A review of the current state of malaria among pregnant women in Papua New Guinea

Besides young children, pregnant women are at high risk of malaria in highly endemic countries. This paper reviews evidence from studies conducted in Papua New Guinea (PNG) in the last 20 years on the burden and prevention of malaria in pregnancy and highlights gaps in our knowledge of malaria in pregnancy in PNG. Overall, primigravidae were found to be at higher risk than multigravidae, with up to 40% of primigravidae but only 10-25% of multigravidae infected with Plasmodium falciparum at delivery. Such infections were found to be associated with a 128-145 g decrease in birthweight. Mean birthweights reported between 1980 and 2003 range from 2.58 to 2.72 kg in primigravidae and 2.84 to 3.09 kg in multigravidae, with 21% to 48% and 9% to 19% of babies born to primigravidae and multigravidae, respectively, of low birthweight (<2500 g). The negative impact of malaria in pregnancy is compounded by relatively low rates of antenatal coverage. The current PNG national treatment policy which prescribes a treatment course of first-line antimalarial treatment (currently chloroquine and sulphadoxine-pyrimethamine) at first antenatal clinic contact, followed by weekly chloroquine prophylaxis and iron and folate supplementation, may no longer be effective given the high levels of resistance to chloroquine in PNG and poor compliance. In order to reduce the burden of malaria in pregnancy in PNG, alternative methods of control such as insecticide-treated nets and intermittent preventive treatment in pregnancy (IPTp), as well as improved modes of delivery of maternal health interventions, are urgently needed.

The epidemiology of malaria in the Papua New Guinea highlands: 5. Aseki, Menyamya and Wau-Bulolo, Morobe Province

Although not strictly a highlands province, Morobe encompasses large highlands areas, the most important being Aseki, Menyamya and Wau-Bulolo. A series of rapid malaria surveys conducted in both the wet and dry seasons found malaria to be clearly endemic in areas below 1400 m in Menyamya and Wau-Bulolo, with overall prevalence rates in the wet season (25.5%, range: 9.1%-39.2%) greatly exceeding those in the dry season (8.3%, range: 2.4%-22.8%; p < 0.001). In the wet season surveys Plasmodium falciparum was the clearly predominant species, accounting for 63% of all infections. P. vivax increased in frequency in the dry season (from 27% to 46%, p < 0.001), while P. falciparum and P. malariae decreased. In line with past surveys a low prevalence of malaria was found in the Aseki area. Malaria was found to be the main source of febrile illness in the wet season with at least 60% of measured or reported fever associated with parasitaemia. Other causes of febrile illness dominated in the dry. In villages with parasite prevalence rates < 20% mean haemoglobin levels and prevalence of severe anaemia were strongly correlated with overall parasite prevalence. In addition concurrent malarial infections were associated with a strong reduction of individual haemoglobin levels (-1.2 g/dl) and there was increased risk of moderate-to-severe anaemia with concurrent malaria. Malarial infections are thus the most significant cause of febrile illness and anaemia in the highlands fringe populations in Morobe. As a consequence all villages below 1500-1600 m in Morobe Province should be included in malaria control activities.

The epidemiology of malaria in the Papua New Guinea highlands: 6. Simbai and Bundi, Madang Province

Although predominantly a lowland province, Madang also includes highland areas such as Simbai and Bundi along the northern highland fringe. While the malaria situation in the coastal lowlands has been studied in great detail, the current malaria situation in the highland fringe communities has not been studied in depth since the 1960s. A series of recent malariological surveys found that the malaria situation has changed little over the last 40 years in both Simbai and Bundi. In the Simbai area there is little malaria transmission in villages above 1400 m, with a prevalence rate (PR) of 2.5-4.2%. Below 1400 m, however, there is moderate to high transmission (PR 8.6-24.7%) with surprisingly little difference in prevalence rates between survey villages, despite large differences in altitude. Prevalence rates of malaria infection were low in all Bundi villages (2.5-8.5%) with most infections occurring in adolescents and adults, which indicates limited acquisition of effective immunity to malaria and the possibility that many infections are acquired when travelling to the highly malarious lowlands area. Based on spleen rates the lower Simbai area would be regarded as mesoendemic, and the upper Simbai and Bundi areas as hypoendemic. Only in the lower Simbai area is malaria a major cause of febrile illness. However, in all areas village mean haemoglobin (Hb) levels were highly correlated with the prevalence of malaria infections, while concurrent parasitaemia reduced individual Hb levels by 1.3 g/dl (CI95 [1.0-1.5], p < 0.001) and significantly increased the risk for moderate-to-severe anaemia (Hb < 8 g/dl) (adjusted odds ratio 5.6, CI95 [3.6-8.6], p < 0.001). Based on the survey results, areas of different malaria epidemiology are delineated and options for control in each area are discussed.

