In vitro and in vivo studies suggested that combination of lipid formulation of amphotericin B (L-AMB) and echinocandins may have a synergistic or additive effect against Aspergillus. Furthermore, clinical studies suggested that this combination may improve response of invasive aspergillosis (IA).
Between August 1993 and June 2008, the authors identified a total of 159 patients with hematological malignancies who received salvage therapy for IA, with L-AMB alone, echinocandins alone, or a combination of L-AMB and echinocandins. Clinical characteristics, response to salvage therapy, and death up to 12 weeks after initiation of salvage therapy were retrospectively determined for all patients.
Seventy patients received salvage therapy with L-AMB, 18 patients received echinocandins alone (89% of whom received caspofungin), and 71 patients received the combination therapy of amphotericin B and echinocandins (90% of who received caspofungin). The 3 salvage treatment groups were comparable in regard to clinical characteristics; graft versus host disease was more frequently encountered in the echinocandin group, whereas more patients in the L-AMB and combination groups had neutropenia and received immunotherapy. The response to salvage therapy was better in the echinocandin group (9% L-AMB, 28% echinocandins, and 21% for combination therapy). The 3 groups had a comparable rate of Aspergillus-related death (58%-64%) and overall mortality (61%-67%).
Invasive aspergillosis (IA) remains the most common opportunistic mold infection and major cause for morbidity and mortality in patients with hematological malignancies.1-4 Historically, amphotericin B formulations have been standard therapy for IA; however, in the current decade new antifungal agents have been introduced. Echinocandins, originally identified in 1974, have been increasingly used in the past several years owing to a unique mechanism of action, high efficacy, and favorable safety profile.5, 6 Because they have a different mechanism of action compared with other classes of antifungal agents, several studies have looked at the potential benefit of combination therapy. Data derived from in vitro studies as well as animal models have suggested that combination of lipid formulation of amphotericin B (L-AMB) and echinocandins may have a synergistic or additive effect against Aspergillus.7, 8 Furthermore, clinical studies suggested that this combination may improve response of IA.9-12
MATERIALS AND METHODS
A retrospective study was conducted at The University of Texas M. D. Anderson Cancer Center from August 1993 until June 2008. A waiver of informed consent and a waiver of authorization were requested and provided by our institutional review board. Inclusion criteria included patients with hematological malignancies, of any age, with proven or probable diagnoses of IA according to criteria of the European Organization for Research and Treatment of Cancer Mycosis Study Group.13 Patients with possible diagnosis of IA were not included.
We identified 315 patients with hematological malignancies, either refractory or intolerant to multiple previous antifungal therapies, who received salvage therapy. We selected a group of 159 patients who received echinocandins alone, L-AMB alone, or a combination of echinocandins and L-AMB. If a patient had >1 episode of salvage treatment, we included only their first episode.
The following demographic and clinical information was recorded for all study patients: age, sex, underlying hematological malignancies, history of hematopoietic stem cell transplant, presence of graft-versus-host disease (GVHD), duration of neutropenia, persistence of neutropenia during therapy, high-dose steroid use (defined as the equivalent of 600 mg of prednisone), tacrolimus use, immunotherapy during infection including white blood cell transfusion, granulocyte-macrophage colony-stimulating factor (GM-CSF), granulocyte colony-stimulating factor (G-CSF), interferon-gamma (IFN-γ), intensive care unit (ICU) stay, and need for mechanical ventilation. Type and site of either definite or probable IA along with prior antifungal prophylaxis and therapy were also recorded. Outcomes were assessed at the end of each salvage episode and up to 12 weeks after initiation of salvage therapy. Autopsy data were not available.
Breakthrough IA was defined as an infection in a patient receiving systemic prophylactic antifungals with known activity against Aspergillus species for at least 7 days before the onset of IA. Primary antifungal therapy included all antifungal drug regimes that patients received before the institution of the salvage therapy consisting of echinocandins alone, L-AMB alone, or the combination of echinocandins and L-AMB.
Favorable response was defined as complete or partial resolution of clinical, radiographic, and microbiologic abnormalities attributable to IA compared with baseline.14 Unfavorable outcome was defined as progression or stable infection based on worsening or nonresolution of clinical and radiographic abnormalities and persistently positive cultures, histopathological findings, or aspergillus antigen assays. Aspergillosis was considered to have contributed to the death of a patient if the patient died with clinical and radiographic evidence of aspergillosis and had not had a favorable response to the salvage therapy. Nephrotoxicity was defined as an increase in creatinine of 2× baseline value. Hepatotoxicity was defined as a 3-fold increase in either bilirubin and transaminase or alkaline phosphatase and transaminase compared with baseline.
