Digital Features for this article can be found at https://figshare.com/s/5e036327f480a030a490.

FormalPara Key Points

The primary measure of efficacy used in this study (global response of success rate) at the end of IV treatment with anidulafungin in patients with invasive candidiasis/candidemia and solid tumors was comparable to that of a representative sample of patients with IC/candidemia.

In this analysis, treatment with anidulafungin was effective irrespective of a past or recent diagnosis of solid tumors.

1 Introduction

Patients with cancer, particularly solid tumors and hematologic malignancies, are at increased risk of invasive candidiasis (IC) [1] or candidemia due to IC [2]. In an analysis of European multi-institutional surveys of patients with candidemia, an underlying pathology of solid tumors was the third-most common factor associated with IC/candidemia, after surgery and intensive care treatment [3]. Candida species are a frequent cause of invasive fungal infection in patients with many different types of solid tumors, including gastrointestinal, breast, ear/nose/throat, genitourinary, head and neck, hematologic, lung, liver, and skin [2, 4,5,6,7,8].

In epidemiology and surveillance studies, morbidity and mortality outcomes of IC in patients with solid tumors and hematologic cancers were poor [3, 4, 9], with mortality rates between 30 and 65% [3,4,5,6,7, 9, 10]. In addition to older age and advanced disease, surgery, admission to the intensive care unit (ICU), and invasive procedures were factors associated with 30-day mortality [5, 7, 9, 10]. Furthermore, C. albicans is reported as the most frequent isolate in patients with solid tumors followed by C. parapsilosis, C. tropicalis, or C. glabrata [3, 4, 6, 9, 11].

Appropriate use of antifungal treatment in patients with cancer and candidemia is a life-saving measure [4, 5, 7, 9]. Echinocandins are an effective treatment against a range of Candida species [12], and the Infectious Diseases Society of America guidelines recommend echinocandins for treatment of IC/candidemia [13]. Also, guidelines from the European Society for Clinical Microbiology and Infectious Diseases strongly recommend echinocandins for the initial treatment of candidemia in adults [14].

Echinocandins, including anidulafungin, target Candida species by inhibiting beta-(1,3)-d-glucan synthase, essential for the synthesis of fungal cell wall glucan [12]. Anidulafungin is approved in the USA and Europe for the treatment of IC/candidemia [15, 16]. Previous pooled analyses of anidulafungin studies have demonstrated effectiveness and tolerability in patients with IC/candidemia, intra-abdominal candidiasis, neutropenia, and infections caused by C. parapsilosis and C. krusei [17,18,19].

To address the relative scarcity of data on the treatment of patients with invasive fungal infections and malignancies, and to explore the temporal effect of tumor diagnosis on the risk of fungal infection, this post hoc analysis investigated whether the efficacy and safety of anidulafungin treatment of patients with IC/candidemia and solid tumors varied when summarized by past or recent diagnosis of solid tumor diagnosis.

2 Patients and Methods

2.1 Study Data

In this post hoc analysis, patient-level data were pooled and analyzed from four open-label, non-comparative studies of anidulafungin [20,21,22,23] and two double-blind, double-dummy, randomized studies (Pfizer data on file) that evaluated anidulafungin and caspofungin (Online Resource 1); all studies had similar protocols and endpoints. Data from all the above studies have been included in a previous analysis of the efficacy of anidulafungin in patients with IC [18], and this analysis followed a similar methodology.

Conduct was in accordance with applicable legal and regulatory requirements, the International Conference on Harmonization Guidelines for Good Clinical Practice, and the Declaration of Helsinki. The Institutional Review Boards or Independent Ethics Committees at each investigational center approved the studies. All patients provided written informed consent.

2.2 Patients

Male or female patients aged ≥ 18 years for the four open-labeled and the two randomized studies were included in the pooled dataset if they had blood culture-confirmed candidemia, or IC confirmed by culture positive for Candida species isolated from a normally sterile site or a newly placed drain, with or without a positive Candida blood culture, within 96 h of entry to the study. Patients could be included based on microbiologic evidence of Candida infection, such as a positive blood (or tissue specimen) culture for yeast, provided confirmation of Candida species was obtained within 96 h.

