Special Feature
Update on Nosocomial Candida Bloodstream Infections
By Jun Takezawa, MD
Nosocomial infections involve 2 million patients per year in the United States, and approximately 250,000 patients develop bloodstream infections (BSI) such as bacteremia and fungemia. Crude mortality from BSI ranges from 12% to 80% and attributable mortality averages 26%. The microorganisms associated with BSI are coagulase-negative staphylococci (32.4%), Staphylococcus aureus (16.7%), Enterococcus spp. (11.7%), Candida spp. (8.0%), and E. coli (6.4%).
BSIs occur 2-7 times more frequently in ICU patients than in patients in general wards. According to the study by Pittet and associates,1 nosocomial BSIs occur in 2.7 per 100 admissions to the surgical ICU, and their case-control analysis revealed that crude mortality rates from the cases and controls were 50% and 15%, respectively. Attributable mortality was 35% (95% CI, 25-45%). The median length of hospital stay was 40 and 26 days for cases and controls, respectively. The extra cost attributable to BSI averaged $40,000 per survivor. The most common pathogens identified in the ICU were surveyed by NNIS and are shown in Table 1.
| Table 1-Eight Most Common Pathogens Associated with Nosocomial Infections in ICU Patients, NNIS January 1989-July 1998 | ||||||
| Relative Percent of Site of Infection | ||||||
| All site n = 235,753 | BSI n = 50,091 | Pneu n = 64,056 | UTI n = 47,502 | SSI n = 22,043 | Other N=52,066 | |
| CNS | 14.3 | 39.3 | 2.5 | 3.1 | 13.5 | 15.4 |
| S. aureus | 11.4 | 10.7 | 16.8 | 1.6 | 12.6 | 13.7 |
| P. aeuginosa | 9.9 | 3.0 | 16.1 | 10.6 | 9.2 | 8.7 |
| Enterococcus spp. | 8.1 | 10.3 | 1.9 | 13.8 | 14.5 | 5.9 |
| Enterobacter spp. | 7.3 | 4.2 | 10.7 | 5.7 | 8.8 | 6.8 |
| E. coli | 7.0 | 2.9 | 4.4 | 18.2 | 7.1 | 4.0 |
| C. albicans | 6.6 | 4.9 | 4.0 | 15.3 | 4.8 | 4.3 |
| K. pneumoniae | 4.7 | 2.9 | 6.5 | 6.1 | 3.5 | 3.5 |
| Others | 30.7 | 21.8 | 37.1 | 25.6 | 26.0 | 37.7 |
| Total | 100 | 100 | 100 | 100 | 100 | 100 |
| Pneu = pneumonia, UTI = urinary tract infection, SSI = surgical site infection. | ||||||
Epidemiology of Bloodstream Infections Caused by Candida
Nosocomial fungemia due to Candida spp. accounts for 8-10% of all BSIs, at a rate of 5-10 per 10,000 hospital admissions, and provides a crude mortality of 50% or more with an attributable mortality of 35-50%. The frequency of BSIs caused by Candida spp. increased by 500% (1.0-4.9/10.000 discharge) from 1980 to 1992, according to the NNIS report. The Surveillance and Control of Pathogens of Epidemiological Importance (SCOPE) study showed that the Candida spp. are the fourth most common isolates in the bloodstream,2 and 90% of the C. albicans isolates causing BSI were susceptible to fluconazole and itraconazole. However, there was a regional difference in susceptibility, where a decreased susceptibility was found in the Southeast and Northeast regions of the United States.
Pittet et al analyzed BSIs caused by Candida spp.3 and reported that Candida spp. were the only microorganisms that independently affected outcome of BSIs (OR 1.84, 95% CI 1.22-2.76). Pittet et al found that the attributable mortality rate was the highest (35% at 28 days, and 69% at discharge) in BSI patients with Candida spp. Lewis and colleagues also reported that the mortality rate of BSIs due to candidemia was 38% (CI 26-49%), a risk ratio of 2.94 (95% CI 1.95-4.43), and a median increase in hospital stay of 30 days for survivors.4
Nguyen and colleagues5 conducted a multicenter surveillance in the United States and reported a significant increase in non-albicans candidemia from 1990 to 1994, especially BSIs caused by C. glabrata. They also reported a decreased susceptibility of C. albicans to fluconazole and itraconazole, and also increased MICs of fluconazole for non-Candida albicans.
