Spatio-Temporal Dynamics of Asymptomatic Malaria: Bridging the Gap Between Annual Malaria Resurgences in a Sahelian Environment

Abstract

In areas of seasonal malaria transmission, the incidence rate of malaria infection is presumed to be near zero at the end of the dry season. Asymptomatic individuals may constitute a major parasite reservoir during this time. We conducted a longitudinal analysis of the spatio-temporal distribution of clinical malaria and asymptomatic parasitemia over time in a Malian town to highlight these malaria transmission dynamics. For a cohort of 300 rural children followed over 2009-2014, periodicity and phase shift between malaria and rainfall were determined by spectral analysis. Spatial risk clusters of clinical episodes or carriage were identified. A nested-case-control study was conducted to assess the parasite carriage factors. Malaria infection persisted over the entire year with seasonal peaks. High transmission periods began 2-3 months after the rains began. A cluster with a low risk of clinical malaria in the town center persisted in high and low transmission periods. Throughout 2009-2014, cluster locations did not vary from year to year. Asymptomatic and gametocyte carriage were persistent, even during low transmission periods. For high transmission periods, the ratio of asymptomatic to clinical cases was approximately 0.5, but was five times higher during low transmission periods. Clinical episodes at previous high transmission periods were a protective factor for asymptomatic carriage, but carrying parasites without symptoms at a previous high transmission period was a risk factor for asymptomatic carriage. Stable malaria transmission was associated with sustained asymptomatic carriage during dry seasons. Control strategies should target persistent low-level parasitemia clusters to interrupt transmission.

Figure 1.

Spatial and temporal pattern of malaria cases. (A) represents the weekly time series of clinical case incidence per 100 children (red). Weekly rainfall is presented in blue. Low and high transmission period are presented, respectively, in grey and white. Choropleth maps of clinical case incidences, (B and C), are presented at the neighborhood level, respectively, for high and low transmission periods (through 2009–2014). Green ellipses represent low risk cluster (according to Kulldorff’s definition) and black ones, high risk clusters. Estimated relative risks of each cluster are presented (as well as the number of neighborhood within each cluster, between brackets). This figure appears in color at www.ajtmh.org.

Figure 2.

Spatial and temporal pattern of asymptomatic carriage. (A) represents the weekly time series of asymptomatic carriage incidence per 100 children (orange). Weekly rainfall is presented in blue. Low and high transmission periods are presented, respectively, in grey and white. Choropleth maps of asymptomatic incidences, (B and C), are presented at the neighborhood level, respectively, for high and low transmission periods (through 2009–2014). Green ellipses represent low risk clusters (according to Kulldorff’s definition) and black ones, high risk clusters. Estimated relative risks of each cluster are presented (as well as the number of neighborhoods within each cluster, between brackets). (D) represents the asymptomatic carriage ratios by clinical cases for each high and low transmission period (red and blue, respectively). Each square is proportional to each period length. The diamonds represent the ratios for the high and low transmission periods (red and blue, respectively). This figure appears in color at www.ajtmh.org.

Figure 3.

Spatial and temporal pattern of gametocytemia. (A) represents the weekly time series of gametocytemia incidence per 100 children (purple). Weekly rainfall is presented in blue. Low and high transmission period are presented, respectively, in gray and white. Choropleth maps of gametocytemia incidences, (B and C), are presented at the neighborhood level, respectively, for high and low transmission periods (through 2009–2014). Green ellipses represent low risk clusters (according to Kulldorff’s definition) and black ones, high risk clusters. Estimated relative risks of each cluster are presented (as well as the number of neighborhoods within each cluster, between brackets). (D) represents the positive gametocytemia ratios by clinical cases for each high and low transmission period (red and blue, respectively). Each square is proportional to each period length. The diamonds represent the ratios for the high and low transmission periods (red and blue, respectively). This figure appears in color at www.ajtmh.org.

Figure 4.

Wavelet analysis of weekly clinical malaria incidence and weekly rainfall. The different panels comprise each time series analysis. The left panels (A1 and B1) are the wavelet power spectrum of the square root transformed and normalized series (2009–2014). The color code for power values ranges from dark blue (low values) to dark red (high values). The dotted black lines show the statistically significant area (threshold of 95% CI). The white lines represent the maxima of the wavelet power spectrum, and the black curves indicate the cone of influence that delimits the region not influenced by edge effects. The right panels (A2 and B2) correspond to the mean spectrum (black solid line) with its threshold value of 95% CI (dotted black line) for the aggregate time series. The third raw panel (C) represents the phase evolution of the two time series, using wavelet analysis. The black dotted line represents the phase differences (lag) between the two series. This figure appears in color at www.ajtmh.org.