In recent years, antibiotics have emerged as important environmental pollutants in aquatic environments because of their potential for adverse effects on ecosystem and human health (Managaki et al. 2007; Pruden et al. 2012). Antibiotics are commonly prescribed as therapeutics to avert or heal infectious diseases in humans and animals worldwide. They are also used in large quantities as growth promoters in agriculture, beekeeping, livestock, and aquaculture. The worldwide yearly production of antibiotics has been estimated at between 100,000 and 200,000 tons (Kümmerer 2007), and more than 25,000 tons are used each year in China (Xu et al. 2007). Antibiotic use has been growing in industrialized and developing countries in recent years, leading to their widespread occurrence in surface water, sediment, and biota worldwide. Antibiotics enter the environment though the effluent from municipal waste water treatment plants (Lindberg et al. 2010; LaPara et al. 2011), drainage from livestock and poultry farming, and manure fertilizer runoff from farmland (Cheng et al. 2016). Antibiotics have been detected in surface waters at concentrations in parts-per-trillion (ng/L) and parts-per-million (μg/L) (Zhang and Li 2011), and have been detected in solid matter (sludge, soil, and sediments) at concentrations in g/kg or mg/kg (Hu et al. 2012). There is therefore an urgent need to monitor the occurrence, partitioning coefficients, and temporal and spatial distribution of antibiotics in watersheds, and to evaluate their environmental behavior and associated risks for the aquatic environment.

We selected the Weihe River in the North China Plain for our catchment-scale study of the characteristics and contamination profiles of antibiotics. We were particularly interested in this watershed as it includes a range of different land uses and levels of economic development. The Weihe River drains an area of 12,921 km2 in Henan Province, China, and is the largest tributary of the Haihe River. The annual average precipitation in the Weihe River Basin is about 500 mm, and rainfall is concentrated into a very short period during the summer. The main inputs of antibiotics into the Weihe River are thought to be surface and groundwater flow, and localized anthropogenic sources such as effluents from urban areas, livestock breeding, aquaculture facilities, paper mills, and water treatment facilities. In this study we examined the partitioning of, and risks from, antibiotics in water and sediment at ten sampling sites in the Weihe River. Our choice of target antibiotic compounds was mainly guided by trends in human and agricultural consumption of antibiotics in China (Tang et al. 2015). The 12 target antibiotics belong to three different families of compounds, namely (1) sulfonamides (SAs) comprising sulfadiazine (SDZ), sulfamethoxazole (SMX), sulfamethazine (SMZ), sulfachloropyridazine (SCP), and sulfadimethoxine (SDM); (2) fluoroquinolones (FQs) comprising norfloxacin (NOR), ciprofloxacin (CIP), ofloxacin (OFL), and enrofloxacin (ENR), and (3) tetracyclines (TCs) comprising tetracycline (TC), oxytetracycline (OTC), and chlortetracycline (CTC). This study will serve as a reference for future risk assessments. We also hope that it will represent a useful contribution to research into antibiotics pollution in this area and that the results will be used to support the development of future pollution control measures.

Materials and Methods

Information about the Weihe River and the sampling sites is presented in Fig. 1 and Table 1. The samples were collected in August 2014. Surface water samples were collected from between 10 and 15 cm below the surface in 2 L brown glass bottles that had been pre-cleaned with methanol and ultrapure water. Samples were stored at −4°C until pretreatment, which was done within 24 h of sample collection. Sediment samples were centrifuged at 2415 g to obtain about 200 mL of pore water. Grab samples of top 10 cm surface sediments were collected in stainless steel containers. The sediment samples were freeze-dried and then sieved through an 80-mesh sieve. All sediment data in this study are reported on a dry weight (dw) basis.

Fig. 1
figure 1

Geographical location of the ten sampling sites along the Weihe River

Full size image
Table 1 Sampling sites description along Weihe River
Full size table

The water samples (1500 mL of surface water and 200 mL of pore water) were processed by solid-phase extraction (SPE) as described by Huang et al. (2013). Briefly, the sample extracts were cleaned up using solid phase extraction vacuum manifolds (Supelco Company, Bellefonte, PA). Oasis hydrophilic–lipophilic balance (HLB) cartridges (6 mL per 500 mg, Waters, Watford, UK) were preconditioned by washing sequentially with 10 mL of methanol and 10 mL of ultrapure water. The pH of the samples was adjusted to three with concentrated sulfuric acid. Subsequently, 50 ng of surrogate standards, namely trimethyl-13C3(TRM-13C3), SAM-13C6 (sulfonamide-13 C6), CIP-D8, and Na2EDTA (0.8 g) were added to the samples. A portion (5 g) of the sieved sediment sample was weighed into a 50-mL polytetrafluoroethylene tube, and spiked with 50 ng of the internal standards CIP-D8, TRM-13C3, SAM-13C6. Then 30 mL of the extraction buffer (pH = 5), made up of 5 mL of 0.1 M Na2EDTA, 15 mL of methanol, and 10 mL of citrate buffer, was added. All surface water, pore water, and sediment samples were analyzed in triplicate and the data were reported as the average of the three analyses.

