Abstract
Chemoprophylaxis with weekly doses of tafenoquine (200 mg/day for 3 days before departure [loading dose], 200 mg/week during travel and 1-week post-travel [maintenance doses]) is effective in preventing malaria. Effectiveness of malaria chemoprophylaxis drugs in travellers is often compromised by poor compliance. Shorter schedules that can be completed before travel, allowing ‘drug-free holidays’, could increase compliance and thus reduce travel-related malaria. In this meta-analysis, we examined if a loading dose of tafenoquine alone is effective in preventing malaria in short-term travellers.
Four databases were searched in November 2020 for randomized controlled trials (RCTs) that assessed efficacy and/or safety of tafenoquine for chemoprophylaxis. Network meta-analysis using the generalized pair-wise modelling framework was utilized to estimate the odds ratio (OR) of malaria infection in long-term (>28 days) and short-term (≤28 days) travellers, as well as adverse events (AEs) associated with receiving loading dose of tafenoquine alone, loading dose of tafenoquine followed by maintenance doses, loading dose of mefloquine followed by maintenance doses, or placebo.
Nine RCTs (1714 participants) were included. In long-term travellers, compared to mefloquine, tafenoquine with maintenance doses (OR = 1.05; 95% confidence interval [CI]: 0.44–2.46) was equally effective in preventing malaria, while there was an increased risk of infection with the loading dose of tafenoquine alone (OR = 2.89; 95% CI: 0.78–10.68) and placebo (OR = 62.91; 95% CI: 8.53–463.88). In short-term travellers, loading dose of tafenoquine alone (OR = 0.98; 95% CI: 0.04–22.42) and tafenoquine with maintenance doses (OR = 1.00; 95% CI: 0.06–16.10) were as effective as mefloquine. The risk of AEs with tafenoquine with maintenance doses (OR = 1.03; 95% CI: 0.67–1.60) was similar to mefloquine, while loading dose of tafenoquine alone (OR = 0.58; 95% CI: 0.20–1.66) was associated with lower risk of AEs, although the difference was not statistically significant.
For short-term travellers, loading dose of tafenoquine alone was equally effective, had possibly lower rate of AEs, and likely better compliance than standard tafenoquine or mefloquine chemoprophylaxis schedules with maintenance doses. Studies are needed to confirm if short-term travellers remain free of infection after long-term follow-up.
The meta-analysis was registered in PROSPERO (CRD42021223756).
Tafenoquine is the latest approved drug for malaria chemoprophylaxis. A loading dose of tafenoquine (200 mg/day for 3 days before departure) is as effective in preventing malaria in short-term (≤28 days) travellers as chemoprophylaxis schedules of tafenoquine or mefloquine with maintenance doses, allowing travellers to have a ‘drug-free holiday’.
Introduction
The World Health Organization (WHO) estimates that in 2019 there were 229 million cases of malaria and 409 000 malaria deaths worldwide.1 Significant progress has been made in malaria control and elimination programs in the last two decades, including a reduction in malaria case incidence (from 80 to 57 cases per 1000 population in 2000 and 2019, respectively) and deaths (from 736 000 to 409 000 in 2000 and 2019, respectively). However, malaria is still endemic in 87 countries, with Africa being the most affected continent, accounting for 94% of the cases globally.1
Travellers to endemic areas are at particularly high risk of malaria infection. Schlagenhauf et al.2 reported that malaria was the most common diagnosis among ill European returned travellers. Using GeoSentinel records, Leder et al.3 identified that acute febrile illnesses accounted for nearly a quarter of travel-related illnesses, with malaria being the most common cause. Furthermore, among the 28 deaths recorded in the GeoSentinel database from 2007 to 2011, one quarter of them was due to malaria.3
Pretravel medical consultations aim to reduce travel-related morbidity and mortality in travellers, and have been associated with lower incidence and severity of malaria.2 During pretravel consultations, the three pillars of malaria prevention are discussed; mosquito avoidance measures (e.g. repellent), chemoprophylaxis, and advice to seek medical attention in the event of fever.4,5 A number of effective medications are available for chemoprophylaxis, and the choice of medication is based on many factors including the traveller’s demographic characteristics (e.g. age), comorbidities (e.g. glucose-6-phosphate dehydrogenase [G6PD] deficiency), destination (e.g. regions with chloroquine-resistant malaria), duration of travel, proposed activities in country, cost of medications and dosing schedule (daily vs weekly, duration). Currently, standard chemoprophylaxis schedules involve taking medications before, during and after travel to an endemic region.5
Despite effective drugs being available for malaria chemoprophylaxis, their effectiveness is undermined by poor compliance (e.g. missing doses, not adhering to recommended schedule, not completing the course of medications) or even complete failure to take chemoprophylaxis.6 Evidence suggests that poor compliance and failure to take chemoprophylaxis are associated with elevated risk of severe malaria and malaria-associated deaths.7 Several studies have found that forgetting to take medications while travelling was a common problem,8–10 thus compliance could potentially be improved by utilizing simpler and shorter duration schedules, or schedules that can be completed before travel. In 2007, a group of travel medicine and malaria experts highlighted the need to explore pretravel malaria chemoprophylaxis schedules, or ‘drug-free holidays,’ to improve compliance.11 Notably, we demonstrated that a short, 3-day schedule of atovaquone/proguanil completed before departure had a very high compliance rate (>95%), and was well tolerated and accepted by travellers.12
Tafenoquine is a synthetic long-acting analogue of primaquine, with an improved therapeutic index and safety profile. Tafenoquine is the newest licensed malaria drug on the market, having been approved by the United States Food and Drug Administration and Therapeutic Goods Administration in Australia in 2018 for malaria prophylaxis and radical cure of Plasmodium vivax malaria. A recent meta-analysis found that the recommended chemoprophylaxis schedule with weekly doses of tafenoquine in travellers (i.e. 200 mg daily for 3 days before departure [loading dose], 200 mg weekly during travel and 1-week post-travel [maintenance doses]) is effective in preventing P. falciparum and P. vivax malaria.13
One of the advantages of tafenoquine over primaquine is its longer half-life (~15 days). Given that the median trip duration in returned travellers diagnosed with malaria was 32 days; we hypothesize that short-term (≤28 days) travellers might benefit from the long half-life of tafenoquine by being protected from clinical infection for the length of their travel with loading dose alone, i.e. a 3-day schedule which can be completed before departure leading to ‘drug-free holidays’. Therefore, we conducted a network meta-analysis (NWMA) to compare the efficacy of a loading dose of tafenoquine alone vs malaria chemoprophylaxis schedules with regular maintenance doses.
