Single-Dose Liposomal Amphotericin B Treatment for Cryptococcal Meningitis

Background Cryptococcal meningitis is a leading cause of HIV-related mortality in sub-Saharan Africa. Based on phase-II data, we performed a phase-III randomized controlled non-inferiority trial to determine the efficacy of a single high-dose liposomal amphotericin B based treatment regimen. Methods HIV-positive adults with cryptococcal meningitis in Botswana, Malawi, South Africa, Uganda and Zimbabwe were randomized 1:1 to induction therapy of either (i) single, high-dose liposomal amphotericin B 10mg/kg given with 14 days of flucytosine 100mg/kg/day and fluconazole 1200mg/day (AmBisome group), or (ii) the current WHO recommended treatment of 7 daily doses of amphotericin B deoxycholate (1mg/kg/day) plus flucytosine (100mg/kg/day), followed by 7 days of fluconazole 1200mg/day (control group). The primary endpoint was all-cause mortality at 10 weeks with the trial powered to show non-inferiority at a 10% margin. Results We randomized 844 participants. None were lost-to-follow-up. In intention-to-treat analysis, 10-week mortality was 24.8% (101 of 407; 95% confidence interval [CI] 20.7-29.3%) in the AmBisome group and 28.7% (117 of 407; 95% CI 24.4-33.4%) in controls. The absolute difference in mortality was -3.9%, with an upper 1-sided 95% confidence interval of 1.2%. Fungal clearance from cerebrospinal fluid was -0.40 log10 CFU/ml/day in the AmBisome group and -0.42 log10 CFU/ml/day in the control group. Fewer participants experienced grade 3 or 4 adverse events in the AmBisome group than the control group (50.0% vs. 62.3%). Conclusions Single dose liposomal amphotericin B (10mg/kg) on a backbone of flucytosine and fluconazole was non-inferior to the current WHO recommended standard of care for HIV-associated cryptococcal meningitis and associated with fewer adverse events. (Trial registration number: ISRCTN72509687.)

ryptococcal meningitis is the most frequent cause of adult meningitis in areas with a high prevalence of human immunodeficiency virus (HIV) 1,2 and is the second leading cause of HIV-related death worldwide, with the majority of deaths occurring in sub-Saharan Africa. 3 Despite widened access to antiretroviral therapy, there is a persistent burden of advanced HIV disease in the sub-Saharan African region, [4][5][6] and the number of cryptococcal meningitis cases remains high. 6,7 Poor outcomes with conventional antifungal treatment regimens are a key driver of the high mortality from cryptococcal meningitis, with a high incidence of toxic effects with the commonly used 2-week amphotericin B deoxycholate-based regimens and poor efficacy with fluconazole monotherapy, which has been associated with a 10-week mortality in excess of 50%. 8,9 In 2018, after the publication of the Advancing Cryptococcal Meningitis Treatment for Africa (ACTA) trial, 10 the World Health Organization (WHO) updated international guidelines to recommend induction therapy with the less toxic and more efficacious 1-week regimen of amphotericin B deoxycholate and flucytosine in resource-limited settings. 11 However, even 1 week of treatment with amphotericin B deoxycholate is associated with anemia, kidney impairment, and electrolyte abnormalities, 8 and administering and monitoring intravenous amphotericin for 7 days poses logistic challenges in many clinical settings.
Liposomal amphotericin B is potentially well suited for use in short-course induction treatments of cryptococcal meningitis because it can be given at higher doses owing to a lower incidence of drug-induced toxic effects, [12][13][14] has a long tissue half-life, 12,13,[15][16][17] and effectively penetrates into brain tissue. 12,18,19 The concept of a single high-dose intravenous infusion of liposomal amphotericin B has been established in the treatment of visceral leishmaniasis, 20 and pharmacokinetic data from animal models and humans indicate that increasing the dose of liposomal amphotericin B from the currently recommended dose of 3 to 4 mg per kilogram of body weight may lead to improved outcomes in patients with cryptococcal meningitis and that shortcourse regimens may be as effective as daily therapy. 15,16,21,22 In a phase 2 clinical trial, investigators assessed the efficacy of a short-course regimen with a single high dose of liposomal amphotericin B, two high doses of liposomal amphotericin B given on days 1 and 3, or three high doses of liposomal amphotericin B given on days 1, 3, and 7, as compared with the control regimen of 14 daily doses of 3 mg per kilogram of liposomal amphotericin B (all four regimens included 14 days of high-dose fluconazole); they showed that the rate of fungal clearance from the cerebrospinal fluid with any of the three short-course, high-dose regimens was noninferior to that in the control group. 23 Maximal fungicidal activity was attained with a single 10-mg-per-kilogram dose of liposomal amphotericin B, and there was no evidence that additional doses led to greater benefit -findings that are in keeping with the data obtained from animal models. 22,24 No safety concerns have been identified with the use of high-dose liposomal amphotericin B, which has a better adverse-effect profile than that observed with amphotericin B deoxycholate in previous trials. 8,23 On the basis of the findings of the phase 2 trial 23 and the data from a phase 3 trial that showed a role for flucytosine in the induction treatment of cryptococcal meningitis, 10 we conducted an open-label, phase 3, randomized, controlled, noninferiority trial (the Ambition trial) to test a single high dose (10 mg per kilogram) of liposomal amphotericin B given with oral flucytosine and fluconazole for 2 weeks 10 against the WHO-recommended first-line induction treatment with 1 week of amphotericin B deoxycholate plus flucytosine followed by 1 week of highdose fluconazole.

