INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Deep-seated fungal infections are important causes of morbidity and mortality in hospitalized patients (1, 2, 5, 13). Disseminated candidiasis is the most common nosocomial fungal infection, and Candida albicans has been reported to account for 50% to more than 70% cases of invasive candidiasis (2, 5, 6, 8). However, recent reports have also suggested the emergence of infections caused by non-C. albicans candidas (3, 14, 21). In addition, less-common pathogenic yeasts (Malassezia, Trichosporon, Hansenula, and Rhodotorula spp.) have recently been reported, with increased frequency, as causes of nosocomial infections (7).

Although a rare clinical isolate, the ascosporogenous yeast, Pichia anomala (formerly Hansenula anomala) has been implicated in causing fungemia in a neonatal intensive care unit (10), interstitial lung disease (19), endocarditis (12), and enteritis (9). In addition, there have been two reports of nosocomial outbreaks due to P. anomala: one in a Neonatal Intensive Care Unit in Liverpool, United Kingdom (10), and the other in an oncology hospital in Brazil (18). We describe here an outbreak of invasive P. anomala infection in the pediatric wards of our medical center that occurred during April 1996 to February 1998, with an attack rate of 4.2%.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Epidemiologic investigation. The Nehru Hospital, affiliated with the Postgraduate Institute of Medical Education and Research, Chandigarh, India, is a 1,200-bed tertiary adult and pediatric referral center. The pediatric department includes the pediatric emergency ward (24 beds), the premature nursery (PMN; 24 beds), the neonatal intensive care unit (NICU; 16 beds), the neonatal surgical intensive care unit (NSICU; 6 beds), the pediatric intensive care unit (PICU; 10 beds), the pediatric medical ward (36 beds), and the pediatric surgical ward (24 beds). After birth, all premature neonates are routinely transferred from the labor-delivery room to either the PMN or the NICU depending on their condition and the severity of the prematurity.

(i) Patients. During April 1996 to February 1998 (the outbreak period), 9,024 patients were admitted to the pediatric department.

(ii) Case-control study. We performed a case-control study to determine the potential risk factors for P. anomala infection in premature neonates born during the period from March to June 1997. A total of 395 neonates was admitted to the PMN and the NICU during this period. We studied consecutive neonates with P. anomala fungemia (33 neonates), neonates with Candida fungemia (24 neonates), and neonates without fungemia who stayed in the hospital for more than 7 days (54 neonates). The latter two groups, neonates with fungemia due to other yeasts (group 1 controls) and neonates without fungemia (group 2 controls) served as controls.

Microbiologic investigation. All patients with clinical signs and symptoms of sepsis had blood cultures collected in biphasic media containing brain heart infusion agar and broth (3). P. anomala isolates were identified on the basis of their inability to produce germ tubes and urease, a standard sugar assimilation pattern, and the production of one to four hat-shaped ascospores (4, 20). Four randomly selected strains were sent to the Mycology Reference Laboratory, Centers for Disease Control and Prevention, Atlanta, Ga., where their identification was confirmed.

(i) Culture survey. We conducted a culture prevalence survey of all premature neonates born during the period August and September 1996. A total of 130 neonates were admitted to the PMN and the NICU during this period. During this same period, we conducted a culture prevalence study of 50 consecutive premature neonates born in our center who were inpatients for more than 7 days. These infants had cultures obtained from the mouth, rectum, umbilicus, and groin. We used sterile moistened cotton-tipped swabs and cultured on day 0 (i.e., within 24 h of birth); on postnatal days 2, 3, 5, and 7; and weekly thereafter until the patient was discharged from the ward. Swabs were inoculated within 1 h of collection into four separate quadrants of a Sabouraud dextrose agar (SDA) plate. The plates were then incubated at 37°C for 7 days. Growth of P. anomala was identified as described above.

Antifungal susceptibility testing. All P. anomala isolates had susceptibility testing performed for amphotericin B, 5-fluorocytosine, ketoconazole, and fluconazole using the macrobroth dilution method (National Committee for Clinical Laboratory Standards) (11).

