On 2023-06-15 13:36:40, user Rachel Gibson wrote:
This Scientific Correspondence has also been submitted to eLife.
Comment on ‘The clinical pharmacology of tafenoquine in the radical cure of Plasmodium vivax malaria: an individual patient data meta-analysis’<br />
Authors: Raman Sharma1, Chao Chen2, Lionel Tan2, Katie Rolfe1, Ioana-Gabriela Fita2, <br />
Siôn Jones2, Anup Pingle3, Rachel Gibson1, Navin Goyal4*, Isabelle Borghini Fuhrer5, <br />
Stephan Duparc5, Hema Sharma2†, Panayota Bird2<br />
Affiliations: 1GSK, Stevenage, UK; 2GSK, Brentford, UK; 3GSK, Mumbai, India; 4GSK, Upper Providence, PA, USA; 5Medicines for Malaria Venture, Geneva, Switzerland<br />
*At the time of submission of this Letter, Navin Goyal is no longer an employee of GSK and is affiliated to Johnson and Johnson<br />
†At the time of submission of this Letter, Hema Sharma is no longer an employee of GSK and is affiliated to AstraZeneca
Abstract<br />
A single 300 mg dose of tafenoquine, in combination with chloroquine, is currently approved in several countries for the radical cure (prevention of relapse) of Plasmodium vivax malaria in patients aged >=16 years. Watson et al.’s recent publication suggests, however, that the approved dose of tafenoquine is insufficient for radical cure and that a higher 450 mg dose should be recommended. In this response, the authors challenge Watson et al.’s assertion based on empirical evidence from dose-ranging and pivotal studies (published) as well as real-world evidence from post-approval studies (ongoing, therefore currently unpublished). The authors confidently assert that, collectively, these data confirm that the benefit–risk profile of a single 300 mg dose of tafenoquine, co-administered with chloroquine, for the radical cure of Plasmodium vivax malaria in patients who are not G6PD deficient, continues to be favourable.
Introduction<br />
The Plasmodium vivax malarial parasite has a major economic and public health impact, especially in regions such as East Africa, Latin America and South and East Asia.1,2 When present in blood, P. vivax can cause acute malaria with episodes of chills, fever, muscle pains and vomiting. The parasite also has a dormant liver hypnozoite stage, which can reactivate after weeks, months or years, leading to relapses and, potentially, to severe anaemia, permanent brain damage and death.1,2 For effective treatment, eradication of both the blood and liver stages of P. vivax is required (radical cure).2<br />
Since 2018, regulators from the United States initially, and subsequently from Australia, Brazil, Colombia, Thailand, Peru and The Philippines, have approved tafenoquine (as a single oral dose of 300 mg in combination with standard doses of chloroquine) for the radical cure (prevention of relapse) of P. vivax malaria in patients aged >=16 years.1,3-5 A paediatric formulation that allows weight-band-based dosing of children (aged >=2 years) and adolescents is also approved in Australia (since 2022).5 Like primaquine, tafenoquine is an 8-aminoquinoline derivative effective against hypnozoites and all other stages of the P. vivax lifecycle; however, although the World Health Organization (WHO) recommends a 7- or 14-day treatment course for primaquine, tafenoquine is the first single-dose treatment for the radical cure of P. vivax malaria and therefore has patient adherence and convenience advantages.1,3,6 Nonetheless, as an 8 aminoquinoline, the safety profile of tafenoquine is similar to that of primaquine, and both agents can cause oxidant haemolysis in people with glucose-6-phosphate dehydrogenase (G6PD) deficiency.7,8 Acute haemolysis is usually short-lived and does not need specific treatment; however, in rare cases, severe haemolysis may lead to life-threatening anaemia (requiring red blood cell transfusions) or haemoglobinuric renal failure.9 In malaria-endemic regions it has been estimated that 8% of the population are G6PD deficient, although significant variation is reported across regions, with the highest country-specific prevalence estimated in Africa and Western Pacific countries.10,11 G6PD deficiency is an X-linked disorder; males are either G6PD deficient or have normal G6PD activity, whereas females exhibit a wide range of G6PD deficiency.2 Females may be symptomatic if they are homozygous, or if they are heterozygous and inactivation of their normal X chromosome (lyonisation) is skewed towards a deficient phenotype.2,12 Caution is needed because inter-individual variability in the pattern of lyonisation may cause heterozygous females with levels of enzyme activity between 30% and 70% of normal to test as normal for G6PD deficiency using qualitative, phenotypic, rapid diagnostic screening tests.13,14 To reduce the risk of haemolysis, the G6PD status of all potential tafenoquine patients must be determined with a quantitative test capable of accurately differentiating deficient, intermediate and normal G6PD activity levels, and tafenoquine should be withheld from patients with G6PD enzyme levels below 70% of normal.3<br />
