Abstract
Cryptosporidium, Shigella, toxin-producing Escherichia coli, and rotavirus were reported to be the most responsible for severe and fatal diarrhea among infants. This study aimed to investigate the presence of these pathogens in infants' drinking water samples and analyzing using water quality determinants in eastern Ethiopia. A molecular (LAMP)-based cross-sectional study design was employed. A total of 410 and 37 water samples were tested from infant point-of-use at household and corresponding water source, respectively, from June 2020 to May, 2021. Cryptosporidium, Shigella, toxin-producing E. coli, and rotavirus were detected in 28.5, 30.0, 26.3, and 32.2%, of water samples tested from infant point-of-use, respectively. About 13.2% of the water samples were positive for all (four) pathogens together. Cryptosporidium, Shigella, toxin-producing E. coli, and rotavirus were detected in 27.0, 32.4, 29.7, and 37.8%, of water samples tested from water sources, respectively. Positive significant correlation was observed between infant point-of-consumption and water sources from which it is drawn toward the presence of each targeted pathogen. Unimproved water source showed a strong significant association with the presence of Cryptosporidium, Shigella and toxin-producing E. coli. Therefore, efforts should be made in development of improved water sources, source protection safety and health education to caretakers of infants.
HIGHLIGHTS
Highest prevalent agent detected in water samples was rotavirus (32.2%), followed by Shigella (30.0%), Cryptosporadium (28.5%), and toxin-producing E. coli (26.5%).
Positive correlation was observed between drinking water at point-of-use and water sources on presence of pathogens.
Significant association was observed between unimproved water source and presence of Cryptosporidium, Shigella, and toxin-producing E. coli.
ABBREVIATIONS
- AOR
adjusted odds ratio
- cDNA
complement deoxyribonucleic acid
- CI
confidence interval
- COR
crude odds ratio
- DNA
deoxyribonucleic acid
- E. coli
Escherichia coli
- EIWR
Ethiopian Institute of Water Resources
- ICR
Information Collection Rule
- LAMP
loop-mediated isothermal amplification
- NCBI
National Center for Biotechnology Information
- PCR
polymerase chain reaction
- RNA
ribonucleic acid
- SD
standard deviation
- SPSS
Statistical Package for Social Sciences
INTRODUCTION
Waterborne diseases are caused by an extensive range of pathogenic microbes, which include viruses, bacteria, and parasites, and are responsible for high morbidity and mortality rates in infants and children (Mokomane et al. 2018). The etiologic agents causing diarrhea are typically transmitted by the fecal–oral route to another person upon ingestion of contaminated water and food (Gerba 2015). Infants (<1 year of age) exposed to etiologic agents are much more likely to develop severe disease than the general population (Peterson 2006). Some of these agents can cause immediate morbidity and mortality, while others may not be noticed for many years.
According to the existing studies, there are pathogenic microbes responsible for causing the most severe and fatal diarrhea among infants worldwide. These studies reported that the most common causes were rotavirus, the protozoan Cryptosporidium, and bacteria, including Shigella and a toxin-producing strain of Escherichia coli (Kotloff et al. 2013; Global Burden of Disease (GBD) & Diarrhoeal Disease Collaborators 2018; McNeil 2013). These four etiologic agents account for more than 60% of diarrheal deaths in children under 5 years of age worldwide, including Ethiopia (Troeger et al. 2017). Poor drinking water quality continues to pose a major threat to human health (Bosch et al. 2008) and remains one of the most significant challenges in infant and child health. In addition, the role of drinking water as a carrier of disease-causing microbes is a major public health concern in many developing countries. Evidence suggests that high numbers of infants and children died from pathogenic organisms that occur in contaminated drinking water (World Health Organization; Magana-Arachchi & Wanigatunge 2020). Reportedly, unsafe water continues to be responsible for 72.1% of diarrhea deaths in children under 5 years old (Troeger et al. 2018). The global initiative needs to conduct extensive studies on pathogenic organisms in drinking water and catalyze action to reduce morbidity and related deaths.
The molecular-based techniques that involve direct DNA or RNA detection have been essential for the correct identification of species-specific pathogenic agents in drinking water samples. Several molecular methods, such as PCR, real-time PCR, and multiplex PCR assays, have been developed to detect the pathogens accurately (Saharan et al. 2014). In recent years, advanced molecular techniques have been developed that guide the use of methods that amplify the nucleic acid material at isothermal conditions. Among these techniques, loop-mediated isothermal amplification (LAMP) is a powerful nucleic acid amplification method that is simple, highly sensitive and specific, less time-consuming than PCR-based methods, and less prone to inhibition from DNA preparations (Mori & Notomi 2009; Varghese et al. 2012). It is ranked as the most frequently used method (Becherer et al. 2020) and also an important constituent for the efficient screening and testing of drinking water samples in resource-limited settings (Martzy et al. 2017).
To date, there are no studies in sub-Saharan countries in general or in Ethiopia in particular that have examined the prevalence of selected waterborne pathogens in drinking water intended for infants. As a result, there is a lack of information in this regard. The occurrence of pathogenic microbes in infant drinking water is affected by a combination of a wide range of natural and human influences on water quality (Usman et al. 2016), lots of which are inadequately understood. Based on the district sector offices report document, the eastern parts of Ethiopia are vulnerable to a recurring lack of access to and use of clean water, and infant diarrheal morbidity and mortality are commonly reported. Therefore, this study aimed to detect the presence of Cryptosporidium oocyst, Shigella species, toxin-producing strains of E. coli, and rotavirus in infants’ drinking water samples at the point-of-consumption and the corresponding water sources using LAMP and analyze their association with the microbial quality of drinking water determinants in eastern Ethiopia. This study will contribute to filling the aforementioned gaps and functioning as a platform for future research. This paper, thus, helps policymakers to design the right interventions to improve the quality of infants’ drinking water at the household level as well as the water source. This may help to reduce the burden of waterborne pathogens that are causing the most severe and fatal diarrhea in infants.
MATERIALS AND METHODS
Statistical methodology
Study area
This study was conducted in four districts that include Chiro, Mieso, Amibara, and Awash fentale, which are located in the eastern part of Ethiopia: Chiro and Mieso districts are located in the West Hararghe administrative zone of Oromia while Amibara and Awash Fentale districts are enrolled in zone 3 administration of the Afar region. As of 2016, the projected population of each district was estimated to be Chiro (223,167), Mieso (176,200), Amibara (80,531) and Awash Fentale (37,892). As per the national population conversion factor, about 3.4% of the total population is under the age of 1 year. This study covered an overall area of 5,186.14 sq·km. The study area has two main climate seasons: a dry season and a rainy season. The dry season occurs from October to February, whereas the rainy seasons have two periods: the period from June to September is the main rainy season, and some rainy periods usually occur from March to May. Acute respiratory infections, malaria, and diarrheal diseases are the most common reportable diseases in the area.
Study design
This study employed a molecular testing-based cross-sectional study design that applied the LAMP technique for the detection of four microbial pathogens (Cryptosporidium, Shigella, a toxin-producing strain ofE. coli, and rotavirus) from infant drinking water samples at the point-of-use in the households and the corresponding water sources, conducted from June 2020 to May 2021 in the four districts in eastern Ethiopia.
Study population
The study population was comprised of infants (<1 year of age). The sampling unit was households with infants, and the basic sampling population units (i.e., elements from which required information had to be ascertained) were their mothers or primary caregivers.
Sample size determination
The sample size (i.e., the number of households to be included in the study to represent the population of interest) was calculated using Openepi version 3.03a with the single population proportion. The sample size was calculated with the following assumptions: an estimated 80% prevalence of feacally contaminated drinking water samples that was taken from the study conducted in one of the districts in eastern Ethiopia (Mengistie et al. 2013), an error risk parameter of 1.96 (for an error risk of 5%, i.e., 95% confidence limits), a desired precision of 5%, and a design effect of 1.5, which resulted in the sample size of 369. About 15% of the non-response rate is added, and thus, 424 drinking water samples were taken from infants' drinking water at the point-of-use in the households.
