JAC Advance Access originally published online on July 2, 2007
Journal of Antimicrobial Chemotherapy 2007 60(3):619-624; doi:10.1093/jac/dkm255
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Antifungal use in intensive care units
1 Institute of Environmental Medicine and Hospital Epidemiology, University Medical Centre Freiburg, Freiburg University Hospital, Breisacher Str. 115B, 79106 Freiburg, Germany 2 Institute of Hygiene and Environmental Medicine, Charité—University Medicine Berlin, Campus Benjamin Franklin, Hindenburgdamm 27, 12203 Berlin, Germany 3 National Reference Centre for Surveillance of Nosocomial Infections, Heubnerweg 6, 14059 Berlin, Germany 4 Institute of Medical Microbiology and Hospital Epidemiology, Medical School Hannover, Carl-Neuberg-Str. 1, 30625 Hannover, Germany 5 Department of Anesthesiology and Intensive Care Medicine, Tubingen University Hospital, Hoppe-Seyler-Str. 3, 72076 Tubingen, Germany
* Corresponding author. Tel: +49-761-270-8271; Fax: +49-761-270-8253; E-mail: elisabeth.meyer{at}uniklinik-freiburg.de
Received 16 April 2007; returned 11 May 2007; revised 13 June 2007; accepted 16 June 2007
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and methodsResultsDiscussionTransparency declarationsSupplementary dataReferences |
|---|
Objectives: To provide benchmarking data on antifungal use in intensive care units (ICUs), to analyse risk factors and to look for correlations with antibiotic use data and structure parameters.
Methods: Antimicrobial use data for 13 ICUs were obtained from computerized databases from January 2004 through June 2005. Antimicrobial usage density (AD) is expressed as daily defined doses/1000 patient-days. Correlations were calculated by the Spearman correlation or for binomic variables by the two-sided Wilcoxon test. A multivariate regression analysis was performed to identify independent risk factors for the outcome ‘antifungal use’.
Results: Mean systemic antifungal drug use was 93.0, the range being between ADs of 18.9 and 232.2. ICUs treating transplant patients had a significantly higher mean antifungal usage at 152.9 compared with ICUs not treating transplant patients where the AD was 46.0. Fluconazole was the most frequently prescribed antifungal (mean AD 69.6) followed by amphotericin B (11.4) and voriconazole (6.2). Antifungal use correlated significantly with the consumption of quinolones, carbapenems and extended-spectrum penicillins, but not with total antibiotic use and not with the type of ICU or university status. In the multivariate linear regression analysis, two parameters, i.e. high quinolone use (P = 0.002) and ICUs which treat transplant patients (P = 0.027), were independent risk factors for a high level of antifungal use.
Conclusions: Antifungal use was heterogeneous in German ICUs with the mean AD lying at 93. Benchmarking data might provide a useful method for assessing strategies that aim to reduce antifungal use in ICUs. However, data should be stratified for ICUs with and without transplant patients.
Keywords: surveillance , defined daily doses , benchmarking data , fluconazole , amphotericin B , voriconazole
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionTransparency declarationsSupplementary dataReferences |
|---|
During the 1980s, numerous reports described a dramatic increase in the frequency of severe life-threatening infections caused by yeasts, and especially Candida spp.1 Candida albicans is currently the seventh most common nosocomial pathogen in German intensive care units (ICUs) and the third most common bloodstream pathogen in US ICUs.2 The introduction of fluconazole in 1990, a well-tolerated antifungal with a broad spectrum of activity against yeast, lowered the threshold for the prescription of antifungals by physicians. Some investigators suggested that the indiscriminate use of fluconazole would select resistant strains and species.3
This collateral damage and the emerging bacterial resistance caused the World Health Organization (WHO) to address this problem in the ‘WHO Global Strategy for Containment of Antimicrobial Resistance’ in 2001 and to make demands for the appropriate use of antimicrobial agents.4 Benchmarking data provide a basis for the analysis and comparison of consumption data. Such surveillance systems are already in place for antibiotic use and bacterial resistance.5–7 However, there is a great scarcity of benchmarking data on antifungal consumption in high-risk areas.8 Surveillance data are essential for developing policies and programmes and for evaluating their effectiveness in preventing and efficiently bringing public health problems under control. If, for instance, ICUs do not have benchmarking data on antimicrobial use, they will not know whether they have high or outlier consumption in comparison to other ICUs.
