Increasing fungal isolation from clinical specimens: experience in a university hospital over a decade

Journal

of Hospital

Infection

(1997) 35, 185-195

Increasing fungal specimens: experience

M. Weinberger

isolation from in a university a decade

*, T. Sacks*, J. Sulkesf, I. Polacheck*

clinical hospital

M. Shapiro*

over

and

*Department of Clinical Microbiology and Infectious Diseases, Hadassah University Hospital, Jerusalem; -/-The Epidemiology Unit, Rabin Medical Center, Beilinson Campus, Petah-Tikva, and The Suckler Faculty of Medicine, Tel-Aviv University, Ramat Aviv, Israel Received

19 June 1996; revised manuscript

accepted 5 September

1996

Summary: The local patterns

of fungal isolates were studied by a retrospective analysis of fungal species isolated from clinical specimens in a university hospital in Jerusalem. Between 1984 and 1993, 5630 fungi [4071 patient unique isolates (PUI)] were isolated and identified. During the study decade, the annual incidence of all isolates increased 2.7-fold, and PUI increased 1.6-fold. Candida albicans accounted for 61% of PUI; urine was the source of 53%. The intensive care units (ICUs) and the Bone Marrow Transplantation (BMT) Department had the highest incidence of fungal isolation. The following trends were observed: (1) a decrease in the relative frequency of C. albicans and increase in Candida tropicalis; (2) increased number of isolates from urine, surgical wounds and intra-abdominal sites; (3) increased number of isolates from ICUs and BMT. Fungi are emerging as important hospital-acquired pathogens in tertiary care and teaching hospitals, and are associated with high rates of morbidity and mortality. It is important to be familiar with the local patterns of fungal isolation in order to improve treatment. Keywords: Fungi; Candida; surveillance;

ecology; hospital-acquired

infection.

Introduction Fungi have become a leading cause of hospital-acquired infection in the 1990s. Data from the National Nosocomial Infection Surveillance System (NNIS) in the USA revealed an almost twofold increase in hospital-acquired fungal infections between 1980 and 1990.’ Candida species currently account for approximately 10-l 5% of nosocomial bloodstream infections in tertiary care hospitals in the USA,2,3 while Aspergillus species are the commonest cause of pneumonia in patients following bone marrow transplantation or with aplastic anaemia.4,s 01956701/97/030185+11

SlZ.OO/O

0 1997 The Hospital

185

Infection

Society

186

M. Weinbetger

et al.

High mortality, delayed diagnosis and the difficulty in treatment have raised major concerns over the increasing frequency of fungal infections.“12 We conducted a retrospective analysis of fungal isolates over a 10 year period in a single tertiary care hospital in Israel. Trends in the isolation patterns in our hospital mirrored those reported for hospital-acquired fungal infections elsewhere.‘*13

Methods

Data about fungal isolates in Hadassah University Hospital were collected, retrospectively, from January 1984 to December 1993. Hadassah (EinKarem) is a 600-bed tertiary care, referral hospital, affiliated to the Hebrew University Hadassah Medical School in Jerusalem. Specimens were submitted either directly to the Mycology Laboratory, or to the General Microbiology Laboratory. When a specimen submitted for bacterial culture yielded a fungal isolate it was referred to the Mycology Laboratory for identification. Specimens from superficial skin lesions for diagnosis of dermatophytes were excluded. For the period 1984-1986, the data were retrieved from the records of the Mycology Laboratory, and from 1987-1993 from the computerized database of the Microbiology Laboratory. For each positive culture the following information was sought: patient’s sex, ward, date of specimen collection and it’s source, and final identification of the fungal isolate. From 1987 the patient’s age was also available. Fungal identification was performed according to routine methods.‘4”5 Specimens were screened for fungi using potassium hydroxide (KOH) preparation, Calcofluor White or Gram stains. Specimens were inoculated onto primary mycologic isolation media: Sabouraud’s dextrose agar (SDA) and brain-heart infusion agar, both containing gentamicin 5 mg/L and chloramphenicol 50 mg/L. Sterile body fluids were centrifuged before inoculation. Cultures were incubated at 30°C for two to 30 days. Positive cultures were subcultured on SDA for the isolation of a pure, single colony for identification. The identification of yeast-like fungi was based on physiological tests (germ-tube formation), biochemical characteristics, including carbohydrate assimilation (API 20C AUX, Bio MCrieux SA, Lyon, France) and morphology (Dalmau plate technique on cornmeal-Tween 80 agar). Identification of filamentous fungi was based primarily on the macroscopic and microscopic morphology (slide culture techniques using SDA, potato dextrose or Czapek-Dox agar as sporulation media). To be included in the study, yeast isolates had to be identified to the level of species, and filamentous fungi at least to genus level. Results such as ‘Candida species’ or ‘mould species’ were excluded.

