The evaluation of feeding, mortality and oviposition of poultry red mite (Dermanyssus gallinae) on aging hens using a high welfare on-hen feeding device

Ethical statement

Every effort was made to emeliorate harm to the hens used in this study: the staff involved with the feeding assays were already experienced having optimised the device previously (Nunn et al, 2019). Plucking of feathers was kept to a minimum, just enough to allow the pouches to be applied and plucking is not required once the hens reach ~22 weeks of age and adult feathers are present. Bandaging and tape was veterinary grade and had previously been shown to cause no harm and both handlers were practised in its usage. Hens were kept together at all times to prevent stress.

All experimental procedures described here were approved by the Moredun Research Institute Animal Welfare Ethical Review Body (approval E20/18) and were conducted under the legislation of UK Home Office Project License (reference P46F495BD) in accordance with the Animals (Scientific Procedures) Act of 1986. The manuscript was written in adherence with the ARRIVE guidelines.

Hens

Five Lohman brown pullets (18-weeks-old at the start of the experiment), purchased from a local pullet breeding facility, were housed in an enclosed, loose litter floor pen of 2.5 m × 2 m with temperature (18 °C) and lighting (16 h light/8 h dark) controlled to mimic commercial hen laying conditions. Hens were in good condition with no evidence of previous exposure to D. gallinae. Hens had ad libitum access to layer’s pellets and water, perches and nest boxes were provided.

Parasite material

For provision of parasite material, mixed stage and sex D. gallinae were collected every two weeks into vented 75 cm² canted tissue culture flasks (Corning, New York, USA) from a commercial egg laying unit. The optimum quantity of collected mites/detritus placed into a tissue culture flask was ~ 2-3 cm deep when the flasks are upright. Less than this quantity and mites desiccated over the subsequent “conditioning period”; more than this and mites were overcrowded and died. Any overcrowded culture flask (becomes wet and dirty inside) had the contents transferred into fresh flask(s) as many times as required during the seven days after collection, when the flasks were stored at room temperature (RT). Adult females and deutonymphs were isolated (see Figure 1) from this ‘mite pool’ for feeding assays following conditioning.

Protonymphs were provided by hatching mite eggs and allowing the emerging larvae to moult, as follows: Mite eggs were harvested by collecting freshly fed mites from the caps of 75 cm² tissue culture flasks and transferring ~1 cm³ fed mites to individual 30 ml polystyrene Universal tubes (Thermo Scientific, Mass. USA), sealed with a double layer of AeraSeal™ film (Sigma, MO, USA) and incubated overnight at 25°C and 85% relative humidity (RH). The tubes were unsealed and placed onto glass Petri dishes in a containment moat of Flash™ multipurpose cleaning fluid (Procter and Gamble, Harrogate, UK) in a 250 ml plastic weighing boat (Scientific Laboratory Supplies, Nottingham UK). Mite eggs were then able to be picked out of the polystyrene Universal tube using a fine paintbrush and collected in 5 ml polystyrene bijoux tubes (Thermo Scientific, UK), sealed with a double layer of AeraSeal™ film (Sigma) and incubated at 25°C and 85% RH for a three-day period to allow eggs to hatch into hexapod larvae and these larvae to then moult into octopod protonymphs. The protonymphs were stored in the bijoux tubes and stored at 4°C until use. This protonymph hatching protocol made counting and handling of the small protonymphs for subsequent use easier than attempting to pick them from the mites and detritus from ‘mite pool’ field collections.

For mite conditioning prior to feeding assays, deutonymphs and adults were stored in detritus in vented 75 cm² canted tissue culture flasks (Corning) at RT for seven days after field collection to allow for complete digestion of the blood meal. All stages of mites were then stored at 4°C for two weeks (deutonymphs) or four weeks (protonymphs and adult females) before being used in feeding assays. These optimal conditioning periods had previously been empirically determined to be necessary for maximum feeding of each stage (Nunn et al., 2019).

