Several cases of Candida bloodstream infections were documented from June to October 2004 at a tertiary care center in Bogotá, Colombia. Since no cases of candidemia had occurred during the preceding four months, an outbreak was declared. As a result, a microbiological study, a revision of infection control practices and a case-control study were performed. In all, 18 cases of candidemia were ascertained. Parenteral nutrition (p=0.04), presence of a central line (p=0.03), and severity of illness (p=0.03) were associated with candidemia in bivariate analysis. Diverse Candida species were observed. Candida parapsilosis contamination was found in plastic containers used for transient intravenous (IV) medication storage at the bedside, plastic bags reused for the transportation of IV medicines and cotton used for disinfection of IV ports. Poor infection control practices were widely observed. The outbreak was controlled after elimination of plastic bags used for transportation, instauration of daily disinfection of IV medication containers, acquisition of sterile alcohol swabs for port disinfection and staff education. It was concluded that candidemia was associated with previously-described risk factors and that poor infection-control practices were likely responsible for the outbreak.
Candida; fungemia; disease outbreaks; developing countries; Colombia; Latin America
An outbreak of Candida spp. bloodstream infection in a tertiary care center in Bogotá, Colombia
Carlos A. DiazGranadosI, II, III; Adriana MartinezII; Ceneth DeazaII; Sandra ValderramaII
IEmory University School of Medicine, Atlanta, GA, USA
IIJorge Piñeros Corpas Clinic
IIIFundación Universitaria de Ciencias de la Salud; Bogotá, Colombia
Address for correspondence
Several cases of Candida bloodstream infections were documented from June to October 2004 at a tertiary care center in Bogotá, Colombia. Since no cases of candidemia had occurred during the preceding four months, an outbreak was declared. As a result, a microbiological study, a revision of infection control practices and a case-control study were performed. In all, 18 cases of candidemia were ascertained. Parenteral nutrition (p=0.04), presence of a central line (p=0.03), and severity of illness (p=0.03) were associated with candidemia in bivariate analysis. Diverse Candida species were observed. Candida parapsilosis contamination was found in plastic containers used for transient intravenous (IV) medication storage at the bedside, plastic bags reused for the transportation of IV medicines and cotton used for disinfection of IV ports. Poor infection control practices were widely observed. The outbreak was controlled after elimination of plastic bags used for transportation, instauration of daily disinfection of IV medication containers, acquisition of sterile alcohol swabs for port disinfection and staff education.
It was concluded that candidemia was associated with previously-described risk factors and that poor infection-control practices were likely responsible for the outbreak.
Key-Words: Candida, fungemia, disease outbreaks, developing countries, Colombia, Latin America.
Candida spp. has emerged as one of the most common causes of nosocomial bloodstream infections (BSI) in many areas of the world [1-4]. These infections are difficult to diagnose and cause significant morbidity and mortality despite antifungal therapy [5,6]. The mortality associated with Candida spp. BSI is consistently high, with estimated figures between 40%-50% . Most Candida spp. BSIs occur in patients with indwelling vascular devices . Established risk factors for invasive candidiasis include mucosal or cutaneous barrier disruption, defects in the number and function of neutrophils or in cell-mediated immunity, metabolic dysfunction, extremes of age, use of broad-spectrum antibiotics, cytotoxic chemotherapies and transplantation .
Outbreaks of candidemia have been associated with contaminated milk bottles , parenteral nutrition , glycerin suppositories , IV medication contamination , syringe reutilization , healthcare worker (HCW) hand-colonization [13,14] and contamination of indwelling vascular devices . Other outbreak investigations did not find a common source, but in many cases the outbreak was controlled after optimization of infection-control practices [16,17].
Several cases of Candida spp. BSI were documented at the Clínica Jorge Piñeros Corpas, a tertiary care center in the city of Bogotá, Colombia between June and October 2004. No cases of Candida spp. BSI had occurred during the four preceding months.
The purpose of this investigation was to determine the factors contributing to Candida spp. BSI in order to allow the development of control measures that would halt the progression of the outbreak.
Material and Methods
We performed a preliminary investigation to confirm the outbreak and a definitive investigation to determine the factors contributing to its occurrence. The preliminary investigation included establishing a case definition, performing case ascertainment, describing the clinical characteristics of cases, building an epidemic curve and performing a literature review. Cases were defined as patients with at least one blood culture (BC) positive for Candida spp. from January to October 2004, at the Clínica Jorge Piñeros Corpas, Bogotá, Colombia. Case ascertainment was performed by reviewing microbiology laboratory and medical records.