Primary endpoint evaluated outcome at the end of salvage therapy, which was defined as the point where the studied salvage treatment was stopped by the primary physician, based on evidence of satisfactory response, progression of infection, intolerance, or death.
Categorical variables were compared by chi-square or Fisher exact tests and continuous variables by Wilcoxon rank-sum tests. Survival analysis was performed on patients' survival within 12 weeks after initiation of the salvage therapy in the study. The survival probability was estimated by the Kaplan-Meier approach for each group of patients, and was compared among different groups by a log-rank test. In addition, multiple logistic regression method was used to evaluate the independent effect of the salvage therapy of interest on patients' response. All tests were 2-sided, and the statistical significance was set at P ≤ .05. All statistical analyses were performed using SAS version 9.1 (SAS Institute Inc., Cary, NC).
Characteristics of Patients
From August 1993 to June 2008, 159 patients for whom salvage therapy consisted of echinocandins alone, L-AMB alone, or a combination of both echinocandins and L-AMB were identified. Patients were divided into 3 groups; 18 patients were treated with echinocandins alone as salvage therapy, whereas 70 patients received L-AMB alone. Seventy-one patients received a combination of both echinocandins and L-AMB. Eighty-nine percent (16 of 18) of the echinocandins salvage therapy episodes consisted of caspofungin and 11% (2 of 18) of anidulafungin. The majority of the combination salvage therapy episodes (90%) consisted of caspofungin and L-AMB.
Among all 3 groups, 36% (57 of 159) of the salvage episode corresponded to patients with a definite diagnosis of IA, whereas 64% (102 of 159) corresponded to patients with a probable diagnosis of IA (Table 1). The majority of IA consisted of invasive pulmonary infection (16 of 18 episodes with echinocandins, 58 of 70 episodes with L-AMB, and 57 of 71 episodes with combination therapy). Disseminated infection was presented in 11% of all salvage episodes. Sinus infection was present only in 4 of 70 (6%) episodes with L-AMB and 4 of 71 (6%) episodes with combination therapy. Furthermore, the distribution of Aspergillus species causing IA was similar in all 3 groups (Table 1).
Table 1. Characteristics of Patients
Combination, n=71, No. (%)
ECH, n=18, No. (%)
L-AMB, n=70, No. (%)
ECH indicates echinocandins; L-AMB, lipid formulation of amphotericin B; IA, invasive aspergillosis; GVHD, graft versus host disease; ANC, absolute neutrophil count; WBC, white blood cells; GM-CSF, granulocyte-macrophage colony-stimulating factor; G-CSF, granulocyte colony-stimulating factor; ICU, intensive care unit.
Some patients had >1 Aspergillus species recovered on cultures.
Duration of mechanical ventilation, median d (range)
All patients in 3 groups were similar in regard to age, sex, underlying malignancies, type of transplant, steroids use, duration of ICU stay, and mechanical ventilation (Table 1). All patients were severely immunocompromised; neutropenia at onset of IA was present in 56% of the combination group, 50% of the L-AMB group, and 17% of the echinocandin group (P = .01). Both L-AMB and combination groups were more likely to receive immunotherapy with granulocyte transfusion during the study period when compared with the echinocandin group (P = .01). A higher percentage (56%) of patients receiving echinocandins had GVHD (P = .008) and were receiving tacrolimus therapy on diagnosis (Table 1).
The majority of patients received antifungal prophylaxis; there was no statistically significant difference in rate of breakthrough infection in all 3 groups (Table 2). The duration of primary therapy was similar in all groups. A comparable proportion of patients experienced toxicity to primary therapy.
ECH indicates echinocandins; L-AMB, lipid formulation of amphotericin B.
Some of the patients received 2 antifungal drugs concomitantly as primary therapy.
Prophylactic antifungal treatment prior to infection
Primary antifungal therapy
Duration of primary therapy, median d (range)
Lipid amphotericin B alone
Lipid amphotericin B-containing regimen
Favorable response to the primary therapy
Toxicity due to primary therapy
Response and Survival
Twenty-one percent of patients receiving combination therapy (L-AMB and echinocandins) achieved a favorable outcome at the end of salvage therapy versus 28% of patients receiving echinocandins therapy and 9% of patients receiving L-AMB therapy (28% vs 9%, P = .04). The patients receiving L-AMB had a significantly shorter duration of therapy (Table 3). Furthermore, the combination therapy with L-AMB and echinocandins offered no advantage in reducing overall mortality. All 3 groups had a comparable rate of overall death, with a range of 61% to 67%. The rate of Aspergillus-related death was also similar among all 3 groups, with a mean of 63%.
