Context.—Granulomatous pulmonary nodules are common in areas endemic for Histoplasma infection, and may require surgical excision to exclude neoplasia. Surgeons may elect to routinely send material directly to the clinical microbiology laboratory for fungal and mycobacterial cultures.
Objective.—To determine if tissue from surgically excised pulmonary granulomatous nodules removed from patients in a geographic area endemic for Histoplasma infection should be routinely submitted for fungal culture.
Design.—Retrospective review and comparison of surgical pathology histochemical findings and clinical microbiology results of 30 surgical (wedge) lung excisions that demonstrated granulomatous nodule at the time of frozen section.
Results.—Twenty cases demonstrated fungal organisms consistent with Histoplasma species via histochemical fungal stains. Of these 20 cases, 17 were tested in the microbiology laboratory using direct smear examination and fungal culture; Histoplasma was detected in 1 case (1/17). Eight cases revealed no organisms by surgical pathology. Of these, 6 were tested in the microbiology laboratory, and all 6 were negative by culture and direct smear (0/6). The remaining 2 cases demonstrated organisms other than Histoplasma by surgical pathology examination.
Conclusions.—Surgical pathology examination of granulomatous pulmonary nodules detected Histoplasma organisms with greater sensitivity than culture and direct smear. There were no false-negative surgical pathology diagnoses when compared with microbiological results. These findings suggest that it is not necessary to routinely send material from solitary pulmonary granulomas for fungal culture when the material is removed from immunocompetent patients in a geographic area endemic for histoplasmosis.
Histoplasma capsulatum is a thermally dimorphic fungus that is found in a wide geographic distribution, but in the United States most commonly infects those living in the Mississippi and Ohio river valleys. H capsulatum (henceforth referred to simply as Histoplasma) is found in the environment, particularly soil enriched with bat and bird droppings, and exposure to the organism is usually via inhalation of microconidia or hyphal fragments.1–3 It is estimated that 50% to 80% of people living in endemic areas are exposed to this organism based on serologic data.4,5 Pulmonary Histoplasma infection can be broadly categorized by clinical presentation and tempo of disease. Acute pulmonary histoplasmosis occurs in those who are acutely exposed to a large burden of organism. This form of Histoplasma infection is typically asymptomatic and most often resolves spontaneously in the immunocompetent host. The minority of patients who do exhibit pulmonary symptoms may require antifungal treatment if disease is persistent and severe. The chronic (cavitary) pulmonary form of histoplasmosis is strongly associated with underlying chronic obstructive pulmonary disease, may be self-limited in some patients, and generally requires antifungal treatment.1 In either the acute or chronic forms, a parenchymal granulomatous nodule may present as a radiographically suspicious pulmonary lesion. Histoplasma granulomatous nodules (histoplasmomas or fibrocaseous nodules) are not always calcified, and they may demonstrate significant metabolic activity on functional imaging, such as with positron emission tomography (PET) studies.6 In cases where a percutaneous or bronchoscopic biopsy is not possible, these lesions may require the intervention of a thoracic surgeon in order to exclude neoplasia. In our practice, wedge resection of a suspicious pulmonary nodule/lesion is generally performed first, followed by intraoperative frozen section analysis in the surgical pathology laboratory. If the frozen section diagnosis yields granulomatous nodule, no further surgical treatment is required, and tissue is routinely sent to the microbiology laboratory for direct smears, fungal cultures, and mycobacterial cultures in order to exclude or document an infectious etiology. Although microbiologic culture is traditionally accepted as the gold standard for the diagnosis of infectious disease, it has been previously noted that fungal cultures of pulmonary granulomatous nodules are often of low yield.7,8 We anecdotally noted that histochemical staining of formalin-fixed, paraffin-embedded material often yields a morphologic diagnosis of Histoplasma in spite of negative direct smears and microbial cultures. In this study, we chose to retrospectively evaluate the diagnostic utility of obtaining routine fungal cultures on surgical lung excisions of solitary granulomatous pulmonary nodules.
