Outbreak of Intestinal Infection Due to Rhizopus microsporus

Setting.The hematological oncology unit at Queen Mary Hospital, a university-affiliated teaching hospital in Hong Kong, serves as a tertiary referral center for treatment of refractory hematological malignancies and bone marrow transplantation. During the outbreak period, there were two pediatric (wards C6 and K8N) and two adult (wards K20N and J8N) medical wards providing care for these patients or for bone marrow transplantation. Except for C6, all these wards were equipped with positive-pressure high-efficiency particulate-filtered air in protective isolation rooms (K8N with 7 rooms, K20N with 14 rooms, and J8N with 10 rooms) for patients with neutropenia.

Outbreak investigation.An outbreak of intestinal infection due to R. microsporus occurred among these inpatients over a period of 6 months. Invasive intestinal zygomycosis was defined as a symptomatic case having an intestinal tissue specimen showing characteristic histological tissue invasion by aseptate hyphae with wide but varying diameter and a positive culture for R. microsporus with or without extraintestinal dissemination. Mucosal zygomycosis was defined as a case with abdominal symptoms, stool culture positive for R. microsporus, and radiological evidence of bowel mucosal thickening but without histological proof of tissue invasion. A case of colonization was defined as an asymptomatic case with surveillance stool culture positive for R. microsporus. Microbiological laboratory data were also retrieved to identify any unrecognized case and the past incidence of this rare infection. The medical records of the case patients were reviewed by clinical microbiologists and the infection control team as described in our previous outbreak investigations (5, 6, 7, 66). The hematologists were interviewed regarding any change in the chemotherapeutic regimen which might have altered the degree of immunosuppression in this group of patients, and the nursing staffs were inspected for any changes in patient care practice. Patients and their relatives, if available, were interviewed in detail regarding food, drinks, eating utensils, and drug history. A case control study was performed to identify the risk factors for the outbreak. Since their immunocompromised state is a known risk factor for zygomycosis, patients staying in the hematology and bone marrow transplant units with stool culture negative for the fungus were chosen as the control.

Screening of stool samples.During the period of the outbreak investigation, stool samples from all hospitalized patients being tested for bacterial culture in Queen Mary Hospital were also screened for Mucorales. To avoid potential contamination by wooden spatulas used during specimen collections (15, 57), stool samples were collected with sterile swabs and transported in sterile bottles. Pea-sized, formed stool of about 1 g or 0.5 ml of watery stool was streaked onto Sabouraud dextrose agar plus chloramphenicol (50 mg/liter) (Becton Dickinson and Company) and incubated at 37°C for 7 days.

Screening of environmental, food, and drug samples.Environmental samples from clinical areas, the kitchen, and the pharmacy and air samples from clinical areas, the kitchen, and the corridors were collected for fungal culture. Food items including fresh fruits, prepackaged ready-to-eat food, and taste additives were tested as well. Food items were cultured as described previously with modifications (51, 52). Tested items and swabs were inoculated into enrichment broth (containing peptone [10 g/liter], yeast extract [1 g/liter], dextrose [25 g/liter], and chloramphenicol [50 mg/liter]) and Sabouraud agar for incubation at 37°C. All broths were checked for mycelium formation every 24 h and subcultured on Sabouraud dextrose agar at day 5 for 48 h of incubation at 37°C. Quantitative culture was performed for the unopened refrigerated items of the same batch number if there were positives on culture for Rhizopus spp.

For screening of drugs, after swabs for fungal culture were taken from the internal surface of an opened vial, three tablets from each bottle were immersed in enrichment broth (containing peptone [10 g/liter], yeast extract [1 g/liter], dextrose [25 g/liter], and chloramphenicol [50 mg/liter]) incubated at 37°C for up to 7 days. Aliquots were taken for subculture on Sabouraud agar either if the broth became turbid or after 7 days, whichever was earlier. Quantitation of mold was performed by dissolving one tablet in 1 ml of enrichment broth, dividing into 10-μl aliquots in triplicate, and spreading on Sabouraud dextrose agar with modifications as described previously (53). The mean CFU count was the average of those from three different tablets in the same batch. The result was expressed as CFU/g of tablet.

Laboratory identification of Rhizopus isolates.Fungal colonies were examined under direct microscopy for morphological characteristic of Rhizopus spp. Isolates of the R. microsporus group were identified according to standard morphological criteria (17) and stored in 1 ml of sterile water at room temperature for genotypic analysis. Scanning electron microscopy was performed to identify the variety of R. microsporus group isolates.

