Mycopathologia DOI 10.1007/s11046-014-9825-6
Fungal ‘‘colonisation’’ is Associated with Increased Mortality in Medical Intensive Care Unit Patients with Liver Cirrhosis Tobias Lahmer • Marlena Messer • Ulrich Mayr • Bernd Saugel • Sebastian Noe Caroline Schultheiss • Philipp Thies • Christoph Spinner • Simon Nennstiel • Christiane Schwerdtfeger • Veit Phillip • Roland M. Schmid • Wolfgang Huber
•
Received: 20 July 2014 / Accepted: 13 October 2014 Ó Springer Science+Business Media Dordrecht 2014
Abstract Objectives Patients with liver cirrhosis are at increased risk for fungal infections. However, distinction of fungal colonisation (FC) and invasive mycoses is difficult. Aim of this study was to analyse the impact of FC on mortality of cirrhotic ICU-patients. Methods Retrospective mortality analysis of a prospectively maintained database on 120 cirrhotic patients with and without FC. Comparison to 120 noncirrhotic controls matched for APACHE-II (24.9 ± 3.7 vs. 25.0 ± 2.6; p = 0.263). Results About 69/120 (58 %) of patients with cirrhosis had FC. These patients had significantly higher APACHE-II score and mortality compared to cirrhotic patients without FC (27 ± 3 vs. 23 ± 4, p \ 0.001; 78 vs. 35 %, p \ 0.001). In multivariate analysis, FC was independently (p = 0.047) associated to mortality. Mortality of noncirrhotic patients with FC (14/31; 45.2 %) was not different to noncirrhotic controls without FC [28/89 (31.2 %; p = 0.168)]. Similarly, in
Tobias Lahmer and Marlena Messer have equally contributed to this article. T. Lahmer (&) M. Messer U. Mayr B. Saugel S. Noe C. Schultheiss P. Thies C. Spinner S. Nennstiel C. Schwerdtfeger V. Phillip R. M. Schmid W. Huber II. Medizinische Klinik und Poliklinik, Klinikum rechts der Isar der Technischen Universita¨t Mu¨nchen, Ismaninger Str. 22, 81675 Munich, Germany e-mail: [email protected]
multivariate analysis of noncirrhotics, APACHE-II (p \ 0.001), but not FC, was independently associated to mortality. Multiple regression analysis of all 240 cirrhotic and noncirrhotic patients demonstrated that APACHE-II (p \ 0.001), cirrhosis (p = 0.001) and FC (p = 0.049) were independently associated with mortality. Conclusion Fungal ‘‘colonisation’’ is independently associated to mortality in cirrhotic ICU-patients. Early antimycotic therapy should be considered in critically ill cirrhotic patients with FC. Keywords Fungal colonisation Liver cirrhosis Critical ill patients Invasive mycoses
Introduction Invasive mycosis (IM) is a life-threatening infection in neutropenic patients, transplant recipients and in patients with malignant diseases. It is also increasing, especially among patients who are immunocompromised in anyway or hospitalised with serious underlying diseases such as critically ill patients [1, 2]. This also applies to patients with liver cirrhosis (LC) who seem to be particularly susceptible to a variety of complications especially to infections including fungal and opportunistic aetiology. The risk of infection has been associated with severity of liver disease and hospitalisation [3].
123
Mycopathologia
IM causes high morbidity and significant mortality. Early diagnosis, specific treatment and also prophylaxis in patients at high risk are obvious strategies to improve outcome. According to the Invasive Fungal Infections Cooperative Group in Europe and the Mycoses Study Group in the USA, a ‘‘proven’’ IM is defined as a positive fungal culture or histological analysis of a tissue specimen taken from a disease site, or identification of fungal or hyphal elements in a biopsy from a sterile site [4]. In addition to obvious and proven IM, a substantial proportion of patients are initially classified as ‘‘colonised’’ with fungi, in most cases with candida species. For this reason, the candida colonisation index or candida score can be helpful tools to estimate the potential risk for IM [5, 6]. Multiple-site colonisation with fungi is commonly recognised as a risk factor for invasive fungal infection in critically ill patients. Some of these patients will subsequently develop severe infection. In addition to mycological evidence from culture, microscopic analysis and indirect tests including antigen detection, classification as ‘‘probable’’ or ‘‘possible’’ fungal infections is based on clinical features of fungal infection, which depends on specific host factors. Moreover, in critically ill patients with diagnosis of fungal colonisation (FC), clinical signs of inflammation might be also related to infection with other microorganisms. Furthermore, fungal species are normal inhabitants of the mouth, throat and the rest of the gastrointestinal tract, and they usually do not cause any harm. However, the use of certain medications or a weakening of the immune system can cause fungi to multiply and result in invasive infections, especially in immunocompromised patients [2]. Unspecificity of most of the findings and interaction with host factors might explain that these criteria help to characterise patients in clinical trials, but are less useful to guide clinical practice in an individual patient [4]. This particularly applies to patients with end-stage liver disease who frequently do not have classical risk factors and do not fulfil standard criteria for IM. In general, the impact of FC on the outcome of patients with LC is not well investigated. We hypothesised that the impact of FC on outcome might be underestimated in cirrhotic patients who might be at particular risk due to endogenous
123
immunosuppression and immunological anergy impeding detectable host response. Therefore, the present study aimed to determine the impact of FC on the mortality rate in patients with LC in a medical ICU, compared to patients without LC with comparable APACHE-II score.
Methods Study Design, Setting and Patients This retrospective study was conducted at the medical ICU at Klinikum rechts der Isar, University Hospital of the Technische Universita¨t Mu¨nchen, Germany. We analysed the medical records of all patients with LC that were treated at the ICU between January 2010 and December 2012. This observational retrospective study without any specific intervention was reviewed and approved by the ethics committee of our university hospital, and all data were processed anonymously. Need for informed consent was waived for this retrospective analysis. We identified 120 patients with LC. In patients who were administrated more than once to the ICU during the study period, only the first episode of ICU administration was analysed. In the control group, we screened 312 patients and included finally 120 patients. Inclusion criteria were based on comparable age and APACHE-II score. Data regarding the patients’ clinical and demographic characteristics were documented according to the patient’s medical records and the electronic patient database of our hospital. Definitions and Microbiological Testings Decisions for microbiological testing were considered on clinical and/or laboratory reasons. Clinical reasons were based on SIRS and/or sepsis criteria. Laboratory reasons were based on elevated infection parameters as declared in the following: leucocyte count \4 or [9 g/l; C-reactive protein [0.5 mg/dl; or procalcitonin [0.1 ng/ml. For definition of fungal disease, we used the criteria proposed by the US Centers for Disease Control and Prevention and the EORTC/MSG. Invasive mycoses are defined as a primarily proof through blood culture, imaging tests such as chest
Mycopathologia
Microbiological Testing
USA)]. Antimicrobial susceptibility assessment was performed using an automated microdilution system (VITEK system BioMe´rieux, Nu¨rtingen, Germany) or standardised discs. During the entire study period, antimicrobial susceptibility testing was performed according to CLSI methodologies, and testing was regularly updated to the latest recommendations.
Blood Culture
Fungal Testing
Twenty millilitres of venous blood was obtained from patients with febrile episodes and inoculated in an aerobic and an anaerobic blood culture bottle (BacTec system, Becton–Dickinson, Heidelberg, Germany). At the Institute of Microbiology, blood cultures were placed in an automated microbiology growth and detection system (BacTec system, Becton–Dickinson, Heidelberg, Germany) and incubated at 37 °C for 5 days.