The definitive investigation included a case-control study, a revision of infection control practices in the hospital and a microbiological investigation. The first 13 cases were included in the case-control study. Controls were defined as patients with BC results different from Candida spp. (positive for other microorganisms or no growth). Patients with positive cultures for Candida spp. from other body sites were excluded. This was decided based on the fact that invasive candidiasis is difficult to diagnose [5,18,19] and that candida colonization of non-sterile sites, although non-specific, can be a marker of systemic disease [18,19]. Including patients with positive Candida spp. cultures from non-sterile sites in the control group would have created the possibility of misclassification bias of outcome, since some of the "control" patients would have potentially been classified as such when in fact they had undiagnosed invasive candidiasis. We planned to use three controls for each case, selected at random from a list provided by the microbiology laboratory that included BCs done during the outbreak period. Exposure variables collected for cases and controls included age, gender, presence of central venous catheter at time of BC, severity of illness (American Society of Anesthesiology physical status score) at the time of BC, parenteral nutrition, enteral nutrition, blood product transfusion, medications administered, invasive procedures performed and exposure to HCW within 7 days of BC. Bivariate analysis was performed using the Wilcoxon rank-sum test for numeric exposure variables and the Chi-square or Fisher exact test for categorical variables. P-values less than 0.05 were considered significant. Statistical analyses were performed using SAS software version 8.2 (SAS Institute, Cary, NC, USA).
A revision of behaviors and infection control practices was carried out by the hospital epidemiology team through staff interviews and observations related to common practices at different hospital locations. Practices related to intravenous (IV) catheter management, and to IV medication transportation, storage, preparation and administration were particularly scrutinized.
The microbiological investigation was comprised of culturing HCW hands, ventilator tubing in the intensive care unit (ICU) and coronary care unit (CCU), latex gloves, surgical tape, bandages, cotton, IV catheters, IV solutions, IV medication containers and IV medication transportation bags. Sheep blood agar cultures and Sabouraud agar cultures were used to isolate the microorganisms. A germ tube assay was performed for speciation of Candida albicans. Speciation was performed on selected non-albicans isolates using the Vitek Yeast Biochemical Card (YBC).
Information obtained from this investigation was utilized to formulate outbreak control measures. Assessment of the impact of outbreak control strategies was performed comparing the candidemia rate (cases per 1,000 patient-days) during and after the outbreak period, using a z-score calculation.
Eighteen cases of Candida spp. were documented between June and October 2004. No cases of Candida spp. BSI had been documented during the previous four months. Case patients were located at the ICU, the CCU or the pediatric ward. The mortality rate for the 18 cases was 33%. Detailed clinical characteristics for cases and controls are presented in Tables 1 and 2. Figure 1 depicts the epidemic curve for the outbreak.
Thirteen cases and 36 controls were included in the case-control study (ration 1:2.8). Parenteral nutrition (p=0.04), presence of a central venous catheter (p=0.03) and severity of illness (p=0.03) were associated with Candida spp. BSI (Tables 1 and 2). No significant associations were observed between Candida spp. BSI and exposure to a particular HCW, IV medication or invasive procedure.
Poor hand hygiene practices were widely observed. IV medications were prepared at the bedside by ICU and CCU nurses and at the nurse station by pediatric ward nurses. IV medications were transported from the pharmacy to the ICU, CCU and other hospital areas in plastic bags that were reused for several weeks or months without disinfection. IV medications were stored in plastic containers at the bedside in the ICU and CCU, which did not undergo regular disinfection. In the pediatric ward, small pieces of non-sterile cotton were prepared at the beginning of the day in non-sterile plastic cups, and an arbitrary amount of alcohol was poured on top. The cotton pieces were taken from a common source that was kept near the floor, directly exposed to the environment. The nurses used the cotton fragments to clean the IV sites and IV ports before medication injection. These practices had been initiated by the nurses during May or early June 2004.
Candida species isolated from the cases were diverse, with a predominance of Candida albicans and Candida parapsilosis. Ninety-five HCW were cultured. Personnel cultures were positive for six HCW (6.3%). More than 150 environmental cultures were made. Six environmental cultures were positive for Candida parapsilosis. Heavy Candida parapsilosis contamination was documented in plastic containers used for transient IV medication storage at the bedside, plastic bags reused for the transportation of IV medicines and cotton used for disinfection of IV sites and ports. Table 3 shows the Candida species for clinical, staff and environmental isolates.