The trend of species distribution among bloodstream isolates of Candida spp. was evaluated by Pfaller,6 where a total of 1579 BSI candidal isolates were included from 50 U.S. medical centers. As shown in Table 2, the proportion of C. albicans has gradually increased from 1992 to 1998, and C. glabrata was the second most common species associated with the BSIs. However, this is not a population-based surveillance.
| Table 2-Species Distribution of Candidal Bloodstream Isolates During 1992-1998 | |||||||
| 1992 (230) | 1993 (313) | 1995 (404) | 1996 (130) | 1997 (383) | 1998 (119) | All Years (1579) | |
| C. albians | 45 | 46 | 52 | 54 | 57 | 60 | 52 |
| C. glabrata | 17 | 14 | 21 | 16 | 16 | 24 | 18 |
| C. parapsilosis | 20 | 24 | 10 | 11 | 15 | 7 | 15 |
| C. tropicalis | 12 | 12 | 12 | 16 | 9 | 7 | 11 |
| C. krusei | 3 | 1 | 4 | 3 | 2 | - | 2 |
| Candida spp. | 3 | 3 | 1 | - | 1 | 2 | 2 |
An international surveillance of BSIs due to Candida spp. (SENTRY program) was carried out by Pfaller et al between January and December 1997, in which the study included the United States, Canada, and South America, and detected 306 episodes of candidemia in 34 medical centers.7 Eighty percent of the BSIs were nosocomial and 50% occurred in ICU patients. The incidence of BSIs caused by various species of Candida is shown in Table 3.
| Table 4-Sensitivity and Specificity of Serodiagnostic Tests | ||||||
| Enlose antigen | Cand-Tec | Mannan antigen | Beta-glucan | D-arabitol | ||
| Sensitivity | 100 | 76.9 | 25.6 | 84.4 | 29 | |
| Specificity | 71.8 | 87.5 | 100 | 87.5 | 00 | |
As shown in Table 3, the frequency of BSIs caused by non-C. albicans was different among the three countries. This difference may be caused by factors such as primary disease, cytotoxic chemotherapy, and the use of antifungal agents. It was found that C. albicans and C. tropicalis developed BSIs in patients who had not received fluconazole, while C. glabrata and C. krusei affected patients who had received fluconazole. BSIs caused by C. parapsilosis were associated with the placement of central venous catheters (hands of health care workers) and not related to the antifungal agents.
Therefore, the species difference among the countries and among the institutes was associated with antifungal use and infection control practice. Among the different species of Candida, resistance to fluconazole (MIC > 64 mcg/mL) and itraconazole (MIC > 1.0 mcg/mL) was observed with C. glabrata and C. krusei. Isolates of C. albicans, C. parapsilosis, C. tropicalis, and C. gilliermondii were susceptible to both fluconazole (99.4-100%) and iotraconazole (95.8-100%). On the other hand, 8.7% of C. glabrata and 100% of C. krusei isolates were resistant to fluconazole and 36.9% of C. glabrata isolates and 66.6% of C. krusei isolates were resistant to itraconazole.
| Table 3-Species Distribution of Candida Bloodstream Infections (SENTRY Program) | ||||||
| USA (203) | Canada (61) | South America (42) | Overall (306) | |||
| C. albicans | 56.2 | 52.5 | 40.5 | 53.3 | ||
| C. glabrata | 18.7 | 11.5 | 2.4 | 15.0 | ||
| C. parapsilosis | 8.9 | 22.9 | 38.1 | 15.7 | ||
| C. tropicalis | 6.9 | 8.2 | 11.9 | 7.8 | ||
| C. krusei | 2.5 | 1.6 | — | 2.0 | ||
| C. guilliermondii | 0.5 | — | 2.4 | 0.7 | ||
| Other Candida spp. | 6.4 | 3.3 | 4.7 | 5.8 | ||
The importance of population-based surveillance on organisms that cause nosocomial infection cannot be overemphasized in prevention of nosocomial infections by watching the overall trend of drug-resistant infections and comparing surveillance data of the given hospital with benchmarking.
Risk Factors for the Development of BIS by Candida spp.
Granulocytopenia and/or damage to oropharyngeal mucosa through aggressive use of cytotoxic drugs in cancer patients and long-term use of broad-spectrum antibiotics have been described as an independent risk factor for the development of candidemia. Other risk factors for the development of candidal BSIs by using multivariate analysis were age, azotemia, colonization, graft vs. host disease, hemodialysis, hyperglycemia, parenteral nutrition, steroids, and surgery (multiple, extended, abdominal).8
Serological Methods for Diagnosing Fungal BSI
Blood cultures for Candida spp. are insensitive indicators unless they are repeated and high-volume cultures are used.9 Although the lysis-centrifugation method is used, several days are still required to make diagnosis. To overcome this delay, serological diagnostic tests for antibody, antigen, and metabolite detection have been developed.
The detection of antibody to Candida poses some limitations because a false-positive result can be obtained in healthy individuals who are exposed to Candida spp. present in the intestinal flora. Additionally, it does not distinguish active from past Candida infections. The imunodominant cytoplasmic 48-kDa antigen of Candida spp. has been recognized as a glycolytic enzyme enlose, which may represent a novel marker of candidemia.10
Mannan is a major component of the fungal cell wall and is the most widely studied antigen. However, sensitivities vary widely, and Pastrex Candida assay does not react with mannan from C. krusei. The Cand-Tec test detects uncharacterized heat-labile fungal antigen. However, it reacts with a nonfungal antigen such as rheumatoid factor. Beta-glucan is another fungal cell wall component and reacts with Factor G, a coagulation enzyme of limulus amoebocyte, causing gelatinization (Fungitec G test). However, false-positive reaction occurs with cellulose-based filter membrane, human immunoglobulin products, and antitumor polysaccharides such as lentinen, sceeroglucan, and schizophyllan.