Rising

incidence

of fungal

isolates

187

1

4o01 350 300 250 200 1 150 100 50 0

Figure 1. Increase University Hospital

1984

1985

1986

in the annual (1984-1993).

1987

1988 1989 Year of study

incidence of fungal ( n ) Patient unique

1990

isolates fungal

1991

1 1992

1993

over a decade at the Hadassah isolates; (0) all fungal isolates.

Definitions Source :

The body site from which the fungal culture was obtained.

When the same fungal strain was isolated more than once from the same patient it was counted only once.

Patient

unique isolates (PUI)

Hospital

admissions:

:

All admissions to the hospital.

Department

admissions: All admissions to a department, admissions and transferrals from other departments. Statistical

including primary

methods

Statistical analysis was done using the SAS statistical analysis software (SAS Institute Inc., Cary, NC). To show the change in patients’ mean age over the study years, analysis of variance was done using the multiple comparison option of Duncan. To analyse the trend in the annual incidence of PUI by species, by source or by department during the study years, Pearson coefficient (r) and the significance (P) were calculated. To compare the annual incidence of PUI in a particular department to that of a reference department (Internal Medicine), odds ratios and significance were calculated. For all calculations, P-values
From 1984-1993 the Mycology Laboratory identified 5630 fungal isolates from 3451 patients with a mean age of 47.9 years (median: 50 years, range

188

M. Weinberger Table

I.

Pattern

et al.

of fungal isolates in Hadassah

University Fungal Number

Yeasts

Candida species Candida albicans Candida tropicalis Candida glabrata Candida parapsilosis Candida &t&i Candida kefwr Candida li&vtica Candida I&taniae Can&da auilliermondii Candida itellatoidea Candida pelliculosa Candida rugosa Candida humicola Candida ciferrii Candida inconspicua

Filamentous

over a decade

isolates* Percent

3853 3791 2480 518 388 247 73 41 15 14

94.6 93-l 60.9 12.7 9.5 6.1 1.8

4 2 2 1

1 1

Other

Trichosporon cutaneum (beige ‘Iii) Cryptococcus neoformans Hansenula anomala Rhodotorula rubra Saccharomyces cerevisiae Sporobolomyces species

Hospital

!i 5 2 2 2

fund

I.5 I.2 0.1 0.1 0.1 0.1 5.4 3.9

Aspergillus species Aspergillus niger Aspergillus flavus Aspergillus fumigatus Aspergillus terreus Aspergillus species Aspergillus glaucus

::; 0.6 0.3

Mucoraceae

Rhizopus species Mucor species Other

Penicillium species Alternaria species Paecilomyces species Fusarium solani Fusarium species Geotrichium candidum Geotrichium penicillatum Microsporum canis Scopularopsis species Acremonium species Cladosporium species Chrysosporium species Pseudoallescheria boydeii Trichphyton rubrum Trichphyton mentagrophytes Verticillium species Total * Patient

;:: 0.1 ;:; 0.1 0.1

1

1 1 4071

unique

fungal

isolates.

100

1Rising

incidence

of fungal

-

-

189

isolates

100 90

10 0

-

-

1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 Year of study Key: El AI1 other Aspergillus niger ES! Candidu kefyr EZI B-ichosporon cutaneum

I I

m

0

Candida Candidu Calulida Can&h

pampsilosis glabrata tropicalis albicans

Figure 2. Patterns of fungal isolates at the Hadassah University Hospital (1984-1993). Changes in the annual incidence of species isolation. PUI, patient unique isolates.

l-97 years), and a male:female ratio of O-86. From 1987-1993 mean age increased from 47.8-49.7 years (P= 0.04, Duncan’s multiple comparisons). The annual incidence of identified fungal isolates per 10 000 hospital admissions increased 2*7-fold and that of PUI l-6-fold (from 137.2 to 355.6 and from 108-l to 168.8, respectively) (Figure 1). There was a strong, statistically significant correlation for the annual incidence of fungal isolates and the study years (Y =0*80, P=O*OOS, Pearson correlation), and a borderline correlation for the annual incidence of PUI and the study years (r=0*56, P=O*O9). Yeasts accounted for 95% of all PUI, and filamentous fungi for only 5% (Table I). Filamentous fungi were not isolated from urine, and on only two occasions (1%) from blood. Sites with a higher proportion of filamentous fungi isolation included nares and sinuses (60% of PUI from that site), ears (5 l%), lungs and pleura (40%), and bronchoalveolar lavage (27%). Candidu albicans was the most frequently isolated species, and accounted for 61% of all PUT (Table I). Four species, C. ulbicuns (61%), Cundidu

190

M. Weinberger

et al.