Antibiotics were separated on a C18 column (150 × 2.1 mm, 3.5 μm; Waters, Milford, MA, USA) in an HPLC system. The target compounds were detected by high-performance liquid chromatography tandem mass spectrometry (HPLC-ESI MS/MS). The MS/MS parameters were optimized as follows: ion spray voltage of 4.5 kV, sheath gas pressure of 25 arbitrary units, auxiliary gas pressure of 8 arbitrary units, ion transfer capillary temperature of 350°C, and source CID of 10 V. The collision gas (argon) was used at an indicated pressure of 1.5 mTorr.

The recovery values of these spiked antibiotics ranged from 87.2% to 108.7% in water (surface water and pore water) and from 71.4% to 111.2% in sediment. The limit of quantification (LOQ) calculated with a signal/noise ratio of 10 was 0.3 to 1.2 ng/L for water and 1.0 to 2.5 μg/kg for sediment. All the samples were analyzed in triplicate, and the relative standard deviation was less than 15%.

Results and Discussion

The concentrations of antibiotics and their detection frequencies in surface water and pore water from the Weihe River are summarized in Table 2. The most frequently detected compounds, with detection frequencies of 100% in the water samples, were TC and two sulfonamides, SDZ and SMX. The detection frequency of antibiotics in the FQ group (ENR, OFL, and DIF) was the lowest, and ranged from 40% to 64% in surface water and from 42% to 56% in pore water samples. Concentrations of all the detected antibiotics in the pore water and surface water samples were mainly at the ng/L level. The total antibiotic concentrations from the three families in surface water and pore water samples ranged from 11.1 to 173.1 ng/L and from 5.8 to 103.9 ng/L, respectively. Among the selected antibiotics, TC and SMX dominated in surface water, with concentrations ranging from 2.2 to 55.6 ng/L and from 0.1 to 47.5 ng/L, respectively. CIP and CTC were evenly distributed throughout the river. The concentrations of the different antibiotics decreased in the following order: TCs > SAs > FQs. Antibiotics were also ubiquitous in pore water samples from the Weihe River. Sulfonamides and tetracyclines had high detection frequencies and were found in more than 80% of samples. Among the detected antibiotics, TC, OTC, CTC, NOR, and SMX were the predominant compounds in the pore water, with the mean concentration of 17.4, 11.3, 8.2, 13.8 and 7.2 ng/L, respectively. The antibiotics observed at higher concentrations in pore water, namely NOR, OFL and OTC, were different than the dominant compounds in surface water (TC and SMX). This difference might be explained by the strong sorption affinities of TCs to organic matter influencing the partitioning between water and sediments. SAs possess high solubility and chemical stability in water. Consequently, the detection frequency and mean concentrations of SAs were lower in pore water than in surface water.

Table 2 Antibiotics concentrations detected in surface water, pore water and sediments samples from Weihe River
Full size table

Summary information about antibiotics in surface sediment samples is presented in Table 2. The concentrations of antibiotics in sediments ranged from 9.5 to 153.4 μg/kg (dry weight), and averaged 106.0 μg/kg. In other of increasing levels in the pore water samples, measured concentrations were in the range of 2.1 to 58.4 μg/kg for SAs, 3.1 to 57.4 μg/kg for FQs, and 4.3 to 63.8 μg/kg for TCs, respectively. The three antibiotic families concentrations in the sediment decreased in the order: TCs > FQs > SAs. TC had the highest mean concentration (20.2 μg/kg), followed by NOR and OTC, with mean concentrations of 16.4 and 15.2 μg/kg, respectively. These results suggest that TCs, NOR, and OFL tend to accumulate more in sediments than the other target antibiotics.

The pseudo-partition coefficient (P-PC) is useful for examining the partitioning behavior of pollutants between water and sediments in aquatic settings (Kim and Carlson 2007). In this study, we applied the organic carbon normalized pseudo-partition coefficient (P-PC oc ) as we were not sure if the concentrations in water and sediment were balanced. The P-PC oc is computed as follows (Carballa et al. 2008):

$$P-P{{C}_{oc}}={}^{\left( {}^{{{C}_{s}}}\!\!\diagup\!\!{}_{Cw}\; \right)}\!\!\diagup\!\!{}_{{{f}_{oc}}}\;$$

where C s (μg/kg) and C w (μg/L) are the concentrations of antibiotics in sediment and water, respectively, and f oc is a dry weight fraction of TOC in sediment (%).

