A systematic review of the efficacy of a single dose artemisinin–naphthoquine in treating uncomplicated malaria

Studies on the assessment of efficacy of single dose ASNQ in treating uncomplicated malaria were searched in electronic databases such as MEDLINE, EMBASE, CINHAL and the Cochrane Library. For ongoing and unpublished trials, the websites of WHO, the clinical trials database, and the drug manufacturer were checked. The reference sections of the selected studies and relevant reviews were also checked for the possibility of any additional papers. The search was limited to human studies, published in English, French and Chinese languages until May 2015. Medical subject headings (MeSH) terms and text words were (artemisinin–naphthoquine OR ARCO) AND (treatment success OR treatment failure OR efficacy OR tolerability OR safety) AND (malaria OR uncomplicated malaria OR asymptomatic malaria OR Plasmodium OR falciparum OR vivax).

Studies following the PICOS criteria [15] were included.

Two authors independently screened the title and abstract yielded from the electronic search. Any discrepancies between the two authors were resolved by consensus. The two authors individually collected information from each included study using the piloted data extraction form. Information collected included characteristics of participants, study design, characteristics of the experimental drug (dosage, route of administration, brand of ASNQ), confirmation of malaria infection, specific speciation, duration of follow-up and the reported clinical outcomes. For studies with overlapping of study population, the one that provided the most comprehensive data was used.

The methodological quality of the included studies was assessed using the risk of bias tool applied for the Cochrane systematic reviews [16]. The six domains for the risk of bias: random sequence generation, allocation concealment, blinding of outcome assessment, incomplete outcome data adequately addressed, free of suggestion of selective outcome reporting and other sources of potential bias addressed. Any discrepancies between the two authors were resolved by discussion. Analysis was done on an intention-to-treat (ITT) basis, whenever possible.

Sensitivity analysis was not done because of the limited number of studies. Data analysis was performed with RevMan 5·3 [17]. The protocol of the current review is registered with PROSPERO [18].

Only five RCTs and three single-arm studies were identified for the current review. The characteristics of the included RCTs are provided in Additional file 2. Of the five RCTs in the current review, one study each was carried out in Benin [21], Indonesia [22], Nigeria [23], Papua New Guinea (PNG) [19], and Uganda [20]. In the present review, single dose ASNQ was compared to AL in three studies [20, 21, 23], CQ plus SP (CQSP) in one study [19] and dihydroartemisinin–piperaquine (DHP) in another study [22]. One single-arm study was done in China [6], Myanmar [11] and Sudan [10] (Table 1). All five RCTs assessed both efficacy and safety outcomes. Most of the RCTs (80 %; 4/5) reported the clinical and laboratory parameters of the participants [19, 21‐23]. Less than half of the RCTs (40 %; 2/5) provided PCR-adjusted parasitaemia [21, 22]. Two RCTs [19, 21] and two single-arm studies [10, 11] were funded by the drug manufacturer, Kunming Pharmaceutical Corp (KPC). Two RCTs revealed that the authors had no conflicts of interest [22] and/or the drug manufacturer had no role in the planning, protocol design and execution of the study [19].

Overall, most of the RCTs (80 %, 4/5) included in the current review had ‘low risk of bias’ as they met five of the six domains assessed [19, 21‐23]. These clinical trials differed in sample size from 97 [23] to 401 participants [22]. The majority of RCTs in this review (60 %, 3/5) were blinded [20‐22] or were open-label studies [19, 22, 23]. This implied both the research investigator and the patient could be aware of the type of treatment that the patient was receiving. Only one study [22] provided the sample size calculation, showing adequate power to detect significant differences between the two treatment arms of the study. Two RCTs were superiority trials, assessing whether ASNQ was more effective than DHP [22] or AL [23]. All these five RCTS reported ITT. While the ITT approach is not ideal, it is considered to be the most appropriate approach for superiority trials since the ITT principle implies a conservative effect on the outcome of the trial [34]. Two clinical trials also reported per protocol analysis [22, 23].

