The required sample size in vaccination-challenge experiments with infectious bronchitis virus, a meta-analysis

Abstract

For statistical, animal welfare and financial reasons the choice of the number of chickens per group in experiments is important. This estimation, together with the number of tracheal organ cultures (TOCs) that need to be examined from each chicken in order to assess protection, should be based on the difference in level of protection that one would like to be able to detect (effect size), the expected variability of the results between and within the chickens, the desired confidence level and the power of the study. To obtain data that would facilitate this estimation, a meta-analysis was performed on the data from 18 infectious bronchitis virus (IBV) vaccination-challenge experiments performed at the Dutch Animal Health Service Deventer, the Netherlands (GD) in order to determine and quantify the source of variation in the mean level of protection of different groups. For the calculations, 137 groups of chickens were subdivided into 10 clusters based on age (young or adult), vaccination (none, homologous or heterologous), challenge (IBV or mock infected) and location of vaccination (isolator at GD or in the field). The results were used to estimate the required number of chickens per group for the different clusters using 2, 5 or 10 TOCs per chicken to be able to detect effect sizes of 6.25%, 12.5%, 25% and 50% between groups of chickens with 95% confidence (PB0.05) and 80% power. The number of chickens that was required for the mentioned effect sizes varied greatly from 2 to 650. This meta-analysis provided data that allow research workers to estimate the number of chickens that should be included in each group in order to obtain reliable results based on particular combinations of infectious bronchitis vaccination and challenge strains as defined by the presented clusters.

Infectious Bronchitis

“…When vaccinations with a single serotype are not providing sufficient protection against the prevailing field strains, vaccines of other serotypes can be added to the program. The broadening of the protection can be achieved by adding vaccines to the program that are homologous to the most important prevailing field strains (80,81,100,104,108) and/or by using combinations of vaccines that are able to induce a broad cross-protection against many strains (45,63,79).…”

Variability in biological behaviour, pathogenicity, protectotype and induction of virus neutralizing antibodies by different vaccination programmes to infectious bronchitis virus genotype Q1 strains from Chile

“…The level of protection of the trachea was determined using the ciliostasis test on five tracheal rings per chicken at 5 days post challenge (De Wit et al, 2013). The level of beating of the cilia in each ring (TOC score) was expressed as 4 (<25% beating of cilia), 3 (25-50% beating), 2 (50-75% beating), 1 (75-99% beating) or 0 (all beating).…”

Factors influencing the outcome of infectious bronchitis vaccination and challenge experiments

“…Generally, protection is assessed at between 5 and 7 d.p.c. (Ratanasethakul & Cumming, 1983;Gelb et al, 1991;Cook et al, 1999;Alvarado et al, 2003;Terregino et al, 2008;Jackwood et al, 2009;Lim et al, 2012;De Wit et al, 2013), although Darbyshire & Peters (1985) sampled 4 d.p.c. Lambrechts et al (1993) investigated levels of cross-protection in the respiratory tract at 1, 3, 5, 7 and 9 days following aerosol challenge with the B1648 IBV variant in chicks that had been vaccinated with different IBV vaccines.…”

“…For statistical, bird welfare and financial reasons, the choice of the number of chickens per group in experiments is important. The estimation of the required number should be based on the difference in level of protection that it is desirable to detect (effect size), the expected variability of the results between and within groups of chickens, the desired confidence level and the power of the study (De Wit, 2000;De Wit et al, 2013). A meta-analysis performed on the data of 18 IBV vaccinationchallenge experiments performed at The Dutch Animal Health Service, Deventer (GD) showed that the number of birds required per group in order to be able to detect effect sizes of 6.25%, 12.5%, 25% and 50% between groups with 95% confidence (P < 0.05) and 80% power varied greatly from two to 650 depending on both the kind of vaccination programme and on the challenge virus.…”