The P-PC OC values of sediment to surface water (Pss-PC OC ) and sediment to pore water (Psp-PC OC ) are presented in Table 3. The P-PC oc values determined using the data from the samples collected along the Weihe River indicated large differences between the target antibiotics,e.g.,from nd to 943 L/kg for SAs, nd to 2213 L/kg for FQs, nd to 2405 L/kg for TCs, respectively. In general, the Psp-PC OC values were larger than those for Pss-PCOC, which reflect the minimal antibiotic concentrations in pore water relative to those in surface water, as new pollution inputs may directly increase the antibiotic concentrations in surface water. In contrast, we found that TC, NOR, and OFL characterized by higher Psp-PC OC values and low Pss-PC OC values were the main antibiotic category in surface sediments, which shows that Psp-PC OC was a good indicator of the sorption characteristics in the aquatic environment. Studies have shown that the pH of water plays a significant role in governing the partitioning characteristics between soils (sediment) and water for different types of antibiotics. It is extremely easy to distribute antibiotics through soils (sediment) with low pH, and increases in the pH could result in decreased antibiotic sorption in soils (sediment) (Gong et al. 2012; Sukul et al. 2008). The pH values of most water samples in this study area were neutral (6.8–7.5), and were generally slightly higher than those of pore water. The Pss-PC OC values of most of the target compounds were generally less than the Psp-PC OC values, and may reflect the higher pH values in surface water and lower pH values in pore water. The pseudo-partitioning coefficients of different antibiotics have been reported for various regions worldwide, but the organic carbon normalized pseudo partition coefficient has rarely been reported. The P-PC OC values calculated for the three groups of antibiotics in this study are much higher than those reported for Victoria Harbour, Hong Kong (82.9 to 123.8 L g) (Xu and Li 2010), but are lower than those reported for stream water in Northern New Jersey (4380 to 39,240 L kg) (Gibs et al. 2013).

Table 3 Calculated organic carbon normalized pseudo-partition coefficients of selected antibiotics
Full size table

The potential ecological risk associated with the presence of these antibiotics was evaluated by calculating risk quotients (RQ), in line with the technical guidance for risk assessments of the European Commission (European Commission 2003). The risk quotient values for surface water (RQw) and sediment (RQs) were calculated from the antibiotic concentrations in surface water and pore water using the following formula:

$$\text{RQ = MEC/PNEC}$$

where MEC is the measured concentration in surface water or pore water, and PNEC is the predicted no-effect concentrationr. PNEC were taken from literature and values were based on aquatic toxicity data of antibiotics to algae, the most sensitive aquatic species (Tang et al. 2015; Andrieu et al. 2015; Rico et al. 2014). The RQ values were divided into three risk levels, namely low (0.01–0.1), medium (0.1–1.0), high (>1.0), as suggested by de Souza et al. 2009.

In this study, we calculated the RQw and RQs for each antibiotic in the Weihe River. The computed RQ values of the 12 antibiotics for algae in surface water and sediment of the Weihe River are presented in Fig. 2. This diagram indicates that the risks from FQs and TCs antibiotics to algae ranged from very low to medium in both surface water and sediment. There were no risks to algae from SDZ, SMZ, SCP, and SDM at any of the sampling sites, but the risk from SMX ranged from low to high at 90% of the sites. The ecological risk assessment also showed that there was quite a high ecological risk to algae from SMX, NOR, TC, OFL, and CIP in this area.

Fig. 2
figure 2

The calculated risk quotients (RQ) of 12 antibiotics of the ten sampling sites along the Weihe River (RQw for surface water; RQs for sediment). a RQw, b RQs

Full size image

In this study, we examined the distribution and occurrence of 12 selected antibiotics in surface water, pore water, and sediment of the Weihe River in northern China, an area of high density livestock breeding and intensive agricultural activity. The results show that there were antibiotic pollutants in surface water, pore water, and sediment of the Weihe River. SDZ, SMX, and TC were found at high frequencies in surface water and pore water, while NOR and OFL were more likely to accumulate in sediments than other antibiotics. The organic carbon normalized pseudo-partition coefficients of pore water, and surface water were also calculated, and may help to forecast future patterns of antibiotics in the aquatic environment. The ecological risk assessment showed that there was a relatively high ecological risk to algae from SMX, NOR, TC, OFL, and CIP in this area. Further studies are needed to address the transformation and fate of those antibiotics and the spread of antibiotic-resistant bacteria or genes in this watershed.