Efficacy of Primaquine for the Radical Cure of Plasmodium vivax Malaria in Northeast Myanmar and the Impact of CYP2D6 Genotypes

Abstract

Background.

Primaquine’s anti-relapse activity for Plasmodium vivax malaria depends on the host cytochrome P450 2D6 (CYP2D6) enzyme activity. This study aimed to assess the efficacy of primaquine for the radical cure of P. vivax malaria in the Greater Mekong Sub-region and determine the association of the post-primaquine treatment recurrences with CYP2D6 genotypes.

Methods.

In 2014–2017, 239 patients with uncomplicated P. vivax malaria in northeastern Myanmar were prospectively enrolled to receive a standard regimen of 3-day chloroquine and 14-day low-dose primaquine (total dose 3.5 mg/kg) as directly observed therapy. Patients were followed up bi-weekly for one year. A nested case-control study was designed to compare 49 patients with P. vivax recurrences and 49 non-recurrence patients. The CYP2D6 genotype-predicted activity score (AS) was used to estimate the unadjusted odds ratio (OR) of recurrence.

Results.

During the one-year follow-up, 54 recurrences were recorded with 65% occurring within 6 months post-primaquine treatment. Cumulative risk of recurrence was 17.3% [95% confidence interval (CI) 11.9–22.7%] at six months and 25.6% (95% CI: 19.3–31. 9%) at one year. Most recurrences were asymptomatic, with reduced parasitemia and gametocytemia. The CYP2D6*36+*10 tandem type with decreased function was the most prevalent allele (34.7%). Patients with an AS ≤ 1.25 had a significantly higher risk of recurrence (OR = 6.53, 95% CI: 2.0–21, P = 0.0007).

Conclusions.

Over 75% of vivax malaria patients in this region carried CYP2D6 alleles with AS ≤ 1.25, suggesting the consideration for higher primaquine doses.

Summary:

This study evaluated chloroquine and 14-day low-dose primaquine as directly observed therapy for Plasmodium vivax malaria in the Greater Mekong subregion and examined the association of low activity scores of host cytochrome P450 2D6 with increased risks of recurrences.

Plasmodium vivax is the most geographically widespread malaria parasite species outside Africa, putting 2.5 billion people at risk of infection [1]. In the Greater Mekong Sub-region (GMS), striving to eliminate malaria by 2030 [2], intensified control efforts resulted in a continued decline in malaria incidence. However, P. vivax persisted and has become the predominant parasite species, occasionally causing outbreaks along international borders [3]. A key feature of P. vivax resisting control measures is the formation of hypnozoites in the liver, which can cause relapses weeks to months later. The fact that more than 80% of the vivax malaria incidence could be due to relapses underscores the significance of anti-relapse therapy [4]. Currently, only two 8-aminoquinolines, primaquine (PQ) and tafenoquine, have been approved as anti-relapse drugs. In most P. vivax-endemic areas, chloroquine (CQ)/PQ remains the frontline radical cure for uncomplicated P. vivax malaria [5]. However, PQ can cause acute hemolysis in glucose-6-phosphate dehydrogenase (G6PD)-deficient individuals, leading to its underuse for P. vivax malaria treatment [6].

The effectiveness of PQ is also influenced by the host hepatic enzyme cytochrome P450 (CYP) 2D6 [7, 8], which is responsible for metabolizing >20% of currently used drugs [9]. The CYP2D6 gene is highly variable, with over 135 distinct star alleles. The CYP2D6 activity is phenotypically categorized as ultrarapid metabolizer (UM), extensive (normal) metabolizer (EM), intermediate metabolizer (IM), and poor metabolizer (PM), which can be predicted based on the diplotype of the CYP2D6 alleles. A value is assigned to each allele, and an activity score (AS) of the diplotype is obtained by combining the values of the two alleles. An AS of ≥ 2.25, 1.25–2.25, 0.25–1.0, and 0 corresponds to the phenotype UM, EM, IM, and PM, respectively [10]. Since PQ is a pro-drug and must be metabolized by the host CYP2D6 to active metabolites [8, 11], CYP2D6 activity is directly linked to PQ’s therapeutic activity.