Importantly, appropriate clinical practice for the use of 8-aminoquinolines in P. vivax malaria has always been precariously balanced between providing adequate activity against hypnozoites and the real risk of haemolytic harm to patients with G6PD deficiency.15 The cautious benefit–risk balance involved with the single 300 mg dose of tafenoquine has been questioned in a recently published paper in which Watson et al., hypothesise that the current recommended dose of tafenoquine 300 mg is insufficient and that a 450 mg dose of tafenoquine would reduce the risk of relapse.16 That dose is 50% greater than the 300 mg dose approved by the US Food and Drug Administration (FDA), Australian Therapeutic Goods Administration (TGA) and other international regulatory authorities.1,3-5 Herein, the authors discuss concerns regarding the conclusions of Watson et al.<br />
• The benefit–risk profile of tafenoquine 450 mg is not appropriately considered. For example, there is minimal discussion of tafenoquine safety data and key findings from a phase 1 study in healthy female volunteers heterozygous for the G6PD Mahidol variant. This important study demonstrated not only that the haemolytic potential of tafenoquine was dose dependent but also that a single 300 mg dose of tafenoquine had the same potential to cause haemolytic harm as the WHO-recommended dose of primaquine for uncomplicated P. vivax malaria (15 mg/day for 14 days).17,18<br />
• The authors acknowledge that data from the phase 2b, paediatric, pharmacokinetic (PK) bridging study TEACH19 were not available before submission of the Watson et al. manuscript. However, in the TEACH study, in which the tafenoquine dosage in paediatric patients was chosen to match blood exposure in adults receiving 300 mg, tafenoquine was efficacious and generally well tolerated: no patients withdrew from the study because of adverse events.19<br />
• The model used by Watson et al. to predict the recurrence-free rate at 4 months after a 450 mg dose is hypothetical and does not consider data regarding the tafenoquine exposure–response relationship. Importantly, tafenoquine exposure achieved with a single 300 mg dose approaches the plateau of the exposure–response curve; therefore, the incremental recurrence-free rate gained by the proposed 50% increase in dose is small and unlikely to be justified by overall benefit–risk considerations.3 In addition, as primaquine and tafenoquine have different PK and metabolic profiles, the authors consider the extrapolation of data from primaquine to tafenoquine to be problematic.2,9<br />
• The authors feel that, overall, some of the conclusions do not acknowledge evidence-based safety concerns for a >300 mg dose of tafenoquine and do not consider additional data from the INSPECTOR study that the recurrence rate of P. vivax infection within 6 months of tafenoquine treatment was not significantly affected by bodyweight.20<br />
Watson et al. mentioned the phase 2b dose-selection study (DETECTIVE) of tafenoquine,21 from which a single 300 mg dose was chosen for phase 3 evaluation in adults. However, the authors did not point out that, in this study, exposure was a significant predictor of efficacy and doubling the tafenoquine dose from 300 mg to 600 mg was associated with only a marginal increase (from 89.2% to 91.9%) in the primary efficacy endpoint, relapse-free efficacy at 6 months.21 Moreover, in addressing the INSPECTOR study of tafenoquine in Indonesian soldiers, the authors did not specify that this was a study of tafenoquine administered with an artemisinin-based combination therapy rather than chloroquine and, as such, is not directly comparable due to poorly understood but confirmed interactions impacting tafenoquine efficacy.20 Watson et al. also suggest that tafenoquine 300 mg is likely inferior to ‘optimal primaquine regimens’, but it is unclear whether such regimens are the WHO-recommended schedules of primaquine or regimens defined as optimal based on non-regulatory studies of primaquine. The authors provided no specific reference or dosage characterising optimised primaquine therapy, so this a priori inferiority cannot be evaluated.<br />
