Direct centrifugation for detecting Giardia spp. cysts in filter backwash water

Waterborne diseases are of emerging concern, and protozoa have been particularly associated with outbreaks worldwide (Karanis et al. 2007; Efstratiou et al. 2017b). For instance, Giardia spp. cysts are infective forms associated with giardiasis outbreaks (Karanis et al. 2007; Baldursson & Karanis 2011). Cyst dimensions range from 7 to 14 μm, and they have low infectious doses in humans, from 1 to 10 organisms (Ortega & Adam 1997). Furthermore, they might present high environmental resistance as their survival has been registered in water for up to 2 months at 8 °C (Cacciò et al. 2003).

Waterborne diseases concern developing countries, especially as they may be neglected and associated with inadequate sanitation infrastructure (Coelho et al. 2017; Sammarro Silva & Sabogal-Paz 2020). In this scenario, water treatment plants (WTPs) are fundamental in decreasing the drinking-water supply's microbiological risks. Although most cysts can be removed in the filtration steps (Heller et al. 2004), further research is still needed to assess the microbial risks of WTP residues, including water treatment sludge (WTS) and filter backwash water (FBW). Moreover, recirculating FBW could reinsert cysts into the treatment system, especially as they resist chlorination, a common disinfection technique (Karanis et al. 1996; Freitas et al. 2010).

Methods for the detection of Giardia spp. cysts can vary significantly among different samples (Franco et al. 2012), and most approaches have challenges concentrating and detecting cysts in high-turbidity water. Concentration methods are classified into three categories: filtration (e.g., membranes, microfibers, ultrafiltration, and nanofiltration), flocculation–sedimentation (e.g., calcium carbonate, aluminum sulfate, ferric sulfate, and formaldehyde with ethyl acetate or ether), and centrifugation (e.g., batch and continuous flow) (Efstratiou et al. 2017a). Calcium carbonate flocculation (CCF) stands out among concentration methods (Feng et al. 2011). This protocol has already been used for high-turbidity water, although pH changes interfere with the viability of the cysts (Giglio & Sabogal-Paz 2018).

The direct centrifugation (DC) with the addition of the detergent dispersion solution ICN 7X (i.e., a neutral and phosphate-free detergent, nontoxic for sensitive organisms) at 1.0% (DC+ICN 7X) concentrates a large volume sample into a pellet (Boni de Oliveira 2012). This protocol was recently applied with the purification by immunomagnetic separation (IMS) for concentrating cysts in high-turbidity water (Giglio & Sabogal-Paz 2018) and WTS (Silva & Sabogal-Paz 2020; Ogura & Sabogal-Paz 2021a). In addition, another study recovered Giardia spp. cysts in the FBW from a bench-scale flotation treatment (Silva & Sabogal-Paz 2020). However, to the best of our knowledge, the literature lacks a protocol for detecting Giardia spp. cysts in the FBW from a high-turbidity water treatment. In this scenario, monitoring cysts in WTP is still a challenge for negligible matrices, especially for FBW that can be recirculated without proper disinfection.

Therefore, this research aimed to assess and validate the DC+ICN 7X+IMS method for the detection of Giardia spp. cyst in the FBW. The samples were generated in a bench-scale drinking-water treatment, and the experiments were performed with a commercial Giardia sp. suspension and the standard EasySeed^® (Giardia lamblia).

This research consisted of two experimental stages: FBW generation by jar test assays and the performance of the DC+ICN 7X+IMS protocol for cyst concentration and detection. In this study, synthetic samples were prepared to minimize interferences from natural freshwater variations (e.g., pH changes, metals, organic matter, and other microorganisms) that might influence experimental conditions (Ogura & Sabogal-Paz 2021b). Physical, chemical, and microbiological parameters were analyzed based on APHA (2012).