Methods
Findings of this study are presented according to PRISMA reporting guidelines for NWMA (Supplementary Material S1).14 The study was prospectively registered in PROSPERO (CRD42021223756).
Search strategy
A systematic search was conducted in PubMed, Embase, Cochrane Central Register of Controlled Trials (CENTRAL), and Web of Science from inception to 9 November 2020 for studies assessing efficacy and/or safety of tafenoquine for malaria chemoprophylaxis. The search string was initially designed in PubMed using a pool of six relevant articles by a research information specialist (JC), and then converted for use in Embase, CENTRAL, and Web of Science using the Polyglot Search Translator.15 The search included the following keywords and subject terms ‘tafenoquine’, ‘Arakoda’, ‘WR-238605’, ‘Krintafel’, ‘malaria’, ‘plasmodium infection’, ‘pre-exposure prophylaxis’, ‘chemoprevention’, ‘prophylaxis’, ‘chemoprophylaxis’, ‘randomised controlled trials’. The complete search strings for all databases are available in the Supplementary Material (S2). To accomplish a comprehensive assessment of the literature, the systematic search was supplemented with a forward and backward citation search as well as retrieving the first 20 similar articles from PubMed for each of the publications included from the initial search. The reference lists of systematic reviews and meta-analyses13,16 were hand-searched for additional publications.
Selection criteria
The inclusion of studies was restricted to randomized controlled trials (RCTs) conducted in humans at risk of malaria infection that compared the efficacy in preventing malaria (i.e. parasitemia) and/or safety profile (i.e. adverse events [AE]) of at least two of the following interventions:
Loading dose of tafenoquine alone
Loading dose of tafenoquine followed by maintenance doses
Other drugs for malaria chemoprophylaxis with maintenance doses
Placebo
In addition, RCTs were only included if malaria diagnosis was confirmed by a laboratory test (e.g. microscopy or polymerase chain reaction). AEs were defined according to the reports of the included studies. No language or date restrictions were applied. RCTs assessing the efficacy of tafenoquine for radical cure, or where only clinical diagnosis (i.e. without laboratory confirmation) was utilized, were excluded. Observational studies, conference abstracts and proceedings were also excluded.
Screening and data extraction
After duplicate records were removed, titles and abstracts of all papers that were extracted by the search engine, were uploaded to Rayyan (http://rayyan.qcri.org/)17 for screening. Two authors (NI and SW) independently screened the titles and abstracts for studies that met the inclusion criteria. The selected studies underwent full-text screening by the same authors.
Data from the included studies were independently extracted by NI and SW, and summarized in a spreadsheet. Discrepancies during the selection of studies and data extraction were resolved through the involvement of a third author (LFK). The extracted information included: (i) authors, year of publication; (ii) country/region where the RCT was conducted; (iii) predominant Plasmodium sp. circulating at the study site; (iv) year(s)/month(s) when the study was conducted; (v) study population characteristics (e.g. age); (vi) chemoprophylaxis schedules used; (vii) total number of participants; (viii) number of participants with malaria infection; (ix) number of participants who presented with AEs; and (x) duration of follow-up period. Data presented in a graphical format (e.g. Kaplan–Meier curves) were extracted using Plot Digitizer version 2.6.8 (http://plotdigitizer.sourceforge.net/).
If RCTs contained more than two arms, inclusion of loading dose of tafenoquine alone was prioritized followed by loading dose of tafenoquine with maintenance doses, other drugs for malaria chemoprophylaxis with maintenance doses, and placebo. Schedules that used the recommended therapeutic doses (e.g. 200 mg daily for loading dose and 200 mg weekly of tafenoquine) were chosen over schedules with infra- or supra-therapeutic doses (e.g. 400 mg weekly of tafenoquine).