In a P. vivax sporozoite challenge study, Bennet et al. reported the failures of PQ anti-relapse therapy in 2/25 human subjects who were CYP2D6 IM and PM [12]. In a nested-control study of P. vivax patients who completed the directly observed therapy (DOT) of PQ, impaired CYP2D6 activity was associated with a significantly increased risk of relapse [13]. In addition, there were reports of individual P. vivax cases where patients with impaired CYP2D6 activity suffered multiple relapses even though receiving adequate doses of PQ [7, 14–16]. In field conditions when the risk of reinfection could not be avoided, some studies identified the association of impaired CYP2D6 function with increased risk of post-PQ recurrences for radical cure of P. vivax malaria [17–20], whereas others found no association or association only for specific CYP2D6 alleles [21–23]. Reasons for such inconsistent findings could be multiple, such as non-DOT, limited CYP2D6 genotyping, and variations in the risk of reinfection. Therefore, the impact of CYP2D6 polymorphisms on PQ efficacy requires further evaluation to guide the deployment of PQ for P. vivax malaria elimination.

The CYP2D6 allele frequency varies substantially in global human populations. While CYP2D6*17 is more prevalent in African descendants, CYP2D6*10 is most common in Asian populations [9]. Surveys in GMS countries confirmed the 40–60% prevalence of the CYP2D6*10 allele [23–26]. Since the *10 allele was recently assigned an AS of 0.25 [10], a large proportion of the predicted phenotype would be IM and PM, which may confer an increased risk of PQ radical cure failure. With limited information about CYP2D6 alleles in Myanmar, this study aimed to determine the clinical efficacy of PQ radical cure in northeast Myanmar, where P. vivax was prevalent and persisted despite intensive control efforts. Using a nested-control study design, we determined whether P. vivax recurrences were associated with the host CYP2D6 metabolizer status.

METHODS

Patient Enrollment

This study was conducted at two settlements for internally displaced people (IDP) near Laiza town (97.56°E, 24.75°N), Kachin State, Myanmar, from April 2014 to September 2017. The settlements were established in 2010 to accommodate approximately 12000 IDP as the result of internal military conflict. This study screened patients aged 3 to 75 years with a fever history within the past 48 hours and malaria symptoms at two clinics serving the IDP settlements. Malaria was diagnosed by light microscopy of Giemsa-stained thin and thick blood smears. Parasite density was determined by two experienced microscopists [27]. Patients with uncomplicated vivax malaria were recruited, excluding pregnant and lactating women and those with underlying diseases. Written informed consent was obtained from adult patients and the guardians of minors. This study was approved by the Ethics Review Committees of Kunming Medical University, the Health Bureau of Kachin, and the Pennsylvania State University.

Treatment and Follow-up

Treatment of uncomplicated vivax malaria followed the local guidelines – three days of CQ (total 25 mg/kg) and 14 daily doses of PQ (0.25 mg/kg/day) [27, 28]. Both drugs were given as DOT. Per local standards, G6PD deficiency was not screened, but patients were reminded of the risk of PQ toxicity, and urine color was monitored. Participants without P. vivax recurrences within 28 days were actively followed every 2 weeks for an additional 11 months. Malaria symptoms were monitored at each follow-up, and finger-prick blood samples were collected for microscopic and molecular examinations. Patients were asked to return to the clinics if they felt sick during the follow-up period.

Parasite Identification

DNA was isolated from filter paper blood using a QIAamp DNA mircrokit (Qiagen, Germany). Plasmodium species at enrollments and recurrences were verified by nested PCR analysis [28]. Parasites were not genotyped, as this cannot discern relapse from reinfection.

