Elsevier

Veterinary Parasitology

Volume 289, January 2021, 109324
Veterinary Parasitology

Revisiting the adequacy of infection criteria recommended in VICH GL 7 for anthelmintic effectiveness studies: Retrospective simulations

https://doi.org/10.1016/j.vetpar.2020.109324Get rights and content

Highlights

  • Conventional criterion for adequacy of infection in current VICH GL 7 could be over- interpreted.

  • Combination of statistical criterion with the conventional criterion can guard against over-interpretation.

  • Statistical criterion may be overly strict under certain situations.

Abstract

Studies conducted to support registration or approval of veterinary anthelmintics generally follow study design recommendations provided by the VICH (International Cooperation on Harmonisation of Technical Requirements for Registration of Veterinary Medicinal Products), “Efficacy of Anthelmintics: General Requirements” (VICH GL7). For dose confirmation studies, VICH GL7 provides recommendations for determining that the control animals had an adequate infection “to permit the appropriate standards of efficacy to be met with acceptable statistical and biological certitude/confidence.” In the simulation studies described in this report, we investigated the performance of one method, the statistical criterion given in Section 4.5 of VICH GL7, for evaluating the adequacy of infection in anthelmintic studies, in combination with the conventional criterion of a minimum of six adequately infected animals. We conducted numerical simulations, based on parasite data from previously conducted dose confirmation studies in dogs and cattle, to investigate how the statistical criterion impacts adequacy of infection determinations when used with the conventional criterion at various sample sizes. Simulation studies in common nematode species in both dogs and cattle indicated that under certain circumstances the statistical criterion can guard against overinterpreting the evaluation of adequacy of infection as sample size is increased. However, the statistical criterion may be overly restrictive for samples with adequate infection but containing multiple zero parasite counts and adding it to the conventional criterion does not provide any additional benefit when the sample contains no zero counts. It is important for investigators designing efficacy studies to understand the potential impact this criterion may have when establishing adequacy of infection criteria in study protocols.

Introduction

Regulatory studies intended to test the effectiveness of anthelmintics are interpretable only if it can be verified that the subjects were sufficiently infected prior to anthelmintic or placebo administration. For placebo-controlled dose confirmation studies, VICH GL 7 “Efficacy of Anthelmintics: General Requirements” (VICH GL 7, 2000) recommends evaluating adequacy of infection in the control group based on the number of animals that are found to have at least the pre-specified minimum parasite load. Examination of adequate infection in the control group is the first step in the regulatory assessment of the effectiveness of the product being investigated. Only if the infection is considered adequate is the control group considered an appropriate comparator to the treated group to test the effectiveness of the investigational veterinary product.

VICH GL 7 stipulates that protocols should address adequacy of infection in terms of the infection level in individual control animals as well as the number of control animals in which infections are established. Adequacy of infection based on the conventional criterion described in VICH GL7 is defined as at least six individual animals in the control group with a prespecified minimum infection level for each parasite species tested. We note that this minimum infection criterion does not consider the dependence of the criterion’s success on the sample size. For typical dose confirmation studies, VICH GL7 recommends at least 6 animals per experimental group. It is common for studies to enroll 8–10 animals per group in order to ensure that they meet the minimum of 6 adequately infected animals in the control group. There is no provision in VICH GL7 for a maximum number of animals per experimental group. However, the probability of observing 6 adequately infected animals would increase as the sample size increases, even when infection rate in the study animals is low. Therefore, under certain circumstances such as natural low parasite burdens or low prevalence of parasites, one way to increase the chance of achieving adequacy of infection based on the conventional criterion is simply to use a larger sample size. Using only the conventional criterion in these situations could lead to an overinterpretation of adequacy of infection in these studies.

