© Lippincott-Raven Publishers. Volume 11(12), 11 October 1997, p 1463–1471
Liposomal amphotericin B (AmBisome) compared with amphotericin B both followed by oral fluconazole in the treatment of AIDS-associated cryptococcal meningitis

Leenders, Alexander C.A.P.1,12; Reiss, Peter2,3; Portegies, Peter2; Clezy, Kate4; Hop, Wim C.J.5; Hoy, Jennifer6; Borleffs, Jan C.C.7; Allworth, Tony8; Kauffmann, Robert H.9; Jones, Philip10; Kroon, Frank P.11; Verbrugh, Henri A.1,5; de Marie, Simon1,5

1University Hospital Rotterdam, Rotterdam, Amsterdam, The Netherlands
2Academic Medical Centre, Amsterdam, The Netherlands
3National AIDS Therapy Evaluation Centre, Amsterdam, The Netherlands
4National Centre in HIV Epidemiology and Clinical Research, Sydney, Australia
5Erasmus University Rotterdam, Rotterdam Medical School, Rotterdam, The Netherlands
6Fairfield Infectious Disease Hospital, Fairfield, Australia
7Academic Hospital Utrecht, Utrecht, The Netherlands
8Royal Brisbane Hospital, Herston, Australia
9Hospital Leyenburg, The Hague, The Netherlands
10Prince Henry Hospital, Little Bay, Australia
11Academic Hospital Leiden, Leiden, The Netherlands.
12Requests for reprints to: Dr A.C.A.P. Leenders, Department of Bacteriology, University Hospital Rotterdam-Dijkzigt, Dr Molewaterplein 40, 3015 GD Rotterdam, The Netherlands.
Sponsorship: This study was supported by a grant from NeXstar Pharmaceuticals, Inc., San Dimas, California, USA.
Date of receipt: 3 February 1997; revised: 11 June 1997; accepted: 19 June 1997.




Objective: Amphotericin B deoxycholate initial therapy and fluconazole maintenance therapy is the treatment of choice for AIDS-associated cryptococcal meningitis. However, the administration of amphotericin B is associated with considerable toxicity. A potential strategy for reducing the toxicity and increasing the therapeutic index of amphotericin B is the use of lipid formulations of this drug.

Design and methods: HIV-infected patients with cryptococcal meningitis were randomized to treatment with either liposomal amphotericin B (AmBisome) 4 mg/kg daily or standard amphotericin B 0.7 mg/kg daily for 3 weeks, each followed by fluconazole 400 mg daily for 7 weeks. During the first 3 weeks, clinical efficacy was assessed daily. Mycological response was primarily evaluated by cerebrospinal fluid (CSF) cultures at days 7, 14, 21 and 70.

Results: Of the 28 evaluable patients, 15 were assigned to receive AmBisome and 13 to receive amphotericin B. Baseline characteristics were comparable. The time to and the rate of clinical response were the same in both arms. AmBisome therapy resulted in a CSF culture conversion within 7 days in six out of 15 patients versus one out of 12 amphotericin B-treated patients (P = 0.09), within 14 days in 10 out of 15 AmBisome patients versus one out of nine amphotericin B patients (P = 0.01), and within 21 days in 11 out of 15 AmBisome patients versus three out of eight amphotericin B patients (P = 0.19). When Kaplan-Meier estimates were used to compare time to CSF culture conversion, AmBisome was effective more (P < 0.05; median time between 7 and 14 days for AmBisome versus > 21 days for amphotericin B). AmBisome was significantly less nephrotoxic.

Conclusions: A 3-week course of 4 mg/kg AmBisome resulted in a significantly earlier CSF culture conversion than 0.7 mg/kg amphotericin B, had equal clinical efficacy and was significantly less nephrotoxic when used for the treatment of primary episodes of AIDS-associated cryptococcal meningitis.