Fungals were detected by direct examination and culture. Standard procedure for respiratory specimens includes direct examination of gram-stained samples and culture onto sheep blood agar (Becton–Dickinson) incubated for 48 h at 37 °C in an aerobic atmosphere and chocolate agar (BioMe´rieux) incubated for 48 h at 37 °C in a CO2-enriched atmosphere. When fungi are specifically suspected, additional procedures include a direct examination using calcofluor (Becton–Dickinson) and culture onto Sabouraud chloramphenicol horse blood agar (Biotrading) incubated for 2 weeks at 37 °C in an aerobic atmosphere. For biologic sterile fluids or biopsy, in addition to the above-mentioned procedures, a Schaedler broth (Becton–Dickinson) Sabouraud dextrose agar slant (Becton–Dickinson) was inoculated and incubated at 37 °C in an aerobic atmosphere. Cultures on sheep blood agar and chocolate agar were interpreted after 24 and 48 h of incubation, and cultures on Sabouraud chloramphenicol, Schaedler and Sabouraud dextrose agar slant were interpreted after 24 h and then twice a week. Colonies of interest were subcultured on Takashio agar (Biotrading) and identified by their macroscopic and microscopic (toluidine blue) characteristics. All samples were both tested by microscope and culture. For each compartment, a specific significant germ count was defined. A positive BAL was defined by more than 10,000 germs/ml (cfu/ml), and a positive urine sample was defined by more than[100 germs/ll urine. Blood cultures were inbred at 37 °C for a maximum of 5 days, and they were defined as positive by chemical reaction. The galactomannan test for Aspergillus galactomannan (GM) was used in serum, using the Platelia Aspergillus enzyme immunoassay (EIA). The result was computed as an index, with the optical density of the patient sample divided by the optical density of the threshold control. If the serum index was superior to 0.5, the tested serum was positive for GM.
X-ray or CT scan of the lung and other organs, biopsies of affected tissue or samples for evidence of the fungus under a microscope or through fungal culture. In all other cases, the evidence of fungal is interpreted as colonisation.
BAL, Ascites and Urine The Bronchoalveolar lavage (BAL)/Ascites specimens were examined for aerobic and anaerobic bacteria. In each case, 50–100 ll of BAL/Ascites was transferred into liquid nutrient media (glucose broth, thioglycollate broth) and onto solid culture media (Columbia sheep blood agar, chocolate agar, McConkey agar, Schaedler anaerobic agar and Schaedler KV anaerobic agar). Urine specimens were also examined for aerobic and anaerobic bacteria. In each case, 10 ll of urine was transferred into liquid nutrient media and onto solid culture media (Columbia sheep blood agar, McConkey agar and a Bile esculin agar). Subsequently, the culture media were incubated at 37 °C. The aerobic cultures were incubated for 48 h, with the first readout taken after 24 h. The anaerobic cultures were monitored for the first time after 48 h and processed further as required. Germ Identification and Antibiotic Susceptibility Testing From 2009–2013, microbial identification was conducted using biochemical testing systems [VITEK system, BioMe´rieux, Nu¨rtingen, Germany and matrixassociated laser desorption/ionisation time of flight (MALDI-TOF, Bruker Corporation, Billerica, MA,
123
Mycopathologia Table 1 Basic demographic data, patient’s characteristics and the reasons for ICU admission of the LC group Patients with LC and FC (N = 69)
Patients with LC without FC (N = 51)
Age (years)
55 ± 14
58 ± 17
Sex (male/female)
48/21
39/12
Aetiology of LC
Table 2 Basic demographic data, patient’s characteristics and the reasons for ICU admission of the control group Control group without LC and FC (N = 31)
Control group without LC and without FC (N = 89)
Age (years)
64 ± 12
63 ± 16
Sex (male/female) Reason for ICU admission
22/9
59/30
Alcoholic
58
44
NASH
5
4
Sepsis
19
28
Viral (Hep B/C) PSC/PBC
4 2
2 1
COPD
3
6
A
2
3
B
5
14
C
62
34
Cardiological disorders (including heart attack)
Sepsis (pneumonia, SBP, etc.)
37
17
Hepato-renal syndrome
19
22
Gastro-intestinal bleeding
7
12
Gastrointestinal disorders (including acute abdomen, pancreatitis)
Alcoholic steatohepatitis
6
Child-Pugh score
Reason for ICU admission
Statistical Analysis All statistical tests were conducted two-sided in an explorative manner on a 5 % level of significance. Analyses were performed using SPSS version 21.0. Categorical variables were analysed using the Chi-square test. For group comparisons of continuous variables between patients with and without FC as well as between patients with and without cirrhosis, the Mann–Whitney U test was used. Multivariate logistic regression analyses were performed to identify risk factors for FC and ICU mortality. Variables considered to be a risk factor in advance or showing p value \0.100 in the univariate analysis were entered into the multivariate model. To identify independent risk factors for FC, the following factors were included in multivariate regression analysis: age, sex, Child-Pugh score, APACHE-II score, ICU mortality, LC and fungal colonisation.
Neurological disorders (including ICB, SAB, Stroke) Oncological disorders (including AML, NHL)
22
3
10 9
6
14
with FC, 89 without FC, were enroled in this study. Basic demographic data, patient’s characteristics and the reasons for ICU admission are stated in Tables 1 and 2. Microbiological Findings In both groups with FC, candida species dominated the microbiological findings. Especially, Candida albicans and Candida glabrata were found predominantly in both groups (see Table 3). In the LC group, six patients had Aspergillus fumigatus and in four cases other fungal species (see Table 3). Noticeable was the finding in the LC group that different fungals could be detected in separate compartments, e.g. urine and lung within one patient explaining the more findings of fungal evidences in the LC group as compared to the control group. The typical localisation of fungal findings were in both groups predominantly the lung followed by urine and blood culture. In the LC group, fungal could be found in five patients in ascites, respectively.
Results Patients Characteristics
Risk Factors, Candida Colonisation Index and Candida Score
About 120 critically ill patients with LC, 69 with FC, 51 without FC, and 120 patients of a control collective, 31
Other risk factors for fungal colonisation/infection are presented in Table 5.
123
Mycopathologia
Outcome of Cirrhotic Patients With and Without FC Overall, 69 out of 120 (57.5 %) patients with LC had fungi. However, only ten of these 69 patients (14.5 %) fulfilled the criteria for IM. Univariate Analyses Cirrhotic patients with FC had significantly higher APACHE-II (27 ± 3 vs. 23 ± 4, p \ 0.001) score and mortality [54/69 (78 %) vs. 18/51 (35 %); Table 3 Fungal species and colonisation site Patients with LC and FC (N = 69)
Control group without LC and FC (N = 31)
p \ 0.001] compared to cirrhotic patients without FC (Table 4). Furthermore, patients with Child C cirrhosis had significantly more frequent FC compared to Child A and B cirrhotics [62/96 (64.6 %) vs. 7/24 (29.2 %); p = 0.002]. Moreover, APACHE-II was higher in nonsurvivors compared to survivors (26.6 ? 2.2 vs. 22.4 ? 4.0; p \ 0.001). By contrast, gender was not associated with mortality (p = 0.242). To further analyse the independent association of several risk factors with mortality, multivariate binary logistic regression analysis was performed. In addition to APACHE-II (p \ 0.001) and male gender (p = 0.042), FC was independently associated with higher mortality rate (p = 0.047). By contrast, age was not independently associated with mortality.
Kind of fungus
Outcome of Noncirrhotic Controls With and Without FC
Candida species C. albicans
49
C. glabrata
15
19 6
C. krusei
7
1
C. tropicalis
4
4
C. parapsilosis
4
1
C. kefyr
3
C. guilliermondii
1
C. nivariensis
1
Aspergillus Asp. fumigatus
6
Microbiological findings were positive for fungi in 31/120 (25.8 %) of noncirrhotic, but only two patients (6.5 %) fulfilled the criteria for IM. Similar analyses as for cirrhotic patients were performed in noncirrhotic controls with comparable APACHE-II. Univariate Analyses
Others Galactomyces cand.
