The detection of rodent resistance to anticoagulants

Bull. Org. mond. Santd 11973, 48, 239-242 Bull. Wld Hlth Org. I The detection of rodent resistance to anticoagulants D. C. DRUMMOND1 & B. D. RENNISON 1 The procedures used in the United Kingdom to detect anticoagulant resistance in the Norway rat are described. Preliminary inspections are made of the suspect rat populations to exclude other possible causes ofcontrolfailure. The inspections arefollowed by carefully supervised warfarin treatments in which the feeding pattern over 2-3 weeks is compared with that of normal rats. If rats continue to take the bait longer than is usual, laboratory tests are undertaken on trapped animals. It is believed that the methods will be found to be generally applicable to the investigation ofanticoagulant resistance in other rodents in many parts of the world. This review is concerned mainly with the methods currently being used in the United Kingdom to detect resistance in the Norway rat (Rattus norvegi- cus). Although the detection techniques described may be somewhat inappropriate for other types of resis- tance and for other rodent species in other situations the general principles of resistance detection should remain the same and it should not be difficult to modify details of the techniques to suit new situa- tions. When cases of apparent resistance to anticoagu- lants first appeared in the United Kingdom, the first action taken was to inspect the site of the trouble and to seek some explanation for it other than resis- tance to the rodenticide. Where no immediately obvious alternative explanation was forthcoming, the inspection was followed by a carefully supervised warfarin treatment. Where this treatment failed, again without any obvious explanation other than resistance, some of the surviving animals were trapped and subsequently tested in the laboratory. These three consecutive procedures of inspection, treatment, and laboratory tests are all still used and will now be considered in turn. INSPECTION Many factors apart from resistance can contribute to failures or apparent failures of treatments with anticoagulants, and many of them come under the heading of poor operating techniques. Those in the following list that are marked with an asterisk have, 1 Principal Scientific Officer, Pest Infestation Control Labo- ratory, Ministry of Agriculture, Fisheries and Food, Tol- worth, Surrey, England. at one time or another, resulted in reports of resis- tance: misidentification of the rodent species;* the use of too low a concentration of poison; insuffi- ciently thorough mixing of the poison with the bait ;* the use of a bait that is not sufficiently attractive to draw all the rodents away from their normal food ;* the placement oftoo little bait at each point;* and the laying of bait in too few places.* In addition there are other factors that may not be so readily control- lable by the operator, such as the continual movement of other rodents into the treated area or the consump- tion of the bait by other animals, such as birds.* The best way to check such factors is obviously field in- spection during the original treatment that leads to the reports. The assessment of the effects of most of these factors in a quick initial inspection is largely a matter of experience and common sense and will not be dealt with here. On the other hand it has been found worth while to ask investigators of suspected cases of resistance to fill in a standard report form so that the most relevant information is set out in a uniform manner. This form not only helps the investigator decide on further action on the basis of the existing information, but also enables the information about possible new cases of resistance to be passed rapidly to all interested parties who may later become in- volved in further investigation or in eradication campaigns. When investigating failures of poisons to control rodents, it is important to recognize that, if the fault lies with the rodent, it may be related either to the animal's behaviour or to its physiology. Some ro- dents, for example, having experienced the effect of sublethal doses of poison, may subsequently not eat 3002 -239- D. C. DRUMMOND & B. D. RENNLSON more of the same poison or of the bait with which it was mixed. This type of behavioural resistance is well documented for a number of acute poisons (see, for example, Drummond, 1970). It is unlikely, however, to play an important role with anticoagu- lants, where the long delay in the onset of illness severe enough to stop normal feeding results in the intake of a lethal dose. In fact, all well authenticated cases of anticoagulant resistance have so far been ofa physiological nature, the animal simply remaining more or less unaffected by the poison