I al

Figure

3. Sources

of fungal isolation. unique isolates;

Changing incidence BAL, bronchoalveolar

over

the study lavage.

years.

PUI,

patient

tropicalis (13%), Candida (Torulopsis) glabrata (lo%), and Candida parapsilosis (6%) accounted for 90% of all PUI. The pattern of blood isolates was slightly different: C. albicans was isolated less frequently (55%), while 19%; the non-albicans Candida species were more common (C. tropicalis, C. parapsilosis, 11%; and C. glabrata, 7%). species accounted for 72% of filamentous fungi (Table I). Aspergihs Aspergillus niger was the commonest species (41%) followed by Aspergillus flavus (29%), Aspergillus fumigatus (15%), and Aspergillus terreus (8%). The were: ears (50 specimens); bronchoalveolar major sources of Aspergihs lavage (32); sputum (23); nares and sinuses (18); and lung and pleural fluid (8). If ear isolates are excluded, A. flavus was the most frequently isolated. The annual isolation (PUI per 10 000 hospital admissions) of several C. tropicalis (r = 0.94, fungal species increased significantly, including: P C. glabrata > C. parapsilosis > C. tropicalis, but this had changed in 1993 to:

C. albicans > C. tropicalis

> C. glabrata

> C. parapsilosis.

Urine was the major source and accounted for 53% of PUI.

Other sources

Table

II.

Department

incidence

Pattern

of fungal isolates according to departments

No.

Medicine Internal medicine Other medicine Surgery General surgery Other surgery Paediatrics* Intensive Care RICU ICCU PICU Haematology/Oncology Haematology Oncology BMT Obstetrics and Gynecology Neonates Outpatients & ER

paediatric

PUI

(%)

478 (11.7) 121 (4.0) 50 (1.2) 159 179 96 80

(3.9) (4.4) (2.4) (2.0)

6 408

(0.1) (10.0)

4071

(100)

of fungal

No. department admissions

817 (20.1) 2.56 (6.3)

Total * Including

Rising

isolates

191

PUI per IO4 admissions

Odds ratio

P-value

(11.1) (11.2)

211.0 65.8

1.0 0.3


27 018 (7.7) 78 905 (22.6) 40 930 (11.7)

127.3 107.9 55.2

0.6 0.5 0.3


(1.3) (l-4) (0.7)

1026.0 253.4 210.5

4.9 I.2 1.0


6 945 (2.0) 14 639 (4.2) 2387 (0.7) 56 654 (16.2)

228.9 122.3 402.2 14.1

1.1 0.8 I.9 0.1

NS
0.01 -


38 728 39 934

(%)

4659 4775 2 375

32 021 348970

(9.2)

-

(100)

I.9

-

-

-

surgery.

PUI, patient unique isolates; RICU, respiratory NICU, neonatal intensive care unit; PICU, transplantation; ER, emergency room.

intensive paediatric

care unit; intensive

ICCU, care

cardiac unit;

intensive BMT,

bone

care unit; marrow

accounted for ~10% of PUI. There was a significant increase in the annual incidence (PUI per 10 000 hospital admissions) of fungal isolation from urine (r = 0.81, P= 0.005, Pearson correlation), intra-abdominal sites (r = 0.83, P=O*OO3), surgical wounds (r=0.84, P=O*OO2) and ears (r=0*88, P~0.01); with a trend to increased isolation from bronchoalveolar lavage (r=0*73, P=O*O6) and intravenous catheter tips (lines) (r=0*60, P=O.O7). Fungaemia increased 1.3-fold from 1984 to 1993, but this was not statistically significant (r = O-48, P = O-16) (Figure 3). More than 55% of positive fungal cultures were obtained from patients in the medical and surgical departments. The ICUs, including BMT Department, contributed another 18% (Table II). However, the relative contributions (PUI per 10 000 departmental admissions) of the Respiratory ICU and the BMT Department were the highest (Table II). There was a significant increase in the annual incidence (PUI per 10 000 departmental admissions) of positive cultures from the Cardiac ICU (r=0.79, P=O*OO6, Pearson correlation), the Respiratory ICU (r=0*72, P=O*O2), and the BMT Department (r=0*90, P=O.O004). On the other hand, there was a

192

et al.

