The Brief Case: Mold Infection of an Indwelling Cranial Device—a Perplexing Combination of “Classic” Laboratory Findings

CASE

A 40-year-old woman with a history of recurrent medulloblastoma presented for outpatient evaluation of a postoperative wound of the posterior scalp. The patient was diagnosed with medulloblastoma 9 years prior and was initially managed with surgical resection and craniospinal radiotherapy. She had since experienced two cerebellar recurrences, each of which was treated by tumor resection, radiation, and/or chemotherapy, with the most recent recurrence being 8 months prior. Surgical management of the last recurrence included the placement of a posterior titanium mesh; the craniotomy site subsequently experienced poor wound closure, requiring multiple revision procedures. She had been prescribed several months of oral amoxicillin-clavulanate to prevent infection of the surgical site.

At the presenting visit to our institution, a grossly dehisced wound was observed, with visible exposure of the titanium mesh and associated fibrinous exudate. A cranial magnetic resonance image (MRI) was obtained, demonstrating adjacent leptomeningeal enhancement (along with further progression of the medulloblastoma). These findings prompted hospital admission for debridement of the site, removal of hardware, and wound closure. During this procedure, gross exudate was observed around the mesh and underlying meninges. The explanted device and an adjacent soft tissue sample (Fig. 1A) were forwarded to the clinical microbiology laboratory for bacterial and fungal cultures (histologic studies were not possible due to the lack of excised tissue with an intact structure). A direct Gram stain of these two specimens revealed abundant neutrophils and filamentous fungal elements notable for their pronounced morphologic diversity. Such forms included branching septate hyphae, ∼4 to 5 μm across with parallel walls, and Gram-negative staining, suggestive of a nonmucoraceous hyphomycete (Fig. 1B). However, intermixed with these elements were broader filaments (quite varied, up to 10 μm across) without observable septa (Fig. 1C); these elements demonstrated a “ghosting” staining pattern and elicited concern for a mucoromycete. The two filament types are depicted together in Fig. 1D. No bacterial forms were evident in the Gram-stained slides.

FIG 1.

(A) Explanted titanium mesh, submitted for Gram stain and culture. Also present within the same specimen cup is adjacent, debrided soft tissue (white arrow). (B) Gram-stained touch-prep of the debrided specimen, demonstrating Gram-negative hyphae with frequent septations and acute-angle branching. (C) Additional filamentous forms within the specimens were broader and aseptate, with a “ghosting” staining pattern. While initially confused for a possible mucoraceous mold, these forms were ultimately recognized as A. niger conidiophores. Also evident in the field (highlighted between arrows) is the terminus of a conidiophore with a vesicular enlargement. (D) Septate Gram-negative hyphae and aseptate ghosting conidiophores are present together within this field. (E) Tape preparation of the resultant cultured isolate of A. niger, stained with lactophenol cotton blue, demonstrating large conidiophores (black arrows) and a terminal vesicle with biseriate conidiogenesis over its entire surface. In contrast, several hyphal fragments are also evident within the field (white arrow). (F) Darkly pigmented, globose forms were also occasionally present in the initial Gram-stained specimens, consistent with A. niger conidia. (G) Yellow-pigmented trapezoidal crystal within the same slide; these forms did not demonstrate birefringence, consistent with hematoidin. (H to J) Nonpigmented, polymorphic crystals (H and I) that demonstrated birefringence under polarized light (J), indicating calcium oxalate.