MATERIALS AND METHODS
Permission to conduct this study was granted by our local institutional review board. Surgical pathology cases that demonstrated a granulomatous nodule in surgical lung resections were found using a computerized query of our laboratory information system (Cerner, Kansas City, Mo). The query was conducted using procedure, anatomic site, and diagnosis SNOMED codes with text-specific strings applied to our laboratory database between January 1, 2001 and July 31, 2004. The search was done to find surgical pathology cases of all lung (T-28000) excisions (P-1100) that contained the text string “granuloma” in the diagnosis, irrespective of the absence or presence of any anatomically identified organism. Once this list of potential cases was generated, the complete anatomic pathology record was reviewed for each patient, and cases wherein the pathologic diagnosis was known preoperatively were excluded. For example, if a wedge lung excision revealed a necrotizing granulomatous nodule, but a preoperative transbronchial biopsy of the nodule showed granulomatous inflammation, the surgical excision was excluded from the list. Each patient's electronic medical record was also reviewed, and patients who were immunocompromised (acquired immunodeficiency syndrome, chronic systemic steroid therapy, bone marrow and solid organ recipients, and those receiving chemotherapy for malignancy) or had culture-proven systemic mycosis were excluded. All surgical pathology material was reviewed in a nonblinded fashion to confirm accuracy of the original diagnosis, including the results of histochemical organism stains (Gomori methenamine silver, periodic acid–Schiff with diastase, and acid-fast); no discrepancies were found. Once these cases were identified, the corresponding clinical microbiology report was reviewed (if available), and the results of direct smear examinations (Giemsa, calcofluor white, and acid-fast stains) and cultures for fungi and mycobacterium were recorded. Clinical information and the preoperative radiographic description of the lesion were also recorded from the medical record where available.
Thirty surgical lung excisions demonstrating a granulomatous nodule were identified. The summarized clinical and radiographic characteristics of these patients are presented in Table 1, and the results of microbiologic culture and surgical pathology examinations are given in Table 2. Clinical microbiology results were available in 24 of these 30 cases. Twenty cases demonstrated ovoid fungal yeast forms 3 to 4 ;gmm in diameter without hyphae, pseudohyphae, thick cell walls, or evidence of large spherules, consistent with Histoplasma species via Gomori methenamine silver stain or periodic acid–Schiff with diastase by surgical pathology examination. Rare budding forms, characterized by narrow-based unipolar buds, were present in 9 cases. Organisms were found mostly in the necrotic center of granulomas, but occasionally were found within the hyalinized rim of the granuloma. Of these 20 cases, 17 were tested in the microbiology laboratory; Histoplasma was detected in 1 case (1/17), and this single case was associated clinically with a cavitary lesion that perforated the pleura. Eight cases revealed no organisms by surgical pathology examination. Of these, 6 were tested in the microbiology laboratory, and all 6 were culture and direct smear negative (0/6). Infectious agents other than Histoplasma were detected in the remaining 2 cases. One of these cases demonstrated acid-fast bacilli by surgical pathology examination, and this required further characterization by microbiologic methods to arrive at a species-specific diagnosis of Mycobacterium avium-intracellulare (MAI) complex. The other case demonstrated Coccidioides species by surgical pathology, but material from this case was not cultured.
