Low-dose cholecalciferol bait for possum and rodent control

Abstract

A new low-concentration formulation of Feracol^® paste containing 0.4% concentration cholecalciferol was tested for efficacy against possums in cage trials, and against possums and rodents in a single unreplicated field trial. A new low-concentration 0.4% cholecalciferol cereal pellet was also tested in the field against possums and rodents. 20 wild-caught possums, held in individual cages, were presented with 15 g of 0.4% cholecalciferol paste. Possum body weight, amount of paste eaten and days until death were recorded. The field trial occurred on Banks Peninsula, Canterbury, from September to November 2009, using a treatment site of c. 100 ha and a non-treatment site of c. 200 ha. The treatment site was divided into two even blocks, one for paste and one for pellets. Pre- and post-monitoring were carried out using leg-hold traps for possums, and tracking tunnels for rodents. Of the 20 caged possums, a kill rate of 89.5% with an average time to death of 7.0 days±0.41 SEM was achieved, similar to earlier work on cholecalciferol using a higher concentration. In the field trial pellet treatment block, possum numbers were reduced by 94±11%, mice numbers by 100±0% and rat numbers by 89±12%. In the paste treatment block, possum numbers were reduced by 93±16%, mice numbers by 75±29% and rat numbers by 100±0%. Under the current study conditions, the effectiveness of the new Feracol^® paste and pellets containing 0.4% concentration cholecalciferol was comparable to the currently registered 0.8% concentration cholecalciferol Feracol^® paste formula, and was effective in controlling multiple pest species while reducing the poisoning risk to non-target species. Replication of this field trial using the new bait matrices is necessary to increase confidence under similar and differing conditions, to ensure that the efficacy seen here is not arbitrary.

Keywords:

• bovine TB
• cholecalciferol
• Feracol^®
• mice
• Mus musculus
• New Zealand
• possums
• rats
• Rattus sp
• Trichosurus vulpecula
• vertebrate pesticide

Introduction

Brushtail possums (Trichosurus vulpecula) destroy native flora and fauna and are also the main wildlife source of bovine tuberculosis (TB) in New Zealand (Ramsey et al. 2002). Transmission of bovine TB from possums to cattle is a major factor preventing New Zealand from attaining TB-free status, as TB can persist in possum populations in the absence of other infected species (Parkes & Murphy 2003) and also infect TB-free herds. It is therefore one of the main reasons that possum control is undertaken in New Zealand (Parkes & Murphy 2003).

Rats (Rattus sp.) and mice (Mus musculus) do not harbour TB, but are very destructive to native flora and fauna. They feed on seeds and foliage from native trees (Empson & Miskelly 1999; Innes 2005) and are known to eat the eggs of native birds from nests (Imber et al. 2000). In some cases, rats directly attack and kill young birds on the nest (e.g. Imber et al. 2000; Innes et al. 2004). Bird population numbers are often observed to recover after rat control (Empson & Miskelly 1999), directly implicating rats, especially Norway rats (Rattus norvegicus) as predators on birds. Rats and mice are also known predators of native lizard populations (Newman 1994) and, as such, are also key targets for vertebrate pest control, alongside possums.

Currently, the most widely used pesticide for broad-scale pest control in New Zealand is sodium fluoroacetate (1080) (Eason et al. 2008), as it is the only non-anticoagulant poison registered for aerial control of terrestrial vertebrate pests on the mainland (Eason et al. 2010, 2011). However, public opposition to the use of 1080 can be high (Weaver 2003) and there is demand to develop suitable alternatives with minimal non-target species risk. Consequently, research and development to investigate alternatives to 1080 for terrestrial vertebrate pest control is ongoing.