Experimental design and statistical analysis

Assays were carried out, on the same hens as they aged, at two-week intervals over an 18-week period. At each assay point hens either wore feeding devices containing 50 adult female mites and 50 protonymph mites or 50 adult female mites and 50 deutonymph mites. Hens were randomly-allocated feeding devices (“haphazard allocation”) and assays were carried out on Tuesday and Thursday mornings in each week in which assays were performed. At the end of the feeding period, hens were selected randomly (“haphazard selection”) for removal of the feeding devices. Blinding of hen identification numbers to the mite counter was considered to be unnecessary during experimental procedures and analysis as the single experimental group had no specific treatments to be compared with another group. Preliminary work carried out during the optimisation of the feeding device, had demonstrated that the nymph stages of D. gallinae feed better when in the presence of adult mites (“co-operative” feeding; Table 1).

Power calculations were based on simulated mite feeding data using a binomial generalised linear mixed model (GLMM) to establish the numbers of hens required to detect a change from expected scenario of 40% to worst case scenario of 15% in mean feeding rate, with baseline values for variability in feeding rate within bird over time (SD = 0.25) and between birds (SD = 0.6) estimated from two pilot previous assays using birds of different ages (Table 2). The estimated time SD was multiplied by two to consider the case of feeding rates doubling in variability over an extended trial. The results indicated that four birds is the minimum to reach the standard 80% statistical power threshold at a 5% significance level under those conditions, however we used five to cover for unexpected deaths in this long study.

To compare mite mortality and feeding rates over time, GLMMs were fitted by the maximum likelihood method, using the Laplace approximation and a logit link function, to the data including lifestage (protonymph, deutonymph or adult), time (weeks after the start of the experiment) and the interaction between them as fixed effects and hen identity as a random effect. No criteria were set a priori for inclusion/exclusion of animals or data points although one data point was excluded as considered an extreme outlier in mite mortality being 21 times the upper limit of the interquartile range for the protonymph mortality data collected n that week (mortality 57.14%, time death observed 48h, week 10). For mite mortality, time at which death post-feeding was observed was included as an additional random effect. The effect of hen aging on oviposition was formally tested using a Poisson GLMM fitted by maximum likelihood, based on the Laplace approximation and a logarithmic link function, including an offset (in logarithmic scale) to account for the number of fed mites, with time as a fixed effect and hen identity as a random effect. Statistical testing from these model fits was conducted using type-II Wald chi-square statistics. Pairwise tests of differences in mean between lifestages and timepoints were conducted based on the predicted marginal means from the GLMM estimates. The corresponding p-values were adjusted for multiplicity using the Benjamini and Hochberg’s method (Benjamini & Hochberg, 1995). Significance tests were assessed at the usual 5% significance level. All analyses were performed using R Core Team, 2018.

Protocol for each mite feeding event

Step 1: Manufacture of feeding devices. Feeding devices were made in advance of each feeding event, as shown in Video 1, described in the text and shown in Figure 2.

Figure 2. Assembled feeding devices ready to fill with poultry red mites.

The tape overhangs assist with handling and the blue ‘brace’ acts both as a refuge for mites and in keeping apart the mesh sides.

To make the pouch (Figure 2), a strip of polyester mesh [aperture width of 105 μm, pore depth 63 μm (Plastok Ltd, Birkenhead, UK)] was cut to 6cm wide by 15 cm long. Using 1.25 cm diameter Leukoplast® tape (BSN Medical, Germany), a seal was made on the long edge to make a tube, ensuring that the Leucoplast tape was adhered to the inside of the mesh tube. A strip of AeraSeal™ (Sigma-Aldrich Co Ltd, Dorset, UK) tape was attached to the outside of the mesh tube, to cover the Leukoplast® tape join. The tube was then cut into 3 cm segments and one of the open ends sealed with c.a. 3 cm strip of 1.25 cm diameter Leukoplast® tape, leaving an overhang at each end of the pouch. To make the plastic insert, the shaft of a blue plastic bacterial spreader loop (Quadloop Sphere, Sterilin) was gently heated in a Bunsen flame and placed onto a hard surface then bent into a U shape. The ends of the loop were trimmed and filed to remove sharp edges. The plastic inserts were then placed in pouches (one per pouch) ready to fill with mites (Figure 2).

