Statement of the Problem

A dog was evaluated for a 5-year history of extensive grabbing and tugging at people's clothing; the behavior had recently intensified, and the dog had begun mounting household objects.

Signalment

The patient was a 7-year-old 20.1 kg (44.4-lb) neutered male English Springer Spaniel.

History

The patient resided with 2 owners and 1 other dog (a 5-year-old neutered male Labrador Retriever) in an apartment with a garden. The dog had been adopted at 20 months of age and, very shortly after adoption, began to take or steal various objects and guard the items with aggressive behavior (eg, staring, growling, and baring teeth).

With the help of a dog trainer, the owners started to limit opportunities for the dog to steal objects in the house (described as dog-proofing the house). They ignored the dog for stealing objects when it was safe to do so and used treats to divert its attention away from such items; they could then secure the dog behind a barrier and safely remove the stolen item without eliciting further aggressive behavior. Additional advice included increasing the dog's activities, so the owners regularly brought the dog to agility classes, taught it to perform tricks in response to verbal cues, and engaged it in scent-tracking work. In the house, food-dispensing toys and interactive dog games were used to provide mental stimulation and to help reduce the amount of time the dog was free to search for objects to steal.

Although the unwanted behaviors of stealing and aggressively guarding items were markedly reduced after these changes were implemented, the owners reported the dog became more active in the evenings and seemed unable to settle down. Instead of stealing items, the dog began to grab and pull at loose-hanging clothing worn by the owners and by friends both inside and outside the home environment. The owners began to separate the dog from people by placing it in an enclosed conservatory for prolonged periods and by taking it out into public less frequently. With additional professional advice from a nonveterinarian animal behavior counselor, the owners attempted to limit opportunities for this new undesirable behavior by implementing a command-response-reward program and diverting attention from grabbing clothing by using food treats to reward alternative activities (teaching a more appropriate behavior [such as lying down on a mat] that was incompatible with the behavior of concern).

The referring veterinarian diagnosed a compulsive behavior disorder, and prescribed a 4-month course of amitriptyline hydrochloride (1 mg/kg [0.45 mg/lb], PO, q 12 h). The medication had no positive effect on the observed behaviors and was discontinued. Despite the described efforts, the clothes grabbing intensified; the dog also started inappropriate mounting of various objects, and referral to the authors' veterinary behavior clinic was made.

During the consultation, the owners reported that they had not been able to have cushions or any small to medium size movable or dangling soft textured items in the house for the past 20 months. They wore closely fitted clothing to minimize the dog's opportunity to grab and pull, but despite this, multiple daily episodes of the dog grabbing and holding onto clothing (jeans buttons, ends of sleeves, or pant legs) occurred. The behavior intensified at times when both owners were sitting on the sofa together or when a conversation was being held. The same behavior was observed with friends visiting the house and when the owners visited others and brought the dog with them; for example, the dog would jump up at people to take hold of a scarf. The owners, their family, and their friends were able to interrupt the behavior by gently pushing the dog away or cueing it to perform an alternative behavior.

Physical Examination Findings and Laboratory Results

During the consultation, the dog initially investigated the room and engaged playfully with toys that were provided. Thereafter, the dog spent most of its time looking for any loose parts of clothing, which it would gently grab onto and attempt to keep hold of without pulling. This behavior could be interrupted by moving away, standing up, pushing a flat hand in between the clothing and the dog's mouth, or providing a verbal cue for an alternative behavior. The behavior was shown toward everyone in the room and continued for the duration of the consultation. Results of a physical examination were unremarkable except that signs indicative of mild pain were elicited on manipulation of both elbow joints and palpation of the lumbar vertebral column, areas in which the previous medical history indicated chronic osteoarthritis. Results of hematologic analysis, including serum biochemical tests, performed by the referring veterinarian after the consultation were within the respective physiologic reference ranges. A parasitological examination (fecal flotation and testing for Giardia spp) by the referring veterinarian also yielded negative results.