CYP2D6 Genotyping

Single nucleotide polymorphisms (SNPs) in the CYP2D6 gene were identified by PCR and sequencing [15]. PCR amplification of the full-length CYP2D6 coding region was performed using the Phanta Max Super-Fidelity DNA polymerase P505d and the following cycling conditions: initial denaturing at 95°C for 3 min, 35 cycles of 95 °C for 15 s, 53 °C (58 °C for nested reactions) for 15 s, and 72 °C for 5 min, and final extension at 72 °C for 5 min. The PCR product was purified and sequenced using BigDye Terminator v3.1. The DNA sequence was assembled using the SeqMan program (DNASTAR, Madison, WI, USA), manually edited, and aligned with the reference sequence. The predicted phenotypes with AS of 0, 0.25–1.0, 1.25–2.25, and ≥ 2.25 are considered poor, intermediate, normal, and ultrarapid metabolizers, respectively.

CYP2D6 Copy Number

The CYP2D6 copy number variation (CNV) was determined by multiplex PCR amplification of three CYP2D genes (CYP2D6/CYP2D7/CYP2D8) as described [29]. The ratios of CYP2D6/CYP2D8 peak heights were used to assign the CYP2D6 copy number. Each PCR contained 0.4 units of KAPA2G™ Fast HotStart DNA polymerase (KAPA Biosystems), 2.4 μl of 5× PCR buffer, 0.4 μl of dimethyl sulfoxide, 0.16 μl of 10 mM dNTPs, 1 μl of genomic DNA (5–20 ng/μl), and one set of primers each at a final concentration of 0.5 μM. PCR cycling parameters were as follows: initial denaturation 2 min at 95°C followed by 22 cycles of 15 s at 95°C, 30 s at 60°C and 10 s at 72°C, and then a final extension at 72°C for 3 min. PCR products were separated on the 3730 DNA analyzer and quantified with GeneMapper® Version 4.0 (Applied Biosystems).

Data Analysis

Statistical analysis was performed using the GraphPad Prism 6 (GraphPad Software, California). The Mann-Whitney test was used to compare the density of asexual parasites between two groups. The log-rank (Mantel-Cox) test was used to compare the cumulative hazards between two age groups. A life table analysis was done to estimate the cumulative incidence rates of recurrence, adjusting for the loss to follow-up of patients. The Chi-square test was used to assess two CYP2D6 metabolizer phenotype proportions, while Fisher’s exact test was used to assess the CYP2D6 gene allele frequency. The odds ratio (OR) of recurrence and 95% confidence interval (CI) were estimated for those with a lower AS.

RESULTS

Study Cohort

This study screened 287 eligible malaria patients, and 239 satisfying the inclusion criteria were enrolled in the PQ efficacy study and actively followed for a year after CQ/PQ treatment as DOT (Fig. 1). All P. vivax infections were confirmed by PCR analysis. The enrolled subjects were 55.6% males, 3–75 years old (Table 1). P. vivax infections occurred more frequently in children, with 49.8% under 14 years. The 3–13 year age group had a slightly higher, albeit statistically insignificant, parasite density than the ≥14 age group (Table 1). At the time of recruitment, 92.5% of the patients had gametocytemia with a median density of 656/μl.

Figure 1.

The study flow chart.

Table 1.

Demographic and Clinical Features of P. vivax Patients

Patient characteristics	At enrollment (day 0)	At recurrence #	P values
Total no. of patients (% male)a	239 (55.6)	54 (68.5)	0.0942
Age in years [median (range)]b	13 (3–75)	14 (6–59)	0.6553
3–13 (N)a	120	26	0.8804
≥ 14 (N)a	119	28	0.8804
Mean temperature (°C) (range)b	38.7 (36.0–41.0)	36.5 (36.0–40.0)	< 0.0001
Asexual parasite density [/μl, median (range)] in age group*
3–13b	1442 (24–84160)	342 (32–61000)	0.0322
≥ 14b	1000 (24–28200)	414 (16–10100)	0.1281
Gametocytemic patients [%]a	92.5	85.2	0.1097
Gametocyte density [/μl, median (range)]b	656 (0–6080)	162 (0–12000)	0.0029

^aChi-square test.