Although the incidence is decreasing, cryptococcal meningitis is still the most common manifestation of systemic fungal infection in HIV-infected patients and remains associated with significant morbidity and mortality [1–3]. Amphotericin B deoxycholate (AMB) given for a period of 4–6 weeks has been considered to be the gold standard for the initial treatment of cryptococcal meningitis in patients with HIV infection. Initial treatment with the triazoles fluconazole and itraconazole is probably less effective [4–6]. Unfortunately, the administration of AMB may be associated with significant toxicity, mostly nephrotoxicity and acute infusion-related reactions. In order to reduce the risk of AMB-related toxicity, many clinicians have attempted to use shorter courses of AMB followed by oral triazoles in the treatment of AIDS-associated cryptococcal meningitis. A study examining initial treatment with 0.7 mg/kg daily AMB with or without flucytosine for 2 weeks, followed by fluconazole or itraconazole has recently been completed in the United States [7,8].

Another potential strategy for reducing AMB-associated toxicity is the use of liposomal and other lipid-based formulations containing AMB [9–12]. These formulations have shown a reduced toxicity and have permitted the use of higher doses of AMB, which potentially could lead to an increased therapeutic index [13]. One of such liposomal formulation is AmBisome (NeXstar Pharmaceuticals, Inc., San Dimas, California, USA), which has shown promising results in a Phase II study for the treatment of AIDS-associated cryptococcosis [14]. An open randomized trial was performed in HIV-infected patients with a primary episode of cryptococcal meningitis comparing a 3-week course of AMB 0.7 mg/kg daily with AmBisome 4.0 mg/kg daily, followed in both arms by treatment with fluconazole 400 mg daily for an additional 7 weeks.

Study population^

Hospitalized HIV-infected patients, aged 18 years or older, with a primary episode of cryptococcal meningitis were eligible for enrolment. Written informed consent had to be given by the patient or his legal guardian. Patients could be enrolled either on the basis of a positive India ink stain or the presence of cryptococcal antigen (CRAG) in the cerebrospinal fluid (CSF), while culture results were pending. Following enrolment, confirmation of cryptococcal meningitis was required either by a positive CSF culture for Cryptococcus neoformans, or a positive test for CRAG in CSF together with a positive blood culture for C. neoformans. Patients were excluded if they had already been treated for cryptococcal infection or if serum creatinine was above 250 µmol/l.

The study protocol was reviewed and approved by the ethics committees of all participating Dutch and Australian centres, and by the National AIDS Therapy Evaluation Centre in The Netherlands and the National Centre in HIV Epidemiology and Clinical Research in Australia.


Patients were randomized 1:1 to receive a 3-week monotherapeutic regimen of either AmBisome (4 mg/kg daily intravenously) or AMB (0.7 mg/kg daily intravenously). Randomization was performed by means of sealed envelopes, centrally by one Dutch and one Australian coordinating centre and was stratified by participating site. All patients completing these initial 3 weeks of study treatment while hospitalized, continued treatment with fluconazole (400 mg daily orally) for 7 weeks. Subsequently, fluconazole 200 mg was recommended to prevent relapses of cryptococcal infection; all patients were followed-up at least for 6 months to document relapses.

AmBisome was administered at full, without using a dose escalation schedule, and was infused over 45 min; no central intravenous catheter was required. For AMB the dose was escalated within 24 h to a full dose of 0.7 mg/kg daily, infused over 6 h. The use of a central intravenous catheter was recommended. Sodium supplementation prior to AMB infusion to prevent nephrotoxicity was recommended and administration of concomitant medication to prevent acute reactions was allowed.

Dosage adjustments were required when serum creatinine rose to >= 300% of a normal or >= 200% of an already elevated baseline value. Study medication was then to be discontinued for 2 days and could subsequently be reinstituted at half dose, when creatinine levels had fallen below 300 or 200% of the baseline values, respectively. If a rise in serum creatinine did not recur, doses were escalated to full. Study medication had to be discontinued if serum creatinine levels had not fallen below the above-mentioned levels after 5 days administration of reduced doses. Dose adjustment was not mandated in case of elevation of liver enzymes.

Clinical evaluation^

Patients were evaluated daily during the first 3 weeks of study treatment for vital signs, temperature, degree of headache (absent, mild, moderate, intractable), Glasgow Coma Scale score and the presence of meningeal symptoms. During this period an additional neurological and general physical examination, as well as determination of the Karnofsky score, was performed weekly. During the next 7 weeks of treatment with fluconazole, patients were similarly examined at least every 2 weeks as outpatients. The following criteria were used for clinical evaluation: (i) clinical response (normalization of all above mentioned pre-treatment signs, symptoms and scores); (ii) time to clinical response; (iii) clinical relapse (following clinical response, recurrence of any signs or symptoms of cryptococcal meningitis during the 10-week study period or the 6-month follow-up).