Overall, these observations indicated a device-associated mold infection but with a broad differential of potential causative species, including consideration of a polymicrobial etiology given the fungal pleomorphism. Based on these preliminary findings, the infectious diseases service initiated combination therapy with intravenous amphotericin B and isavuconazole, agents that cover both mucoraceous and nonmucoraceous fungi. In parallel with Gram staining, both specimens (mesh and debrided tissue) were directly plated for culture, unground, to maximize the isolation of any Mucoromycota present. By 2 days, growth of mold was already visible from each specimen (both fungal and bacterial media). Upon 2 to 3 days of further maturation, all isolates were identified as Aspergillus niger by classic macroscopic and microscopic criteria (see Fig. 1E and the Discussion). All cultures remained negative for any bacterial growth or additional species of fungus, mucoraceous or otherwise. Extended incubation for 4 weeks and further attempts at isolation (including isolation streaking of the initial growth and additional replating of residual primary specimen) failed to yield any organism except A. niger. Antifungal susceptibility testing was performed on this cultured strain via broth microdilution (CLSI reference protocol), with the following MIC/minimum effective concentration (MIC/MEC) values (in μg/ml): amphotericin B, 0.5; micafungin, <0.015; voriconazole, 2; posaconazole, 1; and isavuconazole, 4. (Clinical breakpoints have not been promulgated for differentiating susceptibility/resistance in A. niger, although these MIC/MEC values would all be considered wild type according to CLSI/EUCAST epidemiologic cutoff values [1].)

Given the monomicrobial culture results, the original Gram-stained slides were reexamined to account for the striking pleomorphism that was observed. Upon further review, a number of the broad filamentous forms demonstrated subtle termini with swollen structures at various degrees of enlargement (evident/indicated in Fig. 1C), suggestive of Aspergillus vesicles. Moreover, nonbudding globose bodies (∼2 μm) were noted with darkened pigmentation (albeit rarely; Fig. 1F), a notable feature of melanized A. niger conidia. Crucially, with this additional context, the broad aseptate filaments (originally interpreted as actual hyphae) were instead recognized as conidiophores. Of note, the notoriously large and aseptate A. niger conidiophores, a “classic” morphologic observation during in vitro culture but rarely observed cytologically/histologically, confounded our initial analysis and prompted the consideration of a second mucoraceous species. The formation of these reproductive cells was likely stimulated by the mesh’s exposure to ambient air within the dehisced wound.

During this additional scrutiny of the original Gram-stained slides, we also noted diverse crystal types, including yellow-pigmented trapezoidal bodies (Fig. 1G) and pleomorphic unpigmented forms (Fig. 1H and I). These crystals were reexamined under polarized light to gain insight into their identities. The yellow trapezoidal bodies failed to demonstrate birefringence, suggesting hematoidin crystals. In contrast, the unpigmented pleomorphic crystals were identified as calcium oxalate through their positive birefringence (Fig. 1J); calcium oxalate is a classic by-product of A. niger metabolism in tissue (see the Discussion for additional information). In light of these additional observations, the microbiologic diagnosis was clarified as a monomicrobial A. niger infection associated with the indwelling mesh, with the presence of both in situ conidiogenesis and fungus-associated calcium oxalate crystals.

Clinical care of the patient continued in parallel to these laboratory efforts. Cerebrospinal fluid obtained via lumbar puncture 6 days postoperatively demonstrated normocytosis and yielded no microbial growth or detectable fungal antigens, including beta-d-glucan (Fungitell) and Aspergillus galactomannan (Platelia). These findings strongly suggested that the infection had not progressed from the anatomic site of the mesh (the subarachnoid space) to the deeper subdural space. The patient was discharged from the hospital on postoperative day 15, with ongoing palliative management for the medulloblastoma. At the time, the scalp incision appeared to be healing well and without associated discharge. She was maintained on oral isavuconazole, given the heretofore favorable response, but passed away 4 weeks later from the underlying neoplasm.

