Introduction
The pathogenicity of yeasts isolated in the peritoneal fluid of patients with complicated intra-abdominal infections (and especially community-acquired infections (CAIs)) is subject to debate. There are some data to suggest that yeasts have an impact on the outcome in cases of peptic ulcer perforation [
1]. Non-postoperative nosocomial intra-abdominal infections (NPNIAIs) share certain microbiological characteristics with CAIs [
2], such as the frequency of yeast isolation in peritoneal fluid samples. In view of the high mortality rate observed in patients with post-operative infections [
3] and patients with organ failure admitted to the intensive care unit (ICU) [
4], the pathogenicity of yeast in these contexts has been investigated more. However, the guidelines on the management of these types of infection are essentially based on extrapolation of data on candidemia [
5,
6]. Even though yeast intra-abdominal infections can be candidemic [
7], the frequency is low and the diseases do not have the same course [
8]. A number of scores have been developed in order to predict the occurrence of candidemia in high-risk patients, including the colonization index [
9], Leon
et al.’s
Candida score [
10] and a clinical rule [
11]. However, none of these scores is suitable for complicated intra-abdominal infections. Furthermore, each of these scores has a high negative predictive value (NPV, for ruling out yeast infection) rather than a high positive predictive value (PPV, for initiating treatment) [
12]. Ten years ago, Dupont
et al. developed a score for severe complicated intra-abdominal infections in the ICU [
13]. This is still the only available score with a moderately good PPV and NPV [
13]. However, it was developed in a severe population of ICU patients. There are few data on less severe patients having undergone emergency surgery. The objectives of the present study were to (i) build a predictive score for yeast isolation in the peritoneal fluid of patients with complicated non-postoperative intra-abdominal infections (CNPIAI) in a retrospective cohort of patients and (ii) validate the score in a separate prospective cohort. The new score will be compared with previously described scores. The relationship between intra-abdominal candidiasis (IAC) and the outcomes for patients with complicated CNPNIAIs was also evaluated.
Discussion
Our present results show that a YP peritoneal fluid culture is associated with worst outcomes and increased mortality in patients with CNPIAI. The prevalence of yeast isolation in this context is low (15.6%). Four parameters were independently associated with IAC: length of stay before surgery ≥48 h, per-operative cardiovascular failure, generalized peritonitis and upper gastrointestinal tract perforation. The predictive score has a high NPV and thus can be used to rule out the presence of yeast in the peritoneal fluid. Hence, this score may constitute an easy-to-use bedside tool that enables the physician to avoid the initiation of inappropriate antifungal treatment.
Although yeasts are undoubtedly pathogenic in postoperative infections [
3], this question is subject to debate in the context of CAIs and there are few data on this specific topic. In a case-control study, there was no significant difference in outcome between YP and YN CAIs [
3]. However, the study featured a small number of patients. In a study of patients with organ failure admitted to the ICU, a YP culture was found to be associated with elevated mortality [
4]. Another study reported a significantly greater proportion of septic shock in CAIs when yeast was detected in the peritoneal fluid culture [
23]. The overall prevalence of a YP culture in our study population (15.6%) is similar to the mean value reported in the literature (with values ranging from 4% to 43.4% in studies of CAIs) [
2,
24]. The prevalence clearly depends on the study population in question. For example, the prevalence was very high in two studies that focused solely on peptic ulcer perforations [
1,
24]. Furthermore, yeast isolation was associated with increased morbidity and mortality in these studies (21.7% vs. 3.4 % [
1] and 33.3% vs. 14.6% [
24] for YP and YN patients, respectively).
