extracellular fungi40/40 (100%)
Quantity of fungi on GMS-stained direct smearsRare (0-2): 1/40 (2.5%)
 Few (3-5): 6/40 (15%)
 Moderate (5-10): 11/40 (27.5%)
 Many (>10): 21/40 (52.5%)
 TNTC: 1/40 (2.5%)
Positive GMS-stained direct smears37/37 (100%)
Positive GMS-stained cell blocks10/18 (56%)

Thirty-seven cases were presumptively diagnosed through the identification of organisms on GMS-stained direct smears only. A GMS stain was performed on all available cell blocks (18 cell blocks) for comparison of fungal detection between direct smears versus cell block. Histoplasma was identified in 10 of 18 cell blocks (56%). Areas of bland acellular necrosis and a paucity of inflammatory infiltrate were noted on these cell blocks.

Laboratory Tests and Culture Results

Laboratory findings are summarized in Table 4. Urine Histoplasma antigen (Mira Vista Diagnostics, Indianapolis, Ind) was positive in 2 of 17 cases (sensitivity, 11.8%). Histoplasma yeast antibody as identified by complement fixation (ARUP Laboratories, Salt Lake City, Utah) was undetectable (< 1:8) in 9 of 22 cases (sensitivity, 59.1%). The positive antibody titers were: 1:8 (2 cases), 1:16 (5 cases), 1:32 (2 cases), 1:64 (1 case), 1:128 (1 case), 1:512 (1 case), and 1:1024 (1 case). Histoplasma mycelial antibody, also identified by complement fixation (ARUP Laboratories), was undetectable in 21 of 22 cases (sensitivity, 4.5%); the lone positive titer was 1:16. Histoplasma antibody identified by immunodiffusion (ARUP Laboratories) was positive in 10 of 21 cases (47.6%); in all cases, only the M antigen was identified. Twenty-nine fungal cultures, either from aspirate material, bronchoalveolar lavage, tracheal aspirate, bronchial washing, or blood, were sent from 22 patients; 1 of 2 blood cultures (50%) yielded H. capsulatum. All cultures from aspirate material (none of 11 cases), bronchoalveolar lavage (none of 12 cases), tracheal aspirate (none of 3 cases), and bronchial washing (none of 1 case) were negative for growth.

Table 4. Clinical Laboratory Identification of 40 Cases of Histoplasmosis
Serum Histoplasma Yeast Antibody (CF, Titer)aNo. of Cases
  1. Abbreviations: CF, compliment fixation; EIA, enzyme immunoassay; ID, immunodiffusion.

  2. a

    Testing performed at ARUP Laboratories, Salt Lake City, Utah.

  3. b

    Testing performed at Mira Vista Diagnostics, Indianapolis, Indiana.

Serum Histoplasma Mycelial Antibody (CF, Titer)a 
Serum Histoplasma Antibody (ID)a 
Not detected11
Detected (M antigen only)10
Urine Histoplasma Antigen (EIA)b 
Fungal CulturesPositive/Total (%)
Blood1/2 (50%)
Aspirated specimen0/11
Bronchoalveolar lavage0/12
Tracheal aspirate0/3
Bronchial washing0/1


Histoplasmosis is an endemic disease in a large area of the central United States.[1-3] Although the majority of infected individuals do not develop significant long-term disease, approximately 1% develop chronic cavitary or disseminated disease.[1] Published reports and series in the cytology literature have focused on disease occurring in severely immunosuppressed patient populations, particularly those with cell-mediated immune deficiency caused by HIV or solid-organ transplantation.[15, 16, 19, 23] A summary of selected reports is presented in Table 5.[15-20] In these reports, patients presented with disease dissemination in which the fungal burden was high and yeast were easily noted in routine cytologic material, mirroring the patients' underlying Th1-lymphocyte immune deficit. Other reports have documented disease dissemination in HIV-positive patients with lymphadenopathy, hepatosplenomegaly, and suspected lymphoma.[24, 25] Arghya et al shared a particularly impressive case of an HIV-positive male who presented with multiple umbilicated cutaneous Histoplasma nodules that were confirmed by FNA.[26] It is interesting to note that there are several recent reports of cytologic diagnoses of histoplasmosis in the setting of unilateral or bilateral adrenal enlargement.[17, 18, 20, 27] Eloubeidi et al published their experience with EUS-guided FNA of bilateral adrenal masses that were suspicious for malignancy but were found to be Histoplasma.[29] This particular report strikes a chord similar to our experience in that 40% of the current study cohort presented with masses identified by either routine or PET imaging during evaluation for malignancy.

Table 5. Comparison of Selected Studies Discussing FNA Diagnosis of Histoplasmosis
StudyNo. of PatientsNo. of Immunocompetent PatientsCause for ImmunosuppressionSites of FNA
  1. Abbreviations: FNA, fine-needle aspiration; HIV, human immunodeficiency virus; NA, not available; TNF-α, tumor necrosis factor-α.

