Multicentre surveillance study on feasibility, safety and efficacy of antifungal combination therapy for proven or probable invasive fungal diseases in haematological patients: the SEIFEM real-life combo study
Division of Hematology, University of Udine, Italy
Correspondence: Anna Candoni, MD, Division of Hematology and BMT, University of Udine, Italy.
Note. A part of this study was presented (oral communication) at 21st European Congress of Clinical Microbiology and Infectious Diseases (ECCMID), Milan, 2011.
This multicentre observational study evaluated the feasibility, efficacy and toxicity of antifungal combination therapy (combo) as treatment of proven or probable invasive fungal diseases (IFDs) in patients with haematological malignancies. Between January 2005 and January 2010, 84 cases of IFDs (39 proven and 45 probable) treated with combo were collected in 20 Hematological Italian Centres, in patients who underwent chemotherapy or allogeneic haematopoietic stem cell transplantation for haematological diseases. Median age of patients was 34 years (range 1–73) and 37% had less than 18 years. Acute leukaemia was the most common underlying haematological disease (68/84; 81%). The phase of treatment was as follows: first induction in 21/84 (25%), consolidation phase in 18/84 (21%) and reinduction/salvage in 45/84 (54%). The main site of infection was lung with or without other sites. The principal fungal pathogens were as follows: Aspergillus sp. 68 cases (81%), Candida sp. six cases (8%), Zygomycetes four cases (5%) and Fusarium sp. four cases (5%). The most used combo was caspofungin+voriconazole 35/84 (42%), caspofungin + liposomal amphotericin B (L-AmB) 20/84 (24%) and L-AmB+voriconazole 15/84 (18%). The median duration of combo was 19 days (range 3–180). The overall response rate (ORR) was 73% (61/84 responders) without significant differences between the combo regimens. The most important factor that significantly influenced the response was granulocyte (PMN) recovery (P 0.009). Only one patient discontinued therapy (voriconazole-related neurotoxicity) and 22% experienced mild and reversible adverse events (hypokalaemia, ALT/AST increase and creatinine increase). The IFDs-attributable mortality was 17%. This study indicates that combo was both well tolerated and effective in haematological patients. The most used combo regimens were caspofungin + voriconazole (ORR 80%) and caspofungin + L-AmB (ORR 70%). The ORR was 73% and the mortality IFD related was 17%. PMN recovery during combo predicts a favourable outcome. Clinical Trials Registration: NCT00906633.
Recent data from haematological series suggest that mortality attributable to invasive fungal infections (IFDs) has decreased, probably due to the application of a correct and timely diagnostic evaluation and to the availability of newer, well tolerated and more efficacious drugs (echinocandins, extended spectrum triazoles, lipid formulations of Amphotericin B).[1-4] However, the efficacy of current antifungal therapies is still suboptimal, particularly in some categories of patients (allogenic HSCT recipients) or in some rare fungal infections (zygomycosis, fusariosis).[2, 3] The expansion of the antifungal armamentarium led to the opportunity to revise the traditional approaches to treat IFDs and to the chance to further improve the outcome.
Combination antifungal therapy (combo) is not a new concept, as its role has been well established for various infectious diseases, such as cryptococcal meningitis.[5, 6] Unfortunately, whether combo may affect the outcome of IFDs is still controversial, since only limited data are available. This is even more unclear for the issue of rare non-Aspergillus moulds. A large number of studies performed in vitro or in animal models have examined the effects of combo suggesting potential benefits for difficult-to-treat mycosis.[7-13] Unfortunately, results of preclinical studies cannot be translated into clinical decisions and many unanswered questions do persist. Efficacy and safety still need to be proved, particularly in the setting of a prospective clinical trial.
The aim of this multicentre observational study was to collect data on the use of this off-label treatment for proven and probable IFDs among Italian Hematological Centres and to assess the feasibility, toxicity and efficacy of antifungal combo strategy in patients with haematological malignancies.
Patients and methods
Patients with any type and stage of haematological malignancy and patients undergoing autologous or allogeneic HSCT were considered eligible for this survey. All participating centres had received a specific Case Report Form to retrospectively register all cases treated with combo. Patient data were queried for haematological underling disease, medical history, predisposing factors to IFDs, IFD sites and aetiology, IFD-related therapy and outcome. The presence of neutropenia was assessed at baseline and at the end of combo for all cases. The Platelia Aspergillus enzyme-linked immunoassay (Bio-Rad, Redmont, WA) was used to detect galactomannan and index of 0,5 or greater in two consecutive samples was considered as positive.[4, 14] All reported cases were reviewed by two independent physicians (AC and LP) and only those identified as proven or probable IFDs (according to EORTC criteria) were included in the survey.