Blood urea nitrogen, serum creatinine and potassium were determined before treatment and twice weekly thereafter, and additional chemistry (including liver enzymes) and haematological determinations were determined weekly during the first 3 weeks and every 2 weeks until week 10.

Discontinuation of therapy^

Study drug treatment was prematurely discontinued in case of a sustained rise in serum creatinine levels (see above), in case of other serious adverse events judged to be study drug-related by the treating physician, or if requested by the patient or physician.

Factors predicting outcome^

The following factors were studied for being predictive for clinical outcome: altered metal state, positive extra neural cultures, positive blood cultures, lumbar opening pressure > 30 cmH2O, CSF white blood cell (WBC) count, CSF CRAG titre, and serum CRAG titre. All factors were included in the multivariate analysis.

Mycological evaluation^

CSF (5 ml), blood (10–20 ml) and material from all other sites suspected for cryptococcal infection (e.g., urine, sputum, lymph node) were obtained before starting study treatment. Lumbar punctures were repeated after 7, 14 and 21 days, and after 10 weeks. An additional puncture was performed at day 28 in those patients whose CSF culture at day 21 was the first to remain sterile. Any other previously positive cultures were repeated. In addition, CSF WBC counts, protein, glucose and CRAG titres were determined and lumbar opening pressure was recorded. Serum CRAG levels were measured weekly. Determination of CRAG titres was performed locally at each participating centre using several tests: Pastorex (Sanofi Diagnostics Pasteur, Marnes la Coquette, France), CALAS (Meridian Diagnostics, Cincinnati, Ohio, USA) and Immy (Immunomycologics, Norman, Oklahoma, USA). All tests were performed at local laboratories. The following criteria were used for mycological evaluation: (i) mycological response (CSF culture conversion, meaning the achievement of two consecutive negative cultures from the CSF); (ii) time to mycological response (time to the achievement of the first of two consecutive negative cultures from the CSF); (iii) persistence (after 10 weeks of study medication at least one culture from a specific site had remained positive); (iv) mycological relapse (following a CSF culture conversion, recurrence of a positive culture during the remainder of the study or the 6-month follow-up period).

Overall evaluation^

Overall response was defined as both clinical and mycological response at the completion of the 10-week treatment period.

Serum and CSF concentrations of AMB^

Trough and peak serum samples and CSF samples were obtained after 7, 14 and 21 days of therapy, respectively. The samples of CSF were obtained randomly just before or after infusion of the study drug. Levels of AMB were measured batchwise by high performance liquid chromatography at the University Hospital Rotterdam, as previously described [15]. This assay had a sensitivity of 0.1 µg AMB per millilitre of blood.

Study design and statistical analysis^

The study was primarily powered to show the superiority of AmBisome over standard AMB in achieving CSF culture conversion more rapidly. Assuming a CSF culture conversion rate of 90% for AmBisome versus 40% for AMB after 1 week of therapy, a sample size of 13 evaluable patients in each arm was expected to be sufficient to enable significant conclusions (two-sided P = 0.05, power = 80%) [5,14]. Time to clinical and mycological response was calculated according to the Kaplan–Meier method. Comparison of these curves was performed using the log-rank test. Percentages were compared using Fisher's exact test. Continuous variables were compared using Student's t test. Cox regression was used for multivariate analysis of factors prognostic for clinical outcome. Correlation coefficients given were determined using Spearman's rank correlation test. The change from baseline of logarithmically transformed creatinine values was compared using repeated measurements analysis of variance (ANOVA)[16]. A P value <= 0.05 was considered significant.

All analysis of clinical and mycological efficacy data were performed on an intention-to-treat basis. For analysis of toxicity, patients were evaluated up to the time of definite discontinuation of study drug.

