Chloroquine (CQ) remains the recommended first-line treatment for Plasmodium vivax globally except for Indonesia, Papua New Guinea, the Solomon Islands and Vanuatu where widespread CQ resistance prompted a change in treatment policy 1 2 . Clinical monitoring of CQ efficacy is confounded by relapses derived from the activation of hypnozoites (dormant hepatic forms characteristic of P. vivax), making it difficult to categorize post-treatment episodes as recrudescences, re-infections or relapses 3 . Moreover measurement of the intrinsic sensitivity to CQ has been hampered by the difficulty to maintain P. vivax in culture. Nonetheless cases of CQ-resistant P. vivax have been reported from all continents where malaria is endemic, but never in pregnancy 1 . Previous clinical studies in the non pregnant Thai population where CQ was combined with primaquine did not show cases of highly suspect CQ resistant vivax 4 5 6 . Combined clinical and laboratory data from a closely monitored Karen pregnant woman on the western border of Thailand highly indicative of CQ resistance by molecular genotyping is presented in this report.

A 38 year-old pregnant Karen woman (blood group B, G6PD level normal and HIV negative) in her third pregnancy registered at a gestational age of 20⁺⁵ weeks (confirmed by abdominal ultrasound) 7 . She lived and worked in the forests on the Thai-Myanmar border and provided written informed consent to participate in a "postpartum susceptibility to malaria" study approved by the Ethics Committees of Oxford University (OxTREC (002_007) and Mahidol University (MUTM 2007-023) which included repeated blood sampling and publication of any data. She gave birth to a growth restricted live born singleton boy without congenital abnormality of 2,540 (±10) grams at a gestational age of 41⁺¹ weeks (<10^th percentile) on 19 December 2008. On the day of registration (D0) she presented with fever and was diagnosed with P. vivax malaria on the presence of asexual forms in peripheral blood (parasitaemia 9294/μL). Treatment was with CQ (25 mg base/kg total dose, Government Pharmaceutical Organization, Thailand) as per standard protocol. Over the following 227 days she was treated eight times for recurrent P. vivax parasitaemic episodes.

Parasite species diagnosis of each episode was determined by microscopic examination of Giemsa-stained blood films. This was later confirmed by nested PCR 8 for the episodes that occurred after D21, when suspicion about CQ resistant vivax was raised. Plasmodium vivax parasites were genotyped for three polymorphic markers: defined polymorphic regions in the genes encoding for the circumsporozoite surface protein (Pvcs), the merozoite surface protein 1 (Pvmsp1) or the merozoite surface protein 3α (Pvmsp3-α) 3 9 10 and for point mutations or copy number variation in the multi drug resistance gene 1, Pvmdr1 11 . Blood samples were also obtained from the placenta (including by mechanical extraction), umbilical cord and from the infant. Histopathologic analysis of a placenta biopsy was performed on Giemsa-stained cryo-sections.

The intrinsic ex vivo sensitivity assays were carried out for chloroquine, artesunate, piperaquine, mefloquine and amodiaquine on a leukocyte-depleted P. vivax isolate 12 . The pre-dosed plates employed for these assays were quality assured using a Plasmodium falciparum cloned line (PfK1). Dose response curves and IC₅₀ (50% inhibitory concentration) values were calculated by fitting the data to a sigmoidal inhibitory E-max pharmacodynamic model using WINNONLIN Ver 4.1 (Pharsight Corporation). The assays were duplicated and the data confirmed independently by three experienced microscopists.

Serum CQ and desethylchloroquine (DECQ) concentrations were measured using solid-phase extraction and liquid chromatography coupled with UV-detection 13 .

Over the course of 227 days this woman was seen 26 times (see Figure 1). The parasitological and laboratory findings for the samples collected from the pregnant woman are summarized in Figure 1. The persistence of P. vivax (parasite count 750/μl, non-symptomatic) on D7 post CQ administration prompted a second CQ course but parasites were still present in the blood on her next visit (D21) (parasite count 850/μl, non-symptomatic). Chloroquine was administered again (3^rd course). Symptoms and a higher parasite density (parasite count 27500/μL) were present on D31, and a complete supervised course of CQ was given. An asymptomatic P. vivax parasitaemia on D49 (parasite count 500/μL) pressed the administration of a fifth course of CQ. When vivax parasites re-appeared on D71 (parasite count 10,000/μL, symptomatic) the women was given a course of dihydroartemisinin (DHA) - piperaquine (PPQ), (6.75 mg/kg DHA and 54 mg/kg PPQ, 3 days, Holley Pharm, Peoples Republic China), the most effective treatment against uncomplicated vivax malaria in West Papua 14 , and parasites were undetectable by microscopy over the next five visits (D92-D119). A parasitaemic episode on D126 (parasite count 1000/μL, non-symptomatic) close to the expected delivery date was treated with CQ. But on delivery (D143) parasites were still detected in the mother's peripheral blood (parasite count 500/μl, non-symptomatic) and placenta (detected by PCR of placental blood). Placenta histologic findings were subtle: rare parasites, very mild inflammatory infiltrate and scant pigment deposition in the intervillous space. This infection was treated with artesunate (AS) (2 mg/kg/day, 7 days, Guilin, PRC) and parasites were not detected in blood smears taken on subsequent visits (D150 - D199), but on D206, 63 days postpartum an asymptomatic infection (parasite count 4,000/μl) was detected again. This was treated with DHA-PPQ and primaquine (PQ) (22.5 mg base/day for 14 days, GPO, Bangkok, Thailand). The parasites were promptly cleared and the blood smear remained negative until the end of the follow-up which was three months post-partum (D227). There were no parasites found in the cord blood or in the infant, who remained negative thereafter.