Efficacy of combo was based on the investigator assessment at the end of treatment (considering clinical, radiographic and microbiological response). In line with current standard definitions, both complete (CR) and partial responses (PR) were considered as success; all other responses were classified as failure. In particular, patients were considered refractory to the treatment if clinical and radiological signs indicated a progressive infection after at least 7 days of antifungal combo therapy with adequate drug doses. Toxicity was evaluated and graded in accordance with WHO criteria.
Overall mortality was defined as any death within 12 weeks from the diagnosis of proven or probable IFD. Mortality was considered attributable to the IFD (IFD-attributable mortality) when patients died within 12 weeks from the onset of a fever with microbiological, histological, or clinical evidence of an active IFD and if other potential causes of death could be excluded by the responsible physician.
Univariate analysis was performed using the chi-squared test with the following independent variables: age (cut-off 50 years), paediatric or adult patient (cut-off 18 years), certainty of diagnosis (proven vs. probable), status of haematological malignancies (active or in remission), days of therapy (using cut-off 14 days), neutrophil count (PMN) at the end of therapy (cut-off 500/mmc), type of combo treatment (sequential vs. ex novo). Multivariate analysis was performed using a logistic regression model. Adjusted HRs and 95%-CIs were calculated. Statistical significance was set at P-value < 0.05. Endpoint for univariate and multivariate analysis was response to combo antifungal treatment (partial and complete responses). Survival curves were generated by the Kaplan–Meier method. Overall survival (OS) was calculated from the data of IFD diagnosis to death or to the last date of follow-up. The impact of clinical variables on survival was evaluated using the Cox hazard regression. The statistical analysis was performed using MedCalc version 22.214.171.124 (MedCalc statistical Software bvba, Belgium).
From January 2005 to January 2010, 84 cases of combo were collected among the 20 participating Italian centres. They were 38 females and 46 males with a median age of 34 years (range 1–73). Thirty-one patients were younger than 18 years (paediatric cases), but only 3/31 were less than 3 years old. All cases had haematological disease, as reported in Table 1, and 35% (29/84) of them had undergone HSCT.
Acute leukaemia was the most common underlying haematological disease (68/84; 81%).
The majority of patients had refractory or relapsing disease (45/84, 54%), whereas the remaining were at the onset of their disease (21/84, 25%) or in remission (18/84, 21%).
Causative agents and infection sites
IFDs were classified as proven in 39/84 cases (46%) and probable in 45/84 cases (54%). The site of IFD was lung in 56 of 84 patients (67%) whereas 25/84 (30%) of cases had disseminated infections with two or more sites involved.
Aspergillus species was identified as causative agent in the majority of cases (68/84, 81%) (Table 1). In detail, the causative agents in proven IFD cases were as follows: Aspergillus fumigatus (6), Aspergillus flavus (5), Aspergillus niger (2), Aspergillus terreus (1), Aspergillus sp. (9), Fusarium sp. (4), Mucor sp. (2), Rhizopus (2), Candida not albicans (6), Blastoschizomyces (1) and Peacillomyces (1).
A total of 57/84 (68%) of patients had received previous antifungal prophylaxis with fluconazole (31/57), itraconazole (20/57), posaconazole (3/57) or L-AmB (3/57). Hence, 67/84 (80%) had received a previous antifungal monotherapy (empiric 39/67, preemptive 22/67 or target 6/67), with a median duration of 9 days (range 3–46). Prior antifungal therapies included amphothericin B lipid formulations (45/84, 54%), caspofungin (10/84, 12%), voriconazole (9/84, 11%), posaconazole or itraconazole (3/84, 4%). In addition, 8/84 (10%) patients had previously received more than one line of antifungal monotherapy.
The neutrophil count (PMN) at the start of combo was below 1000/mmc in 75% of cases and 56/84 (67%) patients had PMN less than 500/mmc. Only 21/84 (25%) of patients had more than 1000 PMN/mmc.
Type of combo
Combo antifungal therapy was started de novo (with two new antifungal drugs) in 45/84 patients (54%) whereas in 39/84 cases (46%) it was a sequential therapy, adding a new antifungal agent to the prior antifungal monotherapy. The most frequently used combo was as follows: caspofungin + voriconazole (35/84, 42%), caspofungin + liposomal amphotericin B (L-AmB) (20/84, 24%) and L-AmB + voriconazole (15/84, 18%) (Table 2). The median duration of combo was 19 days (range 3–180).
