Plasmodium knowlesi malaria in Vietnam: some clarifications

A recently published comment on a report of Plasmodium knowlesi infections in Vietnam states that this may not accurately represent the situation in the study area because the PCR primers used may cross-hybridize with Plasmodium vivax. Nevertheless, P. knowlesi infections have been confirmed by sequencing. In addition, a neighbour-joining tree based on the 18S S-Type SSUrRNA gene shows that the Vietnamese samples clearly cluster with the P. knowlesi isolates identified in Malaysia and are distinct from the corresponding P. vivax sequences. All samples came from asymptomatic individuals who did not consult for fever during the months preceding or following the survey, indicating that asymptomatic P. knowlesi infections occur in this population, although this does not exclude the occurrence of symptomatic cases. Large-scale studies to determine the extent and the epidemiology of P. knowlesi malaria in Vietnam are further needed.


Introduction
Cox-Singh has recently published a comment on a paper recently published in the Malaria Journal [1], in which the presence of Plasmodium knowlesi infections in a Raglai community living in a forested area of central Vietnam is reported. In her comments, Cox-Singh states that this report may not accurately represent the P. knowlesi situation in the study area [2]. Such argument is based on the possibility of cross-hybridization of the primers used with Plasmodium vivax.

Discussion
Cross-hybridization of the primers used with P. vivax is a well known problem and Cox-Singh fails to mention that all P. knowlesi samples detected by PCR were cloned and sequenced in order to confirm the presence of this parasite [1]. Indeed, as already reported, several individuals initially positive by PCR for P. knowlesi were subsequently not confirmed by sequencing due to unspecific reactions with human DNA [1], indicating an inherent limitation of the PCR technique originally published by the Sarawack group [3]. This is further confirmed by the cross-hybridization with P. vivax recently published by Imwong et al [4].
Both the accession numbers of sequences used by Van den Eede et al and their similarity with the corresponding P. knowlesi 18S S-Type SSUrRNA gene sequences available on genbank were provided [1]. Additional evidence against the cross-hybridization hypothesized by Cox-Singh is provided by the neighbour-joining tree based on the 18S S-Type SSUrRNA gene showing that the five Vietnamese samples (FJ160750, FJ160751, FJ160752, FJ160753 and FJ871986) clearly cluster with the P. knowlesi isolates identified in Malaysia and are distinct from the corresponding P. vivax sequences ( Figure 1).
Cox-Singh also stated that Van den Eede et al [1] failed to discover symptomatic P. knowlesi cases in the human population. This is a misinterpretation as the objective of the study had never been to look for symptomatic P. knowlesi cases. Indeed, as explained in the initial section, screening for P. knowlesi infections was carried out after finding an unexpectedly high number of Plasmodium malariae infections (mainly as mixed infections), when analysing blood samples collected during the 2004 cross-sectional survey using species-specific PCR [1]. In addition, the local conditions (environment, vector, host and human behaviour) were similar to those previously described by Cox-Singh for the natural transmission of P. knowlesi [5]. Therefore, the aim of the study carried out in Vietnam was not to describe the prevalence or the epidemiology of P. knowlesi, but rather confirm its presence in the study area. Contrary to what has been stated by Cox-Singh, the conclusion by Van den Eede et al was not of a low P. knowlesi prevalence rather that probably this infection is relatively common given the finding of one child being still infected at one year interval.
Anopheles dirus s.s., a member of the Anopheles leucospherus group, is the main malaria vector in Central Vietnam, and one specimen infected with P. knowlesi has been recently reported from the neighbouring province of Khan Hoa [6]. This is not surprising since the southern half of Vietnam belongs to the geographical range of the long-tailed and pig-tailed macaques, the natural hosts of P. knowlesi [5]. Macaques are present in the area where the study reported by Van den Eede et al has been carried out and are in close contact with humans who often keep them as pets. Plasmodium knowlesi-infected monkeys have not been reported yet in Vietnam, but this does not exclude the possibility of a natural transmission cycle. At the time when the first large human focus of P. knowlesi infections in Malaysian Borneo was published in 2004, neither the vector nor the natural host infected with P. knowlesi had been found [3].
Samples collected in Vietnam on symptomatic cases were not genotyped. Nevertheless, samples collected during the two cross-sectional surveys and identified as having a P. knowlesi infection came from individuals who were asymptomatic at the time of the survey and did not consult for fever during the months preceding or following the survey. This is an indication that asymptomatic P. knowlesi infections occur in this population, without excluding the presence of symptomatic cases. This also supports the hypothesis of a P. knowlesi human reservoir of infections raised by Singh et al in 2004 [3].

Conclusions
The three human P. knowlesi infections reported from central Vietnam are just "case reports", without any pretence of describing the epidemiology of this simian species. Nevertheless, these findings, i.e. P. knowlesi occurring in young children and as asymptomatic infection, contribute to the body of knowledge on P. knowlesi epidemiology in South East Asia and cannot be easily dismissed as non-representative. Large-scale studies to determine the extent and the epidemiology of P. knowlesi malaria in Vietnam are further needed.

Conflict of interests
The authors declare that they have no competing interests. Figure 1 Neighbour-joining tree based on the 18S SSUrRNA gene sequences of P. knowlesi, P. vivax and other plasmodia. The tree was constructed with Treecon [7]. Bootstrap percentages based on 1,000 replicates are shown on the branches (if > 75%).