develop invasive fungal infections. 1 4 The diagnosis of fungal infections is particularly difficult in immunocompromised

Transcription

1 Bone Marrow Transplantation, (1997) 20, Stockton Press All rights reserved /97 $12.00 Amphotericin B lipid complex (ABLC) for the treatment of confirmed or presumed fungal infections in immunocompromised patients with hematologic malignancies J Mehta 1, S Kelsey 2, P Chu 3, R Powles 1, D Hazel 2, U Riley 1, C Evans 1, A Newland 2, J Treleaven 1 and S Singhal 1 1 Royal Marsden Hospital, Surrey; 2 Royal London Hospital, London; and 3 Royal Liverpool University Hospital, Liverpool, UK Summary: develop invasive fungal infections. 1 4 The diagnosis of fungal infections is particularly difficult in immunocompromised Sixty-four adult patients (median age 43) with hematologic patients, and most patients are treated presumptively. malignancies who were immunocompromised after Early institution of anti-fungal therapy is associated with allogeneic (n = 23) or autologous (n = 9) blood/marrow better outcome, and withholding therapy until a definitive transplantation, or chemotherapy (n = 32) received 68 diagnosis is made results in dissemination of the infection courses of amphotericin B lipid complex (ABLC, with a high mortality. 4 Abelcet) at the daily dose of 5 mg/kg for presumed Amphotericin B is the primary treatment for systemic (n = 52) or proven (n = 16) fungal infection. The major fungal infections in immunocompromised patients because indications for ABLC were failure of previous antifun- of its broad spectrum of activity. 5,6 However, its use is gal therapy and/or renal dysfunction. Fifty-three treatment limited by a variety of adverse effects including severe courses in 49 patients comprising 4 58 doses systemic reactions such as fever, rigors, phlebitis and (median 10) were considered evaluable. Fourteen bronchospasm. 6 The major problem with amphotericin is courses administered for confirmed infections resulted its dose-related nephrotoxicity which compromises its efficacy in nine complete and one partial responses, and four in treating systemic fungal infections and restricts its failures (71% response). Thirty-nine empiric courses prophylactic use. 7 Drug-induced renal dysfunction is a resulted in 18 complete and six partial responses, and major concern after blood or marrow transplantation 14 failures (64% response). The overall response rate (BMT), 8,9 where the use of a less nephrotoxic formulation was 66%. Five of seven evaluable patients with aspergillus of amphotericin would be particularly attractive. 10 pneumonia responded. Response rates were compa- A number of lipid-based formulations of amphotericin rable for chemotherapy, autograft and allograft recipients. have been developed to improve the tissue distribution of The change in serum creatinine from the beginning amphotericin and to reduce the toxicity associated with to the end of therapy was 284 to +277 mol/l (median conventional amphotericin. 10 Amphotericin B lipid com- +24). The creatinine doubled during seven evaluable plex (ABLC) is derived from a liposomal formulation of courses of therapy, five of which were associated with amphotericin B consisting of dimyristoyl-phosphatidyl- concomitant nephrotoxic therapy. Nephrotoxicity was choline and dimyristoyl-phosphatidylglycerol in a 7:3 comparable for transplant and chemotherapy patients. molar ratio. Unlike the original liposomal formulation, Renal dysfunction necessitated discontinuation of ABLC has a ribbon-like appearance. 10 ABLC in only four patients. These data suggest that Here, we report our experience with the use of ABLC ABLC is effective in presumed or confirmed fungal for the treatment of suspected or proven fungal sepsis in infections in immunocompromised patients after transplantation immunocompromised patients with hematologic malig- or chemotherapy. nancies undergoing BMT or receiving chemotherapy. Keywords: Abelcet; amphotericin B, amphotericin B lipid complex; aspergillosis; candidiasis; fungal infection Patients and methods Approximately one-third of neutropenic patients who Between May 1994 and October 1995, 64 adult patients remain persistently febrile on standard antibiotic therapy with hematologic malignancies who were immunocompromised or neutropenic as a result of chemotherapy or BMT received ABLC (Abelcet, The Liposome Company, Prince- Correspondence: Dr J Mehta, Division of Hematology/Oncology, Slot 508, ton, NJ, USA) at the Royal Marsden Hospital (n = 46), the University of Arkansas for Medical Sciences, 4301 West Markham, Little Royal London Hospital (n = 10), or the Royal Liverpool Rock, Arkansas 72212, USA University Hospital (n = 8). All patients had received some Presented at the 37th Annual Meeting of the American Society of Hema- tology, Seattle, December 1995 Received 4 November 1996; accepted 11 March 1997 form of topical anti-fungal prophylaxis (nystatin with or without amphotericin lozenges).