Quality assessment
The appraisal tool for RCTs developed by the Joanna Briggs Institute was used for quality assessment.18 This assessment of methodological quality examines bias through a list of 13 safeguards including randomization, allocation concealment, similar treatment groups at baseline, blinding of participants, blinding of investigators, blinding of outcome assessors, treatment groups treated equally other than the intervention of interest, complete follow-up, intention-to-treat analysis, similarity in the measurement approach for outcomes in treatment groups, reliable way to measure outcome, and appropriate statistical analysis approach. The higher the number of safeguards present, the more assurance that the findings have been measured in a manner that is free from bias.
Statistical analyses
There were three outcomes of interest, all assessed using odds ratios (OR): (i) malaria infection during the entire follow-up period (i.e. long-term exposure), (ii) malaria infection during the first 28 days (i.e. short-term exposure) and (iii) AEs with different chemoprophylaxis schedules or placebo. The duration of the follow-up in the studies while at risk of malaria infection (i.e. exposure time), was used as a proxy for short- (≤28 days) and long-term (>28 days) travellers.
The NWMA involved: (i) direct meta-analysis using ORs for the three outcomes; (ii) generation of mixed treatment effects against a common comparator using the generalized pair-wise modelling (GPM) framework19; and (iii) meta-analysis of direct and indirect effects of each chemoprophylaxis schedule (or placebo) to estimate a final mixed effect size for each schedule against a common comparator. Other drugs for malaria chemoprophylaxis with maintenance doses was used as the common comparator—mefloquine (loading dose of 250 mg daily for 3 days followed by weekly maintenance doses) was the only drug that met this criterion in the included RCTs, see result section.
PRISMA flowchart.
Network plots (left column) and forest plots (right column) for (A) malaria infection with long-term exposure (>28 days), (B) malaria infection with short-term exposure (≤28 days) and (C) adverse events. Network plots: the size of the circles is proportional to the number of arms included in the analysis, whereas the width of the lines is proportional to the number of pairs. L-TQ loading dose of tafenoquine alone; TQ loading dose of tafenoquine followed by maintenance doses; MQ loading dose of mefloquine followed by maintenance doses.
The inverse variance heterogeneity model was used for all direct meta-analyses.20 When double-arm-zero-events (i.e. zero events for the outcome of interest in both treatment arms) were recorded, an OR = 1 was imputed and a continuity correction with a factor of 0.5 was applied to estimate the standard error of the study. A sensitivity analysis with a follow-up of 6 weeks (28 days of the short-term traveller + 14 days of median incubation period for P. vivax21) was conducted. Additional sensitivity analysis restricted to RCTs with natural exposure to mosquitos (i.e. excluding RCTs with inoculation experiments) was carried out.
The OR was computed from the study data and used in the analyses. The relative risks (RRs) have been shown to be dependent on the baseline prevalence of an outcome, and therefore, are neither ‘portable’ nor should they be used in meta-analysis.22 For ease of interpretation,23 the final results were converted to RRs using a baseline risk of malaria infection (0.11% per month) and AEs (18.8%) for travellers taking mefloquine with maintenance doses, extracted from one of the largest cohort studies (Malpro study) on malaria chemoprophylaxis effectiveness in travellers.24 Computation of the RRs (from the final results on the OR scale) were done using the Stata logittorisk module.25
Statistical heterogeneity across meta-analysis of direct effects was assessed using Cochran’s Q and the H index.19 Transitivity across the network was assessed by examining inconsistency using the weighted pooled H index (|$\overline{H}$|). The minimum value of H and |$\overline{H}$| can take is 1, |$\overline{H}<3$| was considered to represent minimal inconsistency.19 Publication bias was assessed using comparison-adjusted funnel plots.26
All analyses involved in the GPM framework were conducted using MetaXL version 5.3 (EpiGear Int Pty Ltd; Brisbane, Australia). Network and comparison-adjusted funnel plots were created in Stata version 14.1 (College Station, TX, USA).
Results
Yield of search strategy
The initial search retrieved 107 unique records which were screened by title and abstract. A total of 21 articles remained and underwent full-text screening, and 11 articles met the inclusion criteria. The backward/forward citation search and similarity search identified 417 additional unique records that were initially screened by title and abstract and then by full-text; of these, one additional article met the inclusion criteria. A total of 12 articles, reporting findings of nine RCTs (1714 participants), met the inclusion criteria and were included in the NWMA. The yield of the search, screening and selection process of articles is presented in the PRISMA flowchart (Figure 1).
Characteristics of the included studies
The included RCTs were conducted between 1997 and 2017. Among the nine RCTs, six were conducted in malaria-endemic countries (Timor-Leste, Gabon, Ghana, Kenya ×2, and Thailand) and three in non-endemic countries (Australia, UK and USA). Participants of all RCTs living in endemic regions received treatment for pre-existing parasitemia before the start of the chemoprophylaxis schedule that was examined in the trial. All RCTs included healthy adults from the community or military personnel, except for Lell et al.27 which included secondary school students aged 12–20 years. The duration of follow-up of the RCTs ranged from 6 to 26 weeks with a median of 13 weeks (interquartile range 10–24 weeks).