Diagnosis

The most common causes for prolonged periods of repetitive, excessive, and persistent behaviors include stereotypies and compulsive disorders.¹ The performance of repetitive behaviors can be part of a normal species-specific repertoire; they can be environmentally induced, conditioned, or caused by internal factors including medical disorders.^1,2 On the basis of the physical examination and hematologic findings, metabolic medical causes for the patient's behaviors were considered unlikely. Epileptic disorders such as psychomotor epilepsy were excluded because the dog was easily distracted and no apparent postictal phase was described by the owners.² No medical condition associated with abnormal oral behavior could be detected, and this was also considered unlikely, as the dog did not show signs of licking, chewing, sucking, pica, or polyphagia.³ In light of these considerations, further evaluation for potential underlying gastrointestinal diseases (ultrasonography, endoscopy, and collection of intestinal biopsy samples for histologic examination) was declined by the owner.

Although stereotypic behavior is commonly used as a descriptive term to refer to an invariant repeated motor pattern, a compulsive disorder involves repetitive behaviors being shown as a result of fixation toward a specific goal and interfering with normal functioning of the individual. Differentiating between these conditions can be complex, and this topic has been reviewed elsewhere.⁴ Compulsive disorders can initially arise in situations of frustration, acute conflict, or both.^2,5 Redirected behavior or displacement behavior due to acute conflict was considered unlikely in this case on the basis of the history because there was no pattern of external triggers initiating the behavior. In addition, a very specific context (owner presence) was identified, and the dog could easily be distracted from performing the behavior. Therefore, a compulsive disorder was considered unlikely. Although the lack of response to treatment with amitriptyline prompted referral for assessment by a veterinary behaviorist, to the authors' knowledge, there is no medication reported to be 100% efficacious in treating compulsive disorders in dogs. In 1 retrospective study,⁶ the observed success rate for treatment with amitriptyline was significantly lower than that for treatment with clomipramine in dogs with compulsive disorders. Thus, a lack of response to amitriptyline alone cannot be used to rule out a compulsive disorder.

Another differential diagnosis was redirection of an underlying anxiety.⁴ However, it was considered unlikely because none of the physical and behavioral signs such as mydriasis, body tension, hypervigilance, increased startle response, lowered body position, or a tucked tail were observed during the consultation or reported by the owners.

A tentative diagnosis of attention-seeking behavior or conditioned unwanted behavior and inadvertent reinforcement of an inappropriate behavior chain (having learned to perform a number of behaviors in a particular sequence) was made. An inappropriately applied command-response-reward program, together with distraction techniques, was deemed to have motivated the dog to continue performing the unwanted behaviors, which were maintained by attention from the owners. The diagnosis was confirmed by the treatment response.

Treatment

The owners were asked to extinguish the unwanted behavior by consistently withholding all consequences that previously reinforced it.⁷ The possibility of an extinction burst was also discussed. Increased variability of and an abrupt increase in the previously reinforced behavior are key features of extinction^7,8; however, with appropriate management, these changes are typically temporary. Increasing levels of frustration can also cause occurrences of aggressive behavior.^8,9 Because of these concerns, it was decided that the dog's response to the extinction procedure should be observed during the consultation to assess the safety of this approach.

The owners were asked not to interact with the dog, regardless of its behavior. After being ignored while grabbing at clothing for approximately 5 minutes, the dog began to pull cushions off the sofa on which the owners were sitting. After an additional 5 minutes, this was followed by mounting of the cushions, at which point all present left the room but continued to observe the dog through one-way glass. The behavior stopped immediately, but as soon as the consultant reentered the room and continued to ignore the dog, it started to pull the remaining cushions off the sofa and then repeated the mounting behavior. No self-injurious behavior or signs of aggression toward people were observed. After approximately 10 minutes, the dog lay down, and the owners returned to the consultation room.

After reentering the room, the owners were asked to continuously interrupt the unwanted behavior by leaving the room as soon as the dog started to grab any clothing or other item. They were instructed to continue this process at home. To prevent new behavior chains from forming, the use of distraction techniques and rewards for alternative behavior after unwanted behavior occurred was discontinued. The owners were advised to not interact with the dog for a minimum of 15 minutes after each display of unwanted behavior. The 15-minute delay was chosen to allow sufficient time for the dog to calm down after attention was denied (on the basis of observations during the consultation), and the owners were not to look at, talk to, touch, or cue the dog for any alternative behavior during that period. Attention was to be given only when the dog appeared calm, relaxed, and settled, to reinforce desired and more appropriate behavior.