The epidemiology of malaria in the Papua New Guinea highlands: 4. Enga Province

Of all Papua New Guinea provinces, Enga has the largest proportion of people living at altitudes that preclude malaria transmission. However, the first systematic surveys in 1979 showed that malaria was endemic in lower-lying valleys to the north and east of the province. A series of new surveys conducted in both wet and dry seasons showed that these areas remain the main malaria focus in Enga. However, over the last 25 years the risk of malarial infections has increased substantially in areas < 1200 m (from 10% to 37-41%). In these low-lying areas people acquire substantial antimalarial immunity and most infections are asymptomatic. However, people in villages in these areas had significantly lower mean haemoglobin levels (13.0 vs 14.0 g/dl, p < 0.001) than in areas above 1200 m, where overall prevalence rates (0-9%) have not changed much. In areas between 1200 and 1600 m epidemics with parasite prevalence rates in excess of 20% have been found to occur. Malaria was a significant cause of febrile illness only in endemic areas or during outbreaks. Although rarely used, sleeping under a bednet was associated with a significant reduction in risk of malaria infection (adjusted OR = 0.45, p = 0.01). On the other hand, sleeping in garden houses away from the main villages increased the risk of malaria infection (adjusted OR = 1.6, p = 0.03). Malaria control in outlying, malarious areas of Enga province could therefore be based on the distribution of long-lasting impregnated bednets, while at the same time addressing the additional risks posed by the high mobility of many of these populations through targeted health education.

Malaria control in Papua New Guinea results in complex epidemiological changes

With a renewed interest in large-scale malaria interventions, knowledge about the possible long-term effects of such interventions on the nature of malaria transmission is essential. We document complex changes in malaria epidemiology over the last 40 years associated with changing malaria control activities in Karimui, an isolated area in Papua New Guinea. An initially equal distribution of Plasmodium falciparum, P. vivax and P. malariae changed to currently 68% P. falciparum, after passing through a phase of transitory P. vivax dominance, when control started to fail. Initial drops in malaria prevalence proved difficult to sustain and present post-control levels are significantly higher than pre-control levels. The example of Karimui indicates that unsustained control can lead to changes in malaria patterns that may leave a population worse off.

The epidemiology of malaria in the Papua New Guinea highlands: 3. Simbu Province

Two very distinct malaria zones can be found within Simbu Province. The north of the province is characterized by the absence or very low level of local malaria transmission, but there is a considerable risk of epidemics prevalent in the lower-lying parts. During non-epidemic periods, parasite prevalence was usually under 5%, with similar frequencies for Plasmodium falciparum and P. vivax (47% each), and malaria was an only minor source of febrile illness. During epidemics, however, 13-36% of people were infected, predominantly with P. falciparum (64%), and high levels of severe morbidity were present. In south Simbu malaria is clearly endemic with an overall prevalence of 35%, combined with a strong age-dependence of infections, low haemoglobin levels, high rates of enlarged spleen and moderate to severe anaemia (haemoglobin level < 7.5 g/dl) in children. The malaria epidemiology in south Simbu is thus more similar to the lowlands than to other highlands areas. Epidemic prevention, surveillance and response in the north, and bednet distribution and strengthening of curative services in the south, are therefore the priorities for malaria control in Simbu Province.

The epidemiology of malaria in the Papua New Guinea highlands: 1. Western Highlands Province

Despite a resurgence of malaria in many Papua New Guinea highlands and highlands fringe areas after the cessation of control activities in the early 1980s the malaria situation in these areas has received little attention. A series of cross-sectional surveys were therefore carried out to provide accurate and up-to-date information on the prevalence of malaria and the risk of epidemics and to propose adequate malaria control strategies. Studies in 24 villages in Western Highlands Province found the prevalence of malarial infections to be strongly correlated with altitude, ranging from 1.6% at altitudes of 1500-1800 m to over 30% in villages below 900 m. Malaria outbreaks were observed at the end of the rainy season. All four human malaria species were present with P. falciparum infections clearly dominating. The relative importance of P. vivax increased with altitude, while both P. malariae and P. ovale were rare. Many infections were of low density. While malaria is an important source of febrile illness in endemic areas below 1500 m altitude, only few observed or reported fevers are due to malarial infections in higher, nonendemic areas. Rates of enlarged spleens, mean haemoglobin levels and the prevalence of anaemia (Hb <7.5 g/dl) were strongly linked to the level of malaria found in each community and were associated with both altitude and concurrent malarial infection. Based on the survey results, areas of different malaria epidemiology are delineated and options for control in each area are discussed.

The epidemiology of malaria in the Papua New Guinea highlands: 2. Eastern Highlands Province

The epidemiology of malaria in Eastern Highlands Province (EHP) is characterized by generally very low-level or no local malaria transmission but a considerable risk of epidemics. In non-epidemic situations, parasite prevalence was under 5% in all but one area, Lufa Valley, where P. vivax was the most common parasite (prevalence rate of 58%). During epidemics, however, 23% (range 8-63%) of people were infected, with P. falciparum the predominant parasite. Mixed infections were very common (12%) but P. malariae and P. ovale were rare. Outside epidemics, malaria was only a minor source of febrile illness and enlarged spleens and anaemia (haemoglobin level < 7.5 g/dl) were virtually absent. However, epidemics were associated with a high incidence of relatively severe morbidity, both in terms of fevers and moderate to severe anaemia. Epidemic prevention, surveillance and response therefore are priorities in malaria control in EHP.

Complex patterns of malaria epidemiology in the highlands region of Papua New Guinea

A cross-sectional malaria survey of the Pabrabuk area in the Western Highlands Province found that all 4 human malaria species were present in a single village, with an overall parasite prevalence rate of 27%. Plasmodium falciparum was the most frequently detected infection (14%) followed by P. vivax (11%), P. malariae (5%) and P. ovale (3%). 10 of the 51 infections were mixed. Anopheles punctulatus was the most frequent vector species in the area, but both An. farauti no. 6 and An. karwari were also present in low numbers. This diversity in both parasite and vector populations indicates that complex malaria patterns are found in Papua New Guinea even at the moderate transmission levels found in low-lying inter-montane valleys.