Trial Design and Oversight
The trial design has been described previously, 25 and the details are provided in the trial protocol, available with the full text of this article at NEJM.org. The protocol was approved by the London School of Hygiene and Tropical Medicine Research Ethics Committee and by the relevant ethics committees and national regulatory agencies overseeing the trial sites. All the participants provided written informed consent. If a participant had abnormal mental status, written informed consent was obtained from the next of kin; if a participant recovered the capacity to provide consent, written informed consent was obtained from that participant. An independent data-monitoring committee oversaw the trial and reviewed the trial data regularly. The trial funders, suppliers, and drug manufacturers had no role in the design of the trial; in the collection, analysis, or interpretation of the data; or in the preparation of the manuscript or the decision to submit it for publication. Liposomal amphotericin B was donated by Gilead Sciences; amphotericin B deoxycholate was purchased from Bristol Myers Squibb; flucytosine was purchased from Mylan; and fluconazole was purchased from Cipla-Medopharm. At sites where the Pfizer Diflucan Partnership Program was operational, fluconazole donated by Pfizer was used if available. The authors vouch for the accuracy and completeness of the data and for the adherence of the trial to the protocol.

Trial Participants
HIV-positive adults (≥18 years of age) who had a first episode of cryptococcal meningitis, as diagnosed on the basis of a positive India ink stain or cryptococcal antigen test (CrAg lateral flow assay, IMMY) of a cerebrospinal fluid sample, were recruited from eight hospitals: Princess Marina Hospital, Gaborone, Botswana; Queen Elizabeth Central Hospital, Blantyre, and Kamuzu Central Hospital, Lilongwe, Malawi; Mitchells Plain Hospital and Khayelitsha Hospital, Cape Town, South Africa; Kiruddu National Referral Hospital, Kampala, and Mbarara Regional Referral Hospital, Mbarara, Uganda; and Parirenyatwa Central Hospital, Harare, Zimbabwe. Participants were excluded if they had received more than two doses of either amphotericin (at any dose) or fluconazole (at a dose of ≥800 mg) before screening; declined to consent or, if they had impaired capacity to consent, had no legal representative to consent on their behalf; were pregnant or breast-feeding; were taking contraindicated concomitant drugs; or had had any previous adverse reaction to a trial drug. Lateexclusion criteria, which were put in place to enable the rapid enrollment of critically ill participants pending baseline blood test results, were an alanine aminotransferase level greater than 5 times the upper limit of the normal range (>200 IU per liter), a polymorphonuclear leukocyte count of less than 500 per cubic millimeter, or a platelet count of less than 50,000 per cubic millimeter.

