Edited by: Jean Paul Deslypere, Aesculape CRO, Belgium
Reviewed by: Domenico Criscuolo, Italian Society of Pharmaceutical Medicine, Italy; Mmamosheledi Elsie Mothibe, Sefako Makgatho Health Sciences University, South Africa
*Correspondence: Xiaofeng Yan,
This article was submitted to Pharmaceutical Medicine and Outcomes Research, a section of the journal Frontiers in Pharmacology
This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
Our object was to find the most appropriate, most effective, and most readily available of four induction regimens for HIV-associated cryptococcal meningitis (CM) (Regimen A: 1 week of AmB plus 5-FC followed by 1 week of fluconazole, Regimen B: 1 week of AmB plus fluconazole followed by 1 week of fluconazole, Regimen C: 2 weeks of AmB plus 5-FC, Regimen D: 2 weeks of AmB plus fluconazole), given the vast differences between resource-limited and resource-abundant settings regarding therapeutic drug accessibility, availability, and affordability for HIV-associated (CM).
We conducted a network meta-analysis to compare the therapeutic efficacy and safety of four different induction treatment regimens.
The 10-week mortality of Regimen A was significantly lower than that of Regimen B and D, and the 2-week mortality of Regimen A was significantly lower than that of Regimen B. Furthermore, there were no statistically significant differences in 10-week mortality, 2-week mortality, as well as in effective fungicidal activity (EFA) over the first 2 weeks among Regimens B, C, and D. The statistical differences in adverse events between Regimen B and Regimen D, and Regimen C and Regimen D were not calculated to be significant.
Our results indicate that, 1 week of AmB plus 5-FC followed by 1 week of fluconazole is superior to the three other studied regimens, and that when 5-FC is not available, accessible, or affordable, 2 weeks of AmB plus fluconazole or 1 week of AmB plus fluconazole followed by 1 week of fluconazole is an appropriate substitution for 2 weeks of AmB plus 5-FC.
Cryptococcal meningitis (CM) remains a significant contributor to human immunodeficiency virus- (HIV-) associated mortality (
Prompt, rational, and effective antifungal treatment is imperative to HIV-associated CM management (
The WHO guidelines recommendation change was mainly based on the evidence of a network meta-analysis evaluating the best regimen for patients with HIV-associated CM (
In the present network meta-analysis, we compared three WHO recommended regimens (
Randomized controlled trials (RCTs) or cohort studies recruiting HIV-associated CM patients were searched and screened in Pubmed, the Cochrane Library, Web of Science, and EBSCOhost/MEDLINE from inception until Nov 15th, 2019. Keywords connected as “HIV OR AIDS OR human immunodeficiency virus AND cryptococcal meningitis AND (RCT or cohort study)” were used for search in titles and abstracts (
RCTs or cohort studies were included if they satisfied the following criteria:
Study participants were HIV-positive patients with their first episode of CM, and CM was diagnosed by isolation of cryptococcus from CSF cultures, positive CSF India ink staining, and/or positive CSF cryptococcal antigen tests, or isolation of cryptococcus in blood culture with clinical presentations of meningo-encephalitis and typical CSF features (
Patients receiving 1 week of AmB plus 5-FC followed by 1 week of fluconazole, 1 week of AmB plus fluconazole followed by 1 week of fluconazole, 2 weeks of AmB plus 5-FC, and 2 weeks of AmB plus fluconazole were included.
Studies reported 10-week mortality and one of the following outcomes: 2-week mortality, effective fungicidal activity (EFA) over the first 2 weeks, adverse events, and drug cost.
Studies were excluded if:
They were not RCTs or cohort studies.
Study participants were HIV-negative.
The diagnosis of CM was not clearly established by positive CSF cultures, positive India ink staining, or positive blood cultures of cryptococcus in association with clinical presentation and CSF features of CM.
The study regimens were for consolidation treatment, maintenance treatment, pre-emptive antifungal treatment, or secondary prophylaxis, in the absence of data for induction treatment.