Table 3. Response to Therapy, Duration of Therapy, and Death
Combination, n=71, No. (%)
ECH, n=18, No. (%)
L-AMB, n=70, No. (%)
ECH indicates echinocandins; L-AMB, lipid formulation of amphotericin B.
Duration of therapy, median d (range)
Favorable response at end of salvage therapy
Mortality within 12 weeks of initiating salvage therapy
In a multiple logistic regression analysis, the factors that were independently associated with the response included GVHD, immunotherapy during infection, using azole-containing regimen in the primary therapy, the response to primary therapy, and the duration of the salvage therapy during the study. Patients with GVHD (odds ratio [OR], 0.2; 95% confidence interval [CI], 0.04-0.8; P = .027) and patients receiving immunotherapy during infection (OR, 0.2; 95% CI, 0.06-0.9; P = .029) were less likely to have favorable response. Patients who had an azole-containing regimen as their primary therapy (OR, 4.7; 95% CI, 1.2-19.1; P = .03), patients who had a favorable response to the primary therapy (OR, 7.8; 95% CI, 1.3-48.3; P = .028), and patients with a longer duration of the salvage therapy (OR, 1.07; 95% CI, 1.04-1.1; P<.0001) were more likely to have a favorable response to the salvage therapy in the study.
The Kaplan-Meier survival estimate of the 3 groups of patients during a 12-week period of follow-up after the initiation of salvage therapy is illustrated in Figure 1. There has been no significant difference in survival among the 3 groups of patients (P = .9).
Toxicity and Duration of Salvage Therapy
L-AMB alone or in combination with echinocandins was associated with higher rates of adverse events, including hepatic and renal toxicity when compared with echinocandins alone. The rate of overall toxicity in the combination group when compared with the echinocandin group was 31% versus 6% (P = .03) in the combination group compared with the echinocandin group and 26% versus 6% in the L-AMB group when compared with the echinocandin group (Table 4). Only 1 patient in the echinocandin arm experienced elevation in total bilirubin level, whereas patients in the other treatment groups experienced elevation in bilirubin, transaminases, and creatinine, attributed to antifungal therapy.
Table 4. Adverse Events Associated With Salvage Therapy
In this study, combination of L-AMB and echinocandins for proven or probable IA in patients with hematological malignancies offered no survival advantage or improved response when compared with either drug alone. Response rates have overall been lower than in other published reports, likely because our patient population was sicker, almost half of them being critically ill. Also, in contrast to other studies, combination therapy did not appear to be more effective in patients with persistent neutropenia. Addition of immunotherapy such as granulocyte transfusion was associated with less favorable outcome.
Echinocandins, originally identified in 1974, are inhibitors of fungal cell wall synthesis.5 They exert selective antifungal activity by inhibition of beta-glucan synthase, resulting in formation of aberrantly growing fungal hyphae. When administered to animals infected with Aspergillus species, echinocandins appear to prolong survival, although high fungal counts usually persist in the tissues.15 More recent studies have suggested an immunomodulatory role for echinocandins through stage-specific beta-glucan exposure, which may activate not only polymorphonuclear cells but also monocytes and macrophages.16, 17 Because different classes of antifungals use different mechanisms of action, combination therapy appears an attractive concept.
