Antifungal therapy in 'bone marrow failure'.

Abstract

The incidence of severe opportunistic fungal infections in patients with haematological malignancies has increased dramatically over the past 20 years ( 29; 53; 58; 59; 63). In neutropenic patients these infections are a major cause of morbidity and mortality. Between 20% and 40% of mycoses in patients with haematological malignancies are disseminated, and more than 70% of these are fatal. Candidosis and aspergillosis are the most common fungal infections in patients receiving immunosuppressive therapy and organ transplantation. Late-onset Pneumocystis carinii pneumonia is also seen ( 39). These infections are a particular problem in bone marrow transplant (BMT) recipients. Patients receiving conventional-dose cancer chemotherapy alone for a malignancy will often have a more rapid immune recovery and therefore will be less likely to develop opportunistic fungal infections. In contrast, bone marrow transplant patients may be quite immune suppressed for more than a year after transplant, particularly due to delays of lymphocyte recovery and the prolonged use of medications to prevent graft-versus-host disease. Infection remains the leading cause of death in this population. Most clinical centres have adopted an aggressive approach to antimicrobial therapy in these patients ( 50). At the first sign of infection, broad-spectrum agents, covering both gram-negative and gram-positive organisms, are given. An unresponsive fever or clinical deterioration necessitate alternative antibacterial treatment and additional therapy with antifungal or antiviral agents, or a combination of the two. The conventional interval between lack of response to antibacterial agents and the start of antifungal treatment ranges from 72 to 96 h. The risk of an occurrence of a life-threatening fungal infection varies during the course of the malignant haematological disease ( 66; 29). The sooner remission occurs, the lower the risk that a fatal fungal infection will develop ( 11). The risk depends on the recruitment of host defences. Except for those individuals who are cured after undergoing BMT, the host defence of most patients will sooner or later become compromised, thus resulting in a high risk for fatal infections. The period of immune system reconstitution after BMT is variable in progression and duration. This reconstitution is influenced by multiple factors and in turn strongly influences the risk of infection throughout the post-transplantation period ( 9). However, prior fungal infection is not necessarily a contraindication to bone marrow transplantation ( 26). The early period (0–21 d after BMT) is characterized by neutropenia. Chemoradiation therapies and the use of central venous catheters contribute to the breakdown of normal anatomic barriers, and thereby increase the chance of infection. Infections in this period are predominantly nosocomial because of procedures and loss of skin integrity. The pathogens most often encountered are bacteria, Candida species and herpes simplex virus. Fungal infections must be suspected if fever persists despite broad-spectrum antibiotic treatment ( 15). The incidence of these infections is related to the duration of neutropenia following BMT. Patients with suspected fungal infection should have their stool, urine, blood and respiratory secretions cultured; they should undergo biopsy of any skin lesions; and they should have their retinas examined for fungal endophthalmitis since fungaemia is difficult to diagnose. In neutropenic patients with fever not responsive to antibiotics and in patients with positive fungal blood cultures, a course of intravenous amphotericin B is warranted. The middle period (21–100 d) is characterized by either continued neutropenia from graft failure or rejection, and hence possible fungal infection, or successful engraftment with marrow maturation and the gradual return of immune function. The main obstacle to survival after marrow engraftment is acute graft-versus-host disease (GVHD). Conflicting results have been reported in regard to acute GVHD as a risk factor for invasive aspergillosis. In the recent study reported by 29), patients with grade III–IV acute GVHD were very susceptible to invasive aspergillosis. Patients with extensive chronic GVHD were also at high risk of late infections mostly caused by Aspergillus species. The type of GVHD has not been analysed as a possible risk factor for invasive fungal infection in most published series. 