Table 2. Responses (overall response rate, ORR; complete response, CR; partial response, PR) to combo therapy in proven and probable IFD.
Combo therapy and responses
Caspofungin + Voriconazole
L-AMB + Caspofungin
L-AMB + Voriconazole
L-AMB + Caspo + Vorico
L-AMB + Posaconazole
Abelcet + Caspo or Vorico
Caspofungin + Posaconazole
Response to treatment and prognostic factors
The overall response rate (ORR) was 73% (61/84 responders) with 29 (35%) of CR and 32 (38%) of PR, without significant differences between the combo regimens as reported in Table 2.
If we consider the response to combo according to the different causative agents we had: 71% (48/68) of ORR in Aspergillosis, 50% (2/4) of ORR in Zygomycosis, 100% (6/6) in Candidiasis and 75% (3/4) in Fusariosis.
No differences in terms of ORR were recorded between paediatric and adult cases (81% vs. 71% respectively; P 0.4). However, paediatric cases (31) compared to adult patients (53) had these significant differences: more frequent use of caspofungin + voriconazole combination (19/31 vs. 15/53; P 0.006), longer duration of combo therapy (40 vs. 18 days; P 0.005) and more frequent use of surgery (9/31 vs. 5/53, P 0.004).
In 14/84 cases of IFDs (nine paediatric and five adult patients) the combo was associated with a surgical approach including: lung lobectomy (9 cases), liver lobectomy (2 cases), tonsillectomy (1 case), vitrectomy (1 case) and excision of mycotic brain abscess (1 case). The ORR of combo plus surgery in these cases was 79% (11/14).
In univariate analysis, the only factor that significantly influenced the response of combo was PMN recovery (PMN greater than 500/mmc) during combo (P-value 0.001) and the type of IFD (proven vs. probable IFDs, P-value 0.02). Other factors such as patient's age (greater or less than 50 years, paediatric cases vs. adults), status of underlying disease, previous antifungal prophylaxis, strategies for use of antifungal drugs (sequential vs. de novo combo), and duration of combo, did not affect the outcome (Table 3). At multivariate analysis, only PMN recovery during combo remained significant (P-value 0.009).
Table 3. Factors affecting response to combo antifungal therapy.
Overall, combo was well tolerated. No serious side effects were observed. Only one patient discontinued therapy (voriconazole-related neurotoxicity) and 22% experienced mild (grades I–II WHO) and reversible adverse events (hypokalaemia, liver toxicity and transient creatinine increase) (Table 4).
Table 4. Combo therapy and side effects.
Hypokalaemia (I–II WHO)
Liver toxicity (I–II WHO)
Renal toxicity (I–II WHO)
Stop due to toxicity
After combo, 74% (45/61) of responsive patients received a maintenance therapy with oral voriconazole (34) or posaconazole (11) for a median of 62 days (range 14–380).
After a median follow-up of 4 months (range 1–48), 48/84 (57%) of the patients included in this study have died, mainly due to the refractory or progressive underlying haematological disease. In 14 cases, the IFD was the primary cause of death, with an IFDs-attributable mortality rate of 17% (14/84 cases).
At 12- and 24-week probability of survival (OS) of these patients was 62% and 52%, respectively (Fig. 1A). The responsive patients to the combo had better OS than non-responder cases and the patients who recovered PMN (>500/mmc) during combo had significantly better survival than neutropenic (PMN < 500/mmc) cases (75% vs. 34% at 12 week, P 0.005) (Fig. 1B).
The combo antifungal therapy is still an open issue in haematology because the potential advantages compared to monotherapy are not yet demonstrated.[16-20]
Well-designed and randomized trials are required to define the role of this strategy in haematological patients. However, performing a combination trial in the clinical setting of proven/probable IFDs would be expensive and quite challenging.[16-20] In fact, more than 500 cases are required to perform a randomized trial with an adequate power in this context.[16-20]
In vitro and experimental animal models produced very promising results in terms of efficacy and tolerability of this therapeutic approach.[7-13] However, as we do not have any large prospective and randomized clinical trial with an adequate statistical power, combo in IFDs has now a low level of evidence and recommendation in all current international guidelines.[21-24]
In this study, we report the results of a real-life experience in haematological patients.