Interventions included HCW education, elimination of plastic bags used for the transportation of IV medications, standardization of disinfection practices for the IV medication containers, replacement of alcohol-impregnated cotton by sterile alcohol swabs for IV site and IV port disinfection.
The rate of Candida spp. BSI significantly decreased from 1.13 cases per 1000 patient-days during the outbreak period, to 0.25 cases per 1000 patient-days during the post-outbreak period (z-score 2.18, p=0.014). Of relevance, the proportion of patient-days considered "at-risk" for Candida spp. BSI (the sum of number of ICU, CCU and Oncology patient-days over the total number of patient-days) significantly increased from the pre-outbreak period to the outbreak period (35.6% vs. 40.5%, z-score = -7, p<0.0001) and from the outbreak period to the post-outbreak period (40.5% vs. 49%, z-score = -9.3, p<0.0001).
The outbreak was considered controlled after a five month period, during which only one new case of Candida spp. BSI was ascertained each month (Figure 1).
This is one of the largest outbreaks of candidemia described in the literature. Previous large outbreaks reported in the literature accounted for 22 cases of Candida parapsilosis fungemia among adults in the United States , 19 cases of candidemia in a neonatal ICU in the United States , 17 cases of Candida parapsilosis fungemia in a neonatal ICU in China  and 16 cases of Candida tropicalis fungemia in a neonatal ICU in India . Three of these outbreaks were restricted to a single Candida species [13,16,21]. Similarly, three of these outbreaks [13,20,21] were restricted to single hospital areas. We included different candida species in our investigation because the epidemic curve indicated that any episode of candidemia, regardless of species, was an unusual phenomenon before the initiation of the outbreak. The observed number of cases of candidemia during the outbreak period clearly represented an increase over and above the expected number of cases. Additionally, we elected to study candidemia occurring in any hospital area because the preliminary investigation revealed that the condition had emerged almost simultaneously across several hospital units.
The observed crude mortality was 33% in our study, somewhat lower than the 41-63% range described in previous series [6,16]. Candida spp. BSI occurred more frequently in patients with severe illness, exposed to parenteral nutrition and/or central venous catheters. These conditions are previously recognized risk factors for candidemia [1,7,9, 13, 16]. Interestingly, the proportion of "at-risk" patient-days (the sum of ICU and CCU patient-days and oncology patient-days) significantly increased from the pre-outbreak period to the outbreak period and from the outbreak period to the post-outbreak period. This suggests that a change in patient susceptibility may have facilitated the emergence of the outbreak and may also explain why the rate of candidemia did not entirely return to pre-outbreak levels. However, despite the consistent overall increase in patient susceptibility, the rate of Candida spp. BSI significantly decreased after the implementation of outbreak control measures. We believe this is strong evidence in favor of outbreak control as a result of the implemented measures.
Although Candida spp. BSI usually originates from an endogenous source [1,22], several studies have described the relevance of Candida spp. as a nosocomial pathogen . Candida spp. can be transmitted to patients from hospital sources and staff skin. In the setting of patient predisposition to candidemia, poor infection control practices were the most likely culprits of this outbreak. Our documentation of hand colonization in some HCW and our observation of poor hand-hygiene practices suggest that nosocomial transmission could have been facilitated from HCW hands as transient reservoirs. Several previous studies have linked the hands of HCW with Candida BSI outbreaks [13-15,23,24]. Additionally, we found that devices used for IV medication storage and administration were contaminated with Candida parapsilosis. This is the first outbreak to suggest that contaminated non-sterile cotton, plastic transportation bags and plastic IV medication containers can facilitate the emergence of Candida spp. as a nosocomial bloodstream pathogen. Contaminated plastic bags were used for transportation of IV medications to all affected units (ICU, CCU and pediatric ward); contaminated plastic containers were used in the ICU and CCU for bedside medication storage; and contaminated non-sterile cotton was used in the pediatric ward for IV site and IV port disinfection. These environmental sources are more likely to be problematic in resource-limited settings, where a variety of strategies are often improvised as a result of misleading cost-oriented ("money-saving") policies. At the time of this outbreak, the hospital did not have a central pharmacy for medication preparation. Additionally, there were no medication transportation carts, and sterile alcohol preps were unavailable. Because of the lack of a central pharmacy and medication transportation carts, pharmacy and nursing personnel had to improvise methods for IV medication transportation and storage at bedside. Plastic bags and plastic containers were selected and reused without proper disinfection. Likewise, because of the lack of sterile alcohol preps, the nurses improvised alcohol impregnation of non-sterile cotton for IV site and IV port disinfection. These improvisations favored the emergence of the outbreak, resulting in significant hospital cost as well as patient morbidity and mortality. This report is therefore relevant to caution hospital and regional policy makers, particularly in settings with significant resource limitations, about the possible unanticipated and undesirable consequences of misleading cost-oriented policies. Such policies and their possible behavioral and clinical consequences should be discussed with infection control personnel before implementation.