rather than reacting to and avoiding it. Thus the main positive field evidence of the presence of physiological resis- tance is the survival of animals that continue to feed on the poison bait. It is largely because adequate evidence of such feeding is so rarely available at the time of inspection that experimental treatments were introduced to see whether abnormally prolonged feeding could be demonstrated. EXPERIMENTAL TREATMENTS In order to show that unusually prolonged feeding was occurring it was first necessary to measure the feeding of susceptible animals during normal war- farin treatments. Fortunately suitable data were al- ready available from three experimental treatments carried out in 1954 against farmn rats (D. C. Drum- mond & M. Rowe, unpublished data)-that is, data collected from the kind of environment in which resistance is most commonly reported and at a time before it had been known to occur. The three experimental treatments were as similar as possible to normal treatments, the only difference being that all warfarin bait was laid where it was easily accessible to the investigators and the amounts of bait eaten by rats each day was measured to the nearest 5 g for each bait point. In each treatment the amount of bait eaten (Table 1) increased to a Table 1. Records of treatments with bait containing warfarin (0.025 % on farm 1 and 0.005 % on farms 2 and 3) against infestations of Rattus norvegicus in southern England, 1954 Farm I Farm 2 Farm 3 Day of Bait No. of points Bait No. of points Bait No. of points treatment eaten with take eaten with take eaten with take (9) Actual Adjusteda (9) Actual Adjusteda (g) Actual Adjusteda 1 2003 47 - 560 6 - 500 25 - 2 4 121 78 79 645 7 8 950 24 29 3 2819 57 - 405 8 - 1 045 23 - 4 2 089 47 62 170 8 8 915 21 24 5 934 31 - 55 4 - 435 14 - 6 672 1 3 - 35 2 - 400 6 - 7 415 13 35 15 2 6 200 5 15 8 413 10 - 30 3 - 55 3 - 9 182 6 12 25 1 3 80 3 4 10 126 4 - 10 1 - 90 4 - 11 168 4 6 30 1 2 55 3 5 12 20 2 - 25 3 - 35 1 - 1 3 0 0 - 60 2 - 35 2 - 1 4 - - 2 35 2 4 50 4 5 15 - - - 10 1 - 5 1 - 16 - - 0 0 0 1 0 0 1 17 - - - - - - - - - 18 - - - - - 0 - - 0 a The number of points at which ' take ' would have been recorded if the farms had been visited only on days 2, 4, 7, etc.-see text 240 DETECTION OF RODENT RESISTANCE TO ANTICOAGULANTS peak during the first 2 or 3 days, and then decreased rapidly until feeding ceased within 2-3 weeks. In addition, the number of points from which bait was eaten during each day of each treatment was highly correlated (P< 0.001) with the quantity of bait eaten. Thus it was possible to consider the progress of the treatments purely in terms of the numbers of points from which bait was eaten and to express the num- bers noted each day as proportions of the numbers recorded on the days with peak " takes ". After the peak, the proportions decreased in inverse linear re- lationship to the numbers of days on a logarithmic scale and the relationship could be defined by the equationy= 1.21-1.04x, where y is the proportion and x the logarithm of the day of treatment. To make the relationship more useful in practice it was recalculated using the data that would have been col- lected had each treatment started on a Monday and continued with visits only on subsequent Wednes- days, Fridays, and Mondays. With this schedule the peak " take " occurred as before on day 2, and the comparable relationship could be represented by the equation y = 1.41-1.1 x. The resulting regression line, with its 95% confidence limits, is now used as a sampling graph (Fig. 1) on which operators carry- ing out experimental treatments plot their results. To simplify matters still further the data plotted are the number of points with " takes " recorded on any particular visit divided by the number recorded on day 2 (not always in practice the day of peak take). If for two successive visits the proportion of points 1.0 \I \ CO 0 0.2 0 6) N Co 0) 4- ~ 020. z with " takes " lies above the upper limits it is regarded as abnormally high for too long and is taken to indi- cate warfarin resistance. At this stage trapping is begun to obtain surviving rats for laboratory tests. To standardize the treatments as much as possible operators are asked to use medium oatmeal bait con- taining 0.025% warfarin, to make sure that bait is laid at a sufficient number of points at the first visit, and not to lay bait at any more points subsequently. They are also asked to ensure that the number of complete " takes " of bait is kept to a minimum by always relaying, at any point where all the bait has been taken, twice as much bait as has been eaten previously. After the poison bait has been laid, visits to record " takes " must be confined to the second, fourth, seventh, ninth, etc., days. These are