A 49-year-old man who had experienced fever and chills for half a day was admitted to the Shanglin County Hospital in Guangxi Province, China, on December 19, 2016. He showed additional symptoms including headache, body aches and cough. Upon inquiry of travel history, he informed the doctor that he had spent one year and three months in Ghana (8/15/2015–11/10/2016) and returned home 39 days ago. He explained that during his stay in Ghana, he had experienced two episodes of malaria (species unknown); his last episode of malaria was about half a year ago and both times he was self-treated with artemisinin drugs. On admission, he weighed of 70.2 kg, his axillary temperature was 38.0 °C, and his heart rate, blood pressure and respiratory rate were 92 beats/min, 91/60 mmHg and 20 breaths/min, respectively. With his recent travel history, venous blood was drawn for general hematology and blood chemistry analyses, and a drop of the blood was used to make a thin smear for malaria diagnosis by microscopy. Microscopic examination of the Giemsa-stained blood smear revealed P. vivax parasites. Blood tests showed increases in white blood cells (14.25 × 109/L; reference range 4–10 × 109), neutrophil ratio (85.0%; 43–76%), and C-reactive protein (142.29 mg/L). He was conscious and oriented to time, place, and person. He was not dehydrated, pale, or in respiratory distress. Since he did not have any severe symptoms, he was diagnosed as having uncomplicated vivax malaria. He spent three days in the county hospital and was given three days of oral CQ therapy (total 1550 mg). To help resolve the symptoms faster, he was also given intravenous (IV) injections of artesunate (total dose of 420 mg, initial dose of 120 mg, subsequently divided into 5 times of 60 mg at a 12 h interval). Meanwhile, after confirming that he is G6PD normal, an 8-day course of PQ (22.5 mg/day) was begun. Fever was cleared within one day and parasitemia was cleared within two days. The patient was discharged on day four with instructions for follow-up visits if symptoms reappear. Administration of the remaining five days of PQ was directly observed therapy (DOT) by local Center for Disease Control (CDC) personnel to ensure compliance. The second attack of febrile paroxysm occurred 58 days later on February 15, 2017, and he was admitted to the county hospital again with similar symptoms as the first attack and diagnosed with P. vivax malaria by microscopy. The patient had not left Guanxi Province during the 58 days between these two attacks. He was hospitalized for six days and was treated with the same CQ/PQ combination, together with eleven IV injections of artesunate (total dose of 660 mg at an interval of 12 h). Given that the doctor was not sure whether this re-occurrence was due to chloroquine resistance or potential mixed infection with P. falciparum, at discharge he was given additional three days of an artemisinin-based combination therapy (ACT), artesunate-amodiaquine, which has a different aminoquinoline drug. Both ACT and PQ were administered as DOT by local CDC staff. One hundred and thirteen days later, on June 8, 2017, he suffered a third attack of confirmed P. vivax malaria and hospitalized at the county hospital for six days. He received the same therapy as for the second attack including DOT of 8-day PQ. At home, he was further treated with three days of a different ACT, dihydroartemisinin-piperaquine. Eighty-eight days later, on September 4, 2017, he suffered a fourth attack of confirmed vivax malaria. This time he was not hospitalized, while the same CQ/PQ regimen together with three days of oral therapy of dihydroartemisinin-piperaquine was prescribed. All treatments were taken at home and supervised by local CDC staff. Despite the fact that, after returning from Ghana, this patient lived the entire time in a malaria-free area, he had the fifth attack of vivax malaria 232 days later, on April 24, 2018, 491 days from the first attack. He was admitted to the county hospital for three days and was given IV injection of artesunate six times at a 12 h interval (120 mg each at the first three injections, and 60 mg each at the three subsequent injections). PQ was not prescribed since it was judged not effective. Instead, he was treated with azithromycin (500 mg/day) for seven days. At discharge, he was also given three additional days of dihydroartemisinin-piperaquine. At the time of this interview, he had remained healthy for 330 days after this last episode of vivax malaria. Venous blood was collected at the time of diagnosis at the first, second, third, and fifth attacks. Blood samples were used for molecular diagnosis and genotyping at the Kunming Medical University laboratory. For each sample, total DNA was extracted from 0.2 mL of venous blood using the High Pure PCR Template Preparation Kit (Roche, Switzerland) following the manufacturer’s instruction and eluted in 100 μl of water. Plasmodium species were identified by nested PCR targeting the 18S rRNA genes using genus-specific and species-specific primers for P. falciparum, P. vivax, P. malariae and P. ovale []. The PCR results showed that all the samples were positive only for P. vivax (data not shown). To determine whether the relapses were caused by different parasite strains, we genotyped the polymorphic P. vivax merozoite surface protein (PvMSP) 3α gene by the nested PCR and restriction fragment length polymorphism (PCR/RFLP) methods described earlier []. PCR of PvMSP3α alone detected a similar band size for the first three attacks, but the PCR product from the fifth attack was smaller. Digestion of the PvMSP3α by HhaI showed the same restriction patterns for the first three attacks, whereas the fifth attack was clearly different, suggesting that the first three attacks were likely due to the same parasite strain, whereas the last attack was from a different parasite strain. Since the effectiveness of PQ for radical cure of vivax malaria is influenced by host CYP2D6 activity, we wanted to determine whether the failure of PQ in this case might be linked to CYP2D6 genotypes suggestive of poor metabolizer of PQ. The single nucleotide polymorphisms (SNPs) in CYP2D6 were determined by PCR amplification of the full-length CYP2D6 coding region using a high-fidelity enzyme and sequencing of the PCR products, similar to a method described earlier []. Primary PCR was performed using primers P1 (5′-CTGGCAGCACAGTCAACA-3′) and P2 (5′-TTTGTCTTCCGTTTTGGG-3′), while nested reactions were done with primers N1 (5′-ATAAGGGAAGGGTCACGC-3′) and N2 (5′-GGCAAGGGTAACTGACATCT-3′). The following PCR conditions were used: 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. PCR and sequencing of CYP2D6 detected mutations 214G > C, 221C > A, 223C > G, 227 T > C, 232G > C, 233A > C, 245A > G, 310G > T, 745C > G, 842 T > G, 1662G > C, 2851C > T, 3385A > C, 3585G > A, 3791C > T, 4181G > C, and 4482G > A, which are classified as *2A, and mutations 100C > T, 310G > T, 842 T > G, 1038C > T, 1662G > C, 2098A > G, 3385A > C, 3583A > G, 4125G > C, 4129C > G, 4132A > G, 4134 T > C, 4156C > T, 4157A > C, 4159G > C, 4165 T > G, 4167 T > C, 4168G > A, 4169C > G, 4170 T > C, 4173C > T, and 4181G > C, which are classified as *36. According to the CYP2D6 allele naming database (), the patient’s CYP2D6 genotype corresponds to a *2A/*36 allele variant. CYP2D6*2A is predicted to be functionally normal (score 1), but *36 is non-functional (score 0). Thus, the overall genotype activity score was 1, suggesting that this patient’s CYP2D6 was an impaired PQ metabolizer phenotype []. Real-time PCR was performed to determine the CYP2D6 gene copy number using a previously described method [], and the result showed that CYP2D6 gene in this patient was a single copy.