As the RCTs included in the present review did not compare ASNQ with the same anti-malarial drugs, it was difficult to do a pooled analysis. At day 28, a pooled analysis of two RCTs (n = 271) showed a similar efficacy between ASNQ single dose (94.2 %, 130/138) and AL (94 %, 125/133) (RR: 0.99, 95 % CI 0.96–1.02; I-square value: 0 %) [21, 23]. Of note, parasitaemia in the Nigeria study was not confirmed by PCR [23]. Also, a comparable efficacy was reported between ASNQ single dose (95.5 %, 192/201) and DHP (92.5 %, 186/201) (RR: 1.03, 95 % CI 0.98–1.08) [22] on PCR-confirmed parasitaemia, and between ASNQ single dose (94.1 %, 48/51) and CQSP (87.8 %, 43/49) on PCR-unconfirmed parasitaemia (RR: 1.07, 95 % CI 0.95–1.22) [19] (Fig. 2).

At day 42, there was a similar efficacy between ASNQ (90 %, 181/201) and DHP (90 %, 180/200) on PCR-confirmed parasitaemia (RR: 1.0, 95 % CI 0.94–1.07) [22] and between ASNQ (98.2 %, 111/113) and AL (98.2 %, 110/112) on PCR-confirmed parasitaemia (RR: 1.0, 95 % CI 0.97–1.04) [19] (Fig. 3). Due to a paucity of data, stratify analysis by malaria parasite speciation was not possible. All single-arm studies were assessed on patients with P. falciparum malaria.

The mean parasite clearance time (PCT) was comparable between ASNQ (28 ± 11.7) and DHP (25.5 ± 12.2) (MD: 2.5, 95 % CI 0.16–4.64) [22]. The proportion of parasite clearance in the first three days is shown in Table 3. Among adults with malaria, the odds of reduction in parasite clearance within the initial 24 h in the ASNQ group was 1.2 times faster than that in the CQSP [19] or DHP [22]. Almost all parasitaemia was cleared by day 3 [19, 21‐23].

Fever clearance time (FCT) was reported in two RCTs; a comparable FCT was shown between ASNQ and DHP (MD: 1.7, 95 % CI −0.05 to 3.45) [20] or ASNQ and CQSP (MD: −3.6, 95 % CI −7.42 to 0.22) [19].

Only one RCT reported this data. The gametocytaemia by day 3 was reduced from 67.3 to 18.6 % in ASNQ and from 70.3 to 17 % in DHP (p > 0.05) [22]. A single-arm study reported that no gametocytes were detected in the blood smears examined [10].

In all five RCTs, AEs were recorded and compared between treatment groups (i.e., single dose ASNQ group vs comparator group). Table 4 provides the list of AEs in the included RCTs. Due to inconsistency in reporting it was difficult to undertake the pooled estimates of AE incidence. The most commonly reported AEs, which were assumed to be drug-related were nausea, vomiting, itchiness, transient deafness, and skin rashes.

Per protocol analysis, all patients who were not available (e.g., those withdrawn, lost to follow-up) were removed from the denominator, and this was done in two RCTs [22, 23]. At day 28, one RCT reported a comparable efficacy between ASNQ (87 %, 47/54) and AL (81.4 %, 37/43) on PCR-unconfirmed parasitaemia (p = 0.88) [23]. At day 42, another RCT reported a comparable efficacy between ASNQ and DHP (96.3 vs. 97.3 %; p: 0.56) on PCR-confirmed parasitaemia [22]. Due to the limited number of studies, sensitivity analysis was not performed. Publication bias was not investigated as the minimum recommended number of studies required to perform this is ten [16].

Overall, the findings could provide some evidence that there was a comparable efficacy between ASNQ single dose and comparators such as CQSP, DHP and AL. Moreover, the efficacy estimates of ASNQ were consistent at day 28 and at day 42. Using both day 28 and day 42 in this analysis was relevant because a 28-day follow-up captured the majority of treatment failures with drugs inclusive of artemisinin derivatives, while the longer follow-up for 42 days was optimal for these drugs [35]. RCTs, as well as single-arm studies, showed ≥95 % of therapeutic efficacy, which is a desirable level recommended for a new drug by the WHO [36].