29) found that the risk of fungal infection was 14% in patients treated with MTX plus methylprednisolone and as high as 38% in patients treated with ATG. These authors point out that although it is conceivable that the patients who responded poorly to high-dose methylprednisolone had more severe GVHD and hence more immunosuppression, it should be borne in mind that ATG causes profound and prolonged immunosuppression in its own right. The second major risk to the patient during this period is interstitial pneumonia caused primarily by Aspergillus ( 23). This has been seen in up to 10% of BMT patients in some centres and has an overall mortality in excess of 90%. The presentation of increased respiratory rate, dry nonproductive cough, and fever occurring approximately 1.5 months after BMT is suggestive of interstitial pneumonia. Despite the demonstration of infectious aetiological agents in most cases of interstitial pneumonia, its origin is probably multifactorial. In addition to toxicity secondary to chemotherapy and radiation and infection by viral or opportunistic pathogens, including Aspergillus and Candida, direct toxicity against infected cells in the lung is probably implicated ( 23). Furthermore, the patient's immunosuppressed state, the development of GVHD, and indwelling venous catheters all contribute to the development of sepsis with fungi such as Candida. The late period (> 100 d) is associated with fewer infections than the earlier postoperative periods, since most intravenous lines have been removed, fewer immunosuppressive agents are used, and the immune status has progressively recovered. The majority of children undergoing bone marrow transplantation receive T-lymphocyte-depleted allografts from partially matched related or matched unrelated donors. This setting predisposes to opportunistic infections, principally aspergillosis and candidiasis ( 27; 32). Most children who acquire aspergillosis during the period when they remain immunocompromised after a marrow transplant will succumb to aspergillosis. The infection is particularly lethal in patients during the first month or two after a marrow transplant. In a similar way to the adult setting, most transplant units are constructed to reduce patient exposure to fungal spores and thereby reduce the incidence of fatal aspergillosis. The diagnosis of fungal infection in BMT recipients remains a major difficulty for the clinician. In recognition of this problem a series of recommendations have been drawn up in order to give guidelines to clinicians, microbiologists and infectious disease physicians which incorporate currently available diagnostic techniques ( 17). These include relevant invasive procedures, radiological investigations and histological examinations, as well as the conventional, and newly developed, mycological techniques of culture and serology. Specific recommendations and comment will be found in the individual disease sections. The laboratory diagnosis of fungal infection in the BMT recipient involves one or more of four approaches: (i) a thorough examination of respiratory secretions by direct microscopy, (ii) isolation of the organism, (iii) detection of antibody or antigen, and (iv) histopathologic evidence of invasion ( 36; 66). The isolation of fungi from otherwise sterile sites provides vital information to the clinician. Sputum cultures that yield Aspergillus species should be considered significant unless proved otherwise; however, expectorated sputum is generally of little help in the diagnosis of fungal infections. Where there is diffuse disease bronchoalveolar lavage (BAL) or bronchial biopsy should provide a more reliable and accurate means of diagnosing pulmonary infection and may be positive in up to 60% of cases. Culture is invariably negative where there are focal lung lesions. Here, radiographically-guided fine needle biopsy may be helpful. However, if a BAL fluid is the only available specimen available it should be screened very extensively for fungal elements using direct microscopy and Calcofluor white staining. Serological tests have not been particularly successful in the early diagnosis of invasive disease in transplant recipients. Tests designed to detect circulating protein and carbohydrate antigens of Candida have been extensively evaluated but their value remains controversial. Although the detection of circulating antigen may correlate with invasive aspergillosis, the routine use of antigen detection tests has still to be accepted. Persistent fever suggests an occult fungal infection. The likelihood of this increases with the number of preceding febrile episodes. It has been shown that 44% of patients suffering their fourth bout of fever had a fungal infection as the cause. Patients with prolonged neutropenia (> 7–10 d) or receiving high doses of glucocorticoids during neutropenia, and a high probability of an invasive fungal infection should be given empirical systemic antifungal therapy. This recommendation is based on numerous reviews and clinical studies ( 12; 21; 24; 31; 32; 34; 35; 40; 49; 56; 60). Such patients, with fever persisting despite antibiotic therapy or presenting at the onset of a febrile episode with findings suggesting the possibility of an invasive infection, should be started on antifungals. In this situation amphotericin B must be used. If