Methods<br />
The hypothetical causal model proposed by Watson et al. for the clinical pharmacology of tafenoquine for the radical treatment of P. vivax malaria is similarly problematic. Central to this model are methaemoglobin (MetHb) production and active metabolites. However, MetHb is not a validated biomarker of tafenoquine efficacy, and currently there is no evidence, from non-clinical or clinical studies, of circulating active metabolites of tafenoquine; if such metabolites were fleetingly present, they would require extraordinary potency to exert any significant pharmacodynamic effect.22<br />
Regarding radical curative efficacy, Watson et al. selected P. vivax recurrence within 4 months as their primary endpoint. However, the trial-defined primary endpoint at 6 months from the pivotal tafenoquine clinical trials8,21,23 was an FDA requirement and was mandated for analysis purposes. This was to maximise the probability of capturing relapses, including those from regions with longer latency periods. Watson et al. used the INSPECTOR study20 as one of two reasons to justify the selection of a 4-month endpoint. Relapse rates differ greatly from country to country, so the duration of the endpoint should not be based on rates observed in a single country. Moreover, the 6-month rate of loss to follow-up (only 9.1%) does not justify a change of treatment endpoint from 6 months to 4 months.<br />
In their efficacy models, Watson et al. explored the association between the odds of P. vivax recurrence and the following predictors: mg/kg dose of tafenoquine; AUC0–?; peak plasma tafenoquine concentration; terminal elimination half-life; and Day 7 MetHb level. However, details of how the best predictor was selected and how statistical significance was judged were not provided.<br />
Results<br />
Use of a 4-month versus 6-month follow-up period<br />
A key focus of the Watson et al. manuscript is that the authors describe a possible association between tafenoquine mg/kg dose and the odds of recurrence (using logistic regression), with a 4-month rather than the original 6-month follow-up. An odds ratio of 0.66 (95% confidence interval [CI]: 0.51, 0.85) is cited by Watson et al. in their analysis of the effect of tafenoquine mg/kg dose in patients who received tafenoquine 300 mg, but descriptive details for this result and the analysis are limited. Figure 2 in the Watson et al. manuscript shows Kaplan–Meier survival curves for time to first recurrence, based on tafenoquine mg/kg dosing category, but some areas require clarification, such as how the dosing bands were selected.<br />
Rationale for tafenoquine dose selection<br />
Importantly, the classification and regression tree analysis, in which a clinically relevant breakpoint tafenoquine AUC value of 56.4 ug·h/mL was identified, was not discussed.24 Population PK modelling revealed that tafenoquine 300 mg would provide systemic exposure greater than or equal to the AUC breakpoint in approximately 93% of individuals, who would have a high probability (85%; 95% CI: 80, 90) of remaining relapse-free at 6 months.24 Therefore, this ‘… model-based approach was critical in selecting an appropriate phase 3 dose’ for tafenoquine.24 Although data from the TEACH paediatric study19 were not available when Watson et al. conducted their analysis, had the data been available, they would have validated the AUC approach to tafenoquine dose selection, with an overall efficacy of approximately 95%.19 Individuals (aged 2–15 years) were given tafenoquine, based on bodyweight, to achieve the same median AUC as the 300 mg dose in adults (children weighing >10–20 kg received tafenoquine 100 or 150 mg; >20–35 kg received 200 mg; and >35 kg received 300 mg). The recurrence-free rate at 4 months was 94.7% (95% CI: 84.6, 98.3),19 and the TEACH study supported the successful approval of tafenoquine for children aged 2–16 years by the Australian TGA in March 2022.5<br />
Another important counter to the mg/kg-based dose selection is that, when bodyweight categories were fitted as a continuous variable in the INSPECTOR study (using data for the time to recurrence for all participants), neither bodyweight nor bodyweight-by-treatment interactions were statistically significant (p=0.831 and p=0.520, respectively).20<br />
Use of an unvalidated biomarker<br />