In the present study, four parameters were found to be independently associated with IAC of patients with a CNPIAI. Interestingly, upper gastrointestinal tract perforation and per-operative cardiovascular failure were previously included in a predictive score for severe intra-abdominal infections in the ICU [
4]. Another Spanish study reported these two risk factors in 74
Candida peritonitis patients [
25]. The two other parameters in the latter study (female gender and ongoing antimicrobial therapy for more than 48 hours) were not significant in our study [
4]. However, ongoing antimicrobial therapy was associated with yeast isolation in our univariate analysis; this was probably due to the low prevalence of antimicrobial therapy because few patients had hospital-acquired infections. It has been well established that ongoing antimicrobial therapy is a recognized risk factor for candidemia [
26]. However, in the present study, hospital length of stay ≥48 h before surgery was found to be independently associated with yeast isolation. There are no literature data on why yeast isolation is more frequently associated with generalized peritonitis than with localized infection. It could be only due to the major impact of appendicitis infections in CAI. The population investigated in the present study is mild to moderate with only a 33.9% rate of ICU admission and a low mortality of 11.3%. It is very different to the Dupont
et al. [
13] or Leon
et al. [
10] studies with a 100% rate of ICU admission and reported mortality of 43% and 33%, respectively.
The score developed in the present study has a good NPV. It may be important to avoid the inappropriate initiation of antifungal treatments that are costly and whose impact on resistance is not well known [
27]. All the previously published scores or clinical rules were developed to predict the occurrence of candidemia (even in high-risk surgical patients) [
9-
11]. These scores have much the same operational values as our present score - notably with very high NPV for avoiding treatment. The high NPV is essentially due to the low prevalence of the disease. However, the predictive value of a score has been shown to be better for severe intra-abdominal infections in the ICU than previously described for candidemia, with a PPV of 67% and a NPV of 72% [
13]. The colonization index and the
Candida score were recently tested for the prediction of blood culture-negative IAC but had very poor operational values [
28]. Blood levels of ß-glucan may be of value for the diagnosis of postoperative infections in high-risk surgical patients [
28]. However, there are no data on the value of ß-glucan levels in CAIs. Furthermore, it may not be enough to know the change over time in ß-glucan levels in CAIs because the physician has to decide at bedside whether antifungal treatment must be initiated or not. We did not have access to a ß-glucan assay during our study. It has been suggested that a combination of high levels of ß-glucan and the
C. albicans germ tube antibody can differentiate between
Candida colonization and invasive candidiasis in ICU patients with severe abdominal conditions [
29]. In the latter study, fewer than 50% of the patients had CAIs. Furthermore, the study was designed to assess the course of candidiasis in the ICU, rather than to predict the presence of candidiasis on admission.
A worse outcome associated with YP culture of the peritoneal fluid was evidenced in this study. Furthermore, it was independently associated with mortality. It is the first report in the literature of such impact in mild to moderate infections.
Our study had some potential limitations. First, this was a single-center study. Nevertheless, our center is a large tertiary-care hospital with an experienced, trained team for the care of patients with complicated intra-abdominal infections. Furthermore, peritoneal fluid samples were available for all patients and all were sent for microbiological and mycological culture. However, the present study’s results must be validated in multicenter trials. Recently, the sensitivity breakpoints for
Candida spp. were modified according to the species [
30]. This study used previously described breakpoints (>32 μg.ml
−1) that could have underestimated the rate of strains resistant to fluconazole. Our study focused on the development of a predictive score and thus did not address the question of how best to treat IAC. It was only associated with increased morbidity and mortality in our population. The comparison of mortality rates according to treatment or not in the group of patients with IAC is not very relevant due to many confounding factors and a clear lack of power. The American guidelines for the treatment of abdominal candidiasis are essentially derived from candidemia guidelines [
6]. Recently, an Italian group drew up a list of recommendations for the treatment of IAC [
5]. However, in view of the lack of scientific evidence, the vast majority of guidelines are based on expert opinion or extrapolation of data on candidemia [
8]. Lastly, the use of the new score is not well calibrated for critical care patients because only one-third of the cohort was admitted to the ICU.
This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data.
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
HD designed the study, participated in data acquisition and analyses, and wrote the manuscript. MG and EZ participated in data acquisition, data analyses, and preparation of the manuscript. AN, MP, SP, JMR, TC and YM participated in the data analyses and drafting of the manuscript. All authors read and approved the manuscript.