Combined CQ + DECQ serum concentrations obtained at three of the clinical episodes (D31 = 20.6, D71 = 10.8 and D143 = 7.9 ng/ml) indicated adequate drug exposure and were around the 15 ng/ml serum threshold considered therapeutic against CQ-susceptible parasites 15 . Blood samples were also collected on these days for ex vivo sensitivity assays but only one (D31) harboured parasites at a developmental stage suitable for meaningful analysis 12 . The ex vivo susceptibility profile clearly indicated P. vivax with reduced sensitivity to CQ (IC₅₀ = 270 nM) though not to the other anti-malarials tested (Figure 1).

PCR analysis confirmed the microscopic diagnoses and interestingly two additional sub-microscopic infections (D119, D199) were observed the week before they were detected microscopically. Genotyping revealed that the patient was successively infected by three distinct P. vivax populations (Figure 1 and Table 1). Parasites from D21-D49 were of similar genotype to those of D126-D143 but differed from the two distinct populations present on D71 and D206. All parasites had a single copy of Pvmdr1 and a T958M mutation but for those from D21-D49 and D126-D143 an additional Y976F mutation was observed.

This woman had multiple P. vivax episodes over a 227-day period during pregnancy and post-partum, characterized by repeated recurrences of P. vivax parasites following CQ treatment and resulting in a growth restricted neonate. The in vivo observations of P. vivax recurrences were associated to an ex vivo drug sensitivity assay showing a CQ IC₅₀ of 270 nM. This value is indicative of high-grade CQ vivax resistance as it exceeds recently published IC₅₀ medians for CQ of 37 nM for Thai (CQ sensitive) and 114 nM for Papuan (CQ resistant) isolates 12 . Indeed on two occasions (D31 and D71) each with a different genotype the parasitaemia increased substantially despite the presence of therapeutically adequate drug concentrations. The combined CQ+DECQ concentrations were similar to what has been reported for resistant vivax cases in a study in Myanmar 16 . It is interesting that the Pvmdr1 Y967F mutation was found in the parasites from two genotypically distinct episodes (Genotypes A and B). This putative marker of CQ resistant vivax was relatively rare (only 20%) in isolates from the Thai-Myanmar border 17 , but was present in 97% of the isolates from West Papua where CQ vivax resistance is widespread 11 . This could imply that there are at least two strains of chloroquine resistant parasites circulating in this area.

The CQ-resistant P. vivax in this Karen woman might have been acquired in Myanmar where she worked in the forests during her pregnancy, but she could have also acquired the infection in Thailand where she resides. Irrespective of its origin the fact that CQ-resistant vivax malaria appears to be spreading is a matter for concern.

Evidence based treatment guidelines for CQ resistant vivax infections are urgently needed to prevent repeated dosing of an ineffective drug, a practice known to enhance the selection and spread of resistant parasites. However, the choice of safe and effective drugs to replace chloroquine is very limited. Some artemisinin-combinations, such as DHA-PPQ or artesunate-mefloquine, are effective against P. vivax 1 , but artemisinins are contraindicated in the first trimester and primaquine, the sole drug effective against hypnozoites, is also contraindicated in pregnant women. In South East Asia, P. vivax is resistant to sulphadoxine-pyrimethamine 18 , a drug considered to be safe during pregnancy. Careful monitoring of the efficacy of CQ in the treatment of P. vivax in this region is needed especially in pregnant women to reduce the harmful effects on mothers and infants 19 .

Written informed consent was obtained from the patient for publication of this case report and any accompanying images. A copy of the written consent is available for review by the Editor-in-Chief of this journal.

The authors declare that they have no competing interests.

MJR, MEB, APP, and RMG were responsible for clinical work, design of the study, initiated and drafted the manuscript. BR and MLL did the microscopic examination, nested PCR and sensitivity assays. MI carried out the genotyping and drafted the manuscript. AM carried out the histopathology examination and drafted the manuscript. NL carried out the measurements of drug concentrations and drafted the manuscript. MJR, MEB, BR and GS combined all laboratory and clinical results into the manuscript. LR, GS, PS and FN participated in the design of the study, provided decisive comments and finalized the manuscript

All authors read and approved the final manuscript.