Sampling technique
A two-stage cluster sampling along with a probability proportional to the population size sampling technique was employed for this study. In the first sampling stage, thirty ‘kebeles’ (the small administrative unit) or clusters were allocated to each targeted district on the basis of proportional allocation to the available ‘kebele’ size, followed by the selection of eligible ‘kebeles’ using the lottery method. The second sampling stage comprised households with infants within the selected ‘kebele’ using simple random sampling.
Data collection instrument and methods
Data and water samples were collected by a trained enumerator using pre-coded structured questionnaires and a water sample collection format, respectively.
Study variables
Independent variables of the study
Primary source of drinking water (improved or unimproved), drinking water storage hygienic status, practices of water treatment at point-of-use, water fetching time, water point location, and presence of residual chlorine.
Dependent variables of the study
Presence of microbial pathogens (Cryptosporidium oocyst, Shigella species, toxin-producing strain of E. coli, and rotavirus).
Data quality control
Data quality was assured by regulating both random and systematic error. Data quality was maintained through a properly designed data collection tool. The data collectors as well as supervisors were recruited with health-related educational backgrounds and language proficiency. Training was given to the data collectors on data collection and water sample collection procedures, and a pre-test was carried out in a community with similar characteristics. The data collection procedures that were developed and the collected data were reviewed by the principal investigator. Any identified errors were discussed, and immediate measures were taken, such as revisiting the households before leaving the village to make corrections and complete the questionnaire, discarding the water sample if exposed to contamination, and collecting it in a new sterile bottle.
Data management and analysis
The raw data were entered into CSPro version 6.1 then transformed to SPSS version 23 for analysis. Descriptive statistics such as frequency distribution and cross-tabulation were used to summarize the study variables. A phi coefficient correlation was done to analyze the association between point-of-consumption and water sources for the presence of pathogens. Binary logistic regression was applied to establish crude and adjusted odds ratios with 95% confidence intervals for the association between water quality determinants and each targeted pathogen. In bivariate conditional logistic regression, each variable that showed a p-value <0.25 was considered eligible for multivariable logistic regression. Variables that demonstrated a p-value <0.05 in the multivariable logistic regression were declared significantly associated with the outcome variable (presence of targeted pathogens).
Molecular – LAMP laboratory methodology
Selection of target genes
For the detection of each pathogenic microbe, a specific gene target region was selected. The selection was based on the gene most commonly used for identification of the target genes that expressed proteins unique to the organism of choice that helped to differentiate the presence of each pathogenic microbe (Gratacap-Cavallier et al. 2000; Song et al. 2005; Kumar et al. 2016; Crespo-Medina et al. 2020). The specific targeted genes encoding the 18S rRNA presented in Cryptosporidium spp., the ipaH gene in Shigella, the stx2A and stx2B genes in toxin-producing strains of E. coli, and the VP7 glycoprotein gene for rotavirus were chosen. These genes served as sites for specific primer designs that were used in LAMP assays.
Primer design
The primer sequences for LAMP amplification were designed based on the Cryptosporidium 18S rRNA gene accession number GenBank: L16996.1, Shigella ipaH gene accession number GenBank: M76443.1, Toxin-producing E. coli stx2A and stx2B gene accession number GenBank: FN252458.1, and rotavirus VP7 glycoprotein gene accession number GenBank: AB018697.1 from NCBI. The primers were designed using the Primer Explorer Version 5 software, available at https://primerexplorer.jp/e/ (see Table 1).
Primers and sequence used for LAMP assays
Primer name . | Primer sequence . | Target gene . |
---|---|---|
Cryptosporidium | 18S rRNA | |
FIP CCTCGTTCAAGATCAATAATTGCAA-ATGGGTAATCTTTTGAATATGCA | ||
BIP TCCTAGTAAGCGCAAGTCATCAG-ATTCAATCGGTAGGAGCG | ||
F3 GTATATATTCCTGTTTCGAAGGA | ||
B3 TCCGAATAATTCACCGGATC | ||
LB GCTGATTACGTCCCTGCCCTTTG | ||
Corresponding nucleotide position of Cryptosporidium parvum 18 s rRNA gene, 1746 bp (Accession No. 16996.1) | ||
Shigella | ipaH | |
FIP AAGCTCCGCAGAGGCACTGA-CACGCAATACCTCCGGATTC | ||
BIP AGCAGTCTTTCGCTGTTGCTGC-CCGGAGATTGTTCCATGTGA | ||
F3 GCCTTTCCGATACCGTCTCT | ||
B3 TGATGGACCAGGAGGGTT | ||
LF TGCAGCGACCTGTTCACG | ||
LB CACTGAGAGCTGTGAGGACCG | ||
Corresponding nucleotide position of Shigella flexneri ipaH gene, 1312 bp (Accession No. M76443.1) | ||
Toxin-producing strain of E. coli | Stx2A and stx2B | |
FIP AGACGAAGATGGTCAAAACGC-GCAGTTATTTTGCTGTGGA | ||
BIP CGGGTTCGTTAATACGGCAA-CGGGCACTGATATATGTGT | ||
F3 TCGGTGTCTGTTATTAACCA | ||
B3 TGGAAACCGTTGTCACAC | ||
LF TGATAGACATCAAGCCCTCGTA | ||
Corresponding nucleotide position of Escherichia coli stx2A gene for Shiga toxin 2 A-subunit and stx2B gene for Shiga toxin 2 B-subunit, strain CB10686, 1450 bp (Accession No. FN252458.1) | ||
Rotavirus | VP7 glycoprotein | |
FIP TTGGTTGCTAGCTTCAATTGGATAA-CGCATATGCTAACTCTACTCAA | ||
BIP TGGTGAATGGAAAGATACATTGTCA-TGTACTCTTTAAAGTAGACCGAT | ||
F3 CCAATAACAGGATCAATGGATAC | ||
B3 GGATCGACAGAAAATTCAACAA | ||
LB TGTTTCTTACAAAAGGCTGGCCAAC | ||
Corresponding nucleotide position of Human rotavirus A gene for VP7, 981 bp (Accession No. AB018697.1) |
Primer name . | Primer sequence . | Target gene . |
---|---|---|
Cryptosporidium | 18S rRNA | |
FIP CCTCGTTCAAGATCAATAATTGCAA-ATGGGTAATCTTTTGAATATGCA | ||
BIP TCCTAGTAAGCGCAAGTCATCAG-ATTCAATCGGTAGGAGCG | ||
F3 GTATATATTCCTGTTTCGAAGGA | ||
B3 TCCGAATAATTCACCGGATC | ||
LB GCTGATTACGTCCCTGCCCTTTG | ||
Corresponding nucleotide position of Cryptosporidium parvum 18 s rRNA gene, 1746 bp (Accession No. 16996.1) | ||
Shigella | ipaH | |
FIP AAGCTCCGCAGAGGCACTGA-CACGCAATACCTCCGGATTC | ||
BIP AGCAGTCTTTCGCTGTTGCTGC-CCGGAGATTGTTCCATGTGA | ||
F3 GCCTTTCCGATACCGTCTCT | ||
B3 TGATGGACCAGGAGGGTT | ||
LF TGCAGCGACCTGTTCACG | ||
LB CACTGAGAGCTGTGAGGACCG | ||
Corresponding nucleotide position of Shigella flexneri ipaH gene, 1312 bp (Accession No. M76443.1) | ||
Toxin-producing strain of E. coli | Stx2A and stx2B | |
FIP AGACGAAGATGGTCAAAACGC-GCAGTTATTTTGCTGTGGA | ||
BIP CGGGTTCGTTAATACGGCAA-CGGGCACTGATATATGTGT | ||
F3 TCGGTGTCTGTTATTAACCA | ||
B3 TGGAAACCGTTGTCACAC | ||
LF TGATAGACATCAAGCCCTCGTA | ||
Corresponding nucleotide position of Escherichia coli stx2A gene for Shiga toxin 2 A-subunit and stx2B gene for Shiga toxin 2 B-subunit, strain CB10686, 1450 bp (Accession No. FN252458.1) | ||
Rotavirus | VP7 glycoprotein | |
FIP TTGGTTGCTAGCTTCAATTGGATAA-CGCATATGCTAACTCTACTCAA | ||
BIP TGGTGAATGGAAAGATACATTGTCA-TGTACTCTTTAAAGTAGACCGAT | ||
F3 CCAATAACAGGATCAATGGATAC | ||
B3 GGATCGACAGAAAATTCAACAA | ||
LB TGTTTCTTACAAAAGGCTGGCCAAC | ||
Corresponding nucleotide position of Human rotavirus A gene for VP7, 981 bp (Accession No. AB018697.1) |
Oligonucleotide primers synthesis
Laboratory-made DNA or RNA strands or oligonucleotide primers for each targeted gene were prepared based on the designed primer sequence. All oligonucleotide primers were synthesized by Africa's Genomics Company, Inqaba Biotec East Africa Ltd.