Our study has several limitations. Because of resource limitations, we were unable to determine the species for all candida isolates. We believe that some of the non-albicans isolates for which speciation was not performed could have been C. parapsilosis. Therefore, our reported frequency of Candida parapsilosis from clinical samples may represent an underestimation. Additionally, we were unable to carry out molecular testing to demonstrate a link between the presumptive environmental sources and the clinical cases. Usually, a common environmental source is suspected once the same organism is obtained from the clinical infections and the presumptive reservoirs. Further molecular testing then confirms that the clinical isolates and the environmental isolates are related. Because of limited laboratory capacity and resources, we were unable to incorporate molecular testing into our investigation. Despite a predominance of C. albicans in clinical isolates and C.parapsilosis in environmental isolates, we believe that the same infection control deficiencies and similar reservoirs may have resulted in the emergence of the outbreak. This is supported by the evidence of unprecedented clustering of candida BSI during the outbreak period as well as the significant reduction of the candida BSI rate after eradication of identified reservoirs and optimization of infection control practices. Additionally, several studies suggest that different candida species may have similar adherence properties to cotton and other fabrics , as well as to plastic and other synthetic materials [25-27]. It has been demonstrated in-vitro that once adhered to fabrics and synthetic materials, both C. albicans and C. parapsilosis can survive for several days . The adhesion capacity of diverse candida species to different surfaces presumably plays a major role in the pathogenesis of human colonization and invasion  as well as in the establishment and perpetuation of infection . As a result, different Candida species can contaminate similar environmental reservoirs and cause human infection through similar pathophysiological processes. The isolation of diverse candida species from clinical and environmental sources in our study is therefore not in disagreement with available knowledge. For at least certain synthetic materials, the survival time under dry conditions may be longer for C. parapsilosis than for C. albicans, , which may at least partially explain the predominance of C. parapsilosis in our environmental samples.
In conclusion, we believe that candida contamination of non-sterile cotton, plastic transportation bags and plastic IV medication containers, in conjunction with poor infection control practices, facilitated the occurrence of several candida BSIs, predominantly in patients with established clinical predisposition.
Funding for this study was provided by the Clínica Jorge Piñeros Corpas, Bogotá, Colombia. Presented in part at the 43rd Annual Meeting of the Infectious Diseases Society of America, San Francisco, CA, October 6-9, 2005 [Abstract 1145].
Received on 5 April 2008; revised 21 August 2008.
- 1. Pfaller M.A., Diekema D.J. Epidemiology of invasive candidiasis: a persistent public health problem. Clin Microbiol Rev 2007;20:133-63.
- 2. Pfaller M.A., Jones R.N., Messer S.A., et al. National surveillance of nosocomial blood stream infection due to species of Candida other than Candida albicans: frequency of occurrence and antifungal susceptibility in the SCOPE Program. SCOPE Participant Group. Surveillance and Control of Pathogens of Epidemiologic. Diagn Microbiol Infect Dis 1998;30:121-9.
- 3. Pfaller M.A., Jones R.N., Doern G.V., et al. Bloodstream infections due to Candida species: SENTRY antimicrobial surveillance program in North America and Latin America, 1997-1998. Antimicrob Agents Chemother 2000;44:747-51.
- 4. Pfaller M.A., Jones R.N., Doern G.V., Kugler K. Bacterial pathogens isolated from patients with bloodstream infection: frequencies of occurrence and antimicrobial susceptibility patterns from the SENTRY antimicrobial surveillance program (United States and Canada, 1997). Antimicrob Agents Chemother 1998;42:1762-70.
- 5. Perlroth J., Choi B., Spellberg B. Nosocomial fungal infections: epidemiology, diagnosis, and treatment. Med Mycol 2007;45:321-46.
- 6. Gudlaugsson O., Gillespie S., Lee K., et al. Attributable mortality of nosocomial candidemia, revisited. Clin Infect Dis 2003;37:1172-7.
- 7. Hajjeh R.A., Sofair A.N., Harrison L.H., et al. Incidence of bloodstream infections due to Candida species and in vitro susceptibilities of isolates collected from 1998 to 2000 in a population-based active surveillance program. J Clin Microbiol 2004;42:1519-27.