In Section 4.5 of VICH GL 7, an additional statistical criterion was established, as quoted below:

“In cases where all animals in the control group are infected, then one possible statistical method involves the use of calculating the lower 95 % confidence limit of the control group geometric mean burden. If this value is greater than 10 % of the control group geometric mean burden, then the infection can be said to be adequate. In the case where some of the animals in the control group are not infected (counts = zero) then the geometric means should be replaced by the median and the 95 % confidence limit should be based on the control group median burden.”

We note that this statistical criterion is considered an additional criterion because the conventional criterion is a prerequisite for establishing the adequacy of infection, i.e. the study will be voided if the control group does not show a minimum of six adequately infected animals regardless of whether the statistical criterion is met or not. However, VICH GL 7 does not specify when or whether the statistical criterion should be invoked when the sample size is increased beyond the commonly accepted ranges (e.g. 6–10 animals per study group). To our knowledge, this additional statistical criterion is rarely, if at all, used in dose confirmation studies for anthelmintic product development. The performance characteristics of the statistical criterion are not fully understood, and it is unclear whether it can protect against overinterpreting adequacy of infection when the sample size is intentionally increased to achieve a minimum of six infected animals.

The objective of this study is to explore the properties of these two adequacy of infection criteria regarding their performance when the sample size is increased beyond the commonly accepted 6–10 animals per study group. Specifically, we investigated what effect, if any, the statistical criterion has on the overall assessment of adequacy of infection as a function of the sample size. We conducted statistical numerical simulations to evaluate representative scenarios encountered in dose confirmation studies of veterinary anthelmintic products. The studies were conducted using parasite data collected in dose confirmation studies for common nematodes in dogs and cattle. The results from this study will inform investigators of the impact the statistical criterion may have on the infection criteria for dose confirmation protocols.

Section snippets

Methods

In this paper, “control group” denotes the “placebo-treated group”. As mentioned earlier, the requirement for a minimum of six adequately infected control animals is referred to as the “conventional criterion.” When both the statistical and conventional criteria are applied, we call it “combined criteria”. Under the combined criteria, the parasite loads of the control group should pass both criteria to demonstrate adequacy of infection. In our simulations, the parasite counts in the control

Initial findings – general observations

One typical set of simulation results is summarized in Table 3, Table 4. We simulated samples of size 8–28 from the example data sets. The first set of results (columns 2–6) provides pass rates, i.e. the proportion of samples over the 10,000 runs that were identified as having adequate infections, using the conventional criterion alone, and the second set of pass rates (columns 7–11) are calculated based on the combined criteria (conventional and statistical).

Increasing the sample size

Discussion

In standard dose confirmation studies with parasites counted at necropsy post-treatment, the effectiveness of the investigational veterinary product is usually evaluated based on three criteria: 1) a minimum of six adequately infected animals in the non-medicated (control) group; 2) statistically significant differences (p < 0.05) in parasite counts between treated and control animals; and 3) ≥90 % efficacy calculated using transformed (geometric means) data. Passing the first criterion is a

Conclusion

In this study, we investigated the performance of the statistical criterion given in Section 4.5 of VICH GL7 for evaluating the adequacy of infection in anthelmintic studies when it is used in addition to the conventional criterion of six adequately infected animals. Numerical simulation studies indicated that under certain circumstances, the statistical criterion can guard against overinterpreting the evaluation of adequacy of infection as sample size is increased. However, the mechanism of

CRediT authorship contribution statement

Xiongce Zhao: Conceptualization, Formal analysis, Writing - review & editing. Emily Smith: Conceptualization, Writing - review & editing. Aimee Phillippi-Taylor: Conceptualization, Writing - review & editing. Virginia Recta: Conceptualization, Writing - review & editing.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

References (7)

  • Neal Alexander

    Review: analysis of parasite and other skewed counts

    Trop. Med. Int. Health

    (2012)
  • Douglas W. Mitchell

    More on spreads and non-arithmetic means

    Math. Gaz.

    (2004)
There are more references available in the full text version of this article.

Cited by (0)

View full text