Evaluation of anidulafungin in the treatment of intra-abdominal candidiasis: a pooled analysis of patient-level data from 5 prospective studies

The incidence of nosocomial invasive fungal infections involving Candida spp. has increased markedly in recent years in patients undergoing abdominal surgery. This post hoc analysis aimed to determine the efficacy and safety of anidulafungin treatment in patients with intra-abdominal candidiasis (IAC) from five prospective studies (one comparative and four open-label) of adult surgical patients with microbiologically confirmed Candida intra-abdominal infection. Patients received an intravenous (IV) loading dose of anidulafungin 200 mg, followed by a daily 100-mg maintenance dose. Per study protocols, some patients could be switched to an oral azole after ≥ 5 or ≥ 10 days of IV treatment. Antifungal treatment was maintained for ≥ 14 days after the last positive Candida culture and resolution of symptoms. The global response rate (GRR) at the end of IV treatment (EOIVT) was the primary endpoint. GRR at the end of therapy (EOT), all-cause mortality at days 14 and 28, and safety was also evaluated. Seventy-nine patients had IAC from peritoneal fluid or hepatobiliary tract. C. albicans (72.2%) and C. glabrata (32.9%) were the most common pathogens. Overall GRR was 73.4% and 67.1% at EOIVT and EOT, respectively. All-cause mortality was 17.7% at day 14 and 24.1% at day 28 in the modified intent-to-treat population. Anidulafungin was well tolerated in this population, with most adverse events mild or moderate in severity. In these patients with IAC, anidulafungin showed a GRR at EOIVT similar to the anidulafungin registrational trial, and the results of our analysis confirmed the known safety profile of anidulafungin. ClinicalTrials.gov registration number NCT00496197, registered July 3, 2007, https://clinicaltrials.gov/ct2/show/study/NCT00496197; ClinicalTrials.gov registration number NCT00548262, registered October 19, 2007, https://clinicaltrials.gov/ct2/show/record/NCT00548262; ClinicalTrials.gov registration number NCT00537329, registered September 25, 2007, https://clinicaltrials.gov/ct2/show/record/NCT00537329; ClinicalTrials.gov registration number NCT00689338, registered May 29, 2008, https://clinicaltrials.gov/ct2/show/study/NCT00689338; ClinicalTrials.gov registration number NCT00805740, registered November 26, 2008, https://clinicaltrials.gov/ct2/show/NCT00805740 Electronic supplementary material The online version of this article (10.1007/s10096-019-03617-9) contains supplementary material, which is available to authorized users.


Introduction
The incidence of nosocomial invasive fungal infections involving Candida species has increased in recent years because of the increasing number of immunocompromised patients, including patients with cancer, patients undergoing transplants, patients with human immunodeficiency virus and patients in intensive care units (ICUs) [1][2][3][4].
Clinical data on intra-abdominal candidiasis (IAC) are still scarce, although IAC is a common type of invasive candidiasis following candidaemia [5]. IAC is confirmed in patients when (i) there is clinical evidence of intra-abdominal infection and (ii) Candida isolates are collected from an intra-abdominal site under sterile conditions within 24 h [5,6]. IAC, or Candida-associated peritonitis, is a common cause of mortality in patients in the ICU [1,7,8]. Previous studies have demonstrated the predominance of C. albicans isolates (approximately 75%), followed by C. glabrata, in intra-abdominal Candida infections in surgical patients [9,10].
High rates of non-albicans Candida isolates from abdominal samples have been reported in patients in the ICU [7,11]. In patients admitted to ICUs with post-operative peritonitis, multidrug-resistant strains were frequent and increased with the number of reoperations for persistent abdominal sepsis [12]. In a study of patients with post-operative peritonitis following bariatric surgery, 25 out of 61 patients were positive for Candida infections (6 [22%] were fluconazole-resistant C. glabrata) and 36 had no Candida infection [13]. Candida infections were isolated more often in late-onset peritonitis and were associated with multidrug-resistant bacteria. Another study demonstrated that C. albicans was cultured in 11/18 cases in a bariatric surgery group and in 23/46 cases in a conventional surgery group, whereas C. glabrata was cultured in 13/46 cases in the conventional surgery group and 3/6 cases were resistant to fluconazole [14]. A recent, prospective, single-centre, population-based study in a surgical ICU setting observed that IAC was a frequent form of invasive candidiasis (13/22 episodes in 1149 patients) and that antifungal therapy and good source control led to a good outcome. C. albicans and C. parapsilosis were the most common invasive species, and resistance to fluconazole and itraconazole was noted in 3/22 (13.6%) cases [15]. To our knowledge, no studies have been published assessing the efficacy of antifungal agents in patients with microbiologically documented IAC.
Among the echinocandin class of antifungal agents, anidulafungin is a cyclic lipopeptide approved for the treatment of candidaemia and other forms of invasive candidiasis such as intra-abdominal abscesses and peritonitis. In the pivotal study of anidulafungin [16], there were only a few patients with microbiologically confirmed Candida deepseated tissue infections. To better understand the efficacy and safety of treatment with anidulafungin in these patients, data from five prospective studies were pooled and analysed.
Here, we present the results of treatment with anidulafungin in the group of patients with microbiologically documented IAC.