The first step consisted of preparing high-turbidity study water (112 NTU) with uncontaminated groundwater and kaolinite (0.16 g L⁻¹, Sigma-Aldrich^®/Fluka 60609). Then, the bench-scale drinking-water treatability assays were performed in jar tests (2 L) with automatic fast-mixing (1,000 s⁻¹ for 10 s) and slow-mixing (25 s⁻¹ for 30 min). The polyaluminium chloride (PACl) was selected as the coagulant (specific weight of 1.362 g L⁻¹ and 16.68% content of Al₂O₃). Three PACl dosages were tested (24, 25, and 26 mg L⁻¹, corresponding to 2.11, 2.20, and 2.29 mg Al³⁺ L⁻¹, respectively) based on previous studies (Giglio & Sabogal-Paz 2018). The chosen decantation rate was 1.5 cm min⁻¹ (i.e., approximately 10 min 30 s). Then, the clarified water was directed to the filtration stage.

Second, the filtration step was performed in attached laboratory filters (ALF). The ALF consisted of acrylic tubes with a 19 mm internal diameter, a 40 cm height, and previously washed and dried sand (grain size between 0.30 and 0.59 mm, and effective size of 0.42 mm). The clarified water from each jar was directed to one ALF, and the adopted filtration rate was 100 m³m⁻² day⁻¹. Finally, each sand filter media was washed with 300 mL of distilled water to generate the FBW. Multiple treatability assays were performed to obtain a compound FBW sample.

All material was washed with a 0.1% Tween 80 solution to avoid cyst adhesion to solid surfaces. The commercial suspension of Giardia sp. was purified at the Laboratory of Protozoology at the State University of Campinas, Brazil. The suspension was homogenized in a vortex (2 min) before each use. The counting of the average number of cysts followed the recommendations from the Merifluor^® kit (Meridian Bioscience, Inc.) with DAPI (4′,6-diamidino-2-phenylindole) solution (Sigma-Aldrich^®, F6057). The procedures were done with an aliquot of 5 μL of the suspension in triplicate. In addition, potentially viable cysts were identified by the propidium iodide (PI) incorporation (Sigma-Aldrich^®, P4170). The volume of PI solution corresponded to the added suspension (5 μL), and cysts were classified as stained (PI-positive) or not-stained (PI-negative) (Ogura & Sabogal-Paz 2021a). The counting was performed in a microscope, considering DAPI, fluorescence (fluorescein-5-isothiocyanate, FITC), PI, and differential interference contrast (DIC) images.

The volume of suspensions was inoculated directly into the FBW samples (1 L), considering the estimated number of cysts and the target concentration. This procedure was chosen to avoid losses throughout the treatment process (e.g., flocculation and filtration). The samples were then homogenized with a magnetic stirrer for 20 min. First, a preliminary analytical quality assay (n=3) was performed with the commercial suspension of Giardia sp. (with a target concentration of ±2,000 cysts L⁻¹). Second, the analytical quality control (n=4) validated the chosen method with the EasySeed^® (BTF Bio – Australia) suspension (with a pre-determined number of 100±1.9 G. lamblia cysts).

The DC+ICN 7X method was adapted from Boni de Oliveira (2012). First, the contaminated FBW sample (1 L) was divided into 20 tubes (Falcon^®, 50 mL). Then, the centrifugation step was conducted (at 1,500×g for 20 min), and the remaining pellets (0.5 mL) were transferred to a single tube. In addition, all tubes were washed with 0.5 mL of Tween 80 (0.1%), and this content was transferred to a single tube. Next, another centrifugation was carried out, and the supernatant was removed until the remaining 5 mL mark. This pellet was transferred to a flat-sided tube (FST) with 3.0 mL of 1.0% ICN detergent dispersion solution 7X (ICN Pharmaceuticals Inc.^®). After homogenization in a rotatory mixer (20 rpm for 1.0 h), the FST content was transferred to a centrifuge tube (Falcon^®, 50 mL). The last centrifugation (1,500×g for 20 min) concentrated the sample into a 5 mL pellet.

The CCF protocol was adapted from Giglio & Sabogal-Paz (2018). These procedures also aimed to concentrate 1.0 L of FBW into a 5 mL sample. First, 10 mL of calcium chloride (CaCl₂, 1 M) and 10 mL of sodium bicarbonate (NaHCO₃, 1 M) were added to the FBW. After the samples were agitated for 10 min, 1.5 mL of sodium hydroxide (NaOH, 5 M) increased the pH to 10. Then, the samples rested overnight (at room temperature, covered with a watch glass). On the next day, the supernatant was removed until the 100 mL mark, and 20 mL of 10% sulfamic acid (H₃NSO₃) was added. This sample was divided into three centrifuge tubes (Falcon^®, 50 mL). The beaker was washed with 30 mL of Tween 80 (0.1%), and this content was distributed to the tubes. After centrifugation (20 min at 1,500×g), the supernatant was discarded from each tube. The remaining pellets (1.0 mL) were transferred to a single tube. The pH was corrected to neutrality (7.0), and 4.2 mL of PBSS (phosphate-buffered saline solution – Sigma-Aldrich^®) was added. The volume was concentrated by centrifugation (10 min at 1,500×g) into a 5 mL remaining pellet.