1
Sacch. cerevisiae
2
Trichosporon asahii
1
In univariate analysis, the impact of FC was less pronounced in noncirrhotic patients: there was a slight, but significance in APACHE-II (25.4 ± 2.0 vs. 24.8 ± 2.7), but no difference in mortality [14/31 (45.2 %) vs. 28/89 (31.5 %); p = 0.168] between noncirrhotic patients with and without FC. Mortality was not different for male and female patients without cirrhosis [24/76 (31.6 %) vs. 18/44 (40.9 %); p = 0.301].
Colonisation site Lung
56
21
Urine
21
8
Blood culture
10
2
5
–
Ascites
Table 4 Mortality rates and APACHE-II score Patients with LC and FC (N = 69)
Patients with LC without FC (N = 51)
Control group without LC and FC (N = 31)
Control group without LC and without FC (N = 89)
Mortality rate (%)
54/69 (78 %)
18/51 (35 %)
14/31 (45 %)
28/89 (31 %)
APACHE-II score
27 ± 3
23 ± 4
25 ± 3
24 ± 2
Parameters with a statistical significance are in bold (p \ 0.05)
123
Mycopathologia
In multivariate binary regression analysis, APACHE-II was the only parameter with independent association (p \ 0.001). By contrast, FC, age and gender were not independently associated with mortality. ROC-Analyses Regarding Mortality Different impact of FC on mortality in cirrhotic and noncirrhotic patients was further emphasised by ROCanalysis. In cirrhotic patients, FC provided the largest AUC of all investigated parameters (AUC 0.719; 95 % confidence interval (CI) 0.623–0.815; p \ 0.001) next to APACHE-II (AUC 0.848; 95 % CI 0.774–0.921; p \ 0.001). By contrast, Child-Pugh classification provided an AUC of 0.610 (95 % CI 0.504–0.716; p = 0.042). Neither gender (AUC 0.549) nor age (AUC 0.462) were significant predictors. However, in noncirrhotic patients with comparable APACHE-II, FC did not provide a ROC-AUC with significance (AUC 0.558; 95 % CI 0.448–0.667; p = 0.298). APACHE-II was the only significant predictor in mortality analysis (AUC = 0.826; 95 % CI 0.750–0.903; p \ 0.001). Neither age (AUC 0.546) nor gender (AUC = 0.452) were significant predictors in noncirrhotic patients (Figs. 1, 2).
Fig. 2 Predictive parameters for a higher mortality rate in both the LC and control group: the highest AUC values regarding mortality were observed for APACHE-II score followed by FC
Analysis of All 240 Patients To investigate the impact of FC and cirrhosis in the totality of patients, we performed binary regression analysis in the totality of patients. In these patients, in addition to APACHE-II score (p \ 0.001), the diagnosis of cirrhosis (p = 0.001) was the strongest independent predictor for mortality. FC were independently associated with borderline significance (p = 0.049). Neither age nor gender were independent predictors of mortality. These data were in general confirmed by ROCanalysis (Fig. 3): the largest AUC values regarding mortality were also observed for APACHE-II score (AUC 0.830; 95 % CI 0.77–0.88; p \ 0.001), FC (AUC 0.67; 95 % CI 0.602–0.740; p \ 0.001) and cirrhosis (AUC 0.625; 95 % CI 0.554–0.696; p = 0.001).
Discussion
Fig. 1 Predictive parameters for elevated mortality rate in the LC group: the highest AUC values regarding mortality were observed for APACHE-II score followed by FC and Child-Pugh score
123
Discrimination of FC and IM is particularly difficult in patients with nonovert immune incompetency. This also applies to patients with LC. Therefore, we compared the impact of FC on the mortality rate of cirrhotic versus noncirrhotic patients with comparable APACHE-II score in a medical ICU. Our study demonstrates a marked impact of FC on the outcome of cirrhotic patients, which was substantially more pronounced compared to a control group without LC.
Mycopathologia
Fig. 3 Predictive parameters for mortality in the totality of patients: the highest AUC values regarding mortality were observed for APACHE-II score followed by FC and LC
Our data demonstrate significant association of FC with mortality in univariate as well as multivariate analyses for cirrhotics, but not for noncirrhotic patients. This is in contrast to current recommendations that only patients fulfilling the criteria for IM are at risk for an increased mortality and morbidity. In our study, IM defining findings such as positive blood culture or microscopic and/or cultural evidence for fungi and other specimens was restricted to 14.5 % in cirrhosis and 6.5 % in noncirrhotic patients, respectively. However, 57.5 % of cirrhotics had evidence for FC. Among these patients, the APACHE-II score and the mortality rate were significantly higher than in cirrhotics without FC. Furthermore, FC was markedly more frequent in Child C compared to Child A and B cirrhotic patients (65 vs. 29 %; p = 0.002). These data are in line with previous data showing that the risk for FC increases with the severity of the liver disease and with hospitalisation. Furthermore, there is evidence that precipitating complications such as variceal bleeding increase the risk of infection [7, 8]. In our study, the rate of patients fulfilling criteria for IM is comparable to haematological patients. Moreover, the rate of FC was unexpectedly high. Particularly in immunocompromised and critically ill patients, the difficulty to differentiate between colonisation and infection is recognised: Although
clinical signs of severe infection may manifest early, but due to their low specificity the association with fungal infection may remain in-overt until late in the course of the disease. With regard to ‘‘early goaldirected therapy’’ and ‘‘hit hard and early’’ ‘‘in time’’ detection and treatment as known from septic patients or varices bleeding patients, it is quite challenging for IM in cirrhotics [8–10]. In general, the symptoms of fungal diseases depend on the type of infection, localisation within the body and the host’s response. The symptoms of (invasive) mycoses are not specific. Fever and chills that do not improve after antibiotic therapy are the most common symptoms. If the infection spreads to other organs or parts of the body such as kidneys, liver, bones, muscles, joints, spleen or eyes, additional symptoms may develop, which vary depending on the site of infection. Finally, infection not responding to treatment results in organ failure [8, 11]. Interpretation of these findings as a result of fungal infection is difficult in critically ill patients and particularly in cirrhotics, since these patients often have a variety of explanations for inflammation, new infiltrates or progression towards multiple organ failure [9, 12]. Therefore, early and advanced diagnosis of fungal infections in critically ill patients and particularly with LC is difficult. Obviously, this requires a high index of suspicion [13–15]. Therefore, scoring systems such as the candida colonisation index or the candida score can be helpful tools identifying possible risks for fungal infection. In this study, in both groups with FC, the candida colonisation index as well the candida score was elevated, suspicious for possible fungal infection (see Table 5). In our study, FC in patients with LC was most frequently found in ascites, urine and in the lungs, which were also the most frequent specific sites in previous studies. Fungi in ascites or urine can be easily detected by fungal culture or by microscope [16]. However, affection of the lungs is still challenging. IM still is diagnosed by histopathological examination of lung tissue obtained by thoracoscopy or open-lung biopsy as ‘‘gold standard’’ for the diagnosis of IM. Unfortunately, in patients with LC, the benefits of appropriate diagnostics by lung biopsies have to be outweighed against the risks of thrombocytopenia, impaired coagulation and increased respiratory distress [14].