A pooled analysis of studies assessing artesunate–amodiaquine including a single dose regimen in treating uncomplicated malaria documented that rapid P. falciparum clearance continued to be achieved in sub-Saharan African patients treated with ACT [37].

The current analysis showed that ASNQ single dose had a property of rapid reduction in parasite biomass within the initial 24 h. In fact, the rate at which treatment clears parasites within the first few days is the most useful practical test for ACT. This is because early response to treatment relies predominantly on the parasite response to artemisinin, independent of whether parasites are later cleared permanently through the combination of longer-lived companion drug and the host’s immune response [38].

Although ACT reduced the density of gametocytaemia and the proportion of infected mosquitoes, sub-microscopic levels of gametocytes (which were present in a significant number of patients after treatment) appeared to be sufficient to drive post-treatment transmission [39]. The detection of gametocytes by microscopy represents only the ‘tip of the iceberg’. Therefore, relying only on gametocyte carriage as an indicator can result in an overestimation of the effects of treatment regimens [19].

One of the key elements in any drug development and evaluation is the issue of safety for the population for whom the drug is intended [25]. The incidence of AEs in the included studies was presented in various ways due to inconsistency in the symptoms reported or actual variation in the symptoms attributed to the drug. This created difficulties in pooling the results. A study in Nigeria [28] compared three ASNQ treatment regimens in ≥15-year-old children (n = 121), comparing four tablets as a single dose (700 mg, group A), eight tablets as a double dose (1400 mg, group B) and eight tablets as a single dose (1400 mg, group C). Treatment success was not significantly different among these three groups. The occurrence of a blister on the lips was reported in a patient receiving the double dose of ASNQ, but not in patients receiving the single dose [29]. This can reflect potential problems of tolerability with the double dose. Further careful monitoring of ASNQ-related AE and dose response effects are needed. ASNQ-related SAEs were not reported in any of the studies included in the current review.

A study assessing a dose comparison of ASNQ in children with malaria showed similar efficacy and tolerability between single dose with water/with milk and twice daily dose regimens, indicating single dosing can be expected to improve better compliance [28]. A shorter regimen, such as single dose coformulated drug (ASNQ in this case), is usually better than a three-day regimen for ideal compliance (to the treatment schedule) [11]. Although any of the current regimens could be given for 1 day only the three-day regimens were considered. In the three-day regimen, the artemisinin component is present in the body for a short enough time to cover only two asexual parasite life cycles, except in the case of Plasmodium malariae. In each cycle, artemisinin and its derivatives would reduce parasite biomass by a factor of ~10,000 [19, 40].

A dose range study in PNG showed that a lower single ASNQ dose was associated with relatively frequent recurrence of P. vivax infections [28]. The recommended ACT (for instance, AL, artesunate-mefloquine) follows three-day treatment regimens. A clinical trial showed that the three-day ASNQ was non-inferior to AL for the treatment of uncomplicated falciparum malaria among young children in PNG and had greater efficacy than AL against vivax malaria [32, 33].

Any anti-malarial regimen must have robust evidence of both optimal efficacy for patient survival and an ability to reduce the potential for drug resistance [19]. The findings of this review, based on available data, showed evidence on the efficacy and tolerability of the ASNQ single dose. Continuous monitoring of parasite resistance to this ASNQ is still needed.

There are limitations to the present study. A small number of studies with small samples on the assessment of the ASNQ single dose were identified for the present review. In order to demonstrate the non-inferiority (by a margin of 5 %) of an alternative treatment (ASNQ in this case) to a current treatment known to be 95 % effective, at least 299 patients would be necessary in each study arm, with a one-sided test that has a statistical power of 80 % and a significance level of 2.5 % [41]. Using the best efficacy data (e.g., 100 % in ASNQ and 98.7 % in comparator) and the same 5 % non-inferiority margin, type 1 error probability and power as in the original sample size estimates, the total of 52 participants would be required to show non-inferiority for this primary endpoint [32]. If so, all five RCTs [19‐23] seemed to be sufficient to define the efficacy in two groups. However, the PNG study [19] had reported efficacy in the ASNQ (94 %) and comparator (80 %), hence, the required sample size was 348, which was about 3.5 times higher than the reported total of 100 participants.