One alternative toxicant is cholecalciferol (vitamin D3). It was developed in the 1980s as a rodenticide (Marshall 1984; Tobin et al. 1993) and is registered for use in both the US and Europe (Pospischil & Schnorbach 1994). It is synthesized in animal skin by the action of sunlight on its precursor, 7-dehydrocholesterol. Cholecalciferol in toxic doses raises blood calcium levels (hypercalcaemia) causing metastatic calcification of the blood vessels (Marshall 1984; Marsh & Tunberg 1986). Death from heart failure usually results, with animals dying within 3 to 7 days (Marshall 1984; Marsh & Tunberg 1986; Jolly et al. 1995), which is comparatively rapid compared with anticoagulant rodenticides. Fortunately, the toxicity of cholecalciferol to birds is low; for example, the LD₅₀ for cholecalciferol for mallard ducks is 2000 mg/kg (Marshall 1984) compared with an LD₅₀ for 1080 of 9 mg/kg for the duck family Anatidae (Eason et al. 2011). The risk of secondary poisoning to other non-target species is also low (Marshall 1984; Eason et al. 1996; Booth et al. 2004) making cholecalciferol a potential candidate as a suitable alternative to 1080.

Previous research has shown Norway rats and possums are susceptible to cholecalciferol with an acute LD₅₀ for Norway rats of 43.6 mg/kg (Marshall 1984). For possums, the acute LD₅₀ has been reported as 16.8 mg/kg (Jolly et al. 1995), although later research suggests that possums are more susceptible than thought, with an LD₅₀ of 6.4 mg/kg recently calculated (Morgan 2007).

Possum susceptibility to cholecalciferol was demonstrated in the early 1990s (Eason 1991; Eason et al. 1996). Feracol^® was first registered for possum control in New Zealand in 1999 with a concentration of 0.8%. Later possum field trials using a 0.9% cholecalciferol gel showed a high efficacy (81–100%). However, these results were probably underestimated as poachers killed possums in the non-treatment area (Morgan 2006, 2007). The actual kill is calculated based on what happens in the non-treatment site. Accordingly, a low number of possums in the non-treatment site adjusts the actual kill down. Feracol^® was then registered in 2008 as a multi-species bait for the control of possums and rodents.

As possums regularly weigh 2–4 kg compared with rats weighing < 500 g, Feracol^® containing 0.8% (or 8 mg/g) cholecalciferol should kill most rats in a weight range, e.g. 100–500 g, with between 0.5–3.0 g of bait. This amount of bait should be equivalent to or exceed the LD₅₀ for both species. However, Morgan's (2006) research indicated that 0.8% was too high for possums, and the need to retest the efficacy of cholecalciferol using a low-dose concentration of 0.4% arose. Given that rodents are smaller than possums, a lower dose should be equally as effective on them. Therefore, the aim of this research is to conduct first-phase field trials to examine whether low-dose cholecalciferol paste and pellets are as effective as the 0.8% concentration under field conditions encountered here.

Materials and methods

The following cage and field trial were approved by the Lincoln University Animal Ethics Committee (approval #270, #172B). The use of low-dose cholecalciferol (LDC) paste in cage, and paste and pellets in field trials, was approved by ERMA (Approval HSC 100003). Three samples of the LDC bait were analysed by high-performance liquid chromatography (HPLC) (Mauldin et al. 1999) to confirm 0.4% cholecalciferol concentration before the trials were started.

Cage trial

The effectiveness of LDC paste was tested on possums in a cage trial. All possums were wild-captured from a TB-free area on Banks Peninsula, Canterbury, and housed in individual cages. Water, green vegetables and grain pellets were available ad libitum.

20 possums were used for the cage trial. This number was selected as it was the maximum amount of sampling error the authors were willing to accept, while minimizing animal welfare issues and cost. The number of animals used complies with the New Zealand Food Safety Authority guidelines for assessing the efficacy of vertebrate pesticides (NZFSA 2002). Possum body weight was measured before and after the trial and they were fasted for 24 hours before exposure to the toxic paste. Each possum was presented with 15 g of LDC paste, the only food given for the first 2 days of the trial. The amount of LDC paste consumed was recorded and probit analysis (Finney 1971) was used to estimate the LD₅₀ mg/kg dose using Genstat version 12 (VSN, International).