123
Mycopathologia Table 5 Other risk factors for fungal colonisation and/or infection Patients with LC and FC (N = 69)
Patients with LC without FC (N = 51)
Control group without LC and FC (N = 31)
Control group without LC and without FC (N = 89)
27.6 (±16)
20.4 (±12)
14.4 (±8)
11.8 (±10)
Single therapy
5
16
15
54
2 or more than 2
64
35
16
35
3.5 ± 1.4
3.4 ± 1.5
2.7 ± 1.1
2±1
Length of ICU stay (days) Type of antibiotic therapy
Number of medical devices Haemodialysis
48
23
12
19
Mechanical ventilation
69
45
22
44
Mean candida colonisation index
0.65 ± 0.17
0.45 ± 0.17
Mean candida score
2.8 ± 0.9
1.8 ± 0.8
Although the diagnostic value of BAL for fungal evidence is controversial, BAL is considered to be safe and useful in high-risk patients with suspected IM [17]. The complete information derived from microscopy of BAL samples frequently helps to establish the diagnosis of IM. Moreover, demonstration of fungi in culture or direct microscopy, even if obtained from sputum or BAL fluid (a nonsterile site), has a high predictive value for IM, and for almost all practical purposes, it is valuable for diagnosis and treatment [17–19]. Beyond direct detection of fungus as described above, molecular methods are useful when fungal organisms cannot be isolated in culture, when subtle differences within a complex are not readily visible under light microscopy or when unstable phenotypic characteristics have been lost as may happen in chronic infections [20]. However, in most cases, molecular methods are extensive and their results are often imprecise especially in marginal conditions such as colonisation [20, 21]. To date, no sensitive and specific antibody assay has been developed that is useful for diagnosis. Since fungi—e.g., Candida—are part of the normal flora, healthy persons have antibodies against these organisms. This obviously decreases specificity of the assay [21]. On the other hand, an immunosuppressed host may be unable to generate a vigorous antibody response to Candida infection, thereby decreasing the sensitivity of the assay in this high-risk population. Thus, two nonculture-based assays, PCR and the beta-D-glucan assay, when combined with blood cultures, have the potential to increase the sensitivity of culture methods alone for IM [22–24].
123
Despite recent advances in molecular diagnosis, direct microscopy remains an essential tool for both detection of the organisms in the sample, and identification of the growing culture to the species level [22, 23]. Therefore, further studies have to be performed to describe the sensitivity of diagnostic testing compared with blood cultures alone, especially in immunocompromised patients and in conditions of FC [25, 26]. Therefore, we do not routinely perform these tests at present.
Limitations of the Study Although this study analyses one of the largest cohorts of cirrhotic patients with and without FC and compares these data to a noncirrhotic controls with comparable APACHE-II, this study was conducted retrospectively in a single medical ICU (monocentric study).
Conclusion Patients with LC have a high incidence of FC. These patients have a significantly higher mortality rate as compared to patients with LC without FC and matched noncirrhotic controls with comparable APACHE-II. Regrading their increased mortality, early antimycotic therapy should be considered particularly in patients with Child C cirrhosis even without definite prove of IM.
Mycopathologia Conflict of interest None of the authors has any potential financial conflict of interest related to this manuscript.
References 1. Donnelly JP. Symptoms and diagnosis of nosocomial fungal infections -state-of-the-art. Eur J Med Res. 2002;31:192–9. 2. Hsu JL, Ruoss SJ, Bower ND, Lin M, Holodniy M, Stevens DA. Diagnosing invasive fungal disease in critically ill patients. Crit Rev Microbiol. 2011;37:277–312. 3. Bonnel AR, Bunchorntavakul C, Reddy KR. Immune dysfunction and infections in patients with cirrhosis. Clin Gastroenterol Hepatol. 2011;9:727–38. 4. Tortorano AM, Dho G, Prigitano A, et al. ECMM-FIMUA Study Group invasive fungal infections in the intensive care unit: a multicentre, prospective, observational study in Italy (2006–2008). Mycoses. 2012;55:73–9. 5. Pittet D, Monod M, Suter PM, Frenk E, Auckenthaler R. Candida colonization and subsequent infections in critically ill surgical patients. Ann Surg. 1994;220(6):751–8. 6. Leroy G, Lambiotte F, The´venin D, et al. Evaluation of ‘‘Candida score’’ in critically ill patients: a prospective, multicenter, observational, cohort study. Ann Intensive Care. 2011;1(1):50. 7. Iber FL. Patients with cirrhosis and liver failure are at risk for bacterial and fungus infection. Am J Gastroenterol. 1999;94:2001–3. 8. Rolando N, Philpott-Howard J, Williams R. Bacterial and fungal infection in acute liver failure. Semin Liver Dis. 1996;16:389–402. 9. Dellinger RP, Levy MM, Rhodes A, et al. Surviving sepsis campaign: international guidelines for management of severe sepsis and septic shock, 2012. Intensive Care Med. 2013;39(2):165–228. 10. Kumar A, Safdar N, Kethireddy S, Chateau D. A survival benefit of combination antibiotic therapy for serious infections associated with sepsis and septic shock is contingent only on the risk of death: a meta-analytic/meta-regression study. Crit Care Med. 2010;38(8):1651–64. 11. O’Brien J, Triantos C, Burroughs AK. Management of varices in patients with cirrhosis. Nat Rev Gastroenterol Hepatol. 2013;10(7):402–12. 12. Gustot T, Durand F, Lebrec D, et al. Severe sepsis in cirrhosis. Hepatology. 2009;50(6):2022–33.
13. Marchetti O, Eggimann P, Calandra T. Invasive candidiasis in critically ill patients: does progressing knowledge improve clinical management and outcome? Curr Opin Crit Care. 2010;16:442–4. 14. Vazquez JA. Invasive fungal infections in the intensive care unit. Semin Respir Crit Care Med. 2010;31:79–86. 15. Pema´n J, Zaragoza R. Current diagnostic approaches to invasive candidiasis in critical care settings. Mycoses. 2010; 53:424–33. 16. Choi SH, Soo Kim Y, Chung JW, et al. Clinical significance of untreated Candida species isolated from ascites in cirrhotic patients. Scand J Infect Dis. 2004;9:649–55. 17. Navasa M, Rimola A, Rodes J. Bacterial infections in liver disease. Semin Liver Dis. 1997;17:323–33. 18. Yazıcıog˘lu Moc¸in O, Karakurt Z, Aksoy F et al. Bronchoscopy as an indicator of tracheobronchial fungal infection in non-neutropenic intensive-care unit patients. Clin Microbiol Infect. 2013; 19(3):E136–41. 19. Falcone M, Massetti AP, Russo A, Vullo V, Venditti M. Invasive aspergillosis in patients with liver disease. Med Mycol. 2011;49:406–13. 20. Prodanovic H, Cracco C, Massard J, et al. Invasive pulmonary aspergillosis in patients with decompensated cirrhosis: case series. BMC Gastroenterol. 2007;31:2. 21. Charles PE, Castro C, Ruiz-Santana S, Leo´n C, Saavedra P, Martı´n E. Serum procalcitonin levels in critically ill patients colonized with Candida spp: new clues for the early recognition of invasive candidiasis? Intensive Care Med. 2009; 35:2146–50. 22. Badiee P, Alborzi A, Joukar M. Molecular assay to detect nosocomial fungal infections in intensive care units. Euro J Intern Med. 2011;22:611–5. 23. Badiee P, Alborzi A, Joukar M. Molecular assay to detect nosocomial fungal infections in intensive care units. Eur J Intern Med. 2011;22:611–5. 24. Lipke AB, Mihas AA. Non-decompensated cirrhosis as a risk factor for invasive aspergillosis: a case report and review of the immune dysfunction of cirrhosis. Am J Med Sci. 2007;334:314–6. 25. Echeverria PM, Kett DH, Azoulay E. Candida prophylaxis and therapy in the ICU. Semin Respir Crit Care Med. 2011; 32:159–73. 26. Falcone M, Massetti AP, Russo A, Vullo V, Venditti M. Invasive aspergillosis in patients with liver disease. Med Mycol. 2011;4:406–13.