Although Watson et al. state that increases in blood MetHb concentrations after tafenoquine administration were highly correlated with mg/kg dose, no correlation coefficients were presented. It should also be re-emphasised that MetHb is not a validated, surrogate biomarker of antimalarial treatment efficacy as a radical cure for P. vivax malaria and was used as a safety measure in the INSPECTOR study.20<br />
Potential safety concerns<br />
In the Tolerability and safety section, Watson et al. state that severe haemolytic events were rare; however, this is because all the studies were randomised and controlled, which excluded patients with <70% G6PD activity. In addition, no mention was made that, in one of the constituent studies (which examined the dose–response for haemoglobin decline in participants with 40–60% G6PD enzyme activity),17 dose escalation of tafenoquine from 300 mg to 600 mg was not attempted due to safety concerns about potential haemolysis in patients with G6PD deficiency. In tafenoquine-treated patients in the real-world setting, some instances of severe haemolysis might be expected, and it is already known from the previously highlighted phase 1 study that the haemolytic potential of tafenoquine increases with increasing dose.17 Watson et al.’s Tolerability and safety section also mentions that one tafenoquine-treated patient had a >5 g/dL decrease in haemoglobin level, but the baseline haemoglobin level and tafenoquine dose are not mentioned. The section may have benefitted from a holistic discussion of safety parameters per tafenoquine dose group: for example, the occurrence of serious adverse events, gastrointestinal adverse events (beyond the selective discussion of vomiting within 1 hour post dose) and neuropsychiatric adverse events.<br />
Discussion<br />
Watson et al. conclude that ‘the currently recommended adult dose is insufficient … increasing the adult dose to 450 mg is predicted to reduce the risk of relapse’; however, the authors have raised several concerns relating to these conclusions. In particular, the authors feel that the safety concerns associated with a higher-than-approved tafenoquine dose have not been thoroughly considered: the safety analysis is limited, and the increased risk of haemolysis in patients with G6PD deficiency that a 450 mg tafenoquine dose (which is 50% greater than the approved 300 mg dose) would pose in vulnerable populations in limited-resource settings is not adequately discussed. In some malaria-endemic regions, 8% of the population may be G6PD deficient, although wide variability exists, and in sub Saharan Africa and the Arabian peninsula the prevalence of G6PD deficiency may exceed 30%.10,11 Therefore, in regions with fragile healthcare systems and limited availability of relevant testing for G6PD deficiency, potential exists for a significantly increased risk of haemolysis if tafenoquine is administered at an above recommended dose (450 mg). Importantly, off-label use of a dose not robustly evaluated in clinical trials would pose a considerable risk to patient safety.<br />
Regarding tafenoquine efficacy, the rationale for a dose increase to 450 mg has limitations. Watson et al. suggest that a 50% increase in the adult dose of tafenoquine (from 300 mg to 450 mg) would prevent one relapse of malaria for every 11 patients treated. However, this number-needed-to-treat estimate is not balanced by a number-needed-to-harm estimate for acute haemolytic anaemia. In addition, the phase 2b part of the DETECTIVE study21 showed that, in countries where the trial was carried out, single doses of tafenoquine 300 mg and 600 mg had similar relapse-free efficacy at 6 months (89.2% and 91.9%, respectively); therefore, the lack of additional benefit for tafenoquine 600 mg in DETECTIVE and the phase 1 study, which demonstrated dose-dependent haemolytic potential for tafenoquine, favour a 300 mg dose.<br />
In summary, based on currently available data, dosing tafenoquine at the approved 300 mg dose, in combination with chloroquine, carefully balances efficacy and safety in the radical cure of P. vivax malaria; indeed, tafenoquine 300 mg demonstrated a favourable benefit–risk profile in a comprehensive clinical development programme that included at-risk populations in regions with fragile or resource-restricted healthcare systems. The arguments raised by Watson et al. come with the concerns articulated here, and the authors confidently assert that a tafenoquine dose increase from 300 mg to 450 mg is not supported by available fact-based evidence for the radical cure of P. vivax malaria in adults aged >=16 years.