Study intervention arms included in the analyses were loading dose of tafenoquine alone (3 arms, 196 participants), loading dose of tafenoquine followed by maintenance doses (7 arms, 936 participants), loading dose of mefloquine followed by maintenance doses (3 arms, 304 participants) and placebo (5 arms, 278 participants). The network plots for the three outcomes are presented in Figure 2.
Seven RCTs assessed efficacy of tafenoquine chemoprophylaxis. Six were conducted in malaria-endemic countries during malaria transmission season, and one in a non-endemic setting (Australia) where participants were inoculated with P. falciparum. All RCTs except for one reported AEs. The detailed characteristics of the included RCTs are presented in Table 1.
Characteristics of the included studies
| Author, publication year | Study period | Study location | Plasmodium species endemic in study area | Population characteristic | Treatment for pre-existing cases of parasitemia before chemoprophylaxis | Interventions assessed | Follow-up (weeks) | Assessment of malaria infection | Assessment of adverse events |
|---|---|---|---|---|---|---|---|---|---|
| Dow 201539 | May–October 2000 | Village of Kombewa, Nyanza Province, Kenya | P. falciparum | Healthy adults (18–55 years) | Halofantrine 250 mg qd for 3 days | —TQ 200 mg* —MQ 250 mg* —Placebo | 24 | Microscopy (thick and thin blood smear) | Not reported |
| Edstein 200140; Edstein 200341; Walsh 200442 | April–November 1998 | Nam Yun District, Ubol Ratchatani Province, Thailand | P. falciparum and P. vivax | Royal Thai Army and border patrol soldiers in good health (18–55 years) | Artesunate 200 mg qd for 3 days, doxycycline 100 mg qd for 7 days | —TQ 400 mg (monthly)* —Placebo* | 26 | Microscopy (thick blood smear) | Self-reported during monthly follow-ups and medical records |
| Green 201443 | July 2011–June 2012 | California and Maryland, USA | - | Healthy adults (18–65 years) | None | —TQ 300 mg (once) —TQ 600 mg (once) —L-TQ 400 mg* —Moxifloxacin 400 mg —Placebo* | 9 | Not assessed | Self-reported during follow-ups |
| Hale 200344; Dow 201539 | 1998 | Kassena Nankana District, Ghana | P. falciparum | Healthy adults (18–60 years for men and 40–60 years for women) | Quinine 600 mg tid for 4 days, doxycycline 100 mg bid for 7 days, primaquine 30 mg qd for 14 days | —TQ 25 mg —TQ 50 mg —TQ 100 mg —TQ 200 mg* —MQ 250 mg* —Placebo | 12 | Microscopy (thick and thin blood smear) | Self-reported during weekly follow-ups |
| Leary 200945 | Not reported | Maryland, USA and Berkshire, UK | - | Healthy adults (18–55 years) | None | —TQ 200 mg* —Placebo* | 23 | Not assessed | Self-reported during monthly follow-ups and medical records |
| Lell 200027 | February–July 1999 | Lambarene, Gabon | P. falciparum | Secondary school students (12–20 years) | Halofantrine 250 mg qd for 3 days | —L-TQ 25 mg —L-TQ 50 mg —L-TQ 100 mg —L-TQ 200 mg* —Placebo* | 10 | Microscopy (thick blood smear) | Self-reported during weekly follow-ups |
| McCarthy 201928 | January–March 2017 | Brisbane, Australia | P. falciparum (inoculated) | Healthy adults without history of malaria exposure (20–42 years) | None | —TQ 200 mg* —Placebo* | 6 | Quantitative PCR | Self-reported during follow-ups |
| Nasveld 201046; Dow 201447 | October 2000–April 2001 | Bobonaro District, East Timor | P. falciparum and P. vivax | Australian military personnel in good health deployed to East Timor (18–55 years) | None | —TQ 200 mg* —MQ 250 mg* | 26 | Microscopy (thick and thin blood smears) | Self-reported during monthly follow-ups and medical records |
| Shanks 200148; Dow 201539 | May–September 1997 | Village of Ndori Province, Kenya | P. falciparum | Healthy adults (18–55 years) | Halofantrine 250 mg qd for 3 days | —L-TQ 400 mg* —TQ 200 mg* —TQ 400 mg —Placebo | 13 | Microscopy (thick and thin blood smear) – 2 consecutive positive smears | Self-reported during weekly follow-ups |
| Author, publication year | Study period | Study location | Plasmodium species endemic in study area | Population characteristic | Treatment for pre-existing cases of parasitemia before chemoprophylaxis | Interventions assessed | Follow-up (weeks) | Assessment of malaria infection | Assessment of adverse events |
|---|---|---|---|---|---|---|---|---|---|
| Dow 201539 | May–October 2000 | Village of Kombewa, Nyanza Province, Kenya | P. falciparum | Healthy adults (18–55 years) | Halofantrine 250 mg qd for 3 days | —TQ 200 mg* —MQ 250 mg* —Placebo | 24 | Microscopy (thick and thin blood smear) | Not reported |