Study population^

Between June 1992 and June 1995, 26 HIV-infected patients with suspected cryptococcal meningitis (25 on the basis of a positive India ink stain and one with a positive CRAG test of CSF), and four HIV-infected patients with culture-proven cryptococcal meningitis were enrolled in the study. Two patients were subsequently found to be ineligible. One comatose patient was randomized to AmBisome in error because no informed consent from a legal guardian had been obtained. The other patient, assigned to AMB, was withdrawn because of subsequent negative cultures. Twenty-eight patients were analysed further. Fifteen were assigned to AmBisome and 13 to standard AMB. At entry, clinical parameters and laboratory test results were comparable (Table 1), although patients in the AmBisome group had significantly higher serum CRAG titres (P = 0.007, Student's t test); CSF CRAG titres were comparable in both groups. No differences were found when results from the different centres were compared. There were no differences in pretreatment clinical characteristics or laboratory test results between the patients in Australia and The Netherlands. All 28 evaluable patients were included in the intention-to-treat analysis for efficacy.

Table 1. Clinical characteristics and results of laboratory tests in 28 HIV-infected patients with cryptococcal meningitis.

Premature discontinuation of treatment^

Five patients did not receive the full 3-week course of study treatment. Two AmBisome-treated patients were switched to other medication early (both at day 14), one because of persistently elevated CSF pressure, and the other because of mild somnolence, possibly related to disease progression or toxicity of the study medication. One patient treated with AMB was withdrawn prematurely (day 14) because of chorioretinitis, judged to be a sign of progression of cryptococcal infection. Two other patients treated with AMB were withdrawn because of laboratory toxicity (see below). Two of the 23 patients who did complete 3 weeks of study medication (both randomized to AmBisome), were not switched to fluconazole because of persistently positive CSF cultures with C. neoformans, but initially received the AMB-flucytosine combination for 3 and 4 weeks, respectively. One patient assigned to AMB was lost to follow-up 8 weeks after enrolment.

Clinical outcome^

Clinical response rates after the first 3 weeks of treatment were 12 out of 15 [80%; 95% confidence interval (CI), 52–96] and 11 out of 13 (86%; 95% CI, 55–98) in the AmBisome and AMB-treated groups, respectively (P = 1.0). The median time to clinical response was 15 days in both arms. No patients died during this period. During the subsequent 7 treatment weeks one patient treated with AmBisome (at week 6), and two treated with AMB (at weeks 7 and 10, respectively) died. Clinical response rates at week 10 were 13 out of 15 patients (87%) assigned to AmBisome and 10 out of 12 (83%) patients assigned to AMB. The one patient assigned to AmBisome who had failed to respond clinically to therapy, achieved a clinical response at week 14. No clinical relapses were observed during the 10-week study period. No proven clinical relapses occurred during the 6-month or further follow-up. One patient who had been treated with AmBisome died, after having been admitted to hospital with a reduced level of consciousness, for which no definite explanation was found. During the 6-month follow-up two additional patients died (one in each arm), although both deaths were not as a result of cryptococcal disease.

Factors predicting outcome^

Both in univariate and multivariate analysis two factors were shown to be significantly associated with a longer time to clinical response: CSF WBC count of <20 × 106/l and raised lumbar opening pressure > 30 cmH2O (Table 2).

Table 2. Factors significantly associated with a longer time to clinical response in AIDS-associated cryptococcal meningitis.

Mycological outcome^

AmBisome therapy resulted in a CSF culture conversion within 7 days in six out of 15 patients versus one out of 12 AMB-treated patients (P = 0.09). Significantly more patients treated with AmBisome had a CSF culture conversion within 14 days (10 out of 15 patients) when compared with patients treated with AMB (one out of nine patients; P = 0.01). Within 21 days 11 out of 15 patients treated with AmBisome versus three out of eight patients treated with AMB had responded mycologically (P = 0.18). Five patients assigned to AMB did not undergo all scheduled lumbar punctures, because of abnormalities in haemostasis (one patient), refusal (two patients), and unintentional omission (two patients). These patients were not scored as being failures, but were censored in the analysis, explaining the different denominators in weeks 1, 2 and 3. When Kaplan–Meier estimates were used to compare time to CSF culture conversion, AmBisome was significantly more effective than AMB (P < 0.05; Fig. 1). The median time to CSF culture conversion was between 7 and 14 days for AmBisome versus > 21 days for AMB. One patient treated with AmBisome with negative CSF cultures at weeks 1 and 2 had a mycological relapse at week 3, but was nevertheless successfully switched to fluconazole according to the study protocol. In the patients who underwent lumbar punctures after 10 weeks of treatment, all 11 patients initially treated with AmBisome and all eight patients treated with AMB had negative cultures. No mycological relapses were noted during the 6-month and further follow-up period.

