Introduction
Soil-transmitted helminths (STH) claim a high disease burden, especially in the economically-deprived parts of the world (Jourdan et al., Reference Jourdan, Lamberton, Fenwick and Addiss2018). The diseases, while being treatable by common deworming medications and preventable with common hygienic standards (Campbell et al., Reference Campbell, Nery, McCarthy, Gray, Soares Magalhães and Clements2016), are major contributors to worldwide morbidity. Roundworm (Ascaris lumbricoides, ascariasis), whipworm (Trichuris trichiura, trichuriasis) and hookworm (Ancylostoma duodenale and Necator americanus, ancylostomiasis and necatoriasis) are the three major parasitic worms among others that claim the highest disease morbidities. Roundworms and whipworms are transmitted by the faecal–oral route (Centers for Disease Control, 2020a, 2020d) while the hookworm and threadworm (Strongyloides stercoralis, strongyloidiasis) are transmitted when their larval stage penetrates open skin (Centers for Disease Control, 2020b, 2020c). Ancylostoma duodenale can also be transmitted by ingestion of the larvae stage and hookworms may even have a transmammary/transplacental route of transmission in neonates (Sen-Hai et al., Reference Sen-Hai, Ze-Xiao and Long-Qi1995; Papaiakovou et al., Reference Papaiakovou, Gasser and Littlewood2019). Roundworms, hookworms and threadworms have tissue-dwelling stages where their specific larval stages leave the gut and penetrate the vascular tissue and solid tissue. Whipworms complete their life cycles in the intestine and during their adult form, their heads remain embedded in the linings of the gut.
The DNA-based detection of STH in stool samples has proved to be the most sensitive method (Verweij & Rune Stensvold, Reference Verweij and Rune Stensvold2014; O'Connell & Nutman, Reference O'Connell and Nutman2016). Quantitative polymerase chain reaction (qPCR) is a more consistent method compared to other copro-microscopic techniques and is less affected by the experience of the technologists as personnel with basic knowledge about PCR can perform the experiments with similar results. qPCR has various advantages from the detection of the presence of worms in stools to follow-up evaluation after various therapeutic interventions to decrease the prevalence (Clarke et al., Reference Clarke, Llewellyn, Traub, McCarthy, Richardson and Nery2018). But the sensitivity and specificity may vary due to the biology of the helminth eggs (ploidy in roundworms; germination stages in hookworms) (Papaiakovou et al., Reference Papaiakovou, Gasser and Littlewood2019), non-egg DNA (Andersen et al., Reference Andersen, Röser, Nejsum, Nielsen and Stensvold2013) and the PCR technology used (Verweij & Rune Stensvold, Reference Verweij and Rune Stensvold2014; O'Connell & Nutman, Reference O'Connell and Nutman2016). Various initiatives are being made to replace old but reliable microscopic techniques such as Kato–Katz with more sensitive state-of-the-art PCR technology (Cools et al., Reference Cools, Vlaminck and Albonico2019). However, the high cost of equipment and maintenance of contamination-free conditions may prevent the use of this technique in resource-constraint settings. Despite this, in terms of consumables and reagents used (not including the cost and maintenance of qPCR machines), studies have claimed that the cost of multiplex qPCR tests may be equivalent to conventional microscopy techniques (O'Connell & Nutman, Reference O'Connell and Nutman2016). DNA-based techniques have the additional advantage of the feasibility of carrying tests at later time points if the stool samples are preserved in a proper manner (Ayana et al., Reference Ayana, Cools and Mekonnen2019).
While difficulties have not been reported in the extraction of DNA from stool samples containing roundworms and hookworms, the extraction of DNA from Trichuris eggs has been reported to require additional handling (Demeler et al., Reference Demeler, Ramünke, Wolken, Ianiello, Rinaldi, Gahutu, Cringoli, von Samson-Himmelstjerna and Krücken2013). Inefficient DNA extraction may lead to decreased Trichuris detection sensitivity by qPCR compared to detection by microscopy (Clarke et al., Reference Clarke, Llewellyn, Traub, McCarthy, Richardson and Nery2018). Studies have repeatedly pointed out the requirement of a bead-beating (BB) step during DNA extraction for proper extraction of the Trichuris DNA (Verweij & Rune Stensvold, Reference Verweij and Rune Stensvold2014; O'Connell & Nutman, Reference O'Connell and Nutman2016). The BB step is expected to help break the Trichuris eggs during the enzymatic lysis step of DNA extraction, and thus make Trichuris DNA detectable by PCR. This is carried out by mixing stool samples with lysis buffer and beads of a specific size (less than 1 mm in diameter) and type (glass, zirconium, ceramic, garnet, etc.) and oscillating them at high speed for given durations. A search for articles which assessed the relevance of BB during DNA extraction was carried and a meta-analysis was done. While many studies incorporated BB into their DNA extraction methods for the Trichuris PCR test in the early 2010s (Liu et al., Reference Liu, Gratz and Amour2013; Mejia et al., Reference Mejia, Vicuña, Broncano, Sandoval, Vaca, Chico, Cooper and Nutman2013), recently few have done direct comparative studies to show the efficacy of the inclusion of this step. This systematic review was designed to assess the benefit of including the step for higher sensitivity in detecting Trichuris DNA during qPCR. Singleplex or multiplex qPCR for the detection of STH is usually carried out in community settings to determine the prevalence of the helminth and to determine the efficacy of deworming interventions (Farrell et al., Reference Farrell, Coffeng, Truscott, Werkman, Toor, de Vlas and Anderson2018). The research question for this systematic review also envisages the index and reference tests to have high sensitivity to determine the helminth presence in the community setting rather than in the hospital setting.
Material and methods
This study was carried out and reported according to the Preferred Reporting Items for a Systematic Review and Meta-analysis for Diagnostic Test Accuracy Studies (McInnes et al., Reference McInnes, Moher and Thombs2018). Articles were searched in PubMed by using the ‘AND’ Boolean operator to combine two search fields, ‘qPCR OR quantitative polymerase chain reaction OR quantitative PCR OR quantitative real-time PCR OR polymerase chain reaction OR PCR’ and ‘Trichuris OR Trichuriasis OR Whip worm OR Whipworm OR Trichuris trichiura’. An additional search was carried out in the Google search engine with the search term ‘Trichuris bead beating DNA extraction’. Abstracts and/or bodies of the manuscripts were read to identify the comparative studies according to inclusion criteria. Studies using the index and reference tests in community settings were included. The studies should have carried out DNA extraction by two methods: one including the BB step and, two, without including the BB step. Articles referenced in the selected studies were scanned for more relevant articles but no authors of selected articles were contacted for further details. Studies included field-based projects where samples were collected in a blanket fashion in suspected populations irrespective of the presence of any helminth-related symptoms in target individuals. Tests should not be carried out in individual suspect patients either in the hospital or community settings. Articles using qPCR both as an index and reference tests qualified for the review. The index qPCR test was preceded by DNA extraction using an additional BB step to break the T. trichiura egg. The reference test did not have any BB step. Data were directly extracted from the selected studies and entered in Excel for statistical analyses. Risk of bias and applicability was carried out according to the Quality Assessment of Diagnostic Accuracy Studies-2 (QUADAS-2) checklist (Whiting et al., Reference Whiting, Rutjes, Westwood, Mallett, Deeks, Reitsma, Leeflang, Sterne and Bossuyt2011) developed by the QUADAS-2 study group (Whiting, Reference Whiting2011). An additional literature search was carried out to study the biochemical structure of helminth eggs.
Statistical analyses
The odds ratio (OR) was calculated for individual studies/sub-studies. Each of the collected stool samples in individual studies was examined by index and reference tests. Samples tested by index test were regarded as exposed and those tested with reference were regarded as an unexposed group to calculate the OR. The random effects model (REM) OR was estimated for the meta-analysis. Individual ORs of the sub-studies were combined for the meta-analysis. The analysis was carried out in MedCalc software which uses Mantel and Haenszel's (1959) method for the fixed effect model and DerSimonian and Laird's (1986) method for random effects modelling (MedCalc, 2020). The REM was considered more applicable due to heterogeneity in the study as explained in the Results section.
Data sharing statement
No additional data are available.
Results
Search through the PubMed database resulted in 165 results from inception to December 2019 with no barrier to languages. A search through Google resulted in a total of 15 relevant search results. After step-wise sorting (fig. 1), two articles from PubMed and one article from the Google search engine were selected for further analysis. Studies that carried out experiments to compare the effect of the inclusion or exclusion of the BB step during DNA extraction from the stool on the prevalence of T. trichiura were selected. Further reading of the articles led to exclusion of one article for the meta-analysis as it carried DNA extraction in artificially spiked stool samples and did not specifically study the effects of BB in actual field samples. Both of the included studies were carried out in community settings which matched the research question. Study populations were chosen based on prior information that there was some degree of infection prevalence that could be detected by index and reference tests. All of the study participants were healthy and were not included or excluded based on any specific symptoms. While Ayana et al. (Reference Ayana, Cools and Mekonnen2019) enrolled school children (5–14 years) in Ethiopia, Kaisar et al. (Reference Kaisar, Brienen, Djuardi, Sartono, Yazdanbakhsh, Verweij, Supali and Van Lieshout2017) enrolled mothers (20–37 years) and their babies (1–5 years) in Indonesia. Both of the studies disclosed their funding sources but only Ayana et al. (Reference Ayana, Cools and Mekonnen2019) stated in the publication that the funders had no role in study design, data collection and analysis, decision to publish and preparation of the manuscript. But there was no proof that the funders of Kaisar et al. (Reference Kaisar, Brienen, Djuardi, Sartono, Yazdanbakhsh, Verweij, Supali and Van Lieshout2017) had any input in any of the above activities.
Preferred reporting items for a systematic review and meta-analysis for diagnostic test accuracy studies work-flow for selection of studies.
Risk of bias and applicability
There was a low risk of bias and applicability for each of the studies/sub-studies (fig. 2). Perhaps due to the nature of the diagnostic test, many of the risks could be reduced due to aliquotations. As the patient stool samples could be collected in sufficient amounts and aliquoted, both the index and reference tests could be carried out on identical samples and, theoretically, at the same time. Another factor that reduces the risk of bias was the fact that the index and reference tests consisted of two steps. The second step, qPCR, was identical within each of the studies, and thus the threshold of detection (or analytical sensitivity) was the same for the index and reference tests in each of the studies. In all of the studies/sub-studies, the patient samples were taken from specified populations without categorizing them according to any symptoms or physical conditions. All of the healthy-looking participants were enrolled consecutively. This decreased the overall applicability issue of the review question as the qPCR is intended to be used in community settings.
Risk of bias and applicability of included studies.
Qualitative analysis
Three articles assessed the effect of BB during DNA extraction. Andersen et al. (Reference Andersen, Röser, Nejsum, Nielsen and Stensvold2013) artificially spiked healthy stool samples with T. trichiura or Ascaris suum eggs and carried out qPCR to assess the effect of BB. They compared the effect of different types of beads – 0.5 mm glass beads, 0.15 mm Garnet beads and 0.1 mm zirconium beads – using the NucliSENS easyMag DNA extraction system (Biomerieux, USA). They concluded that zirconium beads gave better sensitivity for DNA isolation of A. suum. In the case of T. trichiura, the study concluded similar analytical sensitivity between vortexing without any beads and using zirconium BB for 30 s at 7000 oscillations. They reported that the clinical sample showed a lower (better) limit of detection compared to the artificially spiked sample, hinting at the presence of microscopically invisible extra-cellular DNA in clinical stool samples. The study by Kaisar et al. (Reference Kaisar, Brienen, Djuardi, Sartono, Yazdanbakhsh, Verweij, Supali and Van Lieshout2017) also assessed the effects of five different kinds of beads: 0.5 mm stainless steel; 0.5 mm zirconium oxide; 0.7 mm Garnet; 0.8 mm Garnet; and 0.5 mm yttria-stabilized zirconium oxide. It was found that 0.8 mm Garnet gave the best analytical sensitivity and further experiments were carried out with this bead by vortexing for 3 min at 1800 rotations per minute. Ayana et al. (Reference Ayana, Cools and Mekonnen2019) used 1.4 mm ceramic beads to beat the stool samples for 1 min at 3000 rotations per minute. Just before the BB, the stool samples were freeze-thawed to increase the efficiency of cell disruption during BB.
Quantitative analysis
The study by Kaisar et al. (Reference Kaisar, Brienen, Djuardi, Sartono, Yazdanbakhsh, Verweij, Supali and Van Lieshout2017) divided the 60 collected stool samples into two aliquots and preserved them in a refrigerator or mixed them with ethanol and stored them at room temperature (Kaisar et al., Reference Kaisar, Brienen, Djuardi, Sartono, Yazdanbakhsh, Verweij, Supali and Van Lieshout2017). Both sets of preserved samples underwent two types of DNA extraction protocols, one including the BB step and the other without. As the sample materials had been altered chemically to some degree due to the preservation methods, we considered the two sets of experiments (frozen and ethanol preserved) as two different sub-studies (Frozen_samp and Ethanol_samp) in this analysis. Both sets of experiments used a QIAamp DNA-easy kit from Qiagen, Germany. Overall, the inclusion of BB gave a higher positivity rate. 51.7% (31/60) of frozen stool samples were positive when BB was used during DNA extraction compared to only 40% (24/60) when BB was not used (P > 0.05). Similarly, the percentage of positive samples was 55% (33/60) and 45% (27/60), respectively, for ethanol-preserved samples (P > 0.05). Overall, ethanol-preserved samples performed better. The study by Ayana et al. (Reference Ayana, Cools and Mekonnen2019) used two different kits to extract DNA. QIAamp DNA Stool Mini kit (S_K) and DNeasy Blood & Tissue kit (BaT_K), both from Qiagen, Germany, were used and the Blood & Tissue kit was reported to perform better. Each method was further divided into one using the BB and the other not. To examine the effect of BB, the stool samples, as assessed microscopically, were divided into 15 faecal egg count (FEC) negative samples, 11 low-intensity FEC positive samples and ten moderate to heavy intensity FEC positive samples. For the sake of this meta-analysis, the 15 FEC negative (zero value for egg per gram) samples were grouped as one (Zero_FEC_BaT_K or Zero_FEC_S_K for FEC negative/zero samples analysed by Blood & Tissue kit and Stool Mini kit, respectively) and 21 FEC positive samples (11 low and ten heavy intensity) as other group (Pos_FEC_BaT_K or Pos_FEC_S_K for FEC positive samples analysed by Blood & Tissue kit and Stool Mini kit, respectively). This grouping is logical because only a small number (n = 15) of the total 159/195 (81.5%) microscopically/DNA Trichuris negative samples were used in the analysis and a combination of all three groups of samples (negative, low intensity and moderate to heavy intensity) could have hampered actual sensitivity estimation. When the Blood & Tissue kit was used 87% (13/15) of the microscopy-negative samples were positive by DNA extraction method incorporating BB compared to only 73% (11/15) for method not including BB (P > 0.05). The positivity percentages for microscopy-positive samples were 100% (21/21) and 95% confidence interval (CI) (20/21), respectively. When the Stool Mini kit was used, the positivity percentages were 73% (11/15) and 67% (10/15), respectively, for microscopy-negative samples (P > 0.05). For microscopy-positive samples, the positivity percentages were 90% (19/21) and 76% (16/21), respectively (P > 0.05).
The REM was chosen compared to the fixed effect model even when I2 (inconsistency) was 0.00% (95% CI: 0.00–0.00). As the methods of DNA extraction, types of beads, stool preservation methods and level of background prevalence of STH were different in the studies or within the studies, the REM was considered more appropriate. The REM showed a summary OR of 1.66 (95% CI: 1.059 to 2.209) (table 1, fig. 3). None of the individual sub-studies showed a significant effect of BB while the REM showed a significant effect. This could be due to the small sample sizes in the individual studies. None of the small sub-studies showed a negative effect of BB. Thus, it can be safely concluded that the significant effect seen in the meta-analysis is due to the synergistic effect of all the sub-studies whose results are directed in one direction. When all the FEC zero and positive samples in Ayana et al. (Reference Ayana, Cools and Mekonnen2019) were combined as one and re-analysed (sample sizes of 36 in BaT_K and S_K each) in the meta-analysis (data not shown), comparable significant REM results (summary OR of 1.665) were obtained.
Random effects model summary odds ratio.
Forest plot for summary odds ratio.
Biochemical studies of helminth eggs
The eggs of nematodes have been known to be resistant to various environmental stresses and chemicals. The egg of Ascaris alone has been shown to remain viable against various acids, alkalis and digestive enzymes (Yoshida, Reference Yoshida1920; Martin, Reference Martin1926; Beyhan et al., Reference Beyhan, Yilmaz and Hokelek2016). This extreme resilience is hypothesized to be due to the eggshell, which has to be strong enough to protect the embryo inside (Wharton, Reference Wharton1986). Nematodes’ eggshells, including those of hookworm, Trichuris and Ascaris, in general, are made up of three layers: outer vitelline layer; middle chitinous layer; and inner lipid layer. The middle chitinous layer is a composite layer composed of chitin microfibrils surrounded by a protein matrix, which is said to make the egg resistant to mechanical damage (Slayter, Reference Slayter1962; Neville, Reference Neville1975; Wharton, Reference Wharton1986). The inner lipid layer is impermeable to most chemicals (Barrett, Reference Barrett1981; Wharton, Reference Wharton1986) and may be a reason for the difficulty in DNA extraction using a chemical process. In many nematodes, including Ascaris, the presence of ascaroside esters could be a reason for the impermeability (Nissen et al., Reference Nissen, Al-Jubury, Hansen, Olsen, Christensen, Thamsborg and Nejsum2012; Ludewig & Schroeder, Reference Ludewig and Schroeder2018) and be also responsible for the resistance against chemical actions (Bartley et al., Reference Bartley, Bennett and Darben1996).
Discussion
An increasing number of STH studies are using BB steps in the DNA extraction method. While the possible need for this step has been raised earlier (Verweij & Rune Stensvold, Reference Verweij and Rune Stensvold2014; O'Connell & Nutman, Reference O'Connell and Nutman2016), very few studies have reported the significance of this addition. Though Trichuris is considered to be the only STH that requires the BB, both Ayana et al. (Reference Ayana, Cools and Mekonnen2019) and Kaisar et al. (Reference Kaisar, Brienen, Djuardi, Sartono, Yazdanbakhsh, Verweij, Supali and Van Lieshout2017) have found that this could also be required for hookworms and Ascaris, respectively. They also reported that the blood and tissue kit gave a better result than the stool kit. It may be because stool kits are usually designed for the isolation of bacterial DNA in the stools (Fiedorová et al., Reference Fiedorová, Radvanský, Němcová, Grombiříková, Bosák, Černochová, Lexa, Šmajs and Freiberger2019) and may not be appropriate for helminth eggs. The sizes of helminth eggs for hookworm, Trichuris and Ascaris are comparable and the biochemical structures of the egg shells are similar too. While the hookworm eggs are known to be fragile and can rupture within hours of stool collection at room temperature, the Trichuris and Ascaris eggs remain in the soil for longer durations. The present review which includes a meta-analysis shows that the inclusion of BB makes a significant impact on the DNA extraction of Trichuris DNA and recommends the use of this technique.
Limitations
Only two published studies qualified for analysis due to the paucity of systematic comparative studies. We divided the two studies into a total of six sub-studies. It was not possible to keep the sub-studies combined as the methods of stool preservation, kits used and the cohort of patients studied, if combined, could affect the outcome concluded. Another limitation could be the non-inclusion of the cycle threshold data for the qPCR which could differentiate the slight changes in the yield of Trichuris DNA extracted using various interventions during DNA extraction. Both of the studies did not report the limit of detection of the qPCR. Another limitation of the study was the inherent heterogeneity in study populations. It has been reported that the qPCR positivity or sensitivity may vary according to the intensity of infection in the study population (Cools et al., Reference Cools, Vlaminck and Albonico2019). Thus different study populations differing by background helminth prevalence may give different sensitivities to the same qPCR technique and thus bias the analysis.
Conclusions
More systematic studies assessing the effect of BB on DNA extraction of Trichuris are recommended. As reported by the two studies above, preliminary optimization of the best bead types and/or sizes should also be carried out so the impact is concluded. Uniformity in the type of beads, beating conditions, DNA extraction method, primers, kits, PCR conditions, etc. can help in comparative studies from different parts of the world.
Strengths and Limitations of this study
• While only two studies are included in the analysis, these contain a total of six independent sub-studies.
• The sub-studies are different in terms of the patient population, sample preservation and processing techniques, and thus can't be combined
• The direction of the odds ratio (OR) in all of the sub-studies is in one direction, thus the significance of the combined OR is strengthened.
• While sample sizes in individual studies are small to give significant statistical significance, the combined OR revealed a statistically significant result.
Financial support
This research received no specific grant from any funding agency in the public, commercial or not-for-profit sectors.
Supplementary material
Risk of Bias assessment using QUADAS2 tool kit.
Conflicts of interest
None
Ethics approval
Not applicable.
Consent to participate
Not applicable.
Consent for publication
Both authors agree to the publication.
Authors’ contributions
DRSJBR was involved in the conceptualization of the study. DRSJBR and NP carried out data curation. DRSJBR did formal data analysis. Both authors were responsible for original draft preparation, reviewing and editing.
