Burkitt lymphoma risk shows geographic and temporal associations with Plasmodium falciparum infections in Uganda, Tanzania, and Kenya

Abstract

Endemic Burkitt lymphoma (eBL) is a pediatric cancer coendemic with malaria in sub-Saharan Africa, suggesting an etiological link between them. However, previous cross-sectional studies of limited geographic areas have not found a convincing association. We used spatially detailed data from the Epidemiology of Burkitt Lymphoma in East African Children and Minors (EMBLEM) study to assess this relationship. EMBLEM is a case-control study of eBL from 2010 through 2016 in six regions of Kenya, Uganda, and Tanzania. To measure the intensity of exposure to the malaria parasite, Plasmodium falciparum, among children in these regions, we used high-resolution spatial data from the Malaria Atlas Project to estimate the annual number of P. falciparum infections from 2000 through 2016 for each of 49 districts within the study region. Cumulative P. falciparum exposure, calculated as the sum of annual infections by birth cohort, varied widely, with a median of 47 estimated infections per child by age 10, ranging from 4 to 315 infections. eBL incidence increased 39% for each 100 additional lifetime P. falciparum infections (95% CI: 6.10 to 81.04%) with the risk peaking among children aged 5 to 11 and declining thereafter. Alternative models using estimated annual P. falciparum infections 0 to 10 y before eBL onset were inconclusive, suggesting that eBL risk is a function of cumulative rather than recent cross-sectional exposure. Our findings provide population-level evidence that eBL is a phenotype related to heavy lifetime exposure to P. falciparum malaria and support emphasizing the link between malaria and eBL.

Keywords: East Africa; Plasmodium falciparum malaria; endemic Burkitt’s lymphoma (eBL); epidemiology; non-Hodgkin lymphoma.

Fig. 1.

Cumulative estimated P. falciparum infections by age, country, and year. Box-and-whisker plots show the distribution of the cumulative number of expected P. falciparum malaria infections per year, for three example years, by age and country. Children born prior to 2000 were excluded from this analysis, which explains the absence of data for children 13 and over in 2012. The horizontal line of each box represents the median for each age, while the lower and upper bounds of the box represent the 25th and 75th percentiles, respectively. Vertical lines of each box represent the values from half of the 25th percentile to 1.5 times the 75th percentile, and dots outside this line show outlier values. The number of estimated infections varies both by geography and time. In 2012, the average 10 y-old had 61, 40, and 201 infections in Kenya, Tanzania, and Uganda, respectively. This changed to 33, 32, and 176 infections in 2014 and 36, 25, and 162 infections in 2016 in Kenya, Tanzania, and Uganda, respectively.

Fig. 2.

Spatial distribution of cumulative annual P. falciparum infection and annual eBL incidence. On the left is the average annual estimated number of cumulative P. falciparum infections from 2000 to 2016 by country and district for the areas included in the EMBLEM study. Darker colors indicate higher estimated number of cumulative infections, indicating greater average P. falciparum infection pressure. On the right is the total eBL incidence by country and district for the areas included in the EMBLEM study. Data are from 2010 through September 2016 in Uganda and 2012 through September 2016 in Kenya and Tanzania. Darker colors indicate higher total eBL incidence. In the center are the study regions colored by country, and Lake Victoria is colored blue. Average annual cumulative P. falciparum infections and the average eBL incidence are strongly correlated, with a Spearman correlation coefficient of 0.58 (95% CI: 0.34 to 0.70).

Fig. 3.

Risk of eBL by age and cumulative P. falciparum incidence. Panel A shows the IRR of eBL at all observed levels of estimated lifetime cumulative P. falciparum infections compared to 50 estimated lifetime P. falciparum infections (where the red dotted line crosses the slope), which is approximately the average number of estimated infections of the study population. The risk of eBL increases 39% for every 100 P. falciparum infections, with 150 estimated lifetime infections corresponding to an IRR of 1.39 when compared to an individual with 50 estimated lifetime malaria infections, holding all other characteristics the same. yo, year-old. In panel B, each plot shows the IRR of eBL for different age groups as the estimated number of lifetime P. falciparum infections increases. In each panel of the figure, solid lines indicate the IRR for eBL associated with both age and increasing exposure to P. falciparum infection. In panel 1, for every 1-y increase in age, cumulative P. falciparum exposure increases by 1 infection. Panels 2 and 3 demonstrate the increase in risk when infections increase to 10 and 20 per each year of life. IRRs in each panel are presented relative to 6 cumulative infections at age 5, or how many infections a child would have at the midpoint of age 5, assuming one infection per year, to allow comparison of both age- and exposure-specific differences in eBL risk.