Patients were also required to have one or more clinical signs and symptoms of fungal infection, including fever, hypothermia, hypotension, localized signs and symptoms of inflammation at a site with Candida infection within 48 h of commencing treatment, or radiologic findings indicative of IC. Patients were categorized as neutropenic if they had a baseline absolute neutrophil count (ANC) of ≤ 500 cells/mm3 (cells/µL). Patients included in the analysis had either a past diagnosis of solid tumors (time of diagnosis ≥ 6 months prior to study entry) or a recent diagnosis of solid tumors (time of diagnosis < 6 months before study entry). Solid tumor status was obtained by examining the medical histories of patients using the following search criteria: tumor, neopla* (neoplastic, neoplasia, neoplasm), mass, cancer, malign* (malignant, malignancy), growth, oncol* (oncology, oncological), carcinoma, adenocarcinoma, and hepatoma.

The main exclusion criteria were: more than 48 h of systemic antifungal treatment prior to study entry; previous treatment failure for current episode of IC or candidemia; or a prosthetic device at a suspected site of infection that could not be removed prior to or within 24–48 h of the start of the study.

2.3 Treatment

Patients received a single IV loading dose of anidulafungin 200 mg on day 1, followed by 100 mg once daily and, in all studies except one [18], could be switched to oral voriconazole or fluconazole therapy after ≥ 5 or ≥ 10 days based on pre-specified criteria. IV anidulafungin and oral azole (if required) were maintained for ≥ 14 days (and up to a maximum of 42 days) after the last positive Candida culture and resolution of symptoms.

2.4 Endpoints

The primary efficacy endpoint of this post hoc analysis was global response success (GRS) rate at the end of intravenous therapy (EOIVT) in the modified intent-to-treat (mITT) population. This included patients with solid tumors who received at least one dose of anidulafungin (the ITT population) and had a confirmed diagnosis of candidemia or a positive culture for Candida species at study commencement or within 96 h of study entry. GRS was achieved with clinical success, defined as cure or improvement of clinical signs and symptoms with no additional systemic or oral antifungal therapy, in conjunction with microbiologic success, defined as eradication or presumed eradication of baseline Candida species.

Secondary endpoints included GRS rate by baseline pathogen and site of infection, evaluated at EOIVT and end of all therapy (EOT) in the mITT population. Other secondary endpoints were all-cause mortality at days 14 and 28, duration of therapy (IV and IV + oral), and time to switch to oral azole therapy. Safety was assessed by the incidence and severity of adverse events occurring during or within 30 days of the last dose of treatment, recorded by system organ class and Medical Dictionary for Regulatory Activities (MedDRA) preferred terms. Safety analyses were based on the ITT population.

2.5 Statistical Analysis

All primary and secondary endpoints were summarized by the time of diagnosis. The past diagnosis was ≥ 6 months since the time of diagnosis, and recent was < 6 months. Analyses for baseline demographics and disease characteristics were for descriptive purposes only. Evaluations were primarily descriptive statistics. GRS rates were estimated with exact 95% confidence intervals (CI) for binomial proportion (Clopper–Pearson method), and 95% CIs differences between past and recent diagnosis groups were estimated with exact unconditional limits. Comparisons between past and recent diagnosis groups were determined using a Fisher’s exact test. Indeterminate or missing data were considered to be failures.

3 Results

3.1 Patients

Of 539 patients in the cohort pooled from six studies (Online Resource 1) [18], 150 had confirmed IC and history of any malignancy, and of these, 139 had solid tumors. Among patients with solid tumors (Table 1), baseline demographics and disease characteristics were generally comparable between the two diagnostic subgroups. The most frequent types of solid tumors were gastrointestinal (60.4%), genitourinary (24.5%), and lung (16.5%). Metastases were present in 12.9% of all patients in the mITT population. Approximately one-half of the patients (52.5%) were male and, overall, the mean age (standard deviation [SD]) was 63.1 (13.4) years. The majority of patients (90.6%) were aged ≥ 45 years, and 48.9% were aged ≥ 65 years.

Table 1 Patient baseline demographics and disease characteristics by time of solid tumor diagnosis (mITT, N = 139)
Full size table

The overall mean (SD) Acute Physiology, Age, Chronic Health Evaluation (APACHE) II Score was 14.9 (5.5), and most patients (90.9%) were non-neutropenic (Table 1). The most frequent baseline pathogens were C. albicans (48.9%) and C. glabrata (21.6%). The main sites of infection were blood only (75.5%), and other sterile sites (20.1%), followed by blood and other sterile sites (4.3%) [data not shown].

