To test potential bilateral causalities relation between HIV-AIDS mortality and GDP, we propose a simple Granger noncausality test for heterogeneous panel data models. 44 African countries are selected for annual pooled data from 1990 to 2009. Results are presented for the heterogeneous noncausality hypothesis (HENC), which tests, for each cross-section unit, the nullity of all the coefficients of the lagged explanatory variable. Bilateral causality relation is observed for 5 countries out of 44 (11% of the countries in our data set). We have 18 countries of unidirectional causality, which 14 are from HIV mortality rate to GDP (43% from total), and 4 are from GDP to HIV mortality rate (9% from total). These results alert for the risk of epidemic trap, initiated first by the deleterious effect of HIV-Aids on countries income.