Ethics
All primary studies were conducted in compliance with the Declaration of Helsinki and Good Clinical Practice Guidelines established by the International Conference on Harmonization. The final protocols, amendments and informed consent documentation were reviewed and approved by the Institutional Review Boards and the Independent Ethics Committees of the investigational centres. All patients provided written informed consent.

Study design and treatment
Patient-level efficacy and safety data were retrospectively pooled from four open-label, non-comparative studies [17][18][19][20] and one comparative study (Pfizer data on file) with the aim of analysing the safety and efficacy of anidulafungin in the treatment of invasive candidiasis in adult patients. The comparative study was a double-blind, double-dummy, randomised, multicentre trial comparing anidulafungin with caspofungin in patients with microbiologically confirmed deep-seated tissue infection due to Candida species. All the studies had similar protocols and endpoints, and investigated the use of anidulafungin in a broad range of patients, including patients in the ICU and patients with neutropenia (absolute neutrophil count [ANC] < 500 cells/mm 3 ) with microbiologically confirmed Candida intra-abdominal infections ( Table 1). All studies were registered with ClinicalTrials.gov (NCT00496197, NCT00548262, NCT00537329, NCT00689338, and NCT00805740).
Patients initiated the treatment with a 200-mg loading dose of intravenous (IV) anidulafungin on day 1 of the study, followed by a maintenance dose of 100 mg IV once daily. In all but one of the studies (A8851022 [NCT00805740]), patients could be switched to an oral azole after ≥ 5 days (A8851011 [NCT00496197], A8851015 [NCT00548262], A8851016 [NCT00537329]) or ≥ 10 days (A8851019 [NCT00689338]) of IV treatment, according to pre-specified criteria in each protocol. Antifungal treatment (IV anidulafungin plus subsequent oral azole if used) was maintained for ≥ 14 days after the last positive Candida culture and resolution of symptoms.

Patients-criteria for the pooled analysis
The pooled analysis included male or female patients aged ≥ 18 years with culture-confirmed IAC from a culture specimen obtained within the preceding 96 h from a normally sterile site or newly placed drain. Patients could enter the studies based on microbiological evidence suggestive of Candida infection (e.g. positive blood culture for yeast). However, confirmation of Candida species was required within 96 h to remain in the study. Patients were also required to have more than one of the following clinical signs and symptoms of fungal infection: fever, defined as an oral or tympanic temperature > 38°C; hypotension, defined as systolic blood pressure < 100 mg Hg or a decrease in systolic blood pressure of > 30 mmHg from baseline; and clinical signs of localised or generalised peritonitis including the presence of intraabdominal abscess or purulent fluid from drains. Patients who had received > 48 h of prior antifungal therapy had the presence of confirmed or suspected Candida osteomyelitis, endocarditis or meningitis, had prosthetic devices at infection sites that could not be removed within 24 h of study entry or who had previously failed treatment for the current episode of Candida infection were excluded. From the pooled database, only patients with peritoneal fluid or hepatobiliary tract as sites of infection, as reported by the investigators in the case report form (CRF), were included in this analysis.