Both methods of concentration were followed by the IMS purification protocol (USEPA 2012) according to the Dynabeads™ GC-Combo (IDEXX^®) procedures and two acid dissociations. The United States Environmental Protection Agency Method 1623.1 for drinking water was used as guidance to validate the recovery efficiency (USEPA 2012). Finally, statistical analyses were performed on Statistica^®. First, the Shapiro–Wilk verified the normality and homogeneity, respectively. Then, statistical differences were evaluated by Student's t-test and the analysis of variance (ANOVA), with 95% confidence (p-value<0.05).

The groundwater used in this study had total alkalinity of 15.3 mg CaCO₃ L⁻¹, conductivity of 33.2 μS cm⁻¹, turbidity of 0.2 NTU, an apparent color of 0.0 HU, and a pH of 6.78. After adding 0.16 g L⁻¹ of kaolinite, the study water had total alkalinity of 8.8 mg CaCO₃ L⁻¹, conductivity of 53.2 μS cm⁻¹, turbidity of 112 NTU, apparent color of 114 HU, true color of 3.3 HU, and a final pH of 7.15.

The filtered water's apparent color and turbidity were endpoints to evaluate the treatment efficiency (Table 1). The dosage of 2.11 mg Al³⁺ L⁻¹ had lower average apparent color (1.5 HU) for filtered water in 10 min but with a higher standard deviation (±1.4 HU). Therefore, the dosage of 25 mg PACl L⁻¹ (2.20 mg Al³⁺ L⁻¹) was applied to achieve low-turbidity filtered water (0.18±0.1 NTU, after 20 min); then, this dosage was chosen for the following research steps. The reference study used to estimate this dosage (Giglio & Sabogal-Paz 2018) under similar treatability conditions and study water (125 and 130 NTU, respectively) also considered 2.20 mg Al³⁺ L⁻¹ as the optimal PACl dosage for obtaining 0.25 and 0.18 NTU filtered water, respectively.

After the water treatment assays, the generated FBW (compound sample) presented turbidity of 14.5 NTU, apparent color of 9.1 HU, true color of 1.2 HU, pH 6.10, conductivity of 52.1 μS cm⁻¹, 113 mg L⁻¹ of total solids, 0.94 mg L⁻¹ of total organic carbon, and alkalinity of 12.3 mg CaCO₃ L⁻¹, 0.55 mg Al L⁻¹, and 0.374 mg Fe L⁻¹. Further information on drinking-water treatment conditions and FBW characterization was previously presented in Ogura & Sabogal-Paz (2021b).

Approximately 1,900 cysts L⁻¹ were inoculated into the FBW samples, considering the cysts counting on the suspension. The preliminary assay (Table 2) showed recovery rates of 14.4 and 10.7% for CCF and DC+ICN 7X, respectively. Thus, the recoveries from our study complied with Method 1623.1 (USEPA 2012), which establishes at least an 8% recovery rate. Furthermore, the DC+ICN 7X recovered more PI-negative cysts (86 cysts, 83.5% of the total) than CCF (43 cysts, 27.4%). This observation may result from CCF method damage to cyst cell walls through pH changes (Giglio & Sabogal-Paz 2018). In these cases, indicator dyes (e.g., PI) can be incorporated into Giardia spp. cysts through cell wall disruptions (Campbell et al. 1992). Thus, the DC+ICN 7X method was selected as it showed cleaner wells and recovered more PI-negative cysts (Figure 1). Nonetheless, this was a preliminary assay with only one sample for each method.