Fig. 1. Kaplan–Meier estimates of the proportion of patients who had positive cerebrospinal fluid cultures during the first 3 weeks of treatment, according to treatment group.

Blood cultures for cryptococcus became negative after a median of 7 days in both treatment groups. All other cultures taken from extraneural sites that had been positive at enrolment (Table 1) became negative within 21 days.

CRAG titres in CSF showed a consistent decrease throughout the first 3 weeks of treatment with no differences between the two treatment groups. Serum CRAG titres in both groups remained at about the same level throughout this period. Lumbar opening pressures showed a consistent decrease in both arms throughout the first 3 weeks of treatment. No relationship between clinical or mycological response and the course of lumbar opening pressures was found.

A significant correlation was found between the time to CSF culture conversion and the time to clinical response (r = 0.63; P < 0.001; Fig. 2).

Fig. 2. Correlation between the time to mycological response and the time to clinical response. The times to mycological and clinical response show a significant correlation in AIDS patients with crytococcal meningitis treated with AmBisome or amphotericin B (P = 0.0009 by Spearman rank correlation test).


Both treatment regimens were generally well tolerated (Table 3). Three patients experienced clinical adverse events possibly related to AmBisome. Mild somnolence was observed in one patient after 2 weeks of therapy. Two patients developed acute reactions during the first infusion of AmBisome. Tachycardia, hypotension, facial flushing and a tickling cough was observed in one patient, and tachycardia, shortness of breath and fever in the other. In both patients these reactions disappeared when infusion was interrupted and did not recur when the infusion continued at a lower infusion rate. Concerning nephrotoxicity, when increases from baseline of serum creatinine levels at the various timepoints were analysed with repeated measurements ANOVA, it was found that this increase was on average a factor of 1.37 (P = 0.003) greater in the AMB-treated patients. Three patients treated with AmBisome and four patients treated with AMB experienced hypokalaemia, but none of these patients had to discontinue therapy for this reason. The number of patients with major changes in biochemical parameters are shown in Table 3, and the median percentage increases are shown in Table 4. Only the median percentage increase in bilirubin was significantly smaller in patients treated with AmBisome (Table 4). Two patients treated with AMB were withdrawn because of laboratory toxicity; withdrawal of one patient (at day 10) was protocol-mandated by an increase in serum creatinine, and of the other (at day 14) by an increase in liver enzymes.

Table 3. Number of patients with adverse events during treatment with amphotericin B or AmBisome.

Table 4. Changes in haemoglobin, aspartate aminotransferase (AST), alanine aminotransferase (ALT), alkaline phosphatase and bilirubin after 3 weeks of therapy.

Serum and CSF concentrations of AMB^

Median trough levels (17.6 µg/ml; range, 2.4–65.6) and peak levels (44.4 µg/ml; range, 20.6–100) of serum AMB in patients treated with AmBisome were higher than those in patients treated with standard AMB (0.9 µg/ml; range, 0.5–2.1; and 1.8 µg/ml; range, 1.2–5.2, respectively; both P < 0.001). However, CSF AMB levels could not be detected in any of the CSF samples obtained just before or after infusion with either AmBisome or standard AMB.


Our study is the first randomized comparative trial to show superior antifungal activity of liposomal AMB (AmBisome) over standard AMB in the treatment of a specific fungal infection. Liposomal or lipid-based formulations of AMB had already been shown to be less toxic than standard AMB and might have enhanced antifungal efficacy when high dosages are used. Indeed, an improved therapeutic index has been reported in several animal models, including models for cryptococcosis [13,17–20]. In patients with AIDS-associated cryptococcal meningitis, a randomized study comparing up to 5 mg/kg Abelcet with 0.7–1.2 mg/kg AMB, no evidence of improved activity was found [21]. However, on the basis of differences in pharmacological properties, the results obtained for one of the lipid-based formulations of AMB cannot be applied to other formulations, unless comparative studies are available [10,11].