| Edstein 200140; Edstein 200341; Walsh 200442 | April–November 1998 | Nam Yun District, Ubol Ratchatani Province, Thailand | P. falciparum and P. vivax | Royal Thai Army and border patrol soldiers in good health (18–55 years) | Artesunate 200 mg qd for 3 days, doxycycline 100 mg qd for 7 days | —TQ 400 mg (monthly)* —Placebo* | 26 | Microscopy (thick blood smear) | Self-reported during monthly follow-ups and medical records |
| Green 201443 | July 2011–June 2012 | California and Maryland, USA | - | Healthy adults (18–65 years) | None | —TQ 300 mg (once) —TQ 600 mg (once) —L-TQ 400 mg* —Moxifloxacin 400 mg —Placebo* | 9 | Not assessed | Self-reported during follow-ups |
| Hale 200344; Dow 201539 | 1998 | Kassena Nankana District, Ghana | P. falciparum | Healthy adults (18–60 years for men and 40–60 years for women) | Quinine 600 mg tid for 4 days, doxycycline 100 mg bid for 7 days, primaquine 30 mg qd for 14 days | —TQ 25 mg —TQ 50 mg —TQ 100 mg —TQ 200 mg* —MQ 250 mg* —Placebo | 12 | Microscopy (thick and thin blood smear) | Self-reported during weekly follow-ups |
| Leary 200945 | Not reported | Maryland, USA and Berkshire, UK | - | Healthy adults (18–55 years) | None | —TQ 200 mg* —Placebo* | 23 | Not assessed | Self-reported during monthly follow-ups and medical records |
| Lell 200027 | February–July 1999 | Lambarene, Gabon | P. falciparum | Secondary school students (12–20 years) | Halofantrine 250 mg qd for 3 days | —L-TQ 25 mg —L-TQ 50 mg —L-TQ 100 mg —L-TQ 200 mg* —Placebo* | 10 | Microscopy (thick blood smear) | Self-reported during weekly follow-ups |
| McCarthy 201928 | January–March 2017 | Brisbane, Australia | P. falciparum (inoculated) | Healthy adults without history of malaria exposure (20–42 years) | None | —TQ 200 mg* —Placebo* | 6 | Quantitative PCR | Self-reported during follow-ups |
| Nasveld 201046; Dow 201447 | October 2000–April 2001 | Bobonaro District, East Timor | P. falciparum and P. vivax | Australian military personnel in good health deployed to East Timor (18–55 years) | None | —TQ 200 mg* —MQ 250 mg* | 26 | Microscopy (thick and thin blood smears) | Self-reported during monthly follow-ups and medical records |
| Shanks 200148; Dow 201539 | May–September 1997 | Village of Ndori Province, Kenya | P. falciparum | Healthy adults (18–55 years) | Halofantrine 250 mg qd for 3 days | —L-TQ 400 mg* —TQ 200 mg* —TQ 400 mg —Placebo | 13 | Microscopy (thick and thin blood smear) – 2 consecutive positive smears | Self-reported during weekly follow-ups |
*Intervention arms included in the analysis.
qd once daily; bid twice daily; tid three times daily; L-TQ loading dose of tafenoquine alone; TQ loading dose of tafenoquine followed by maintenance doses; MQ loading dose of mefloquine followed by maintenance doses.
Characteristics of the included studies
| Author, publication year | Study period | Study location | Plasmodium species endemic in study area | Population characteristic | Treatment for pre-existing cases of parasitemia before chemoprophylaxis | Interventions assessed | Follow-up (weeks) | Assessment of malaria infection | Assessment of adverse events |
|---|---|---|---|---|---|---|---|---|---|
| Dow 201539 | May–October 2000 | Village of Kombewa, Nyanza Province, Kenya | P. falciparum | Healthy adults (18–55 years) | Halofantrine 250 mg qd for 3 days | —TQ 200 mg* —MQ 250 mg* —Placebo | 24 | Microscopy (thick and thin blood smear) | Not reported |
| Edstein 200140; Edstein 200341; Walsh 200442 | April–November 1998 | Nam Yun District, Ubol Ratchatani Province, Thailand | P. falciparum and P. vivax | Royal Thai Army and border patrol soldiers in good health (18–55 years) | Artesunate 200 mg qd for 3 days, doxycycline 100 mg qd for 7 days | —TQ 400 mg (monthly)* —Placebo* | 26 | Microscopy (thick blood smear) | Self-reported during monthly follow-ups and medical records |
| Green 201443 | July 2011–June 2012 | California and Maryland, USA | - | Healthy adults (18–65 years) | None | —TQ 300 mg (once) —TQ 600 mg (once) —L-TQ 400 mg* —Moxifloxacin 400 mg —Placebo* | 9 | Not assessed | Self-reported during follow-ups |
| Hale 200344; Dow 201539 | 1998 | Kassena Nankana District, Ghana | P. falciparum | Healthy adults (18–60 years for men and 40–60 years for women) | Quinine 600 mg tid for 4 days, doxycycline 100 mg bid for 7 days, primaquine 30 mg qd for 14 days | —TQ 25 mg —TQ 50 mg —TQ 100 mg —TQ 200 mg* —MQ 250 mg* —Placebo | 12 | Microscopy (thick and thin blood smear) | Self-reported during weekly follow-ups |
| Leary 200945 | Not reported | Maryland, USA and Berkshire, UK | - | Healthy adults (18–55 years) | None | —TQ 200 mg* —Placebo* | 23 | Not assessed | Self-reported during monthly follow-ups and medical records |