Drinking water sample collection
The drinking water samples were taken from two locations: the primary infant drinking vessels at the site of intake in the home and the corresponding water source. One liter of water was collected in a sterile bottle with no chemical additives. Water samples were collected and stored in accordance with the ICR (Information Collection Rule) guidance (Fout et al. 1996). Briefly, before moving on to the next phase, the obtained water samples were kept at a temperature of 1–4 °C in a dark, cold box and transported within 8 h after the time of collection.
Filtration
Once the water sample was received, filtration was carried out for each sample using a Zeta Plus 1MDS electropositive microfilter media disc with a 47-mm size (3 M Purification Inc., Meriden, CT, U.S.A.). This filter can simultaneously capture and recover multiple microbes from a water sample (Polaczyk et al. 2007). This filter disc was placed on the sterile funnel unit of the membrane filter support disc assembly. The filtration was carried out by applying a vacuum pump until the water sample was finished, and finally, the filter disc was removed using sterile forceps and placed in a sterile petri dish. This is followed by storing at 4°C in the dark for not more than 12 days before undergoing the elution process (http://www.lenntech.com).
Elution
An eluting solution was used to remove the adsorbed pathogenic microbes from the filter disc surface. For a single elution, a 200-mL eluting/backflushing solution was made by mixing [50 mL 1.5% beef extract (Accumix, CA-69574, India) with 0.05 M glycerin (Fluka Chemika, CA-49780, Switzerland) – autoclaved at 121 °C for 15 min + 50 mL of 0.01% Tween 80 (Mana Scientific Product, CA-9005-65-6, India) + 50 mL of 0.1% sodium polyphosphate/NaPP (Sigma Aldrich, CA-305553, USA) + 50 mL of 0.001% antifoam agent (Bio-Rad Laboratories, CA-94547, USA) to reduce foaming effect]. The pH of the eluting solution was adjusted to 8 using an Accumet pH benchtop meter (Thermo Fischer Scientific, Canada), which could reduce the potential destruction of bacteria (Polaczyk et al. 2007). The cleaned and sanitized bronze filtration support disc housing was used to hold the adsorbed filter disc once it had been mated with the filter funnel. The eluent solution was allowed to come into contact with the filter disc for 10 min prior to starting the elution pumping procedure. The eluent solution was then added to the filter funnel and pumped into the sterilized filter housing in the opposite direction of the sample flow until it became clear. The 200-mL eluent was finally put into four sterile 50-mL Falcon tubes (Merck KGaA, Corning 430290, Germany) to be centrifuged (Eppendorf AG- 22331, Hamburg, Germany) for concentration of targeted pathogens in the filtered eluate.
Concentration of pathogens from the filter eluate (centrifuge)
The eluted effluent underwent a secondary concentration stage over several centrifugation cycles. A preliminary centrifugation (Eppendorf AG- 22331, Hamburg, Germany) was performed on four tubes (Merck KGaA, Corning 430290, Germany) containing 50 mL of eluate at 4,100 × g for 30 min. Following that, the Falcon tubes (Merck KGaA, Corning 430290, Germany) were carefully withdrawn from the centrifuge (Eppendorf AG- 22331, Hamburg, Germany), and the supernatant was properly eliminated ∼25 mL from each one using a serological pipette. The aliquots of the re-suspended materials were pipetted up and down each tube to transfer them into two 50 mL tubes (Merck KGaA, Corning 430290, Germany), after which a second centrifugation (Eppendorf AG- 22331, Hamburg, Germany) was performed under the identical circumstances. The residual material was gently shacked and transferred into two sterile 15 mL tubes before being centrifuged at 4,100 g for 30 min (AG- 22331, Hamburg, Germany) with the supernatant, which had been carefully aspirated to a volume of 35 mL. In order to avoid disrupting the pellets, the supernatant was once again aspirated to a total of 12 mL from each of the 15 mL tubes. After being centrifuged at 4,100 g for 30 min, the leftover supernatant (3 mL) was placed into a single 15 mL tube (Merck KGaA, Corning 430055, Germany) and removed. The supernatant was then gently aspirated into each of the 15 mL tubes (Merck KGaA, Corning 430055, Germany) until around 300 μL remained, and the tubes were then kept at −20°C pending DNA/RNA extraction.
DNA/RNA extraction
Genomic DNA for Cryptosporidium, Shigella, and a toxin-producing strain of E. coli and RNA for rotavirus were extracted from 300 μL eluate (concentrated water samples) using the DaAN Gene RNA/DNA Purification Kit (MDSS GmbH, Germany), according to the manufacturer's instructions. The RNA genome for rotavirus was converted to cDNA using first-strand cDNA synthesis protocols [E6300] (New England BioLabs Inc.). This conversion is needed for the reason that cDNA is a more convenient way to work with the coding sequence than RNA, which could be very easily degraded by omnipresent RNases. The extracted DNA/RNA and cDNA samples were stored at –20°C until amplification was done.
LAMP assay
Reaction mixture for LAMP (operated on ice)
The reaction mixture consisted of all the components necessary to make new strands of DNA/cDNA in the LAMP process. The reaction mixture is composed of the following reagents: (1) Master Mix, containing 1× Thermopol reaction buffer (10 mM KCl, 10 mM (NH4)2SO4, 20 mM Tris-HCl, 2.0 mM MgSO4, 0.1% Triton X-100; New England Biolab), 6 mM MgSO4 (100 mM;New England Biolabs Inc.), 0.8 M betaine (Glentham Life Sciences Ltd, Corsham, UK), 1.4 mM dNTP's (10 mM; New England Biolabs Inc.), Bst 2.0 DNA polymerase large fragment (8 U; New England Biolabs Inc), 50× LAMP fluorescent dye (New England Biolabs Inc.), Primer mixture [1.6 μM FIP, 1.6 μM BIP, 0.4 μM LPF, 0.4 μM LFB, 0.2 μM F3 and 0.2 μM B3; Africa's Genomics Company, Inqaba Biotec East Africa Ltd], Nuclease-free H2O (Fisher Scientific, CAS 7732-18-5, USA) and (2) Template genomic DNA/cDNA. The reaction mixture for each target pathogen was prepared in an ice-filled box to a final volume of 25 μL (see Table 2).