Antifungal susceptibility and treatment outcome.Susceptibilities of R. microsporus to amphotericin B and posaconazole were determined by Etest (AB Biodisk, Sweden) according to the manufacturer's instructions. Aspergillus flavus ATCC 204304 was used as a control. The treatment outcome of our cases was recorded.

DNA extraction.Fungal DNA extraction was performed as described in our previous publications (60, 61). Briefly, DNA was extracted from 1 g of fungal cells in 10 ml of distilled water using the DNeasy plant minikit according to the manufacturer's instructions (Qiagen, Hilden, Germany). The extracted DNA was eluted in 50 μl of buffer AE (10 mM Tris-Cl, 0.5 mM EDTA [pH 9.0]); the resultant mixture was diluted 10 times, and 1 μl of the diluted extract was used for PCR.

ITS1-5.8S-ITS2 rRNA gene cluster (ITS) sequencing.PCR amplification and DNA sequencing of the ITS were performed according to published protocols (59, 62). Briefly, DNase I-treated distilled water and PCR master mix, which contains deoxynucleoside triphosphates (dNTPs), PCR buffer, and Taq polymerase, were used in all PCRs by adding 1 U of DNase I (Roche) to 40 μl of distilled water or PCR master mix. The mixture was incubated at 25°C for 15 min and subsequently at 95°C for 10 min to inactivate the DNase I. The fungal DNA extract and controls were amplified with 0.5 μM primers (ITS1, 5′-TCCGTAGGTGAACCTGCGG-3′; and ITS4, 5′-TCCTCCGCTTATTGATATGC-3′) (Gibco BRL, Rockville, MD). The PCR mixture (25 μl) contained fungal DNA, PCR buffer (10 mM Tris-HCl [pH 8.3], 50 mM KCl, 2 mM MgCl₂, and 0.01% gelatin), 200 μM [each] dNTP, and 1.0 U Taq polymerase (Applied Biosystems, Foster City, CA). The mixtures were amplified using 40 cycles of 94°C for 1 min, 55°C for 1 min, and 72°C for 2 min, and a final extension at 72°C for 10 min in an automated thermal cycler (Applied Biosystems, Foster City, CA). DNase I-treated distilled water was used as the negative control. Ten microliters of each amplified product was electrophoresed in 1.5% (wt/vol) agarose gel, with a molecular size marker (φX174 HaeIII digest; Boehringer Mannheim, Germany) in parallel. Electrophoresis in Tris-borate-EDTA buffer was performed at 100 V for 1.5 h. The gel was stained with ethidium bromide (0.5 μg/ml) for 15 min, rinsed, and photographed under UV light illumination.

The PCR products were purified using the QIAquick gel extraction kit (Qiagen, Hilden, Germany). Both strands of the PCR products were sequenced with an ABI Prism 3700 DNA analyzer (Applied Biosystems), using the PCR primers. The sequences of the PCR products were compared with sequences of closely related species in GenBank by multiple sequence alignment using the Clustal_X 1.83 software program (49).

Phylogenetic characterization.Phylogenetic tree construction was performed using the neighbor-joining method with Clustal_X 1.83. A 648-bp segment of ITS was included in the analysis.

Typing by random amplification of polymorphic DNA (RAPD).The DNA of the Rhizopus isolates was extracted using the PrepMan Ultra sample preparation kit (Applied Biosystems) according to the manufacturer's instructions. PCR was carried out as described previously with modifications (55). Reaction mixtures contained 1× PCR buffer, 3 mM MgCl₂, 200 μM (each) dNTP, 0.2 μM primer, 0.625 U of AmpliTaq Gold DNA polymerase (Applied Biosystems), and 50 ng of genomic DNA in a 25-μl volume. Amplification was carried out in a GeneAmp 9700 thermal cycler (Applied Biosystems) with cycle conditions as follows: 95°C hot start for 10 min and 45 cycles of 92°C for 1 min, 35°C for 1 min, and 72°C for 2 min, with a final extension at 72°C for 7 min. Ten μl of PCR products from each reaction were analyzed by electrophoresis in 1.2% agarose gel in 1× Tris-borate-EDTA buffer and stained with 1× SYBR gold nucleic acid stain (Invitrogen).