Study endpoints
The primary endpoint of this analysis was the successful global response at the end of IV treatment (EOIVT), based on the modified intent-to-treat (MITT) population. The MITT population included all patients who received ≥ 1 dose of anidulafungin (intent-to-treat population) and who had a confirmed diagnosis of invasive candidiasis/candidaemia and a positive culture for Candida species within 96 h of study entry. A global response was considered successful if patients achieved both clinical success (defined as the resolution of signs and symptoms of Candida infection and no additional systemic or oral antifungal therapy required) and microbiological success (defined as eradication or presumed eradication of Candida species present at baseline).
Secondary endpoints of this analysis included the global response at the end of all therapy (EOT) in the MITT population, and the rate of all-cause mortality at day 14 and day 28 after initiation of IV therapy in the MITT populations. Treatment-emergent adverse events (TEAEs) and serious adverse events (SAEs) of all causalities in the MITT population were recorded in all studies and were summarised by Medical Dictionary for Regulatory Activities preferred term.

Statistical analysis
Efficacy and safety analyses were for descriptive purposes, and no hypotheses for efficacy endpoints were tested. Evaluation of the data comprised primarily of summary descriptive statistics. Success rates for global response were estimated with exact 95% confidence intervals (CI) for binomial proportion (Clopper-Pearson method). In efficacy analyses, indeterminate or missing was considered failure.
Multivariate logistic regression was used to calculate odds ratios and 95% CI to identify factors that were significantly related to treatment failure in the MITT population. Risk factors for developing candidiasis were included in the protocols of all studies. Available patient characteristics data included in the multivariate analysis were age, baseline Acute Physiology and Chronic Health Evaluation (APACHE) II score, body mass index and baseline Candida species (mono-vs polypathogenic infections).

Patients
The 5 studies, summarised in Table 1, included 129 anidulafungin-treated patients with deep-seated tissue infection (the MITT population). Of these, 79 patients had microbiologically confirmed intra-abdominal infection with Candida spp. The most frequent baseline pathogens ( Table 2) were C. albicans (72.2%) and C. glabrata (32.9%), although patients could have more than one pathogen. In vitro minimum inhibitory concentration data for anidulafungin, along with susceptibility to anidulafungin, fluconazole and voriconazole, are shown in Online Resource 1. The peritoneal cavity (70/79; 88.6%) was the most common site of infection. Among the nine patients (9/79; 11.4%) with infections in the hepatobiliary tract, two had post-liver transplant complications and seven had biliary infections following liver or biliary tract interventions for malignancies. Ten patients (10/79; 12.7%) also had candidaemia, and three patients had more than one site of infection (peritoneal cavity plus abdomen; peritoneal cavity plus other; peritoneal cavity plus pleural cavity plus other). Eight patients (two hepatobiliary, six peritoneal cavity) had concomitant bacterial infections with bacteria isolated from microbiological culture and reported in the CRF. However, almost all the patients (77/79) were also receiving systemic concomitant antibiotic treatment.

Outcomes
In the MITT population, the overall global response rate (GRR) was 73.4% at EOIVT and 67.1% at EOT, as shown in Table 3. The GRRs at EOIVT and EOT by most frequent pathogens were 73.7% and 68.4% for C. albicans, and 76.9% and 73.1% for C. glabrata (Table 3).
All-cause mortality at EOIVT and EOT is shown in Table 3. All-cause mortality at day 14 and day 28 was 17.7% (14/79) and  24.1% (19/79), respectively, in the MITT population. Due to limited data for neutropenia, sub-analysis was not possible. Multivariate logistic regression was used to identify factors that were significantly related to failure. For the analysis, age, body mass index, baseline APACHE II score and baseline Candida species (infections caused by single vs multiple Candida species) were evaluated. None of the assessed factors were associated with failure.