In vitro studies have suggested that caspofungin and amphotericin B combination was synergistic or additive in >50% of the 14 species of Aspergillus tested.7 Subsequent animal studies have suggested that combination of echinocandins with liposomal amphotericin B may have some benefit.8
Early clinical reports have suggested that the combination of caspofungin and L-AMB appears safe and generally well tolerated and translated into an overall favorable response in 60% of 26 patients with acute leukemia and refractory pulmonary aspergillosis.9 However, only 6 of the 26 patients enrolled had a diagnosis of proven aspergillosis. A larger, multicenter study involving 53 patients with documented IA receiving salvage therapy in many combinations including caspofungin-triazole and caspofungin-polyene suggested that the 50% of the patients receiving the combination caspofungin-polyene had a favorable response at the end of combination therapy and Day 84 visit, but the overall number of patients enrolled in this arm was relatively small (16 patients).10
A randomized, open, pilot study of 30 patients comparing high-dose L-AMB versus combination of low-dose L-AMB with caspofungin found significantly more favorable responses in the combination arm, which translated into survival benefit at 12 weeks; the small size and the open label design of the study make the interpretation of the results more difficult.11 Lastly, a report from our institution evaluated 31 patients who received the combination caspofungin and L-AMB for refractory aspergillosis; 35% of these patients responded to this combination.12 When caspofungin was studied as a single agent for treatment of refractory aspergillosis, an overall response rate of 45% was noted; however, only 22% of these patients were allogeneic stem cell transplant recipients, and only 26% had neutropenia.18
Herein we report the results of a retrospective analysis of 159 patients receiving salvage therapy for proven and probable IA over a time span of 15 years, which to our knowledge is among the largest series published to date. Thirty-six percent of these episodes were because of proven disease. Our results indicate that combination therapy appears to offer no advantage over L-AMB or echinocandins as single drugs, with a response rate ranging from 9% to 28%. The favorable response rate in our study is lower than in the previously reported data, probably due to the fact that we enrolled a higher overall number of patients with documented severe infection, in the setting of advanced hematological malignancies; almost half of the study patients required ICU admission, for a median of 8 to 11 days; over a third of the patients required mechanical ventilation. Aspergillus-related mortality and overall mortality appeared no different between the 3 groups, ranging 58% to 64% and 61% to 67%, respectively, significantly higher when compared with other studies.
The 3 groups were similar in regard to underlying host factors, including receipt of a hematopoietic stem cell transplant and steroid use, as well as in regard to characteristics of IA. Some of the patients received immunotherapy, aimed to enhance the host immune response in clearance of invasive fungal infections, including G-CSF, GM-CSF, INF-γ, and granulocyte transfusions.19, 20
Although the proportion of patients with persistent neutropenia was similar between the 3 groups, neutropenia at onset of IA was overall more frequently encountered in the L-AMB and combination arm, whereas GVHD was present more often in the echinocandin group. As a result, usage of G-CSF and granulocyte transfusions was significantly higher in L-AMB and combination groups, without additional benefit in favorable response at the end of therapy. Moreover, in multivariate analysis both GVHD and receipt of immunotherapy were associated with less favorable response. One possible explanation is that addition of immunotherapy, particularly granulocyte transfusion, may induce significant pulmonary inflammatory responses with clinical and radiographic deterioration. However, among all salvage episodes with favorable response, the use of additional immune-enhancing therapy made no difference. Alternatively, echinocandins may have exerted their therapeutic effect in the absence of additional immune-enhancing strategies, particularly in patients with GVHD.
Previous studies suggested that combination therapy appears beneficial especially in patients with poor prognostic features such as neutropenia or refractory leukemia. In our study, data regarding refractory leukemia were not available; however, for patients with persistent neutropenia, combination therapy offered no advantage compared with single-agent therapy.
Predictably, there were fewer adverse events in the echinocandins arm compared with the L-AMB and combination arms. It is well accepted that echinocandins have a favorable adverse events profile and are generally very well tolerated. In the echinocandins arm, fewer salvage episodes were complicated by development of liver function abnormalities and none by renal impairment. The duration of salvage therapy with L-AMB alone was shorter, perhaps reflecting the increased toxicity and poor tolerability associated with L-AMB. Hence, polyenes used alone or in combination with echinocandins add to the toxicity, without offering a therapeutic advantage.
Our study has several limitations. The retrospective design allows for selection bias; its 15-year span allows for changes in supportive care and heterogeneity in respect to the antifungal agents used as primary therapy. There was overall a smaller number of salvage episodes with echinocandins alone; the contribution of IA to death was not diagnosed by autopsy.
In conclusion, the study we conducted is among the largest reported to date, including a patient population with refractory IA in the setting of advanced hematological malignancy, of whom almost 50% were critically ill. Our data suggest no additional benefit of combination therapy versus single-agent echinocandins or L-AMB as salvage therapy for documented or probable IA, but rather a negative impact related to toxicity and cost. Addition of immune-enhancing strategies has not favored 1 group over others. As expected, echinocandins have demonstrated a favorable profile of adverse events. Larger, prospective studies are needed for validation.
CONFLICT OF INTEREST DISCLOSURES
Dr. Raad has received grants from Wyeth, Cubist, Pfizer, and Cook, is a consultant for Cubist and Cook, is on the speaker's bureau for Pfizer and Cook, and has received royalties related to patents licensed to Cook on which Dr. Raad is an inventor/co-inventor. Dr. Hachem has received grant support from both Pfizer and Wyeth.