Field trial

Trial site

A single unreplicated field trial was conducted. The trial site consisted of two north-facing gullies at Kinloch near Little River on Banks Peninsula, Canterbury. One gully was used as a non-treatment site and the other a treatment site. The two sites were approximately 800 m apart and, for the purpose of this experiment, considered independent of each other (see below). The treatment site was approximately 100 ha, divided into two 50 ha blocks; one for the application of LDC paste and one for LDC pellets; the control site was approximately 200 ha. The two treatment blocks had a 300 m no-monitor zone in the middle, i.e. control occurred to the boundary of both blocks, but no monitoring occurred for 150 m either side of the boundary line. Considering the short time frame of the trial, the blocks within the treatment site were also considered independent of each other, as previous research has shown that very few of the same possums frequent bait stations that are spaced more than 200 m apart (Thomas & Fitzgerald 1995). Possum home-range studies and current possum best practice dictates that bait stations should never be further than 150 m apart (i.e. one per 2.25 ha) in forest habitats (Thomas 1994) given that the average home range for male possums is 1.9 ha, and 1.3 ha for female possums (Cowan & Clout 2000). Vegetation consisted of open pasture, manuka (Leptospermum scoparium), the native tree nettle ongaonga (Urtica ferox) and regenerating scrubland. LDC pellets were placed in 120 Sentry^® bait stations; LDC paste was placed in 123 Romark^® bait stations. Bait stations were set up on a 100×100 m grid in areas where vegetation was sparse, and a 50×50 m grid in areas of thick scrub.

Pre-feeding and toxic baiting

Pre-feeding was carried out at the treatment site. On two occasions, 100 g of non-toxic pellets were placed in the Sentry^® bait stations, and 80 g of non-toxic paste was placed in each Romark^® bait station, 1 week apart, from September to November 2009. Toxic baiting was conducted 1 week after the November pre-feeding, when equivalent amounts of LDC pellets and paste (100 g and 80 g) were placed in the respective bait stations. Toxic bait take was checked 3 and 7 days after placement. If bait stations were empty, they were refilled with the same amount of paste or pellets.

Monitoring

Possums

Possum abundance was measured pre- and post-control using the residual trap catch (RTC) method with #1 leg-hold traps (NPCA 2008). Lines were placed 250 m apart to be independent sampling units, with 10 traps, at 20 m intervals, per line. Four lines were placed in the treatment site, two per block, and four lines in the non-treatment site. These were set for 3 consecutive nights, both pre- and post-control. All possums trapped during pre-control monitoring were released and percentage reduction was calculated using the NPCA (2008) possum monitoring protocol formula below:

where Mean_pre is the mean catch of the trap lines pre-control, and Mean_post is the mean catch of the trap lines post-control. A full explanation of how the Mean_pre&post estimates were calculated for possums is given in NPCA (2008).

Following calculation of the percentage kill, changes in the non-treatment site were used to adjust the percentage kill estimate, enabling estimation of the mean percentage change directly attributable to the treatment. For example, if no possums had been removed, then similar Mean_pre:Mean_post ratios for trap catch would be expected in the treatment and non-treatment site. This generates the following equation:

where subscripts _t and _nt indicate treatment and non-treatment sites, respectively. By rearranging the formula we get:

Finally, the adjusted kill is estimated by:

The standard error of the kill was then calculated using the formula below:

where C² _pre and C² _post is the coefficient of variation (squared) for the pre- and post-estimates; pptn refers to possums per trap night on each line; and Correlation is the correlation coefficient between the pre- and post-control lines (see NPCA 2008). A 95% confidence limit is the standard error multiplied by two. The authors acknowledge that some statistical assumptions may be violated using these equations, so the 95% confidence intervals reported here are approximations only.

Rodents

Rat and mice abundance was measured pre- and post-poisoning using Black Trakka™ tracking tunnels. Tracking tunnels were made of polypropylene, measured 55 cm long and contained a cardboard tracking card with ink in the middle. Peanut butter bait was placed in the middle of the tunnels. Eight lines of five tunnels were placed in both the treatment and non-treatment sites; therefore, four lines were placed in both the pellet and paste blocks. Tracking tunnels were placed at 50 m intervals over a 2-night period and followed a standardized rodent monitoring protocol (King & Edgar 1977; Blackwell et al. 2002). The percentage kill was estimated for rats and mice. These were adjusted and a standard error for the kill estimated using the same equations detailed above. A full explanation for how the Mean_pre&post estimates were calculated for rodents is given in Gillies & Williams (2006a). It was noted that no tracking tunnel lines recorded 100% activity pre-control, i.e. there was no evidence of saturation. All rodent footprints were identified from Gillies & Williams (2006b).