DISCUSSION

Although a standard histologic practice (2), the direct visualization of molds is far less common via direct Gram-staining or cytologic preparations (3). In all of these settings, an ability to recognize mycotic infections and perform limited characterization of the pathogen can add significant value to patient care. Of note, while species-level identification of molds in situ is challenging (often impossible), differentiating mucoraceous from nonmucoraceous features is important. This is especially true given potential differences in empirical therapy, with the intrinsic resistance of Mucoromycota to many triazole class agents (posaconazole and isavuconazole being notable exceptions). Nevertheless, atypical clinicopathologic scenarios can create distinct obstacles for direct morphologic assessment. For instance, although the current case highlights key diagnostic features of A. niger, several aspects of the case initially placed these findings out of context, including the site of infection (A. niger has not previously been reported from an indwelling cranial device), the formation of conidiogenic structures in situ, and the presence of multiple crystal types.

A. niger is a ubiquitous saprophyte found in organic matter such as decaying vegetation and stored grain. Like other members of the genus, A. niger can elicit opportunistic infections of immunocompromised individuals, including invasive hyphomycoses (i.e., aspergillosis) of the lungs and other anatomic sites (2–4). In immunocompetent hosts, it is typically only associated with non-tissue-invasive pathologies, including superficial mycoses (otomycosis and onychomycosis), fungus balls of existing cavities (physiologic or nonphysiologic spaces), and allergic disease (2). Its growth can also represent a simple laboratory contaminant. The current finding of A. niger in a device-related infection is unexpected, although it underscores the ability of diverse environmental organisms to exploit opportunities for infection, in this case, an implanted abiotic surface with partial external exposure.

Taxonomically, A. niger actually encompasses a group of closely related genomospecies that share the same morphologic features, the A. niger complex or Nigri section. In standard diagnostic practice, identification only occurs to the level of species complex, often simplified in laboratory reports as just A. niger. Genomospecies characterization requires single/multilocus sequencing (alpha-tubulin and/or calmodulin genes), nonclinical analyses that do not currently inform routine patient management (5). These fungi grow rapidly from clinical specimens, often within 2 to 3 days. Macroscopic growth initially demonstrates white obverse coloration, eventually turning brown to black with conidial development and melanization. However, the reverse of a plate (the hyphal mat) remains white or yellow, maintaining the functional characterization of A. niger as a hyaline mold. Microscopically, it demonstrates extremely long (even millimeter length, visible to the unaided eye), smooth-walled conidiophores with round terminal vesicles. The phialides are typically biseriate, with a vesicle-annellide-phialide arrangement (although phialide-only, uniseriate genomospecies are described). Elongating, unbranched chains of spherical conidia arise from phialides over the entire surface of the vesicles (6). The conidia are produced enteroblastically (i.e., with the conidial wall generated from the inner layer of the phialide wall), with the youngest conidium of a given chain adjacent to the phialide. Although often unnecessary in light of this classic morphology, identification of the complex is possible via matrix-assisted laser desorption ionization–time of flight (MALDI-TOF) mass spectrometry (including on commercial systems).

One must note that the above-mentioned features pertain to A. niger when cultured on agar surfaces exposed to the ambient atmosphere. When invading human tissue (in essence, submerged growth), conidiation is not well supported, and Aspergillus spp. will typically only demonstrate septate hyphae with parallel walls and regular, acute-angle branching. Without their asexual reproductive structures, individual Aspergillus complexes/species cannot be differentiated from one another, as well as from other genera of hyaline molds. At the same time, this invasive hyphal morphology is still distinguishable from that of mucoraceous molds (i.e., members of the phylum Mucoromycota, formerly Zygomycota), whose tissue features include broader pauciseptate hyphae (10+ μm across, although occasionally less) with ribbon-like walls and right-angle branching (2, 3). The initial Gram stain observations of the current case thus posed a dilemma, in particular, the large conidiophores without internal septations. Once culture results became available, and with more extensive analysis of the original slides, these ambiguities were ultimately resolved, in particular with the recognition of additional reproductive structures.