Interventions and Randomization
Participants underwent randomization individually and were assigned in a 1:1 ratio to receive the experimental regimen that included a single dose (10 mg per kilogram of body weight) of liposomal amphotericin B (AmBisome, Gilead Sciences) plus 14 days of flucytosine (100 mg per kilogram per day) and fluconazole (1200 mg per day) 26 or the current WHO-recommended regimen, which includes amphotericin B deoxycholate (1 mg per kilogram per day) plus flucytosine (100 mg per kilogram per day) for 7 days, followed by fluconazole (1200 mg per day) on days 8 through 14 (the control group). Randomization was performed with the use of a computer-generated randomization list with block sizes of four and six, stratified according to site. Randomization was performed electronically with a bespoke electronic data-capture tool in which the randomassignment sequence was concealed from all trial investigators involved in participant recruitment. The treatment-group assignments were provided to the recruiting teams after consent had been obtained and the participant enrolled. The trial medications were administered on an open-label basis.
All the participants were treated in-hospital for a minimum of 7 days. The single 10-mg-perkilogram dose of liposomal amphotericin B was suspended in 1 liter of 5% dextrose and administered over the course of 2 hours, and the 1-mgper-kilogram doses of amphotericin B deoxycholate were dissolved in 1 liter of 5% dextrose and administered over the course of 4 hours. Participants received 1 liter of intravenous normal saline before any amphotericin dose, plus at least 1 additional liter of intravenous fluid (5% dextrose or normal saline) on each day of amphotericin therapy. Potassium and magnesium supplements were given on each day that the participants received amphotericin and then for 2 additional days. Oral medications were administered through a nasogastric tube if participants were unable to swallow.
The results of laboratory blood tests were monitored regularly during the first 2 weeks and again at week 4. The monitoring schedule is provided in Table S1 in the Supplementary Appendix, available at NEJM.org. Lumbar punctures for quantitative cryptococcal cultures were performed at the time of diagnosis and on days 7 and 14. Participants with increased intracranial pressure received additional daily therapeutic lumbar punctures until the pressure was controlled at less than 20 cm of water.
Participants were followed at outpatient clinics for 10 weeks and were contacted by telephone n engl j med 386;12 nejm.org March 24, 2022 at week 16. If a participant missed a clinic appointment, follow-up was performed by the trial teams either by telephone or in person. After the 2-week induction period, all the participants received fluconazole at a dose of 800 mg per day for 8 weeks and then at a dose of 200 mg per day thereafter. Antiretroviral therapy was initiated, reinitiated, or switched to a new antiretroviral therapy with a different agent during weeks 4 to 6 and was chosen in accordance with national guidelines.

End Points
The primary end point was death from any cause at 10 weeks after randomization. As prespecified in the statistical analysis plan, the primary end point was tested for superiority after noninferiority was established. Secondary end points were death from any cause at 2 weeks, 4 weeks, and 16 weeks; overall mortality in a time-to-event analysis; the rate of fungal clearance from the cerebrospinal fluid per day over the course of 14 days of induction therapy; the percentage of participants in each trial group with clinical or laboratory-defined adverse events of grade 3 or 4, as determined according to the criteria of the Division of AIDS 27 ; and the median absolute or percentage change from baseline in laboratory values.

Statistical Analysis
Assuming 35% mortality at 10 weeks in both treatment groups, we calculated that a sample size of 390 per group (780 in total) would provide the trial with 90% power to show noninferiority of a single high dose of liposomal amphotericin B given with flucytosine and fluconazole to the current WHO recommended standard of care, with a specified noninferiority margin of 10 percentage points (the upper boundary of the one-sided 95% confidence interval of the absolute difference in mortality). The primary analysis was performed in the intention-to-treat population, which included all the participants who had undergone randomization and had not met any late-exclusion criteria. A generalized linear model with a binomial distribution was used to calculate the differences in mortality.
We performed two sensitivity analyses. First, a per-protocol analysis was performed in which participants were excluded if they had missed more than 1 day of any single treatment in the first 2 weeks or had missed more than 2 weeks of fluconazole consolidation treatment between weeks 2 and 10. Second, we performed analyses that adjusted for the prespecified covariates of trial site, age, sex, baseline Glasgow Coma Scale score, CD4+ cell count, cryptococcal colonyforming units (CFU) per milliliter of cerebrospinal fluid, antiretroviral therapy status, hemoglobin level, and cerebrospinal fluid opening pressure. In the superiority, secondary end-point, and sensitivity analyses, no adjustments were made for multiple comparisons. Analysis of log-transformed longitudinal fungal counts in the cerebrospinal fluid was performed with the use of a linear mixed-effects model, in which undetectable measurements were left-censored (i.e., sterile cultures from day 7 onward were excluded if the values lessened the slope, because sterility would have been achieved before lumbar puncture on that day and use of these values would have therefore led to an underestimation of the true slope). 28 Adverse events were evaluated in the safety population, which included all the participants who had received one or more doses of a trial medication. Analyses were conducted with the use of SAS statistical software, version 9.4 (SAS Institute). The full statistical analysis plan is provided in the protocol.