123
Fungal ‘‘colonisation’’ is Associated with Increased Mortality in Medical Intensive Care Unit Patients with Liver Cirrhosis Tobias Lahmer • Marlena Messer • Ulrich Mayr • Bernd Saugel • Sebastian Noe Caroline Schultheiss • Philipp Thies • Christoph Spinner • Simon Nennstiel • Christiane Schwerdtfeger • Veit Phillip • Roland M. Schmid • Wolfgang Huber
•
Received: 20 July 2014 / Accepted: 13 October 2014 Ó Springer Science+Business Media Dordrecht 2014
Abstract Objectives Patients with liver cirrhosis are at increased risk for fungal infections. However, distinction of fungal colonisation (FC) and invasive mycoses is difficult. Aim of this study was to analyse the impact of FC on mortality of cirrhotic ICU-patients. Methods Retrospective mortality analysis of a prospectively maintained database on 120 cirrhotic patients with and without FC. Comparison to 120 noncirrhotic controls matched for APACHE-II (24.9 ± 3.7 vs. 25.0 ± 2.6; p = 0.263). Results About 69/120 (58 %) of patients with cirrhosis had FC. These patients had significantly higher APACHE-II score and mortality compared to cirrhotic patients without FC (27 ± 3 vs. 23 ± 4, p \ 0.001; 78 vs. 35 %, p \ 0.001). In multivariate analysis, FC was independently (p = 0.047) associated to mortality. Mortality of noncirrhotic patients with FC (14/31; 45.2 %) was not different to noncirrhotic controls without FC [28/89 (31.2 %; p = 0.168)]. Similarly, in
Tobias Lahmer and Marlena Messer have equally contributed to this article. T. Lahmer (&) M. Messer U. Mayr B. Saugel S. Noe C. Schultheiss P. Thies C. Spinner S. Nennstiel C. Schwerdtfeger V. Phillip R. M. Schmid W. Huber II. Medizinische Klinik und Poliklinik, Klinikum rechts der Isar der Technischen Universita¨t Mu¨nchen, Ismaninger Str. 22, 81675 Munich, Germany e-mail: [email protected]
multivariate analysis of noncirrhotics, APACHE-II (p \ 0.001), but not FC, was independently associated to mortality. Multiple regression analysis of all 240 cirrhotic and noncirrhotic patients demonstrated that APACHE-II (p \ 0.001), cirrhosis (p = 0.001) and FC (p = 0.049) were independently associated with mortality. Conclusion Fungal ‘‘colonisation’’ is independently associated to mortality in cirrhotic ICU-patients. Early antimycotic therapy should be considered in critically ill cirrhotic patients with FC. Keywords Fungal colonisation Liver cirrhosis Critical ill patients Invasive mycoses
Introduction Invasive mycosis (IM) is a life-threatening infection in neutropenic patients, transplant recipients and in patients with malignant diseases. It is also increasing, especially among patients who are immunocompromised in anyway or hospitalised with serious underlying diseases such as critically ill patients [1, 2]. This also applies to patients with liver cirrhosis (LC) who seem to be particularly susceptible to a variety of complications especially to infections including fungal and opportunistic aetiology. The risk of infection has been associated with severity of liver disease and hospitalisation [3].
123
Mycopathologia
IM causes high morbidity and significant mortality. Early diagnosis, specific treatment and also prophylaxis in patients at high risk are obvious strategies to improve outcome. According to the Invasive Fungal Infections Cooperative Group in Europe and the Mycoses Study Group in the USA, a ‘‘proven’’ IM is defined as a positive fungal culture or histological analysis of a tissue specimen taken from a disease site, or identification of fungal or hyphal elements in a biopsy from a sterile site [4]. In addition to obvious and proven IM, a substantial proportion of patients are initially classified as ‘‘colonised’’ with fungi, in most cases with candida species. For this reason, the candida colonisation index or candida score can be helpful tools to estimate the potential risk for IM [5, 6]. Multiple-site colonisation with fungi is commonly recognised as a risk factor for invasive fungal infection in critically ill patients. Some of these patients will subsequently develop severe infection. In addition to mycological evidence from culture, microscopic analysis and indirect tests including antigen detection, classification as ‘‘probable’’ or ‘‘possible’’ fungal infections is based on clinical features of fungal infection, which depends on specific host factors. Moreover, in critically ill patients with diagnosis of fungal colonisation (FC), clinical signs of inflammation might be also related to infection with other microorganisms. Furthermore, fungal species are normal inhabitants of the mouth, throat and the rest of the gastrointestinal tract, and they usually do not cause any harm. However, the use of certain medications or a weakening of the immune system can cause fungi to multiply and result in invasive infections, especially in immunocompromised patients [2]. Unspecificity of most of the findings and interaction with host factors might explain that these criteria help to characterise patients in clinical trials, but are less useful to guide clinical practice in an individual patient [4]. This particularly applies to patients with end-stage liver disease who frequently do not have classical risk factors and do not fulfil standard criteria for IM. In general, the impact of FC on the outcome of patients with LC is not well investigated. We hypothesised that the impact of FC on outcome might be underestimated in cirrhotic patients who might be at particular risk due to endogenous
123
immunosuppression and immunological anergy impeding detectable host response. Therefore, the present study aimed to determine the impact of FC on the mortality rate in patients with LC in a medical ICU, compared to patients without LC with comparable APACHE-II score.
Methods Study Design, Setting and Patients This retrospective study was conducted at the medical ICU at Klinikum rechts der Isar, University Hospital of the Technische Universita¨t Mu¨nchen, Germany. We analysed the medical records of all patients with LC that were treated at the ICU between January 2010 and December 2012. This observational retrospective study without any specific intervention was reviewed and approved by the ethics committee of our university hospital, and all data were processed anonymously. Need for informed consent was waived for this retrospective analysis. We identified 120 patients with LC. In patients who were administrated more than once to the ICU during the study period, only the first episode of ICU administration was analysed. In the control group, we screened 312 patients and included finally 120 patients. Inclusion criteria were based on comparable age and APACHE-II score. Data regarding the patients’ clinical and demographic characteristics were documented according to the patient’s medical records and the electronic patient database of our hospital. Definitions and Microbiological Testings Decisions for microbiological testing were considered on clinical and/or laboratory reasons. Clinical reasons were based on SIRS and/or sepsis criteria. Laboratory reasons were based on elevated infection parameters as declared in the following: leucocyte count \4 or [9 g/l; C-reactive protein [0.5 mg/dl; or procalcitonin [0.1 ng/ml. For definition of fungal disease, we used the criteria proposed by the US Centers for Disease Control and Prevention and the EORTC/MSG. Invasive mycoses are defined as a primarily proof through blood culture, imaging tests such as chest
Mycopathologia
Microbiological Testing
USA)]. Antimicrobial susceptibility assessment was performed using an automated microdilution system (VITEK system BioMe´rieux, Nu¨rtingen, Germany) or standardised discs. During the entire study period, antimicrobial susceptibility testing was performed according to CLSI methodologies, and testing was regularly updated to the latest recommendations.
Blood Culture
Fungal Testing
Twenty millilitres of venous blood was obtained from patients with febrile episodes and inoculated in an aerobic and an anaerobic blood culture bottle (BacTec system, Becton–Dickinson, Heidelberg, Germany). At the Institute of Microbiology, blood cultures were placed in an automated microbiology growth and detection system (BacTec system, Becton–Dickinson, Heidelberg, Germany) and incubated at 37 °C for 5 days.