Individual reaction mixture composition of the LAMP assays for the detection of the selected pathogens
Prepare Mixtures . | Quantity (μL) . | |||
---|---|---|---|---|
Cryptosporidium . | Shigella . | Toxin-producing strain of E. coli . | Rotavirus . | |
1. Primer mixture ingredient (per reaction) | ||||
1.6 μM FIP | 1.0 μL | 0.8 μL | 1.0 μL | 1.0 μL |
1.6 μM BIP | 1.0 μL | 0.8 μL | 1.0 μL | 1.0 μL |
0.4 μM LPF | – | 0.4 μL | 1.0 μL | 1.0 μL |
0.4 μM LFB | 1.0 μL | 0.4 μL | – | – |
0.2 μM F3 | 1.0 μL | 0.1 μL | 1.0 μL | 1.0 μL |
0.2 μM B3 | 1.0 μL | 0.1 μL | 1.0 μL | 1.0 μL |
Total | 5.0 μL | 2.6 μL | 5.0 μL | 5.0 μL |
2. Create the reaction mix ingredients (per reaction) | ||||
1 × Isothermal Buffer | 2.5 μL | 2.5 μL | 2.5 μL | 2.5 μL |
6 mM MgSO4 | 1.5 μL | 1.5 μL | 1.5 μL | 1.5 μL |
1.4 mM DNTPs | 3.5 μL | 3.5 μL | 3.5 μL | 3.5 μL |
0.8 M Betaine | 4.0 μL | 4.0 μL | 4.0 μL | 4.0 μL |
Bst 2.0 DNA polymerase | 1.0 μL | 1.0 μL | 1.0 μL | 1.0 μL |
Nuclease-free H2O | 4.0 μL | 6.4 μL | 2.0 μL | 2.0 μL |
Total | 16.5 μL | 18.9 μL | 14.5 μL | 14.5 μL |
3. Template DNA | 3 μL | 3 μL | 5.0 μL | – |
4. Template cDNA | – | – | – | 5.0 μL |
5. LAMP fluorescent dye | 0.5 μL | 0.5 μL | 0.5 μL | 0.5 μL |
Overall reaction mixture | 25 μL | 25 μL | 25 μL | 25 μL |
Prepare Mixtures . | Quantity (μL) . | |||
---|---|---|---|---|
Cryptosporidium . | Shigella . | Toxin-producing strain of E. coli . | Rotavirus . | |
1. Primer mixture ingredient (per reaction) | ||||
1.6 μM FIP | 1.0 μL | 0.8 μL | 1.0 μL | 1.0 μL |
1.6 μM BIP | 1.0 μL | 0.8 μL | 1.0 μL | 1.0 μL |
0.4 μM LPF | – | 0.4 μL | 1.0 μL | 1.0 μL |
0.4 μM LFB | 1.0 μL | 0.4 μL | – | – |
0.2 μM F3 | 1.0 μL | 0.1 μL | 1.0 μL | 1.0 μL |
0.2 μM B3 | 1.0 μL | 0.1 μL | 1.0 μL | 1.0 μL |
Total | 5.0 μL | 2.6 μL | 5.0 μL | 5.0 μL |
2. Create the reaction mix ingredients (per reaction) | ||||
1 × Isothermal Buffer | 2.5 μL | 2.5 μL | 2.5 μL | 2.5 μL |
6 mM MgSO4 | 1.5 μL | 1.5 μL | 1.5 μL | 1.5 μL |
1.4 mM DNTPs | 3.5 μL | 3.5 μL | 3.5 μL | 3.5 μL |
0.8 M Betaine | 4.0 μL | 4.0 μL | 4.0 μL | 4.0 μL |
Bst 2.0 DNA polymerase | 1.0 μL | 1.0 μL | 1.0 μL | 1.0 μL |
Nuclease-free H2O | 4.0 μL | 6.4 μL | 2.0 μL | 2.0 μL |
Total | 16.5 μL | 18.9 μL | 14.5 μL | 14.5 μL |
3. Template DNA | 3 μL | 3 μL | 5.0 μL | – |
4. Template cDNA | – | – | – | 5.0 μL |
5. LAMP fluorescent dye | 0.5 μL | 0.5 μL | 0.5 μL | 0.5 μL |
Overall reaction mixture | 25 μL | 25 μL | 25 μL | 25 μL |
Optimization of LAMP assay for each pathogen
To establish the best amplification reaction conditions, the concentrations, temperature, and time of the reaction mixture were optimized for the LAMP assay. The reaction mixture for each targeted pathogenic microbe was tested under different conditions of concentration, amplification temperature, and time. The temperature tested ranged from 60 to 65°C for 60 min, followed by a heat inactivation step at 80°C for 5–10 min. The genomic DNAs/RNA-cDNA extracted from the positive control of each targeted pathogen and DNA/RNA extracted from nuclease-free water (Fisher Scientific, CA-7732-18-5, USA) as a negative control were used for the optimization of LAMP. After several trials, the optimized LAMP reaction mixture condition for each targeted pathogen was presented in Table 2. The optimum temperature and time for the best amplification result for Shigella species, Cryptosporidium oocysts, and rotaviruses were 60 °C for 60 min followed by 80 °C heat inactivation for 5 min while a toxin-producing strain of E. coli was best worked at 64°C for 60 min followed by inactivation at 80°C for 5 min. In all optimization procedures, there was no amplification result observed in the negative control.
Positive and negative control
To determine the positive control, the live pathogen strains (Cryptosporidium, ATCC PRA-67DQ; Shigella, ATCC 12022; a toxin-producing strain of E. coli, ATCC 43895; and rotavirus, a positive stool sample stored in a laboratory) were obtained from the Ethiopian Institute of Public Health (EPHI) and Ethiopian Biodiversity Institute (EBI) biobanking repository laboratories. This strain was diluted in one liter of distilled water and processed by filtration, elution, and centrifuging (Eppendorf AG-22331, Hamburg, Germany), followed by DNA/RNA extraction (MDSS GmbH, Germany), and amplification. A negative control was done using nuclease-free water (Fisher Scientific, CA-7732-18-5, USA) instead of the DNA template. A positive control and a negative control were included in every LAMP reaction to ensure that false-negative and false-positive results were eliminated (see Figure 2).
LAMP operational procedure and reaction
The final concentration of the LAMP reaction mixture of 25 μL was dispensed into each labeled Loopamp reaction tube (EIKEN Chemical Co., Ltd Tochigi, Japan). The reaction was carried out by inserting the reaction tubes in a Bio-RAD iCycler Thermal Cycler (Bio-Rad Laboratories, Inc. USA). Based on the optimization result for Shigella, Cryptosporidium, and rotavirus, the reaction was run by heating at 95°C for 3 min, subsequently incubating at 60°C for 60 min, inactivating at 80°C for 5 min and then cooling at 4°C for 5 min to terminate the reaction. The same procedure was undertaken for the detection of a toxin-producing strain of E. coli by activating at 64°C for 60 min and inactivating at 80 °C for 5 min.
Detection using UV transilluminator
After completion of LAMP, amplified DNA/cDNA were detected using a UV transilluminator (SYNGENE, Synoptics Ltd, UK) and photographed, and finally the numbers were noted down for all positive samples.
Sensitivity and specificity of LAMP
The LAMP reaction sensitivity greatly relies on the previous studies that explored the high sensitivity of LAMP for Cryptosporidium oocyst (Shi et al. 2012), Shigella species (Song et al. 2005), toxin-producing strains of E. coli (Song et al. 2005; Mahony et al. 2016; Ranjbar et al. 2016), and RNA-cDNA of rotavirus (Malik et al. 2013).
The specificity of the LAMP assays for Shigella was examined with three closely related bacterial species: Salmonella species, Vibrio cholerae, and toxin-producing E. coli. The result demonstrated a positive for only Shigella species and a negative for the aforementioned bacterial species. The specificity of the toxin-producing strain of E. coli was determined by testing it with three other bacterial species: Salmonella species, Vibrio cholerae, and Shigella species, which solely resulted in a positive for the toxin-producing strain of E. coli and a negative for all the other bacterial species. The LAMP assay for Cryptosporidium oocyst and rotavirus was examined with Giardia cysts and adenovirus, respectively, which demonstrated only positive results for Cryptosporidium oocyst and rotavirus. Therefore, the assay established in this study was found to detect only the sequence of the targeted pathogens and no cross-reaction with other pathogenic microbes, representing its high specificity.