^bMann-Whitney U test.

^*P = 0.4273 (at enrollment) and P = 0.7934 (at recurrence) for comparison between the two age groups.

^#Recurrence recorded during days 28 – 364.

All P. vivax patients at enrollment had at least one of the malaria-related symptoms (Table 2), which were mostly mild, with fever, chills, headache, and loss of appetite as the dominant symptoms. A few patients had more severe symptoms, including vomiting and convulsions.

Table 2.

Patients with Malaria-related Symptoms at Enrollment and Recurrence

No. (%) of patients with symptoms	At enrollment (n=239)	At recurrence (n=54)	P values
Fever (axillary temperature > 37.5 °C) [n,(%)]	220 (92.0)	9 (16.7)	< 0.0001 *
Chills	135 (56.5)	5 (9.3)	< 0.0001*
Headache	128 (53.6)	5 (9.3)	< 0.0001*
Loss of appetite	63 (26.4)	1 (1.9)	< 0.0001#
Vomiting	17 (7.1)	-
Joint pain	15 (6.3)	-
Dizziness	14 (5.9)	-
Nausea	9 (3.8)	-
Convulsions	3 (1.3)	-
Coughing	3 (1.3)	-

^*Chi-square test;

^#Fisher’s exact test

Recurrence of P. vivax Malaria During One Year of Follow-up

Of the 239 patients, 186 and 180 completed the 6- and 12-month follow-up, and 59 patients were lost to follow-up (Fig. 1). In total, 54 of the 180 patients who completed the one-year follow-up had recurrent P. vivax parasitemia, with 64.8% occurring within 6 months after treatment. The cumulative incidence rates of recurrent P. vivax infections were 17.3% [95% confidence interval (CI), 11.9–22.7%] at six months and 25.6% (95% CI, 19.3–31. 9%) at one year (Fig. 2). Of the patients with recurrences, 10 had more than one recurrent episode during the follow-up period.

Figure 2.

Cumulative incidence of recurrent parasitemia in P. vivax patients treated with chloroquine/primaquine. Dashed lines indicate 95% confidence interval bounds.

A higher proportion of male patients had recurrences. Age did not affect the risk and cumulative hazard of P. vivax recurrence: 26 in the 3–13-year group, compared to 28 in the ≥ 14-year group (Fig. 3). Asexual parasitemia at recurrence was much lower than those at enrollment (Table 1). Although a similarly high proportion of the recurrences (85.2%) were gametocytemic, the gametocyte density was significantly lower than at enrollment (Table 1).

Figure 3.

Kaplan-Meier’s survival curves for children 3–13 years and patients ≥14 years old. The two groups had no significant differences in the risk of recurrence (P = 0.3583, log-rank Mantel-Cox test). Dashed lines represent 95% confidence interval bounds.

Notably, only 9/54 (16.7%) subjects were ill when recurrent parasitemia was detected (Table 1). Compared to 92% of patients being febrile at enrollment, 16.7% of those with recurrent parasitemia had a fever (Table 1).

CYP2D6 Alleles in P. vivax Patients

For the nested case-control design, 49 recurrent patients (39 with single and 10 with multiple recurrences) were designated as cases, age- and sex-matched with 49 non-recurrence patients in the same study cohort as controls. The two groups did not differ in body weight (Fig. S1). Genotyping CYP2D6 genes in these patients detected seven CYP2D6 alleles with normal (*1, *2, and *35), decreased (*10 and *36+*10), and no (*15 and *69) function (Table 3). The CYP2D6*36+*10 tandem type was the predominant allele, accounting for 34.7% (Table 3). Among all alleles, 48.5% had decreased or no functions. The frequency of the *35 allele in the case group was significantly lower than in the control group (P = 0.0070).

