Malaria morbidity and mortality following introduction of a universal policy of artemisinin-based treatment for malaria in Papua, Indonesia: A longitudinal surveillance study

Abstract

Background: Malaria control activities can have a disproportionately greater impact on Plasmodium falciparum than on P. vivax in areas where both species are coendemic. We investigated temporal trends in malaria-related morbidity and mortality in Papua, Indonesia, before and after introduction of a universal, artemisinin-based antimalarial treatment strategy for all Plasmodium species.

Methods and findings: A prospective, district-wide malariometric surveillance system was established in April 2004 to record all cases of malaria at community clinics and the regional hospital and maintained until December 2013. In March 2006, antimalarial treatment policy was changed to artemisinin combination therapy for uncomplicated malaria and intravenous artesunate for severe malaria due to any Plasmodium species. Over the study period, a total of 418,238 patients presented to the surveillance facilities with malaria. The proportion of patients with malaria requiring admission to hospital fell from 26.9% (7,745/28,789) in the pre-policy change period (April 2004 to March 2006) to 14.0% (4,786/34,117) in the late transition period (April 2008 to December 2009), a difference of -12.9% (95% confidence interval [CI] -13.5% to -12.2%). There was a significant fall in the mortality of patients presenting to the hospital with P. falciparum malaria (0.53% [100/18,965] versus 0.32% [57/17,691]; difference = -0.21% [95% CI -0.34 to -0.07]) but not in patients with P. vivax malaria (0.28% [21/7,545] versus 0.23% [28/12,397]; difference = -0.05% [95% CI -0.20 to 0.09]). Between the same periods, the overall proportion of malaria due to P. vivax rose from 44.1% (30,444/69,098) to 53.3% (29,934/56,125) in the community clinics and from 32.4% (9,325/28,789) to 44.1% (15,035/34,117) at the hospital. After controlling for population growth and changes in treatment-seeking behaviour, the incidence of P. falciparum malaria fell from 511 to 249 per 1,000 person-years (py) (incidence rate ratio [IRR] = 0.49 [95% CI 0.48-0.49]), whereas the incidence of P. vivax malaria fell from 331 to 239 per 1,000 py (IRR = 0.72 [95% CI 0.71-0.73]). The main limitations of our study were possible confounding from changes in healthcare provision, a growing population, and significant shifts in treatment-seeking behaviour following implementation of a new antimalarial policy.

Conclusions: In this area with high levels of antimalarial drug resistance, adoption of a universal policy of efficacious artemisinin-based therapy for malaria infections due to any Plasmodium species was associated with a significant reduction in total malaria-attributable morbidity and mortality. The burden of P. falciparum malaria was reduced to a greater extent than that of P. vivax malaria. In coendemic regions, the timely elimination of malaria will require that safe and effective radical cure of both the blood and liver stages of the parasite is widely available for all patients at risk of malaria.

Fig 1. Map of Mimika district in Papua Province, eastern Indonesia.

Map adapted from WorldOfMaps (https://www.worldofmaps.com).

Fig 2. Community entomological and meteorological surveillance.

Predicted number of Anopheles (‘An’) mosquito bites per year by month for the three main Plasmodium vectors in Mimika District (top) and mean daily rainfall by month (bottom).

Fig 3. Malaria surveillance gathered from community healthcare facilities.

Absolute number (‘N.’) of malaria cases detected within the community surveillance network by Plasmodium species and month (top) and proportion of all blood films examined in the community with gametocytes (‘gams’) by Plasmodium species and time (bottom).

Fig 4. Antimalarial drug prescriptions at Mitra Masyarakat hospital.

Number of oral blood schizontocidal drug prescriptions (top), PQ prescriptions (middle), and IV blood schizontocidal drug prescriptions (bottom) at Mitra Masyarakat Hospital by month. Low-dose PQ is defined as a total dose of ≥1.5 to <5 mg/kg. High-dose PQ is defined as a total dose of ≥5 mg/kg. IV, intravenous; PQ, primaquine.

Fig 5. Malaria cases at Mitra Masyarakat hospital.

Number of outpatient (top) and inpatient (middle) malaria cases at Mitra Masyarakat Hospital by Plasmodium species and month and proportion of all hospital malaria cases due to the different Plasmodium species (bottom).

Fig 6. Malaria-related morbidity and mortality at Mitra Masyarakat hospital.

Proportion of all patients presenting to Mitra Masyarakat Hospital who were admitted by Plasmodium species and month (top). Total number of days of inpatient bed occupancy × 100/month by Plasmodium species (second from top). Proportion of all patients presenting to Mitra Masyarakat Hospital who were anaemic (Hb < 7 g/dL) (third from top) and who died (bottom) by Plasmodium species and month. Hb, haemoglobin.

Fig 7. Proportion of malaria morbidity and mortality attributable to Pv mono- or mixed infection.

In the community and hospital (top). In the hospital stratified by age group (middle). In the hospital: malaria-related hospital admissions, inpatient bed days, malaria-related anaemia, and malaria-related deaths (bottom). Pv, P. vivax.