Malaria, a tropical disease caused by infection with single-celled parasites of the genus Plasmodium, is one of the most deadly parasitic diseases in the world. According to the World Malaria Report 2011, an estimated 3.3 billion people were at risk of malaria in 2010. Of this total, 2.1 billion people were at low risk whereas 1.2 billion people were at high risk (>one case per 1,000 population) and living mostly in the WHO African (47%) and Southeast Asian regions (37%) [1]. Malaria is the most common vector-borne parasitic disease in Malaysia, responsible for 20–40 deaths per year over the last decade, and it is estimated that 3% of total Malaysian population live in malaria risk areas. Around 80% of nationwide cases are found in Malaysian Borneo with 58% occurring in the northern state of Sabah and less than 20% of total malaria cases occurring in Peninsular Malaysia [2].

In Sabah, malaria is transmitted by various Anopheles mosquitoes, such as Anopheles balabacensis, Anopheles sundaicus and Anopheles flavirostris[3]. Traditionally, human malaria can be caused by four Plasmodium species, which include Plasmodium falciparum, Plasmodium vivax, Plasmodium malariae and Plasmodium ovale. These four Plasmodium species are well-recognized to cause human malaria worldwide [4]. Infection caused by P. falciparum is severe and may be fatal in the absence of prompt recognition of the disease and its complications, while the disease caused by P. vivax, P. ovale and P. malariae is usually milder and rarely fatal.

Recently, a simian malaria parasite, Plasmodium knowlesi, has been recognized as the fifth human malaria species by WHO [5]. Plasmodium knowlesi is a malaria parasite of Old World monkeys, which was formerly known to cause malaria only in macaques [6]. However, this parasitic infection is widespread in humans in Malaysia and has been reported from Malaysian Borneo [7–11] and Peninsular Malaysia [8, 12–14]. This infection has also been reported in Singapore [15–17], Thailand [18, 19], Myanmar [20, 21], Philippines [22], Indonesia [23, 24], Vietnam [25] and Cambodia [26], pinpointing that the incidence of P. knowlesi was generally high in Southeast Asian countries especially Malaysia. Unfortunately, this parasite is life-threatening and has caused mortality in humans [27, 28].

The diagnosis of P. knowlesi infection can be challenging for an inexperienced microscopist because the early trophozoite stages of this parasite are morphologically similar to those of P. falciparum[7, 29] whilst the later erythrocytic stages (i.e., late trophozoites, schizonts and gametocytes) of P. knowlesi are indistinguishable from P. malariae[7], including the presence of the typical band-forms. Due to these morphological similarities between P. knowlesi and P. malariae, naturally acquired P. knowlesi infections are often misdiagnosed via microscopy as P. malariae[7, 29].

This was evident when six out of 14 malaria deaths reported in Sabah during 2010–2011 [28] were actually caused by P. knowlesi (later diagnosis done with molecular tools) but were diagnosed initially as P. malariae infections via microscopy detection. Hence, in order to overcome the limitations of microscopy for the detection of malaria, molecular tools such as polymerase chain reaction (PCR) are necessary for accurate identification of P. knowlesi as well as other Plasmodium parasites in order to avoid mortality. This method has proven to be more specific and sensitive than conventional microscopy [30]. Recognizing this advantage, the utilization of molecular techniques in diagnosis and research in molecular epidemiology has become more common. Recently, in a study based on nested PCR assay, a total of 107 malaria cases (44% of 243 samples) have been reported in four districts in the interior division of western Sabah (i.e., Keningau, Tenom, Tambunan, Nabawan) [31].

However, information on malaria incidence in eastern districts of Sabah is scarce. Therefore, the aim of the present epidemiological study was to investigate the incidence and distribution of the Plasmodium species in these districts of Sabah, Malaysian Borneo based on a most recent developed hexaplex PCR system in order to obtain accurate epidemiological data which serves as an important foundation in the effectiveness of malaria control strategy. The straightforward single-step hexaplex PCR system (commercial name PlasmoNex™) targeting five human Plasmodium 18 s small subunit rRNA (ssu rRNA) was developed by Chew et al.[32]. This system can detect all five human Plasmodium species simultaneously as well as mixed infections in a single step PCR with high sensitivity and specificity. Therefore, this system was the method of choice for the detection of malaria parasites in this study instead of nested PCR because it is rapid and less labour-intensive.