M. Weinberger

3

1800 y/,

YICCUHaematology ~~~~ T-4 4 Yearofstudy

CJrl

z

$j

Figure 4. Annual incidence of fungal isolation according to department. PUI, patient unique isolates; RICU, respiratory intensive care unit; BMT, bone marrow transplantation department; ICCU, cardiac intensive care unit; PICU, paediatric intensive care unit.

decrease in the number of specimens obtained from medicine departments other than internal medicine (Y= -0.63, P= 0.05) (Figure 4). Discussion

Our retrospective analysis of the fungal isolates from clinical specimens in a tertiary care hospital in Jerusalem revealed an alarming increase in the annual incidence of fungal isolation between 1984 and 1993. The major source was urine (55%), which had a major impact on the pattern of fungal isolation. However, there was a significant increase in the annual incidence of fungal isolation from sources such as surgical wounds and intra-abdominal sites, and a less significant increase in that from intravenous catheter tip cultures and blood. C. albicans was the commonest isolate from all sources, accounting for 61% of all PUI and 5.5% of blood isolates. During the study the relative frequency of C. albicans isolation decreased, and the relative frequency of non-albicans Can&da species, mainly C. tropic&, increased. Similar trends have been observed elsewhere.‘3”6 Most of the patients with positive fungal cultures were in the medical or surgical departments, which also had the largest number of admissions. However, the relative contribution from the ICUs and the BMT Department was far greater than from other wards (Table II), and has increased significantly during the study period.

Rising Table

III.

incidence

of fungal

isolates

193

Pattern of fungal isolates over a decade in three studies Reference: Weinberger*

Setting Years of study No. isolates or infections Study focus Secular Fungal

trends isolates

600-bed hospital 1984-1993 40711 Fungal 60%

Weber university

isolates increaseq

Candida

(13)

600-bed hospital 198Gl989 861$

Beck-Sag&

university

Nosocomial infections 40% increase

(1)

US National surveillance 1980-1990 26 553s

study

Nosocomial infections 90% increase1

11

albicans:

C. albicans: 54%**

C. albicans: 60%

tropicalis:

C. tropicalis:

Candida

glabrata:

C. glabrata:

61%

Candida 13% Candida

16%

spp.:

19%

9%

10%

Increase site

in source/

Candida parapsilosis: 6% Candida krusei: 1.8%

C. krusei: ~1%

Wound: Urine: Blood:

Wound: Urine: Blood:

5.67 X 1.7 x 1.3 x

C. parapsilosis:

* Present study. t Number of patient unique fungal isolates. $ Number of fungal isolates from patients with hospital-acquired $ Number of nosocomial fungal infections. 7 Last study year compared to first study year. 1 1 The study period 1980-1984 compared to the study period ** Percent of 240 yeast isolates from blood. UTI, urinary tract infection.

10%

x 2.11 1 X 1.3 x I.5

Wound infection: x 3.151 UTI: x 2.3 Fungaemia: x 4.9

infections.

1985-1989.

The pattern of fungal isolation in our study resembles that reported in two recent studies from the USA1”3 (Table III). These studies, which were carried out during a similar time frame, analysed the pattern of fungal isolates from patients with documented fungal infections. By contrast, most of the isolates in our study probably represented colonization. This is why the similarity of the secular trends between the studies is important, and may indicate that invasive fungal infections are only the tip of the iceberg. Colonization patterns may indeed reflect, and even influence, the pattern of invasive infection.“~‘* The increasing annual incidence of fungal infections during the last two decades may correlate with the increasing immunosuppressed population on one hand and the more sophisticated life-support systems and surgical procedures on the other.8~‘9-21 What are the factors responsible for the increase in fungal isolation during the study years in our institution? Although our study was not designed to answer these questions, several findings may provide a clue.

M. Weinberger

194

et al.

First, the patients’ age has increased significantly during the second half of the study period. Second, the highest contribution of the positive fungal specimens was from the ICUs and the BMT Department, and this contribution has increased over the study years. Third, there has been a significant increase in sources of fungal isolates, which are better correlated with invasive infection, and could be attributed to surgery, e.g., intraabdominal specimens and surgical wounds. All these may point to an older and enlarging population of patients undergoing immunosuppression and surgical procedures, which require more intensive care in specialized units. Indeed, the BMT Department in our hospital only opened in 1988, and the liver transplantation program became active in the last two years of the study. Fungi are emerging as important hospital-acquired pathogens in tertiary care and teaching hospitals. Because they are associated with high morbidity and mortality, and are difficult to diagnose, it is important to be familiar with the local patterns, as well as the universal trends of fungal pathogens and invasive fungal infections. This work Health.