2 40 Diagnosis of fungal infections ABLC in immunocompromised patients Refractory fever not responding to at least 72 h of broad- spectrum combination antibiotic therapy and/or coarse pulmonary infiltrates with negative cultures from all appropri- ate sites were considered evidence of a presumptive fungal infection. Patients with coarse pulmonary infiltrates usually received therapy directed against atypical bacteria, Pneumocystis carinii, and cytomegalovirus in varying combinations depending upon the clinical situation in addition to anti-fungal therapy. Confirmed fungal infections were diagnosed on the basis of positive cultures or histologic results. Growth of a Candida species organism from stool or throat in the absence of symptoms was not considered an evidence of candidiasis. ABLC (usually from the morning of the day ABLC was started) was considered the baseline creatinine, and the first creatinine level after the last dose of ABLC (usually from the morning of the day after ABLC was stopped) was considered the final creatinine. The effect of the drug on the kidneys was expressed as the difference in these two values. Concomitant therapy with other nephrotoxic drugs (cyclosporine, ganciclovir, foscarnet, high-dose intravenous acyclovir, vancomycin and aminoglycosides) was specifi- cally recorded. Data were collected using a standard proforma at the time of treatment or retrospectively from the notes (Royal London Hospital and Royal Liverpool University Hospital) and the prospective Leukaemia/Myeloma Research Data- bases of the Royal Marsden Hospital. 12,13 ABLC therapy The indications for ABLC therapy 11 were as follows: (1) proven invasive fungal infection; (2) strong clinicoradiologic picture of fungal infection in patients who were neu- tropenic and expected to remain neutropenic for the next few days, and who were on other nephrotoxic drugs; (3) refractory fever with compromised renal function ( 50% elevation of creatinine from baseline) and/or high intravenous potassium supplementation requirements ( 120 meq per day); (4) failure to respond to conventional amphotericin or fluconazole (persistent fever for 3 days after introduction of antifungal therapy in presumed infections, or failure to show any clinical, radiologic or microbiologic sign of improvement for 3 days in confirmed infections); or (5) intolerance of conventional amphotericin (severe systemic reactions despite adequate premedication). ABLC was administered intravenously at a daily dose of 5 mg/kg. Premedication with corticosteroids or antihista- mine agents was not usually employed. Treatment courses comprising at least four doses of ABLC were considered evaluable for efficacy as well as toxicity. Treatment courses comprising less than four doses of ABLC were not considered evaluable for efficacy, but were evaluated for nephrotoxicity and systemic adverse reactions. Results Table 1 shows the baseline characteristics of the 64 patients. A total of 68 treatment courses was administered (four patients received two treatment courses each). The dose was 5 mg/kg during 67 courses, and 3 mg/kg during one. The drug was administered daily during 63 courses, and daily followed by every alternate day during five. Sixteen of the treatment courses were for microbiologically confirmed fungal infections (six in BMT recipients), and 52 for presumptive infections (28 in BMT recipients). Ten of the confirmed infections were due to Aspergillus spp. (microscopic/culture examination of sputum, bron- choalveolar lavage or sinus drainage), five due to Candida spp., and one due to Cryptococcus neoformans (Table 2). Hepatosplenic candidiasis was diagnosed in one patient on the basis of fever, neutrophilia, liver function abnormalities, and a characteristic appearance of the liver and the spleen on ultrasonography. However, a liver biopsy was not per- formed. This patient has been included in the presumptive infections group. The indications for anti-fungal therapy in patients with- Table 1 Patient characteristics Evaluation of response Number of patients 64 In patients with confirmed infections, complete response Sex was defined as clearance of all clinical, radiologic, and lab- Male 47 oratory evidence of disease. Partial response was defined Female 17 Median age (range) years 43 (17 68) as clearance of all clinical evidence and 50% clearance Diagnosis of the radiologic evidence of disease with or without Acute leukemia 35 microbiologic response. Failure was defined as persistence Chronic leukemia 7 or progression of signs or symptoms on therapy, or 50% Lymphoma 7 resolution of the radiologic findings with persistently posi- Multiple myeloma 9 Other 6 tive culture. In patients with presumed infections, lack of Number of courses of therapy 68 resolution of fever within 72 h or 50% clearance of lung Therapeutic intervention infiltrates was considered failure, 50% clearance of pul- Allogeneic bone marrow transplantation 24 monary infiltrates partial response, and complete resolution Autologous blood or marrow transplantation 10 Chemotherapy 34 of infiltrates or defervescence of fever complete response. Indication for anti-fungal therapy (%) Presumed fungal sepsis 52 (76) Nephrotoxicity Proven fungal infection 16 (24) Median neutrophil count at the initiation of /l (0 12.2) All patients were monitored with daily serum creatinine ABLC (range) measurements. The last creatinine level prior to starting