Despite the limits of our study design (observational, not randomized) we think it may be of help in displaying the current role of combo in daily clinical practice and in analysing its tolerability and feasibility in haematological patients with proven/probable IFDs. Overall, combo antifungal therapy has been rarely used in the 20 participating centres over 2005–2010 (less than one case/year/centre), mostly in patients with acute leukaemia or after allogeneic HSCT. The most common combinations were those including caspofungin (included in 76% of combinations), voriconazole (in 67% of combinations) or L-AmB (in 52% of combinations). The ORR was 73% (with 35% CR and 38% PR) and the only factor significantly associated with response to combo, in multivariate analysis, was the PMN recovery during therapy.
In our experience, combo resulted to be well tolerated, both in children and in adult patients, and only one of them had to withdraw of treatment (1.2%) for voriconazole-related and reversible neurotoxicity. It is worth noticing that the IFDs-related mortality in this study was low (17%, 14/84 cases) and, as expected, that patients responsive to the combo had a significantly better survival than non-responsive cases.
The literature data on combo in haematology focus mainly on invasive aspergillosis (IA) and mucormycosis and seem to confirm the utility of this therapeutic approach. A summary of recent published clinical studies evaluating combination therapy for IA in haematological patients is reported in Table 5.[25-35]
Table 5. Summary of recent published studies evaluating combination therapy for invasive aspergillosis in haematological patients.
In the context of IA, only two prospective studies on antifungal combo have been published.[25, 26] In the first one, 53 patients with prevalence of pulmonary IA (81%) were treated with caspofungin in combination with other antifungals; the most common underlying disease was acute leukaemia (53%) and most patients (87%) were refractory to prior antifungal therapy. The use of combo resulted in an ORR of 55% (29/53). Fifty-seven per cent of patients with neutropenia and 54% who had received an allogeneic HSCT responded favourably with proportion of survivors being 55% by day 84. Combo was well tolerated. Two (4%) serious drug-related adverse events were reported, both attributed to voriconazole. None of the patients discontinued caspofungin due to toxicity. This analysis showed that caspofungin in combination with a triazole or polyene was an effective alternative to salvage monotherapy for patients with IA.
In the second prospective study (Combistrat pilot trial) only 30 patients with haematological malignancies were analysed (15 in each arm). The median duration of treatment was 18 days for the combo group (caspofungin plus standard dose of L-AmB) and 17 days for the high-dose monotherapy group (L-AmB 10 mg Kg−1 per day). At the end of treatment, the ORR in the combo arm was significantly superior (67%) compared to monotherapy (27%; P 0.028). Survival rates at 12 weeks were 100% and 80% respectively. Infusion-related reactions occurred in three patients in the high-dose monotherapy arm. A twofold increase in serum creatinine occurred in four of 17 patients (23%) who received high-dose monotherapy and one of 15 patients (7%) who received combo.
As for retrospective studies in IA, Kontoyiannis et al.  analysed 48 cases of documented (23) or possible (25) progressive IA, treated with caspofungin plus L-AmB (65%) as salvage therapy, following 7 or more days of a previous L-AmB monotherapy. The ORR was 42% and no significant toxic effects were seen. In the retrospective analysis by Aliff et al. , 30 haematological patients were treated with caspofungin plus L-AmB. The median duration of combo was 24 days and the ORR was 60%. In a small, single-centre study, Marr et al.  evaluated the outcome of 47 patients with IA who experienced failure of first-line therapy with AmB formulations and received either voriconazole (31) or a combination of voriconazole and caspofungin (16) for salvage therapy. The combo of voriconazole plus caspofungin was associated with an improved 3-month survival compared to voriconazole alone.
Another very recent retrospective analysis by Rojas et al.  reported the outcome of 61 haematological patients treated with different combinations in Spanish Hematologic Centers (L-AmB + caspo, 20; L-AmB + triazole, 20; voriconazole+echinocandin, 21). Combo was well tolerated and 38 (62%) patients achieved a response (with 35 CR), regardless of the combo type.
We have also to underline that Marr et al. presented at ECCMID 2012 the final results of the largest (459 patients) prospective, randomized 1:1, phase III trial that compares efficacy and safety of voriconazole plus anidulafungin or placebo in newly diagnosed IA. This study was conducted at 93 sites from 2008 to 2011 and the primary endpoint was 6-week all-cause mortality. Results demonstrated that in the modified intent-to-treat population, 6-week mortality was 19.3% for the patients in voriconazole plus anidulafungin group and 27.5% for the patients in voriconazole plus placebo group (P 0.09). This study showed, for the first time, that a combination of two antifungals determines a reduction in early mortality.