Technical tips:

1. Use fabric shears or a paper guillotine to cut the mesh to prevent the mesh from fraying.

2. Keep all edges straight to minimise the amount of sticky Leukoplast® tape adhesive that the mites come into contact with.

3. Covering the outside of the mesh tube long join with AeraSeal™ tape, covers the Leukoplast glue and makes mite recovery easier after the feeding assay.

4. Making the bottom and top seals longer than the pouch width (overhangs) enables handling of the pouches without squashing mites-it also helps in pouch removal from the hen.

Step 2: Filling the devices with mites. To fill the feeding devices (pouches) with mites, flasks containing conditioned mites were removed from storage at 4°C, and then kept at RT for 20–30 minutes to allow the mites to become mobile. Two centimetres of wet ice were placed into two large (280ml) chemical weigh boats (Fisherbrand). The cap of the flask was unscrewed and tapped out onto a glass Petri dish and the mites allowed to disperse for a few seconds before placing Petri dish on the ice in the weigh boat (Figure 3a). Mites were sorted using a fine paint brush and transferred to the second glass Petri dish on ice until sufficient numbers for one feeding device were obtained. The mites are then gently gathered into a ball using the brush and placed into the feeding device which was then quickly sealed with a piece of Leukoplast® tape (BSN Medical, Germany) (Figure 3b). Filled feeding devices were kept overnight in a loosely sealed plastic container placed in polystyrene box containing wet ice. The plastic container was raised above the wet ice on a metal support, so that it was not in direct contact with the ice. This kept the temperature sufficiently low (around 10°C) and the relative humidity level high, so that the mites did not desiccate.

Figure 3. Showing poultry red mite handling using weighboats and ice (Panel A), filled pouches (Panel B) and pouch application kit (Panel C).

Technical tips.

1. When transferring mites from the flask cap to Petri dish, do not use a cold Petri dish or put Petri dish on ice first as the mites will ‘ball up’ and make it very difficult to pick individuals.

2. Work quickly to limit the amount of condensation on Petri dishes - once a dish gets very wet it makes picking mites up more difficult and they won’t be in the best condition going into the feeding device.

3. When transferring the selected mites from the Petri dish to the device, place the device close to the mites on the cold Petri dish and try to transfer them in one movement. Work quickly! The blue plastic brace is a convenient place to quickly wipe the mites off the paint brush.

4. If it is convenient to fill the devices the day before application to the hens, store the devices overnight in a loosely sealed plastic container placed in an insulated box containing wet ice. Raise the container above the wet ice on a support, so that it isn’t in direct contact with the ice. This keeps the temperature sufficiently low (around 10°C) and the RH level high, so that the mites do not desiccate.

Step 3. Application of feeding devices to hens. The pouch application kit is shown in Figure 3c. A video clip demonstrating pouch application can be seen in Video 1.

Prior to the first feeding assay, feathers were plucked from the outer thigh areas of each hen and thereafter whenever needed. Before laying the hen on its side, the pouch was prepared for application by cutting four strips of 1.25cm wide Leukoplast® tape. Two of those strips were cut slightly shorter for the shorter sides of the pouch. Holding the pouch with the central join facing upwards, a strip of tape was placed on each of the four sides of the pouch. The size of the pouch with the tape strips added was not larger than the plucked area of the hen’s leg (Figure 4, panels A and B), thus preventing the tape from causing irritation to the hen while it is attached and stops the tape pulling on feathers when it is removed. The device was further secured with cohesive bandage (Central Medical Supplies Ltd., Pontefract, UK) wrapped over the feeding device which not only secures the device onto the hen but also provides dark conditions to encourage mite feeding. Feeding devices were left in place for three hours at 9am in the morning, to allow for mite processing in the afternoon. During this time hens were allowed to carry out normal foraging and nesting behaviours. Handfuls of mixed corn were scattered in the litter to encourage foraging behaviour while the pouches were attached. Following the completion of feeding (three hours after the application of the device), the devices were removed and fully engorged and partially fed mites were recovered from the devices as described below.

