Plasmodium knowlesi is the most common cause of malaria admissions to Kudat District Hospital and affects all ages from young children to the elderly. Although the greater proportion of males, particularly among adults, is likely to be consistent with occupational forest or plantation exposure as a risk factor for knowlesi malaria, the wide age distribution suggests that this is not the only determinant of transmission. Furthermore, this report of two family clusters, one of which had not travelled outside Kudat town, suggests that transmission may be occurring close to or inside people’s homes. These findings differ from those reported from two studies in Sarawak, where a smaller proportion of knowlesi malaria cases occurred in children (8/121 [6.6%] in one study
 and 9/106 [8.5%] in the other
), and no clustering of cases was reported despite the presence of communal longhouses in the study areas
[4, 10]. Epidemiological risk factors of knowlesi malaria have not been previously investigated in either of these areas and require further evaluation.
It was notable that in contrast to the age-distribution of other malaria species, 25% of all P. knowlesi cases occurred in adults over 50 years of age. This is consistent with a lack of past immunity to P. knowlesi in this age group and/or a relatively recent increase in the risk of exposure to this species in the Kudat region. It may also relate to greater forest exposure among older individuals, with farmers and plantation workers over-represented among this age group
. The finding that adult females with knowlesi malaria are older than adult males is consistent with a previous report
 and requires further investigation, but may relate to less forest and/or vector exposure among younger females. The older age of female adults with P. knowlesi may also explain the finding from two previous studies that female patients are at greater risk of severe disease
In Kapit, Sarawak, An. latens was recently identified as the primary P. knowlesi vector, with the mosquitoes biting both monkeys and humans, and four (0.4%) infected mosquitoes found in forest and farm locations
. Although 126 (11.7%) An. latens were found in longhouses, none were infected, suggesting transmission occurs away from homes
. A subsequent study found evidence that in Kapit P. knowlesi remains a zoonosis, based on an extremely high prevalence (87%) of P. knowlesi in long-tailed macaques, and sequencing of the csp gene and mtDNA showing a greater number of P. knowlesi genotypes per monkey infection than human infection, with genotypes common to both hosts and no genotype exclusive to either
In Kudat, the wide age distribution and clustering of cases suggest that the P. knowlesi vector may be biting humans close to or inside people’s houses. Previously An. balabacensis was known to be the primary vector of human malaria in Sabah
, with Anopheles flavirostris also identified as a potent P. falciparum vector on Banggi Island
. Anopheles balabacensis have been shown to readily bite monkeys at the canopy level
, and a P. knowlesi-infected An balabacensis was recently found in Ranau
. Of significance for potential human-to-human malaria transmission, An. balabacensis have been shown to exhibit “learning behaviour” in relation to host preference, and may also demonstrate “habitat loyalty” by returning to previous feeding sites
. Furthermore, An. balabacensis has been shown to be a highly efficient vector
, and is closely related to Anopheles dirus, the primary malaria vector in the Mekong region, which maintains high levels of malaria endemicity at low population densities
, is capable of adapting to deforestation
, and in Southern Vietnam was found to be positive for P. knowlesi DNA
More recently however, An. donaldi was shown to have replaced An. balabacensis as the primary species in the Kinabatangan region of Sabah
. The peak outdoor biting time for this mosquito occurs from 18.00-19.00 hours, when humans are often outside their homes, and indoor biting occurs throughout the night. Although sporozoites were detected in An. donaldi in the Kinabatangan region
, species identification was not performed and further studies are required to determine the role of this species in the transmission of knowlesi malaria.
A strong association was demonstrated between knowlesi malaria cases and rainfall, with a lag time of three to five months. Anopheles mosquitoes depend on rainfall to provide aquatic breeding sites, and An. balabacensis and An. donaldi have both been shown to have increased parity rates during the months of greatest rainfall
. An association between rainfall and falciparum and vivax malaria prevalence has been well documented
[25–29], and can be used to assist with malaria early warning systems and prediction of epidemics
[30, 31]. This association should also be considered when planning for control of knowlesi malaria. However, alternative explanations for seasonal variation in knowlesi malaria such as other climatic variables and seasonal plantation work were not explored in this study and require further investigation.
Although the numbers of P. vivax and P. falciparum in this study were small, and could only be assessed over a three-year period, the number of P. vivax cases increased during the study period along with P. knowlesi, while a modest decrease was seen in P. falciparum cases. Larger prospective studies will be required to further evaluate the interaction between malaria species, which may have implications for malaria control.
This study had several limitations. First, the retrospective design did not allow the collection of information on where patients acquired their malaria infection. This prevented detailed assessment of forest exposure as a risk factor for knowlesi malaria, and may also have influenced the association between malaria cases and rainfall recorded at Kudat Meteorological Station. In particular, information was lacking on which patients came from Banggi Island, which, although only a short distance from Kudat town, may represent a different geographical setting with different transmission dynamics. Second, identification of case clusters was based only on patients with the same family name recorded in microscopy books, and this may have led to an underestimation of case clustering. Third, only a small number of patients with microscopy-diagnosed P. falciparum or P. vivax had PCR performed, and microscopy reports were therefore relied on for analyses involving these species. It is possible that some of the patients diagnosed by microscopy as P. falciparum or P. vivax had P. knowlesi. However, despite the potential for this to dilute age-related differences among species, a significant difference in age distribution was noted between P. knowlesi and microscopy-diagnosed P .falciparum or P. vivax, making the estimates of age-related differences conservative. Finally, this study involved the use, prior to June 2011, of a nested PCR assay (using Pmk8-Pmk9 primers) that has been associated with cross-reactivity with P. vivax DNA, and false-positive PCR findings of mixed P. knowlesi/P. vivax in true P. vivax mono-infections
. In this study this may account for the one third of patients diagnosed by microscopy with P. vivax mono-infection prior to June 2011, and subsequently found to have mixed P. vivax/P. knowlesi infections by PCR. Cross-reactivity with P. vivax isolates has not been reported with the real-time PCR assay
, and it was notable that after the introduction of this method there were no P. knowlesi/P. vivax mixed infections diagnosed by PCR despite an increase in microscopy-diagnosed P. vivax mixed and mono-infections over this time.
Malaysia has made impressive progress in reducing malaria prevalence in recent years; however, P. knowlesi cases appear to be increasing, with the species now accounting for the majority of malaria admissions to Kudat Hospital as well as other district hospitals throughout Sabah and Sarawak
[3, 4, 9]. The simian host reservoir of this zoonotic species presents particular challenges for malaria control, and will hinder the progress of malaria elimination in Malaysia. In addition, the increasing dominance of P. knowlesi and the possibility of transmission occurring close to or inside people’s homes may increase the possibility of a switch to the human host. Prospective studies on epidemiological risk factors, vectors and transmission dynamics of P. knowlesi in Sabah, including the possibility of human-to-human transmission, are required in order to develop strategies for knowlesi malaria control.