possible, the full therapeutic dosage level (1.0–1.5 mg/kg/d of the conventional formulation) should be reached within the first 24 h of treatment. There is no need for gradual escalation of dosage, nor is there evidence to support the clinical prejudice that a lower dose can be used in suspected candidiasis. Should the conventional formulation be contraindicated, one of the lipid-complexed formulations should be used instead ( 8). For example, in a Europe-wide, randomized, double-blind comparative trial of the liposomal formulation of amphotericin B (AmBisome) versus conventional amphotericin B (cAMB) in the empiric treatment of both paediatric and adult febrile neutropenic patients AmBisome (3 mg/kg/d) was as effective as cAMB in the prevention of proven treatment-emergent fungal infections ( 52). Of significance was the reduction in overall drug-related toxicity by 2–6-fold in patients treated with AmBisome compared with cAMB. Severe drug-related adverse reactions were almost absent in patients treated with AmBisome. Nephrotoxicity, in patients not receiving concomitant nephrotoxic agents, was seen in only 3% of patients treated with AmBisome, compared to 23% of patients on conventional amphotericin B. Time to develop nephrotoxicity was longer in patients on AmBisome compared to cAMB. A significantly larger study, which formed the basis of a successful new drug application to the Federal Drug Agency (FDA) for the use of AmBisome in empirical therapy of presumed fungal infections in febrile neutropenic patients, has been reported recently ( 64). This study was a multicentre double-blind trial that compared AmBisome (3 mg/kg/d) with cAMB (0.6 mg/kg/d). Nearly 700 adult and paediatric patients with neutropenia and pyrexia of unknown origin (PUO) were randomized to receive either cAMB or AmBisome. Therapeutic success, measured by a combination of factors including resolution of fever, absence of emergent fungal infection and patient survival for at least 7 d after therapy, was equivalent between the two groups. There were significantly fewer proven treatment-emergent fungal infections in patients treated with AmBisome than in patients treated with cAMB. Similar to the 52) report, there was a significant reduction in the frequency of infusion-related fever and in the development of nephrotoxicity. The conclusions that can be drawn from the 64) study are that AmBisome was equivalent to cAMB for empirical antifungal therapy in neutropenic patients, but superior in reducing proven treatment-emergent fungal infections, nephrotoxicity and infusion-related toxicity. The duration of treatment will differ from patient to patient. If the patient responds and a diagnosis of fungal infection is established, a full course of treatment should be given. More often, however, the patient responds and/or the neutrophil count recovers, but a firm diagnosis is not obtained. In this situation it is reasonable to discontinue amphotericin B when the neutrophil count goes above 1 × 109/l, the fever and other symptoms and signs resolve, and relevant radiological abnormalities return to normal. Neutropenic patients who recover from a deep fungal infection, such as aspergillosis, may suffer from reactivation of the infection during subsequent periods of immunosuppression. One solution to this problem is to begin empirical treatment with amphotericin B (1 mg/kg/d) not less than 48 h before antileukaemic treatment is commenced. This drug should be continued until the neutrophil count has recovered. It is important to remember that empirical antifungal therapy does not preclude further investigation for other occult causes of fever (drug-induced fever, viral infection, possibly Pneumocystis, rarely parasites or Mycobacteria). In addition, a change in fever pattern, including abatement of previously continuous fevers, which frequently follows initiation of amphotericin B therapy, cannot be taken as guarantee of a favourable response to empirical antifungal therapy. Furthermore, efforts to document fungal infection during empirical therapy in order to adapt antifungal therapy to specific agents and their clinical symptoms should not be abandoned. A sharp distinction between empirical therapy and treatment of an established infection is often blurred. Frequently, amphotericin B is added to an empirical antibacterial regimen in patients with persisting fever of unknown cause during prolonged aplasia, but in many instances additional diagnostic information can be made available in this situation, making a fungal infection highly probable and possibly indicating the most likely pathogen. Thus in a patient with cough, pleuritic chest pain and/or a new pulmonary infiltrate, or in a patient developing sinus tenderness during broad-spectrum antibiotic therapy in an appropriate clinical and epidemiological setting, aspergillosis is likely. If in such a patient typical hyphal elements are seen in BAL or