Plasmodium falciparum Recrudescence Two Years after Treatment of an Uncomplicated Infection without Return to an Area Where Malaria Is Endemic

ABSTRACT

We report evidence, confirmed by the lack of travel activity outside of France and genetic diversity analysis using polymorphic microsatellite markers, that Plasmodium falciparum malaria infection effectively treated with an artemisinin-based combination can remain dormant and relapse during pregnancy at least 2 years after treatment.

TEXT

In 2015, 4,750 imported malaria cases were reported in France, with a mean of 6 days between the end of travel to an area where malaria is endemic and diagnosis (Centre National de Référence du Paludisme, personal communication, B. Pradines). Declaration of symptoms associated with Plasmodium falciparum generally does not occur beyond 2 months after returning from a country where malaria is endemic (1). However, some cases of late detection of P. falciparum infections have been reported for up to 9 years (2). We describe a molecular-documented case of P. falciparum recrudescence 2 years after an uncomplicated infection was first treated.

A woman in her 30s was admitted with febrile presentation to an emergency unit in Bordeaux, France, in 2014. Originally from Cameroon, she moved to France during her childhood and returned to her native country for 10 days in 2014, returning to France 2 weeks before hospital admission. She was under mefloquine prophylaxis during her stay but did not use appropriate protective measures against mosquito bites. She complained of abdominal pain, followed by fever (39°C), headache, and vomiting. She was diagnosed by microscopy with P. falciparum (1.6% parasitemia), hospitalized in the infectious diseases department, and treated with dihydroartemisinin-piperaquine 320 mg/40 mg for 3 days (Eurartesim). Drug intake was monitored by nurses, and the routine parasitemia controls by blood smear showed no parasites at days 3, 7, 14, and 21 after treatment. She was examined 9 and 30 days after admission and had no complaints or signs of any disease.

In 2016, the woman was admitted at 24 weeks of pregnancy to the gynecology department of the same hospital in Bordeaux, presenting with fever, vomiting, and asthenia. Since her last malaria manifestation in 2014, the patient did not travel to an area where malaria was endemic and did not receive any blood transfusions. Nevertheless, she tested positive for malaria by microscopy, with 0.75% parasitemia by P. falciparum, including 50% late trophozoites. The unexpected diagnosis was confirmed in Bordeaux by real-time PCR using the FTD malaria diagnostic test and by the Centre National de Référence du Paludisme in Marseille by real-time PCR using species-specific primers and probes against the aquaglyceroporin gene of P. falciparum. No other infection was documented. She did not support a first intake of atovaquone-proguanil (Malarone) and was perfused with a single dose of intravenous artesunate, followed by a consecutive 3-day dihydroartemisinin-piperaquine oral treatment. The patient recovered well and delivered her baby under normal conditions.

To confirm the recrudescence of P. falciparum infection, the genetic diversity of blood samples from 2014 and 2016 was assessed using multilocus variable-number tandem-repeat analysis. These samples were genotyped using eight polymorphic microsatellite markers specific for P. falciparum (poly α, TA109, TA1, TA81, TA42, ARA2, PfPK2, and Pfg377). Microsatellite amplification was performed as previously described (3). To determine lengths, PCR products were analyzed using an ABI 3130XL capillary sequencer (Applied Biosystems, Foster City, CA). Microsatellite allele length was determined using internal size standards (GeneScan 500 LIZ size standard, Applied Biosystems) via GeneMapper v4.0 software (Applied Biosystems). Diversity of the P. falciparum samples was analyzed by comparing the length variation of polymorphic microsatellite markers between the 2014 and 2016 samples (Table 1). The two samples were successfully genotyped using the eight specific polymorphic microsatellite markers. The number of alleles per microsatellite marker varied from 1 to 4 for the 2014 sample; only 1 allele per microsatellite marker (i.e., a unique genotypic profile) was found in the 2016 sample. The initial P. falciparum population contained at least four parasitic clones, whereas a single clone was found in the 2016 sample. This clone contained 7 identical alleles from the initial P. falciparum infection of 2014. The allele length was different for the TA1 locus; a 3-bp difference for locus TA1 can define a distinct clone. Only one clone was identified for locus TA1 in the initial isolate. However, allele differences between the initial and recrudescent parasites were already shown, probably due to undetectable parasites present in the initial isolate (4). The eight microsatellite markers were genotyped three times for each sample to ensure reproducibility of the results. Very low levels of polymorphisms were found between the P. falciparum populations obtained from the patient in 2014 and 2016. The isolate clonality for the eight microsatellites, the identity of the fragment sizes for seven microsatellites, and the absence of travel to an area where malaria is endemic suggested that the clone from 2016 was recrudescent from the strain of 2014 and did not result from a new infection by a mosquito bite.