Quality control for LAMP
For the laboratory quality assurance of the LAMP assay, the ICR microbial laboratory manual was followed (Fout et al. 1996). The operating procedures were kept exactly as stated in the protocol. All LAMP laboratory equipment was maintained in a safe and working condition. Each DNA/RNA extraction and LAMP test was carried out by knowledgeable and skilled personnel in a secured separate room set-up to avoid contamination. Samples with positive results were confirmed by analyzing them with the same procedures as the positive control. All test results were photographed and recorded directly into an electronic registry format. The methods of recovery for each organism comprised of filter material, eluent, and centrifuge steps were applied considering the other study (Polaczyk et al. 2007), which can yield the highest recovery efficiency.
Operational definitions
Oligonucleotide: The term ‘oligonucleotide’ or ‘oligo’ usually refers to a synthetic laboratory-made DNA or RNA strand (Bio-Synthesis Inc. 2014).
Molecular methods: Can be defined as those methods that target macromolecules containing information about the identity of the microorganisms that produce them (Nocker et al. 2009).
Loop-mediated isothermal amplification (LAMP): Is a rapid molecular method in which a reaction is performed in a single tube under isothermal condition for the amplification of nucleic acid (DNA or RNA) by using Bst DNA polymerase with strand displacement for the detection of certain pathogens (Notomi et al. 2000).
Optimization: Determination of optimum concentrations, temperature and time at which the LAMP functions to its best capacity both in terms of yield and specificity.
Primer: Is a short single-stranded DNA fragment used in molecular laboratory techniques in which pair of primers hybridizes with the sample DNA and defines the region that will be amplified, resulting in millions and millions of copies in a very short timeframe (National Human Genome Research Institute).
Drinking Water Treatment: Boiling, add bleach/chlorine, water filter (ceramic, sand, composite), and solar disinfection (World Health Organization).
Safe Water Storage: Water stored in plastic, clay or metal pot narrow mouth (usually diameter of 3 cm or less), with a lid or secured cover and a tap (spigot), cleaned and kept covered (Centers for Disease Control and Prevention (CDC) 2011), while its reverse could be taken as unsafe water storage.
Improved water source: Defined as those that are likely to be protected from outside contamination, and from fecal matter in particular. This includes piped water into dwelling/yard/plot, public tap/standpipe, tubewell/borehole, protected dug well, protected spring, rainwater and bottled water; whereas unimproved water source includes unprotected dug well, unprotected spring, surface water (river, dam, lake, pond, stream, canal, irrigation channels), cart with small tank/drum, tanker-truck (World Health Organization).
RESULTS
General characteristics of study population and households
This study included 410 infant-containing households, resulting in a 97.0% response rate. The reasons for the non-response were a refusal to engage in the interview and the disposal of some water samples. The study subjects’ mean age (±SD) was 8.35 (±2.59) months. The majority of the study population fell within the age group of 6–12 months at 73.4%. The sex composition consisted of 217 (52.9%) males and 193 (47.1%) females, making the overall sex ratio 1.1:1.
Among the 410 households with infants surveyed, the majority (37.3%) were getting their drinking water from a piped public tap/standpipe. This was followed by unprotected springs, used by 28.8% of the households; protected springs (12.0%); and piped water in their own yard or premises (10.2%). Almost one-third of the households (32.7%) have no access to an improved source of drinking water. Nearly all (99.8%) of the households did not treat drinking water before they provided it to the infant. The majority of the households (94.9%) were observed to be storing their drinking water in an unsafe manner. About 71.7% of the households had fetched water at the present time prior to the survey. Most of the households (83.7%) collected their water from water points in public spaces. Almost all (99.8%) of the household water samples tested had residual chlorine that was not within the standard required level (between 0.2 and 0.5 mg/L) (see Table 3).
Demographic and drinking water characteristics in the study population
Characteristics . | Frequency (n) . | Percentage (%) . |
---|---|---|
Age of infant | ||
< 6 months | 109 | 26.6 |
6–12 months old | 301 | 73.4 |
Sex of infant | ||
Male | 217 | 52.9 |
Female | 193 | 47.1 |
Primary water source type | ||
Piped water into dwelling | 4 | 1.0 |
Piped water into yard or plot | 42 | 10.2 |
Piped water, public tap/standpipe | 153 | 37.3 |
Piped water kiosk or retailer | 7 | 1.7 |
Protected dug well | 8 | 2.0 |
Protected spring | 49 | 12.0 |
Bottled water | 13 | 3.2 |
Unprotected dug well | 1 | 0.2 |
Unprotected spring | 118 | 28.8 |
River water | 2 | 0.5 |
Irrigation canal | 13 | 3.2 |
Water source improvement status | ||
Improved water source | 276 | 67.3 |
Unimproved water source | 134 | 32.7 |
Point-of-use water treatment practices | ||
Do not treat | 409 | 99.8 |
Treat Water | 1 | 0.2 |
Drinking water storage | ||
Unsafe | 389 | 94.9 |
Safe | 21 | 5.1 |
When water is fetched? | ||
Same day | 294 | 71.7 |
One day ago | 87 | 21.2 |
Days ago | 29 | 7.1 |
Where the water point is found? | ||
In the dwelling | 17 | 4.1 |
Private yard/plot | 37 | 9.0 |
Neighbor's yard/shared compound | 13 | 3.2 |
Public Space | 343 | 83.7 |
Residual chlorine | ||
Not between 0.2 and 0.5 mg/L | 409 | 99.8 |
Between 0.2 and 0.5 mg/L | 1 | 0.2 |
Characteristics . | Frequency (n) . | Percentage (%) . |
---|---|---|
Age of infant | ||
< 6 months | 109 | 26.6 |
6–12 months old | 301 | 73.4 |
Sex of infant | ||
Male | 217 | 52.9 |
Female | 193 | 47.1 |
Primary water source type | ||
Piped water into dwelling | 4 | 1.0 |
Piped water into yard or plot | 42 | 10.2 |
Piped water, public tap/standpipe | 153 | 37.3 |
Piped water kiosk or retailer | 7 | 1.7 |
Protected dug well | 8 | 2.0 |
Protected spring | 49 | 12.0 |
Bottled water | 13 | 3.2 |
Unprotected dug well | 1 | 0.2 |
Unprotected spring | 118 | 28.8 |
River water | 2 | 0.5 |
Irrigation canal | 13 | 3.2 |
Water source improvement status | ||
Improved water source | 276 | 67.3 |
Unimproved water source | 134 | 32.7 |
Point-of-use water treatment practices | ||
Do not treat | 409 | 99.8 |
Treat Water | 1 | 0.2 |
Drinking water storage | ||
Unsafe | 389 | 94.9 |
Safe | 21 | 5.1 |
When water is fetched? | ||
Same day | 294 | 71.7 |
One day ago | 87 | 21.2 |
Days ago | 29 | 7.1 |
Where the water point is found? | ||
In the dwelling | 17 | 4.1 |
Private yard/plot | 37 | 9.0 |
Neighbor's yard/shared compound | 13 | 3.2 |
Public Space | 343 | 83.7 |
Residual chlorine | ||
Not between 0.2 and 0.5 mg/L | 409 | 99.8 |
Between 0.2 and 0.5 mg/L | 1 | 0.2 |
LAMP results in infants' point-of-use water samples
Prevalence of waterborne pathogens from infant drinking water samples.