Table 3.

Frequencies of CYP2D6 Alleles in P. vivax Patients

		Number (%) of alleles
Allele	Functional status	Total (N=98)	Case group (N=49)	Control group (N=49)	P values
*1	Normal function	21 (10.7)	13 (13.3)	8 (8.2)	0.3560*
*2	Normal function	33 (16.8)	16 (16.3)	17 (17.3)	1.0000*
*10	Decreased function	4 (2.0)	2 (2.0)	2 (2.0)	1.0000#
*15	No function	22 (11.2)	12 (12.2)	10 (10.2)	0.8215*
*35	Normal function	47 (24.0)	15 (15.3)	32 (32.7)	0.0070 *
*36+*10	Decreased function	68 (34.7)	39 (39.8)	29 (29.6)	0.1766*
*69	No function	1 (0.5)	1 (1.0)	0	1.0000#

^*Chi-square test,

^#Fisher’s exact test

Table 4 summarizes the CYP2D6 diplotypes and their predicted metabolizer phenotypes. The seven CYP2D6 alleles resulted in 18 diplotypes with a predicted AS ranging from 0.25 to 2.5 (Table 4). The most abundant diplotype was *35/*36+*10, reaching 29.6% of the sample. The CNV analysis identified ten individuals with CYP2D6 gene amplification. The predicted phenotypes were UM (9.2%), EM (63.3%), and IM (27.6%). Notably, 13 of the 62 patients with predicted normal genotypes had an AS of 2.0, whereas the remaining 49 had an AS of 1.25 (Table 4). Further, no patients with an AS of ≥2 suffered multiple recurrences.

Table 4.

Distribution of Cytochrome P450 2D6 Genotypes and Predicted Phenotypes in P. vivax Case (with One or Multiple Recurrences) and Control (No Recurrence) Patients

Genotype	Predicted activity score	Predicted metabolizer phenotype	Total number (%) of patients	Controls (no recurrence), # (%)	Cases # (%)
					With one recurrence	With multiple recurrences
*1×2/*36+*10×2	2.5	Ultrarapid	4 (4.1)	4 (8.2)	0	0
*2×2/*36+*10×2	2.5	Ultrarapid	1 (1.0)	1 (2.0)	0	0
*35×2/*36+*10×2	2.5	Ultrarapid	4 (4.1)	3 (6.1)	1 (2.6)	0
*1/*1	2	Normal	2 (2.0)	0	2 (5.1)	0
*1/*2	2	Normal	1 (1.0)	1 (2.0)	0	0
*1/*35	2	Normal	1 (1.0)	1 (2.0)	0	0
*2/*35	2	Normal	8 (8.2)	7 (14.3)	1 (2.6)	0
*15/*35×2	2	Normal	1 (1.0)	1 (2.0)	0	0
*2/*10	1.25	Normal	2 (2.0)	1 (2.0)	1 (2.6)	0
*1/*36+*10	1.25	Normal	11 (11.2)	2 (4.1)	8 (20.5)	1 (10.0)
*2/*36+*10	1.25	Normal	7 (7.1)	0	4 (10.3)	3 (30.0)
*35/*36+*10	1.25	Normal	29 (29.6)	18 (36.7)	9 (23.1)	2 (20.0)
*2/*15	1	Intermediate	13 (13.3)	7 (14.3)	4 (10.3)	2 (20.0)
*2/*69	1	Intermediate	1 (1.0)	0	0	1 (10.0)
*15/*35	1	Intermediate	4 (4.1)	2 (4.1)	2 (5.1)	0
*10/*36+*10	0.5	Intermediate	2 (2.0)	1 (2.0)	0	1 (10.0)
*36+*10/*36+*10	0.5	Intermediate	3 (3.1)	0	3 (7.7)	0
*15/*36+*10	0.25	Intermediate	4 (4.1)		4 (10.3)	0
Total # (N)			98		39	10