TABLE 1.

Genotyping of two samples obtained from a patient from 2014 and 2016 using eight polymorphic microsatellite markers specific for P. falciparum

Sample yr	Microsatellite allele lengtha (bp) for:
	PfPK2	TA109	Poly α	ARA2	TA42	Pfg377	TA1	TA81
2014	162.21	166.56	156.44	67.72	191.43	101.68	169.26	121.73
	168.21	169.58	159.65	73.60	206.12	107.40		136.52
	177.03	181.51	168.83
	191.60		174.94
2016	177.03	166.56	159.65	73.59	191.38	107.41	166.26	136.63

^aAlleles present in both samples are shown in boldface.

The ex vivo susceptibility to antimalarial drugs was evaluated for the two samples as previously described (5). Additionally, the K13 propeller gene involved in artemisinin resistance, pfmdr1 (P. falciparum multidrug resistance 1) gene, and pfcrt (P. falciparum chloroquine resistance transporter) gene involved in quinoline resistance were sequenced as previously described (6). The copy number amplification of pfmdr1, involved in mefloquine resistance, was assessed as previously described (7). To explore piperaquine resistance, plasmepsin II copy number amplification and single nucleotide polymorphisms in pfcrt (codons 145 and 350) and in the exonuclease gene (PF3D7_1362500, codon 415) were assessed (8–11). The sample from 2014 was found to be highly resistant to piperaquine 174 nM and mefloquine 90.9 nM and weakly resistant to chloroquine 126 nM and doxycycline 37 μM (Table 2) in vitro. The sample from 2016 failed to cultivate. Sequencing of the two samples revealed wild-type 72-76 haplotype (CVMNK) and no mutations at codons 145 (F145) and 350 (C350) for the pfcrt gene, wild-type haplotype for the pfmdr1 gene (N86, Y184, S1034, N1046, and D1246 alleles), wild-type exonuclease gene, and wild-type K13. The wild-type 72-76 haplotype may explain the weak in vitro chloroquine resistance. The two samples also showed only one copy for pfmdr1 and plasmepsin II genes.

TABLE 2.

In vitro susceptibility to standard antimalarial drugs of the initial P. falciparum isolate from 2014 versus P. falciparum W2 clone tested with the same plate batch

Drug	IC50a		Resistance cutoff
	2014 Isolate	W2 clone (SD)
Piperaquine (nM)	174	53.8 (9.0)	135
Dihydroartemisinin (nM)	6.3	2.6 (0.6)	10.5
Artesunate (nM)	2.7	1.5 (0.3)	10.5
Chloroquine (nM)	126	483 (30)	100
Quinine (nM)	434	388 (30)	800
Monodesethylamodiaquine (nM)	38.4	92 (16)	80
Lumefantrine (nM)	18.6	0.93 (0.39)	150
Pyronaridine (nM)	24.8	21.0 (2.8)	60
Mefloquine (nM)	90.9	25.5 (3.2)	30
Doxycycline (μM)	37.0	11.3 (2.1)	35

^aIC₅₀s of W2 are means of 7 independent experiments. Threshold values for reduced in vitro susceptibility or resistance were estimated previously (5, 20).