RAPD banding patterns were compared based on the presence or absence of observed amplification products in gel electrophoresis. Numerical analysis was performed using the GelCompar II software program, version 3.0 (Applied Maths). Cluster analysis was performed with calculation of Jaccard coefficients, and a dendrogram was constructed using the unweighted pair-group method using arithmetic averages. However, no interpretative standards for typing of filamentous fungus by RAPD currently exist.

Statistical analysis.Statistical analysis was carried out to show the absence of differences in the clinical characteristics between the patient and control groups, apart from the risk factors proposed. The nonparametric Mann-Whitney U test was chosen for the continuous variables because there is no assumption of normal distribution of values in the sample variable as distinct from parametric tests. Fisher's exact test was used to compare independent categorical variables between groups and is considered superior to the chi-square goodness-of-fit test because it does not utilize a normal approximation to the binomial distribution, which could lead to errors when the sample numbers are small. All reported P values were two sided. A P value of <0.05 was considered statistically significant. Computation was performed using the SPSS (Statistical Package for the Social Sciences) software program, version 15.0 for Windows.

Outbreak investigation.Since the opening of a bone marrow transplant center at Queen Mary Hospital in 1990, a total of 44,940 fungal surveillance cultures, including 5,319 stool samples, were performed. Up to 31 December 2005, only 1 Mucorales isolate (an Absidia sp.) was recovered, from a sputum sample in the year 2004. During the same period, a total of 100,312 fungal cultures were performed for non-bone marrow transplant patients. Twenty-five specimens from 11 patients were positive for Mucorales, but none of them was isolated from stool or intestinal tissue. During the outbreak period, 14 surveillance stool samples from 7 patients out of a total of 557 stool samples were found to have R. microsporus. For septic workup of symptomatic cases, 1 wound swab from 1 patient, 1 abdominal fluid sample from 1 patient, and 20 operative tissue specimens from 3 patients were obtained. After the retrospective review of clinical and laboratory data, a total of 12 cases (5 invasive, 2 mucosal, and 5 colonization) due to R. microsporus were identified during the study period (Fig. 1a to d). They were of ages between 6 and 73 years, and the male-to-female ratio was 9:3 (Table 1). All patients had solely intestinal involvement without clinical or radiological evidence of sinopulmonary involvement. The cases were scattered in the four hematology-oncology wards (C6, K8N, K20N, and J8N) and in one other general adult medical ward (E3), where these patients may stay temporarily before moving into hematology-oncology wards. In a case control study, the use of allopurinol during hospitalization and the intake of commercially packaged ready-to-eat food in the preceding 2 weeks were found to be independent statistically significant risk factors for developing intestinal infection by R. microsporus (Table 2).

• Open in new tab
• Download powerpoint

FIG. 1.

Microbiological and histopathological examination of the resected intestinal specimen of case 5. (a) Potassium hydroxide (KOH)-digested wet mount of tissue stained by calcofluor white, showing bluish-white fluorescence of branching fungal elements without septum. (b) Hematoxylin-and-eosin-stained section of the mesoappendix with a vessel showing extensive thrombosis, with the lumen filled by blood cells and fungal elements. (c) Pale-staining hyphae appear as holes or bubbles in the section. Only red cells are seen in the thrombosed vessel due to severe neutropenia. Invasion into the vessel wall by fungal hyphae is seen. (d) Grocott silver-stained section of a vessel, showing wide wavy ribbon-like aseptate hyphae with broad-angle branching.

A total of 709 nonpatient specimens were obtained for fungal culture. These included 378 environmental and air samples taken from the wards (246 samples), pharmacy (17 samples), kitchen and catering service (63 samples), storeroom (40 samples), corridor (5 samples), and microbiology laboratory (7 samples). Another 181 food samples and 150 drug samples were also tested (Table 3). Cultures revealed Paecilomyces species (8 samples), Penicillium species (8 samples), Exophiala species (3 samples), Aspergillus niger (5 samples), Aspergillus fumigatus (1 sample), Absidia species (1 sample), and Rhizopus microsporus (1 sample) from environmental specimens, 5 samples positive for Mucorales (one isolate [each] of Rhizopus oryzae and Mucor species, and three isolates of R. microsporus) from food specimens, and 16 samples positive for Rhizopus species from drug specimens. Further analysis demonstrated that all 16 samples of allopurinol tablets, 3 prepackaged ready-to-eat food items, and 1 pair of wooden chopsticks were positive for R. microsporus. None of the patients had ever used wooden chopsticks, and therefore, this factor was not included in further analysis. The mean viable fungal counts of allopurinol tablets obtained from ward and pharmacy were 4.22 × 10³ CFU/g of tablet (range, 3.07 × 10³ to 5.48 × 10³) and 3.24 × 10³ CFU/g of tablet (range, 2.68 × 10³ to 3.72 × 10³), respectively, whereas the mean viable counts in the food items were about 2 × 10² CFU/g.