References<br />
1. GSK. US FDA approves Krintafel (tafenoquine) for the radical cure of P. vivax malaria [press release]. July 20, 2018. https://www.gsk.com/en-gb/media/press-releases/us-fda-approves-krintafel-tafenoquine-for-the-radical-cure-of-p-vivax-malaria/ (accessed 26 April 2023).<br />
2. Hounkpatin AB et al. Clinical utility of tafenoquine in the prevention of relapse of Plasmodium vivax malaria: a review on the mode of action and emerging trial data. Infect Drug Resist 2019;12:553–570.<br />
3. GSK. Krintafel. Highlights of prescribing information. https://www.accessdata.fda.gov/drugsatfda_docs/label/2018/210795s000lbl.pdf (accessed 26 April 2023).<br />
4. GSK, Medicines for Malaria Venture. Perú becomes second malaria-endemic country in Latin America to approve single-dose tafenoquine for radical cure of P. vivax malaria [press release]. https://www.vivaxmalaria.org/sites/pvivax/files/content/attachments/2021-01-25/GSK%20-%20MMV%20PRESS%20RELEASE%20TAFENOQUINE%20APPROVED%20IN%20PERU.pdf (accessed 26 April 2023).<br />
5. Medicines for Malaria Venture. Single-dose Kozenis (tafenoquine) approved for children with Plasmodium vivax malaria by Australian Therapeutic Goods Administration. https://www.mmv.org/newsroom/press-releases/single-dose-kozenis-tafenoquine-approved-children-plasmodium-vivax-malaria (accessed 26 April 2023).<br />
6. World Health Organization. WHO guidelines for malaria, 14 March 2023. https://www.who.int/teams/global-malaria-programme (accessed 26 April 2023).<br />
7. Milligan R et al. Primaquine at alternative dosing schedules for preventing relapse in people with Plasmodium vivax malaria. Cochrane Database Syst Rev 2019;7:CD012656.<br />
8. Llanos-Cuentas A et al. Tafenoquine versus primaquine to prevent relapse of Plasmodium vivax malaria. N Engl J Med 2019;380:229–241.<br />
9. Baird JK. 8-Aminoquinoline therapy for latent malaria. Clin Microbiol Rev 2019;32.<br />
10. Howes RE et al. G6PD deficiency prevalence and estimates of affected populations in malaria endemic countries: a geostatistical model-based map. PLoS Med 2012;9:e1001339.<br />
11. P. vivax information hub. G6PD global prevalence. https://www.vivaxmalaria.org/diagnosis-treatment/g6pd-deficiency/g6pd-global-prevalence#:~:text=G6PD%20Global%20Prevalence,-Photo%3A%20Jaya%20Banerji&text=G6PD%20deficiency%20affects%20around%20400%20million%20people%20globally (accessed 26 April 2023).<br />
12. Domingo GJ et al. Addressing the gender-knowledge gap in glucose-6-phosphate dehydrogenase deficiency: challenges and opportunities. Int Health 2019;11:7–14.<br />
13. Chu CS et al. Haemolysis in G6PD heterozygous females treated with primaquine for Plasmodium vivax malaria: a nested cohort in a trial of radical curative regimens. PLoS Med 2017;14:e1002224.<br />
14. Baird JK et al. Noninferiority of glucose-6-phosphate dehydrogenase deficiency diagnosis by a point-of-care rapid test vs the laboratory fluorescent spot test demonstrated by copper inhibition in normal human red blood cells. Transl Res 2015;165:677–688.<br />
15. Shanks GD. Historical 8-aminoquinoline combinations: not all antimalarial drugs work well together. Am J Trop Med Hyg 2022;107:964–967.<br />