| Lell 200027 | February–July 1999 | Lambarene, Gabon | P. falciparum | Secondary school students (12–20 years) | Halofantrine 250 mg qd for 3 days | —L-TQ 25 mg —L-TQ 50 mg —L-TQ 100 mg —L-TQ 200 mg* —Placebo* | 10 | Microscopy (thick blood smear) | Self-reported during weekly follow-ups |
| McCarthy 201928 | January–March 2017 | Brisbane, Australia | P. falciparum (inoculated) | Healthy adults without history of malaria exposure (20–42 years) | None | —TQ 200 mg* —Placebo* | 6 | Quantitative PCR | Self-reported during follow-ups |
| Nasveld 201046; Dow 201447 | October 2000–April 2001 | Bobonaro District, East Timor | P. falciparum and P. vivax | Australian military personnel in good health deployed to East Timor (18–55 years) | None | —TQ 200 mg* —MQ 250 mg* | 26 | Microscopy (thick and thin blood smears) | Self-reported during monthly follow-ups and medical records |
| Shanks 200148; Dow 201539 | May–September 1997 | Village of Ndori Province, Kenya | P. falciparum | Healthy adults (18–55 years) | Halofantrine 250 mg qd for 3 days | —L-TQ 400 mg* —TQ 200 mg* —TQ 400 mg —Placebo | 13 | Microscopy (thick and thin blood smear) – 2 consecutive positive smears | Self-reported during weekly follow-ups |
| Author, publication year | Study period | Study location | Plasmodium species endemic in study area | Population characteristic | Treatment for pre-existing cases of parasitemia before chemoprophylaxis | Interventions assessed | Follow-up (weeks) | Assessment of malaria infection | Assessment of adverse events |
|---|---|---|---|---|---|---|---|---|---|
| Dow 201539 | May–October 2000 | Village of Kombewa, Nyanza Province, Kenya | P. falciparum | Healthy adults (18–55 years) | Halofantrine 250 mg qd for 3 days | —TQ 200 mg* —MQ 250 mg* —Placebo | 24 | Microscopy (thick and thin blood smear) | Not reported |
| Edstein 200140; Edstein 200341; Walsh 200442 | April–November 1998 | Nam Yun District, Ubol Ratchatani Province, Thailand | P. falciparum and P. vivax | Royal Thai Army and border patrol soldiers in good health (18–55 years) | Artesunate 200 mg qd for 3 days, doxycycline 100 mg qd for 7 days | —TQ 400 mg (monthly)* —Placebo* | 26 | Microscopy (thick blood smear) | Self-reported during monthly follow-ups and medical records |
| Green 201443 | July 2011–June 2012 | California and Maryland, USA | - | Healthy adults (18–65 years) | None | —TQ 300 mg (once) —TQ 600 mg (once) —L-TQ 400 mg* —Moxifloxacin 400 mg —Placebo* | 9 | Not assessed | Self-reported during follow-ups |
| Hale 200344; Dow 201539 | 1998 | Kassena Nankana District, Ghana | P. falciparum | Healthy adults (18–60 years for men and 40–60 years for women) | Quinine 600 mg tid for 4 days, doxycycline 100 mg bid for 7 days, primaquine 30 mg qd for 14 days | —TQ 25 mg —TQ 50 mg —TQ 100 mg —TQ 200 mg* —MQ 250 mg* —Placebo | 12 | Microscopy (thick and thin blood smear) | Self-reported during weekly follow-ups |
| Leary 200945 | Not reported | Maryland, USA and Berkshire, UK | - | Healthy adults (18–55 years) | None | —TQ 200 mg* —Placebo* | 23 | Not assessed | Self-reported during monthly follow-ups and medical records |
| Lell 200027 | February–July 1999 | Lambarene, Gabon | P. falciparum | Secondary school students (12–20 years) | Halofantrine 250 mg qd for 3 days | —L-TQ 25 mg —L-TQ 50 mg —L-TQ 100 mg —L-TQ 200 mg* —Placebo* | 10 | Microscopy (thick blood smear) | Self-reported during weekly follow-ups |
| McCarthy 201928 | January–March 2017 | Brisbane, Australia | P. falciparum (inoculated) | Healthy adults without history of malaria exposure (20–42 years) | None | —TQ 200 mg* —Placebo* | 6 | Quantitative PCR | Self-reported during follow-ups |
| Nasveld 201046; Dow 201447 | October 2000–April 2001 | Bobonaro District, East Timor | P. falciparum and P. vivax | Australian military personnel in good health deployed to East Timor (18–55 years) | None | —TQ 200 mg* —MQ 250 mg* | 26 | Microscopy (thick and thin blood smears) | Self-reported during monthly follow-ups and medical records |
| Shanks 200148; Dow 201539 | May–September 1997 | Village of Ndori Province, Kenya | P. falciparum | Healthy adults (18–55 years) | Halofantrine 250 mg qd for 3 days | —L-TQ 400 mg* —TQ 200 mg* —TQ 400 mg —Placebo | 13 | Microscopy (thick and thin blood smear) – 2 consecutive positive smears | Self-reported during weekly follow-ups |
*Intervention arms included in the analysis.
qd once daily; bid twice daily; tid three times daily; L-TQ loading dose of tafenoquine alone; TQ loading dose of tafenoquine followed by maintenance doses; MQ loading dose of mefloquine followed by maintenance doses.