The PCR-confirmation of parasites was done in only two RCTs in the current review [21, 22], hence, the chance of diagnostic bias is a concern. This was more prominent in areas where P. falciparum and P. vivax co-exist. For instance, the efficacy of ASNQ and CQSP were comparable in the PNG study where P. falciparum and P. vivax are equally important [19]. Thus, P. vivax was likely to be missed in PCR-uncorrected parasitaemia. This could be more pronounced in the CQSP group as parasite resistance to this combination drug is already confirmed in PNG [19]. The efficacy reported in the PNG study, therefore, needs to be interpreted in the light of these important biases. Moreover, PCR is useful for classification of re-infection or recrudescence. Hence, misclassification bias is a concern in the absence of PCR confirmation.

Three RCTs in this review were open-label trials [19, 22, 23]. As both the research investigator and the patient could be aware of the type of treatment this might lead to reporting bias of AEs in open-label trials [42]. Furthermore, true anti-malarial drug resistance may not be the same as treatment failure. If the administration of anti-malarial drug does not reach a sufficient blood concentration level, it will cause an inability to clear parasites in that malaria episode [1, 43]. Due to a lack of information on blood concentrations of the drugs administered, the current findings should be interpreted with a high degree of caution.

Due to limited data, the current review is unable to provide evidence for different age groups or different Plasmodium species. All single-arm studies were conducted on patients having P. falciparum [6, 10, 11]. It is likely that studies which were not published in the peer-reviewed literature might have been missed. However, not being in peer-reviewed literature raises questions about the quality of the study and its evidence. Another concern is that an assessment of therapeutic efficacy between ASNQ and CQSP [19] needs caution as CQ is a relatively old drug with known resistance by P. falciparum.

Based on available data, the current analysis shows comparable efficacy and safety profiles of the single dose ASNQ. Single dose ASNQ can improve treatment compliance and is simple and practicable. This is because this single dose can be delivered as directly observed treatment, immediately after confirmation of malaria [11]. Should mass drug administration be required in an outbreak of malaria, the (ASNQ) single dose could be a key element in community deployment of ACT. However, a single dose (even with a repeated dose on day 2) is insufficient as it can expose only one asexual cycle to the arteminisin component. Although a high cure rate could be attained from an ACT containing an effective partner drug, this will not provide sufficient protection from drug resistance [44]. In fact, malaria parasites will be eliminated, and the patient cured, if the plasma concentration of the free anti-malarial drug (ASNQ in this case) still exceeds the concentrations required to maintain the parasite multiplication rate below one (the minimum inhibitory concentration) [45]. The risk of the development of de novo resistance is increased with the duration of time the dividing asexual parasites are exposed to drugs. This makes the long-term risk of developing resistance a concern for single dose ACT [22]. As the ultimate goal of combination therapy is to prevent resistance developing, the usefulness of single dose ASNQ for treating uncomplicated malaria in endemic countries needs further evaluation.

The relatively long terminal half-life and wide therapeutic index of NQ [29] could contribute to better efficacy in three-day ASNQ regimen. This is because NQ can suppress the re-appearance of both P. falciparum and P. vivax for a longer period than piperaquine and, in particular lumefantrine. This is because the peak plasma concentration of benflumetol (i.e., the chemical name for lumefantrine) is attained slowly (8 h plus a 2-h absorption lag time), and the elimination half-life is estimated to be 4.5 days [46]. Although the resistance mechanisms of NQ are unknown, there is a potential for it to develop cross-resistance between NQ and CQ (both are 4-aminoquinoline derivatives) and a slight cross-resistance between NQ and artemisinin, lumefantrine or pyrimethamine (these drugs are unrelated to NQ both chemically and in terms of mechanism of action). The decreasing efficacy of amodiaquine single-agent treatment have been reported in some sub-Saharan African settings. If artemisinin resistance occurs there, it is expected to result soon in treatment failures, particularly in semi-immune populations such as children in high-transmission areas or patients of all ages in low transmission areas [37]. There is no study on single dose ASNQ with special groups such as pregnant women and infants and these are still needed.