Figure 4.

The stages of pouch application with Panel A showing hen restraint and plucked thigh. Hens are placed gently on their side with the handler using one hand to restrain the hen and the other to carefully straighten the leg being used. Panel B shows how the four strips of Leukoplast® tape are used to loosely adhere the pouch to the skin (described in Step 3) before the final bandaging using the cohesive bandage (Panel C). The cohesive bandage should be firm enough to be secure but not so tight as to cause any discomfort.

Technical tips

1. Practise bandaging on yourself or a colleague first –too tight will be uncomfortable for the hen, too loose and mites might not be able to attach for feeding.

2. A piece of leucoplast tape can be used to secure the end of the cohesive bandage if necessary.

3. If required, cohesive bandage can be cut in half (i.e. from 5 cm width to 2.5 cm width). This can be done while bandage is still on the roll, using a Stanley knife or scalpel blade.

Step 4. Mite recovery from the feeding device. A video demonstrating how to dissect the pouch can be found in Video 1. The bandage can be unwound and removed completely. Slowly and carefully remove the device from the hen’s skin, peeling it off in the same direction of feather growth as this prevents damage to the hen’s skin.

Pouches can then be dissected as shown in Figure 5a and 5b.

Figure 5.

a) A pouch post feeding assay; b) partially dissected pouch; c) tissue culture plates containing fed mites for monitoring.

When taking the feeding device apart, the tape that was used to stick the pouch to the hen was removed first. Then, one of the tape seals was removed by holding one end tab on one side of the device and gently pulling the tab on the opposite side as shown (see Figure 5a and 5b). Once the seal on one side of the device had been peeled off, the blue plastic brace was then removed using forceps. The blue plastic brace and the semi-dismantled device was placed onto a glass Petri dish on ice to prevent mites escaping. Then fed/unfed mites were recovered from it. Mites generally crawled out from the feeding device during this time and were recovered from the surface of the device. The feeding device was then carefully dissected with scalpel and forceps to allow collection of and remaining mites.

Fed mites, identified by the presence of internalised hen blood, were counted, any dead mites (judged by curled legs and lack of movement) counted and then placed individually into wells (Figure 5c) of 96-well tissue culture plates (Costar, Corning, NY, USA), sealed with AeraSeal™ tape (Sigma-Aldrich Co Ltd, Dorset, UK) and incubated at 25°C with 85% relative humidity and monitored for mortality and fecundity, 48-hour intervals for six days. Egg counting per fed mite was used as the measure for fecundity.

Technical tips:

1. If visualisation of the mites under light microscopy is required following feeding, isolating them in a 96-well tissue culture plates (Costar, Corning, NY, USA) allows clear microscopical visualisation of the mites through the underside of the plate. Moulting, oviposition and mortality can be scored using this system.

2. We have found AeraSeal™ to be the best tape to use as mites are less likely to stick to than to other porous tapes.

Comparison of feeding by mite life stage and time point

There were statistically significant overall effects of mite lifestage (when comparing protonymph, deutonymph or adult) and of the interaction between this and time (i.e. hen aging) on the proportions of mites feeding on hens when applied to the hens in the feeding devices (p < 0.0001 in each case) (Nunn et al., 2020a). The differences in predicted feeding rate (proportion of all mites in each feeding device expected to feed successfully) by mite lifestage are shown in Figure 6 (and the raw data supporting this in Table 3).

Figure 6. Feeding of each lifestage of Poultry Red Mites PRMs in on-hen feeding devices placed on hens across an 18-week period.

Each data point represents the predicted feeding rate of mites, aggregated across all timepoints in the trial (± 95% CI) based on recorded percentages of 100 adult mites or 50 protonymph or deutonymph mites which had fed on each of five replicate hens at each timepoint.