was partially

supported

by a grant

from

the

Chief

Scientist,

Israeli

Ministry

of

References 1. Beck-Sag& C, Jarvis WR, and the National Nosocomial Infection Surveillance System. Secular trends in the enidemiologv of nosocomial fungal infections in the United States, 1980-1990. J Infect D;s 1993; 16% 1247-1251. 2. Pfaller M, Wenzel R. Impact of the changing epidemiology of fungal infections in the 1990’s. Eur J Clin Microbial Infect Dis 1992; 11: 287-291. 3. Jarvis WR. Epidemiology of nosocomial fungal infections, with emphasis on Can&da snecies. Clin Infect Dis 1995: 20: 1526-1530. 4. Pannuti C, Gingrich R, Pfaller MA, Kao C, Wenzel RP. Nosocomial pneumonia in patients having bone marrow transplant. Attributable mortality and risk factors. Cancer 1992; 69: 2653-2662. 5. Weinberger M, Elattar I, Marshall D, et al. Patterns of infection in patients with anlastic anemia and the emergence of AsPergilZus as a maior cause of death. Medicine _ (Baltimore) 1992; 71: 24-l3. Wey SB, Mori M, Pfaller MA, Woolson RF, Wenzel RP. Risk factors for hospitalacauired candidemia. A matched case-control studv. Arch Int Med 1988: 149: 2349-2353. We; SB, Mori M, Pfaller MA, Woolson RF, We&e1 RP. Hospital acquired candidemia: the attributable mortality and excess length of stay. Arch Int Med 1988; 148: 2642-2645. Pfaller MA. Epidemiology and control of fungal infections. Clin Infect Dis 1994; 19 (Suppl. 1): S8-13. Komshian SV, Uwayda AK, Sobel JD, Crane LR. Fungemia caused by Candida species and Torulopsis glubruta in the hospitalized patient: frequency, characteristics, and evaluation of factors influencing outcome. Rev Infect Dis 1989; 11: 379-390. 10. Fraser VJ, Jones M, Dunkel J, Storfer S, Medoff G, Dunagan WC. Candidaemia in a tertiary care hospital: epidemiology, risk factors, and predictors of mortality. Clin Infect Dis 1992; 15: 414-421. 11. Sternberg S. The emerging fungal threat. Science 1994; 266: 1632-1634. 12. Wenzel RP. Nosocomial candidemia: risk factors and attributable mortality. Clin Infect Dis 1995; 20: 1531-1534. 13. Weber DJ, Rutala WA, Samsa GP, Wilson MB, Hoffmann KK. Relative frequency of

Rising nosocomial

incidence

of fungal

isolates

195

pathogens at a university hospital during the decade 1980 to 1989. 1992; 20: 192-197. KJ, Bennett JE. Medical Mycology. Philadelphia: Lea & Febiger,

Am J

Infect Control 14. 15.

16.

17.

Kwon-Chung 1992: 4+71. Merz WG, Roberts GD. Detection and recovery of fungi from clinical specimens. In: Murray PR, Baron EJ, Pfaller MA, Tenover FC, Yolken RH, Eds. Clinical Microbiology. 6th edn. Washington: ASM Press, 1995: 709-722. Borg-von Zepelin M, Eiffert H, Kann M, Rtichel R. Changes in the spectrum of fungal isolates: results from clinical specimens gathered in 1987/88 compared with those in 1991/92 in the University Hospital Gottingen, Germany. Mycoses 1993; 36: 247-253. Voss A, Hollis RJ, Pfaller MA, Wenzel RP, Dobbeling BN. Investigation of the sequence of colonization and candidemia in nonneutropenic patients. J Clin Microbial 1994; 32:

975-980. 18.

19.

Martin0 P, Girmenia Prospective study of velopment of invasive 1994; 13: 797-804. Bodey GP. Hematogenous

C,

Micozzi A, De colonization, mycosis in neutropenic

Candidu

organ candidiasis. 2nd edn. New

In: Bodey GP, Ed. York: Raven Press

20. Bodey

as major

pathogens.

21.

H, et al. New spectrum 1989; 11: 369-378.

Puthogenesis, 279-329.

and major

Bernardis F, Boccanera M, Cassone A. use of empiric amphotericin B and depatients. Eur J Clin Microbial Znfect Dis

Diagnosis and Treatment.

GP. The emergence of fungi 11 (Suppl. A): 411-426. Anaissie E, Bodey GP, Kantarjian patients with cancer. Rev Infect Dis

hospital

J

Candid&is: Ltd,

1993:

Hasp Infect 1988;

of fungal

infections

in