D-Arabinitol is a major metabolic product of Candida spp. Although it is a quantitative diagnostic marker for candidemia without being affected by gastrointestinal colonization, sensitivity is usually low. Recently, the PCR assay to detect small subunits of the rRNA gene sequence in fungal organisms has been developed. However, as in the serodiagnostic tests, it is impossible to differentiate fungemia from colonization or active and past infections. The quantitative (nested) PCR may solve this limitation. The sensitivity and specificity of those tests are shown in Table 4.11,12
Standard Method for Testing Antifungal Susceptibility
The scope of antifungal resistance has not been well described because tests for antifungal susceptibility have not been standardized and, furthermore, each institute has different antifungal breakpoint concentrations for defining resistance. To reduce interlaboratory variability, the National Committee for Clinical Laboratory Standards (NCCLS) has developed a standard method for testing antifungal susceptibility.13 The guideline for in vitro susceptibility tests for isolates of Candida spp. is shown in Table 5.
| Table 5-Guideline for In Vivo Susceptibility Tests for Candida spp. | ||||||
| Agent | MIC (mcg/mL) Susceptible | S-DD | Intermediate | Resistance | ||
| Fluconazole | < 8 | 16-32 | — | > 64 | ||
| Itraconazole | < 0.125 | 1.25-0.5 | — | > 1 | ||
| Flucytosine | < 4 | — | 8-16 | > 32 | ||
| *MIC = Minimum inhibitory concentration. S-DD = Susceptible-dose-dependent indicates isolate may respond, if high levels of drug can be maintained at the infection site. | ||||||
Perspectives of BSIS Caused by Drug-Resistant Candida spp. Antifungal resistance is characterized as primary (innate) and secondary (acquired) resistance. Primary resistance is intrinsic to Candida spp., such as fluconazole resistance to C. krusei, and amphotericin B and flucytosine resistance to C. lustianiae. Repeated exposure to antifungal agents may select resistant species and replace susceptible species with intrinsically antifungal agent selects resistant species, replacing susceptible species with intrinsically resistant species such as C. krusei or C. glabrata.
The secondary resistance occurs during antifungal treatment. Resistance to amphotericin B is extremely rare; however, resistance of flucytosine is well known when this agent is used alone. Although several mechanisms are proposed to explain the development of secondary resistance, resistance to one azole may not translate into cross-resistance to other azoles or amphotericin B14 because unlike bacteria, fungi are not able to transfer resistance genes from one species to another.15 Therefore, dissemination of antifungal resistance may not occur as rapidly as the spread of antibacterial resistance.
Prudent use of antifungal agents is most effective to prevent the development of BSIs due to drug-resistant candidal organisms as well as the daily practice to prevent nosocomial infection along with hospital-based and population-based surveillance for nosocomial infection due to drug-resistant organisms.
References
1. Pittet D, et al. JAMA 1994;271:1598-1601.
2. Jones RN, et al. Diagn Microbiol Infect Dis 1997; 29:95-102.
3. Pittet D, et al. Clin Infect Dis 1997;24:1068-1078.
4. Lewis RE, Klepser ME. Am J Health-Syst Pharm 1999;56:525-535.
5. Nguyen MH, et al. Am J Med 1996;100:617-623.
6. Pfaller MA, et al. Diagn Microbiol Infect Dis 1999; 33:217-222.
7. Pfaller MA, et al. J Clin Microbiol 1998;36:1886-1889.
8. Verduyn Lunel FM, et al. Diagn Microbiol Infect Dis 1999;34:213-220.
9. Miller PJ, Wenzel RP. J Infect Dis 1987;156:471-477.
10. Walsh TJ, et al. N Engl J Med 1991;324:1026-1031.
11. Mitsutake K, et al. J Clin Microbiol 1996;34:1918-1921.
12. Kohno S, et al. Microbiol Immunol 1993;37:207-212.
13. Approved standard M27A. Wayne, PA: NCCLS;1997.
14. Venkateswarlu K, et al. Antimicrob Agents Chemother 1996;40:2443-2446.
15. Manual of clinical microbiology. Am Soc Microbiol 1995.
Which of the following Candida spp. is most resistant to fluconazole?
a. C. albicans
b. C. glabrata
c. C. parapsilosis
d. C. tropicalis
e. C. krusei
You have reached your article limit for the month. Subscribe now to access this article plus other member-only content.
- Award-winning Medical Content
- Latest Advances & Development in Medicine
- Unbiased Content