Trial Population
From January 2018 through February 2021, a total of 844 participants underwent randomization ( Fig. 1). Of these participants, 30 were excluded -24 met the prespecified late-exclusion criteria (13 had a low platelet count, 5 had a low neutrophil count, 2 had an increased alanine aminotransferase level, 3 had a low platelet count and a low neutrophil count, and 1 had a low platelet count and an increased alanine aminotransferase level [ Table S2]), 5 did not have cryptococcal meningitis, and 1 was HIV-negative, which left 814 participants (407 in each treatment group) in the intention-to-treat population. None were lost to follow-up. An additional 30 participants were excluded from the per-protocol population (20 had missed more than 1 day of treatment in the first 2 weeks, 6 had received incorrect treatment, and 4 had missed more than 2 weeks of fluconazole consolidation treatment between weeks 2 and 10). The baseline characteristics of the participants were similar in the trial groups ( Table 1).

Primary End Point
In the intention-to-treat analysis, 10-week mortality was 24 Fig. 2A). The absolute difference in mortality at 10 weeks between the liposomal amphotericin B group and control group was −3.9 percentage  points, and the upper boundary of the one-sided 95% confidence interval was 1.2 percentage points, which was within the prespecified 10-percentage-point noninferiority margin (P<0.001 for noninferiority) (Fig. 2B) The results of the prespecified adjusted analyses (Table 2 and Fig. 2B) and key subgroup analyses (Table S3B) were consistent with those of the primary end-point analysis. In prespecified superiority analyses performed at the 10-week time point, the between-group difference in mortality was −3.9 percentage points with the 95% confidence interval crossing zero (95% CI, −10.0 to 2.2) in the unadjusted analysis and −5.7 percentage points with the 95% confidence in-

1116
T h e ne w e ngl a nd jou r na l o f m e dicine terval not crossing zero (95% CI, −11.4 to −0.04) in the analysis that was adjusted for the covariates associated with cryptococcal mortality.

Secondary End Points
Mortality at 2 weeks, 4 weeks, and 16 weeks is shown in Table S3. The results were consistent with the result of the primary end-point analysis of 10-week mortality, with upper boundaries of the one-sided 95% confidence intervals of less than 10 percentage points. The results of timeto-event analyses of mortality that were performed with the use of Cox regression are shown in Table S4 and Figure 2A.
The mean rate of fungal clearance from the cerebrospinal fluid over the course of 14 days was −0.40 log 10 CFU per milliliter per day in the liposomal amphotericin B group and −0.42 log 10 CFU per milliliter per day in the control group, for a difference of 0.017 log 10 CFU per milliliter per day (95% CI, −0.001 to 0.036) ( Table 2 and Fig. S2). Paradoxical immune reconstitution inflammatory syndrome was reported in 15 of 407 participants (3.7%) in the liposomal amphotericin B group and in 19 of 407 participants (4.7%) in the control group (Table S6). There were no cases of culture-positive relapse in the liposomal amphotericin B group. One case of relapse occurred in a participant in the control group who had received full induction therapy and had initial clearance of cryptococcus from the cerebrospinal fluid but subsequently had poor adherence to consolidation-phase fluconazole. During the initial 10 weeks of follow-up, 71 of 407 participants (17.4%) in each treatment group were readmitted to the hospital at least once (Table S7).