LAMP results in water sources
The study included a total of 37 water samples that were collected from various types of water sources used by the households for infant ingestion in various geographic locations. Among the 37 water sources, the majority of 16 (43.2%) were a public tap/standpipe designated from borehole followed by an unprotected spring 11 (29.7%), a protected spring 5 (13.5%) and a reservoir designated from boreholes 2 (5.4%). A few water sources were protected dug well, an irrigation canal, and a river, each accounting for 1(2.7%). About 35.1% of the water samples were observed to be drawn from unimproved water sources. The majority of the reported water sources (83.8%) did not undergo routine cleaning. The majority (97.3%) of the water sources were untreated on a regular basis, and 73% lacked catchment or fencing.
Out of the 37 water source samples tested, the result showed that 10 (27.0%), 12 (32.4%), 11 (29.7%), and 14 (37.8%) were positive for Cryptosporidium oocysts, Shigella species, toxin-producing strains of E. coli, and rotavirus, respectively. In total, 13.5% of the water samples tested from water sources was positive for all four pathogens (see Figure 2).
Correlation of pathogens presence between water source and point-of-use water sample
The phi coefficient correlation showed a significant positive relationship between the primary water source and infant drinking water at the point-of-use in the household for the presence of targeted pathogens. The presence of Cryptosporidium in water source samples and drinking water at the point-of-use, thus, showed a statistically significant and moderately positive correlation (Phi = 0.527; p = 0.000). The presence of Shigella species showed a strong positive correlation between the water source and the point-of-use (Phi = 0.524; p = 0.000). The presence of a toxin-producing strain of E. coli also significantly correlated between water source samples and point-of-use at the household level (Phi = 0.424; p = 0.000). Between the water source and the point-of-use drinking water at the household level, rotavirus exhibits a significant positive correlation, although one that is weak in magnitude (Phi = 0.113; p = 0.023).
Multivariable's conditional logistic regression analysis of targeted pathogens with selected explanatory water quality determinants
In binary logistic regression, bivariate analysis indicated that an unimproved water source and the length of time water was retained before use were significantly associated with the occurrence of Cryptosporidium, Shigella, and toxin-producing strains of E. coli. Variables such as unsafe water storage, the retention time of water before use, and fetching water in public spaces showed a significant association with the presence of rotavirus. An unimproved water supply and the water retained before an infant drinks were significantly associated with the presence of overall pathogens in drinking water.
After adjustment, in the multivariable conditional logistic regression, households with infants who used water from unimproved sources were significantly more likely to increase the presence of Cryptosporidium oocysts (p = 0.000, AOR (adjusted odds ratio): 4.02, 95% CI: 2.29–7.04), Shigella species (p = 0.000, AOR: 4.21, 95% CI: 2.43–7.29) and toxin-producing strains of E. coli (p = 0.004, AOR: 2.14, 95% CI: 1.27–3.59). However, no statistically significant association was observed in households using an unimproved water source with the occurrence of rotavirus. The household water samples that tested positive for Cryptosporidium were significantly associated with those who fetched water days ago (p = 0.001, AOR: 0.03, 95% CI: 0.00–0.26) and 1 day ago (p = 0.010, AOR: 0.51, 95% CI: 0.30–0.85). Similarly, the presence of a toxin producing strain of E. coli showed a significant relationship with fetching water on the preceding days (p = 0.008, AOR: 0.18, 95% CI: 0.05–0.64). The presence of rotavirus showed a significant association with those households that fetched water 1 day ago (p = 0.002, AOR: 0.46, 95% CI: 0.28–0.76) prior to the survey time. No significant relationship is observed between the presence of each targeted microbial pathogen in drinking water samples and the household's drinking water storage status as well as the household's fetching water places (see Table 4).
Multivariable logistic regression analysis for the presence of targeted pathogens in water samples and water quality determinants in eastern Ethiopia
Variables . | Model 1 AOR (95% CI) . | Model 2 AOR (95% CI) . | Model 3 AOR (95% CI) . | Model 4 AOR (95% CI) . | Final Model AOR (95% CI) . |
---|---|---|---|---|---|
Water Source | |||||
Unimproved | 4.02 (2.29, 7.04)* | 4.21 (2.43, 7.29)* | 2.14 (1.27, 3.59)* | – | 3.96 (1.73, 9.07)* |
Improved | 1 | 1 | 1 | 1 | |
Drinking Water Storage | |||||
Unsafe | – | – | – | 1.78 (0.68,4.68) | – |
Safe | 1 | ||||
When water is fetched | |||||
Days ago | 0.03 (0.00,0.26)* | – | 0.18 (0.05,0.64)* | 1.08 (0.46,2.53) | – |
Yesterday | 0.51 (0.30,0.85)* | 0.66 (0.39,1.11) | 0.65 (0.39,1.09) | 0.46 (0.28,0.76)* | 0.44 (0.23,0.81)* |
Today | 1 | 1 | 1 | 1 | 1 |
Where the water point is found? | |||||
Public Space | – | – | – | 1.07 (0.59,1.97) | – |
In the dwelling/yard | 1 |
Variables . | Model 1 AOR (95% CI) . | Model 2 AOR (95% CI) . | Model 3 AOR (95% CI) . | Model 4 AOR (95% CI) . | Final Model AOR (95% CI) . |
---|---|---|---|---|---|
Water Source | |||||
Unimproved | 4.02 (2.29, 7.04)* | 4.21 (2.43, 7.29)* | 2.14 (1.27, 3.59)* | – | 3.96 (1.73, 9.07)* |
Improved | 1 | 1 | 1 | 1 | |
Drinking Water Storage | |||||
Unsafe | – | – | – | 1.78 (0.68,4.68) | – |
Safe | 1 | ||||
When water is fetched | |||||
Days ago | 0.03 (0.00,0.26)* | – | 0.18 (0.05,0.64)* | 1.08 (0.46,2.53) | – |
Yesterday | 0.51 (0.30,0.85)* | 0.66 (0.39,1.11) | 0.65 (0.39,1.09) | 0.46 (0.28,0.76)* | 0.44 (0.23,0.81)* |
Today | 1 | 1 | 1 | 1 | 1 |
Where the water point is found? | |||||
Public Space | – | – | – | 1.07 (0.59,1.97) | – |
In the dwelling/yard | 1 |
Factors that hadn't p-value of <0.2 from bivariate analysis and not eligible in multivariable.
Model 1: Association of factors with the presence of Cryptosporidium oocyst in the water sample.
Model 2: Association of factors with the presence of Shigella species in water sample.
Model 3: Association of factors with the presence of toxin-producing strain of E. coli in water sample.
Model 4: Association of factors with the presence of rotavirus in water sample.
Final Model: Association of factors with the presence of all targeted pathogens (Cryptosporidium, Shigella, toxin-producing strain of E. coli and rotavirus) in the sample water sample.
*Statistically significant at p < 0.05.
In the final model, the presence of the overall (four) pathogens among water samples showed a significant association with households using unimproved water sources, which were approximately four times (p = 0.001, AOR: 3.96, 95% CI: 1.73–9.07) higher than those households that used improved water sources. The water samples tested from the household's drinking water that were fetched 1 day ago were significantly associated with the presence of all targeted pathogens in infants' drinking water (p = 0.009, AOR: 0.44, 95% CI: 0.23–0.81) (see Table 4).
DISCUSSION
The detection of Cryptosporidium oocysts, rotavirus, Shigella species and a toxin-producing strain of Escherichia coli has significant public health implications because worldwide these waterborne pathogens are most frequently associated with severe infantile diarrhea (McNeil 2013). Molecular analytical techniques are useful tools for evaluating the microbial quality of water (Li et al. 2015). Among a wide range of molecular techniques, nucleic acid-based amplification methods comprised of PCR have been developed to detect microbial pathogenic species directly from drinking water samples (Botes et al. 2013). Subsequently, a technology termed loop-mediated isothermal amplification (LAMP) has been developed, which is highly sensitive and accurate for the replication of DNA in isothermal conditions (Notomi et al. 2000). The present study was focused on the detection of the aforementioned pathogens in infants' drinking water at the point-of-use in the household and the water sources from which they were collected using the LAMP technique.