Cases were those of recurrence or retreatment, and not those of first episode of CM.
They were not published in English.
We extracted the following data for further analysis: (1) characteristics of studies: first author, date of publication, study type, study duration, and country/area; (2) characteristics of patients: number, age, timing of ART initiation after antifungal therapy; (3) interventions included induction therapeutic regimens and consolidation therapeutic regimens; (4) outcomes included 10-week mortality, 2-week mortality, EFA over the first 2 weeks (
Two review authors (YL and YQ) independently evaluated the methodological quality of the six RCTs included in our network meta-analysis by means of the Cochrane “risk of bias” tool (
Ten-week mortality and 2-week mortality for Regimens A, B, C, and D was assessed using forest plots. Risk ratios (RRs) with 95% confidence intervals (CIs) in each pair comparison were generated for treatment effect measurements (
The cumulative ranking probabilities were summarized using the surface under the cumulative ranking area (SUCRA) curve (
Consistency of network meta-analysis was assessed by global and local consistency (
Statistical heterogeneity was assessed using the inconsistency factor (IF) and
Reporting bias was assessed by examining asymmetry in funnel plots of pairwise comparisons (
Pairwise meta-analysis was conducted with the traditional frequentist approach (
For dichotomous outcome measures, RRs were calculated with 95% CIs. For continuous outcomes, MDs were calculated with 95% CIs. These two types of outcomes were both analyzed by Review Manager Version 5.3 (The Nordic Cochrane Centre, The Cochrane Collaboration, Copenhagen) (
Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines were followed through all phases in our study (
A total of 926 articles were retrieved from four electronic databases through our search strategy, and 89 of these articles were eligible for inclusion after screening titles and abstracts. Among the 89 articles, 82 were excluded due to not reporting 10-week mortality, or one of the following criteria: 2-week mortality, EFA, adverse events and drug cost, absence of data for induction treatment, or recurrence and retreatment. Finally, seven articles (
Flow chart of study selection.
Characteristics of six included studies.
Study (author year) | Number | Study type | Study duration | Country/area | Age (years old) | Induction therapeutic regimens | Consolidation therapeutic regimens | ART initial time (after antifungal therapy) | 2-week mortality | 10-week mortality | EFA over the first 2 weeks (log CFU/ml/day) | Adverse events (AEs) over the first 2 weeks | Drug cost |
---|---|---|---|---|---|---|---|---|---|---|---|---|---|
( |
453 | a RCT | From 2013/01 to 2016/11 | 9 African |
≥18 | (1) AmB (1 mg/kg/day, 1 week) + 5-FC (100 mg/kg/d), 2 weeks; (2) AmB (1 mg/kg/day, 1 week) + fluconazole (1,200 mg/d), 2 weeks; (3) AmB (1 mg/kg/day) + 5-FC (100 mg/kg/d), 2 weeks; |
Fluconazole (800 mg/d until ART initiation at 4 weeks, at 400 mg/d until 10 weeks) | At 4 weeks | (1) 36 out of 111; |
(1) 54 out of 111; |
(1) −0.41 ± 0.26; |
128 grade 3 or 4 AEs occurred in group (1) and (2); |
Not report |
( |
79 | RCT | From 2014/10 to 2016/9 | Botswana and Tanzania | ≥18 | (1) L-AmB (10 mg/kg, on day 1) + fluconazole (1,200 mg/d), 2 weeks; |
fluconazole (800 mg/day until 10 weeks) | ART naive and exposed | (1) 2 out of 18; |
(1) 4 out of 18; |
(1) −0.52 ± 0.35; |
(1) 2 grade 3 or 4 AEs out of 18; |