Safety and Adverse Events
During the initial 21 days of treatment in the safety population, there were 382 grade 3 or 4 adverse events in 210 of 420 participants (50.0%) in the liposomal amphotericin B group and 579 grade 3 or 4 adverse events in 263 of 422 participants (62.3%) in the control group (P<0.001). A summary of clinical and laboratory-defined adverse events is provided in Table 3, and a detailed list is provided in Table S8. The mean decrease in hemoglobin level during the first week of the induction period was 0.3 g per deciliter in the liposomal amphotericin B group and 1.9 g per deciliter in the control group (P<0.001); blood transfusion was performed Panel A shows the cumulative all-cause mortality up to week 10 according to treatment strategy in the intention-to-treat population. Panel B shows a noninferiority graph for the differences in all-cause mortality at 10 weeks (calculated as the value in the liposomal amphotericin B group minus the value in the control group). The mean absolute difference in 10-week mortality between the liposomal amphotericin B group and the control group and the two-sided 90% confidence intervals in both unadjusted and adjusted intention-to-treat and per-protocol analyses are shown. The dashed line indicates the prespecified 10-percentage-point noninferiority margin. The analysis was adjusted for prespecified baseline covariates of trial site, age, sex, Glasgow Coma Scale score, CD4+ cell count, cryptococcal colony-forming units per milliliter of cerebrospinal fluid, HIV therapy status, hemoglobin level, and cerebrospinal fluid opening pressure. in 32 of 420 participants (7.6%) in the liposomal amphotericin B group and in 76 of 422 participants (18.0%) in the control group. A grade 3 or 4 increase in the creatinine level developed in 22 of 420 participants (5.2%) in the liposomal amphotericin B group and in 25 of 422 participants (5.9%) in the control group. The mean relative increase in the serum creatinine level from baseline to day 7 was 20.2% in the liposomal amphotericin B group and 49.7% in the control group (P<0.001). Thrombophlebitis leading to antibiotic therapy occurred in 8 of 420 participants (1.9%) in the liposomal amphotericin B group and in 28 of 422 participants (6.6%) in the control group (P = 0.001). A low incidence of grade 4 thrombocytopenia, neutropenia, and elevated alanine aminotransferase level was observed in both treatment groups.