The aim of this study was to provide such benchmarking data on antifungal use in ICUs and to look for the correlations and risk factors relating to antifungal use against antibiotic use and against structure parameters of ICUs.
|
Materials and methods |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionTransparency declarationsSupplementary dataReferences |
|---|
From January 2004 through June 2005, data on adult use of antifungals and antibiotics in 13 ICUs were obtained from computerized hospital databases. Initially, 45 ICUs participating in project SARI (Surveillance on Antibiotic Use and Resistance in Intensive Care Units) were invited to provide data, but only 13 ICUs agreed to do so. Data on the use of antifungals were collected according to the SARI protocol.7 Data on fungal isolates from primary sterile compartments were obtained directly either from the microbiology laboratory or from the person responsible for SARI. This might have been the intensive care physician, the microbiologist or the infection control physician. None of the ICUs treating transplant patients used fluconazole to change anti-rejection levels.
Consumption was expressed as daily defined doses (DDDs) and normalized per 1000 patient-days (AD=antimicrobial usage density), one DDD being the standard adult dose of an antimicrobial agent for 1 day treatment defined by the WHO [Table S1, available as Supplementary data at JAC Online (http://jac.oxfordjournals.org)]. The following structure and process parameters were reported: size of hospital (< or 
1000 beds), number of ICU beds (< or 12), type of ICU (medical, surgical, interdisciplinary), affiliation status (university hospital yes/no), treatment of transplant patients in the ICU (yes/no), incidence density of fungi, i.e. the number of non-duplicate fungal isolates from primary sterile compartments/1000 patient-days, length of stay (days), bed utilization and (endotracheal) tube utilization. ‘Primary sterile compartment’ was defined as isolates originating from blood cultures, liquor and intra-abdominal or pleural biopsies or punctures.
In the univariable analysis, associations were tested between antifungal use and antibiotic use, structure and process parameters. We determined the Spearman correlation rs for the association between antifungal use and all continuous parameters, e.g. antibiotic use of all antibiotic groups. For dichotomous variables, e.g. university hospital yes or no, we used the two-sided Wilcoxon test. The significance level was set at P = 0.05.
A multiple linear regression model was carried out to identify independent risk factors for the outcome ‘antifungal use’. All continuous variables were logarithmically transformed. The stepwise forward selection was used with significance level P = 0.05 for entering the variable into the model and P = 0.10 for removing the variable. The R2-statistic was used to measure how well the regression explains the data.
Data analysis was performed using SAS version 9.01 and SPSS version 12.0.1.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionTransparency declarationsSupplementary dataReferences |
|---|
Five of the 13 ICUs were located in university and seven in teaching hospitals. Five were surgical and four each medical and interdisciplinary. Six belonged to hospitals with more than 1000 beds. Systemic antifungal drug use ranged between an AD of 18.9 and 232.2, with the pooled mean lying at 93.0 (Table 1). Cumulative data from January 2004 to June 2005 revealed that fluconazole was the most frequently prescribed antifungal (mean AD 69.6) followed by amphotericin B (mean AD = 11.4) and voriconazole (mean AD = 6.2) (Table 2).
|
|
Correlation of antifungal use data with antibiotic use showed a significant positive correlation with the consumption of quinolones (rs = 0.830, P < 0.001), carbapenems (rs = 0.621, P = 0.024) and penicillins with extended spectrum (rs = 0.588, P = 0.035) (Table 3). Additionally, there was a significant correlation between the total incidence density of fungi from primarily sterile compartments (rs = 0.693, P = 0.026) and ICUs treating transplant patients (P = 0.030). The statistically significant results with a correlation coefficient > 0.6 are demonstrated in Figure 1(a–c). In Figure 1(c), the ICU with the highest incidence density of fungi (2.13 isolates/1000 patient-days) uses antifungals with an AD of 152.9. In contrast, the ICU with the highest antifungal use, i.e. 232.2, has an incidence density of fungi of 0.81.
|
|
The correlation was also carried out for ICUs (n = 10) that had provided data on the number of non-duplicate fungal isolates obtained from primary sterile compartments. The correlation with antifungal use remained statistically significant for the use of quinolones (rs = 0.794, P = 0.006) and carbapenems (rs = 0.661, P = 0.038) and for the incidence density of fungi (rs = 0.693, P = 0.023) and for ICUs treating transplant patients (P = 0.019). However, it was no longer statistically significant for extended-spectrum penicillins (rs = 0.527, P = 0.117).
