FormalPara Key Points

Tafenoquine is approved for Plasmodium vivax malaria relapse prevention.

In this placebo-controlled study in healthy volunteers, single-dose tafenoquine 300 mg had no pharmacodynamic effect on the retina or clinically relevant effects on ophthalmic safety.

These findings are reassuring given that tafenoquine will be used in resource-poor areas, where access to ophthalmological care is often limited.

1 Introduction

Tafenoquine is an antimalarial drug for the prevention of relapse in Plasmodium vivax malaria—termed radical cure [1,2,3]. More than a third of the world’s population are at risk of P. vivax malaria and it is the most prevalent form of malaria outside Africa [4]. Relapses are caused by reactivation of a dormant parasite liver state—the hypnozoite [5]. Patients can suffer multiple relapses recurring weeks or months after first becoming infected. Relapses also generate gametocytes, the sexual phase of the parasite, allowing onward human-to-mosquito transmission and maintaining the parasite in the population [5]. As hypnozoites are undetectable by any diagnostic method and cause no symptoms, infected individuals can silently harbor the parasite for extended periods. In this manner, the parasite evades malaria control efforts and can be transported to areas where malaria has been previously eliminated [6]. Thus, radical cure of P. vivax malaria, which kills both the blood-stage parasites that cause clinical disease and the hypnozoites that propagate relapse, is necessary to both reduce the burden of disease and accelerate malaria elimination [6]. Tafenoquine was recently approved by the United States Food and Drug Administration and the Australian Therapeutic Goods Administration, with approval from these stringent authorities supporting regulatory submissions in P. vivax-endemic countries to enable patient access achieved through their National Malaria Control Programs.

The 8-aminoquinolines primaquine and tafenoquine are the only available treatments that can clear P. vivax hypnozoites from the liver. These drugs must be given with a schizonticide to clear blood-stage parasites, usually 3 days of chloroquine. The recommended regimen for primaquine lasts 14 days (15 mg/kg/day), and outside clinical trials, adherence is poor and effectiveness compromised [7]. In contrast, tafenoquine is given as a single 300-mg dose [1,2,3], a significant improvement in convenience, which should promote real-world clinical effectiveness. Chloroquine, which has known effects on the retina [8], is a 4-aminoquinoline, and although tafenoquine is an 8-aminoquinoline, as these two compounds will be co-administered, any theoretical risk of retinal adverse events with tafenoquine requires investigation.

Tafenoquine has been studied in malaria prophylaxis and treatment as well as P. vivax radical cure. Ophthalmic findings were reported in a study conducted in Australian soldiers randomized 3:1 to receive either a loading dose of 200 mg daily tafenoquine for 3 days followed by weekly tafenoquine prophylaxis (200 mg) or a loading dose of 250 mg mefloquine/day for 3 days followed by weekly mefloquine prophylaxis (250 mg) for 6 months [9]. Mild vortex keratopathy, which did not affect visual acuity, was observed in 69 out of 74 tafenoquine subjects (93%) versus none out of 21 with mefloquine, and was fully resolved in all subjects by 1 year [9]. Retinal abnormalities were reported in 39.1% of tafenoquine subjects (27/69) and 23.5% (4/17) with mefloquine [9]. However, an expert ophthalmology review board convened to examine the data could not determine the relevance of the retinal findings because no baseline retinal photography data were available [9]. The board also noted that the retinal findings may have been normal variations [9]. Consequently, a follow-up ocular safety study was conducted comparing high-dose tafenoquine monotherapy (400 mg for 3 days) versus chloroquine (1500 mg for 3 days) followed by primaquine (15 mg for 14 days) in P. vivax malaria patients [10]. In this study, baseline data were available, and mild post-baseline keratopathy was observed in 31.8% of patients (14/44) with tafenoquine versus 0% (0/24) with chloroquine/primaquine. On clinical examination, mild post-baseline retinal findings were reported in ten out of 44 patients (22.7%) receiving tafenoquine and two out of 24 (8.3%) receiving chloroquine/primaquine (P = 0.15; treatment difference 14.4%, 95% confidence interval [CI] − 5.7 to 30.8), but with no clinically relevant changes in visual acuity or macular function tests [10]. Masked digital photograph review by an independent reviewer indicated no changes from baseline at day 28 for either group [10]. At day 90, one tafenoquine patient had a slight increase in atrophy versus baseline and one chloroquine/primaquine patient had retinal atrophy at baseline and a slight increase in atrophy versus baseline at day 90, with no effect on visual acuity in either case [10].