16. Watson JA et al. The clinical pharmacology of tafenoquine in the radical cure of Plasmodium vivax malaria: An individual patient data meta-analysis. Elife 2022;11:e83433.<br />
17. Rueangweerayut R et al. Hemolytic potential of tafenoquine in female volunteers heterozygous for glucose-6-phosphate dehydrogenase (G6PD) deficiency (G6PD Mahidol variant) versus G6PD-normal volunteers. Am J Trop Med Hyg 2017;97:702–711.<br />
18. World Health Organization. Guidelines for the treatment of malaria, 3rd ed. https://apps.who.int/iris/handle/10665/162441 (accessed 26 April 2023).<br />
19. Velez ID et al. Tafenoquine exposure assessment, safety, and relapse prevention efficacy in children with Plasmodium vivax malaria: open-label, single-arm, non-comparative, multicentre, pharmacokinetic bridging, phase 2 trial. Lancet Child Adolesc Health 2022;6:86–95.<br />
20. Sutanto I et al. Randomised, placebo-controlled, efficacy and safety study of tafenoquine co-administered with dihydroartemisinin-piperaquine for the radical cure of Plasmodium vivax malaria (INSPECTOR). Lancet Infect Dis [2023 May 23:S1473-3099(23)00213-X doi: 101016/S1473-3099(23)00213-X Epub ahead of print PMID: 37236221].<br />
21. Llanos-Cuentas A et al. Tafenoquine plus chloroquine for the treatment and relapse prevention of Plasmodium vivax malaria (DETECTIVE): a multicentre, double-blind, randomised, phase 2b dose-selection study. Lancet 2014;383:1049–1058.<br />
22. GSK. Investigator brochure. Data on file.<br />
23. Lacerda MVG et al. Single-dose tafenoquine to prevent relapse of Plasmodium vivax malaria. N Engl J Med 2019;380:215–228.<br />
24. Tenero D et al. Exposure-response analyses for tafenoquine after administration to patients with Plasmodium vivax malaria. Antimicrob Agents Chemother 2015;59:6188–6194.
Authors’ contributions<br />
Hema Sharma, Lionel Tan, Katie Rolfe, and Navin Goyal contributed to the conception or design of the studies the paper contains data from. All authors contributed to data analysis or interpretation. All authors contributed to the development and writing of this correspondence and approved the final submitted version.
Conflicts of interest statements <br />
Raman Sharma, Siôn Jones, Rachel Gibson, Katie Rolfe, Lionel Tan, Ioana-Gabriela Fita, Chao Chen, Panayota Bird, and Anup Pingle are employees of, and shareholders in GSK.<br />
Hema Sharma is a former employee of GSK, a shareholder in GSK and a current employee of AstraZeneca. Navin Goyal is a former employee and shareholder in GSK and a current employee of Johnson and Johnson. Isabelle Borghini Fuhrer and Stephan Duparc have no conflict of interest to report. <br />
Acknowledgements <br />
Medical writing support was provided by David Murdoch, a contract writer working on behalf of Apollo, and Alex Coulthard of Apollo, OPEN Health Communications, funded by GSK Biologicals SA, in accordance with Good Publication Practice 3 (GPP) guidelines (www.ismpp.org/gpp-2022) "www.ismpp.org/gpp-2022)").
Funding<br />
Funding for this article was provided by GSK Biologicals SA.
Data availability<br />
Data sharing is not applicable to this article as no datasets were generated or analysed.