In our study, two-thirds of patients given AmBisome, had CSF culture conversion within 14 days, and more than one-half of these patients within 7 days of treatment. In comparison, the median time to CSF culture conversion was more than 21 days for patients assigned to AMB, with only one patient having achieved CSF culture conversion within 7 days. Unfortunately, several patients in the AMB group did no undergo lumbar puncture at days 14 and 21. However, we felt that this probably did not influence the overall outcome of the study, because time to clinical response of these patients was the same as in patients who underwent all lumbar punctures (median time, 14 days; range, 10–19 days). Furthermore, these patients were not scored as failures, which should have been appropriate in an intention-to-treat analysis, but were censored, so they could not negatively influence the results of the AMB group. Our findings are remarkable when considering that the median time to the first negative CSF culture was as long as 42 days for patients treated with AMB in the largest trial of AIDS-associated cryptococcal meningitis to date, although only moderate doses were used [4]. The short median CSF culture conversion time following AmBisome treatment could not be explained by a predominance of less severe cases. The severity of illness in our patient population as indicated by factors known to be predictive of worse outcome (altered mental state, lumbar opening pressure > 30 cmH2O, CSF WBC count < 20 × 106/l, and CSF CRAG titre >1:1024) was comparable to that in other studies [4,5,22]. The relatively low number of AMB-treated patients who had CSF culture conversion at day 14 when compared with the most recent AIDS Clinical Trials Group trial might be explained by the fact that we also included patients with severe infections [7]. We were unable to explain the differences that were found in the serum CRAG titres; they were not correlated to positive blood cultures, nor to the CRAG titre in the CSF. However, high CRAG titres are only known to be predictive for a worse outcome in non-HIV-infected patients and not in AIDS patients with cryptococcal meningitis, so we do not consider this to be a confounding factor [23].

Our strategy of initiating therapy with intravenous AMB, and consolidating the treatment with an oral triazole antifungal agent was very similar to the strategy used in a large, recently completed, trial of AIDS-associated cryptococcal meningitis [7,8]. The high rates of clinical and mycological responses in our study after 10 weeks of treatment, as well as the low mortality during the initial period of intravenous therapy, is comparable to the overall results obtained for patients in this study who had received fluconazole as consolidation therapy. We are not aware of reports concerning the influence of the addition of flucytosine to AmBisome, but this might be an option to further increase efficacy. We have been unable to show a superior clinical response for patients receiving AmBisome, but this may well have been due to the limited number of patients in our study.

As has been shown for Abelcet (The Liposome Company, Princeton, New Jersey, USA), AmBisome was significantly less nephrotoxic than AMB in this group of patients [21]. Although, as in patients treated with AMB, hypokalaemia was frequently observed in the patients receiving AmBisome, it did not lead to premature discontinuation of treatment in any of these patients. This study also showed that high doses of AmBisome can be given to HIV-infected patients without severe adverse events. More than one-half of the patients treated with AmBisome did not experience any side-effects during the 3-week treatment. Coker et al. [14] showed that changes in liver enzymes occurred during treatment with AmBisome. We showed that there was no significant difference in changes in liver enzymes between patients treated with AmBisome and AMB. The minor but significant difference between changes in bilirubin may have been caused by the interference of lipids in the determination of the levels of bilirubin [24]. The only other relevant toxicities were observed in two patients during the first infusion of AmBisome. However, restarting drug administration at a lower infusion rate was uncomplicated. These events are quite similar to the adverse events described by Arning et al. [25]. We therefore suggest that patients receiving AmBisome for the first time should be closely monitored during this infusion. When acute infusion-related reactions occur, infusion should be stopped, but can be restarted at a lower rate.

Although serum peak and trough levels of AMB in patients receiving AmBisome were much higher than in those receiving standard AMB, the relevance of this finding is unclear because the assay used cannot distinguish between lipid-bound and free AMB [10]. Since no levels of AMB in CSF were detectable, this agent appears to be bound to the meninges. This is consistent with the observation in animals treated with AmBisome [26]. The lack of correlation between the presence of AMB in the CSF and the outcome of cryptococcal meningitis following treatment with standard AMB and other AMB-containing formulations, is understandable if one considers this to be an infection of the brain and meninges. Large numbers of cryptococci can indeed be seen in tissue sections of brain and meninges from patients with cryptococcal meningitis [27]. A similar lack of correlation has been reported for itraconazole [28].