Rotaviruses were detected in drinking water supplies used for infants in about 32.2% of households (Figure 1) which is higher than previously reported for Shandong, China, at 16.7% using a similar technique (Yang et al. 2013), southeast France at 7.1% (Gratacap-Cavallier et al. 2000), Karachi, Pakistan, at 5% using ELISA (Yousuf et al. 2017), and Colombia at 27.3% and 20.5% (Toranzos et al. 1986; Pelaez-Carvajal et al. 2016). The presence of rotavirus in the 37.8% of drinking water source samples observed in this study was much higher than previously reported results tested using PCR in Benin at 2.1% (Verheyen et al. 2009), Peshawar, Pakistan, at 9.47% (Ahmad et al. 2016), Karachi, Pakistan, at 23% (Rashid et al. 2021), Beijing, China, at 20.3% (He et al. 2009), Costa Rica at 8.1% applying somatic coliphage (Barrantes et al. 2022), Faisalabad, Pakistan, at 26.6% using Latex agglutination test (Shoaib et al. 2019), Southern Africa at 2.0% (Van Zyl et al. 2006), and Egypt at 15.6, 8.3, and 23.3% using RT-PCR and multiplex semi-nested RT-PCR (Mahmoud et al. 2019; Rizk & Allayeh 2018; Shaheen 2019), while lower than the study in Ghana (Dongdem et al. 2010), where 48.1% of water samples tested by multiplex RT-PCR were positive for rotaviruses. This disparity in the prevalence status might be explained by differences between geographic areas with various contributing factors such as socio-economic and cultural factors, access coverage of water, sanitation, and hygiene behavioral practices, and also by the use of different techniques for rotavirus detection having dissimilar sensitivity. Remarkably, the rotavirus prevalence in our study appeared to be relatively higher than the other detected microbial pathogens, both in water samples at the point-of-consumption and from water sources. This might be attributed to the highest spreading phenomena of rotaviruses in the environment of the study area. Rotavirus is excreted in enormous quantities in the feces of infected person, at a rate of up to 1011 virus particles per gram (Gratacap-Cavallier et al. 2000). Rotavirus is very resistant to diverse environmental conditions and physicochemical treatment processes that are able to stabilize the virus and make it present in large amounts in environmental water (Yang et al. 2013). Studies indicated that rotavirus survive well enough in chlorine-based conventionally treated drinking water to make it a possible vehicle for their transmission (Sattar et al. 1984). In addition, rotavirus has high infectivity and an increased risk of transmission in comparison with protozoa and bacteria (Francy et al. 2011).
Furthermore, our study revealed no significant association between households using drinking water from an unimproved water source and the presence of rotavirus in a water sample, as observed with other detected protozoa (Cryptosporidium) and bacteria (Shigella and a toxin-producing strain of E. coli). Thus, it is reasonable to assume that households' access to an improved water source is not a guarantee that it is always safe (Shaheed et al. 2014). The only significant relationship was observed in retention time, implying a household's fetching water 1 day prior to the survey has a protective effect for the presence of rotavirus in the water sample. This might be due to the influence of temperature and water movement, such as large, slow-moving, or stagnant sources, on rotavirus survival and transmission (Kraay et al. 2018).
Our study also detected Cryptosporidium oocysts in 28.5% of the household's drinking water provided to infants. This may have implications for the poor drinking water quality status in the study area as the World Health Organization (WHO 2009) categorizes Cryptosporidium as a reference pathogen for the assessment of drinking water quality (Sattar et al. 1984). Regardless of the population group that consumed the water, the result of our study is much higher than previous findings of the studies conducted in Shanghai, China, at 0% applying Immunoflurescent assay (Zhang et al. 2010) and Mongolia at 2.0% using multiplex real-time PCR (Barnes et al. 2021). In the corresponding water sources, about 27% of the sample were found positive for Cryptosporidium oocyst, which is higher than the prevalence study that was conducted using flouresent microscopic examination reported in Tigray at 5% (Kifleyohannes & Robertson 2020), a 2008 study in Addis Ababa at 21% (Atnafu et al. 2012), and less than the study in Dire-Dawa at 58.9% (Amenu et al. 2013), while a 2012 study in Addis Ababa found that all (100%) water samples were positive (Fikrie et al. 2008). Our result appeared to be lower than the studies conducted elsewhere, such as in Egypt at 52.6% (Sakran et al. 2017) and Switzerland at 40% (Fuchslin et al. 2012). On the contrary, our result is higher than the study conducted in four countries in Southeast Asia, which came in at 24.4% using real-time PCR assays (Kumar et al. 2016). The authors who reviewed varying studies showed the prevalence of Cryptosporidium species in drinking water ranged from 1.4 to 100% (Johnson et al. 2008). Numerous reasons can cause the observed differences. One of the reasons for this rate discrepancy might be attributed to the practices of animal grazing in the area, which may be contributing to the excretion of oocysts from infected hosts in the environment, which can have a chance of contaminating water sources. Another reason could be that the LAMP technique for the detection of pathogens might also influence the magnitude of the result due to its high sensitivity.
Cryptosporidium is a protozoan parasite that exhibits various characteristics that support its extended survival in the environment (Armon et al. 2016). It can be transmitted through water in the oocyst form and is more resistant to environmental conditions and disinfection (Plutzer et al. 2010; Kothavade 2012). The infective dose is low (10–100 oocysts) (Lowery et al. 2000), and evidence suggests Cryptosporidium requires as few as 10–30 oocysts per 100 L of water for the possible existence of an outbreak (Haas & Rose 1995). It is evident that the households whose water was drawn from an unimproved water source for their infants' use had a significantly higher association with the presence of Cryptosporidium oocysts by four-fold than those households sourcing water from an improved source for their infants' consumption. This is indirectly in line with a study that suggested households, use of an improved water source had a protective effect on the presence of Cryptosporidium in drinking water (Barnes et al. 2021). Studies suggested that Cryptosporidium oocysts were found more in surface water than in other water sources (Kolören et al. 2017) and can survive for months (World Health Organization 2009), indicating that there could be a chance of having several moments to ingest. This is certainly attributable to the fact that unimproved water sources are more exposed to contamination by human and animal waste. Our study also further investigated the fact that households that fetched drinking water other than on the day of our survey had a significant relationship with the presence of Cryptosporidium oocysts in their drinking water. The possible reason for this finding might be the presence of unsafe water sources in the study area.
In terms of Shigella species detection, our results showed that 30% of the infant drinking water sample at the household tested positive for Shigella species. Apart from the specific study subject in the households, this result is higher than the previous studies conducted using biochemical tests in Shashemene rural districts at 0% (Negera et al. 2017) and Jigjiga city at 3.33% (Asfaw et al. 2016), Ethiopia, and somewhere else such as Quetta, Pakistan, at 22% (Saima et al. 2018), Peshawar, Pakistan, at 6.47% using PCR (Nisa et al.), Yaounde, Cameroon, at 0.24% (Nguendo-yongsi 2011), and the Vanda region of South Africa at 5% applying culture method (Potgieter et al. 2005), while slightly lower than a study in Egypt at 34.7% (Bahy et al. 2019). These indicate that the drinking water in the study area has a high chance of contamination due to poor water quality. On the other hand, Shigella species were detected in 32.4% of the water samples tested from the corresponding water sources. This finding is higher than the study conducted in Jijiga, which found 8% (Asfaw et al. 2016), 4.8% in Ziway (Mekonnen et al. 2014), and 6.5% in the rural districts of Shashemene (Negera et al. 2017) in Ethiopia. Several studies elsewhere, such as Kampala, Uganda (10%), using microfluidic quantitative PCR (Sadik et al. 2017), Sri Lanka (0%) (Mahagamage et al. 2020), Bangladesh (30%) (Sarker et al. 2019), Nigeria (9.4%) (Okpasuo et al. 2020) and Kenya (6.9%) (Wahome et al. 2014), which used culture and biochemical examination, have also shown lower contamination prevalence of Shigella species in drinking water sources than the present study. The highest prevalence of Shigella species in drinking water either at the point-of-use or from water sources in our study might be due to the use of an unimproved water source, a lack of water source protection, and unhygienic practices such as poor handling of water, no hand- washing, and unsafe water storage in the study area. In addition, the differences in testing methods among the studies might be able to influence the prevalence of pathogens detected. Most of the earlier studies detected this pathogen using the biochemical test and conventional PCR, which differ greatly from the LAMP technique in sensitivity (Song et al. 2005; Shi et al. 2012; Mahony et al. 2016; Ranjbar et al. 2016).