In the development program for tafenoquine plus chloroquine in P. vivax radical cure, a single 300-mg tafenoquine dose was used, a therapy with a lower dose and shorter duration than that used in the studies outlined above that were conducted in prophylaxis or monotherapy treatment. Across three randomized clinical trials with tafenoquine 300 mg plus chloroquine in P. vivax radical cure, ophthalmic findings were uncommon based on clinical examination: keratopathy was noted in 0.9% (1/107) of patients, retinal hyperpigmentation in 1.0% (1/102), and retinal hypopigmentation in 1.0% (1/102) [2]. However, in these studies, retinal changes were assessed using clinical examination and masked reading of digital fundus photography. These methods are not consistent with recent recommendations [8], and have inadequate sensitivity to detect potential subtle changes in retinal health. Additional confounders included co-administration of chloroquine, which has known ophthalmic safety effects [8], and the disease under study, as symptomatic P. vivax malaria can also include ocular manifestations [11].

The current study assessed the pharmacodynamic effects of a 300-mg single oral dose of tafenoquine on the retina in comparison to placebo using spectral domain optical coherence tomography (SD-OCT) and fundus autofluorescence (FAF) in healthy volunteers using procedures consistent with American Academy of Ophthalmology guidelines [8]. Secondary objectives were to assess the overall ophthalmic safety of tafenoquine and general safety. The study is registered at ClinicalTrials.gov (identifier: NCT02658435).

2 Methods

2.1 Study Design

This was a phase I, prospective, multicenter, randomized, single-masked, placebo-controlled, parallel-group study of a 300-mg single oral dose of tafenoquine in adult healthy volunteers. The study was conducted between 2 February 2016 and 14 September 2017 at three study centers in the USA. A central reading center (Digital Angiography Reading Center, Great Neck, NY, USA) masked to treatment allocation conducted the evaluation of SD-OCT, FAF and fundus photography outcomes. The study was conducted in accordance with the Declaration of Helsinki and all applicable local laws and followed Good Clinical Practice procedures. An independent review board at each participating center reviewed and approved the study protocol. All subjects provided written informed consent.

2.2 Population

Eligible participants were of either sex, aged ≥ 18 to ≤ 45 years, weighed ≥ 35 to ≤ 100 kg, and were healthy as determined by the investigator, with hematology and chemistry values within the normal range, and capable of giving signed informed consent. Full inclusion and exclusion criteria are listed in detail in Electronic Supplementary Material 1. Key general exclusion criteria were as follows: current or chronic history of liver disease, or known hepatic or biliary abnormalities, hemoglobin values outside the lower limit of normal range, phenotypic glucose-6-phosphate dehydrogenase (G6PD) deficiency (determined by a quantitative assay of enzyme activity; defined as < 70% of locally defined median value), and a Fredericia-corrected QTc interval of > 450 ms. Subjects had to be of non-reproductive potential or capable of adhering to contraceptive requirements. Pregnant and lactating females were excluded. Key ophthalmic exclusion criteria were bilateral best-corrected visual acuity (BCVA) of ≤ 72 letters, eye disease that could compromise ophthalmic assessments, intraocular surgery or laser photocoagulation within 3 months of dosing, high myopia (equal to, or worse than, − 6.00 diopters), anterior, intermediate or posterior uveitis or history of significant intraocular infectious disease or another active inflammatory disease, SD-OCT central subfield thickness < 250 µm or > 290 µm, presence of significant abnormal patterns on FAF or ocular abnormalities on fundus photography at screening, or uncontrolled intraocular pressure > 22 mmHg.

2.3 Randomization and Masking

Eligible subjects were randomized in a 2:1 ratio to receive either a single 300-mg oral dose of tafenoquine (GlaxoSmithKline, Harlow, United Kingdom) or matched placebo on day 1 of the study (Fig. 1). Treatment was directly observed, administered with 240 mL of water and given with food. The randomization schedule was generated by GlaxoSmithKline Quantitative Sciences using validated internal software. The GlaxoSmithKline online randomization system (RAMOS NG) was used to allocate subjects to treatment. Subjects were masked to treatment; as were the readers at the independent retinal reading center, who reviewed the data. Although not required by the protocol, investigators were also masked to treatment allocation, with only pharmacy staff aware of treatment allocation throughout the study.