Fungals were detected by direct examination and culture. Standard procedure for respiratory specimens includes direct examination of gram-stained samples and culture onto sheep blood agar (Becton–Dickinson) incubated for 48 h at 37 °C in an aerobic atmosphere and chocolate agar (BioMe´rieux) incubated for 48 h at 37 °C in a CO2-enriched atmosphere. When fungi are specifically suspected, additional procedures include a direct examination using calcofluor (Becton–Dickinson) and culture onto Sabouraud chloramphenicol horse blood agar (Biotrading) incubated for 2 weeks at 37 °C in an aerobic atmosphere. For biologic sterile fluids or biopsy, in addition to the above-mentioned procedures, a Schaedler broth (Becton–Dickinson) Sabouraud dextrose agar slant (Becton–Dickinson) was inoculated and incubated at 37 °C in an aerobic atmosphere. Cultures on sheep blood agar and chocolate agar were interpreted after 24 and 48 h of incubation, and cultures on Sabouraud chloramphenicol, Schaedler and Sabouraud dextrose agar slant were interpreted after 24 h and then twice a week. Colonies of interest were subcultured on Takashio agar (Biotrading) and identified by their macroscopic and microscopic (toluidine blue) characteristics. All samples were both tested by microscope and culture. For each compartment, a specific significant germ count was defined. A positive BAL was defined by more than 10,000 germs/ml (cfu/ml), and a positive urine sample was defined by more than[100 germs/ll urine. Blood cultures were inbred at 37 °C for a maximum of 5 days, and they were defined as positive by chemical reaction. The galactomannan test for Aspergillus galactomannan (GM) was used in serum, using the Platelia Aspergillus enzyme immunoassay (EIA). The result was computed as an index, with the optical density of the patient sample divided by the optical density of the threshold control. If the serum index was superior to 0.5, the tested serum was positive for GM.
X-ray or CT scan of the lung and other organs, biopsies of affected tissue or samples for evidence of the fungus under a microscope or through fungal culture. In all other cases, the evidence of fungal is interpreted as colonisation.
BAL, Ascites and Urine The Bronchoalveolar lavage (BAL)/Ascites specimens were examined for aerobic and anaerobic bacteria. In each case, 50–100 ll of BAL/Ascites was transferred into liquid nutrient media (glucose broth, thioglycollate broth) and onto solid culture media (Columbia sheep blood agar, chocolate agar, McConkey agar, Schaedler anaerobic agar and Schaedler KV anaerobic agar). Urine specimens were also examined for aerobic and anaerobic bacteria. In each case, 10 ll of urine was transferred into liquid nutrient media and onto solid culture media (Columbia sheep blood agar, McConkey agar and a Bile esculin agar). Subsequently, the culture media were incubated at 37 °C. The aerobic cultures were incubated for 48 h, with the first readout taken after 24 h. The anaerobic cultures were monitored for the first time after 48 h and processed further as required. Germ Identification and Antibiotic Susceptibility Testing From 2009–2013, microbial identification was conducted using biochemical testing systems [VITEK system, BioMe´rieux, Nu¨rtingen, Germany and matrixassociated laser desorption/ionisation time of flight (MALDI-TOF, Bruker Corporation, Billerica, MA,
123
Mycopathologia Table 1 Basic demographic data, patient’s characteristics and the reasons for ICU admission of the LC group Patients with LC and FC (N = 69)
Patients with LC without FC (N = 51)
Age (years)
55 ± 14
58 ± 17
Sex (male/female)
48/21
39/12
Aetiology of LC
Table 2 Basic demographic data, patient’s characteristics and the reasons for ICU admission of the control group Control group without LC and FC (N = 31)
Control group without LC and without FC (N = 89)
Age (years)
64 ± 12
63 ± 16
Sex (male/female) Reason for ICU admission
22/9
59/30
Alcoholic
58
44
NASH
5
4
Sepsis
19
28
Viral (Hep B/C) PSC/PBC
4 2
2 1
COPD
3
6
A
2
3
B
5
14
C
62
34
Cardiological disorders (including heart attack)
Sepsis (pneumonia, SBP, etc.)
37
17
Hepato-renal syndrome
19
22
Gastro-intestinal bleeding
7
12
Gastrointestinal disorders (including acute abdomen, pancreatitis)
Alcoholic steatohepatitis
6
Child-Pugh score
Reason for ICU admission
Statistical Analysis All statistical tests were conducted two-sided in an explorative manner on a 5 % level of significance. Analyses were performed using SPSS version 21.0. Categorical variables were analysed using the Chi-square test. For group comparisons of continuous variables between patients with and without FC as well as between patients with and without cirrhosis, the Mann–Whitney U test was used. Multivariate logistic regression analyses were performed to identify risk factors for FC and ICU mortality. Variables considered to be a risk factor in advance or showing p value \0.100 in the univariate analysis were entered into the multivariate model. To identify independent risk factors for FC, the following factors were included in multivariate regression analysis: age, sex, Child-Pugh score, APACHE-II score, ICU mortality, LC and fungal colonisation.
Neurological disorders (including ICB, SAB, Stroke) Oncological disorders (including AML, NHL)
22
3
10 9
6
14
with FC, 89 without FC, were enroled in this study. Basic demographic data, patient’s characteristics and the reasons for ICU admission are stated in Tables 1 and 2. Microbiological Findings In both groups with FC, candida species dominated the microbiological findings. Especially, Candida albicans and Candida glabrata were found predominantly in both groups (see Table 3). In the LC group, six patients had Aspergillus fumigatus and in four cases other fungal species (see Table 3). Noticeable was the finding in the LC group that different fungals could be detected in separate compartments, e.g. urine and lung within one patient explaining the more findings of fungal evidences in the LC group as compared to the control group. The typical localisation of fungal findings were in both groups predominantly the lung followed by urine and blood culture. In the LC group, fungal could be found in five patients in ascites, respectively.
Results Patients Characteristics
Risk Factors, Candida Colonisation Index and Candida Score
About 120 critically ill patients with LC, 69 with FC, 51 without FC, and 120 patients of a control collective, 31
Other risk factors for fungal colonisation/infection are presented in Table 5.
123
Mycopathologia
Outcome of Cirrhotic Patients With and Without FC Overall, 69 out of 120 (57.5 %) patients with LC had fungi. However, only ten of these 69 patients (14.5 %) fulfilled the criteria for IM. Univariate Analyses Cirrhotic patients with FC had significantly higher APACHE-II (27 ± 3 vs. 23 ± 4, p \ 0.001) score and mortality [54/69 (78 %) vs. 18/51 (35 %); Table 3 Fungal species and colonisation site Patients with LC and FC (N = 69)
Control group without LC and FC (N = 31)
p \ 0.001] compared to cirrhotic patients without FC (Table 4). Furthermore, patients with Child C cirrhosis had significantly more frequent FC compared to Child A and B cirrhotics [62/96 (64.6 %) vs. 7/24 (29.2 %); p = 0.002]. Moreover, APACHE-II was higher in nonsurvivors compared to survivors (26.6 ? 2.2 vs. 22.4 ? 4.0; p \ 0.001). By contrast, gender was not associated with mortality (p = 0.242). To further analyse the independent association of several risk factors with mortality, multivariate binary logistic regression analysis was performed. In addition to APACHE-II (p \ 0.001) and male gender (p = 0.042), FC was independently associated with higher mortality rate (p = 0.047). By contrast, age was not independently associated with mortality.
Kind of fungus
Outcome of Noncirrhotic Controls With and Without FC
Candida species C. albicans
49
C. glabrata
15
19 6
C. krusei
7
1
C. tropicalis
4
4
C. parapsilosis
4
1
C. kefyr
3
C. guilliermondii
1
C. nivariensis
1
Aspergillus Asp. fumigatus
6
Microbiological findings were positive for fungi in 31/120 (25.8 %) of noncirrhotic, but only two patients (6.5 %) fulfilled the criteria for IM. Similar analyses as for cirrhotic patients were performed in noncirrhotic controls with comparable APACHE-II. Univariate Analyses
Others Galactomyces cand.
