-D-Glucan Detection Test: A Step Toward Preemptive Therapy for Fungal Infections in Leukemic Patients?

Invasive fungal infections are still a challenge in patients with prolonged neutropenia. The incidence is often >10% during neutropenia after induction or consolidation therapy for acute leukemia [1, 2]. Invasive fungal infections still carry a 30%–50% mortality rate for invasive aspergillosis and an ∼40% mortality rate for invasive candidiasis, despite the recent development of new antifungal agents. The death rates in monocentric retrospective studies are always higher than in clinical trials. For instance, we observed a 52% 12-week survival rate for patients with invasive aspergillosis in our medical center, compared with a 71% survival rate for patients treated with voriconazole and a 72% survival rate for patients treated with liposomal amphotericin B in 2 recent clinical trials [3–5]. Similarly, Pagano et al. reported a 12-week survival rate as low as 26% in hematological patients with various types of invasive fungal infection [6]. The differences in survival between retrospective reviews and prospective clinical trials can be explained by exclusion from clinical trials of the patients with the worst prognosis—very short life expectancy due to refractory malignancy; significant renal or hepatic dysfunction; concomitant infection with another fungal, bacterial, or viral pathogen; intubation; and mechanical ventilation are all usual exclusion criteria for clinical trials and are known to adversely affect patient outcome [3, 7]. Outcome is improving with advances in clinical practice. In a large review of invasive aspergillosis after hematopoietic stem cell transplantation, Upton et al. [7] underscored the beneficial impact on survival of the improvement in the management of the underlying condition, the earlier diagnosis of fungal infection, and the widespread use of new antifungal agents. Treatment strategies in leukemic patients. Early onset of therapy has already been associated with a better survival in patients with invasive aspergillosis [8]. Starting the therapy at the stage of possible aspergillosis, before the criteria for probable or proven aspergillosis are met, is associated with a reduced overall mortality [3]. According to the European Organisation for Research and Treatment of Cancer/Mycoses Study Group (EORTC/MSG) consensus criteria, possible invasive fungal infections are defined in an immunocompromised patient by clinical and or radiological signs without mycological criteria or by mycological criteria without clinical or radiological signs [9]. Therefore, starting therapy in possible case patients means starting preemptive therapy, because the diagnosis is likely but not documented. The concept of preemptive therapy has been developed for cytomegalovirus disease in allogeneic hematopoietic stem cell transplant recipients. Positive surveillance antigen detection test results or PCR results appropriately detect patients at highest risk of developing cytomegalovirus diseases. Early therapy is associated with an improved outcome and reduces significantly the number of patients exposed to antiviral drugs, compared with the number exposed to prophylaxis. There is not yet consensus about the definition of preemptive therapy for fungal infections in neutropenic patients, as opposed to prophylactic, empirical, or directed therapy. Directed therapy applies to documented—either probable or definite—fungal infections, as defined by the EORTC/MSG criteria [9]. Empirical therapy has been largely promoted in neutropenic patients, and several major clinical trials have used the same inclusion criteria: persistent fever in a neutropenic patient, despite broad-spectrum antibiotic therapy, in the absence of any other sign suggestive of fungal infection (especially no focal sign of infection and no positive mycological test results) [10–12]. Prophylaxis clearly applies to patients at high risk of fungal infection but is started before any sign of infection is apparent. These strategies seem to form a continuum from prophylaxis to directed therapy. However, preemptive therapy is not yet clearly placed on this continuum (figure 1). Preemptive therapy can be placed on the continuum after empirical therapy: the patient is febrile and presents some other clinical signs consistent with an invasive fungal infection. The best example is a neutropenic febrile patient with cough, pleural pain, and nodules on chest CT scan but no positive microscopy findings or culture results from respiratory samples and no test results positive for Aspergillus galactomannan in serum or bronchoalveolar lavage samples. Preemptive therapy can also be placed on the continuum before empirical therapy. In this case, the patient has no fever and no focal sign of infection, but surveillance tests, such as Aspergillus galactomannan or Candida mannan detection tests or fungal PCR, show 1 or several positive results. Actually, both places on the continuum are not mutually exclusive; this suggests that empirical therapy has a more hazy definition and might be absorbed into preemptive therapy, with a future improvement in a diagnostic test. Therapeutic strategies in leukemic patients at high risk of invasive fungal infections. *Biomarkers include Aspergillus galactomannan, Candida mannan, β-d-glucan, or PCR. #Positive microscopy, culture, or histopathology result from a clinically relevant sample. Mycological diagnostic tests applied to leukemic patients. Microscopy and culture techniques have their limitations. The samples are difficult to obtain in respiratory tract or other deep-seated infections and may require invasive procedures. In addition, sensitivity of these tests is insufficient; the Gruppo Italanio Malattie Ematologiche dell'Adulto group showed, in a large retrospective review, that sputum, bronchoalveolar lavage fluid, and nasal swab