Antibiotics have been detected in surface waters worldwide at concentrations up to 1.9 micrograms/L, but are typically detected at low nanogram/L concentrations. The potential health effects of exposure to low levels of these compounds via tap water are not known, but there is significant concern among water consumers regarding the occurrence of antibiotics and other pharmaceutical compounds in water supplies. Thus, a significant amount of research has been performed recently to investigate the removal of pharmaceuticals via conventional and advanced water treatment processes. While conventional treatment processes (i.e., coagulation, flocculation, sedimentation, and filtration) are generally not effective, oxidation processes (e.g., chlorination, ozonation) and granular activated carbon exhibit some effectiveness at removing pharmaceuticals. As expected, removals are highly dependent on compound structure. Furthermore, some oxidants, such as chloramines, are not effective at oxidizing pharmaceuticals.
Slow-rate biofiltration processes (SRBF), such as slow sand filtration (SSF) and riverbank filtration (RBF), are drinking water treatmeant systems comprised of two stages in sequence: 1) a relatively shallow biotic region where media (i.e., filter sand or aquifer material) is colonized by biofilm bacteria, followed by 2) an deeper abiotic filtration zone. These processes are extensively used in Europe and developing global regions and are seeing increased usage in the United States. There is evidence in the literature that SRBFs can remove a wide variety of trace organic pollutants including: pesticides, disinfection byproducts, and some pharmaceuticals. Little is known regarding the ability of SRBF processes to remove antibiotics from water supplies nor has any work been done to investigate the potential adverse effects of antibiotics on the biofilm bacteria that are critical to SRBF system performance. Thus, this research was performed to determine the extent and mechanisms (i.e., sorption versus biodegradation) of antibiotic removal in SRBF processes and the effects of antibiotics on biofilm bacteria (i.e., activity and community composition).
The effect of antibiotics on bacterial activity and community structure was investigated by growing biofilm in the presence and absence of a mixture of antibiotics in a continuous-flow rotating annular bioreactor (CFRAB) with acetate as substrate. Three representative compounds were selected for use in this research: sulfamethoxazole (SMX), erythromycin (ERY), and ciprofloxacin (CIP). These antibiotics were selected because they: 1) represent three prominent classes of antibiotics with differing mechanisms of action against bacteria, 2) have been detected in surface water, 3) exhibit different chemical characteristics, and 4) have differing levels of biodegradability. Areal acetate utilization rates for a constant feed of antibiotics were similar to the control experiments, and utilization rates did not change during an antibiotic shock loading experiment. Attached biomass levels were greater for experiments involving a high CIP concentration (3.33 micrograms/L), however, yielding comparatively lower steady-state biomass-normalized substrate utilization rates. Microbial community analyses via automated ribosomal intergenic spacer analysis (ARISA) revealed shifts in community structure for the high dose CIP experiments.
A CFRAB was also used to investigate antibiotic sorption to bacterial biofilm. The extent of sorption, as indicated by the organic carbon partition coefficient (Koc), was 15 to 23 times greater for CIP compared to ERY and SMX. The Koc values did not correlate with experimentally-determined Kow values, suggesting that the sorption of relatively hydrophilic (i.e. Kow < 1.7) and charged antibiotics to typically negatively charged biofilm is driven by ionic interactions (i.e. ion exchange) rather than hydrophobic interactions.
The attenuation and impact of antibiotics in SRBF systems was investigated by conducting bench-scale filter column experiments with mixtures of SMX, ERY, and CIP at high (3.33 micrograms/L, each) and low (0.33 microgram/L, each) antibiotic feed conditions. Consistent with the CFRAB experiments, antibiotic breakthrough times were greatest for CIP, with very little uptake of SMX or ERY. Biodegradation was not observed for any antibiotic during 6-weeks of filter column operation or in complementary batch experiments. A one-dimensional advection-dispersion equation (with linear sorption) model was validated against experimental results and used to compare antibiotic retardation in SSF, RBF, and rapid gravity biofiltration (RGBF) systems. Of the modeled systems, antibiotic retardation was greatest in RBF, with little antibiotic removal expected for SSF. Based on analysis of ARISA data, the community structure of bacterial biofilm was not affected in filters exposed to antibiotics at low concentrations (i.e. 0.33 microgram/L, each) similar to those found in surface waters, with a few species impacted under high concentration conditions (3.33 microgram/L, each).
The results of this work will help those interested in understanding and predicting antibiotic fate in engineered and natural systems where biofilm is present. The results indicate that antibiotic removal in SRBF processes will be dictated by compound properties such as charge and hydrophobicity, and that limited removal of antibiotics in SRBF processes can be expected. Finally, the results suggest that that mixtures of antibiotics at concentrations typically observed in surface waters are unlikely to adversely affect SRBF biofilm bacteria or process performance.