The evidences on clinical efficacy and tolerability of the combo in Mucormycosis are even more limited, given the rarity of this IFD. The SEIFEM and FUNGISCOPE registries recently published the analysis of 32 cases of proven/probable invasive mucormycosis in haematological patients treated with a combination of L-AmB and posaconazole. Thirteen of them (41%) had also received surgical excision of the infected tissue. After a median follow-up of 3 months the ORR was 78% with 11 CR (34%), 7 PR (22%) and 5 stable diseases (16%). These results suggest a beneficial role of combo plus surgery in invasive mucormycosis.
In summary, all these experiences, although difficult to compare, suggest that combo is well tolerated and it seems to be effective either as first or second line, with an ORR greater than 50% in aspergillosis and mucormycosis.[25-37]
However, the use of combo antifungal therapy may be limited by some drawbacks, like the higher costs when compared to single antifungal agents.
The emergence of resistant strains to antifungal drugs is also strictly related to the issue of combo because a resistant fungal pathogen should be suspected in patients not responsive to first-line antifungal monotherapy.[38, 39] This could justify, in selected unresponsive patients the use of combo, especially when in vitro tests of resistance are not available.
In conclusion, this multicentre observational study on antifungal combo, as treatment of proven or probable IFDs in haematological patients, confirms the feasibility and good tolerability of combo and supports previous data on its efficacy.
The use of combo may be justified, in patients with responsive haematological cancer and proven or probable IFD, as antifungal bridge until PMN recovery occurs. In clinical practice, the combo therapy can also be used in combination with surgical approach with the aim of eradicating an IFD that could delay the entire therapeutic programme of underling haematological disease.[38, 39] Clearly, a better characterization of patients with a very high risk of IFDs is therefore important for the selection of those who may be candidate to this kind of treatment as first-line approach. Finally, given the high costs of combo treatment, it may also suggest to reserve combo to those patients without refractory or progressive underlying haematological disease for whom we might expect a very good life expectancy if IFD is cured.
In addition, well-controlled studies are still required to adequately determine the most efficacious antifungal combination regimens for specific fungal infections and to evaluate the pharmaco-economic impact of this strategy.
AC, MC and LP conceived the study and drafted the manuscript. AC and MC collected the data and performed the statistical analysis. All authors read and approved the final manuscript.
AC has received honoraria from Gilead Sciences, Merck, and Pfizer Pharmaceuticals, she has been speaker for Gilead Sciences, Merck, Pfizer Pharmaceuticals. MC has received honoraria from Gilead Sciences and Merck; she has been speaker for Gilead Sciences, Merck, Pfizer Pharmaceuticals, Schering-Plough, and Astellas Pharma. LP has received honoraria from Gilead Sciences, Schering-Plough, Astellas Pharma, Merck, and Pfizer Pharmaceuticals, he has been speaker for Gilead Sciences, Schering-Plough, Merck, Pfizer Pharmaceuticals, Astellas Pharma.
S. Cesaro: Pediatric Hematology Oncology, AOU Integrata, Verona, It. A. Busca: Stem Cell Transplant Center, Ematologia II, AOU San Giovanni Battista, Turin, It. M. Giacchino: Pediatric Onco-Hematology, SCT and Cellular Therapy Division, Regina Margherita Children's Hospital, Turin, It. Fanci R: Haematology Unit, Careggi Hospital and University of Florence, It. Delia M: Hematology Section, Department of Emergency and Organ Transplantation, University of Bari, It. A. Nosari: Divisione di Ematologia, Niguarda Ca’ Granda Hospital, Milano, It. A. Bonini: Transfusion Medicine Unit, IRCCS-Arcispedale Santa Maria Nuova, Reggio Emilia, It. C. Cattaneo: UO Ematologia, Spedali Civili, Brescia, It. L. Melillo: Hematology and SCT Unit, IRCCS, Casa Sollievo della Sofferenza Hospital, San Giovanni Rotondo, It. C. Caramatti: Ematologia-CTMO AOU Parma, It. G. Milone: Division of Hematology and BMT, Ospedale “Ferrarotto”, Catania, It. R. Scimè: Department of Oncology, Hematology Unit, Ospedale “La Maddalena”, Palermo, It. M. Picardi: Department of Clinical Medicine and Surgery, Federico II University Medical School, Naples, It.