The DC+ICN 7X+IMS method recovered 19.5±0.6 and 7.8±2.9 cysts for the commercial and EasySeed^® suspensions (statistically different as p<0.05), respectively (Table 3). Thus, our study's protocol for cyst detection complied with Method 1623.1 (USEPA 2012). However, the commercial suspensions presented more consistent results, with a higher recovery rate (19.5%) and a lower coefficient of variation (3.2%) than EasySeed^® (7.8 and 36.9%, respectively). In addition, most cysts were recovered in the first dissociation for the commercial suspensions (p<0.05), while the number was higher from the second dissociation with EasySeed^® (p<0.05). Ultimately, Andreoli & Sabogal-Paz (2021) indicated that a third acid dissociation increases the number of recovered cysts up to 31.9% from flotation FBW samples. The present investigation was limited to two acid dissociations in the IMS protocol due to the high costs of analyses.

Other studies performed several protocols for detecting cysts from FBW with turbidity ranging from 7 to 27 NTU (Table 4). For instance, Silva & Sabogal-Paz (2020) applied the DC+ICN 7X+IMS method and achieved a 43.9% average recovery rate of Giardia spp. (from a commercial suspension) for FBW (26.7 NTU) from a flotation bench-scale jar test. However, these authors reported a recovery rate of only 3.8% for the EasySeed^® suspension, which was lower than our study (7.8% for a 14 NTU FBW). This variability can result from different FBW samples generated from specific water treatment processes (i.e., decantation and flotation). In addition, the ferric sulfate flocculation method was applied for Giardia muris in FBW (6.6 NTU) and reached an 1.8% recovery rate (Sammarro Silva & Sabogal-Paz 2021). In another study, these authors applied membrane filtration and obtained a recovery rate of 17.4% of G. muris cysts from the same FBW sample, although adding the IMS protocol did not improve the recovery rate of this concentration method (13.0%) (Sammarro Silva & Sabogal-Paz 2020).

Furthermore, for the FBW from swimming pools, Greinert et al. (2004) had a 3.3% recovery rate for Giardia spp. cysts when combining the CCF and DC protocols. Although Ladeia et al. (2018) did not find Giardia duodenalis in their FBW samples, cysts were detected in 6.1% of their centrifuged and WTS samples. In addition, high-turbidity water has also been studied (Table 4). For example, Giglio & Sabogal-Paz (2018) achieved an 11.5% cyst recovery rate for this protocol for high-turbidity study water (130 NTU). These authors showed that the CCF and DC+H₂SO₄ methods presented higher average recovery rates (46.1 and 26.0%).

Most research has focused on artificial FBW samples under controlled laboratory conditions (Table 4). Accordingly, in the present investigation, synthetic water was used to minimize possible adverse effects and interferences of the matrix on recovery, as the objective was to evaluate the detection method itself. However, the studied protocol should be tested and validated in other matrices; notably, various FBW samples generated from raw freshwater (e.g., water samples with different organic matter content and turbidity). In addition, the DC+ICN 7X+IMS protocol could be applied to actual contaminated FBW samples as the inoculation of protozoa suspensions into artificial samples from the laboratory might not represent the conditions from WTPs. Finally, the variations from commercial and EasySeed^® suspensions are still unclear, although different results from our research could be related to the number of inoculated cysts in each protocol. Studies have already claimed that higher initial concentrations of cysts would likely result in better recovery efficiencies (Sammarro Silva & Sabogal-Paz 2020 and references therein). These aspects could be interesting for future studies regarding the gaps in detecting Giardia spp. cysts in FBW samples.

The WTS is another residue from WTP that has been studied with several detection protocols (Table 4). For instance, Ogura & Sabogal-Paz (2021a) concentrated Giardia spp. cysts from a WTS sample (3,187 NTU) with 17.0 and 24.8% recoveries, respectively, for the commercial and EasySeed^® suspensions. Furthermore, these authors applied the DC+ICN 7X+IMS protocol for detecting cysts after the alkaline treatment with 27 mg CaO/100 mL. As a result, they achieved 2.05 and 2.14 logs of cyst inactivation for 3 and 5 days, respectively. The interferences on the viability of cysts can pose challenges for evaluating the disinfection efficiency (Silva & Sabogal-Paz 2020; Ogura & Sabogal-Paz 2021a). These studies considered PI-negative cysts to indicate potential viability when assessing alkaline and ozone treatments for disinfection of water treatment residues. On the other hand, Silva & Sabogal-Paz (2020) showed 32.5 and 49.0% recoveries for the floated residue (380 NTU) by applying the DC+ICN 7X+IMS protocol. This detection protocol is also viable for other challenging matrices. For example, Boni de Oliveira (2012) obtained 37.6 and 33.2% recovery rates for concentrating soil samples with 500 and 1,000 cysts of Giardia spp., respectively. Accordingly, Orlofsky et al. (2013) applied centrifugation and IMS for soil samples, recovering up to 89±11% of G. lamblia cysts.