A frequently mentioned significant drawback of AmBisome and other lipid-based formulations of AMB is their high cost [11,29]. However, as we have demonstrated, a significant correlation between the time to CSF culture conversion and the time to clinical response, with CSF culture conversion being accomplished in a substantial proportion of AmBisome-treated patients within as little as 7 days, short treatment with AmBisome (e.g., for 1 week), followed by oral fluconazole may be appropriate. Reduction of the duration of hospitalization, low toxicity rates, the shortened duration of infusion and the possibility of administering AmBisome through peripheral veins could result in a cost–effective regimen. We suggest that comparative studies should address these possibilities and should include a formal cost-benefit analysis for the use of AmBisome in the treatment of AIDS-associated cryptococcal meningitis.

Our results indicate that a high-dose AmBisome regimen combines a reduced toxicity with an increase in antifungal activity compared with standard AMB in HIV-infected patients with cryptococcal meningitis, when used in a dose as high as 4 mg/kg daily for 3 weeks. It was therefore concluded that the AmBisome regimen used in this study is probably the most effective single agent regimen for the treatment of AIDS-associated cryptococcal meningitis.


The authors are indebted to K. Kerstens for assistance with the data collection.


1. Dismukes WE: Cryptococcal meningitis in patients with AIDS. J Infect Dis 1988, 157:624–628. [Context Link]

2. Chuck SL, Sande MA: Infections with Cryptococcus neoformans in the acquired immunodeficiency syndrome. N Engl J Med 1989, 321:794–799. Bibliographic Links [Context Link]

3. Zuger A, Louie E, Holzman RS, Simberkoff MS, Rahal JJ: Cryptococcal disease in patients with the acquired immunodeficiency syndrome: diagnostic features and outcome of treatment. Ann Intern Med 1986, 104:234–240. Bibliographic Links [Context Link]

4. Saag MS, Powderly WG, Cloud GA, et al.: Comparison of amphotericin B with fluconazole in the treatment of acute AIDS-associated cryptococcal meningitis. N Engl J Med 1992, 326:83–89. Bibliographic Links [Context Link]

5. Larsen RA, Leal MA, Chan LS: Fluconazole compared with amphotericin B plus flucytosine for cryptococcal meningitis in AIDS. Ann Intern Med 1990, 113:183–187. Bibliographic Links [Context Link]

6. Gans de J, Portegies P, Tiessens G, et al.: Itraconazole compared with amphotericin B plus flucytosine in AIDS patients with cryptococcal meningitis. AIDS 1992, 6:185–190. [Context Link]

7. Horst van der CM, Saag MS, Cloud GA, et al.: Treatment of cryptococcal meningitis associated with the acquired immunodeficiency syndrome. N Engl J Med 1997, 337:15–21. [Context Link]

8. Saag M, Horst van der C, Cloud G, et al.: Randomized double blind comparison of amphotericin B plus flucytosine to amphotericin B alone (step 1) followed by a comparison of fluconazole to itraconazole (step 2) in the treatment of acute cryptococcal meningitis in patients with AIDS: part 2. 35th Interscience Conference on Antimicrobial Agents and Chemotherapy. San Francisco, September 1995 [abstract 1217]. [Context Link]

9. Janknegt R, Marie de S, Bakker-Woudenberg IAJM, Crommelin JA: Liposomal and lipid formulations of amphotericin B: clinical pharmacokinetics. Clin Pharmacokinet 1992, 23:279–291. [Context Link]

10. Marie de S, Janknegt R, Bakker-Woudenberg IAJM: Clinical use of liposomal and lipid-complexed amphotericin B. J Antimicrob Chemother 1994, 33:907–916. [Context Link]

11. Graybill JR: Lipid formulations of amphotericin B: does the emperor need new clothes? Ann Intern Med 1996, 124:921–923. Bibliographic Links [Context Link]

12. Ng TTC, Denning DW: Liposomal amphotericin B (AmBisome) therapy in invasive fungal infections. Arch Intern Med 1995, 155:1093–1098. [Context Link]