Fig. 1
figure 1

Study schematic

Full size image

2.4 Assessments

Subjects were followed up for 90 days, with safety follow-up on day 7 (± 1 day) and telephone contact for determination of adverse events and concomitant medication on days 30 and 60 (± 7 days) (Fig. 1). Ophthalmic assessments were conducted at baseline (day − 7 to day 1 pre-dose) and at day 90 (− 7 to + 14 days) (Fig. 1).

A complete eye exam was conducted, i.e., pupil, motility and confrontation visual field examination; slit lamp evaluation of anterior segment structures (including cornea); and intraocular pressure measurement and dilated fundus examination (indirect ophthalmoscopy and slit lamp biomicroscopy, including the lens). BCVA was measured at a distance of 4 m using Early Treatment Diabetic Retinopathy Study (ETDRS) visual acuity charts.

SD-OCT, FAF and color fundus images were obtained by an appropriately trained technician using a protocol provided by the central reading center using approved equipment (SD-OCT and FAF: SPECTRALIS® system, Heidelberg Engineering, Inc., Franklin, MA, USA; fundus photography: Zeiss FF4, Carl Zeiss Meditec, Dublin, CA, USA). Images were evaluated by the investigator and confirmed by the reading center for protocol inclusion/exclusion criteria and safety monitoring. For SD-OCT and FAF, pharmacodynamic effect was evaluated by the central reading center using protocol-defined criteria (see Electronic Supplementary Material 2 for details).

Planned safety evaluations were conducted at screening and days 7, 30, 60 and 90 (Fig. 1). Adverse events were recorded and categorized using MedDRA (version 20) preferred terms.

Triplicate 12-lead electrocardiograms (ECGs) were performed at screening, and standard clinical laboratory tests were performed at screening and day 7 (clinical chemistry, urinalysis and hematology, including methemoglobin).

2.5 Objectives and Endpoints

The primary objective of the study was to assess the pharmacodynamic effects of a 300-mg single oral dose of tafenoquine on the retina via SD-OCT and FAF in healthy adult volunteers. Secondary objectives were to assess the overall ophthalmic safety of tafenoquine and general safety.

The primary endpoint was the proportion of subjects in the tafenoquine group with significant protocol-defined retinal changes at day 90 compared with baseline. A subject was considered to have met the primary endpoint if any one of five parameters evaluated using SD-OCT or FAF indicated a change from baseline in either eye: (1) central subfield thickness change from baseline ≥ 40 µm (automated reading); (2) total macular volume change of ≥ 10% from baseline (automated reading); (3) central retinal/lesion thickness change from baseline ≥ 40 µm (manual reading); (4) ellipsoid zone disruption (manual reading); or (5) if there was any abnormal FAF pattern at follow-up (see Electronic Supplementary Material 2 for details). If the automated readings of central subfield thickness or total macular volume at either visit were missing or deemed unreliable by the independent masked reading center, then corresponding manual readings of the central retinal area in the same eye were used for the primary endpoint.

Secondary endpoints for ophthalmic safety were (a) the mean change from baseline at day 90 in key SD-OCT parameters, including central retinal/lesion thickness, central subfield thickness, total macular volume, central retinal/lesion thickness, and ellipsoid zone disruption width; (b) the proportion of subjects with abnormal changes at day 90 versus baseline observed on FAF; (c) the mean change at day 90 versus baseline in BCVA; and (d) the proportion of subjects with vortex keratopathy from slit lamp examination. Changes for quantitative measures were defined as a change from baseline by at least 10% or by 2 standard deviations (SDs) from historical mean values for normal subjects. For qualitative assessments, new findings or worsening of findings from baseline exams considered to be of clinical importance by the masked reading center were reported as a change from baseline.

For general safety, secondary endpoints were the frequency of adverse events and serious adverse events, the severity of adverse events, and abnormal values or clinically important changes in laboratory tests. A serious adverse event was defined as one that resulted in death, was life-threatening, required hospitalization, caused disability/incapacity, led to a congenital abnormality or birth defect, or was deemed serious by the investigator. Additionally, liver injury and impaired liver function with alanine transaminase ≥ 3 times the upper limit of normal (ULN) plus either total bilirubin ≥ 2 times the ULN or an international normalized ratio (INR) > 1.5 was defined as a serious adverse event. A severe adverse event was defined as an event that prevents normal everyday activities as assessed by the investigator or a change in laboratory values graded for severity using the Common Terminology Criteria for Adverse Events (CTCAE) (version 4.03).

