Invasive fungal infections and (1,3)-β-d-glucan serum concentrations in long-term intensive care patients

Article Outline

• Summary
• Introduction
• Methods
  • Patients
  • Microbiology and serology
  • Statistics
• Results
• Discussion
• References
• Copyright

Summary 

Objective

Invasive fungal infections are associated with high morbidity and increased mortality. This study was performed to assess the epidemiology of fungal infections and to determine (1,3)-β-d-glucan serum concentrations in patients admitted to intensive care units (ICUs).

Patients and methods

Overall 197 patients were admitted to nine medical and surgical intensive care units (ICUs) at a 2200-bed university hospital during a 3-month period. Retrospectively, the patients were split into three groups: group A comprised 24 patients with proven invasive fungal infections admitted for a median of 40 days. Group B comprised 58 patients who were admitted to the ICU for 30 days but without fungal infection. One hundred and fifteen post-operative patients served as controls (group C). The levels of (1,3)-β-d-glucan were monitored in all patients twice weekly during their ICU admittance.

Results

Average (1,3)-β-d-glucan concentrations were significantly higher in the patients with fungal infections compared to group B and group C (median 44 vs. 22 and 12.9pg/ml, respectively; p<0.001). For a serum (1,3)-β-D-glucan level of 40 pg/ml, the sensitivity, the specificity, the positive predictive value, the negative predictive value, the area under the curve of the receiver operating characteristics (AUC ROC) curve, the likelihood ratio (LR)+ and LR− were 52.2, 75.9, 46.2, 80, 0.7, 2.16, and 0.63, respectively, on day 7. Patients in group A had bacterial infections significantly more often than patients in group B (p=0.003). The hospitalization before ICU admittance for group A was significantly longer than for groups B and C (median 19 (group A) vs. 6 (group B) vs. 10 (group C) days; p<0.05).

Conclusions

Longer hospitalization and multiple bacterial infections were found to be the main risk factors for invasive fungal infections. Long-term ICU patients have elevated (1,3)-β-d-glucan levels, not only due to invasive fungal infections, but also due to the serious underlying diseases and conditions, inter-current complications, and intensive care measures. Yet, persistently high serum levels of (1,3)-β-d-glucan in ICU patients may be indicative of invasive fungal infections and warrant additional diagnostic efforts.

Keywords: Fungal infection, Diagnosis, (1,3)-β-d-Glucan, Intensive care unit

Back to Article Outline

Introduction 

Invasive fungal infections are of increasing relevance for severely ill and immunocompromised patients. The attributable mortality is up to 50%.¹, ² Given the high mortality of invasive infections, early and adequate treatment is vital. Candida species are the predominant pathogens causing invasive disease in intensive care, however invasive disease caused by Aspergillus in non-neutropenic intensive care unit (ICU) patients is increasing.³ The basic problem associated with adequate treatment of fungal infections in ICU patients is accurate diagnosis, with blood cultures being notoriously insensitive. There have been intensive efforts to identify markers of fungal invasion, including serological methods to detect surface antigens, metabolites, or antibodies against fungal antigens.⁴ Both, the sensitivity and specificity of these methods are highly variable and depend on the patients studied, but in short depend on the test and patient population investigated.⁴ The (1,3)-β-d-glucan assay (which is based on the recognition of (1,3)-β-d-glucan, a component of the cell wall of Candida and Aspergillus, by the innate immune system of horseshoe crabs⁵), was developed about 10 years ago. Given the limits of the present diagnostic armamentarium, this test may be useful for the diagnosis of invasive Candida and Aspergillus infections in ICU patients.

In this observational study, the clinical and epidemiological data of patients admitted for ≥7 days to the ICU were analyzed for the presence of invasive fungal infections. To evaluate the usefulness of (1,3)-β-d-glucan levels in serum for the detection of invasive fungal infections, and for invasive Candida infection in particular, (1,3)-β-d-glucan concentrations were measured sequentially during the patients’ admittance to the ICU.