normally the Wednesdays, Fridays, and Mondays following the Monday on which the bait has been laid, but can of course equally well be succeeding Thursdays, Saturdays, and Tuesdays when the bait is laid ini- tially on a Tuesday. This type of standard treatment, as well as being used to identify resistant rat popu- lations, can also be used to compare the effects of other anticoagulants with those of warfarin in the field (B. D. Rennison, unpublished data). Although the sampling graph was originally based on only three early treatments, the examination of later results from many other farms, for which the graph gave no evidence for resistance, has not re- vealed any need to alter it. One of its greatest merits has been that it has quickly shown a number of sus- 7 9 11 Day of visit (bait first laid on day 0). Fig 1. Test treatment graph for anticoagulant resistance. 241 242 D. C. DRUMMOND & B. D. RENNISON pect rat populations to be nonresistant and has there- by eliminated the need to trap rats and conduct labo- ratory tests. It is perhaps unnecessary to add that this system of field screening has been made possible only because of the availability of a large number of officers stationed throughout the country who are trained to deal with vertebrate pest control problems. Unfortunately, however, because of the possibility that continuous feeding on the bait might be, for example, the result of the continuous immigration of new rats or even perhaps the result of a very high vitamin KI content in the normal food, the results of field treatments are rarely entirely satisfactory and final evidence of physiological resistance to anti- coagulants must generally be sought from laboratory tests. In some cases where evidence from an inspection seemed sufficiently conclusive and when laboratory testing facilities were available at the time, no inter- vening treatments were required. LABORATORY TESTS Laboratory tests to confirm suspected cases of re- sistance are similar to the experimental field treat- ments in that they aim to show whether the suspected resistant animals can consume warfarin bait over a longer period than can susceptible animals. The main difference is that the results of the laboratory tests, in which conditions are more strictly controlled, are rather more easily interpreted. The methods of testing rodents for anticoagulant resistance in the la- boratory suggested by Drummond (unpublished re- port to WHO, 1966) were subsequently modified and adopted by the WHO Expert Committee on Insecti- cides (1970). These methods should be tried out on a variety of rodent species throughout the world and the results should be reported to the World Health Organization. RtSUMt DETECllON DE LA RESISTANCE DES RONGEURS AUX ANTICOAGULANTS On decrit les methodes utilisees au Royaume-Uni pour d6celer la resistance aux anticoagulants chez le rat de Norvege (Rattus norvegicus). Le processus de d6tection comporte g6neralement trois stades. On procede d'abord a une inspection pr6liminaire des endroits oiu les anti- coagulants se revelent impuissants a assurer une destruc- tion satisfaisante des rongeurs. Ensuite, dans tous les cas oii une resistance physiologique peut etre suspect6e, on applique un traitement experimental par la warfarine, sous surveillance stricte. Enfin, lorsque l'echec de ce traitement semble confirmer qu'il s'agit bien d'une resis- tance physiologique, on capture un certain nombre de rats survivants et on les soumet a des epreuves de labo- ratoire. Pour l'inspection pr6liminaire, on fait surtout con- fiance l'exp6rience et au bon sens de l'enqu8teur. Neanmoins, pour etre certain que celui-ci a envisag6 toutes les causes possibles de l'6chec du traitement par les anti- coagulants, on lui demande de remplir une formule standard de rapport. Pour evaluer les resultats du traitement experimental, on se base sur le fait que dans une population de rats sen- sibles a la warfarine la quantit6 d'appat ingeree par les rongeurs, apres avoir atteint un maximum pendant les 2 ou 3 premiers jours, decroit rapidement selon une courbe caract6ristique. Lorsque la courbe obtenue a la suite du traitement experimental differe de cette courbe de reference, on en conclut que la population survivante renferme des individus resistants. L'epreuve de laboratoire consiste a administrer, pen- dant une periode donn&e, un appat renfermant un anticoa- gulant a des rats encagds individuellement et a comparer leur mortalit6 avec celle observ&e chez des animaux sensibles soumis au meme rdgime. Cette methode, l6gere- ment modifi6e, a 6t6 adoptee par le Comite OMS d'experts des Insecticides (1970). REFTERENCES Drummond, D. C. (1970) Symp. zool. Soc. Lond., 26, 351-367 WHO Expert Committee on Insecticides (1970) Wid Hlth Org. techn. Rep. Ser., No. 443