Malaria is one of the most severe parasitic infections and it is also a major public health problem in many countries in Southeast Asia. In the 1980s, the prevalence of malaria drastically decreased and was eliminated in many areas of Peninsular Malaysia due to the efficient malaria control programme. However, malaria is still prevalent in Malaysian Borneo and among the ethnic minority groups [39]. Recently, significant reductions in malaria cases have been observed and achieved in Malaysia [1]. Malaysia is now working to be malaria-free in the peninsula by 2015 and in Malaysian Borneo by 2020 [1, 40]. However, recently, a significant increase (>ten-fold) of P. knowlesi cases was observed in Sabah between 2004 and 2011 and this trend threatens malaria elimination [41]. Since the early 2000s, it was reported that most districts of Sabah experienced an increase of P. knowlesi cases and it appeared to have begun in the south-west of Sabah (Interior division), followed by West Coast division and gradually progressed north-easterly (Kudat division). This increase appears to have begun initially in the Interior division and there was a steady increase of P. knowlesi cases from 2000 to 2011 in this region. West Coast division experienced an increase of P. knowlesi cases later, from 2001 to 2009. After that, the tip of Borneo, Kudat division experienced the most noticeable increase in P. knowlesi cases with more than 200 cases for the years 2009 and 2011. Plasmodium knowlesi infection has also been reported in eastern districts of Sabah (Sandakan and Tawau), although the cases reported in these two districts were fewer compared to other districts, it has been increasing since 2008 [41]. Findings of the present study showed that P. knowlesi cases were also detected in the districts of Sandakan division. This shows that P. knowlesi is generally prevalent in Sabah. Therefore, the main objective of this study is to investigate the incidence and distribution of P. knowlesi as well as other human Plasmodium species in eastern Sabah.

Hexaplex PCR (or PlasmoNex™) findings of the present study demonstrated that P. knowlesi were identified in 42 (23.6%) samples, however, microscopic examination did not show P. knowlesi in any samples examined. Thirty-five infections diagnosed as P. malariae by microscopic examination were found to be P. knowlesi. This finding was similar to the studies carried out by other researchers in Malaysian Borneo [7, 10, 28, 35], reporting that most microscopy-positive P. malariae infections were actually proven to be P. knowlesi infections when molecular approach was used. However, the transmission dynamics of this infection still remain unknown despite wide distribution of this parasite in Malaysian Borneo. The most prevalent Plasmodium species identified in the study was P. falciparum followed by P. vivax and P. knowlesi. It was reported that P. vivax was the most prevalent human malaria parasites in Malaysia [42], however, P. falciparum was found to be the most prevalent species in this study.

Conventional diagnostic technique, microscopic examination of asexual stages of Plasmodium on thin and thick blood films has been considered as the “gold standard” for the diagnosis of Plasmodium parasite infection because it is cost-effective and simple [43]. However, it can be rather challenging to diagnose mixed infections and infections with low level parasitaemia [44]. The diagnostic accuracy depends greatly on the experience of microscopists. Sensitivity of microscopy examination of malaria parasites can be affected by various factors such as the quality of staining, skills of the laboratory technician, expertise of microscopists and types of microscope used. In addition, microscopic identification of P. knowlesi is rather challenging due to its morphology, which is indistinguishable from P. falciparum and P. malariae. Underdiagnosis of P. knowlesi by microscopy examination was observed in the present study. Thus, molecular tool, such as PCR, is necessary to overcome the limitations of microscopy for malaria detection especially in generating epidemiology data.

The recently developed PlasmoNex™ [32], was used for the detection of malaria parasites and identification of the five human Plasmodium species in this study. Molecular methods based on DNA amplification have been used for malaria diagnosis since 1980s [45–49]. Several reports have shown that the DNA-based amplification methods had higher sensitivity (as low as 1 parasite/μl of blood) compared to microscopic examination of thin blood films, especially in cases of low parasitaemia or mixed infections [50–52]. Besides that, they are able to characterize each human malaria species [45, 53–55].

PlasmoNex™ is a single-step hexaplex PCR system [32], which is able to detect all five species of human malaria with sensitivity below 0.5 parasites/μl of blood sample. To date, nested PCR is the molecular gold standard for the diagnosis of malaria and species identification. This method has been proven to be more specific and sensitive than conventional microscopy [30]. However, malaria diagnosis and species identification of malaria by nested PCR or semi-nested PCR are time consuming. The time required for hexaplex PCR is only three hours compared to semi-nested PCR which requires about five hours and approximately 14 hours for nested PCR, including the time for PCR preparation and gel electrophoresis. Besides, there are some other advantages of hexaplex PCR over the available nested PCR as it is less labour intensive, less contamination and reduced usage of reagents and consumables. Hexaplex PCR is a single-step system whereas semi-nested PCR required at least two PCR reactions and for nested PCR at least six PCR reactions are needed for the identification of all five Plasmodium species. The fewer PCR reactions reduce the usage of reagents and consumables and are also less labour intensive. Mixed infection detected by single-step hexaplex PCR in this study indicated that this system is robust and capable of detecting at least up to two species level mixed infections.