1
Sacch. cerevisiae
2
Trichosporon asahii
1
In univariate analysis, the impact of FC was less pronounced in noncirrhotic patients: there was a slight, but significance in APACHE-II (25.4 ± 2.0 vs. 24.8 ± 2.7), but no difference in mortality [14/31 (45.2 %) vs. 28/89 (31.5 %); p = 0.168] between noncirrhotic patients with and without FC. Mortality was not different for male and female patients without cirrhosis [24/76 (31.6 %) vs. 18/44 (40.9 %); p = 0.301].
Colonisation site Lung
56
21
Urine
21
8
Blood culture
10
2
5
–
Ascites
Table 4 Mortality rates and APACHE-II score Patients with LC and FC (N = 69)
Patients with LC without FC (N = 51)
Control group without LC and FC (N = 31)
Control group without LC and without FC (N = 89)
Mortality rate (%)
54/69 (78 %)
18/51 (35 %)
14/31 (45 %)
28/89 (31 %)
APACHE-II score
27 ± 3
23 ± 4
25 ± 3
24 ± 2
Parameters with a statistical significance are in bold (p \ 0.05)
123
Mycopathologia
In multivariate binary regression analysis, APACHE-II was the only parameter with independent association (p \ 0.001). By contrast, FC, age and gender were not independently associated with mortality. ROC-Analyses Regarding Mortality Different impact of FC on mortality in cirrhotic and noncirrhotic patients was further emphasised by ROCanalysis. In cirrhotic patients, FC provided the largest AUC of all investigated parameters (AUC 0.719; 95 % confidence interval (CI) 0.623–0.815; p \ 0.001) next to APACHE-II (AUC 0.848; 95 % CI 0.774–0.921; p \ 0.001). By contrast, Child-Pugh classification provided an AUC of 0.610 (95 % CI 0.504–0.716; p = 0.042). Neither gender (AUC 0.549) nor age (AUC 0.462) were significant predictors. However, in noncirrhotic patients with comparable APACHE-II, FC did not provide a ROC-AUC with significance (AUC 0.558; 95 % CI 0.448–0.667; p = 0.298). APACHE-II was the only significant predictor in mortality analysis (AUC = 0.826; 95 % CI 0.750–0.903; p \ 0.001). Neither age (AUC 0.546) nor gender (AUC = 0.452) were significant predictors in noncirrhotic patients (Figs. 1, 2).
Fig. 2 Predictive parameters for a higher mortality rate in both the LC and control group: the highest AUC values regarding mortality were observed for APACHE-II score followed by FC
Analysis of All 240 Patients To investigate the impact of FC and cirrhosis in the totality of patients, we performed binary regression analysis in the totality of patients. In these patients, in addition to APACHE-II score (p \ 0.001), the diagnosis of cirrhosis (p = 0.001) was the strongest independent predictor for mortality. FC were independently associated with borderline significance (p = 0.049). Neither age nor gender were independent predictors of mortality. These data were in general confirmed by ROCanalysis (Fig. 3): the largest AUC values regarding mortality were also observed for APACHE-II score (AUC 0.830; 95 % CI 0.77–0.88; p \ 0.001), FC (AUC 0.67; 95 % CI 0.602–0.740; p \ 0.001) and cirrhosis (AUC 0.625; 95 % CI 0.554–0.696; p = 0.001).
Discussion
Fig. 1 Predictive parameters for elevated mortality rate in the LC group: the highest AUC values regarding mortality were observed for APACHE-II score followed by FC and Child-Pugh score
123
Discrimination of FC and IM is particularly difficult in patients with nonovert immune incompetency. This also applies to patients with LC. Therefore, we compared the impact of FC on the mortality rate of cirrhotic versus noncirrhotic patients with comparable APACHE-II score in a medical ICU. Our study demonstrates a marked impact of FC on the outcome of cirrhotic patients, which was substantially more pronounced compared to a control group without LC.
Mycopathologia
Fig. 3 Predictive parameters for mortality in the totality of patients: the highest AUC values regarding mortality were observed for APACHE-II score followed by FC and LC
Our data demonstrate significant association of FC with mortality in univariate as well as multivariate analyses for cirrhotics, but not for noncirrhotic patients. This is in contrast to current recommendations that only patients fulfilling the criteria for IM are at risk for an increased mortality and morbidity. In our study, IM defining findings such as positive blood culture or microscopic and/or cultural evidence for fungi and other specimens was restricted to 14.5 % in cirrhosis and 6.5 % in noncirrhotic patients, respectively. However, 57.5 % of cirrhotics had evidence for FC. Among these patients, the APACHE-II score and the mortality rate were significantly higher than in cirrhotics without FC. Furthermore, FC was markedly more frequent in Child C compared to Child A and B cirrhotic patients (65 vs. 29 %; p = 0.002). These data are in line with previous data showing that the risk for FC increases with the severity of the liver disease and with hospitalisation. Furthermore, there is evidence that precipitating complications such as variceal bleeding increase the risk of infection [7, 8]. In our study, the rate of patients fulfilling criteria for IM is comparable to haematological patients. Moreover, the rate of FC was unexpectedly high. Particularly in immunocompromised and critically ill patients, the difficulty to differentiate between colonisation and infection is recognised: Although
clinical signs of severe infection may manifest early, but due to their low specificity the association with fungal infection may remain in-overt until late in the course of the disease. With regard to ‘‘early goaldirected therapy’’ and ‘‘hit hard and early’’ ‘‘in time’’ detection and treatment as known from septic patients or varices bleeding patients, it is quite challenging for IM in cirrhotics [8–10]. In general, the symptoms of fungal diseases depend on the type of infection, localisation within the body and the host’s response. The symptoms of (invasive) mycoses are not specific. Fever and chills that do not improve after antibiotic therapy are the most common symptoms. If the infection spreads to other organs or parts of the body such as kidneys, liver, bones, muscles, joints, spleen or eyes, additional symptoms may develop, which vary depending on the site of infection. Finally, infection not responding to treatment results in organ failure [8, 11]. Interpretation of these findings as a result of fungal infection is difficult in critically ill patients and particularly in cirrhotics, since these patients often have a variety of explanations for inflammation, new infiltrates or progression towards multiple organ failure [9, 12]. Therefore, early and advanced diagnosis of fungal infections in critically ill patients and particularly with LC is difficult. Obviously, this requires a high index of suspicion [13–15]. Therefore, scoring systems such as the candida colonisation index or the candida score can be helpful tools identifying possible risks for fungal infection. In this study, in both groups with FC, the candida colonisation index as well the candida score was elevated, suspicious for possible fungal infection (see Table 5). In our study, FC in patients with LC was most frequently found in ascites, urine and in the lungs, which were also the most frequent specific sites in previous studies. Fungi in ascites or urine can be easily detected by fungal culture or by microscope [16]. However, affection of the lungs is still challenging. IM still is diagnosed by histopathological examination of lung tissue obtained by thoracoscopy or open-lung biopsy as ‘‘gold standard’’ for the diagnosis of IM. Unfortunately, in patients with LC, the benefits of appropriate diagnostics by lung biopsies have to be outweighed against the risks of thrombocytopenia, impaired coagulation and increased respiratory distress [14].