Gupta 2010[19]72HIV (5 cases)Lymph nodes, various sites
Rana 2011[18]44NAAdrenal glands
Jaiswal 2011[20]55NAAdrenal glands
Singh 2012[15]21HIV (1 case)Lung, cervical lymph node
Singh 2012[16]63HIV (3 cases)NA
Ahuja 2012[17]33NAAdrenal glands
Current study4037Incident to age (1 case), anti-TNF-α therapy for rheumatoid arthritis (1 case), chemotherapy for lymphoma (1 case)Various sites including lung, lymph nodes, adrenal gland, and thyroid bed

Perhaps the 2 greatest differences in our experience compared with that of others are the number of cases investigated and the characteristics of the patients we describe. Thirty-seven (92.5%) of the 40 cases were individuals who were immunocompetent by our criteria. Furthermore, no one in the cohort was infected with HIV. The majority of patients in the current study (75%) came to clinical attention either by high-resolution imaging (either CT or PET) for staging a primary malignancy or because of pulmonary/systemic symptoms. In all cases, the differential diagnosis included malignancy.

The characteristics of the specimens in the current study also appears to be different from that typically reported. In those reports focused on immunosuppressed patients, the cytologic description of histoplasmosis includes abundant intracellular and extracellular yeast that is apparent on the routine stain.[15-20] As we noted previously, this presentation is typical in those with defective cell-mediated immunity in whom macrophages are unable to transform and form granulomas. Of the 40 cases in the current study, only 1 presented with these features. Much more commonly, our specimens consisted of macroscopically necrotic and granular material and smears revealed bland acellular necrosis with scattered epithelioid granulomas. Necrosis was bland, uniform, and finely granular without evidence of cellular debris. This is as opposed to tumor necrosis, which typically contains cellular and/or nuclear debris and may also have a significant inflammatory cell component. In the vast majority of immunocompetent cases, we could not identify yeast without the aid of GMS staining and only identified rare intracellular yeast in 2 cases. Nearly all cases required additional smears (air-dried or ethanol fixed) dedicated to fungal stains or destaining of routine cytologic stains with subsequent restaining with GMS.

These differences in cytologic findings from those previously reported may be explained by the predominant immunocompetent cohort in the current study, reflecting a localized fibrocaseous disease process and a lower overall fungal burden. In addition, our location in an endemic region for histoplasmosis predicts that the majority of our cases reflect chronic pulmonary disease. Gupta et al published what to our knowledge is the next largest series of patients diagnosed with histoplasmosis using cytology to date.[19] Five of 7 patients in their report were immunosuppressed with documented HIV. However, of the 2 patients listed as immunocompetent, 1 presented with skin nodules and hepatosplenomegaly, features that others have reported in disseminated forms of the disease and that appear more likely to occur with some form of unidentified immunosuppression.[25] In the long clinical follow-up of the current study cohort, no cause of immune suppression was noted in those patients we classified as immunocompetent.

Also noteworthy is our finding that GMS staining performed on direct smears has a higher diagnostic sensitivity than cell block. All 37 cases for which direct smears were evaluated contained yeast consistent with Histoplasma compared with slightly over one-half of cases (10 of 18; 56%) in which cell block was evaluated by GMS stain and organisms identified. Although one may argue regarding the quality of the specimen contained in cell blocks, given our preference for staining direct smears, we noted in comparing these cases that there was a similar degree of bland acellular necrosis and a paucity of inflammatory cell infiltrate (ie, the character of the specimens was very similar). One explanation for the increased diagnostic sensitivity of direct smears is that the aspirate material is spread relatively evenly across the slide, allowing the cytopathologist to evaluate a greater percentage of the specimen. With cell block, the material is compressed into small pellet for paraffin embedding and only a small fraction of the specimen is examined with each level on a glass slide. We believe this finding may allow those in endemic regions to increase their diagnostic rates.

We note that untreated Histoplasma has a relatively specific morphology and must be differentiated from other infectious organisms depending on tissue site, such as Pneumocystis jiroveci, Cryptococcus neoformans, Blastomyces dermatitidis, Candida species, Toxoplasma gondii, and Leishmaniasis. Histoplasma is an ovoid yeast measuring 2 μm to 5 μm on Romanowsky or silver stain.[1] Typically, it has 1 pole that is slightly tapered, resembling a sesame seed. Pneumocystis is a 4- μm to 6-μm, oval- to cup-shaped, often arranged in clusters, and has a dot-like intracellular trophozoite.[15] Cryptococcus generally has larger yeast forms (5μm-10 μm), more size variability, and a thick polysaccharide capsule that stains with mucin stains or India ink.[8, 19, 20, 26] Blastomyces is larger still (7μm −20 μm) and characterized by broad base budding of daughter yeast.[8, 19, 20] Candida glabrata is a small yeast with narrow base budding, but it too demonstrates variability in yeast size and generally does not cause the same disease manifestations as Histoplasma.[1, 20, 26] Toxoplasma is an intracellular parasite whose tachyzoites may mimic Histoplasma, but they are negative for fungal stains and do not bud.[15, 20] Lastly, the intracellular amastigotes of Leishmania are differentiated from Histoplasma by their nucleus and characteristic kinetoplast.[1, 19, 20, 26] An important caveat in presumptively identifying fungi in tissue or cytologic preparations is that once an organism has been treated, morphology can change significantly and definitive diagnosis may depend on culture or other testing for characterization.