Recently, P. knowlesi cases had been reported in the Interior division and north-eastern state of Sabah [10, 28, 35]. It is obvious that the prevalence of P. knowlesi in Sabah has increased markedly and studies reported on the discovery of this parasite by using molecular approaches have greatly increased recently. In the early 1990s, there was around 50,000 malaria cases reported in Sabah and a continuous decrease was observed from the year 2001 to 2003. There were 6,050 cases in 2001, and this number fell to 5,096 cases in 2002 and subsequently to 1,770 cases in 2003 [56]. Previous report also showed that the notification of P. falciparum and P. vivax in Sabah had drastically decreased over the past decade [41]. The notification of P. falciparum and P. vivax had decreased to around 600 for the year 2011, however, a significant increase of P. knowlesi had occurred following the reduction of P. falciparum and P. vivax in Sabah. It showed that when human malaria cases decreased, P. knowlesi cases were on the rise. It is postulated that environmental changes together with decreasing rates of P. falciparum and P. vivax were likely to have contributed to the rise of P. knowlesi in Sabah [41]. Human activities such as extensive deforestation due to road construction work, oil palm plantation, timber or eco-tourism that occur in Sabah have brought human, macaques and mosquito vector into close contact and this propagates successful transmission of this parasite. In the present study, P. knowlesi infection was detected in all four districts of Sandakan divisions with the highest incidence of this parasite in the Kinabatangan region.

The wide distribution of P. knowlesi in this region is not surprising as this region is surrounded by dense primary and secondary forests which act as suitable habitats for mosquito vectors (i.e., forest-dwelling Anopheles balabacensis) [3] and macaques (i.e., long-tailed macaques, M. fascicularis and pig-tailed macaques, M. nemestrina) which are the natural hosts for P. knowlesi in Sabah. Previously, An. balabacensis was well recognized as the predominant vector of human malaria in Sabah [36]. However, Anopheles donaldi was reported to have replaced An. balabacensis in the Kinabatangan region in 2005 [56]. The peak outdoor feeding time for An. donaldi mosquitoes occurs between 18:00 and 19:00. This is the period most people are being bitten by this mosquito when the adults are working outside and children are having outdoor activities. Although An. balabacensis was recognized as the vector of human malaria in Sabah, more studies should be carried out to investigate the current vector situation in this region. Perhaps, the situation of vectors in this region has changed due to recent development and mass forest clearance in the area. Nevertheless, communities should be informed to take necessary preventive measures against this parasite in order to prevent outbreaks of P. knowlesi infection.

Malaria patients who participated in the study comprised ages from young children to the elderly, however, this infection was more prevalent among adults especially the males. The epidemiological data showed that P. knowlesi malaria was diagnosed in male patients more often than in female patients, and only small proportion of knowlesi malaria cases occurred in children, this may be related to the limited outdoor or forest activities by this group of patients. These findings were in concordance with those reported from two studies in Sarawak [7, 9], where a smaller proportion of knowlesi malaria occurred in children and no clustering of cases was reported. Findings of these two studies suggested that P. knowlesi infection occurs outside people’s home and adults were the high risk group of this infection due to greater forest exposure, and it is postulated that human to human transmission did not occur. However, previous study postulated that the transmission of P. knowlesi may occur from human to human due to the discovery of family clustering cases [57]. Nevertheless, no clustering of P. knowlesi cases was found in the present study.

Microscopic findings showed that there were more gametocytes found in the P. falciparum and P. vivax samples compared to P. knowlesi samples. This could infer that infected humans were getting the knowlesi infection from mosquitoes that fed on monkeys and not from humans. Consistent with previous studies [9, 35, 57], patients with P. knowlesi were older than those with P. falciparum or P. vivax. The possible reason for the older age group being infected by P. knowlesi was discussed in a study done by William and his colleagues [41]. It was suggested that older individuals had greater forest exposure compared to the young. Patients aged between 21 to 30 years old were the group with the highest number of malaria cases detected in this study, possibly due to greater forest activity by this group, who either work or stay in the forest, such as timber and oil palm plantations workers, since the economic activities in Kinabatangan are the timber industry and oil palm plantations and the highest cases of P. knowlesi were detected in this district. Long working hours in the forest may increase the risk of exposure to the mosquito vectors, resulting in successful transmission of P. knowlesi. Despite these observations, epidemiological risk factor of P. knowlesi malaria still remains unclear, hence further investigations are required in order to develop successful control strategies for knowlesi malaria.