(iii) MLEE. Multilocus enzyme electrophoresis (MLEE) analysis was performed on 40 P. anomala outbreak isolates. These isolates were selected to include at least one isolate per month and at least one strain from each of the following hospital areas or services: the pediatric emergency ward (including the first isolate), the PMN, the NICU, the PICU, and the children's wards. One isolate each from the hands of healthcare personnel, from intravenous cannulae, and from the contaminated wash basin were also included for this analysis. P. anomala reference strain MTCC 237 (Microbial Type Culture Collection, Chandigarh, India)-NCYC 1509 (National Collection of Yeast Culture [United Kingdom]) was also included for comparison with outbreak strains. The strains were maintained on an SDA slant at 4°C and were plated on the same media after incubation at 30°C for 4 days. The MLEE analysis was performed according to a combination of the method used in our laboratory for Candida spp. and a modification of the standard method (16). Strains were subcultured separately in sterile 50-ml portions of liquid media containing yeast nitrogen base without amino acids (Difco Laboratories, Detroit, Mich.) but including 0.03 M D-mannitol and 0.1 M sucrose in 250-ml flasks and were incubated in an incubator shaker for 24 h at 37°C at 100 rpm. The cell suspensions from each flask were then harvested by centrifugation at 2,000 × g for 10 min and washed twice in 0.1 M Tris-HCl buffer (pH 8.0). The yeast cells were lysed in an ice-cooled cup with sonicator (SONIPREP 150) for 150 cycles of 20 s each at intervals of 20 s. The soluble extracts were obtained by centrifugation (13,000 × g, 2 min at 4°C) and removal of the cellular debris. A 350-µg sample of each soluble extract was subjected to polyacrylamide gel electrophoresis (Bio-Rad Miniprotean II). This system was maintained by Tris-HCl-glycine discontinuous buffer system at pH 9.5 under nondissociating conditions using a 2.5% (wt/vol) stacking gel and a 7.5% (wt/vol) separating gel. The gels were run in the cold (4°C) at 75 V and 40 mA for 4 to 4.5 h. After electrophoresis the gels were washed twice with Tris-HCl-glycine buffer (at least 15 min for each wash). In this study four enzymes were assayed, including glucose-6-phosphate dehydrogenase (GDH), malate dehydrogenase, alpha-glucosidase, and alkaline phosphatase. The sites of enzyme activity were recorded, and the R_fs were calculated for bands of activity by using the following formula: R_f = the distance of enzyme migrated in the separating gel/distance the dye front migrated in the separating gel.

Environmental investigation. (i) Environmental cultures. During these studies, multiple environmental cultures were also obtained, including cultures of unused intravenous cannulae and all intravenous cannulae after use, intravenous fluid from the bottles before and after infusion, and swabs from various fomites and inanimate environmental sources, including the wash basins, soap, floor, window panes, incubator, ventilator, suction machine, suction catheter, stethoscope, thermometer, medicine tray, electrodes, beds, and mattresses. When broth cultures of hand wash specimens were received in the laboratory, 0.1 ml was subcultured onto a plate containing SDA with chloramphenicol and gentamicin (SDA⁺). The remaining 19.9 ml of broth specimens and agar plates were incubated at 30°C for 6 days. Additional subcultures were performed after 6 days. Cultures of intravenous cannulae involved rolling them directly onto SDA⁺ plates for two and one-half turns. Liquid specimens for culture were initially centrifuged (3,000 × g for 15 min), and the sediment was inoculated directly onto SDA plates. All swabs for fungal culture were inoculated onto SDA⁺ plates and incubated at 30°C for 6 days. Any yeast growth observed was identified as described above. Intravenous fluid specimens were also cultured for fungi. Specimens were obtained aseptically from 22 intravenous infusion bottles, including a variety of different plasma, colloid, and electrolyte preparations from different manufacturers, and obtained from different areas of the hospital.

(ii) Personnel hand cultures. During July 1996, hand wash cultures of all health care workers, including physicians, residents, nursing staff, and ward assistants assigned to the PMN, NICU, and PICU, were collected by standard bag broth technique (17). Infection control personnel conducted two visits to obtain hand wash cultures. To ensure the accuracy of this screening, health care and ancillary personnel were not notified about the requirement for hand wash cultures prior to these visits.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Epidemiologic investigation. P. anomala blood isolates were identified from 379 (4.2%) babies during the outbreak period (Table 1). This represents the attack rate among all pediatric patients in the entire hospital. During this period, P. anomala was identified in 39.8% of all cases of fungemia, non-C. albicans candidas accounted for 53.4%, and C. albicans accounted for 6.3% of all cases. In 32% of the P. anomala-infected patients, the same pathogenic yeast species was isolated from multiple samples (maximum, seven times) collected over 48 h. P. anomala was first isolated in April 1996 from a patient in the pediatric emergency ward. In May, P. anomala-infected patients were first found in other neonatal wards. By July and August 1996, the outbreak had spread to include other pediatric wards (Table 2, Fig. 1). In 1997, a total of 46 infected patients were detected in the adult wards. No new cases of P. anomala were isolated from March 1998 onward.