2.6 Statistical Analysis

The safety population included all subjects who received at least one dose of study medication. The primary endpoint was evaluated in the ophthalmic safety population, in which all subjects had valid baseline and day 90 SD-OCT and FAF measurements within the visit window that allowed derivation of the primary outcome. For the primary endpoint, the point estimate and the corresponding 95% CI were calculated for the proportion of subjects with significant protocol-defined retinal changes from baseline to day 90. As a secondary endpoint, the difference in the proportion of subjects with clinically significant changes between tafenoquine and placebo was estimated. Sensitivity analyses for these endpoints were performed on the per-protocol analysis that excluded subjects with major protocol violations, and separately on an expanded ophthalmic safety population, including subjects with follow-up beyond the day 90 +14-day assessment window. A missing equals failure analysis of the primary endpoint was also conducted using the safety population.

For quantitative secondary ophthalmic endpoints, the change from baseline was summarized for the safety population. Adverse events, ECGs, vital signs and laboratory parameters were evaluated in the safety population and summarized descriptively.

There were no formal hypotheses tested in this study. A sample size of 300 subjects was planned to be treated with tafenoquine based on a 95% probability of detecting an event when the underlying risk of clinically significant retinal findings at 90 days follow-up was 1%. A placebo control group of 150 subjects was used to compare the ophthalmic assessments between the two groups. To allow for 10% lost to follow-up, a total sample size of up to 500 subjects was planned to be enrolled in the study. With a smaller sample size of 275 subjects treated with tafenoquine, there was a 95% probability of detecting an event if the underlying risk was considered to be 1.1%.

3 Results

3.1 Patient Disposition

Of the 1522 subjects screened, 500 subjects were enrolled and randomized, of which 498 were treated and 486 (98%) completed the study (Fig. 2). Two subjects randomized to tafenoquine did not receive study medication and were not included in any analyses. Baseline characteristics were balanced between the two groups (Table 1). The ophthalmic safety population included 306 out of 330 subjects (92.7%) in the tafenoquine group and 161 out of 168 (95.8%) in the placebo group (Fig. 2). Major protocol deviations were observed in 21 subjects in the ophthalmic safety population, which led to their exclusion from the per-protocol population (Fig. 2).

Fig. 2
figure 2

Subject disposition

Full size image
Table 1 Baseline characteristics
Full size table

3.2 Ophthalmic Safety

3.2.1 Primary Outcome: Retinal Changes from Baseline

One subject in each group met the primary endpoint for retinal changes identified with SD-OCT or FAF, i.e., one out of 306 (0.3%) with tafenoquine and one out of 161 (0.6%) with placebo (Fig. 3) (see Electronic Supplementary Material 3). Both cases had unilateral focal ellipsoid zone disruption, with no changes in either eye for the other four parameters in the primary endpoint and no visual adverse events or changes in BCVA. The upper limit of the 95% CI for the proportion of subjects with retinal changes in either eye with tafenoquine was 1.5%, and the treatment difference between tafenoquine and placebo was − 0.3% (95% CI − 3.1 to 1.3). There were no subjects with bilateral retinal changes in the study. The upper limit of the 95% CI for the proportion of subjects with retinal changes in both eyes following tafenoquine (0/306) was 0.9%, with a treatment difference of 0% (95% CI − 2.3%, 1.2%). The sensitivity analysis which included all subjects with day 90 data irrespective of the assessment window did not identify any further subjects with protocol-defined retinal changes.

Fig. 3
figure 3figure 3

Spectral domain optical coherence tomography (SD-OCT) images in the two subjects meeting the primary endpoint for retinal changes who received a 300-mg single-dose tafenoquine (note that the subject had this abnormality at baseline and was enrolled in error) or b placebo

Full size image

The tafenoquine-treated subject with focal ellipsoid zone disruption at day 90 also had this finding at baseline and was, therefore, enrolled into the study in error and was excluded from the per-protocol population. The focal disruption was related to small retinal pigment epithelial detachments. Thus, in the per-protocol analysis, none of 294 subjects (0%; upper limit for 95% CI 0.9%) had retinal changes with tafenoquine and one of 152 (0.7%) had such changes with placebo; treatment difference − 0.7% (95% CI − 3.6%, 0.7%). For the subject in the placebo group with ellipsoid zone disruption during the study, there was no detectable or known cause for this disruption.