In addition, the owners were asked to continue with their management of dog-proofing the house and to preemptively occupy the dog in situations where grabbing of clothing was reported to be most intense (eg, when both owners were present and while people were having conversations). This included more strategic use of food-dispensing toys and introduction of new food-dispensing toys in rotation. The dog was to be kept separate from unfamiliar visitors, and familiar visitors were to be asked to follow the same pattern of actions as the owners if the unwanted behaviors occurred in their presence.

Follow-up

In a telephone conversation 2 weeks after the consultation, the owners reported marked improvement. The dog would attempt to grab clothing 2 to 3 times daily for a cumulative period of no longer than 15 minutes, after which it would settle and lie down without showing signs of frustration. The owners no longer needed to place the dog in the conservatory during the evening (when the behaviors had previously been intensified). They were able to have conversations while the dog was engaged with food-dispensing toys. Familiar visitors had been invited for structured sessions, and the owners reported that there were no problems with visitors following their instructions.

Eight weeks later, further improvements were reported; the owners were able to wear clothing with buttons and walk the dog in public places again. On rare occasions in the house, the owners had to stand up to interrupt an attempt to grab at clothing, but they no longer needed to leave the room to stop the behavior. There was no mounting of objects reported during the follow-up period.

However, the owners reported the occurrence of short periods of barking (1 to 2 times weekly), which was interpreted as resurgence (the reappearance of a previously extinguished behavior when a previously reinforced alternative behavior is subsequently placed on extinction).¹⁰ The owners were advised to continue ignoring this unwanted behavior. Spontaneous recovery (ie, reappearance of the learned association placed on extinction, in this case mounting behavior and tugging at clothing) was not reported during the 3 months of follow-up.

Discussion

A command-response-reward program is meant to create structure and predictability for a dog. The foundation of this program is for the dog to respond to a command to receive a reward. It is used to help redirect a dog from unwanted behavior and focus it on performing more appropriate behaviors.² However, following learning theory, chaining is a potential risk of this approach. Performing a number of behaviors in a particular sequence forms a behavior chain. Each link in the chain is reinforced in part by the opportunity to perform the next step in the chain, with only the last act in the chain rewarded with a reinforcer.⁷ This means that dogs have the potential to learn that unwanted behavior elicits a request to perform a different behavior, which is ultimately followed by a reward in the form of attention, food treats, or both, leading to the formation of multiple behavior chains all elicited by the initial display of unwanted behavior. For patients with a high tendency toward attention-seeking behavior, the risk of development of behavior chains when desirable items are introduced as distractions or as part of a command-response-reward program should be carefully assessed and monitored. Any deterioration or inappropriate change of the behavior during treatment of such patients should be reevaluated for the development of an unwanted behavior chain and managed accordingly.

Extinction procedures involve withholding all consequences that reinforce a behavior,⁷ as was done for the dog of this report. Although these can be valuable tools to help reduce unwanted behavior, the procedures need to be carefully assessed in terms of risk to the patient and others. The immediate response to extinction is often an abrupt increase of the unwanted behavior or so-called extinction burst. Increased variability of behavior and additional behaviors that might elicit the desired consequence are also features of extinction. As mentioned, increased signs of frustration can be observed, potentially leading to aggressive behavior (direct and redirected).⁹ In addition, the frustration that might be caused by implementing extinction must be carefully assessed as a welfare consideration for the dog. For these reasons, a risk evaluation must be performed, and owners must be carefully informed about risk mitigation and monitoring before starting an extinction program.

This case illustrated that, when an unwanted behavior arising as a learned response for reward is correctly identified, correct application of learning theory (while ensuring the dog's welfare is maintained and potential risks are mitigated) can result in rapid resolution of the problematic behavior and its replacement with appropriate dog-owner interactions. In addition, the simple plan was easily implemented with consistent actions on the part of the owners, and the results observed over the following 3 months supported that such consistency can help to maintain this positive change in the long term.