In the multivariate linear regression analysis, high quinolone use (P = 0.002) and ICUs treating transplant patients (P = 0.027) were independent risk factors for high antifungal use. The goodness of fit of the model was R2 = 0.79.
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionTransparency declarationsSupplementary dataReferences |
|---|
Although a major increase has been documented in the consumption of antifungals since 1980, benchmarking data on antifungal consumption in ICUs are scarce. Surveillance data on other antimicrobials, i.e. antibiotics, exist and can be used for comparison and analysis of consumption data in one's own hospital.5,9,10 For reasons of national and international comparison, all these projects use DDDs as recommended by the WHO. However, the drawback is that DDDs do not necessarily reflect the doses of some antifungals currently administered and that this in turn may result in an overestimation of antifungal use compared with the expression of antifungal use as prescribed daily doses (PDDs).
In our study, the mean systemic antifungal AD was 93.0, and featured a broad range from 18.9 to 232.2. This is much lower than that reported by Cook et al. for a 730 bed hospital in the USA. Before the introduction of an antimicrobial intervention programme, hospital-wide total antifungal use lay at 144 DDD/1000 patient-days. After implementation of the programme, consumption dropped by 28% to 103 DDD/1000 patient-days.11 de With et al.8 calculated the antifungal use for five German university hospitals by PDDs/100 patient days. PDDs for most drugs were set twice as high as WHO DDDs (e.g. fluconazole PDD was 400 mg). They found a mean antifungal PDD/100 patient-days of 18.3, i.e. 183/1000 patient-days, in five medical ICUs and 10.7/100 patient-days in five surgical ICUs. This is in line with our data on mean antifungal consumption, as well as the fact that fluconazole remained the most widely used drug despite the availability of new antifungal drugs. However, we could not see a positive correlation between antifungal use and structural parameters such as ‘medical ICU’ or ‘ICU located in a university hospital’. In contrast, the parameter ‘ICU treating transplant patients’ was revealed to be an independent risk factor for high antifungal use, which is used for these patients both for therapy and for prophylaxis. Therefore, it might be useful to differentiate between ICUs with and without transplant patients if data are generated for comparison.
Obviously, a positive correlation between the incidence density of fungi and antifungal use was expected. However, there was still a broad range between those two parameters in different ICUs. The clinical difficulty in discriminating between fungal infection and colonization is a well-known dilemma, especially in the vulnerable group comprising ICU patients, although there are scoring systems or new 1,3 beta-glucan blood tests for the diagnosis of invasive fungal infection.12,13 Nevertheless, benchmarking data can give a reason to analyse the indication for antifungal use in the ICU.
Interestingly, our data showed a significant positive correlation between total antifungal consumption and that of the broad-spectrum antibiotics such as quinolones, carbapenems and extended-spectrum penicillins, but not with the total antibiotic use. In the multivariate analysis with stepwise forward selection, we found quinolone use to be an independent risk factor. In general, broad-spectrum antibiotic use is a well-described risk factor for fungal colonization and infections.3,14 In an Israeli teaching hospital, the correlation of antibiotic use and candiduria was studied: for broad-spectrum antibiotics, the corresponding correlation coefficient was 0.66 (P = 0.001), but the strongest correlations were shown for the use of meropenem (r = 0.79, P < 0.001) and ceftazidime (r = 0.66, P = 0.001).15
The limitations of this study are: first, the incidence density of fungi from primary sterile sites might be overestimated because not only histologically confirmed intra-abdominal cultures were included; secondly, the results of this study might constitute a local phenomenon that might not extrapolate to other ICUs or countries; thirdly, correlations cannot prove a causal relationship and neither can quantitative data give information on quality. The appropriateness of therapy can only be evaluated by audits as was performed by Natsch et al.16
Nevertheless, significant associations might help in identifying ‘useful versus not useful’ interventions. These benchmarking data on antifungal use in ICUs might provide a useful method for assessing ICU strategies that aim to reduce antifungal use. Furthermore, surveillance data are essential for developing policies and programmes and for evaluating their effectiveness in either preventing public health problems or effectively bringing them under control. However, if surveillance data for antifungal use are generated, it might be reasonable to differentiate between ICUs with and without transplant patients.