- 8. Nedret Koc A., Kocagoz S., Erdem F., Gunduz Z. Outbreak of nosocomial fungemia caused by Candida glabrata. Mycoses 2002;45:470-5.
- 9. Zancope-Oliveira R.M., James M.J., Derossi A.P., et al. Strain characterization of Candida parapsilosis fungemia by molecular typing methods. Eur J Clin Microbiol Infect Dis 2000;19:514-20.
- 10. Welbel S.F., McNeil M.M., Kuykendall R.J., et al. Candida parapsilosis bloodstream infections in neonatal intensive care unit patients: epidemiologic and laboratory confirmation of a common source outbreak. Pediatr Infect Dis J 1996;15:998-1002.
- 11. Bennett S.N., McNeil M.M., Bland L.A., et al. Postoperative infections traced to contamination of an intravenous anesthetic, propofol. N Engl J Med 1995;333:147-54.
- 12. Sherertz R.J., Gledhill K.S., Hampton K.D., et al. Outbreak of Candida bloodstream infections associated with retrograde medication administration in a neonatal intensive care unit. J Pediatr 1992;120:455-61.
- 13. Huang Y.C., Lin T.Y., Leu H.S., et al. Outbreak of Candida parapsilosis fungemia in neonatal intensive care units: clinical implications and genotyping analysis. Infection 1999;27:97-102.
- 14. Diekema D.J., Messer S.A., Hollis R.J., et al. An outbreak of Candida parapsilosis prosthetic valve endocarditis. Diagn Microbiol Infect Dis 1997;29:147-53.
- 15. Levin A.S., Costa S.F., Mussi N.S., et al. Candida parapsilosis fungemia associated with implantable and semi-implantable central venous catheters and the hands of healthcare workers. Diagn Microbiol Infect Dis 1998;30:243-9.
- 16. Clark T.A., Slavinski S.A., Morgan J., et al. Epidemiologic and molecular characterization of an outbreak of Candida parapsilosis bloodstream infections in a community hospital. J Clin Microbiol 2004;42:4468-72.
- 17. Masala L., Luzzati R., Maccacaro L., et al. Nosocomial cluster of Candida guillermondii fungemia in surgical patients. Eur J Clin Microbiol Infect Dis 2003;22:686-8.
- 18. Colombo A.L., Guimarães T. [Candiduria: a clinical and therapeutic approach] Rev Soc Bras Med Trop 2007;40:332-7.
- 19. Toya S.P., Schraufnagel D.E., Tzelepis G.E. Candiduria in intensive care units: association with heavy colonization and candidaemia. J Hosp Infect 2007;66:201-6.
- 20. Vazquez J.A., Boikov D., Boikov S.G., Dajani A.S. Use of electrophoretic karyotyping in the evaluation of Candida infections in a neonatal intensive-care unit. Infect Control Hosp Epidemiol 1997;18:32-7.
- 21. Chowdhary A., Becker K., Fegeler W., et al. An outbreak of candidemia due to Candida tropicalis in a neonatal intensive care unit. Mycoses 2003;46:287-92.
- 22. Cole G.T., Halawa A.A., Anaissie E.J. The role of the gastrointestinal tract in hematogenous candidiasis: from the laboratory to the bedside. Clin Infect Dis 1996;22 Suppl 2:S73-88.
- 23. Sanchez V., Vazquez J.A., Barth-Jones D., et al. Nosocomial acquisition of Candida parapsilosis: an epidemiologic study. Am J Med 1993;94:577-82.
- 24. Weems J.J., Jr. Candida parapsilosis: epidemiology, pathogenicity, clinical manifestations, and antimicrobial susceptibility. Clin Infect Dis 1992;14:756-66.
- 25. Traore O., Springthorpe V.S., Sattar S.A. A quantitative study of the survival of two species of Candida on porous and non-porous environmental surfaces and hands. J Appl Microbiol 2002;92:549-55.
- 26. Rotrosen D., Calderone R.A., Edwards J.E., Jr. Adherence of Candida species to host tissues and plastic surfaces. Rev Infect Dis 1986;8:73-85.
- 27. Klotz S.A., Drutz D.J., Zajic J.E. Factors governing adherence of Candida species to plastic surfaces. Infect Immun 1985;50:97-101.
- 28. Shoham S., Shaffer R., Sweet L., et al. Candidemia in patients with ventricular assist devices. Clin Infect Dis 2007;44:e9-12.
Publication in this collection
09 Feb 2009
Date of issue
05 Apr 2008
21 Aug 2008