Frequent risk factors for IC included the use of broad-spectrum antibiotics (88.5%), central venous catheters (80.3%), and total parenteral nutrition (54.1%) (Table 1). The median (range) duration of IV + oral therapy for the mITT population was 15.0 (1–56) days. Fifty-four of 139 patients (38.8%) in the mITT population switched to oral-only therapy for a median (range) duration of 9.0 (5–35) days, and a median time to switch of 8.0 days (range: 4–34). Among these patients, the proportion of patients switching to oral azoles was lower in patients with a past diagnosis of solid tumors than a recent diagnosis (42.6% vs 57.4%, respectively), and the median time to switch was shorter (6 vs 10 days, respectively).

3.2 Efficacy Endpoints

The GRS rate in the mITT population was 73.4% (102/139) at EOIVT (primary endpoint) and 65.5% (91/139) at EOT (Fig. 1). GRS rates at EOIVT and EOT in patients with a past tumor diagnosis were 75.5% (37/49) and 71.4% (35/49); corresponding GRS rates were 72.2% (65/90) and 62.2% (56/90), respectively, in those with a recent diagnosis of solid tumors (Fig. 1). GRS rates at both EOIVT and EOT were not significantly different between those with recent or past diagnosis of solid tumors (p > 0.05).

Fig. 1
figure 1

Anidulafungin global response success rates at EOIVT and EOT by the time of solid tumor diagnosis (mITT, N = 139). EOIVT end of intravenous treatment; EOT end of treatment; m months; ns not significant based on Fisher’s exact test. Past diagnosis of solid tumors, ≥ 6 months prior to study entry; recent diagnosis, < 6 months before study entry

Full size image

GRS rates at EOIVT and EOT in the overall population were 77.9% and 69.1%, respectively, for C. albicans, and 67.1% and 61.2%, respectively, for non-albicans Candida species (Table 2). GRS rate by the site of infection was 75.2% for blood-only infections at EOIVT and 65.7% at EOT (Table 2). For patients with infection in the blood only, those with recent solid tumor diagnosis had a significantly smaller GRS at EOT, compared with patients with past solid tumor diagnoses (p = 0.049).

Table 2 Anidulafungin global response success rates at EOIVT and EOT by the time of solid tumor diagnosis, pathogen and by the site of infection (mITT, N = 139)
Full size table

All-cause mortality in the overall mITT population was 20/139 (14.4%) at day 14 and 28/139 (20.1%) at day 28. All-cause mortality rates were not significantly different between populations with a recent and past diagnosis of solid tumors at day 14 or overall. At day 28, the all-cause mortality rate was significantly higher in those with a recent diagnosis of solid tumors (25.6%), compared with those with a past diagnosis (10.2%) (p = 0.045).

3.3 Safety Endpoints

Incidence and severity of treatment-emergent adverse events (TEAEs) are shown in Table 3. A total of 787 TEAEs were observed in 123 patients, and the proportion of patients with events was similar in those with a recent (88.9%) or past diagnosis of solid tumors (87.8%) (data not shown). The majority of all-causality TEAEs in the overall population were mild or moderate in severity (642/787; 81.6%). The most common TEAEs were gastrointestinal, including diarrhea (21/139; 15.1%), nausea (13/139; 9.4%), and vomiting (13/139; 9.4%); the pattern was similar in the two tumor subgroups (data not shown).

Table 3 Incidence and severity of treatment-emergent adverse events during all treatment (IV anidulafungin + oral azole) by system organ class (ITT)
Full size table

4 Discussion

This analysis of patient-level data pooled from six prospective studies of anidulafungin showed that in patients with IC/candidemia and solid tumors, the GRS rate at EOIVT (73.4%; primary endpoint) was comparable to the overall GRS rate (75.6% at EOIVT) in patients with IC in the registrational study [24]. The GRS rates were also comparable to a randomized, double-blind comparison of caspofungin and amphotericin B in 74 patients with IC and active malignancies. In that investigation, response rates in patients with solid tumors were 80% and 59% in the two treatment arms, respectively [2].

Additionally, in patients with solid tumors, all-cause mortality rate by 6–8 weeks following discontinuation of IV treatment was 40% in caspofungin recipients and 21% in recipients of amphotericin B [2]. By comparison, the present analysis included a higher number of patients with IC/candidemia and solid tumors, and the all-cause mortality rate was 20.1% at day 28.