Results

Cage trials

Assay tests conducted on the cholecalciferol used for this research confirmed the concentration was 0.4%. One possum tipped the food out on both nights, and therefore did not eat any paste; this possum was excluded from the results. 17 of the remaining 19 possums exposed to LDC paste died (89.5%) in an average of 7.0 days±0.41 SEM. Of the 19 possums that ate paste, 16 ate the maximum amount presented (15 g), while three possums consumed 12, 13 and 14 g, respectively. As the cage trial used a fixed paste amount (15 g) and 16 of 19 possums ate the maximum amount presented, correlation between possum live weight and amount of paste bait eaten was not examined. The three possums eating less than 15 g of bait all died, while two possums that ate the full 15 g did not die. Possum mean weight loss over the trial duration was 15.2±1.4% SEM. An LD₅₀ of 16.7 mg/kg was estimated; however, as the study was designed using a fixed-concentration dosing regimen (i.e. only 15 g of 0.4% concentration bait was presented to each possum), the LD₅₀ is considered indicative only. This value is similar to an earlier reported value of 16.8 mg/kg (Jolly et al. 1995), but higher than that calculated by Morgan (2007) of 6.4 mg/kg; however, different cholecalciferol sources and bait matrices were used.

Although rodents were not tested here, parallel cage studies using rats and mice confirmed that doses as low as 0.075% effectively killed rats, but this dose was less effective for mice (Hix 2009a). A further captive trial confirmed that a dose of 0.4% effectively killed both rats and mice (Hix 2009b) and is therefore likely to be an effective multi-species bait in the field. Based on these results, and the above caged possum trial, it was decided to carry out field trials using possums, rats and mice.

Field trials

Amounts of LDC paste and pellets taken from the respective bait stations were recorded after 3 and 7 nights. Nine Romark^® stations (7.3%) containing LDC paste and 22 Sentry^® stations (18.3%) containing LDC pellets were empty at 3 days, and refilled. None of the Romark^® or Sentry^® stations was empty at the day 7 check. Reductions in possum trap catch exceeding 90% were recorded in both paste and pellet blocks; whereas, a reduction of only 12±2.2% was recorded in the non-treatment site. Based on these results, the corrected percentage kill for possums in the treatment site was 92.5±15.7% for paste and 94.4±11% for pellets (Table 1).

Table 1  Actual and corrected percentage reductions from both paste and pellet treatment areas based on residual trap-catch of possums pre- and post-control. Mean pre- and post-control trapping line rates are shown±SEM. All corrected% reductions are shown±95% CI.

	Pre-control	Post-control	% reduction	Corrected% kill
Paste	25.6±8.9	1.7±1.7	93.8±7.6	92.5±15.7
Pellet	34.6±4.0	1.7±1.7	95.1±5.2	94.4±11.0
Control site	13.7±3.7	12.0±2.2	12.2

A reduction of greater than 90% was seen in rats for both paste and pellet post-treatment (Table 2), whereas mice numbers reduced by 50% for the paste treatment, but by 100% for the pellet treatment (Table 2). In the non-treatment site, rat numbers reduced by 5.6%. In contrast, mice numbers showed a 100% increase in the non-treatment site. Therefore, the corrected percentage kills for rats and mice, respectively, is 100±0% and 75±28.9% for paste, and 89.4±11.7% and 100±0% for pellets.

Table 2  Actual and corrected percentage reductions from both paste and pellet treatment areas based on tracking tunnel rates of rats and mice pre- and post-control. Mean pre- and post-control tracking tunnel rates are shown±SEM. All corrected% reductions are shown±95% CI.

Rats	Pre-treatment	Post-treatment	% reduction	Corrected% kill
Paste	60.0±24.5	0.0±0.0	100.0±0.0	100.0±0.0
Pellet	50.0±23.8	5.0±5.0	90.0±12.0	89.4±11.7
Control	45.0±9.1	42.5±5.9	5.6
Mice
Paste	10.0±5.8	5.0±5.0	50.0±62.7	75.0±28.9
Pellet	45.9±17.1	0.0±0.0	100.0±0.0	100.0±0.0
Control	2.5±2.5	5.0±3.3	−100.0

Certainly a lack of replication weakens the confidence of these results, as seen by the large confidence intervals. Nonetheless, it does not detract from the fact that this is valuable early-stage research using a toxin with lower non-target risks than 1080, and shows promising results.