specimens had only 58%, 62%, and 32% sensitivity rates, respectively, in pulmonary fungal infections associated with hematological malignancies [6]. Biomarkers of invasive fungal infections will play a critical role in the evolution of treatment strategies. Thus far, A. galactomannan and C. mannan detection tests in serum lack sensitivity and specificity to guarantee appropriate diagnosis of Aspergillus and Candida infections. Although these tests may have positive results before clinical signs have developed, the sensitivity is only ∼40% for C. mannan and ∼70% for A. galactomannan [13, 14]. PCR is not yet standardized for routine use and is expensive for routine monitoring. However, the negative predictive value of Aspergillus PCR is high, and the costs could be defrayed by the savings on expensive antifungal agents [15, 16]. Obviously, other non-culture-based rapid diagnostic tools are needed. Detection of β-d-glucan might be one these tools. β-d-glucan is a component of the cell wall of a wide variety of fungi and can be detected by its ability to activate factor G of the horseshoe crab coagulation cascade [17]. It has been widely assessed in Japan, with promising results [18–20]. Odabasi et al. [21] confirmed these encouraging results in a large population of leukemic patients. The potential role for β-d-glucan has been acknowledged by its inclusion in the revised EORTC/MSG definitions. Performance of β-d-glucan detection test. In this issue of Clinical Infectious Diseases, Senn et al. [22] report the assessment of β-d-glucan for early diagnosis of invasive fungal infections in neutropenic patients with acute leukemia. β-d-glucan antigenemia was measured twice weekly in the absence and daily in the presence of fever during 190 consecutive neutropenic episodes in 95 different patients. All patients were undergoing induction or consolidation chemotherapy for acute leukemia and experienced prolonged neutropenia. Patients were hospitalized in high-efficiency particulate air-filtered air isolation rooms, and those with mucositis and known digestive tract fungal colonization received prophylaxis with fluconazole. Overall, 62 episodes of invasive fungal infection were diagnosed, with more than one-half of the cases reaching the level of probable or proven infection. A median of 16 samples were collected for β-d-glucan measurement per neutropenic episode. The best performance of the test was obtained with 2 consecutive samples ⩾7 pg/mL. Sensitivity and specificity were 63% and 96%, respectively. Reducing the cutoff value to <7 pg/mL resulted in a major decrease in specificity. Similarly, consideration of only 1 positive β-d-glucan test result decreases dramatically the specificity (75% with the same cutoff value of ⩾7 pg/mL). Very interestingly, the performance of the test was similar for aspergillosis and candidiasis. The median time elapsed between fever onset and first β-d-glucan result was <1 day, with earlier results in aspergillosis (median days after onset of fever, 0), compared with candidiasis (median days after onset of fever, 2). Positive β-d-glucan tests result were obtained before the results of any other conventional diagnostic method, including histopathology, cultures, and radiological signs. Close monitoring of fungal biomarkers or prophylaxis? This study by Senn et al. [22] has a limitation. The incidence of invasive fungal infection is especially high for both candidiasis and aspergillosis: 62 episodes (including the possible cases) occurred in 58 different patients, out of a total of 95 patients exposed to 190 episodes of neutropenia. More than 60% of the leukemic patients treated with a mean of 2 cycles of induction, reinduction, or consolidation chemotherapy developed an invasive fungal infection, despite hospitalization in an isolation room with high-efficiency particulate air-filtered air for all patients and prophylaxis with fluconazole for some of the patients. With such a high incidence of fungal infection, efforts should probably be focused more on prophylaxis than on earlier diagnosis. At the time the study was initiated and conducted, there was no demonstration of effective prophylaxis against aspergillosis. Now, 2 large studies—1 involving allogeneic stem cell transplant recipients with graft versus host disease and the second involving patients with acute leukemia or myelodysplastic syndrome receiving intensive chemotherapy—have been completed, and results have been published [23, 24]. In both studies, posaconazole has shown efficacy in preventing fungal infections, including invasive aspergillosis. We now must decide whether to monitor twice-a-week biomarkers of fungal infection and to order a CT scan of the chest if any of the tests are positive or to administer routine prophylaxis with posaconazole for leukemic patients undergoing intensive chemotherapy. Thus far, no general recommendation can be given; the decision must be based mainly on local incidence of fungal infections and on the ease of access to the biological tests and to a CT scan. The decision must also integrate the very high cost of prophylaxis; the number needed to treat with posaconazole to prevent 1 case of aspergillosis in leukemic patients was 16 in the comparative prophylaxis trial [23]. Maertens et al. [25] demonstrated the safety of a preemptive strategy based on galactomannan assessment and a CT scan of the chest, and they proposed a comparison of empirical and preemptive therapy in a randomized trial. In fact, the new data on biomarkers, including the report of Senn et al. [22], and on prophylaxis call more for a comparison between prophylaxis and preemptive therapy driven by positive serum biomarker and/or PCR results. Potential conflicts of interest. R.H. has served as consultant for Gilead, Merck Sharp & Dohme, Novartis, Pfizer, and Schering-Plough and has received a grant from Pfizer. A.B.: no conflicts.