View larger version (28K): [in this window] [in a new window]   FIG. 1.   Distribution of P. anomala cases in pediatric wards from April 1996 to February 1998

Case control study. When compared with group 1 (other yeast type isolated) and group 2 (no fungemia) control infants, case infants (P. anomala isolation) had a lower mean birth weight (P < 0.05, Student's t-test), lower mean gestational age (P < 0.05), and a longer duration of hospital stay (P < 0.05). However, no significant difference was observed in the mortality rate amongst the patients with P. anomala (42.4%) or other yeast fungemia (37%) (Table 3). Two premature neonates who died with P. anomala infection had a postmortem examination performed. For both of these neonates, histopathologic examination identified yeast cells with accompanying acute inflammatory reactions in the kidneys, liver, and spleen. No culture was obtained from those tissues.

Microbiologic investigation. (i) Culture survey. Among 50 premature neonates, P. anomala colonization was identified in 14 (28%) neonates, and 20% of these neonates subsequently developed P. anomala fungemia. In the remaining 80% of the colonized infants, no adverse effect of the colonization was observed and no specific therapy was initiated for these patients. In P. anomala fungemic neonates, the most common site of prior colonization was the umbilicus (80%), followed by the mouth (60%), the rectum (30%), and the groin (20%). Neonates with documented sepsis were treated with fluconazole or amphotericin B; a few patients received itraconazole. The specific antifungal therapy prescribed varied according to the decision of individual clinicians, and all neonates on therapy had a good response. One patient who did not respond to fluconazole therapy was successfully treated with amphotericin B.

(ii) Antifungal susceptibility test result. Only one P. anomala isolate was found to be resistant to fluconazole by broth dilution, and one strain was intermediate in activity against all of the antifungals tested. The MIC of the resistant strain against fluconazole was >64 µg/ml.

(iii) MLEE analysis. No differences were observed in enzyme activity at different loci of the four enzymes tested except that the first three P. anomala strains, obtained during April and May 1996, expressed one additional site of GDH activity (data not shown). The non-epidemiologically related reference strain's pattern was different from the pattern for the outbreak isolates.

Environmental investigation. (i) Environmental cultures. P. anomala was cultured from one postinfusion drip set and postinfusion drip bottle (both neonates had P. anomala infection). P. anomala was also isolated from a swab collected from one wash basin. However, the yeast was not isolated from any of the 22 infusion bottles tested from different parts of the hospital.

(ii) Hand culture survey. Hand cultures were obtained from 14 physicians, 17 nurses, 7 trainee nurses, 5 ward assistants, and 5 mothers. Four physicians had hand cultures that were positive for yeast; one each had evidence of C. albicans, C. tropicalis, C. parapsilosis, and P. anomala. Yeasts were also isolated from the hand cultures of four nurses; one each had C. tropicalis, C. parapsilosis, C. guilliermondii, and P. anomala, C. parapsilosis was isolated from one trainee nurse's hand cultures. Two ward assistants had positive hand cultures; one grew C. tropicalis, and the other grew C. guilliermondii. One infant's mother's hand cultures grew C. albicans.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

Traditionally, C. albicans has been considered to be the most important yeast species to cause human infection. However, modern medical therapy and improved methods for detecting and identifying yeasts have shown that many other species may be important clinically (7). In a previous study, we showed that the isolation late for non-C. albicans candidas from blood was higher than that for C. albicans in our medical center (3). We recently observed an outbreak of one of these non-C. albicans species, P. anomala, during routine microbiologic workup of our septic pediatric patients. P. anomala (anamorph Candida pelliculosa) is a free-living environmental yeast, rarely isolated from humans, which grows well in a high-sugar-containing medium (15).

One hypothesis for this shift in species distribution was that it might be due to an increase in the number of infections caused by newer species that have demonstrably increased resistance to antifungal agents (13, 14). However, this does not appear to explain the emergence of P. anomala in our hospital since all except one of the P. anomala outbreak isolates were susceptible to the four antifungal agents tested.