13. Leenders ACAP, de Marie S: The use of lipid formulations of amphotericin B for systemic fungal infections. Leukemia 1996, 10:1570–1575. [Context Link]

14. Coker RJ, Viviani M, Gazzard BG, et al.: Treatment of cryptococcosis with liposomal amphotericin B (AmBisome) in 23 patients with AIDS. AIDS 1993, 7:829–835. Bibliographic Links [Context Link]

15. Etten van E, Heuvel van den-de Groot C, Bakker-Woudenberg IAJM: Efficacies of amphotericin B-desoxycholate (Fungizone), liposomal amphotericin B (AmBisome) and fluconazole in the treatment of systemic candidosis in immunocompetent and leukopenic mice. J Antimicrob Chemother 1993, 32:723–739. [Context Link]

16. Dixon WJ (Ed): BMDP Statistical Software Manual, BMDP-5V. Oxford: University of California Press; 1990:1207–1244. [Context Link]

17. Adler-Moore JP, Chiang SM, Satorius A, et al.: Treatment of murine candidosis and cryptococcosis with unilamellar liposomal amphotericin B formulation. J Antimicrob Chemother 1991, 28 (suppl B):S63-S71. [Context Link]

18. Hostetler JS, Clemons KV, Hanson LH, Stevens DA: Efficacy and safety of amphotericin B colloidal dispersion compared with those of amphotericin B deoxycholate suspension for treatment of disseminated murine cryptococcosis. Antimicrob Agents Chemother 1992, 36:2656–2660. Bibliographic Links [Context Link]

19. Perfect JR, Wright KA: Amphotericin B lipid complex in the treatment of experimental cryptococcal meningitis and disseminated candidosis. J Antimicrob Chemother 1994, 33:73–81. Bibliographic Links [Context Link]

20. Albert MM, Stahl-Carroll TL, Luther MF, Graybill JR: Comparison of liposomal amphotericin B to amphotericin B for treatment of murine cryptococcal meningitis. J Mycol Med 1995, 5:1–6. Bibliographic Links [Context Link]

21. Sharkey PK, Graybill JR, Johnson ES, et al.: Amphotericin B lipid complex compared with amphotericin B in the treatment of cryptococcal meningitis in patients with AIDS. Clin Infect Dis 1996, 22:315–321. Bibliographic Links [Context Link]

22. Leenders AC, Ende van der ME, van der Ree TC, de Marie S: Cryptococcal meningitis in HIV-infected patients [letter]. AIDS 1994, 8:1741–1743. [Context Link]

23. Diamond RD, Bennett JE: Prognostic factors in cryptococcal meningitis. Ann Intern Med 1974, 80:176–181. Bibliographic Links [Context Link]

24. Pekelharing JM (Ed): Handboek Klinisch-Chemische Tests. Utrecht: Bunge; 1995:98–100. [Context Link]

25. Arning M, Heer-Sonderhoff AH, Wehmeier A, Schneider W: Pulmonary toxicity during infusion of liposomal amphotericin B in two patients with acute leukemia. Eur J Clin Microbiol Infect Dis 1995, 14:41–43. Bibliographic Links [Context Link]

26. Proffitt RT, Satorius A, Chiang SM, Sullivan L, Adler-Moore J: Pharmacology and toxicology of a liposomal formulation of amphotericin B (AmBisome) in rodents. J Antimicrob Chemother 1991, 28 (suppl B):S49-S61. [Context Link]

27. Chandler FW, Kaplan W, Ajello L: A Colour Atlas and Textbook of the Histopathology of Mycotic Diseases. Weert: Wolfe Medical Publications; 1989:190–198. [Context Link]

28. Denning DW, Tucker RM, Hanson LH, Hamilton JR, Stevens DA: Itraconazole therapy for cryptococcal meningitis and cryptococcosis. Arch Intern Med 1989, 149:2301–2308. Bibliographic Links [Context Link]

29. Tollemar J, Ringden O: Lipid formulations of amphotericin B. Less toxicity but at what economic cost?. Drug Safety 1995, 13:207–218. Bibliographic Links [Context Link]

Keywords: Liposomal amphotericin B; amphotericin B; AmBisome; cryptococcal meningitis; Cryptococcus neoformans; HIV; randomized comparative trial

Accession Number: 00002030-199712000-00010