Quantitative synthesis
The pooled estimate of seven RCTs (when loading dose of mefloquine followed by maintenance doses was used as the common comparator) revealed that for long-term exposure, tafenoquine with maintenance doses (OR = 1.05; 95% confidence interval [CI] 0.44–2.46) was equally effective as mefloquine in preventing malaria infection. However, there was an increased odds with loading dose of tafenoquine alone (OR = 2.89; 95% CI 0.78–10.68) or placebo (OR = 62.91; 95% CI 8.53–463.88) (Figure 2A, Supplementary Material S3). When the analysis was restricted to the first 28 days (i.e. short-term exposure), loading dose of tafenoquine alone (OR = 0.98; 95% CI 0.04–22.42) and tafenoquine with maintenance doses (OR = 1.00; 95% CI 0.06–16.10) were as effective as mefloquine, but there was an increased odds of infection with placebo (OR = 35.94; 95% CI 0.80–1617.28) (Figure 2B, Supplementary Material S3).
The pooled analysis of AEs in eight RCTs during the entire follow-up period, revealed that tafenoquine with maintenance doses (OR = 1.03; 95% CI 0.67–1.60) had the same odds of AEs as mefloquine, while loading dose of tafenoquine alone (OR = 0.58; 95% CI 0.20–1.66) and placebo (OR = 0.57; 95% CI 0.28–1.17) presented lower odds of AEs although these were not statistically significant (Figure 2C, Supplementary Material S3). Severe AEs were uncommon (0% for loading dose of tafenoquine alone; 4.6% for tafenoquine with maintenance doses; 1.0% for mefloquine with maintenance doses) and it was not possible to provide a pooled estimate. Sensitivity analyses revealed similar results when the follow-up of short-term travellers was extended to 6 weeks (Supplementary Material S4) and when the analysis was restricted to RCTs with natural exposure to mosquitoes (i.e. exclusion of the RCT by McCarthy et al.28).
Interpretation of findings in terms of risks
Efficacy in long-term exposure
At a baseline risk of malaria infection with mefloquine of 0.33%24 [0.11% per month for 3 months (median follow-up period of the included RCTs)], the estimated effect of tafenoquine with maintenance doses (RR = 1.05; 95% CI 0.45–2.45) suggested a slight increase in infection risk when compared to mefloquine, but there was almost no evidence against the model hypothesis of zero difference in risk at this sample size. With loading dose of tafenoquine only, the estimated effect suggested almost 3-fold increase in malaria risk compared to mefloquine (RR = 2.88; 95%CI 0.78–10.35), but with limited evidence against the model hypothesis of zero difference in risk at this sample size.
Efficacy in short-term exposure
At a baseline risk of infection of 0.11% per month with mefloquine,24 both loading dose of tafenoquine alone (RR = 0.98; 95% CI 0.04–21.90) and tafenoquine with maintenance doses (RR = 1.00; 95% CI 0.06–15.84) suggested no difference in malaria risk when compared with mefloquine, and there was also almost no evidence against the model hypothesis of zero difference in risk at this sample size.
Adverse events
At a baseline risk of AEs of 18.8% with mefloquine,24 the estimated effect with loading dose of tafenoquine alone (RR = 0.63; 95% CI 0.24–1.48) suggested lower risk of AEs though with limited evidence against the model hypothesis of zero difference in risk at this sample size. The estimated effect of tafenoquine with maintenance doses (RR = 1.02; 95% CI 0.71–1.44) suggested a slight increase in AEs compared to mefloquine, but with almost no evidence against the model hypothesis of zero difference in risk at this sample size.
Risk of bias and publication bias assessment
Overall, the quality of the RCTs was high; the maximum number of safeguard deficiencies was 3 out of 13 in two RCTs. The most common deficiencies in safe-guarding against bias were not blinding of outcome assessors (55.6%, five RCTs); unclear or no allocation concealment (44.4%, four RCTs), and not blinding of investigators (22.2%, two RCTs) (Supplementary Material S5).
The comparison-adjusted funnel plots revealed no asymmetry for risk of malaria infection during the entire follow-up or during the first 28 days. There was minor asymmetry for AEs (Supplementary Material S6).
Discussion
After pooling data from nine RCTs (including 1714 participants) our study provides valuable evidence on the efficacy of a loading dose of tafenoquine alone for short-term (≤28 days) chemoprophylaxis against P. falciparum and P. vivax. Our findings revealed that loading dose of tafenoquine alone was not effective for long-term travellers, and maintenance doses are required to achieve a protective effect, due to the decrease in plasma concentration of tafenoquine. However, in short-term travellers, a loading dose of tafenoquine alone had similar efficacy in preventing malaria infection for the length of the travel (up to 28 days) with a better safety profile than schedules with weekly doses of tafenoquine or mefloquine. The benefits of completing a course of chemoprophylaxis medications before travel include; better compliance, removing the need to take medications while overseas especially in countries where gastro-intestinal upset may impair drug absorption, potential for AEs to be managed before departure including change in chemoprophylaxis medication if necessary, and convenience (no need to carry or buy medications while overseas).