Back to Article Outline

Methods 

The study was conducted from February 2004 to April 2004. Patients from nine intensive care units were enrolled, including units of neurosurgery, trauma surgery, solid organ transplantation, medical cardiology, cardiothoracic surgery, and general abdominal surgery/anesthesiology. After ≥7 days of admission in the ICU, patients were screened for the presence of invasive fungal infection using the Candida hemagglutination test. Patient sera were also tested for the presence of (1,3)-β-d-glucan using the commercially available Glucatell test (Cape Cod, USA).

Patients 

During the observation period, 197 patients were admitted to the ICUs. The patients were retrospectively classified into three groups based on the presence of an invasive fungal infection and the length of admittance. Group A (N=24) included patients with severe underlying diseases, extensive surgery and other underlying conditions, and a stay of at least 7 days. These patients had proven invasive fungal infections with growth of fungi in blood cultures and abscesses or aspirates from otherwise sterile compartments (pleura, cerebrospinal fluid). All patients from group A received antifungal treatment. Group B (N=58) included patients with severe underlying diseases and extensive surgery and other underlying conditions, who were admitted to the ICU for at least 7 days, but did not have evidence of fungal infection based on the clinical and microbiological data. Group B patients did not receive antifungal therapy. Group C (N=115) comprised post-operative patients without evidence of fungal infection admitted to the ICUs for post-operative monitoring for ≤7 days.

The following data were collected from the patient charts: age, gender, underlying disease, reason for ICU admission, concomitant infections, presence and duration of risk factors for Candida colonization and infection, antifungal treatment, and vital status at discharge (survival vs. death).

Microbiology and serology 

Specimens for microbiological cultures were taken from peripheral blood, aspirates of abscesses, intravascular lines, wound exudates, surgical drains, or other infectious foci, if clinically indicated. Specimens were processed at the clinical microbiology laboratory for the detection of Candida according to standard procedures. These included inoculating onto ChromAgar (Becton & Dickinson, Germany) and processing of blood cultures using an automated system (BacTAlert FA and FN, BioMerieux, France). Identification of yeasts to the species level was carried out using the API 32ID whenever possible.

During the 3-month period, blood samples for Candida serology were collected twice weekly in endotoxin-free tubes (Greiner) to avoid endotoxin contamination, and serum was frozen at −72°C and examined in bulk using the (1,3)-β-d-glucan assay (Glucatell). The assay was performed according to the protocol supplied by the manufacturer on the Internet (2004). Two procedures can be used to measure the (1,3)-β-d-glucan, the end-point assay and the kinetic rate assay, and in this study we used the end-point assay. To detect Candida antibodies the Candida hemagglutination test (Hemkit Candida IHA, Ravo Diagnostika, Germany) was used. To ascertain reproducibility of the tests, controls were run with each test.

Statistics 

Significance of difference was assessed by means of the Chi-square test for categorical variables and the Wilcoxon rank sum test for continuous variables. For subgroups and non-normally distributed variables the Mann–Whitney U test was used. All tests were two-sided, and p<0.05 was considered significant. The following indices of diagnostic performance and their 95% confidence intervals (CI) were calculated: sensitivity, specificity, the positive predictive value (PPV), the negative predictive value (NPV), likelihood ratio test (LR) positivity (LR+), and LR negativity (LR−). LR+ and LR− are independent of prevalence, as are sensitivity and specificity, and are therefore increasingly used to evaluate diagnostic test performance. The predictive values of the LR+/LR− are: >10/<0.1 very good, 5–10/0.1–0.2 useful, 2–5/0.2–0.5 medium, and 1–2/0.5–1 negligible. The discriminatory power of the (1,3)-β-d-glucan levels in serum was evaluated by the area under the receiver operating characteristics (AUC ROC) curve and the 95% CI. Descriptive and correlation statistics were obtained using SPSS for Windows, release 13 (SPSS Inc., Chicago, USA).