Feeding rates averaged over all 10 timepoints across the 18-week period were approximately 1.5 times higher for adult mites than for either deutonymphs or protonymphs (p < 0.001 in both cases) but there were no statistically significant differences between the feeding rates of the two nymphal stages (p = 0.2437). When examined between timepoints in the experiment (i.e. as the hens aged) there was notable variation in feeding rates at different timepoints for deutonymphs and protonymphs (Figure 7, panels B and C with supporting data in Table 4). For example, at week 6 deutonymphs had much higher feeding rates than at every other timepoint (p < 0.0001 in each case), though no overall trend of feeding rates across time was evident (Figure 7B). There was substantially less variation in adult feeding across timepoints (Figure 7A), though feeding rates at week 16, for example, were significantly lower than on all other dates (p < 0.05). Again, no overall trend of feeding rates across time was obvious for adult mites.

Figure 7. Feeding of each lifestage of Poultry Red Mite (PRM) on hens across an 18-week period.

Panel A shows data for adult females, Panel B for deutonymphs and Panel C for protonymphs. Each data point represents the predicted feeding rate of mites (± 95% CI), based on recorded percentages of 100 adult mites or 50 protonymph or deutonymph mites which had fed on each of five replicate hens at each timepoint.

Comparison of mite mortality by mite life stage and timepoint

There was no statistically significant overall effect of mite lifestage (when comparing protonymph, deutonymph or adult) on mite mortality post-feeding (p = 0.2907). However, there was a significant effect of the interaction between lifestage and time (i.e. hen aging) (p = 0.0006). This was examined further by investigating post-feeding mortality rates by time point for each lifestage (Figure 8 and Table 5).

Figure 8.

Mortality of each lifestage of Poultry Red Mite (PRM) which had fed on aging hens across an 18-week period with Panel A showing data for adult females, Panel B for deutonymphs, and Panel C for protonymphs. Each data point represents the predicted mortality rate of mites (± 95% CI), where a value of 1 would represent 100% mortality, based on recorded percentages of 100 adult mites or 50 protonymph or deutonymph mites which had fed on each of five replicate hens at each timepoint

Mortality rates were variable for adults and deutonymphs across the time period of the experiment. For adults (Figure 8A), mortality at week 4 was significantly lower than at weeks 0, 6, 8 and 18 (p < 0.05) and, in addition, mortality rates for adults at week 18 was significantly higher than at weeks 4 and 16 but no clear pattern over time could be discerned across the whole experiment. For deutonymphs (Figure 8B), mortality at week 12 was significantly higher than at weeks 0, 4, 10, 14 and 16 (p < 0.05) and mortality at week 18 was significantly higher than at week 16 (p < 0.05). Again, no discernible trend in mortality over time could be determined for deutonymphs. In contrast, for protonymphs, there appeared to be a weak positive trend in mortality as the hens aged (Figure 8C), with significant increases in mortality of protonymphs between week 0 and weeks 6, 10, 14, 16 and 18 (p < 0.05). Overall, levels of mortality were very low in each lifestage albeit with variability in mortality rates observed for all lifestages as indicated by the wide confidence intervals.

Comparison of mite oviposition by timepoint

Analysis of the oviposition of adult mites (total eggs per surviving, fed female mite monitored over the six-day period for each time point) over time demonstrated that there was a statistically significant overall effect of hen age on the capacity of adult mites to lay eggs after feeding on these hens (p < 0.0001; Figure 9 and Table 6), with a trend of oviposition dropping as the hens aged.

Figure 9. Oviposition of Poultry Red Mite (PRMs; n = 1964) which had fed on aging hens across an 18-week period (boxplots and actual values of egg counts in logarithmic scale).

Data shows the total number of eggs laid per fed, surviving female mite as monitored over a six-day period following each assay week.

Thus, average mite oviposition at week 0 was statistically significantly higher than that at all subsequent weeks (p < 0.0001) and average oviposition at week 18 was significantly lower than that at all previous timepoints (p ˂ 0.0012).

Underlying data

Figshare: The evaluation of feeding, mortality and oviposition of poultry red mite (Dermanyssus gallinae) on aging hens using a high welfare on-hen feeding device_DATA.xlsx. https://doi.org/10.6084/m9.figshare.12848432.v1 (Nunn et al., 2020a).

Data are available under the terms of the Creative Commons Attribution 4.0 International license (CC-BY 4.0).