3 ABLC in immunocompromised patients Table 2 Confirmed fungal infections and response pneumonia. However, on retrospective review, the creatinine was found to have decreased after an initial rise while Total No. Evaluable Complete Partial Failure the patient was still on ABLC. The incidence of systemic response response adverse reactions was similar for chemotherapy and transplant patients. Aspergillus a 1 2 pneumonia Candidemia Aspergillus Response in confirmed infections sinusitis Nine complete responses were seen, including two patients with pulmonary aspergillosis in whom there was over 90% candidiasis clearance of the lung infiltrates with complete microbio- Cryptococcal meningitis logic resolution and resolution of the symptoms after 4 weeks of ABLC. In these two patients, complete clearance Total of the lung infiltrates on CT scanning occurred over the Oropharyngeal One patient with the characteristic clinical and radiologic picture of hepa- ensuing several weeks with intraconazole maintenance. tosplenic candidiasis but in whom microbiologic documentation was not Another patient with pulmonary aspergillosis in whom 7 available has been included in the presumptive therapy group. The syn- days of ABLC resulted in over 50% clearance of lung infildrome resolved completely on ABLC (see text). a trates and resolution of all clinical symptoms was classified Two patients were switched to prolonged oral itraconazole when all clinias a partial responder because nephrotoxicity necessitated cal symptoms had resolved, repeat sputum cultures were negative, and the radiologic changes had improved by 90 95%. discontinuation of ABLC. Four patients did not respond, and the overall response rate was 71% (10 of 14). Three of the four non-responders out confirmed fungal infections were: refractory fever died as a result of the fungal infection, and one died (n = 17), pulmonary infiltrates (n = 16), combination of the because of complications of the underlying disease. Four two (n = 17), oral mucositis (n = 1), and a characteristic cli- of six transplant recipients, including three of four with nicoradiologic picture of hepatosplenic candidiasis (n = 1). Aspergillus pneumonia responded. The indications for the use of ABLC (rather than conven- Table 2 shows response by site of infection and species. tional amphotericin or fluconazole/itraconazole) are shown All five patients with candidiasis responded, while the in Table 3. response rate in pulmonary aspergillosis was 71% (five of Fifty-three courses comprising at least four doses (4 58 seven). None of the evaluable patients with aspergillosis doses, median 10) were considered evaluable for efficacy, had received conventional amphotericin prior to receiving whereas 15 courses of one to three doses were considered ABLC, but all the patients with candidiasis had received non-evaluable for efficacy. Fourteen of the evaluable conventional amphotericin for 3 7 days prior to starting courses were for confirmed fungal infections (six in BMT ABLC. recipients), and 39 for presumed infections (23 in BMT Three of the seven evaluable patients with pulmonary recipients). aspergillosis had neutrophil counts of /l at the onset of therapy; partial response was seen in one, and the Non-evaluable courses drug was changed to liposomal amphotericin (AmBisome, NexStar, Boulder, CO, USA) in the other two due to Five patients developed systemic reactions to ABLC (fever nephrotoxicity after 4 and 7 days of therapy. All four with or without rigor and sweating) on the first day, and did not receive a second dose of the drug. None of these had received any premedication for ABLC administration. Eight patients (including one with Aspergillus pneumonia) died as a result of the underlying malignancy or complications of therapy after receiving one to three doses of ABLC. The drug was discontinued after two doses in a patient who was being treated empirically and whose gen- eral clinical condition improved significantly. The drug was discontinued due to an apparent elevation in serum creati- nine after three doses in another patient with Aspergillus patients with neutrophil counts of /l at the onset of treatment responded. Response in presumptive infections Nineteen of the 39 empiric courses resulted in complete response (resolution of fever or pulmonary infiltrates, or complete resolution of the clinicoradiologic picture of hepatosplenic candidiasis), and six in partial response (incomplete resolution of pulmonary infiltrates). Fourteen courses failed, resulting in a 63% response rate (24 of 38). 41 Table 3 Indication for ABLC Overall response The overall response rate was 66% (35 of 53), and was Failure of previous amphotericin or fluconazole therapy 28 comparable for chemotherapy (18 of 30; 60%), autograft Pre-existing renal dysfunction 16 (7 of 9; 77%) and allograft (10 of 14; 71%) recipients. Sev- Combination of the above factors 12 Intolerance of conventional amphotericin 6 enteen of 20 courses administered with initial neutrophil Previous invasive fungal infection 4 counts of more than /l resulted in a response com- Unknown 2 pared with 18 of 33 courses with neutrophils less than /l (P = 0.069, Fisher s exact test).