Back to Article Outline

Results 

During the 3-month surveillance period, 24 patients with invasive fungal infections (22 with Candida infections and two with invasive aspergillosis; group A) and 58 patients without fungal infection who were admitted to the ICU for at least 7 days (group B) were identified. Group C consisted of 115 post-operative patients (71 males, 44 females) who had undergone cardiac surgery (n=29), abdominal surgery (n=40), orthopedic surgery (n=19), or neurosurgery (n=27). Demographic data are given in Table 1. The overall incidence of fungal infections during the 3-month surveillance period was 12.2%.

Table 1. Demography and predisposing conditions for invasive fungal infections

	Group A	Group B	Group C	p-Value
Number of patients	24	58	115	-
Male/female	17/7	39/19	71/44	NS
Overall mortality	6/24 (25%)	10/58 (17.2%)	3/115 (2.6%)	0.003
Temperature, median (range)	37.9°C (35.3–40.1)	37.8°C (36.1–40.3)	37.1°C (35.7–38.9)	NS
Admittance to ICU, median (range)	40 days (30–246)	30 days (7–167)	-	NS
Hospitalization before, median (range)	19 days (4–107)	6 days (1–69)	10 days (1–66)	0.047

Groups A and B, long-term ICU patients; group C, post-operative patients; ICU, intensive care unit; NS, not significant; OR, odds ratio; 95% CI, 95% confidence interval.

Patients in group A had bacterial infections significantly more often than patients in group B (22/24 patients in group A vs. 30/58 patients in group B; p=0.003; Table 2). Although there was no difference in the number of patients receiving broad-spectrum antimicrobial therapy (A, n=24/24; B, n=37/58), group A received a significantly greater number of different antimicrobial agents then group B (mean 3.8/patient vs. 0.83/patient; p<0.05). There was no difference between group A and group B with regard to the underlying diseases and predisposing conditions for invasive fungal infections. The underlying diseases comprised primarily trauma, cardiovascular disease, and organ transplantation (Table 2). There was no significant difference in the incidence of conditions considered risk factors for invasive fungal infections (Table 1) between groups A and B.

Table 2. Underlying diseases und infections in the long-term intensive care unit patient groups A and B

	Group A (N=24)	Group B (N=58)
Underlying diseases
Cardiovascular disease	3	7
Trauma	6	16
Acute respiratory distress syndrome	1	2
Solid neoplasm	5	5
AIDS	1	0
Past organ transplantation	0	2
Recent organ transplantation	5	10
Cerebral hemorrhage	2	12
Liver disease	2	9
Lung disease	4	4
Leukemia/lymphoma	2	1
Renal disease	0	1
Bacterial infections	22	30
Pneumonia	9	10
Septicemia	4	8
Osteomyelitis	3	3
Catheter-related bacteremia	3	1
Peritonitis	3	1
Urinary tract infection	3	1
Cholangitis	1	1
Abscess	2	1
Fasciitis	1	0
Wound infection	0	1
Abscess	1	0

In group A, the fungal infections comprised candidemia (n=11) and hepatic candidiasis (n=9), Candida peritonitis (n=2), and invasive aspergillosis of the lung (n=2). The fungal pathogens were Candida albicans (n=14), Candida glabrata (n=4), Candida tropicalis (n=2), Candida krusei (n=2), and Aspergillus fumigatus (n=2). All patients received antifungal treatment comprising fluconazole (n=13), caspofungin (n=6), amphotericin B (n=4), and voriconazole (n=3). Three patients received a combination with amphotericin B and caspofungin, and two patients received fluconazole and caspofungin sequentially. Nine patients died within 14 days after the onset of the fungal infection and 15 patients (62%) survived.