Given the shorter duration of the schedule, it is expected that a loading dose of tafenoquine alone will have higher compliance rate than chemoprophylaxis schedules with daily or weekly maintenance doses. Although there are no post-marketing studies documenting compliance rate with the standard weekly doses of tafenoquine, its compliance may be similar to the standard mefloquine schedule (52–80%),29,30 which also requires weekly maintenance doses. In our study noted earlier, we found that a 3-day schedule of atovaquone/proguanil had an impressive compliance rate (97.7%)12; thus, we anticipate that the 3-day schedule required for the loading dose of tafenoquine alone would likely have a similar compliance rate in travellers.
In non-endemic countries, malaria is increasing being reported from immigrants originating from malaria-endemic countries who travel to their countries of origin to visit friends and relatives (VFRs).31,32 Mischlinger et al.33 found that people who are VFRs have a different risk perception of malaria, leading to lower adherence to recommendations for chemoprophylaxis. Therefore, uptake and compliance rates for pretravel chemoprophylaxis in VFR travellers may differ from other travellers. Future research should investigate the efficacy and compliance rate of short chemoprophylaxis schedules in VFRs.
Despite the comprehensive search and the robust methods utilized in the meta-analysis, we acknowledge that there were certain limitations. There was some degree of uncertainty in the findings (i.e. wide CIs) because of the low number of RCTs included in the meta-analysis. The criteria used to define AE was not homogenous across RCTs; this difference could lead to heterogeneity across studies, but we did not find evidence of such heterogeneity (i.e. inconsistency) during the analysis. More importantly, the length of the follow-up while participants were at risk of malaria infection was used as a proxy to define short- and long-term travellers. Although, the average incubation period for P. falciparum and P. vivax is 9–14 days and 12–17 days, respectively21 (thus the sensitivity analysis was extended to 6 weeks to account for the 2 weeks of incubation period), it could prolong for months specially for P. vivax. It is also possible that participants who remained free of attacks during the first 28 days could have had suppressed the infection, rather than killing the asexual blood stages and later emerges and experience delayed attacks.34 Therefore, studies with long follow-up periods (e.g. 6 months) are needed to confirm the efficacy/effectiveness of loading dose of tafenoquine alone in short-term travellers after they return to a non-endemic setting/country.
Malaria infection is a complex condition that includes many stages i.e. patent, sub-patent and latent hepatic stage. The RCTs included in the current meta-analysis focused in the prevention of patent blood-stage infection, thus future studies are needed that investigate the efficacy of a loading dose of tafenoquine alone in sub-patent stage infections. Although the results suggest that in short-term travellers the risk of malaria attacks within the first 28 days is low with a loading dose of tafenoquine alone, future studies also need to consider the possibility of post-exposure attacks after 28 days due to reactivation of hypnozoites.35 These studies specifically designed to compare loading dose of tafenoquine alone vs schedules with maintenance doses at multiple endpoints (e.g. 4, 8, 12 weeks), and with extended follow-up periods are needed to strengthen the evidence and determine the maximum duration of protection after a loading dose alone.
Children (especially VFRs) are at particularly high risk of malaria; Angelo et al.36 found that among 325 children with malaria reported to GeoSentinel, 10% presented with severe malaria. In our systematic search, we only found one RCT27 that included secondary school students, thus results cannot be generalized to younger children. Tafenoquine is contraindicated in individuals with G6PD deficiency,37 in pregnancy and breastfeeding women (if the infant’s C6PD status is unknown),38 and not approved for chemoprophylaxis in persons aged <18 years37; thus studies are needed to develop safe and effective short chemoprophylaxis schedules for these specific group of people.
In conclusion, our study support that a loading dose of tafenoquine alone (200 mg daily for 3 days) could be effective for chemoprophylaxis against P. falciparum and P. vivax malaria in G6PD normal travellers going on short duration trips. Compared to standard chemoprophylaxis schedules with weekly maintenance doses, a loading dose of tafenoquine alone seemed to be equally effective and had a lower rate of AEs. Fewer medications would also have the advantage of being cheaper and likely will have a higher compliance rate. RCTs specifically designed to examine loading dose of tafenoquine alone in short-term travels, but with long follow-ups are needed to confirm the findings in this meta-analysis given the possibility of post-exposure attacks.
Authors’ contribution
CLL, SARD and LFK did the conception and design of the study. JC and LFK developed the search strategy. NI, SW and LFK did the collection and assembly of the dataset. NI, SARD and LFK did the analysis of the dataset. NI, SW, CLL, SARD, DJM, ACAC and LFK did the interpretation of results. All authors had written the manuscript and did the final approval.
Acknowledgements
CLL and LFK were supported by Australian National Health and Medical Research Council Fellowships (APP1193826 and APP1158469).
Conflicts of interest
None declared.
References
Author notes
Nazmul Islam and Sophie Wright contributed equally to this manuscript.