Magnetic microspheres were detected in the reading wells even after two acid dissociations (Figure 2(d)). This limitation was also observed by Andreoli & Sabogal-Paz (2021), who detected cysts that were still attached to the magnetic microspheres after three acid dissociations. Furthermore, Fava et al. (2021) applied the microfiltration followed by the IMS protocol with 50% fewer beads for water with 40 NTU. As a result, these authors obtained a 56.1% recovery rate for a commercial G. duodenalis cyst suspension, while 47.5% for the analytical quality control with ColorSeed^®. Thus, 50 μL of the immunomagnetic beads complied with Method 1623.1 for G. duodenalis (USEPA 2012), which is an advantage regarding cost reduction.

Cryptosporidium spp. oocysts are another concern for protozoa monitoring and detection in developing countries. In a previous study (Ogura & Sabogal-Paz 2021b), Cryptosporidium spp. oocysts were recovered from FBW samples generated under the same treatment conditions. However, oocyst recovery rates (15.4±3.3 and 2.8±0.8%) were lower than the ones obtained for Giardia spp. cysts in the present research (19.5±3.2 and 7.8±36.9%), considering commercial and Easyseed^® suspensions, respectively. Although Method 1623.1 requires a higher oocyst recovery rate (>32%) than cysts (>8%), this observation was also reported by another evaluation of both protozoa in FBW samples (Silva & Sabogal-Paz 2020).

The costs incurred by cyst detection in WTP residues were challenging in the present research. Four studies performed the cost analysis for the IMS protocol, ranging from US$ 118 to US$ 212 (Giglio & Sabogal-Paz 2018; Ogura & Sabogal-Paz 2021a). In addition, Andreoli & Sabogal-Paz (2021) indicated that a third acid dissociation had a 23.5% increase in costs compared to the second one, which might not be feasible. Therefore, high costs concern and limit protozoa environmental monitoring, especially in developing countries with limited or unevenly distributed financial resources.

This study performed the DC+ICN 7X+IMS protocol to detect Giardia spp. cysts from an artificially contaminated FBW, which was generated by the bench-scale water treatment. The cyst recovery rate (7.8%, considering the EasySeed^® suspension) from the applied protocol was slightly below the required range by Method 1623.1 (8–100%). On the other hand, the recovery rate was superior for the commercial suspension (19.5%), meeting the minimum standards. Nonetheless, other studies from the literature presented recovery efficiencies generally below 50%, highlighting some challenges for Giardia spp. cyst detection in water treatment residues. Therefore, the high costs and low efficiency show an evident limitation of this methodology. Moreover, developing feasible detection protocols for Giardia spp. cysts can contribute to studies on protozoa outbreaks and giardiasis epidemiology. In addition, future studies should apply this methodology to natural samples (e.g., surface water and effluents from the wastewater treatment).

The authors are grateful to the São Paulo Research Foundation (FAPESP) (Process 12/50522-0), the Global Challenges Research Fund (GCRF), UK Research and Innovation (SAFEWATER; EPSRC Grant Reference EP/P032427/1), the Royal Society (ICA/R1/201373 – International Collaboration Awards 2020), and the National Council for Scientific and Technological Development (CNPq-Brazil, process no. 308070/2021-6) for the research support. The authors are also thankful to the National Council for Scientific and Technological Development (CNPq-Brazil) for the master's scholarship awarded to Allan Pretti Ogura.

The authors certify that they have no affiliations with or involvement in any organization or entity with any financial or non-financial interest in the subject matter or materials discussed in this manuscript.

All relevant data are included in the paper or its Supplementary Information.