Rapid decline in the susceptibility of Plasmodium falciparum to dihydroartemisinin–piperaquine in the south of Vietnam

Over the past 30 years there have been substantial reductions in malaria mortality and morbidity across the Greater Mekong sub-region. This has been driven largely by increased deployment of artemisinin-based combination therapy (ACT) for falciparum malaria which have replaced the monotherapies which successively fell to resistance. In Vietnam, the number of clinical cases of malaria decreased from 1,672,000 with 4650 deaths in 1991 to 19,252 with 3 deaths in 2015 [1]. Dihydroartemisinin-piperaquine (DP) was adopted in 2005 as first-line treatment and is the most extensively used anti-malarial drug in Vietnam. The success of the National Malaria Control Programme (NMCP) led to the initiation of “Malaria Elimination by the year of 2025,” a project endorsed with strong political commitment. Neighbouring countries have also set malaria elimination targets for the near future. These laudable objectives are now threatened by the emergence and spread of artemisinin resistant Plasmodium falciparum. First recognized in Western Cambodia 10 years ago, artemisinin resistant malaria parasites are now prevalent across the Greater Mekong subregion from the Myanmar-India border in the west to the coast of Vietnam in the East [2, 3]. To characterize the extent of artemisinin resistance and its effects on therapeutic responses to ACT, the Global Plan for Artemisinin Resistance Containment [4] recommended regular in vivo and in vitro assessments. This study reports on the results of sequential studies conducted in the southern provinces Binh Phuoc, Ninh Thuan, and Gia Lai of Vietnam.

Between August 2011 and December 2015, 489 patients with uncomplicated P. falciparum malaria were recruited: 253 in Binh Phuoc; 188 in Ninh Thuan; and 48 in Gia Lai. The 42-day follow-up schedule was completed by 423 (87%) subjects (Fig. 2). Adults (>15 years of age) represented 77% (375/489) of the patients, and 95% (471/489) were febrile upon admission (Table 1). The proportion of patients with a microscopy detected parasitaemia on day 3 (D3) post-DP treatment was highest in the Binh Phuoc site and increased progressively with time: 38, 41, 54, and 57% in 2012, 2013, 2014, and 2015, respectively (P < 0.001) (Fig. 3). In Ninh Thuan, the percentage remained low (<10%), and the differences were not significant (P = 0.29). In Gia Lai, patients were only recruited during 2011 and 2015 The D3 positivity rates were 0% in Iato versus 55% in Krong Pa (P < 0.001). The parasite clearance half-life in Binh Phuoc increased from 3.75 h in 2011 to 6.60 h in 2015 (P < 0.001) (Tables 2, 3, 4; Fig. 3). One early treatment failure occurred at Binh Phuoc (Table 2). This patient was febrile with a positive blood smear on day 3, but recovered fully by the end of treatment. Treatment failure rates were highest in Binh Phuoc, and increased significantly over time (P < 0.001) as follows: 0% in 2012 and 2013 and 7 and 26% in 2014 and 2015, respectively (Table 2). There were only two late treatment failures in Ninh Thuan (Table 3) and none in Gia Lai (Table 4).

Samples for evaluation of mutations were available for 481 subjects. K13-propeller mutations (n = 151, 31%) were as follows: 109, C580Y; 24, Y493H; 10, P553L; 7, I543T and 1 V568G (Additional File 2). Baseline parasitaemia was associated with the K13-propeller mutation (P value = 0.37 [linear regression, adjusted for site]). The median (IQR) parasite clearance half-life in patients with a K13-propeller mutation was 6.59 h (5.14–7.42 h) compared with 2.05 h (1.66–2.90 h) in subjects without the mutation [P < 0.001 (linear regression, adjusted for site)]. Half-lives were longer in subjects with a C580Y mutation compared with those with Y493H mutations, although the difference was not statistically significant (median 6.70 and 5.42 h, P = 0.07). Among the 17 subjects with PCR-corrected treatment failure in Binh Phuoc, 15 had a K13-propeller mutation (all C580Y), one did not have the mutation, and one did not provide a sample.

Ex vivo responses to DHA were measured in 64% (9/14) of P. falciparum isolates collected from recrudescent patients in 2015 by RSA and all were resistant to DHA (survival rate >1), with a median survival rate of 14.29% (IQR 4.76%, 23.51%). The 9 isolates were also tested in the PSA and all were resistant to piperaquine, with survival rate ≥10% (Fig. 4). The median survival rate after exposure to 200 nM of piperaquine was 20.72% (IQR 8.28%, 30.07%). The RSA and PSA were also carried out on isolates from two patients who experienced an ACPR response to DP.

The exoE415G mutation, which has been reported associated with increased tolerance to piperaquine [12] were found in 93% (13/14) isolates collected from recrudescent patients in 2015. Due to limited DNA amount of the recrudescent samples, plasmepsin 2 amplification was measured in 71% (10/14) of the samples. Eight of them (80%) had multi-copies of this gene (Fig. 5).

The widespread use of artemisinin and its derivatives in Vietnam has been a major factor underlying the marked reduction in malaria mortality and morbidity in the country. Until recently the nationally recommended ACT, dihydroartemisinin-piperaquine, retained excellent efficacy throughout the country.