A comparison of the distribution of different genotypes between the case and control groups revealed a significant difference (χ²=12.029, P = 0.0073) (Table 5). Similarly, genotypes between the control and case group with single recurrences were significantly different (χ²=8.60, P=0.0351). Among the 27 patients with a genotypic AS of ≤ 1.0, 17 had recurrences. In contrast, for the 22 patients with an AS of ≥ 2.0, only four had recurrences. The OR of recurrence for patients with an AS of ≤ 1.0 was 2.07 (95% CI, 0.83–5.14) when patients with an AS of ≥ 1.25 were used as the reference. Remarkably, the OR of recurrence for patients with an AS of ≤ 1.25 increased to 6.53 (95% CI, 2.0–21) (Table 5). The timing of recurrences after the low-dose PQ therapy did not significantly correlate with the CYP2D6 genotype AS (P > 0.05, Fig. S2).

Table 5.

Association of Post-primaquine Recurrences with Predicted CYP2D6 Metabolizer Phenotypes

		Number (%) of phenotypes
Predicted metabolizer phenotype	Predicted activity score	Control group (N=49)	Case group (with single recurrence, N=39)	Case group (with multiple recurrences, N=10)
Ultrarapid	>2.25	8 (16.3)	1 (2.6)	0
Normal	2	10 (20.4)	3 (7.7)	0
Normal	1.25	21 (42.9)	22 (56.4)	6 (60.0)
Intermediate	0.25–1.0	10 (20.4)	13 (33.3)	4 (40.0)
P values*		χ2=12.03 P=0.0073a	χ2=8.60 P=0.0351b	χ2=5.55 P=0.1357c

^*Pearson’s Chi-square test:

^abetween the control group and total cases;

^bbetween the control group and single-recurrence cases;

^cbetween the control group and multiple-recurrence cases.

Note: For comparison between predicted activity scores of ≥2 and ≤1.25, χ²=5.286, P=0.0215.

DISCUSSION

Vivax malaria presents a formidable challenge for the GMS countries in their pursuit of regional malaria elimination. Our earlier work in northeast Myanmar detected a decline in CQ efficacy and potentially high-grade CQ resistance [28, 30]. This study evaluated the efficacy of a 14-day low-dose PQ administered via DOT for the radical cure against P. vivax relapses. The cumulative risk of recurrence at one year was 25.6%, comparable to the 28.2% reported in other low-dose PQ efficacy studies [31]. However, it was much higher than the 9.1% rate observed in our previous study [27]. This variance could stem from the 2016–2017 vivax malaria outbreaks in the IDP settlements [3, 32], which likely increased reinfection rates. Also, as most P. vivax relapses in the GMS occur in the first six months [27, 33], a significant portion of the detected recurrences in the latter half-year were likely reinfections. Given that over 60% of vivax cases in the GMS experience recurrences within a year [33], a conservative estimate is that low-dose PQ could mitigate 57% of vivax recurrences in the study area.

A worrisome finding is that more than 80% of the recurrences lacked symptoms, meaning that the treated clinical cases represented only a small fraction of the infected population. Moreover, 85.2% of the recurrences were gametocytemic, highlighting the existence of a large infectious reservoir, sustaining transmission and driving malaria outbreaks. Clearly, treating the clinical vivax cases alone is insufficient to prevent continuous transmission of P. vivax, necessitating supplementary strategies like mass drug administration.

PQ’s therapeutic efficacy for the radical cure of relapsing malaria is affected by many factors, including dosing, duration, adherence, and human genetics. DOT addresses adherence issues and enhances PQ efficacy [34]. However, studies in Indonesia demonstrated PQ treatment failures despite DOT and the absence of reinfection risks [35, 36]. After the initial discovery of the association of IM and PM CYP2D6 genotypes with PQ therapeutic failures [12], Baird et al. further established that P. vivax patients with a phenotypic AS of < 1.5 had 9.4 × odds of post-PQ relapse [13]. Our current study extends this finding, showing a correlation between lower CYP2D6 genotypic scores and higher recurrence likelihoods post-DOT with low-dose PQ in the GMS.