Non-albicans Candida infection also represents concern in daily clinical practice [1]. In our analysis, C. albicans was overall the most common isolate, and GRS rates were slightly lower in patients with baseline C. albicans and a past solid tumor diagnosis compared with a recent diagnosis at EOIVT (74% vs 80%), but not at EOT (70% vs 69%). Overall, anidulafungin was effective in patients with baseline C. albicans and non-albicans Candida infection (78% and 67% at EOIVT, respectively). The characteristics of patients in our analysis were generally consistent with other studies [4, 6, 9], including 49% in our mITT population with IC/candidemia due to C. albicans, 60% with solid tumors that were located in the gastrointestinal tract, and around half (53%) having undergone abdominal surgery. In a separate pooled analysis of 79 patients with intra-abdominal candidiasis, the overall global response rate following anidulafungin treatment was 73.4% at EOIVT [19].

Neutropenia is a more common characteristic of patients with hematologic cancer than patients with solid tumors [9]. Our population was comparable with this observation in that the majority with solid tumors were non-neutropenic at baseline (91%).

There is a paucity of studies that have evaluated the time of solid tumor diagnosis as a factor that may determine the clinical outcome, and in our analysis, the GRS rate was comparable at EOIVT regardless of past or recent diagnosis of solid tumors. At EOT, the pattern of global response rates was comparable. In a retrospective cohort study of 144 non-neutropenic patients with candidemia, multivariate analysis showed that one of the predictors of non-albicans candidemia was a recent history of solid tumor [25]. A prospective, multicenter, surveillance program in Spain, conducted in part to identify predictors of death in candidiasis, found that 7- versus 30-day mortality in patients with hematologic malignancies and solid tumors was 12.8% (93/726) and 30.6% (220/720). When adjusted by the primary source of candidemia and severity of infection (severe sepsis or septic shock) in the multivariate regression analysis, appropriate antifungal treatment within the first 48 h was the only factor independently associated with lower mortality [26]. This highlights the importance of early antifungal treatment of candidiasis in patients with solid organ tumors.

De-escalation of antifungal treatment has been assessed in a study of 190 critically ill patients with suspected IC [27]. The approach, which was undertaken in 20% of the patients included, was considered to be safe and was associated with a shorter median duration of antifungal treatment compared with no de-escalation (6 vs 13 days). In our study, a higher proportion of patients (38.8%) were switched to oral therapy, and the median total duration of therapy (IV + oral) was similar in past and recent solid tumor diagnostic sub-groups (15 vs 14 days, respectively).

Elderly patients (aged ≥ 65 years) formed almost half (48.9%) of the patients in this post hoc analysis, and older age is a factor associated with IC/candidemia in patients with cancer [9]. In another post hoc analysis of anidulafungin treatment in critically ill patients, a similar proportion (47%) were elderly patients with IC/candidemia, and the median duration of IV anidulafungin treatment was comparable in elderly and non-elderly patients (14.0 vs 14.5 days) [28], similar to the present analysis. Overall, switch to oral fluconazole or voriconazole occurred in 58/170 (34%) of patients in the earlier investigation [28], also similar to the proportion in our analysis.

Overall, anidulafungin was well tolerated, and the majority of TEAEs were mild-to-moderate in severity; no new safety signals arose from this analysis [24]. The number of TEAEs experienced by patients with a past or recent diagnosis of solid tumors was generally similar.

One limitation of this analysis was that four studies were open-label and conducted at different times. However, all studies had a similar design that permitted the pooling of data. Our analysis was post hoc and not pre-planned by the individual study protocols. Susceptibility testing was not evaluated as part of this pooled analysis but has been described for a number of the individual prospective studies of anidulafungin [20, 22, 23, 28], and in pooled populations [17, 18]. Also, due to the limited number of patients, no multivariate analysis was performed, and patients were included either with solid tumors or a history of solid tumors. Selection of the time of past or present diagnosis of solid tumors did not lead to the identification of a difference between the effect of treatment in the two sub-groups, but we cannot exclude the possibility that further investigations using this or a different definition and a larger group of patients may be needed. Our study did not evaluate the potential for drug–drug interactions (DDIs). In recent years, several new immunotherapeutic agents have been developed for the treatment of patients with hematologic malignancies and solid tumors, and, for some of these new agents, the use of azoles that are strong inhibitors of cytochrome P450 3A4 is precluded due to the potential for DDIs [29]. Anidulafungin does not undergo hepatic metabolism, nor does it interact with the cytochrome P450 enzymes involved in DDIs [30]. No significant interactions between anidulafungin and cyclosporine, tacrolimus, rifampin, voriconazole, or liposomal amphotericin B are currently known [16].

5 Conclusion

Anidulafungin was effective for the treatment of IC/candidemia in patients regardless of past or recent time of solid tumor diagnosis. The safety results were consistent with the known profile of anidulafungin.