4 42 ABLC in immunocompromised patients Among the 40 patients who had failed prior anti-fungal therapy with fluconazole or conventional amphotericin, four were not evaluable, 12 did not respond, and 24 responded (67% of 36 evaluable patients). Table 4 Change in renal function as a result of ABLC therapy Group Change in creatinine % change ( mol/l) median and median and range range Systemic adverse reactions Chemotherapy Seven patients developed febrile reactions to ABLC. None 284 to to +212 Autograft of them had been premedicated for ABLC administration. 60 to to +141 The drug was stopped in five patients after a single dose, Allograft to to +270 but was continued uneventfully in two after premedication P (chemotherapy vs with steroids. Subsequently, all patients with a previous his- autograft) tory of reactions to conventional amphotericin received preallograft) P (autograft vs medication prior to ABLC administration and no reactions P (allograft vs were seen. Another patient developed tachycardia and chemotherapy tachypnea on the 8th day of therapy resulting in the discon- On cyclosporine tinuation of ABLC. 64 to to +270 Not on cyclosporine Renal function 284 to to +212 P On other nephrotoxic The baseline serum creatinine was mol/l (median drugs a 284 to to ). The change in serum creatinine over all the treatment Not on other courses was 284 to +277 mol/l (median +24), and over nephrotoxic drugs 217 to to +212 P the 53 evaluable courses, 110 to +277 mol/l (median Baseline creatinine ). The creatinine doubled over seven evaluable courses to to +270 of therapy, during five of which patients were receiving Baseline creatinine concomitant therapy with one to four other nephrotoxic to to +256 agents (cyclosporine, aminoglycosides, vancomycin, highdose P intravenous acyclovir, ganciclovir or foscarnet). Nephrotoxicity necessitated discontinuation of ABLC in a Including cyclosporine. only three patients. The creatinine increased by 50% of the baseline value in 10 of 24 allograft, two of 10 autograft, amphotericin in 178 patients with invasive aspergillosis. Of and seven of 34 chemotherapy recipients (P = 0.18, 2 test). the 111 patients considered evaluable for response, 40% of As Table 4 shows, there was a trend towards patients on the ABLC-treated patients responded compared with 23% cyclosporine and those with initially normal creatinine to of those treated with conventional amphotericin develop more elevated creatinine. (P = 0.002). Despite the high concentrations in the liver and the lungs, concentrations in the kidneys are comparable to conventional Discussion amphotericin and plasma levels are consistently lower, 16 which may explain the kidney-sparing effect of Our data suggest that ABLC is effective for the treatment ABLC. of microbiologically proven or clinically suspected fungal We found that although serum creatinine levels tended infections in immunocompromised patients with hematologic to increase, significant elevations (to more than double the malignancies including patients failing previous ther- baseline) were seen in less than 15% of the patients, most apy with fluconazole or conventional amphotericin. It may of whom were receiving other nephrotoxic agents as well. thus provide a useful first-line or salvage alternative to con- More importantly, renal toxicity resulted in discontinuation ventional amphotericin in the setting of BMT or chemotherapy. of the drug in only four of 68 cases (6%). As Table 4 As observed by others, 14,15 our data show that shows, no factor was consistently associated with elevation nephrotoxicity is not a major problem with ABLC. in creatinine levels. The pharmacokinetic properties of ABLC can explain its Because of their high cost, drugs such as ABLC are usually favorable toxicity and efficacy profile. The results of tissue used when the patient s clinical situation is poor, somefavorable distribution studies with ABLC in animals have shown a times when significant multi-organ dysfunction is present. dose-dependent increase in mean amphotericin concentrations In some of these cases, failure of therapy is due to death in the liver, spleen