Furthermore, households that collected water from unimproved water sources were strongly associated with the presence of Shigella species by four times more than those that collected from improved sources. This finding is supported by the view expressed in other studies that unimproved water sources are more likely to be exposed to fecal contamination than improved water sources (Bain et al. 2014). This fecal contamination poses a greater risk due to the potential source of pathogens.
The other detected bacterial type, known as a toxin-producing strain of E. coli, was found positive in 26.3% of the drinking water sample from infant point-of-consumption in the households. This prevalence appeared to be slightly higher than the study conducted using conventional PCR in South Wollo, Ethiopia at 17% (Gemeda et al. 2022), and much higher prevalence than the studies in Ouagadougou, Burkina Faso at 1% (Bonkoungou et al. 2020), Trinidad at 2% using agglutination test (Welch et al. 2000), and Malaysian villages at 19% applying ELISA (Vadivelu et al. 1989). However, it is much lower than the result reported from a study conducted in Bangladesh, where 87.5% tested positive using real-time PCR (Lothigius et al. 2008). With regard to the water sources sampled, about 30% were contaminated with a toxin-producing strain of E. coli. This finding is higher than the studies in Ziway at 5.5% tested with culture method (Mekonnen et al. 2014) and Modjo at 4.2% (Mersha et al. 2010) in Ethiopia, South Africa at 25.6% (Momba et al. 2008), and Brazil at 6.2%, 1.0%, and 0.65% (Lascowski et al. 2013; Caroline et al. 2018; Moreira et al. 2020) using conventional PCR, and slightly lower than the study conducted in Uganda at 33% using real-time PCR (Sadik et al. 2017) and India at 33.3% (Ram et al. 2008), while it was much lower than studies in northern Ghana at 75% (Kichana et al. 2022) and southeast of the island of Puerto Rico at 52% applying multiplex PCR (Crespo-Medina et al. 2020). Similar reasons that were provided for other pathogens could explain the prevalence variance among studies. Evidently, the application of the LAMP technique may result in a high prevalence of pathogens in drinking water due to its higher sensitivity than other conventional methods such as PCR (Song et al. 2005; Shi et al. 2012; Mahony et al. 2016; Ranjbar et al. 2016). To our knowledge, a few or no similar studies were done with more recent molecular methods such as real-time PCR and digital PCR.
Our study also confirmed that households getting water from unimproved sources for infant consumption were more likely to be exposed to a toxin-producing strain of E. coli by two-fold than improved sources. This finding is indirectly consistent with the result reported in the previous studies, which suggested that improved water sources significantly reduce the detection of E. coli virulence genes in stored drinking water (Mattioli et al. 2014). This is because getting water from improved sources could reduce environmental contamination.
A positive significant correlation was observed between the water samples tested from infant point-of-use at household level and those of the corresponding water sources for the presence of Cryptosporidium, Shigella, and a toxin-producing strain of E. coli. This could be explained by the fact that the water sources might be subjected to fecal contaminations, which increase opportunities for the presence of pathogens in household drinking water. However, rotavirus shows a negligible positive correlation between the water source and drinking water at the household level. This circumstance might be explained by the fact that this pathogen holds additional favorable conditions to be present in drinking water at the household level through home-based contamination.
Generally, our results showed that 13.2% of the drinking water of infants at the household level was positive for all four pathogens. The simultaneous presence of these pathogens in drinking water is closely linked with the status of the water source. This is implying that households collecting water from unimproved sources for infant consumption were nearly four times more likely to have the presence of these pathogens as compared to those sourcing water from improved supplies. Unimproved water sources may have a chance to be contaminated by feces as well as other matters that could allow different pathogens to exist in drinking water. Studies indicated that the quality of water is much more reduced at the point-of-use in the household where source water is largely contaminated (Wright et al. 2004). In addition, a lack of practices for any form of water treatment at the household level in the study area may also contribute to the presence of these pathogens in drinking water (Mengistie et al. 2013).
Strength and limitations
One of the strengths of this study is that we used the LAMP technique, which is the most powerful nucleic acid amplification method that could be used to detect pathogenic agents from drinking water with high sensitivity (Notomi et al. 2000). This is the first study to demonstrate an individual assay for detecting the four selected waterborne pathogens from an infant's drinking water sample. Our study result could provide new insights for building evidence based health policies and strategies by defining the magnitude of selected pathogens that cause severe and fatal diarrhea in infants' drinking water and strong contributing factors such as an unimproved water source. In addition, it helps with a better outlook in planning for the best solutions and generates ideas for further research.
This study has certain limitations that include the study design that we used, which demonstrated a snapshot of the water sample test result. Although one-time sampling information is very useful, it does not allow us to capture the burden of their presence in drinking water at multiple points in time. Our study is based only on the presence-absence test of organisms in drinking water, and there is no indication of the quantitative number of organisms. Our study was limited to detecting only four pathogens, and no total coliforms and E. coli were measured to recognize the degree of drinking water pollution. In addition, since no studies have been conducted on the molecular detection of pathogenic agents from infants' drinking water samples, comparisons between studies were difficult.
CONCLUSIONS
The detection of pathogenic microbial genes in drinking water samples plays a significant role in infant health. In our study, the results have demonstrated a high level of exposure of infants to contaminated drinking water by those recognized pathogens that cause the most severe and fatal diarrhea. The high prevalence of these pathogens in the water samples in our study seems to be robustly influenced by the technique used to detect the pathogens and various contributing factors, such as unimproved water sources, poor water source protection, and a lack of point-of-use water treatment practices. With the exception of rotavirus, unimproved water sources remained the only strong significant determinants for contamination of drinking water by these pathogens at the point-of-consumption. The presence of pathogens in drinking water used for infants at the household level is positively correlated with the water sources, which implies that the presence of pathogens at the point-of-use mainly depends on the water source used for collection. Therefore, efforts should be made towards the development and rehabilitation of improved water sources and water treatment at the household level. Water safety protection and other sanitary measures should be implemented to mitigate contamination from human and animal wastes. Health education must be undertaken to increase awareness among mothers or caretakers of infants on the prevention of waterborne pathogens in drinking water aimed at keeping the water safe for infants.
ACKNOWLEDGEMENTS
The authors would like to thank Addis Ababa University, EIWR, Ethiopian Public Health Institution (EPHI), and Ethiopian Biodiversity Institution (EBI) for their technical and in-kind resources assistance. Our deepest acknowledgement goes to the overall district's health offices (Chiro, Mieso, Amibara, and Awash Fentale) for their dedication and commitment in providing information and facilitating the survey. We also extend our sincere thanks to the study participants, such as the data collectors, supervisors, and respondents.
AUTHOR CONTRIBUTIONS
S.M. and A.W. were involved in conceptualization, data curation, format analysis, investigation, methodology, supervision, validation, visualization, writing-review, and editing. S.M. contributed in writing-original draft. S.M., H.S., A.A. contributed in project administration, resource, molecular laboratory (LAMP) technical methodology, software, and writing-review and editing.
DATA AVAILABILITY STATEMENT
Data cannot be made publicly available; readers should contact the corresponding author for details.
CONFLICTS OF INTEREST
The authors declare there is not conflict.