The median (1,3)-β-d-glucan concentrations were significantly higher in the group A patients with fungal infections compared to group B and group C (median 44 vs. 22 and 12.9 pg/ml, respectively; p<0.001) (Table 3). The (1,3)-β-d-glucan concentrations remained significantly higher in group A than in group B over 7–10 days, but then decreased to a similar level (Table 3, Figure 1). For a serum (1,3)-β-d-glucan level of 40 pg/ml, the sensitivity, the specificity, the positive predictive value, and the negative predictive value of the (1,3)-β-d-glucan assay used were 52.2, 75.9, 46.2, and 80%, respectively, on day 7 after admission to the ICU. The descriptive parameters for the efficiency of (1,3)-β-d-glucan levels ≥40 pg/ml to detect a fungal infection are given in Table 3. The LR+ ranged from 2.16 to 3.09, the LR−, however, was persistently greater than 0.5. On days 7, 10, and 14, the AUC ROC curve was greater than 0.7 indicating an acceptable association of increasing (1,3)-β-d-glucan serum levels with the presence of invasive fungal infection (p=0.005).

Table 3. (1,3)-β-d-Glucan serum levels and presumed efficiency parameters of (1,3)-β-d-glucan serum concentrations ≥40 pg/ml in long-term ICU patients with proven fungal infections (group A, N=24) and in long-term ICU patients without fungal infections (group B, N=58)

ICU, intensive care unit; CI, confidence interval; PPV, positive predictive value; NPV, negative predictive value; LR+, likelihood ratio positivity; LR−, likelihood ratio negativity; AUC ROC, area under the receiver operating characteristics curve.

• 
  • View Large Image
  • Download to PowerPoint

• Figure 1. 

  Mean concentrations of (1,3)-β-d-glucan over time (squares: group A, patients with fungal infections; circles: group B, patients without fungal infections).

On day 7, 14 patients from group B (without fungal infections) had (1,3)-β-d-glucan levels ≥40 pg/ml: six patients had concurrent bacterial infections including pneumonia (n=2; Streptococcus pneumoniae, Staphylococcus aureus), osteomyelitis (S. aureus), peritonitis, and catheter-associated bacteremia (n=2; Staphylococcus epidermidis). The (1,3)-β-d-glucan levels were not persistent, and by day 14, eight other patients from group B had (1,3)-β-d-glucan levels ≥40 pg/ml. Two of these patients had catheter-associated bacteremia with S. epidermidis. In one patient, Pneumocystis jirovecii pneumonia was suspected and treated after the study. The other patients had no proven bacterial infections. All of these patients had central venous catheters and various other invasive ICU devices and gear.

With regard to the Candida hemagglutination titer, there was no difference between the three patient groups on day 7 (group A: 0 (0–20 480); group B: 0 (0–2560); group C 0 (0–10 240); median (range)) and between the two long-term ICU patient groups A and B on days 10, 14, and 17 [group A: day 10, 0 (0–1280); day 14, 0 (0–1280); day 17, 0 (0–1280); group B: day 10, 0 (0–2560); day 14, 0 (0–2560); day 17, 0 (0–2560); median (range)].

Back to Article Outline

Discussion 

Compared to most studies evaluating (1,3)-β-d-glucan levels at the time of infection only, the present study has documented the course of (1,3)-β-d-glucan levels in a ‘normal’, heterogeneous ICU population during the period of their admittance using the (1,3)-β-d-glucan research test. At 7 days post-admission, patients with acute invasive Candida infections had significantly higher serum concentrations of (1,3)-β-d-glucan than patients without fungal infections or the short-term ICU patients. In patients with fungal infections the (1,3)-β-d-glucan levels were persistently high, and decreased slowly (Figure 1). However, (1,3)-β-d-glucan levels of long-term ICU patients without fungal infections also increased starting at day 14, but importantly not to serum levels detected in patients with proven fungal infections (Figure 1). There may be several explanations: Elevated concentrations of (1,3)-β-d-glucan have been reported to be associated with Gram-positive infections, biofilms on vascular catheters, hemodialysis, or the administration of intravenous immunoglobulins.⁶, ⁷, ⁸ Out of the long-term ICU patients without fungal infections, five had hemofiltration, another five had Gram-positive infections due to S. aureus or S. pneumoniae, and four patients had catheter-associated bacteremia due to coagulase-negative staphylococci. None of them received immunoglobulins. After the end of the study, a patient was treated for Pneumocystis pneumonia suspected because of his underlying immunodeficiency syndrome. Only recently, Persat et al. described the usefulness of determining (1,3)-β-d-glucan in the diagnosis of Pneumocystis pneumonia patients.⁹ These interactions may have influenced the sensitivity, specificity, PPV, NPV, LR+, and LR−. Moreover, Candida colonization and occult candidemia cannot be totally excluded however evaluation of Candida colonization was not systematically performed or documented limiting the information obtained in a retrospective study.