aspirates and Aspergillus fumigatus or A. flavus is grown from respiratory secretions, a diagnosis of invasive aspergillosis must be assumed. In contrast, persistent or recurrent fever at the time of recovery from neutropenia, possibly associated with right upper quadrant tenderness, an increase in serum alkaline phosphatase activity or development of focal hepatic and splenic lesions on CT scan, magnetic resonance imaging or ultrasonography indicate a diagnosis of disseminated candidiasis. Based on these findings, antifungal therapy can be adapted for optimal activity against a presumed pathogen. However, the possibility of an infection with a less common fungal species must be kept in mind. An optimal microbiological and histological identification of the pathogen responsible for a presumed mycosis is very much dependent on invasive procedures that provide tissue specimens. Identification of these potential pathogens is particularly important, because this might lead to deviation from the main therapeutic recommendations for candidiasis and aspergillosis. For example, Fusarium species and Trichosporon species are not reliably sensitive to amphotericin B. Invasive aspergillosis (IA) is one of the most common fungal infections in haematological neoplastic disease (14–20% of cases) ( 47; 59; 61). The presentation of acute pulmonary aspergillosis often mimics an acute bacterial pneumonia. Cough, usually unproductive, and fever are the most frequent presenting symptoms. Pleuritic chest pain is common and a pleural friction rub is not unusual. Pleuritic pain in a leukaemic patient with persistent cough and fever suggests invasive aspergillosis even in the absence of a chest X-ray abnormality. Cavitation is strongly suggestive of invasive pulmonary aspergillosis in neutropenic patients. CT scanning of the chest has made a major impact on the management of this patient group ( 22; 10). It is more sensitive than chest radiography and is particularly valuable when the chest X-ray is negative or shows only subtle changes. In one study intrathoracic complications of BMT were found using CT scanning in 57% of patients in whom the chest X-ray was negative ( 22). It can often differentiate between invasive pulmonary aspergillosis and bacterial and viral infections. Typical features of invasive aspergillosis include wedge-shaped peripheral areas of consolidation, usually extending to the pleural surface, with or without cavitation, or nodular areas of consolidation, often related to blood vessels. CT scanning is also helpful in guiding further invasive diagnostic procedures, such as the best location for needle biopsy or open lung biopsy, or in defining whether bronchoscopy is the best modality for confirming diagnosis. Confirmation of the diagnosis of Aspergillus infection ideally requires culture of tissue from deep sites: invasive procedures to obtain this type of clinical specimen are difficult to perform in neutropenic and thrombocytopenic patients. The use of Aspergillus antigen tests (latex agglutination and ELISA) have proved to be useful ( 47); some are commercially available. PCR methods are currently being developed and evaluated. The poor sensitivity for the early recovery of Aspergillus species from routine culture specimens limits the use of surveillance methods for the early detection of infection ( 38; 47). Surveillance of nasal mucosal surfaces is of limited value. When an Aspergillus species is found in a respiratory specimen, its significance is often questioned, because it may indicate colonization only. However, the conclusion drawn from several studies is that a positive respiratory culture for Aspergillus in a profoundly immunocompromised host is highly suggestive of invasive disease. It has also been noted that, although false-positive sputum cultures are common, multiple positive cultures are more indicative of true infection than is a single positive culture. Therefore a positive sputum culture in a neutropenic BMT patient or in a solid organ transplant recipient undergoing treatment for rejection should prompt a thorough evaluation and institution of antifungal therapy. The successful management of acute invasive aspergillosis in the neutropenic patient depends on the prompt initiation of antifungal treatment (within 96 h of the onset of infection) ( 18; 16). The response rate differs from one host group to another and depends on the length of time for which treatment is administered. In BMT recipients the response rate is about 10% and in neutropenic cancer patients it is around 30%. Patients who have survived have all received at least 2 weeks of antifungal treatment. The prognosis is poor if the neutrophil count does not recover. Limited clinical experience suggests that shortening the duration of neutropenia with colony-stimulating factors may be beneficial in treating invasive aspergillosis. Surgery appears to be of some benefit in patients with