Discussion

Possums are known carriers of bovine TB, able to retain infection in the absence of other infected species (Parkes & Murphy 2003), thus acting as TB reservoirs and posing a constant threat to New Zealand's cattle and meat export trade. Possums, rats and mice are also well-known consumers of native flora and fauna. As such, considerable time, effort and money are spent on possum control (Livingston 1994; Morgan 2004) to reduce economic losses to the meat industry and also to protect native flora and fauna. To date, 1080 has been the most heavily relied upon vertebrate pesticide used in New Zealand for widespread vertebrate pest control (Eason et al. 2008), but constant research into low-risk alternatives has led to some successful breakthroughs. In this preliminary trial, an environmentally safer bait has been trialled to target possums, rats and mice. Results from this first field trial suggest that bait containing 0.4% cholecalciferol was successful in reducing all three target species. If results can be replicated with equal success at other sites, then the use of low-dose bait matrices will provide the added bonus of increasing benefits to native species by reducing multiple predators with one control operation.

The cage trial conducted here shows that possums are highly susceptible to LDC paste, dying in an average of 7 days; this is similar to the trial reported by Jolly et al. (1995), where possums died in an average of 7 days even though cholecalciferol concentrations were higher than that used here. This shows that under the chosen laboratory conditions, cholecalciferol at 0.4% instead of 0.8% effectively kills possums in a similar time frame as that previously reported (a mean of 9 days, O'Connor et al. 2003).

Previous cage trials have shown that cholecalciferol ingestion can cause excessive weight loss. In O'Connor et al. (2003), it was suggested that 21% of possums lost more than 30% of their body weight. Little information was given on how cage trials were carried out by O'Connor et al. (2003), thus it is difficult to speculate on the differences in possum body weight loss seen between the results from O'Connor et al. (2003) and those reported here. Here, mean weight loss of possums in the cage trial was 15.2%. Cholecalciferol has a ‘stop feed effect’ whereby the toxin causes a loss of appetite, usually within 24 hours of first being consumed (Morgan & Rhodes 2000). Undue suffering caused by the stop feed effect can be exhibited by excessive weight loss in the animal in a short period of time. Excessive weight loss was a concern in lowering the toxic concentration to 0.4%. However, the captive possum trial indicates that the weight loss caused by LDC does not cause extended suffering when compared with the 0.8% concentration.

The substantial reduction in possums and rodents in the treatment site indicates that the new 0.4% cholecalciferol Feracol^® paste formulation has not reduced the efficacy of the product under these field conditions; however, as a small number of both paste and pellet bait stations were found empty on day 3, a slight increase in the amount of toxic bait provided could be warranted. Cholecalciferol has a lower risk of poisoning domestic pets such as dogs when compared with 1080 (Eason et al. 2000); therefore, if the low-dose formulation proves equally successful under other field conditions, use of these low-dose bait matrices will contribute further towards reducing the risk of primary poisoning to dogs and non-target bird species. Effective control of possums and rodents within the same control operation overcomes some consequences of traditional, single species control, such as meso-predator release (Tompkins & Veltman 2006).

In conclusion, although previous research has demonstrated that the 0.8% concentration effectively kills rats, mice and possums (Eason et al. 2010), the current field trial is the first to demonstrate the field effectiveness of 0.4% LDC paste and pellets for possum and rodent control. However, until further replication is carried out, the field efficacy results should only be accepted as preliminary for the conditions tested here. Although very positive results were obtained, they should be viewed as the initial phase of ongoing research; replication of low-dose cholecalciferol at other field trials should strengthen the results seen here and confirm that 0.4% cholecalciferol is effective in reducing populations of possums, rats and mice.

Acknowledgements

Thank you to Lyn Booth for undertaking the HPLC bait analyses and the Animal Health Board for contracting this work (Contract R-80706).