One tafenoquine subject had a missing central subfield thickness measurement at baseline. Seven other subjects (four placebo; three tafenoquine) had large changes from baseline recorded in central subfield thickness as a result of misalignment of images between baseline and day 90 visits using the automated software algorithm. In these eight cases, for the primary endpoint, the results were imputed from the categorical change from baseline measured manually in central retinal/lesion thickness for the same eye.

3.2.2 Secondary Ophthalmic Safety Outcomes

Quantitative results of SD-OCT parameters did not show any treatment differences in central subfield thickness, central retinal/lesion thickness, macular cube volume, or subretinal fluid thickness (Table 2; see also Electronic Supplementary Material 4 and 5). For intraocular pressure, there was no clinically important change from baseline to day 90 or any difference between treatment arms (Electronic Supplementary Material 6). Vortex keratopathy was present in one subject in the placebo group at baseline, but resolved by day 90. New-onset vortex keratopathy at day 90 was recorded in both eyes for one subject in the tafenoquine group (1/307; upper limit of 95% CI 1.4%), but not in the placebo group (0/162; treatment difference 0.3% [95% CI − 0.2 to 1.8]). However, post-database lock, it was discovered that, because of a data entry error, the case of vortex keratopathy in the tafenoquine group had in fact been noted by the ophthalmologist as a LASIK scar with calcium deposits, present at baseline and day 90. This subject did not, therefore, have vortex keratopathy, but the results have been reported as recorded in the study.

Table 2 Treatment comparison for tafenoquine versus placebo of the change from baseline to day 90 in quantitative optical confocal tomography results (safety population)
Full size table

3.2.3 Best-Corrected Visual Acuity

Absolute values at day 90 and change from baseline to day 90 for BCVA did not indicate any clinically meaningful effect of tafenoquine (Table 3). There was no treatment difference for BCVA (logMAR) between tafenoquine and placebo (0.012 [95% CI − 0.002 to 0.027]; i.e., < 1 letter on the Snellen scale), and the categorical results of BCVA were also similar between the two groups (see Electronic Supplementary Material 7 and 8). Scatter plots of SD-OCT parameters versus BCVA were produced to explore any association of structural and functional changes (Fig. 4). No patterns indicating any effect of tafenoquine were identified from these analyses.

Table 3 Best-corrected visual acuity at baseline and day 90 and change from baseline to day 90
Full size table
Fig. 4
figure 4

Change in baseline for best-corrected visual acuity versus optical coherence tomography parameters. No subjects had subretinal fluid thickness. Excludes the central subfield thickness results of subjects with imputed change from baseline and one tafenoquine (TQ) subject that had a missing baseline value (4 placebo; 4 TQ)

Full size image

3.3 General Safety

The frequency of adverse events of any cause was similar between the tafenoquine (26.1% [86/330]) and placebo (25.0% [42/168]) groups (Table 4; see also Electronic Supplementary Material 9). Headache, nausea and vomiting were more common with tafenoquine than placebo (Table 4). There were no deaths and no serious or severe adverse events during the study, nor any adverse events leading to study withdrawal. All adverse events in the tafenoquine group were mild in severity. In the placebo group, two adverse events were of moderate severity (tooth abscess and limb injury). One pregnancy was reported in the tafenoquine group (7 weeks pregnant at the day 90 assessment), the outcome of the pregnancy was unknown as the patient was lost to follow-up, despite multiple attempts at contact.

Table 4 Most common adverse events of any cause
Full size table

The incidence of drug-related adverse events was higher with tafenoquine (9.7% [32/330]) than with placebo (2.4% [4/168]). This was mainly because of a higher rate of nausea (3.6% [12/330] vs 0.6% [1/168]) in subjects receiving tafenoquine (see Electronic Supplementary Material 10). There was no difference between treatment groups in the incidence of ophthalmic adverse events (Table 5). ECGs were only performed at baseline, with no clinically relevant findings (data not shown).

