A country-wide malaria survey in Mozambique. II. Malaria attributable proportion of fever and establishment of malaria case definition in children across different epidemiological settings

Background Protection against clinical malaria episodes is acquired slowly after frequent exposure to malaria parasites. This is reflected by a decrease with increasing age in both parasite density and incidence of clinical episodes. In many settings of stable malaria transmission, the presence of asymptomatic malaria parasite carriers is common and the definition of clinical malaria remains uncertain. Methods Between February 2002 and April 2003, a country-wide malaria survey was conducted in 24 districts of Mozambique, aiming to characterize the malaria transmission intensities and to estimate the proportion of fever cases attributable to malaria infections in order to establish the malaria case definition. A total of 8,816 children less than ten years of age were selected for the study. Axillary temperature was measured in all participating subjects and finger prick blood collections were taken to prepare thick and thin films for identification of parasite species and determination of parasite density. The proportion of fever cases attributable to malaria infection was estimated using a logistic regression of the fever on a monotonic function of the parasite density and, using bootstrap facilities, bootstrapped estimated confidence intervals, as well as the sensitivity and specificity for different parasite density cut-offs were produced. Results Overall, the prevalence of Plasmodium falciparum was 52.4% (4,616/8,816). The prevalence of fever (axillary temperature ≥ 37.5°C) was 9.4% (766/8,816). Fever episodes peaked among children below 12 months of life [15.1% (206/1,517)]. The lowest fever prevalence of 5.9% (67/1,224) was recorded amongst children between five and seven years of age. Among 4,098 parasitized children, 498/4,098 (13.02%) had fever. The prevalence of malaria infections associated with fever peaked among children in the less than twelve months age group and thereafter decreased rapidly with increasing age (p < 0.001). High parasite densities were significantly associated with fever (p < 0.04). The proportion of fever attributed to malaria was 37.8% (95% CI 32.9% – 42.7%). An age-specific pattern was observed with significant variations across different regions in the country. In general, among children less than 12 months of life, the proportion of fever attributed to malaria infection was 43.5% (95% CI 25.8% – 61.2%), in children aged between 12 and 59 months of age was 39.6% (95% CI 30.3% – 48.9%), and among children aged between 5 and 10 years old was 21.5% (95% CI 11.6% – 31.4%). Conclusion This study confirms that malaria remains a major cause of febrile illness during childhood. It also defines the relation between parasite density and fever and how this varies with age and region. This may help guide case definition for clinical trials of preventive tools, as well as provide definitions that may improve the precision of measurement of the burden of disease.