The Hong Kong government and the manufacturer, which is a local pharmaceutical company, were notified of this important public health incident involving Rhizopus-contaminated allopurinol. All production lines of this pharmaceutical company were immediately suspended for further investigation. The raw materials and finished tablets of allopurinol were retrieved for microbiological culture by our laboratory and the government laboratory. Territory-wide recall and suspension of the use of this brand of allopurinol tablets were urgently implemented. Subsequent laboratory tests also showed that the cornstarch used in the excipients for allopurinol was found to contain 2 CFU/g of R. microsporus. During the site visit, it was found that the active and inactive ingredients of the drug granules were mixed and heated up to 50°C in an oven for 4 h and subsequently held at 20°C with a 3% water content for 5 to 14 days. This prolonged holding time was due to inevitable delays as a result of recent overbooking of the tableting machine. Such a long holding time provided an excellent opportunity for this thermotolerant fungus to multiply to a high level.

In view of this unusual occurrence of fatal cases due to intestinal zygomycosis in our hospital, the Hospital Authority of Hong Kong initiated a retrospective review and a 2-week prospective enhanced surveillance for all hospitalized hematology patients at all regional hospitals in Hong Kong. Ten other patients with intestinal infection by R. microsporus were identified in four regional hospitals. Two patients who succumbed in two other regional hospitals were confirmed to have invasive intestinal zygomycosis due to R. microsporus by postmortem histopathological and microbiological examination.

Laboratory identification of Rhizopus isolates.All patient-derived strains were identified as R. microsporus by their morphological appearance of fluffy, rapidly growing mold with aseptate, broad, readily twisted, ribbon-like hyphae, the presence of rhizoids directly under sporangiophores with unbranched sporangiophores, sporangium size of less than 100 μm, sporangiophore length of less than 1,000 μm (Fig. 2a to c), the absence of azygospores, and optimal growth at 42°C. R. microsporus var. chinensis and R. microsporus var. rhizopodiformis were differentiated by the spore appearance on electron microscopy (Fig. 2d to e). However, morphological transitions of spores between R. microsporus var. chinensis and R. microsporus var. rhizopodiformis were observed in the purified colonies under electron microscopy, which rendered the differentiation useless for typing purposes (Fig. 2f).

• Open in new tab
• Download powerpoint

FIG. 2.

(a) Clinical specimen was cultured for 48 h at 37°C; a fluffy colony with a pale gray color filled up the whole Sabouraud dextrose agar plate (80 mm). (b) With lactophenol cotton blue preparation, globose sporangia (up to 80 μm in diameter) were found with simple rhizoids directly under unbranched, short (up to 500 μm in length) sporangiophores. (c) In three-week-old culture, azygospores (up to 30 μm in diameter) were not found in all patient, food, and drug isolates except in the control strain. (d) For scanning electron microscopy examination of spore wall ornamentation, R. microsporus var. rhizopodiformis was observed as small, regularly spaced, and cylindrical spines formed on the wall of globose spores. R. microsporus var. rhizopodiformis was identified in patients' isolates (P3, P7, and P9) and in allopurinol (D2). (e) R. microsporus var. chinensis; rows of blunt warts were arranged mainly in line from pole to pole on the wall of ellipsoidal spores. R. microsporus var. chinensis was identified in patients' isolates (P1, P2, P4 to P6, P8, P10, and P11), food isolates (F3 and F4), and raw material (cornstarch) of allopurinol (R1). The presence of R. microsporus var. chinensis and that of R. microsporus var. rhizopodiformis were found simultaneously in isolates from patient (P12) and allopurinol (D1, D3, and D4). (f) Morphological transition of spores between R. microsporus var. chinensis and R. microsporus var. rhizopodiformis was observed in the purified colonies by electron microscopy.