Since the first recognition of artemisinin resistance in Cambodia, studies have been conducted to assess whether resistance had emerged in or spread to Vietnam, and what impact that would have on treatment efficacy. In 2010–2011, a randomized controlled clinical trial of 166 patients with uncomplicated P. falciparum malaria in the Binh Phuoc area showed that the proportion of patients with parasite clearance times >72 h ranged from 22 to 25%, depending on the dose of artesunate given (2 or 4 mg/kg/day). The corresponding median parasite clearance half-lives (PC_1/2) for were <5 h. Both indicated a high prevalence of slow parasite clearance. Although there were two ETFs (4%) and one late clinical failure (2%) in the lower dose artesunate (2 mg/kg/day) group, the PCR-corrected ACPR rates in the three groups were 94, 100, and 100% [14, 15]. This showed that artemisinin resistance was present in Vietnam but had not yet affected the efficacy of the first-line treatment.

This study reports longitudinal clinical, parasitological and molecular surveillance over a five year period from malaria endemic areas in the south of the country close to Eastern Cambodia. In Binh Phuoc province between 2012 and 2015 the proportion of patients with parasites still present on day 3 after starting DP treatment has progressively increased from 38 to 57%, with a corresponding reduction in cure rates from 0 to 26% (PCR corrected as per protocol). It provides unequivocal evidence of worsening artemisinin resistance and it shows that this has recently been associated with a rapid deterioration in the efficacy of DP. This strongly suggests that piperaquine resistance has arisen upon a background of artemisinin resistance reported recently from Cambodia [16]. The in vitro and molecular studies provided support for this; all patients who failed treatment were infected with K13 mutated parasites, and the parasites from recrudescent infections had median in vitro PSA survival rate of 39.2% compared with 0.17% in the patients who were cured with DP. Artemisinin-resistant K13 mutants with ex vivo PSA survival rates ≥10% were associated with a 32-fold higher risk of recrudescence [11]. It is notable that all 14 recrudescent patients in 2015–2016 were infected with the C580Y PfK13 genotype, which may indicate a recent hard selective sweep by a successful multi-drug resistant lineage.

While piperaquine was effective the DP combination provided high cure rates, but the reduced anti-malarial effect associated with artemisinin resistance places greater selection pressure on the partner drug, and when resistance to that occurs, the combination predictably fails. The detection of markers for piperaquine and DP treatment failure in the recrudescent patients also supported this. Among all the tested recrudescent parasites collected in 2015–2016, 93% (13/14) of them have exoE415G mutations and 80% (8/10) have multi copies of the plasmepsin 2 gene. It may indicate that resistance to piperaquine occurred. There has been insufficient attention to the protection of new anti-malarial medicines. Strategies such as deployment of multiple first line therapies are needed when effective alternatives are available [17]. In Vietnam, alternatives to DP are limited.

These findings present policymakers in Vietnam with a dilemma. The currently recommended second-line regimen combines quinine with doxycycline. But this regimen is not well tolerated and adherence to the required 7-day regimen is likely to be poor. The combination of artesunate with mefloquine may serve as an alternative. Vietnam used mefloquine both as a monotherapy and in an ACT (artesunate-mefloquine) since the 1990s. In 2005, foreign financial support ceased, and the NMCP of Vietnam, therefore, switched to locally produced DP as first-line treatment for falciparum malaria. In 2002–2003, a 2-day regimen of DHA-mefloquine (300 mg DHA at 0 h, then 300 mg DHA plus 750 mg mefloquine at 24 h) was evaluated at Binh Thuan and Binh Phuoc provinces and found to be 88.6% efficacious with a 42-day follow-up period. However, the limitation of the study was that it was not PCR corrected for differentiating rescrudescence from reinfection, and it is quite likely that a total dose of artesunate (12 mg/kg) plus mefloquine (25 mg/kg) given over 3 days would have been more effective. A 3-day artesunate-mefloquine regimen is presently being used in Cambodia in areas of DP treatment failure with cure rates of >99% (Chanaki et al. pers. comm.). A disadvantage of the mefloquine containing regimens is that they are not as well tolerated as DP [18‐20]. Atovaquone–proguanil (ATQ–PG) was suggested as a non-artemisinin containing anti-malarial treatment option. ATQ–PG was assessed in a small trial conducted in 2003 by the NMCP of Vietnam (unpublished data) and the overall cure rate (CI) was 86% (76–98%) [21]. This is still lower than the 95% benchmark recommended by WHO. Recently, an in vitro study provided evidence that ATQ–PG may be a useful stopgap for DP failed cases [22], but a previous study has also shown that this combination remains susceptible to developing resistance [23]. Furthermore, because of the high costs of this drug ATQ–PG has not also been considered as an alternative treatment by the Vietnam NMCP.

Artemether-lumefantrine (AL, Coartem®) may be another treatment option, but no clinical trial to assess AL has been conducted in Vietnam over the last 20 years. A pilot trial was conducted in 2003 that resulted in 100% (45/45) APCR rate with 28 day follow-up (personal communication). However, Vietnam’s NMCP has not considered AL as viable treatment alternative to DP due to financial constraints. The future may belong to newer anti-malarials, such as cipargamin (KAE 609), KAF 156, and artefenomel (OZ 439), but these are in phase II of development [24] and will likely not reach the market until 2020.

The authors conclude that there is no current medicine available to Vietnam’s national treatment guideline committee that meets the WHO criteria of being, “based on an average cure rate of >95%, as assessed in clinical trials.” There is an urgent requirement to evaluate alternative treatments with other available forms of ACT and to develop new anti-malarial drugs given standard courses of ACT are failing.