The clinical laboratory diagnosis of histoplasmosis has traditionally relied on the use of multiple tests, including urine enzyme immunoassay for Histoplasma antigen, complement fixation, and/or immunodiffusion assays to detect either the yeast or mycelial antibody, or culture and rapid stains because the sensitivity and specificity for any one test is not 100%.[30, 31] The rate of detection depends on disease manifestation and immune status. Urine antigen testing has been shown to be superior to serum antigen testing.[32, 33] In a multicenter evaluation, Hage et al showed an urine antigen positivity rate of 41.3% and 50% in pulmonary histoplasmosis in immunocompetent and immunosuppressed patients, respectively.[34] That rate falls within the range of 25% to 75% presented by Wheat based on his experience.[35] More recently, Swartzentruber et al reported that antigenuria was detected in 60% of cases reported with acute pulmonary histoplasmosis.[36] However, in our experience, urine antigen detection was not as helpful in immunocompetent patients with pulmonary histoplasmosis and failed to detect antigen in the 15 immunocompetent patients tested. In contrast, it was detected in both of the immunosuppressed patients (2 of 2; 100%) who presented with disease dissemination. These findings fall more in line with those early studies by Wheat et al, who showed much lower antigenuria rates in cases of self-limited or fibrocavitary histoplasmosis.[32] Serum antigen tests were not available in the current study cohort. Although it has not been the practice at our institution, there are reports that propose the testing of bronchoalveolar lavage specimens for Histoplasma antigen.[37, 38]

In conjunction with the decreased detection of urine antigen, serologic tests by complement fixation and immunodiffusion demonstrated much lower rates of antibody detection than reported by large studies. For example, it is reported that immunodiffusion should detect antibodies in 71% to 100% and 57% to 77% of histoplasmosis in immunocompetent and immunosuppressed individuals, respectively.[35] Similarly, complement fixation (for both yeast and mycelial antibody) should detect antibodies in 85% to 97% and 60% to 79% of cases, respectively. Histoplasma yeast antibody was detected in 13 of 22 cases (59.1%) and the mycelial antibody was detected in 1 of 22 cases (4.5%) included in the study. Immunodiffusion was positive in only 10 of 21 cases (47.6%). Although others have proposed using polymerase chain reaction-based assays for the nucleotide sequence of the Histoplasma M antigen, to the best of our knowledge this method of detection has not caught on clinically.[39]

Finally, although culture is considered the diagnostic gold standard, fungal cultures were not helpful in establishing the diagnosis in the cases in the current study for 2 main reasons. First, only 1 of 29 fungal cultures sent cultivated Histoplasma. This was a blood culture from an immunosuppressed male with disseminated disease that was collected after a presumptive FNA diagnosis was made and was confirmatory given the patient's clinical presentation. All other cultures, including from aspirate specimens, bronchoalveolar lavage, tracheal aspirates, and bronchial washings, were negative. We noted that over the study period, the fungal culture-ordering practice of our clinicians changed such that over the last 3 years, with the exception of a case of dissemination, FNA specimens were not cultured and only the occasional bronchoalveolar lavage specimens was cultured. A second reason culture may be less helpful is the time duration from specimen collection to productive growth and identification, which can range from several days to weeks, during which time other test modalities have either confirmed or excluded significant disease.


Histoplasmosis must be considered in the differential diagnosis of immunocompetent patients who present with constitutional symptoms, lymphadenopathy, or mass lesions that are suspicious for malignancy. In our experience in an endemic region, histoplasmosis is not infrequent in patients who undergo staging procedures, especially with high-resolution CT and PET imaging, for malignancy or those who present with an incidental finding of a mass or lymphadenopathy. We believe that increased use of EUS and endobronchial ultrasound allows for the detection and biopsy of more deep-seated lesions than reflected in previous reports. Greater than 90% of the patients diagnosed with histoplasmosis in our 10-year experience were immunocompetent, which is significantly different from other series or reports. The character of specimens in cases of histoplasmosis is creamy to granular necrotic debris. The typical smear demonstrates bland acellular necrosis with or without granulomas and is not frequently accompanied by a significant neutrophilic or giant cell infiltrate. Organisms are extracellular, in contrast to those cases in immunosuppressed individuals, and typically are noted only on GMS-stained smears. The cytopathologist must be aware of the macroscopic and microscopic nature of specimens with Histoplasma so that additional direct smears or cell blocks can be made for fungal stains. In our experience, GMS-stained direct smears yield higher diagnostic rates. This is especially important because laboratory tests and fungal cultures are not sensitive in detecting histoplasmosis in this patient population. Fortunately, FNA allows for direct, presumptive diagnosis.


No specific funding was disclosed.


The authors made no disclosures.