Such features can serve as important clues for the presence of Aspergillus spp. given the possibility of in situ conidiation when mold-infected sites are exposed to the air (2, 7). Additional clinical scenarios where this is relevant and more commonly encountered in routine practice, especially in anatomic pathology, include pulmonary/sinus processes that abut the airway (e.g., fungus balls), chronic superficial wounds (e.g., burns), and otitis externa (in the specific case of A. niger). The presence of birefringent crystals within the Gram-stained specimens represented an additional clue as to the organism’s identity. Oxalic acid is a fermentation product of A. niger and can react with host-derived calcium ions in tissue, precipitating as crystals. The association of A. niger with calcium oxalate formation (although not absolutely specific to this species complex) is well described and typically encountered by histopathologists (2) or cytopathologists (7). In the current case, these forms were visible together with hematoidin crystals. Representing porphyrin breakdown products from extravasated erythrocytes, the concomitant presence of hematoidin likely reflected the chronic nature of the patient’s wound. While clinical microbiologists do not commonly evaluate crystals, an ability to recognize these forms on Gram-stained slides can occasionally provide valuable insight. Also, of logistical note, we conducted plane-polarized microscopy for this case within an adjacent anatomic pathology laboratory at our institution. Like many microbiology sections, we do not routinely install polarizers on our microscope condensers, although familiarity with (and access to) the technique remain important for atypical situations.

In several ways, therefore, this case illustrates an important overarching theme, that certain microscopic findings more commonly relevant to other diagnostic settings are likewise relevant to Gram staining of clinical specimens, although they can be significantly more challenging to recognize when encountered outside their traditional context. In parallel, the case illustrates several additional considerations that apply broadly to diagnostic mycology and its impact on patient management. For instance, it was noted that specimens were not mechanically ground prior to plating for culture. This practice is broadly advisable for fungal cultures, as any mucoraceous species present can be sterilized by this physical force, due to the coenocytic nature of their hyphae (i.e., pauciseptate with an extensively contiguous cytoplasm). Also noteworthy, carbohydrate antigen testing of the cerebrospinal fluid (CSF) was utilized to evaluate the progression of infection several days after the removal of the mesh. Representing cell envelope components, beta-d-glucan is synthesized broadly by many fungal clades (with Mucoromycota as a notable exception), while galactomannan production is more specific to Aspergillus spp. (although production by related genera and assay cross-reactivity can occur). While the commercial Fungitell and Platelia assays utilized here are not FDA approved for CSF (but rather serum), they have been widely investigated, validated, and employed as laboratory-developed tests to help evaluate central nervous system infections with this specimen type.

In summary, we present the case of a 40-year-old woman with a cranial device-associated A. niger infection, a novel presentation for this organism. The case highlights the expanding diversity of opportunistic infections encountered in clinical practice. It likewise demonstrates the importance of a specimen’s anatomic source when evaluating microscopically for molds, whose morphologic clues can vary with their environment.

SELF-ASSESSMENT QUESTIONS

1. Which of the following infections is commonly associated with the Aspergillus niger complex in otherwise healthy individuals?

   • a.
     Soft tissue hyphomycoses
   • b.
     Outer ear infections
   • c.
     Meningitis
   • d.
     Invasive rhinocerebral infections

1. Which of the following characteristics, when visualized by a tape preparation of a cultured isolate, would be expected for Aspergillus niger?

   • a.
     Hyphae that are pauciseptate and broad
   • b.
     Short conidiophores with ovoid terminal vesicles
   • c.
     Phialides that are uniseriate with adherent conidia in clusters
   • d.
     Melanized conidia

1. What is the underlying morphologic reason that clinical laboratories should not grind tissue specimens when homogenizing them for fungal culture?

   • a.
     The hyphae of mucoraceous molds are pauciseptate
   • b.
     The conidia of Aspergillus niger are melanized
   • c.
     The tissue morphology of Histoplasma and Blastomyces spp. is yeast phase
   • d.
     In situ conidiation can occur when molds infect air-exposed tissues

For answers to the self-assessment questions and take-home points, see https://doi.org/10.1128/JCM.01117-19 in this issue.