and lung. 16 The liver, spleen resulting from causes other than the fungal infection. It is and lungs are often the sites of deep fungal infections such possible that the use of ABLC earlier on, in the period of as candidiasis and aspergillosis. Targeted drug delivery to febrile neutropenia or after confirmed fungal infection, may these sites could potentially be beneficial for local eradication allow a higher proportion of patients to be treated success- of the infection. This could account for the high fully. response rates in invasive fungal infections, as seen with We conclude that ABLC is effective for presumed or pulmonary aspergillosis in our study. Hiemenz et al 17 confirmed fungal infections in high-risk immunocompromised reported a historical comparison of ABLC and conventional patients including those who have failed treatment

5 ABLC in immunocompromised patients with conventional amphotericin or fluconazole. Nephrotoxwith ing autologous bone marrow transplantation in adult patients icity requiring discontinuation of ABLC is rare. The drug acute leukemia. Acta Oncol 1996; 35: requires further evaluation in the management of fungal 10 Mehta J. Do variations in molecular structures affect the cliniinfections, especially at earlier stages of infection when the cal efficacy and safety of lipid-based amphotericin B prep- arations? Leuk Res 1997; 21: patient s general condition is better. 11 Franklin IM, Mehta J, Root T. The use of amphotericin B lipid complex/ablc (Ablcet) as an alternative to conventional References amphotericin B. J Antimicrob Chemother 1997; 39: Powles R, Milan S, Horton C et al. A prospective haematooncology 1 Goodrich JM, Reed EC, Mori M et al. Clinical features and patient database: structure and description. Bone analysis of risk factors for invasive candidal infection after Marrow Transplant 1996; 17 (Suppl. 1): S140. marrow transplantation. J Infect Dis 1991; 164: Powles R, Milan S, Horton C et al. A prospective haemato- 2 Uzun O, Anaissie EJ. Antifungal prophylaxis in patients with oncology patient database for use in routine care and clinical hematologic malignancies: a reappraisal. Blood 1995; 86: research. Bone Marrow Transplant 1996; 17 (Suppl. 1): S Walsh TJ, Hiemenz JW, Seibel N, Anaissie EJ. Amphotericin 3 Morrison VA, Haake RJ, Weisdorf DJ. The spectrum of non- B lipid complex in the treatment of 228 cases of invasive Candida fungal infections following bone marrow transplan- mycosis. Proceedings of the 34th Interscience Conference on tation. Medicine Baltimore 1993; 72: Antimicrobial Agents and Chemotherapy, October 1994, 4 Pizzo PA. Management of fever in patients with cancer and Orlando, p 247 (Abstr. M69). treatment-induced neutropenia. New Engl J Med 1993; 328: 15 Anaissie EJ, White M, Uzun O et al. Amphotericin B lipid complex (ABLC) vs amphotericin B (AMB) for treatment of 5 Bowler WA, Oldfield EC. New approaches to amphotericin B hematogenous and invasive candidiasis: a prospective, ranadministration. Infect Med 1992; 9: domized, multicenter trial. Proceedings of the 35th Interscience 6 Warnock DW. Amphotericin B: an introduction. J Antimicrob Conference on Antimicrobial Agents and Chemo- Chemother 1991; 28 (Suppl. B): therapy, September 1995, San Francisco, p 330 (Abstr. 7 Bodey GP, Anaissie EJ, Elting LS et al. Antifungal prophylaxis LM21). during remission induction therapy for acute leukemia 16 Janoff AS, Perkins WR, Saleton SL, Swenson CE. Ampho- fluconazole versus intravenous amphotericin B. Cancer 1994; tericin B lipid complex (ABLC): a molecular rationale for the 73: attenuation of amphotericin B related toxicities. J Liposome 8 Bearman SI, Appelbaum FR, Back A et al. Regimen-related Res 1993; 3: toxicity and early post-transplant survival in patients undergo- 17 Hiemenz JW, Lister J, Anaissie EJ et al. Emergency-use ing marrow transplantation for lymphoma. J Clin Oncol 1989; amphotericin B lipid complex (ABLC) in the treatment of 7: patients with aspergillosis: historical control comparison with 9 Glynne P, Powles R, Steele J et al. Renal dysfunction follow- amphotericin B. Blood 1995; 86: (Suppl. 1): 849a. 43