Human Pathogenic Bacteria Detected in Rainwater: Risk Assessment and Correlation to Microbial Source Tracking Markers and Traditional Indicators

Abstract

Roof-harvested rainwater (RHRW) was investigated for the presence of the human pathogenic bacteria Mycobacterium tuberculosis (M. tuberculosis), Yersinia spp. and Listeria monocytogenes (L. monocytogenes). While Yersinia spp. were detected in 92% (n = 25) of the RHRW samples, and L. monocytogenes and M. tuberculosis were detected in 100% (n = 25) of the samples, a significantly higher mean concentration (1.4 × 10³ cells/100 mL) was recorded for L. monocytogenes over the sampling period. As the identification of appropriate water quality indicators is crucial to ensure access to safe water sources, correlation of the pathogens to traditional indicator organisms [Escherichia coli (E. coli) and Enterococcus spp.] and microbial source tracking (MST) markers (Bacteroides HF183, adenovirus and Lachnospiraceae) was conducted. A significant positive correlation was then recorded for E. coli versus L. monocytogenes (r = 0.6738; p = 0.000), and Enterococcus spp. versus the Bacteroides HF183 marker (r = 0.4071; p = 0.043), while a significant negative correlation was observed for M. tuberculosis versus the Bacteroides HF183 marker (r = -0.4558; p = 0.022). Quantitative microbial risk assessment indicated that the mean annual risk of infection posed by L. monocytogenes in the RHRW samples exceeded the annual infection risk benchmark limit (1 × 10^-4 infections per person per year) for intentional drinking (∼10^-4). In comparison, the mean annual risk of infection posed by E. coli was exceeded for intentional drinking (∼10^-1), accidental consumption (∼10^-3) and cleaning of the home (∼10^-3). However, while the risk posed by M. tuberculosis for the two relevant exposure scenarios [garden hosing (∼10^-5) and washing laundry by hand (∼10^-5)] was below the benchmark limit, the risk posed by adenovirus for garden hosing (∼10^-3) and washing laundry by hand (∼10^-3) exceeded the benchmark limit. Thus, while the correlation analysis confirms that traditional indicators and MST markers should be used in combination to accurately monitor the pathogen-associated risk linked to the utilisation of RHRW, the integration of QMRA offers a more site-specific approach to monitor and estimate the human health risks associated with the use of RHRW.

Keywords: QMRA; human pathogenic bacteria; microbial source tracking markers; rainwater; traditional indicator organisms.

FIGURE 1

Box and whiskers plot of the concentration (cells/100 mL) for L. monocytogenes, M. tuberculosis, Yersinia spp., E. coli, Enterococcus spp., adenovirus, Bacteroides HF183, and Lachnospiraceae. The whiskers illustrate the minimum and maximum, the outer box illustrates the 1st and 3rd quartiles, and the inner line illustrates the median.

FIGURE 2

Dendrogram of the Cluster Analysis with Ward’s Methods of target pathogens versus the MST markers and indicator organisms detected in the RHRW samples.

FIGURE 3

(A) Annual health risk associated with the use of RHRW in the community for ingestion scenarios based on the presence of E. coli (EC; blue) and L. monocytogenes (LM; orange). (B) Annual health risk associated with the use of RHRW in the community for inhalation scenarios based on the presence of E. coli (EC; blue), L. monocytogenes (LM; orange), M. tuberculosis (MT; green) and adenovirus (AV; pink). The whiskers illustrate the minimum and maximum, the outer box illustrates the 1st and 3rd quartiles, and the inner line illustrates the median. The benchmark limit (1 × 10^{– 4}) is indicated by the dashed red line.