paranasal infection ( 33). The basic approach to the treatment of invasive aspergillosis is outlined in Table I. Amphotericin B is the standard treatment for invasive aspergillosis but has limited success. Reconstitution of bone marrow function is only one major determinant of a successful outcome. Persistent steroid-induced immunosuppression (as seen in allogeneic bone marrow grafts) is associated with a poor outcome, even in the absence of granulocytopenia. There are numerous regimens for the administration of this drug, but there is widespread agreement that in neutropenic patients it is important to give the full dose of amphotericin B from the outset. High doses must be used (at least 1.0 mg/kg/d). The optimum duration of treatment has not been established, but amphotericin B should be continued at least until the neutrophil count is > 0.5 × 109/l. Thereafter treatment should be continued until symptoms resolve and relevant radiological abnormalities (on X-rays and CT scans) disappear. The shortcomings of current methods of diagnosis often require clinicians to proceed to amphotericin B treatment without waiting for formal proof that a neutropenic patient has persistent fever (> 72–96 h duration), and is unresponsive to antibacterial drugs. Empirical treatment should be initiated with the usual test dose (1 mg) of amphotericin B. If possible, the full therapeutic dosage level should be reached within the first 24 h of treatment ( 21). Rapid escalation of the dose carries a greater risk of acute renal failure, but immunocompromised patients often tolerate these regimens well. During the 1980s several investigators incorporated amphotericin B into lipid vehicles and showed a reduction in the toxic adverse effects without loss in the effectiveness of amphotericin B. Initially, such formulations were produced locally and used at single hospitals. In more recent years, lipid complexes of amphotericin B have become commercially available. These formulations include: the liposome AmBisome (Vestar, Nexstar Inc., Nexstar Pharmaceuticals, Boulder, Colorado, U.S.A.); a lipid vesicle–amphotericin B colloidal dispersion: Amphocil (Amphotec) (Liposome Technology Inc., Menlo Park, California, U.S.A.); an amphotericin B–lipid complex: Abelcet (The Liposome Company Inc., Princeton, New Jersey, U.S.A.). All of these preparations differ in size, structure and pharmacokinetics (reviewed in 25), and, to a certain extent, in the clinical efficacy in the treatment of invasive aspergillosis (reviewed in 54; 57). At the present time it is impossible to give recommendations regarding doses until a consensus of the most appropriate dosage indications for maximum efficay and agreement on toxicity tolerances have been established on the basis of published and unpublished studies. The recommendations given below are an attempt to adhere to the dosages given in the appropriate licence documents. The most-well established of the lipid-complexed formulations of AMB is AmBisome (Nexstar Pharmaceutical Inc.). This is the only true lipsomal formulation. The use of AmBisome clinically is supported extensively by published data. Over 150 publications on safety and efficacy of AmBisome illustrate how beneficial this agent has been. Liposomal amphotericin B (AmBisome) is well tolerated and doses as high as 10 mg/kg/d have been administered without significant side-effects. Some 10 000 patients have been treated worldwide over the past 6 years. Administration of the drug in this form has sometimes eradicated Aspergillus infection in neutropenic patients, and it should be considered in patients who have failed to respond to the conventional parenteral formulation, or who have developed side-effects that would otherwise necessitate discontinuation of the drug. A number of substantive studies have shown that AmBisome (typcal dose 3 mg/kg) is as effective as conventional amphotericin B in the treatment of invasive aspergillosis (reviewed by 44; 54; 57). Three key studies are summarized here. 14) showed that in seven patients (two AML, one ALL, two autologous BMT and one allogeneic BMT) three were cured (42%) and that four patients failed and died. In a compassionate phase II–III multicentre study where126 patients were included, 12 patients with proven aspergillosis were evaluable: nine (32%) were cured, eight improved and 11 failed ( 55). Long-term survival was not specified. A further study described 11 haematological patients, seven of whom had confirmed invasive aspergillosis. Resolution was observed in 7/11 cases (63%). The largest series of patients with IA treated with AmBisome was reported by 43): 116 patients were incorporated into the study, 21 of whom had proven IA; 18 of these had received previous cAMB. A complete resolution was observed in 10 patients and a partial resolution in three, giving a combined response rate of 62% even in the context of CAB failure. It was pointed out that eight