|
Transparency declarations |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionTransparency declarationsSupplementary dataReferences |
|---|
None to declare.
|
Supplementary data |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionTransparency declarationsSupplementary dataReferences |
|---|
Table S1 is available as Supplementary data at JAC Online.
|
Acknowledgements |
|---|
The authors would like to thank all the ICUs that provided data for this study. All the ICUs participated in SARI (Surveillance of Antimicrobial use and antimicrobial Resistance in German Intensive Care Units), a project which is supported by a grant from the Federal Ministry of Education and Research.
|
References |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionTransparency declarationsSupplementary dataReferences |
|---|
1 Jarvis WR, Edwards JR, Culver DH, et al. Nosocomial infection rates in adult and pediatric intensive care units in the United States. National Nosocomial Infections Surveillance System. Am J Med (1991) 91:185–91.[CrossRef]
2 Wisplinghoff H, Bischoff T, Tallent SM, et al. Nosocomial bloodstream infections in US hospitals: analysis of 24,179 cases from a prospective nationwide surveillance study. Clin Infect Dis (2004) 309–17.
3 Ostrosky-Zeichner L, Pappas PG. Invasive candidiasis in the intensive care unit. Crit Care Med (2006) 34:857–63.[CrossRef][Web of Science][Medline]
4 World Health Organisation. WHO Global Strategy for Containment of Antimicrobial Resistance 2001. http://www.who.int/drugresistance/en (13 April 2007, date last accessed).
5 National Nosocomial Infections Surveillance System. National Nosocomial Infections Surveillance (NNIS) system report, data summary from January 1992 through June 2004. Am J Infect Control (2004) 32:470–85.[CrossRef][Web of Science][Medline]
6 The Danish Integrated Antimicrobial Resistance Monitoring, Research Programme. DANMAP 2004—Use of Antimicrobial Agents and Occurrence of Antimicrobial Resistance in Bacteria from Food Animals, Foods and Humans in Denmark. http://www.dfvf.dk/Default.aspx?ID=9604 (13 April 2007, date last accessed).
7 Meyer E, Jonas D, Schwab F, et al. Design of a surveillance system of antibiotic use and bacterial resistance in German intensive Care units (SARI). Infection (2003) 31:208–15.[Web of Science][Medline]
8 de With K, Steib-Bauert M, Knoth H, et al. Hospital use of systemic antifungal drugs. BMC Clin Pharmacol (2005) 5:1.[CrossRef][Medline]
9 Meyer E, Schwab F, Jonas D, et al. Surveillance of antimicrobial use and antimicrobial resistance in intensive care units (SARI): 1. Antimicrobial use in German intensive care units. Intensive Care Med (2004) 30:1089–96.[CrossRef][Web of Science][Medline]
10 Vander Stichele RH, Elseviers MM, Ferech M, et al. European surveillance of antimicrobial consumption (ESAC): data collection performance and methodological approach. Br J Clin Pharmacol (2004) 58:419–28.[CrossRef][Web of Science][Medline]
11
Cook PP, Catrou PG, Christie JD, et al. Reduction in broad-spectrum antimicrobial use associated with no improvement in hospital antibiogram. J Antimicrob Chemother (2004) 53:853–9.
12 Leon C, Ruiz-Santana S, Saavedra P, et al. A bedside scoring system (‘Candida score’) for early antifungal treatment in nonneutropenic critically ill patients with Candida colonization. Crit Care Med (2006) 34:730–7.[CrossRef][Web of Science][Medline]
13 Lipsett PA. Surgical critical care: fungal infections in surgical patients. Crit Care Med (2006) 34:215–24.[CrossRef]
14 Dancer SJ. How antibiotics can make us sick: the less obvious adverse effects of antimicrobial chemotherapy. Lancet Infect Dis (2004) 4:611–9.[CrossRef][Web of Science][Medline]
15 Weinberger M, Sweet S, Leibovici L, et al. Correlation between candiduria and departmental antibiotic use. J Hosp Infect (2003) 53:183–6.[CrossRef][Web of Science][Medline]
16
Natsch S, Steeghs MH, Hekster YA, et al. Use of fluconazole in daily practice: still room for improvement. J Antimicrob Chemother (2001) 48:303–10.
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||