Antifungal susceptibility and treatment outcome.Susceptibilities of 12 patient-derived R. microsporus isolates to amphotericin B and posaconazole were determined by Etests. The MICs of amphotericin B and posaconazole ranged from 0.06 μg/ml to 1.5 μg/ml and 0.5 μg/ml to 2 μg/ml, respectively. The MICs of the control strain of A. flavus ATCC 204304 against amphotericin B and posaconazole were 0.5 μg/ml and 0.25 μg/ml, respectively.

Based on these MIC results and the previous reports of in vitro synergism between liposomal amphotericin B (AmBisome) and posaconazole (34) and between caspofungin and posaconazole (11), our patients were treated with these three antifungal agents, either in combination or with posaconazole alone, in addition to surgical intervention where possible, depending on the extent of involvement (Table 1). For the five patients with invasive intestinal disease, three (cases 1, 2, and 5) were identified prospectively and were given the combination therapy. Cases 1 and 5 succumbed despite surgery and use of antifungal agents due to the extensive involvement and failure of neutrophil recovery, whereas case 2 survived with an appendectomy, antifungal therapy, and recovery of neutrophils. The other two patients with invasive intestinal disease (cases 3 and 8) were identified retrospectively and were treated with fluconazole and voriconazole, respectively, as empirical antifungal therapy. The two patients (cases 4 and 12) with mucosal zygomycosis were treated with combination therapy (case 4 received anidulafungin instead of caspofungin), and both responded clinically with microbiological clearance of Rhizopus microsporus in subsequent stool samples. One of them (case 4) succumbed 44 days after the onset of symptom due to severe veno-occlusive disease with liver failure and nonengraftment. All except one (case 9) of the five patients (cases 6, 7, 9, 10, and 11) with asymptomatic colonization were treated with posaconazole because of upcoming chemotherapy or conditioning for bone marrow transplant. All, including the patient who did not receive treatment, remained asymptomatic, with documented clearance of R. microsporus in subsequent stool cultures. After the use of this contaminated brand of allopurinol was stopped by switching to that produced by another manufacturer and after the strict enforcement of taking only boiled food items by this group of patients, there were no reports of any further cases in the subsequent 4 months of surveillance.

ITS sequencing.PCR of the ITS region of the fungal isolates produced amplicons of about 600 bp. Sequencing of the amplicons showed unambiguous sequence in 24 isolates. Phylogenetic analysis showed that there was >98% nucleotide identity between the sequence of any of the 24 isolates and those of Rhizopus microsporus var. rhizopodiformis (DQ641306), Rhizopus azygosporus (DQ119008), Rhizopus homothallicus (AF115728), and Rhizopus oryzae (AB097378). Among the 24 isolates, there were 8 sequence types (Fig. 3). The first one consisted of four patient isolates (P2, P3, P6, and P7), five drug isolates (D1-1, D1-2, D1-3, D2, and D3-2), two food isolates (F3 and F5), and one cornstarch isolate (R1); the second one consisted of two patient isolates (P8 and P9), one drug isolate (D4-2), and two food isolates (F4-1 and F4-2); the third one consisted of two food isolates (F2-1 and F2-2); and the other five consisted of one isolate each (P1, P4, P5, P10-1, and D5).

• Open in new tab
• Download powerpoint

FIG. 3.

Phylogenetic tree showing the relationship of 24 strains isolated from patients, foods, drugs and cornstarch, inferred from ITS (648 nucleotide positions) sequence data by the neighbor-joining method and rooted using sequence of Pirella circinans (AM778602). The scale bar indicates the estimated number of substitutions per 50 bases. Numbers at nodes indicated levels of bootstrap support calculated from 1,000 trees. All names and accession numbers are given as cited in the GenBank database.

RAPD.PCR was performed with four random primers (OPC-01, TTCGAGCCAG; OPC-02, GTGAGGCGTC; OPC-11, AAAGCTGCGG; OPC-19, GTTGCCAGCC). Each set of reaction was performed in duplicate. Weak and indeterminate bands were excluded from the numerical analysis. For the R. microsporus isolates P9 and F1, we failed to produce consistent band patterns using the above primers and hence they were excluded from RAPD analysis. A total of 61 RAPD markers were established from the present study (unpublished data). We attempted to type the isolates by RAPD analysis, with limited evidence of clonal distribution.