In most P. falciparum infections, the incubation period between an infected mosquito bite and symptomatic malaria is ∼9 to 30 days. Clinical manifestations of P. falciparum malaria 2 to 9 years after infection were previously described (2, 12, 13). Generally, recurrence of P. falciparum infection in Europe affects (i) migrants from areas where malaria is endemic in whom malaria parasites were not initially detected (1); (ii) patients with immunocompromised conditions, such as pregnancy or coinfection with HIV or sickle cell disease (13, 14); and (iii) patients undergoing chemoprophylaxis (5, 15).

In our clinical case, malaria symptoms with fever developed in a pregnant patient 2 years after exposure to malaria during a trip to Cameroon in 2014 and treatment of a P. falciparum infection with dihydroartemisinin-piperaquine. This patient was infected in 2016 with the same strain as that found in 2014 and, more specifically, with one P. falciparum clone identified in the initial polyclonic sample of 2014. In contrast to Plasmodium vivax and Plasmodium ovale infections, P. falciparum has no hypnozoite form that causes relapse after several years. A specific localization of P. falciparum parasites causing relapse is not known. One of the most likely hypotheses explaining recrudescence in this patient is that the dihydroartemisinin-piperaquine treatment in 2014 did not completely clear the infection, and one clone survived from the initial parasite population. The polyclonal parasites from 2014 showed reduced susceptibility to piperaquine but not to artemisinin derivatives. Additionally, no K13 mutation involved in artemisinin resistance and no polymorphism (mutation or copy number amplification) in Asia was observed in either of the samples. However, these data on artemisinin resistance should be considered with caution; the standard in vitro test we used is not well adapted to monitor resistance to artemisinin derivatives, and K13 mutation is not yet associated with artemisinin resistance in Africa, in contrast to that in Asia. Additionally, the molecular markers involved in piperaquine resistance in Asia are not yet validated in Africa. Recent publications reported that parasites collected from African patients still parasitemic at day 3 and beyond after artemether-lumefantrine treatment did not carry mutation on K13 (16). In our case, the patient originated from Cameroon and reported many malarial infections during her childhood, thus leading to the development of partial immunity against malaria parasitemia and symptoms. Indeed, indigenous people acquire a naturally partial and unstable protective immunity against symptoms related to P. falciparum, called premunity, which is mainly associated with repeated infections (17). In the case of this Cameroonian patient, premunity acquired during childhood probably prevented the onset of severe clinical symptoms for 2 years. In 2016, the pregnancy of the patient may have caused decreased immunity, resulting in elevated parasitemia- and malaria-associated symptoms. The pregnancy state creates an antigenic variability in the placenta (18), and the switching of parasite surface antigens allows malaria parasites to escape the immune response, resulting in an increase in parasitemia. This clinical case of P. falciparum malaria, manifested 2 years after exposure and after the first-treated uncomplicated infection, raises questions about the chronicity and activation of P. falciparum infection. The presence of late trophozoites (50%), which are not normally detected in the peripheral blood of patients, suggests altered sequestration. Recent studies suggested that P. falciparum infection may induce dormant stages, such as persisting nidus of sequestered parasites described in the placenta (19). Some recurrences of P. falciparum malaria may be caused by dormant erythrocytic forms that are newly activated as a consequence of decreased immunity.

Therefore, P. falciparum infection should not be excluded in patients with symptoms of malaria, even a long time after travel to an area where malaria is endemic, especially when patients are under mild immunosuppression, as occurs with pregnancy.

Ethics statement.

Bio-banking of human clinical samples used for malaria diagnostics and secondary uses for scientific purposes is possible as long as the corresponding patients are informed and have not indicated any objections. This requirement was fulfilled here by giving verbal information to the patient, and no immediate or delayed patient opposition was reported to the hospital clinicians. Informed consent was not required for this study because the sampling procedures and testing are part of the French national recommendations for the care and surveillance of malaria.