The diagnosis of invasive Candida infections in ICU patients is still an enigma. Scores to quantitate Candida colonization for the decision to start antifungal treatment at the bedside may be helpful, but are inaccurate tools for the diagnosis of invasive Candida infection.¹⁰, ¹¹ Methods under investigation for the diagnosis of invasive fungal infections involve the detection of fungal antigens or metabolites, or PCR-based assays.⁴, ¹² However, assays for Candida metabolites, mannan and enolase, and antibodies against Candida have never proved to be fully diagnostic for the invasive Candida infection.¹³, ¹⁴, ¹⁵ Increased (1,3)-β-d-glucan serum levels have been reported to be very useful in the diagnosis of experimental candidiasis and various fungal infections in humans.¹⁶, ¹⁷ The study of Digby et al. comparing a small number of patients with Candida infection and patients with bacterial infections, described a low specificity of the (1,3)-β-d-glucan assay.⁶ However, in patients with hematological malignancies and transplant recipients, monitoring of (1,3)-β-d-glucan antigenemia may be a useful method for the early diagnosis of invasive fungal infections including even aspergillosis.¹⁸, ¹⁹, ⁸

The patient population admitted at these ICUs was rather heterogeneous. The high incidence of fungal infections in almost one third of the long-term ICU patients may be explained by the long pre-ICU admittance of the patients and serious underlying diseases necessitating ICU stays of up to 246 days. For early diagnosis of invasive fungal infections, additional non-invasive methods may be useful. But this study highlights the difficulties in the interpretation of increased (1,3)-β-d-glucan levels and the importance of evaluating this serological method in the ICU patient population. In high-risk ICU patients, when results of cultures of blood or sterile body sites are still pending or negative, high serum concentrations of (1,3)-β-d-glucan may be another useful piece of evidence in the diagnosis of invasive fungal infections.

Conflict of interest: No conflict of interest to declare.