patients had a ANC of < 1000 at discontinuation of therapy and that none had relapsed. Long-term survival was not specified. In an additional group of 36 patients with suspected aspergillosis, 12 responded completely to AmBisome treatment and seven partially responded. It is usual to begin treatment with AmBisome with a dose of 1.0–3.0 mg/kg, or even higher. This formulation is infused over a 30–60 min period. AmBisome has been administered to individual patients for up to 3 months, to a cumulative dosage of 15 g without significant toxic side-effects. These and other studies have emphasized that patients who responded to AmBisome tended to have received a larger (cumulative) dose for a longer period of time than those in whom treatment failed. Among the patients who had proven invasive aspergillosis, those in whom AmBisome was used as salvage therapy had a less favourable prognosis than those who had never been treated with cAMB. A notable feature of AmBisome is the very low incidence of acute infusion-related adverse reactions; in fact, where such reactions do occur, patients appear to tolerate subsequent doses much more readily. There is no requirement for test dosing, slow escalation or premedication. Furthermore, AmBisome appears to have an underlying immunomodulating effect by enhancing effector cell function against fungal cells. The studies on the accumulation of AmBisome around fungal lesions and its mode of action at the fungal cell membrane are well accepted. Amphocil, Amphotec (amphotericin B colloidal dispersion, ABCD) is formed from equimolar amounts of amphotericin B and cholesterol sulphate; it has a disc-like form with a mean size of 122 nm and is rapidly taken up by the liver. The tolerance and efficacy of this formulation have been assessed in three clinical trials and independent studies. However, the clinical data is still too limited to allow firm conclusions to be drawn. Nevertheless the collective data on compassionate use of Amphocil in patients with invasive aspergillosis is encouraging, but the optimal dosage and duration of the use of Amphocil has not been established. It is recommended an initial dose of 1 mg/kg is given as a single infusion over a 60–90 min period, with a step-wise increase to 3–4 mg as required The collective U.S.A. and European experience has been presented by 45). Where 168 patients with a variety of underlying conditions (haematological malignancies, 92) were enrolled and 97 were evaluable, 32 patients had IA and six were infected with Aspergillus and another pathogen. In the patient group with IA a complete resolution was seen in 16%, an improvement in 19%, and 66% failed. The clinical response in relation to underlying disease was not analysed specifically for aspergillosis. The mortality rate within 4 weeks of the last dose was given for the entire patient population (88/168) but not for the patients with IA. In an analysis where 82 patients with proven or probable IA were treated in clinical trials with Amphocil were compared retrospectively with 261 patients with aspergillosis treated with amphotericin B, it was shown that Amphocil caused fewer nephrotoxic effects than cAMB and the efficacy of Amphocil was at least comparable with that of the conventional formulation. Further U.S.A. experience is reported by 65). Abelcet (amphotericin B lipid complex, ABLC) consists of amphotericin B complexed with two lipids, dimyristoylphosphatidylcholine and dimyristoylphosphatidylglycerol, in a 1:1 drug:lipid molar ratio. In animal models, ABLC has been shown to be at least as effective as amphotericin B and substantially less toxic. The drug is indicated for the treatment of severe invasive candidiasis and as second-line treatment of invasive aspergillosis, cryptococcosis in HIV patients and miscellaneous fungal infections ( 20; 42; 46). The drug appears to be cleared rapidly from the serum. High tissue levels have been recorded in animals, especially in the lung, followed by rapid clearance of the drug from the lung area over 24 h. Human data regarding accumulation of Abelcet in lung tissue appears to be based on post-mortem tissue from one patient. The initial clinical trials were conducted in the U.S.A. (reviewed in 37). The size and quality of these studies varied considerably and the conclusions that can be drawn are limited in the absence of confirmatory studies. The recommended dosage of ABLC is 5 mg/kg and this should be infused over a 2 h period for at least 2 weeks. Response to treatment times are highly variable. This formulation of amphotericin B has been administered to individual patients for up to 11 months, to a cumulative dosage of 50 g without significant toxic side-effects. Very few studies evaluating the effectiveness of Abelcet in invasive aspergillosis have been described. 42) reported that five of seven evaluable patients with Aspergillus pneumonia responded to Abelcet (daily dose of 5 mg/kg). T