Not report |
( |
57 | RCT | From 2002/11 to |
Thailand | >13 | (1) AmB (0.7 mg/kg/day, 1 week) + fluconazole (400 mg/d), for 2 weeks; |
fluconazole |
On ART | Not reported | (1) 2 out of 30; |
Not reported | All regimens were well tolerated and no drawn within the first two weeks because of side effects. | Not reported |
( |
561 | System-atic review | No reported | South Africa, Uganda, and Thailand | No specified | (1) AmB (1 mg/kg/d, 5–7 days) + fluconazole (1,200 mg/d), for 2 weeks; (2) AmB (0.7–1 mg/kg/d) + fluconazole (800 mg/d), for 2 weeks; (3) AmB (0.7–1 mg/kg/d) + 5-FC (100 mg/kg/d), for 2 weeks | Not reported | On ART | Not reported | (1) 33 out of 127; (2) 61 out of 203; |
Not reported | Not reported | Total cost of care: |
( |
67 | RCT | From 2006/08 to 2008/10 | South Africa | ≥18 | (1) AmB (0.7–1 mg/kg) + 5-FC (100 mg/kg/d), for 2 weeks; (2) AmB (0.7–1 mg/kg) + fluconazole (800 mg/d), for 2 weeks; |
fluconazole (400 mg/d for 8 weeks and 200 mg/d thereafter) | 2 weeks | (1) 1 out of 20; (2) 7 out of 45 | (1) 6 out of 20; |
(1) -0.41 ± 0.22; |
(1) 9 grade 3 or 4 AEs in group 1; |
Not reported |
( |
32 | RCT | From 2002/05 to |
Thailand | No | (1) AmB (0.7 mg/kg/d) + 5-FC (100 mg/kg/d), for 2 weeks; |
fluconazole (400 mg/d for 8 weeks, and 200 mg/d |
After 10 weeks | (1) 1 in 16; (2) 5 in 16; | (1) 1 in 16; (2) 7 in 16 | (1) −0.54 ± 0.19; |
No withdrawn within the first 2 weeks because of side-effects | Not reported |
( |
199 | RCT | From 2004/4 to |
Vietnam | >14 | (1) AmB (1 mg/kg) + 5-FC (100 mg/kg/d), for 2 weeks; (2) AmB (1 mg/kg) + fluconazole (400 mg twice daily), for 2 weeks | (1) fluconazole (400 m/d for 8 weeks); |
Some at study entry; some within 2 weeks; some between 2 to 10 weeks | (1) 15 out of 100; (2) 20 out of 99 | (1) 30 out of 100; (2) 33 out of 99 | (1) −0.42 ± 0.20; (2) −0.32 ± 0.25 | 85 adverse events both in group (1) and group (2) | Not reported |
aRCT, randomized controlled trial.
bArm (1), (2), and (3), which have same dose of L-AmB, and different course (single, 2, or 3 dose) of L-AmB but all less than 1 week, were combined as Regimen B in Jarvis et al.’s study. The mean EFA of arm (1), (2), and (3) was calculated as the EFA of Regimen A in Jarvis et al.’s study.
cArm (2) and arm (3), which have the same course and same dose of AmB and same course but different dose of fluconazole (800 mg/d vs. 600 mg/d), were combined as Regimen D in Loyse et al.’s study. The mean EFA of arm (2) and arm (3) was calculated as the EFA of Regimen D in Loyse et al.’s study.
The risk bias of 6 RCTs described as “low risk,” “high risk,” or “unclear risk” with supporting evidence is presented in
Ten-week mortality of seven trials with six pairwise comparisons (1,448 persons) was assessed by STATA 15.1, forming a rhombus-shaped network (
Network plot for 10-week mortality among Regimens A, B, C, and D. Regimen A: 1 week of AmB plus 5-FC followed by 1 week of fluconazole; Regimen B: 1 week of AmB plus fluconazole followed by 1 week of fluconazole; Regimen C: 2 weeks of AmB plus 5-FC; Regimen D: 2 weeks of AmB plus fluconazole. Note: The thicker the line, the more studies of the comparison between 2 interventions.
The 10-week mortality of the six pairwise comparisons was displayed in a forest plot (
Treatment effect.
Two-week mortality of four trials was compared and presented in forest plots (
Cumulative ranking probabilities of 10-week mortality for each treatment regimen are shown in
Cumulative ranking probability of 10-week mortality for each regimen.