Submission of surgical material for microbiologic tests is typically assumed to be of clinical value in the workup of potential infectious disease, and furthermore is traditionally seen as the gold standard of infectious disease diagnosis. These assumptions may not be applicable to a specific patient population: those who present with suspicious pulmonary nodules in a geographic region hyperendemic for pulmonary Histoplasma infection. In this series, 66% of nonneoplastic, granulomatous pulmonary nodules in our practice were found to be secondary to Histoplasma infection, and this documentation was better achieved with histologic examination rather than with microbiologic techniques such as direct smear and culture (Table 3). When compared with culture, histologic examination alone was more sensitive in detecting the organism, and in our series there were no false positives or false negatives compared with the microbiologic results (Table 4). Histologic detection of the organism had the added benefit of relatively faster turnaround time (1–2 days) compared to the culture interval required to detect or exclude Histoplasma (several days to several weeks). Similar observations were reported by Ulbright and Katzenstein7 25 years ago, also in an area endemic for Histoplasma. In their series of 86 solitary pulmonary granulomas, 24 cases of Histoplasma were identified, and fungal cultures of 19 of these cases were negative. Of the 25 cases in which no organism was seen on histology, fungal cultures were performed on 15 of these granulomas and were all negative. We conclude that routine submission of fungal cultures is of minimal diagnostic benefit in this specific clinical setting, in contrast to histochemical fungal organism staining. Consideration of fungal culture testing may be given in a case-by-case manner, as there are clinical settings in which it would be appropriate, such as those with a previous history of a systemic mycosis, those patients who present with an acute pulmonary or systemic illness, and the immunocompromised.
There are important limitations to our study. First, the sensitivity of histochemical staining for fungal organism detection was limited by the number of sections examined. In our practice, only 1 block containing 1 to 3 cross sections of the granulomatous nodule is selected and stained using Gomori methenamine silver stain or periodic acid– Schiff with diastase. Previous studies have shown increased diagnostic yield when 2 or more blocks are stained.7 Therefore, our results may actually underestimate the sensitivity of histochemical organism detection in this regard. Second, histochemical detection does not reliably discriminate between viable and nonviable organisms, which may also explain the marked detection discrepancy; however, budding organisms were visualized in 9 cases, only 1 of which was culture positive. Related to this, distinguishing Histoplasma yeast from the endospores of Coccidioides immitis can be difficult in the absence of budding. The incidence of coccidiomycosis is extremely low in our geographic region, and no patient with a diagnosis of Histoplasma via histochemical stain had a travel history to areas endemic for Coccidioides species in our study. The third limitation is the nature of the population studied. We examined the relative yield of histochemistry versus culture in patients who had a radiographically suspicious pulmonary nodule without clinical evidence of systemic infection or evidence of immune compromise. In contrast to this relatively healthy group of patients, many studies have shown that culture is very sensitive in detecting Histoplasma in the bone marrow of patients with disseminated infection.9–13 Therefore, the yield of culture in pulmonary granulomatous material is likely dependant on the extent of infection, immune competence of the host, and presence or absence of viable organism within the clinical sample. Finally, our study was retrospective, so we could not control for sampling bias; the size of the portion of tissue sent for microbiologic studies certainly varied from case to case.
The incidence of tuberculosis is low in our institution; however, a comment regarding the use of mycobacterial culture is appropriate here. As opposed to Histoplasma, which is not transmitted from human to human, an undetected or misclassified case of mycobacterial infection may have serious public health consequences. In our series, there were no cases of Mycobacterium tuberculosis, and there was 1 case of non-tuberculous mycobacterial infection (MAI). Although acid-fast organisms were detected by surgical pathology techniques in this single case, definitive classification required microbiologic techniques. Therefore we feel material should always be submitted for mycobacterial culture if granulomatous inflammation is present at the time of frozen section.
In summary, the routine practice of sending material for fungal studies is of marginal diagnostic utility in the clinical setting of an immunocompetent patient demonstrating a solitary pulmonary granulomatous nodule in an area hyperendemic for Histoplasma. Unless otherwise indicated by the clinical presentation, routine fungal cultures in this setting may contribute to inappropriate laboratory use and increase patient charges. Although the incidence of mycobacterial infection is low in our patient population, routine ordering of mycobacterial studies is required in order to ensure optimal detection and classification of acid-fast bacilli.
The authors have no relevant financial interest in the products or companies described in this article.
Reprints: Jamie A. Weydert, MD, University of Iowa Carver College of Medicine, Pathology, 5239C RCP, 200 Hawkins Dr, Iowa City, IA 52242 (firstname.lastname@example.org)