Safety and tolerability
The incidence and severity of TEAEs during anidulafungin IV treatment are shown in Table 4. Anidulafungin was well tolerated in this population, and most adverse events were mild or moderate in severity. SAEs were noted in 39 patients (49.4%), with the most frequent being septic shock (n = 8), multipleorgan dysfunction syndrome (n = 3), cardiac arrest (n = 3), hypotension (n = 3) and haemorrhagic shock (n = 3).

Discussion
To our knowledge, this study represents the largest collection of patients treated with anidulafungin in microbiologically documented intra-abdominal infections in a clinical trial setting. The pooled analysis demonstrates that in the population of surgical patients with IAC treated with anidulafungin, the GRR was similar to that in the anidulafungin registrational trial [16]. The low incidence of candidaemia (12.7%) was     comparable with that in a recent epidemiological study [5]. Our findings are also consistent with other investigations [5,9], in terms of C. albicans (72.2%) and C. glabrata (32.9%) being the most common pathogens causing IAC at baseline. However, C. parapsilosis was not as common in this study (only 2 isolates at baseline), despite being one of the most common pathogens in another investigation (22.7%) [15]. In this pooled analysis, all Candida species were susceptible to anidulafungin, 84.2% were susceptible to fluconazole and 94.7% were susceptible to voriconazole. The superior susceptibility of anidulafungin over fluconazole was consistent with a study in a murine model of candidiasis [21]. Our study confirmed that abdominal surgery and use of broad-spectrum antibiotics are the most important risk factors associated with intra-abdominal invasive candidiasis.
Interestingly, in this study only 8 out of 79 patients had bacteria isolated from microbiological culture and recorded in the CRF. However, almost all patients (77/79) were receiving systemic antibiotic treatments, indicating concomitant bacterial infections, although it was not possible to know more about the sites of bacterial infection as this information was not collected in the CRF.
The median duration of IV anidulafungin in patients with intra-abdominal infections was 14 days, which was longer than the 10 days of IV anidulafungin duration described in the pooled analysis of the overall general population [22]. In the real-world AmarCAND2 study, the median duration of antifungal therapy was 14 days for targeted therapy and 17 days for empirical therapy [11]. Furthermore, in our population, a lower percentage of the patients (22% and 19%; data not shown) switched to oral therapy compared to that found by Kullberg et al. [22], where 51.8% and 34.8% of patients switched when permitted after ≥ 5 or ≥ 10 days, respectively.
Across the five studies, anidulafungin treatment was well tolerated, with most TEAEs mild or moderate in severity, consistent with the known safety profile of anidulafungin demonstrated since the registrational trials.
The anidulafungin registration trial included few patients with IAC. The pooled analysis of anidulafungin studies included 129 patients with deep-seated tissue infection [22], of whom 79 in the present study had confirmed IAC from peritoneal fluid and hepatobiliary tract. In general, other studies with echinocandins have included low numbers of patients with IAC. In one investigation where 16 patients had Candida peritonitis, the success rate was 87.5% (7/8) with amphotericin B and 100% (8/8) for caspofungin [23]. Success rates of 72.7% with liposomal amphotericin B and 82.4% with micafungin were observed in 28 patients with invasive candidiasis in the peritoneum [24]. In a study of 18 patients with Candida peritonitis, a 40% success rate was seen with caspofungin compared with 66.7% and 57.1% with micafungin 100 mg and 150 mg, respectively [25]. In a prospective study of 279 adult patients in the ICU, C. albicans was involved in 67% and C. glabrata in 15.6% [11]. Of the 179 Candida strains cultured from peritoneal samples, C. albicans was cultured in 61/96 (64%) isolates in empirically treated patients and in 59/83 (71%) receiving targeted therapy. Of the 179 empiric and targeted patients, 3 had candidaemia. Risk factors that influenced a higher mortality in the AmarCAND2 study included healthcare-associated Candida peritonitis, Sequential Organ Failure Assessment score ≥ 8 at ICU admission and Simplified Acute Physiologic Score II ≥ 45 at systemic antifungal treatment initiation [11]. Findings from the current analysis are similar but with lower mortality.
