Table 1.
Description of the alcohol samples used in this study.
Table 2.
Exposure parameters used for assessing the health risk of consuming 40% (v/v) sachet ethanol in Acoli, Uganda.
Fig 1.
Sum of all metals detected in each sachet is shown as solid rectangles.
(a) Cu is detected at high levels in locally distilled alcoholic drinks (AWE, 421 μg/100 mL (97%); BOL, 123 μg/100 mL (92%); NSB, 140 μg/100 mL (94%); TEB, 1210 μg/100 mL (100%) and in the Single Malt Scotch Whisky (TGL, 41 μg/100 mL (76%)); the former due to long copper tubes employed in the distillation process, and the latter is due to copper pots used in the brewing and distillation processes. (b) An expansion of the plot in (a) without Cu, an outlier, to confirm the disproportionate level of copper in the Lira-Lira spirits. Note that levels of all metals are overrepresented in REX. Means are very low and shown as angle grids.
Table 3.
Summary statistics associated with 20 metals detected in 13 documented and four undocumented alcoholic beverage brands sold in northern Uganda (in μg/L).
Details are in S1 Table.
Table 4.
The levels of metals (μg/L) in each brand of alcohol and their associated health indices (HI).
HI is listed from the most significant (top) to the least (bottom).
Fig 2.
Semi-log plots showing pictorial overview of (a) the density of contaminating metals in each sachet brand, and (b) the frequency of detecting an individual metal across the brands. Qualitatively, REX contains the highest combination of different metals. The frequency of detecting As, Al, Ba, Be, Cd, Cr, Mn, Mo, Pb, Sb, Sn, Sr and Zn in all the brands is relatively higher than of detecting Ag, Co, Cu, Ni, Se, V and Tl.
Fig 3.
Box-Whisker plots using Plotly software for the target hazard quotient (THQ) data of the 20 metal ions detected in the 18 waragi studied.
THQ values were row-percentage transformed to accommodate for the high levels of Cu in the Lira-Lira spirits). Outliers are shown as solid circles and suspected outliers as open circles. Whisker type: Standard Error; Outlier method: rounding.
Fig 4.
Cluster analysis of the target hazard quotient (THQ) data with Cu and ethanol data excluded (to provided better signal strength).
(a) At an arbitrary distance of 2.4, three distinct metals clusters are observed. Cluster 1 has only one metal, As; cluster 2 contains 13 metals; cluster 3 contains Se, V, Sb, Co and Tl (Algorithm Ward’s, cophenetic correlation coefficient: 0.8952). Note that at a distance of 1.0, clusters 3 divides into three sub-clusters. (b) At an arbitrary distance of 1.0, the spirit brands cluster into five different groups (Algorithm: Ward’s, cophenetic correlation coefficient: 0.9494). REX is the only member of cluster 1; NBS and TGL are in cluster 2; cluster 3 contains eight brands: CW1, CW2, V6T, BRG, KPW, ROV, UGW, SAG; cluster 4 has GOV, BEG and B7W; and cluster 5 contains BOL, B5V and AWE.