Since the PQ active metabolite(s) remains unclear, genotyping CYP2D6 star alleles and assessing the AS serve as proxies for PQ activation [14, 17, 18, 20]. However, most PQ efficacy and CYP2D6 studies only genotyped a limited number of CYP2D6 SNPs. Here, we adopted a sequencing approach to interrogate all SNPs in CYP2D6 among the Kachin ethnicity in northeast Myanmar. Our analysis revealed that the *36+*10 tandem type was predominant in the Kachin population, diverging from the high prevalence of the *10 allele observed in other earlier studies in the GMS [24, 25]. This disparity is likely due to the omission of *36-specific mutations in earlier studies. Since *36 allele also includes the *10 allele mutations (100C>T and 4180G>C), relying solely on these two mutations will overestimate the *10 allele frequency. This distinction is critical as *36 has a null function while *10 exhibits a decreased function. Our sequencing results are consistent with a recent study in southern Thailand, which also reported the prevalence of the *36+*10 tandem type (51.1%) [37].

Although CYP2D6 AS facilitates genotype-phenotype translation [38], its substrate-specific nature remains challenging. Baird et al. measured the CYP2D6 activity in metabolizing dextromethorphan to dextrorphan and found that the metabolic ratio is predictable of post-PQ treatment relapses and strongly correlated with the genotype-derived AS [13]. However, the AS system may serve only as a surrogate for PQ metabolizer activity, so further elucidation is needed. Especially with the downgrade of the *10 allele AS to 0.25 [10], the AS threshold for inferring the CYP2D6 phenotype in PQ metabolism should be refined. Using the AS of ≤1.0, we found that patients in the IM group had an OR of 2.07, similar to the findings for the temperate zone P. vivax infection in Korea [20]. However, a markedly higher OR of 6.53 was obtained for patients with an AS of 1.25 or less, suggesting impaired PQ metabolism in this subgroup. This has significant implications for PQ deployment. Although screening CYP2D6 in P. vivax malaria patients is impractical, knowledge of the CYP2D6 information in local populations of P. vivax-endemic areas can inform PQ therapy. Currently, the GMS nations use the low-dose PQ regimen for the radical cure of vivax malaria. Since >75% of the P. vivax patients had a CYP2D6 AS of ≤ 1.25 associated with significantly higher odds of recurrences after the low-dose PQ treatment, augmenting PQ doses may improve PQ’s efficacy in this population [13]. This approach has proven to overcome impaired CYP2D6 metabolism in two vivax malaria cases [39].

Our study has several limitations. Firstly, we did not genotype the parasites to determine the probability of true relapses in the recurrent cases. Secondly, patients after PQ treatment remained in the endemic area under the risk of reinfection. Besides, the study was undertaken when P. vivax transmission surged in the IDP community, thus increasing the chances of reinfection. Thirdly, we used patients’ CYP2D6 genotypes instead of measuring metabolic activity, which may not faithfully represent the CYP2D6 activity for metabolizing PQ. Finally, patients with no recurrences may have been due to the lack of hypnozoites at enrollment and radical cure. However, this would make our observed association between PQ failure and CYP2D6 genotypes more conservative.

Conclusions

While low-dose PQ demonstrates efficacy in northeastern Myanmar, a sizeable portion of recurrences occurred, which may be attributed to patients’ impaired CYP2D6 function. A large proportion (>75%) of vivax malaria patients had CYP2D6 genotype scores ≤1.25, suggesting impaired PQ-metabolizing activity and underlining potential benefits of higher PQ doses for the radical cure of P. vivax malaria in this region.