Intra-abdominal candidiasis is still poorly understood compared with candidaemia. To date, data and studies on the efficacy of echinocandins in intra-abdominal invasive candidiasis are scarce and, although IAC is still burdened by high mortality rates, all current international guidelines mainly address candidaemia. Source control is of paramount importance in the treatment of IAC, but adequate antifungal therapy is essential to eradicate the pathogen and to treat candidaemia, when present. For this reason, guidelines recommend an echinocandin as the drug of choice [26][27][28], although the Infectious Diseases Society of America does not include guidance on duration of therapy [26].
The concept of early adequate systemic antifungal therapy in patients with IAC is a point of debate based on the assumption that delayed initiation of antifungal therapy is associated with a poor outcome, as reported for patients with IAC, particularly if they also have candidaemia [29][30][31][32][33]. The deleterious impact of delayed initiation of systemic antifungal treatment has not yet been demonstrated for Candida intraabdominal infection [5], and further evidence is needed.
A limitation of the current investigation was that four studies were open-label and were conducted in different centres within different timeframes, and yielded a small number of patients. In mitigation, all studies had similar protocols and endpoints that allowed the data to be pooled. Although two pharmacokinetic studies have demonstrated adequate anidulafungin plasma concentrations [34,35], the lack of data to demonstrate that anidulafungin achieves adequate penetration of the peritoneal cavity may also be considered a limitation of this analysis. The analysis did not permit the assessment of whether anidulafungin would behave differently from other echinocandins in this population. Furthermore, this post hoc analysis did not include a control group and was not preplanned by the protocols of the individual studies. It was, therefore, not possible to evaluate the adequacy of source control or of antibacterial therapy. In addition, assessment of global response and mortality in relation to the neutropenic status of patients could not be performed because data on neutropenic status were not available for all patients. Finally, only a small number of factors (age, body mass index, baseline APACHE II score and baseline Candida species) were evaluated by multivariate logistic regression to assess any association with failure, and none of these factors were found to be associated with failure.
This analysis of pooled data showed that the epidemiology of IAC is similar to that in previous studies, with C. albicans and C. glabrata being the most common pathogens. Also, IAC seemed to be accompanied by a lower incidence of candidaemia compared with other studies [5,36]. Anidulafungin provided good efficacy and tolerability, which is an important consideration in critically ill surgical patients with IAC in the ICU.
Nevertheless, further studies are needed to evaluate the clinical efficacy of echinocandins, as well as their tissue penetration into the peritoneum and possibly in specific intra-abdominal districts such as bile, the liver and the pancreas.
Research data sharing Upon request, and subject to certain criteria, conditions and exceptions (see https://www.pfizer.com/science/clinicaltrials/trial-data-and-results for more information), Pfizer will provide access to individual de-identified participant data from Pfizer-sponsored global interventional clinical studies conducted for medicines, vaccines and medical devices (1) for indications that have been approved in the US and/or EU or (2) in programmes that have been terminated (i.e. development for all indications has been discontinued). Pfizer will also consider requests for the protocol, data dictionary and statistical analysis plan. Data may be requested from Pfizer trials 24 months after study completion. The de-identified participant data will be made available to researchers whose proposals meet the research criteria and other conditions, and for which an exception does not apply, via a secure portal. To gain access, data requestors must enter into a data access agreement with Pfizer.
Funding This work was supported by Pfizer. Medical writing support was funded by Pfizer. Informed consent All patients provided written informed consent.

Compliance with ethical standards
Open Access This article is distributed under the terms of the Creative Comm ons Attribution 4.0 International License (http:// creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.