Table 5 All ophthalmic adverse events of any cause
Full size table

Clinical chemistry values and changes from baseline showed no values or patterns of concern and were consistent with the expected tafenoquine safety profile (see Electronic Supplementary Material 11 and 12) [12]. There was no clinically important effect of tafenoquine on hematological parameters, except for methemoglobin, which was mildly increased from baseline to day 7 with tafenoquine (mean change 0.41% [SD 0.58]) compared with placebo (0.06% [SD 0.40]) (see Electronic Supplementary Material 13 and 14).

4 Discussion

This placebo-controlled study reports the pharmacodynamic effect of 300-mg single-dose tafenoquine in healthy subjects followed up for 90 days. Retinal changes from baseline were assessed using SD-OCT and FAF using procedures consistent with guidelines from the American Academy of Ophthalmology [8]. Additionally, clinical examination and BCVA evaluation was conducted. The methods used had the required sensitivity to detect early retinal changes that might have resulted from tafenoquine treatment. Ophthalmic safety and general safety were secondary objectives.

One subject in the tafenoquine group met the primary endpoint of a retinal change from baseline, with ellipsoid zone disruption associated with small retinal pigment detachment seen on SD-OCT in one eye at day 90. However, re-examination of the subject’s screening images clearly showed that he had this pre-existing abnormality, visible in the images, noted by the masked reading center, but not flagged as a screen fail. Therefore, this subject should have been excluded from the study, but was inadvertently enrolled. There were no other retinal or ophthalmic findings for this subject, and the observed abnormality was in a single eye. Thus, no tafenoquine subject met the primary endpoint in the per-protocol population. One subject in the placebo group had unilateral ellipsoid zone disruption that fulfilled the criteria for the primary endpoint.

Benign, reversible vortex keratopathy has been reported in previous tafenoquine studies following high doses and/or long-term therapy [9, 10]. In clinical trials with single-dose tafenoquine 300 mg in P. vivax patients, there was one case out of 107 subjects (0.9%) of unilateral vortex keratopathy, and this is included on the product labelling as an uncommon potential adverse reaction [12]. In this study, although vortex keratopathy was reported in one tafenoquine subject, the subject had a LASIK scar with calcium deposits that was observed at both visits which was inaccurately coded to keratopathy. Thus, there were no confirmed vortex keratopathy reported in this study.

No serious or severe adverse events were reported. There was a higher percentage of participants reporting nausea in the tafenoquine group; otherwise there were no clinically meaningful differences between the two treatment arms in the nature, frequency or severity of adverse events reported, including ophthalmic adverse events. There were no unexpected clinical laboratory abnormalities.

One subject in the tafenoquine arm reported mydriasis and photophobia, which were self-limiting and required no intervention. These adverse events were temporally associated with tafenoquine administration (on the tafenoquine dosing day and day 2). As similar events of this nature are recognized with other quinoline antimalarial drugs, tafenoquine causality cannot be excluded, and consequently, tafenoquine product labelling includes photophobia as a potential adverse event [12].

The primary limitation of this study is that it was conducted in adult healthy individuals without pre-existing eye conditions and so does not reflect the patient population likely to receive tafenoquine. This was necessary to fully investigate the pharmacodynamic potential of tafenoquine to cause retinal effects in the absence of significant confounding factors. However, it does not exclude the possibility of an effect of tafenoquine on the retina in individuals with pre-existing retinal damage or in the presence of P. vivax malaria. Ophthalmic assessments using clinical examination and digital fundus photography were included in phase IIb and phase III studies in P. vivax radical cure, with mild retinal hypo/hyperpigmentation observed in two out of 102 patients (2.0%) treated with tafenoquine and one out of 56 (1.8%) treated with primaquine [1,2,3]. There are also potential methodological limitations, as although the methods used were sensitive to detect retinal changes, further investigations, such as microperimetry and multifocal electroretinography, were not included.

In conclusion, consistent with data from the phase IIb and phase III studies evaluating a single 300-mg dose of tafenoquine for the radical cure of P. vivax malaria [1,2,3], there was no indication of any clinically important short-term ocular risk from the use of 300-mg tafenoquine single-dose treatment in this placebo-controlled study in adult healthy volunteers. These findings are reassuring given that tafenoquine will be used in resource-poor areas, where access to ophthalmological care is often limited.