Discussion
This trial showed that induction therapy with a single 10 mg-per-kilogram dose of liposomal amphotericin B in combination with oral flucytosine and fluconazole was noninferior to the WHO-recommended standard of care that included 1 week of amphotericin B deoxycholate given with flucytosine and was associated with significantly fewer adverse events. Because this clinical trial involving HIV-positive adults with cryptococcal meningitis was conducted in a range of health care settings across five countries in southern and eastern Africa with no loss to follow-up, our results are likely to be generalizable to other African settings with a high prevalence of HIV (Table S9). The 10-week mortality of 24.8% observed in the liposomal amphotericin B group in our trial is among the lowest reported from a major cryptococcal meningitis trial in Africa, despite more than a quarter of participants presenting with very severe disease and abnormal baseline mental status. Our trial showed that either strategy (a single dose of liposomal amphotericin B plus 14 days of therapy with flucytosine and fluconazole or short-course treatment with 7 days of amphotericin B deoxycholate plus flucytosine followed by 7 days of fluconazole therapy) can reduce 10-week mortality from cryptococcal meningitis to below 30%. This finding represents a notable improvement on the rates of 40 to 45% reported in trials of 2-week amphotericin B deoxycholate-based regimens that were conducted in resource-limited settings 10,[29][30][31] and is consistent with the relatively favorable outcomes with the 1-week regimen of amphotericin B deoxycholate plus flucytosine that were reported in the ACTA trial. 10 Our trial builds on phase 2 data 23 showing that a single 10-mg-per-kilogram dose of liposomal amphotericin B is effective in clearing cryptococcus from the cerebrospinal fluid. The effect on fungicidal activity with a single high dose of liposomal amphotericin B given with flucytosine and fluconazole matched that of 7 days of treatment with amphotericin B deoxycholate (1 mg per kilogram per day) plus flucytosine. In addition, the regimen that included a single high dose of liposomal amphotericin B led to fewer adverse effects than the 1-week amphotericin B deoxycholate regimen, with fewer adverse events overall, fewer life-threatening grade 4 events, fewer episodes of grade 3 or 4 anemia, a reduced need for blood transfusion, and less severe thrombophlebitis. These findings reflect the toxicity profile of liposomal amphotericin B that is known to be better than that of amphotericin B deoxycholate. 12,14 In this trial, we administered preemptive fluid and electrolytes to all the participants to reduce the risk of amphotericin B-related toxic effects, adopted an intensive blood-monitoring schedule, and actively managed adverse events when they occurred. The reality of routine care in resource-limited settings is that the necessary resources are often not available to implement measures to reduce toxic effects and an intensive monitoring and management approach.
An additional potential benefit of the liposomal amphotericin B regimen is that it may be possible to shorten the length of hospital stay needed to safely administer effective treatment. For the evaluation of safety in this trial, our protocol required that all participants be hospitalized for a 7-day period of inpatient monitoring. However, when scaled-up in real-world situations, earlier discharge will probably be possible for some patients. A cost-effectiveness comparison is under way. Given our results, a single high dose of liposomal amphotericin B may be worth investigating in the treatment of other systemic fungal infections that are prevalent in resourcelimited settings, such as histoplasmosis and talaromycosis. 32,33 Our trial was open label, and clinical care of the critically ill participants with advanced HIV disease was complex. However, both the primary  The adverse event data are presented for the safety population, which included all the participants who underwent randomization and received at least one dose of a trial medication. One participant in the liposomal amphotericin B group withdrew consent after randomization but before receiving any trial medication, and one participant in the control group died after randomization but before receiving any trial medication. Both participants were excluded from the safety analysis. ALT denotes alanine aminotransferase. † P values were derived from chi-square or Student t-tests as appropriate. ‡ Grade 3 anemia was defined as a hemoglobin level of 7.0 to less than 9.0 g per deciliter in men and of 6.5 to less than 8.5 g per deciliter in women, and grade 4 as a hemoglobin level of less than 7.0 g per deciliter in men and of less than 6.5 g per deciliter in women. § Data regarding grade 3 events are reported for the participants who had both baseline and day 7 values available. Data were missing for 50 participants in the liposomal amphotericin B group and 61 in the control group. ¶ Grade 3 neutropenia was defined as a neutrophil count of 400 to 599 per cubic millimeter, and grade 4 as a neutrophil count of less than 400 per cubic millimeter. ‖ Grade 3 thrombocytopenia was defined as a thrombocyte count of 25,000 to 49,999 per cubic millimeter, and grade 4 as a thrombocyte count of less than 25,000 per cubic millimeter. ** Grade 3 creatinine increase was defined as creatinine level of 2.47 to 4.42 mg per deciliter (216 to 400 μmol per liter), and grade 4 as a creatinine level of greater than 4.42 mg per deciliter. † † Data regarding grade 4 events are reported for participants who had both baseline and day 7 values available. Data were missing for 42 participants in the liposomal amphotericin B group and 50 in the control group. ‡ ‡ Grade 3 hypokalemia was defined as a potassium level of 2.0 to 2.4 mmol per liter, and grade 4 as a potassium level greater than 2.4 mmol per liter. § § A grade 3 elevation in ALT level was defined as an ALT level of 200 to 400 IU per liter, and a grade 4 elevation as an ALT level greater than 400 IU per liter. ¶ ¶ During the course of the trial there were two infusion reactions that met the grade 3 criteria, both of which occurred in the liposomal amphotericin B group. Both cases responded to simple supportive measures. There were no participants in whom the prescribed dose of either liposomal amphotericin B or amphotericin B deoxycholate could not be given owing to infusion-related adverse events. We did not collate data on milder infusion reactions. end point of death from any cause and the key safety end points of laboratory-confirmed toxic effects were objectively measured, and a consistent approach to HIV management and antiretroviral therapy was agreed on by the investigators and applied throughout the trial (Table S10) in order to avoid differential management strategies or outcome assessments in the treatment groups. This trial showed that a single high dose of liposomal amphotericin B given with flucytosine and fluconazole was noninferior to the current WHO recommended standard of care for cryptococcal meningitis and offers a practical treatment for the management for HIV-associated cryptococcal meningitis that is easier to administer and associated with fewer drug-related adverse effects. Continued efforts to ensure access to liposomal amphotericin B and flucytosine are needed to enable the implementation of this treatment.