The symptoms of human P. knowlesi infection are almost similar to the other four human malaria species, however, it is more severe than the disease caused by P. vivax and more virulent because it has a short life-cycle of 24 hours [58] which leads to hyperparasitaemia enabling a fast progression of the disease. Malaria is frequently associated with a variety of haematological complications such as thrombocytopaenia and anaemia and both have high mortality rates [9, 35, 57, 59–61]. A healthy human platelet count ranges from 150,000 to 450,000 platelets per microlitre of blood, levels lower than this range is considered as thrombocytopaenia. It is one of the common features of acute malaria and this occurs in P. knowlesi[9, 15, 35, 62–64], P. falciparum[65, 66] and P. vivax[66] infections regardless of severity of infection.

The present study showed similar results with previous studies whereby thrombocytopaenia occurred in all species of human malaria parasites indicating that thrombocytopaenia is a common feature of malaria and its occurrence should increase the suspicion of malaria. However, there are several other causes of thrombocytopaenia such as leukaemia, hereditary syndromes, HIV-associated thrombocytopaenia and also dengue fever. Thrombocytopaenia can be caused by decreased platelet production or peripheral destruction. It is postulated that malarial antigen caused the lysis of immune complexes, which leads to sequestration of the injured platelet by macrophages in the spleen [67, 68]. However, the exact mechanism of thrombocytopaenia in malaria is not clear and requires further investigation. Findings of the present study demonstrated that most of the P. falciparum patients suffered severe and moderate thrombocytopaenia followed by P. knowlesi and P. vivax patients. This might explain the role of platelet activation in the pathogenesis of liver-attacking malaria. Platelets are produced by a process called thrombopoiesis, which occurs in the bone marrow by budding off from megakaryocytes. Thrombopoietin is a hormone that is secreted for the regulation of platelet production and megakaryocytes. This hormone is produced by the liver and kidneys. Hyperparasitaemia of P. falciparum and P. knowlesi parasites due to their shorter lifecycle may destruct the production of this hormone in the liver and thus decreased the production of platelet and caused thrombocytopaenia. However, this hypothesis requires further investigation. Many studies have shown that P. knowlesi patients suffered severe thrombocytopaenia [28, 35, 69, 70]. Findings of the present study were in parallel with the study done by Barber and her colleagues in 2011 [35] where platelet counts of patients were lower in P. knowlesi than in P. falciparum malaria. Besides, findings of the present study are also in line with the study conducted by Patel and his colleagues [66] which reported that patients with P. falciparum malaria were found to have lower platelet counts than patients with P. vivax malaria. Therefore, findings of present study demonstrated that the platelet count of P. knowlesi-infected patients were lower than those in P. falciparum-infected patients and P. vivax-infected patients, which, in turn, were lower than those in the uninfected patients and a platelet count below 75,500 cells/cmm should increase the suspicion of P. knowlesi infection. Nonetheless, it is suggested that in order to verify this hypothesis, more samples need to be included in future. It is believed that thrombocytopaenia is more common in P. falciparum malaria compared to P. vivax malaria. However, recent studies have shown the trend that thrombocytopaenia was found to be equally common in P. vivax malaria compared to P. falciparum[71–74]. A study conducted by George and Alexander [75] reported significant thrombocytopaenia in P. vivax malaria. The findings were in accordance with those in Brazil [76] and Qatar [77]. Although the relation between thrombocytopaenia and the species of malaria is unclear, it is a fact that thrombocytopaenia is universal in malaria-infected patients and should act as a significant indicator of malaria especially knowlesi malaria.

Anaemia is also another haematological complication that is usually associated with malaria especially in P. knowlesi and P. falciparum infections [35, 57, 78]. The present study also highlighted that anaemia was observed in all species of malaria patients including P. knowlesi patients indicating that anaemia is common in human malaria infections. However, malarial anaemia appears to be multifactorial, factors such as genetic disorders (e.g., thalassaemia), haemoglobinopathies, poor dietary intake and socio-economic causes may also contribute to anaemia. It was observed that anaemia is significantly higher in female patients (X² = 5.1; p = 0.025) compared to male patients in the present study. This might also be due to blood loss of female during menstruation. With regard to age, the present study demonstrated that the prevalence of anaemia was significantly higher (X² = 6.30; p = 0.012) in children compared to adults. This is in agreement with other previous studies where malaria-related anaemia is more severe in children rather than in adults [79, 80] due to increased demand of iron by the body as growth is rapid in children. In 2002, Praba-Egge and her colleagues [81] carried out experimental infection of non-human primates with P. knowlesi and Plasmodium cynomolgi. The study reported that anaemia was observed in P. knowlesi- infected monkeys after primary and/or repeated infection. Severe anaemia was notable in macaques after repeated infection by P. knowlesi[81]. This showed that anaemia was common in both P. knowlesi-infected humans or macaques. Malaria-related anaemia is a consequence of hyperparasitaemia and it has been associated with immune-mediated lysis of erythrocytes by schizonts or erythrophagocytosis in the spleen. Findings of the present study showed that both anaemia and thrombocytopaenia are common in malaria-infected patients, including P. knowlesi patients.