As displayed in
Forest plot of comparisons of EFA over the first 2 weeks. SD, standard deviation.
As displayed in
Forest plot of comparisons of adverse events.
Consistency was assessed using 10-week mortality data, and was shown in
Loop-specific heterogeneity was evaluated by calculating IF
The reporting bias funnel plot of six comparisons is displayed in
In this study, we conducted a network meta-analysis to compare the appropriateness of three WHO-recommended induction therapeutic regimens (regimens A, C, and D), and a 5-FC-free induction therapeutic regimen (regimen B) usually chosen in clinical practice in resource-limited settings. All trials included in our study were carried out in resource-limited settings, and therefore, our results would be of clinical significance for the treatment of HIV-associated CM in those specific settings.
The 10-week mortality of Regimen A was significantly lower than that of Regimen B and Regimen D, and the 2-week mortality of Regimen A was significantly lower than that of Regimen B. The above results are consistent with the results of another network meta-analysis conducted by
Our results of drug effectiveness suggest that 5-FC as a second drug is superior to fluconazole when combined with AmB, because Regimen A (1 week of AmB plus 5-FC followed by 1 week of fluconazole) was found to be more efficacious than Regimen B (1 week of AmB plus fluconazole followed by 1 week of fluconazole). Unfortunately, 5-FC remains neither readily accessible/available, nor affordable in resource-limited settings despite of its superior efficacy (
Furthermore, 5-FC is not licensed in 89 (71.2%) of 125 countries, and unavailable in 95 (76.0%) of 125 countries (
Some limitations of our meta-analysis have to be pointed out. Firstly, one systematic review pooling the 10-week mortality and drug cost of 18 trials and cohorts was incorporated in our meta-analysis, which unfortunately had a paucity of outcome data for direct comparisons between regimens. Secondly, we did not perform an analysis of drug costs due to the limited data for medical expenditure in the included studies. Thirdly, because of the limited number of eligible studies and the limited data extracted from these studies, we did not conduct an analysis of ART use for patients in each study, nor did we evaluate the doses of antifungal medications prescribed to subjects in each study. We also did not assess the utilization of therapeutic lumbar puncture for subjects in each study, which may well have had an impact on heterogeneity in the present study. Finally, the liposomal AmB incorporated in Regimen B, and the high risk of detection bias and attrition bias may have contributed to a degree of bias to our results. Additionally, the SUCRA rankings should be interpreted cautiously, even when accompanied by statistically significant or clinically meaningful effects, due to the imprecision of treatment effects on mortality throughout the network meta-analysis (
Regimen A (1 week of AmB plus 5-FC followed by 1 week of fluconazole) remains the most appropriate induction regimen for the treatment of HIV-associated CM in resource-limited settings, and is superior to Regimen B (1 week of AmB plus fluconazole followed by 1 week of fluconazole) and Regimen D (2 weeks of AmB plus fluconazole) in several aspects. Substituting Regimen B or Regimen D for Regimen C is appropriate in terms of efficacy in resource-limited settings, where 5-FC is likely to be unavailable or unaffordable.
YL, XY, and YC conceived and designed the protocol and study. YL and YQ identified the studies to be screened. XH identified studies for eligibility, extracted data, and assessed the methodological quality of included studies. YL performed the data analysis, with assistance from XY and YC. All authors contributed to the article and approved the submitted version.
This work was supported by the National Science and Technology Major Project of China during the 13th Five-year Plan Period (2018ZX10302104), Key Project of Joint Medical Research Project of Science and Health in Chongqing in 2019 (2019ZDXM012).
All authors declare that this research was conducted in the absence of any commercial or financial relationships that could be construed as potentially causing conflict of interest.
We wish to thank Dr. Vijay Harypursat of the Division of Infectious Diseases, Chongqing Public Health Medical Center for his assistance in proofreading and language improvement for this manuscript.
The Supplementary Material for this article can be found online at: