Research paperHow to improve the standardization and the diagnostic performance of the fecal egg count reduction test?
Introduction
The fecal egg count reduction test (FECRT) remains the recommended assay to assess anthelmintic drug efficacy against gastrointestinal nematodes in animals, and hence anthelmintic resistance (Kaplan and Vidyashankar, 2012). Guidelines for performing a FECRT were provided in the 1992 World Association for the Advancement of Veterinary Parasitology (WAAVP) publication on how to detect anthelmintic resistance in nematodes of veterinary importance (Coles et al., 1992). Since the publication of these guidelines, a variety of studies have provided novel insights into how to best design (e.g., Torgerson et al., 2005, Torgerson et al., 2012; McKenna, 2006; Dobson et al., 2012; Levecke et al., 2012; Calvete and Uriarte, 2012), analyze (e.g., Vercruysse et al., 2001; Cabaret and Berrag, 2004; Dobson et al., 2009, Levecke et al., 2011; Vercruysse et al., 2011; Vidyashankar et al., 2012; Calvete and Uriarte, 2012) and interpret a FECRT (e.g., Vidyashankar et al., 2007, Torgerson et al., 2005, Torgerson et al., 2014; Denwood et al., 2010; Dobson et al., 2012; Lyndal-Murphy et al., 2014; Geurden et al., 2015). These new insights point to the need for an update of the recommendations for the FECRT.
However, it is not straightforward to provide guidance that allows improving both the standardization and the performance of the FECRT across a variety of both animal and parasite species. For example, it is well known that the required sample size and the lower analytic sensitivity of the fecal egg count (FEC) method (≈1 / mass of feces in gram examined under the microscope) will depend on the underlying level (or intensity) and aggregation of egg excretion (Levecke et al., 2012). However, these two egg excretion parameters vary considerable between and within both parasite and animal species, and are often unknown prior the FECRT, making it difficult to recommend one study design that applies to all possible scenarios of egg excretion while assuring the reliable performance of a FECRT. An approach that may untangle this complex issue is to make recommendations on the total number of eggs that need to be counted under the microscope at baseline, a strategy that has been previously applied by Dobson et al., 2012. Traditionally, the eggs counted are converted into eggs per gram (EPG), and those data are then used for calculating percent reduction in FECs and the corresponding uncertainty intervals (UI, confidence intervals in case of frequentist methods and credible intervals in case of Bayesian methods). However, the eggs counted under the microscope are the actual data recorded in a FECRT; hence it is an important parameter with regard to both analytical issues and study design. For example, when FECR is based on the pre- and post-treatment FECs of the same animals and when all pre- and post-treatment samples are examined applying a FEC the same analytic sensitivity, the formula can be reduced to the ratio of the total number of eggs counted under the microscope at pre- and post-treatment (see Equations 1 and 2; Supplementary Material 1). Additionally, the total number of eggs counted under the microscope nicely grasps the variations in both study design (analytic sensitivity and sample size) and host-parasite interactions (level and aggregation of egg excretion), and hence recommending a minimum number of eggs to be counted under the microscope (also dependent on the examined sample size) allows avoiding stringent recommendations on the other parameters. As an example, if it were recommended to count at least 200 eggs under the microscope across at least 10 animals prior treatment and if animals were excreting on average 500 EPG, one could either screen 10 animals using a diagnostic technique with a analytic sensitivity of 25 EPG (10 animals x 500 EPG / 25 EPG = 200 eggs) or screen 20 animals with a analytic sensitivity of 50 EPG (20 animals x 500 EPG / 50 EPG = 200 eggs). In addition, it would also allow avoiding underpowered trials when egg excretion at baseline revealed to be lower as anticipated or when the level of egg excretion is even unknown. When we now assume that the animals excrete 250 EPG instead of 500 EPG, the total number of eggs at baseline would be 100 for both study designs (=10 animals x 250 EPG / 25 EPG = 20 animals x 250 EPG / 50 EPG), and hence the trials may not allow to readily draw conclusions on the drug efficacy. At this point, one could easily adapt the diagnostic strategy until a sufficient number of eggs are counted at baseline (e.g. reexamination of the same samples with the same diagnostic technique would in principle double the eggs counted). Although these examples illustrate the elegance of recommending a minimum number of eggs to be counted under the microscope we currently lack the evidence to recommend the number of eggs required to yield a reliable assessment of drug efficacy by means of the FECRT. Lastly, it is also important to avoid having most eggs counted come from only few animals, and hence guidance on the allowed distribution of egg counts across animals is also needed.
Other aspects of analysis and interpretation of the FECRT that need further research are the methodology of calculating the corresponding UIs and the criteria for classifying drug efficacy into ‘normal’, ‘suspected’ and ‘reduced’. The UI methodology in the current guidelines has two important limitations. First, it can only be applied on a randomized controlled study design using post-treatment counts of both treatment and control groups (Coles et al., 1992). This experimental design, however, has proven to be less sensitive at detecting reduced efficacy compared to FECRT based on pre- and post-treatment counts from the same animals (McKenna, 2006; Dobson et al. 2012; Calvete and Uriarte, 2012). Second, the uncertainty of the estimates (the UI) cannot be calculated when the observed FECR is 100%, and hence it is impossible to draw conclusions on the reliability of the estimate (Denwood et al., 2010; Dobson et al., 2012; Torgerson et al., 2014). This is relevant because seeing no eggs following treatment does not mean the efficacy was 100%; depending on how many eggs were counted in the pre-treatment FECs the true efficacy may be approaching 100%, but may never reach 100% (or be considerably less than 100%). Alternative UI methodologies can be applied to FECRT based on counts of the same animals (Denwood et al., 2010; Lyndal-Murphy et al., 2014; Torgerson et al., 2014; Levecke et al., 2015; Geurden et al., 2015; Peña-Espinoza et al., 2016). Other methods allow assessment of the uncertainty of estimates when FECR is 100% (Denwood et al., 2010; Dobson et al., 2012; Torgerson et al., 2014; Geurden et al., 2015; Peña-Espinoza et al., 2016). However, at present these methodologies need further research, thus the accuracy and precision of these methods requires further clarification. Currently, the efficacy of an anthelmintic is classified as ‘reduced’, ‘suspected’ and ‘normal’ based on how the obtained FECR and the lower limit (LL) of the 95% UI relates to established thresholds, a drug efficacy for sheep and goats being ‘reduced’ when FECR <95% and LL of UI <90%; as ‘suspected’ when either FECR <95% or LL <90%, and as ‘normal’ when FECR ≥ 95% and LL ≥ 90% (Coles et al., 1992). Whether alternative classification criteria that are solely based on the UI (e.g. drug efficacy being ‘reduced’ when upper limit (UL) <95%, ‘normal’ when LL ≥ 95%, and ‘suspected’ in any other cases; El-Abdellati et al., 2010) or combining both the FECR-estimate and the UI (Lyndal-Murphy et al., 2014; Geurden et al., 2015; ‘normal’: FECR ≥ 95%, UL of UI ≥ 95% and LL of UI ≥ 90%; ‘reduced’: FECR <95%, UL of UI <95% and LL of UI <90%; ‘suspected’ in all other cases), or only the FECR estimate (e.g., drug efficacy being ‘reduced’ when FECR <90%, ‘normal’ when FECR ≥ 95%, and ‘suspected’ in any other cases; in analogy with World Health Organization, 2013) would allow for reliable detection of reduced drug efficacy is unclear.
The aim of this study is to provide insights complementary to the current knowledge on how to design, analyze and interpret FECRTs, and to ultimately provide guidance in the development of new standardized guidelines for the FECRT that lead to improving both the standardization and the analytical performance of the FECRT across a variety of both animal and parasite species.
Section snippets
Methods
The study consisted of two consecutive procedures. First, data were generated using a simulation in which the ‘true’ drug efficacy (TDE) was evaluated by the FECRT under varying scenarios of sample size, analytic sensitivity of the diagnostic technique and level of both intensity and aggregation of egg excretion. Second, the obtained data were analyzed with the aim (i) to verify which classification criteria allows for reliable detection of reduced drug efficacy, (ii) to identify the UI
Comparison of the diagnostic performance of four classification criteria for drug efficacy
The sensitivity and the specificity of detecting a truly reduced efficacy for each of the four classification criteria are summarized in Table 1. As illustrated in Fig. 1 the sensitivity and specificity varied across different values of TDE, with the probability of correctly classifying drug efficacy decreasing as the TDE approached the threshold of 95% (surface of grey zone increases). For example, when applying criteria 1, an efficacy of 70% and 99% are correctly classified, with high
Discussion
Despite increasing criticism of the existing WAAVP recommended guidelines for FECRT published almost 25 years ago in 1992, and the recent progress made on how to best design, analyze and interpret a FECRT, there remains a lack of important evidence to support the revision of the current guidelines. In this study we aimed to address this lack of evidence, with the goal of providing insights complementary to the current knowledge, and to ultimately provide guidance that allows improving both the
Acknowledgements
BL is funded by the Fund for Scientific Research-Flanders (Belgium) (F.W.O.-Vlaanderen; www.fwo.be, grant no 1285316N). SMT was partly funded by the EMIDA-project CARES. We also want to thank Matthew J. Denwood (Department of Veterinary and Animal Sciences, Faculty of Health and Medical Sciences, University of Copenhagen, Denmark) for applying the Bayescount methodology on the toy example described in Supplementary Material 2.
References (44)
- et al.
Strategic control of gastrointestinal nematodes in grazing sheep with a long-acting moxidectin formulation
Small Ruminant Res.
(2015) - et al.
Multispecies resistance of cattle gastrointestinal nematodes to long-acting avermectin formulations in Mato Grosso do Sul
Vet. Parasitol.
(2015) - et al.
Faecal egg count reduction test for assessing anthelmintic efficacy: average versus individually based estimations
Vet. Parasitol.
(2004) - et al.
World Association for the Advancement of Veterinary Parasitology (W.A.A.V.P.) methods for the detection of anthelmintic resistance in nematodes of veterinary importance
Vet. Parasitol.
(1992) - et al.
Comparison of three alternative methods for analysis of equine faecal egg count reduction test data
Prev. Vet. Med.
(2010) - et al.
Geometric means provide a biased efficacy result when conducting a faecal egg count reduction test (FECRT)
Vet Parasitol.
(2009) - et al.
Monitoring macrocyclic lactone resistance in Cooperia oncophora on a Belgian cattle farm during four consecutive years
Vet Parasitol.
(2010) - et al.
Anthelmintic resistance to ivermectin and moxidectin in gastrointestinal nematodes of cattle in Europe
Int. J. Parasitol. Drugs Drug Resist.
(2015) - et al.
Control of Endoparasitic Nematode infections in goats
Vet. Clinics North. Am. Food Animal Practice
(2011) - et al.
An inconvenient truth: global worming and anthelmintic resistance
Vet. Parasitol.
(2012)
Application of a poisson distribution quality control measure to the analysis of two human hookworm drug treatment studies in Ghana
Int. J. Parasitol. Drugs Drug Resist.
Mathematical inference on helminth egg counts in stool and its applications in mass drug administration programmes to control soil-transmitted helminthiasis in public health
Adv. Parasitol.
Novel insights in the faecal egg count reduction test for monitoring drug efficacy against gastro-intestinal nematodes of veterinary importance
Vet. Parasitol.
Methods to determine resistance to anthelmintics when continuing larval development occurs
Vet. Parasitol.
Novel insights into the pathogenic importance, diagnosis and treatment of the rumen fluke (Calicophoron daubneyi) in cattle
Vet. Parasitol.
Effects of aggregation and sample size on composite faecal egg counts in sheep
Vet. Parasitol.
First report of closantel treatment failure against Fasciola hepatica in cattle
Int. J. Parasitol. Drugs Drug Resist.
Efficacy of ivermectin against gastrointestinal nematodes of cattle in Denmark evaluated by different methods for analysis of faecal egg count reduction
Int. J. Parasitol. Drugs Drug Resist.
Evaluating faecal egg count reduction using a specifically designed package “eggCounts” in R and a user friendly web interface
Int. J. Parasitol.
The contribution of simple random sampling to observed variations in faecal egg counts
Vet. Parasitol.
Detection of anthelmintic resistance: a comparison of mathematical techniques
Vet. Parasitol.
International harmonisation of anthelmintic efficacy guidelines
Vet. Parasitol.
Cited by (44)
Integration of ITS-2 rDNA nemabiome metabarcoding with Fecal Egg Count Reduction Testing (FECRT) reveals ivermectin resistance in multiple gastrointestinal nematode species, including hypobiotic Ostertagia ostertagi, in western Canadian beef cattle
2023, International Journal for Parasitology: Drugs and Drug Resistance