Back to Article Outline

References 

1. Wenzel RP. Nosocomial candidemia: risk factors and attributable mortality. Clin Infect Dis. 1995;20:1531–1534
   • View In Article
   • MEDLINE
   • CrossRef
2. Leleu G, Aegerter P, Guidet B. Systemic candidiasis in intensive care units: a multicenter, matched-cohort study. J Crit Care. 2002;17:168–175
   • View In Article
   • Abstract
   • Full Text
   • Full-Text PDF (95 KB)
   • CrossRef
3. Meersseman W, Vandecasteele SJ, Wilmer A, Verbeken E, Peetermans WE, Van Wijngaerden E. Invasive Aspergillosis in Critically Ill Patients without Malignancy. Am J Respir Crit Care Med. 2004;170:621–625
   • View In Article
   • MEDLINE
   • CrossRef
4. Yeo SF, Wong B. Current status of nonculture methods for diagnosis of invasive fungal infections. Clin Microbiol Rev. 2002;15:465–484
   • View In Article
   • MEDLINE
   • CrossRef
5. Iwanaga S, Miyata T, Tokunaga F, Muta T. Molecular mechanism of hemolymph clotting system in Limulus. Thromb Res. 1992;68:1–32
   • View In Article
   • Abstract
   • Full-Text PDF (2365 KB)
   • CrossRef
6. Digby J, Kalbfleisch J, Glenn A, Larsen A, Browder W, Williams D. Serum glucan levels are not specific for presence of fungal infections in intensive care unit patients. Clin Diagn Lab Immunol. 2003;10:882–885
   • View In Article
   • MEDLINE
7. Nett J, Lincoln L, Marchillo K, Andes D. Beta-1,3 glucan as a test for central venous catheter biofilm infection. J Infect Dis. 2007;195:1705–1712
   • View In Article
   • MEDLINE
   • CrossRef
8. Pickering JW, Sant HW, Bowles CA, Roberts WL, Woods GL. Evaluation of a (1→3)-beta-d-glucan assay for diagnosis of invasive fungal infections. J Clin Microbiol. 2005;43:5957–5962
   • View In Article
   • MEDLINE
   • CrossRef
9. Persat F, Ranque S, Derouin F, Michel-Nguyen A, Picot S, Sulahian A. Contribution of the (1→3)-beta-d-glucan assay for diagnosis of invasive fungal infections. J Clin Microbiol. 2008;46:1009–1013
   • View In Article
   • CrossRef
10. Pittet D, Monod M, Suter PM, Frenk E, Auckenthaler R. Candida colonization and subsequent infections in critically ill surgical patients. Ann Surg. 1994;220:751–758
    • View In Article
    • MEDLINE
    • CrossRef
11. Leon C, Ruiz-Santana S, Saavedra P, Almirante B, Nolla-Salas J, Alvarez-Lerma F, et al. A bedside scoring system (‘Candida score’) for early antifungal treatment in non-neutropenic critically ill patients with Candida colonization. Crit Care Med. 2006;34:730–737
    • View In Article
    • MEDLINE
    • CrossRef
12. Bille J, Marchetti O, Calandra T. Changing face of health-care associated fungal infections. Curr Opin Infect Dis. 2005;18:314–319
    • View In Article
    • MEDLINE
13. Yera H, Sendid B, Francois N, Camus D, Poulain D. Contribution of serological tests and blood culture to the early diagnosis of systemic candidiasis. Eur J Clin Microbiol Infect Dis. 2001;20:864–870
    • View In Article
    • MEDLINE
    • CrossRef
14. Sendid B, Caillot D, Baccouch-Humbert B, Klingspor L, Grandjean M, Bonnin A, et al. Contribution of the Platelia Candida-specific antibody and antigen tests to early diagnosis of systemic Candida tropicalis infection in neutropenic adults. J Clin Microbiol. 2003;41:4551–4558
    • View In Article
    • MEDLINE
    • CrossRef
15. van Deventer A, van Vliet H, Hop WC, Goessens WH. Diagnostic value of anti-Candida enolase antibodies. J Clin Microbiol. 1994;32:17–23
    • View In Article
    • MEDLINE
16. Odabasi Z, Mattiuzzi G, Estey E, Kantarjian H, Saeki F, Ridge RJ, et al. Beta-d-glucan as a diagnostic adjunct for invasive fungal infections: validation, cutoff development, and performance in patients with acute myelogenous leukemia and myelodysplastic syndrome. Clin Infect Dis. 2004;39:199–205
    • View In Article
    • CrossRef
17. Obayashi T, Yoshida M, Mori T, Goto H, Yasuoka A, Iwasaki H, et al. Plasma (1→3)-beta-d-glucan measurement in diagnosis of invasive deep mycosis and fungal febrile episodes. Lancet. 1995;345:17–20
    • View In Article
    • Abstract
    • CrossRef
18. Ostrosky-Zeichner L, Alexander BD, Kett DH, Vazquez J, Pappas PG, Saeki F, et al. Multicenter clinical evaluation of the (1→3) beta-d-glucan assay as an aid to diagnosis of fungal infections in humans. Clin Infect Dis. 2005;41:654–659
    • View In Article
    • CrossRef
19. Senn L, Robinson JO, Schmidt S, Knaup M, Asahi N, Satomura S, et al. 1,3-Beta-d-glucan antigenemia for early diagnosis of invasive fungal infections in neutropenic patients with acute leukemia. Clin Infect Dis. 2008;46:878–885
    • View In Article
    • CrossRef