123
Mycopathologia Table 5 Other risk factors for fungal colonisation and/or infection Patients with LC and FC (N = 69)
Patients with LC without FC (N = 51)
Control group without LC and FC (N = 31)
Control group without LC and without FC (N = 89)
27.6 (±16)
20.4 (±12)
14.4 (±8)
11.8 (±10)
Single therapy
5
16
15
54
2 or more than 2
64
35
16
35
3.5 ± 1.4
3.4 ± 1.5
2.7 ± 1.1
2±1
Length of ICU stay (days) Type of antibiotic therapy
Number of medical devices Haemodialysis
48
23
12
19
Mechanical ventilation
69
45
22
44
Mean candida colonisation index
0.65 ± 0.17
0.45 ± 0.17
Mean candida score
2.8 ± 0.9
1.8 ± 0.8
Although the diagnostic value of BAL for fungal evidence is controversial, BAL is considered to be safe and useful in high-risk patients with suspected IM [17]. The complete information derived from microscopy of BAL samples frequently helps to establish the diagnosis of IM. Moreover, demonstration of fungi in culture or direct microscopy, even if obtained from sputum or BAL fluid (a nonsterile site), has a high predictive value for IM, and for almost all practical purposes, it is valuable for diagnosis and treatment [17–19]. Beyond direct detection of fungus as described above, molecular methods are useful when fungal organisms cannot be isolated in culture, when subtle differences within a complex are not readily visible under light microscopy or when unstable phenotypic characteristics have been lost as may happen in chronic infections [20]. However, in most cases, molecular methods are extensive and their results are often imprecise especially in marginal conditions such as colonisation [20, 21]. To date, no sensitive and specific antibody assay has been developed that is useful for diagnosis. Since fungi—e.g., Candida—are part of the normal flora, healthy persons have antibodies against these organisms. This obviously decreases specificity of the assay [21]. On the other hand, an immunosuppressed host may be unable to generate a vigorous antibody response to Candida infection, thereby decreasing the sensitivity of the assay in this high-risk population. Thus, two nonculture-based assays, PCR and the beta-D-glucan assay, when combined with blood cultures, have the potential to increase the sensitivity of culture methods alone for IM [22–24].
123
Despite recent advances in molecular diagnosis, direct microscopy remains an essential tool for both detection of the organisms in the sample, and identification of the growing culture to the species level [22, 23]. Therefore, further studies have to be performed to describe the sensitivity of diagnostic testing compared with blood cultures alone, especially in immunocompromised patients and in conditions of FC [25, 26]. Therefore, we do not routinely perform these tests at present.
Limitations of the Study Although this study analyses one of the largest cohorts of cirrhotic patients with and without FC and compares these data to a noncirrhotic controls with comparable APACHE-II, this study was conducted retrospectively in a single medical ICU (monocentric study).
Conclusion Patients with LC have a high incidence of FC. These patients have a significantly higher mortality rate as compared to patients with LC without FC and matched noncirrhotic controls with comparable APACHE-II. Regrading their increased mortality, early antimycotic therapy should be considered particularly in patients with Child C cirrhosis even without definite prove of IM.
Mycopathologia Conflict of interest None of the authors has any potential financial conflict of interest related to this manuscript.
References 1. Donnelly JP. Symptoms and diagnosis of nosocomial fungal infections -state-of-the-art. Eur J Med Res. 2002;31:192–9. 2. Hsu JL, Ruoss SJ, Bower ND, Lin M, Holodniy M, Stevens DA. Diagnosing invasive fungal disease in critically ill patients. Crit Rev Microbiol. 2011;37:277–312. 3. Bonnel AR, Bunchorntavakul C, Reddy KR. Immune dysfunction and infections in patients with cirrhosis. Clin Gastroenterol Hepatol. 2011;9:727–38. 4. Tortorano AM, Dho G, Prigitano A, et al. ECMM-FIMUA Study Group invasive fungal infections in the intensive care unit: a multicentre, prospective, observational study in Italy (2006–2008). Mycoses. 2012;55:73–9. 5. Pittet D, Monod M, Suter PM, Frenk E, Auckenthaler R. Candida colonization and subsequent infections in critically ill surgical patients. Ann Surg. 1994;220(6):751–8. 6. Leroy G, Lambiotte F, The´venin D, et al. Evaluation of ‘‘Candida score’’ in critically ill patients: a prospective, multicenter, observational, cohort study. Ann Intensive Care. 2011;1(1):50. 7. Iber FL. Patients with cirrhosis and liver failure are at risk for bacterial and fungus infection. Am J Gastroenterol. 1999;94:2001–3. 8. Rolando N, Philpott-Howard J, Williams R. Bacterial and fungal infection in acute liver failure. Semin Liver Dis. 1996;16:389–402. 9. Dellinger RP, Levy MM, Rhodes A, et al. Surviving sepsis campaign: international guidelines for management of severe sepsis and septic shock, 2012. Intensive Care Med. 2013;39(2):165–228. 10. Kumar A, Safdar N, Kethireddy S, Chateau D. A survival benefit of combination antibiotic therapy for serious infections associated with sepsis and septic shock is contingent only on the risk of death: a meta-analytic/meta-regression study. Crit Care Med. 2010;38(8):1651–64. 11. O’Brien J, Triantos C, Burroughs AK. Management of varices in patients with cirrhosis. Nat Rev Gastroenterol Hepatol. 2013;10(7):402–12. 12. Gustot T, Durand F, Lebrec D, et al. Severe sepsis in cirrhosis. Hepatology. 2009;50(6):2022–33.
13. Marchetti O, Eggimann P, Calandra T. Invasive candidiasis in critically ill patients: does progressing knowledge improve clinical management and outcome? Curr Opin Crit Care. 2010;16:442–4. 14. Vazquez JA. Invasive fungal infections in the intensive care unit. Semin Respir Crit Care Med. 2010;31:79–86. 15. Pema´n J, Zaragoza R. Current diagnostic approaches to invasive candidiasis in critical care settings. Mycoses. 2010; 53:424–33. 16. Choi SH, Soo Kim Y, Chung JW, et al. Clinical significance of untreated Candida species isolated from ascites in cirrhotic patients. Scand J Infect Dis. 2004;9:649–55. 17. Navasa M, Rimola A, Rodes J. Bacterial infections in liver disease. Semin Liver Dis. 1997;17:323–33. 18. Yazıcıog˘lu Moc¸in O, Karakurt Z, Aksoy F et al. Bronchoscopy as an indicator of tracheobronchial fungal infection in non-neutropenic intensive-care unit patients. Clin Microbiol Infect. 2013; 19(3):E136–41. 19. Falcone M, Massetti AP, Russo A, Vullo V, Venditti M. Invasive aspergillosis in patients with liver disease. Med Mycol. 2011;49:406–13. 20. Prodanovic H, Cracco C, Massard J, et al. Invasive pulmonary aspergillosis in patients with decompensated cirrhosis: case series. BMC Gastroenterol. 2007;31:2. 21. Charles PE, Castro C, Ruiz-Santana S, Leo´n C, Saavedra P, Martı´n E. Serum procalcitonin levels in critically ill patients colonized with Candida spp: new clues for the early recognition of invasive candidiasis? Intensive Care Med. 2009; 35:2146–50. 22. Badiee P, Alborzi A, Joukar M. Molecular assay to detect nosocomial fungal infections in intensive care units. Euro J Intern Med. 2011;22:611–5. 23. Badiee P, Alborzi A, Joukar M. Molecular assay to detect nosocomial fungal infections in intensive care units. Eur J Intern Med. 2011;22:611–5. 24. Lipke AB, Mihas AA. Non-decompensated cirrhosis as a risk factor for invasive aspergillosis: a case report and review of the immune dysfunction of cirrhosis. Am J Med Sci. 2007;334:314–6. 25. Echeverria PM, Kett DH, Azoulay E. Candida prophylaxis and therapy in the ICU. Semin Respir Crit Care Med. 2011; 32:159–73. 26. Falcone M, Massetti AP, Russo A, Vullo V, Venditti M. Invasive aspergillosis in patients with liver disease. Med Mycol. 2011;4:406–13.
123