There was a significant inverse correlation between parasite density with Hb and platelet value for all the species of malaria in the present study. This was in line with other studies where malaria-infected patients tended to have significantly lower Hb levels and platelet counts [9, 82]. The trend of increasing parasite density with decreasing platelet count observed in the present study has been previously reported for P. falciparum[83, 84] but not P. vivax[85]. However, detailed study on the relation between parasite density and platelet count for human P. knowlesi infection was scarce. This showed that more investigations have to be conducted in order to study the correlation of platelet count with parasite density in malaria infection. Low Hb level due to increased parasite density might be explained by increased haemolysis or decreased rate of erythrocyte production due to the invasion of erythrocytes by malaria parasite [80, 86].

White blood cell count (WBC) during malaria are usually characterized as being low to normal [59, 87], a decrease in the number of WBC (leukocytes) in the blood is known as leukopaenia whereas elevated WBC refers to leukocytosis. Findings demonstrated that only a small proportion of the malaria patients in the present study suffered leukopaenia and leukocytosis, however, this may be due to the small sample size in the present study. Results showed that the WBC for P. vivax malaria patient was lower than P. falciparum malaria patient, which was in contrast to previous study where WBC in P. falciparum patients were lower than those in P. vivax-infected patients and, in turn, lower than those in healthy patients [87]. Overall, the TWBC of patients for all the malaria species detected in the present study were generally within the normal range.

The present study had several limitations. Firstly, haematological profiles such as Hb level, platelet counts and TWBC were only available for 85 patients. The statistical analyses of these parameters were mostly not significant due to the small sample size. Secondly, the baseline demographic data is incomplete and, unfortunately, detailed clinical features of patients could not be retrieved. There were several reasons for the incomplete data in the study. Demographic data and haematological profiles were unable to be retrieved for the samples that were sent from other districts of the division to the clinic. Case notes of all these patients were kept in the respective hospital or clinic therefore the data required were not available. In addition, sometimes, patients with severe malaria were referred to the district hospitals immediately before the full blood profiles were tested in the clinic. There were also some patients with uncomplicated malaria who have refused admission and, therefore, full blood counts were unable to be performed. Thirdly, travel history of patients was not available and therefore the information on where patients acquired malaria infection, especially P. knowlesi infection, could not be deduced. Nevertheless, despite all these limitations, this study is one of the few comprehensively analysed studies on the occurrence and epidemiology study of malaria in the eastern districts of Sabah.

Present study revealed that most of the P. knowlesi infection in Sandakan division was misdiagnosed as P. malariae by microscopic examination. The wide distribution of human P. knowlesi infection indicated that this parasite is generally able to infect humans and there is high potential for knowlesi outbreak to occur given that the full knowledge of the dynamics of transmission between human-monkey-mosquito is not adequately understood. This study also observed that there was no clustering of cases of P. knowlesi suggesting that transmission may occur in forested areas away from people’s homes. In addition, given that microscopic findings showed that gametocytes were only detected in two P. knowlesi patients, indicating that there was less chance of human to human transmission occurring. With regard to haematological analysis, thrombocytopaenia and anaemia was shown to be the most frequent malaria-associated haematological complications. However, larger prospective studies are needed to investigate the relations between parasite species as well as the parasite density to the haematological complications associated with malaria. It is crucial that prospective studies on epidemiological risk factors and transmission dynamics of P. knowlesi in these areas are comprehensive in order to accrue accurate epidemiological data, which serves as an imperative foundation in formulating effective malaria control measures that will subsequently contribute to the successful elimination of malaria in Malaysia by 2020. In order to reduce mortality, it is pertinent for malariae suspected samples to be confirmed by PCR. The availability of advanced diagnostic tool PlasmoNex™ enhanced the accuracy and accelerated the speed in the diagnosis of malaria. Given the severity of knowlesi malaria and its fatality, confirmed P. knowlesi cases should be reported as such by the doctors treating these cases so that the actual disease burden is known.