Cost-effectiveness of single, high-dose, liposomal amphotericin regimen for HIV-associated cryptococcal meningitis in five countries in sub-Saharan Africa: an economic analysis of the AMBITION-cm trial

Abstract

Background: HIV-associated cryptococcal meningitis is a leading cause of AIDS-related mortality. The AMBITION-cm trial showed that a regimen based on a single high dose of liposomal amphotericin B deoxycholate (AmBisome group) was non-inferior to the WHO-recommended treatment of seven daily doses of amphotericin B deoxycholate (control group) and was associated with fewer adverse events. We present a five-country cost-effectiveness analysis.

Methods: The AMBITION-cm trial enrolled patients with HIV-associated cryptococcal meningitis from eight hospitals in Botswana, Malawi, South Africa, Uganda, and Zimbabwe. Taking a health service perspective, we collected country-specific unit costs and individual resource-use data per participant over the 10-week trial period, calculating mean cost per participant by group, mean cost-difference between groups, and incremental cost-effectiveness ratio per life-year saved. Non-parametric bootstrapping and scenarios analyses were performed including hypothetical real-world resource use. The trial registration number is ISRCTN72509687, and the trial has been completed.

Findings: The AMBITION-cm trial enrolled 844 participants, and 814 were included in the intention-to-treat analysis (327 from Uganda, 225 from Malawi, 107 from South Africa, 84 from Botswana, and 71 from Zimbabwe) with 407 in each group, between Jan 31, 2018, and Feb 17, 2021. Using Malawi as a representative example, mean total costs per participant were US$1369 (95% CI 1314-1424) in the AmBisome group and $1237 (1181-1293) in the control group. The incremental cost-effectiveness ratio was $128 (59-257) per life-year saved. Excluding study protocol-driven cost, using a real-world toxicity monitoring schedule, the cost per life-year saved reduced to $80 (15-275). Changes in the duration of the hospital stay and antifungal medication cost showed the greatest effect in sensitivity analyses. Results were similar across countries, with the cost per life-year saved in the real-world scenario ranging from $71 in Botswana to $121 in Uganda.

Interpretation: The AmBisome regimen was cost-effective at a low incremental cost-effectiveness ratio. The regimen might be even less costly and potentially cost-saving in real-world implementation given the lower drug-related toxicity and the potential for shorter hospital stays.

Funding: European Developing Countries Clinical Trials Partnership, Swedish International Development Cooperation Agency, Wellcome Trust and Medical Research Council, UKAID Joint Global Health Trials, and the National Institute for Health Research.

Translations: For the Chichewa, Isixhosa, Luganda, Setswana and Shona translations of the abstract see Supplementary Materials section.

Figure 1:

Using the in-trial scenario: A) Cost-effectiveness plane of deaths prevented (%) and incremental health service cost (US2020$). B) Cost-effectiveness acceptability curve of AmBisome treatment against the control. Based on the potential real-life resource use scenario: C) Cost-effectiveness plane of deaths prevented (%) and incremental health service cost (US2020$). D) Cost-effectiveness acceptability curve of AmBisome treatment against the control. Note: Figures A and C. 10,000 bootstrap iterations with 2,000 shown with 95% confidence interval ellipse. The red dot denotes the mean. Figures B and D. The red line denotes the cumulative probability of the AmBisome arm being more cost effective than the control treatment at a particular willingness to pay threshold (WTP) i.e. cost per life-year gained. The dashed line denotes the cumulative probability of positive health benefits.

Figure 2:

‘Tipping Point’ Scenario analysis demonstrating the change in mean treatment costs per participant if the admission duration in AmBisome arm was reduced by one or two days in each of the five country settings. The hospital “hotel” cost of one day in hospital was $13·85 in Uganda, $15·90 in Zimbabwe $18·36 in Malawi, $80·66 in South Africa and $88·80 in Botswana. Figure A is based on in-trial resource use and Figure B is based on potential real-life resource use.

Figure 3:

Bivariate sensitivity analysis demonstrating the impact of fluctuations in antifungal medication prices on the Incremental Cost Effectiveness Ratio (ICER) with all other parameters held constant using the potential real-life resource use in Malawi. A) Fluctuations in the price of AmBisome (L-AmB) with costs ranging from US$5 to US$30 per 50mg vial and Amphotericin B deoxycholate (AmB costs ranging from US$2 to US$20 per 50mg. B) Fluctuations in the price of AmBisome (L-AmB) values ranging from US$5 to US$30 per 50mg vial and Flucytosine (5FC) values ranging from US$0·30 to US$1·50 per 500mg pill. In this analysis the price of AmB remains constant at US$8·10.