P. anomala had not previously been detected in our hospital from blood, other specimens, or environmental sources. This was true despite an intensive culture survey conducted among hospitalized premature neonates during the period from November 1995 to February 1996, which followed a perceived higher incidence of candidemia among these patients. Our review of clinical and laboratory procedures and practices in the hospital found that there had been no changes in either the blood culture collection method or the laboratory workup of these cultures in the hospital's mycology laboratory. We did not isolate P. anomala from any of the infusion bottles obtained from different parts of the hospital during the outbreak. The only positive environmental culture for P. anomala was from a single bottle of haemacel obtained after infusion of a patient who had P. anomala fungemia. We hypothesize that the bottle became contaminated with P. anomala either during the infusion due to backflow or from cross-contamination via the hands of health care personnel. Two premature neonates who died during the initial part of the outbreak had histopathologic evidence of an invasive yeast infection on autopsy. Presumably, it was P. anomala; however, no cultures were taken at the time to definitively identify the organism.

The first case of systemic P. anomala infection was detected in April 1996, in the pediatric emergency ward, where patients from the community are admitted. By May 1996, seven cases of P. anomala infection had been diagnosed in the premature neonatal wards; these wards admit only babies born in the hospital. However, there was no record of a patient being transferred from the pediatric emergency ward to these neonatal wards. The emergency service is also located some distance from either of these wards, making it difficult to identify patient-to-patient transfer as the method of spread. We identified resident medical staff as a major group of health care personnel who rotated duties between these areas on a monthly roster and, subsequently, conducted a hand culture prevalence survey of the resident medical staff and nurses from the neonatal wards during July 1996. We isolated P. anomala from two hospital staff members: one resident and one nurse (4% of 48 hospital personnel investigated) from these areas, suggesting that a lack of rigorous handwashing was responsible for the spread of this infection. The colonized resident worked in the pediatric emergency ward in April 1996 and was transferred to the PMN and the NICU the following month; the colonized nurse had not worked outside the PMN and NICU. Although this finding is consistent with the resident physician disseminating the infection outside the pediatric emergency ward, we cannot exclude other scenarios, including transmission by other medical personnel who may have been transiently colonized and thus would not have been detected by our culture survey. From July 1996, P. anomala infections occurred in other hospital areas, including the PICU, the children's ward, the children's ward nursery, and the pediatric surgical ward. These areas are widely dispersed in the hospital, and we found no association with colonized hospital personnel.

In January 1997, following detection of the outbreak, hospital infection control personnel began intensive educational efforts to improve handwashing practices. Special emphasis was placed on the need to observe strict aseptic technique in caring for high-risk patients, and the cohorting of P. anomala-infected patients was begun. The attack rate decreased in February 1997. However, the rate increased again in March 1997. Therefore, in August 1997, a prophylactic antifungal regimen of nystatin (oral) or fluconazole (oral and/or intravenous) was prescribed for all patients in the PMN, NICU, and PICU, i.e., units identified as having majority of the cases. However, in November 1997, a further increase in cases was noted, and so we started routine antifungal prophylaxis of high-risk infants and an intensified hospital-wide health education campaign with poster displays, formal lectures, and informal group discussions. The campaign repeatedly stressed the need for health care personnel to adhere to strict handwashing procedures following all patient contacts.

Our analysis of potential risk factors for P. anomala infection in premature neonates found that a lower gestational age, a very low birth weight, and a longer duration of hospital stay were all significantly associated with these infections. These findings support an earlier report of a P. anomala nosocomial outbreak in pediatric patients, in which Murphy et al. reported that all patients had multiple problems associated with very low birth weight and prematurity (10). These authors found that 52 infants were colonized and, of the 8 infants (13%) who developed infection, all except one were heavily colonized prior to the invasive episode. In our study, of 50 premature neonates cultured, 28% (14 of 50) were colonized with P. anomala. We also observed, as did Murphy et al., that hand washing and cohorting of infected babies limited transmission but that elimination of the organism from the unit was possible only with the introduction of an oral nystatin prophylactic regimen together with topical application of iodophore at venipuncture sites (10).

Except for the first three patient isolates obtained during April and May 1996 that differed in expression of one additional site for GDH activity, we found that the patterns of outbreak isolates from both patients and the environment were identical, suggesting a common point source for the various outbreaks. The MLEE enzyme pattern for all of the outbreak strains was clearly different from that of the P. anomala reference strain MTCC 237-NCYC 1509. The epidemiologic relevance of the different enzyme pattern of the first three outbreak isolates requires further confirmation with another molecular typing method.

In conclusion, cross-contamination via the hands of hospital personnel and the possible role of the inanimate hospital environment as a reservoir were most likely to have contributed to this outbreak. Of importance, outbreak control could only be achieved by the combination of the introduction of prophylactic antifungal therapy in high-risk neonates and an intensive educational effort that emphasized a strict handwashing procedure after all patient contacts.