Behavioral Problems Associated with Canine Aggression

Treatment of Behavioral Problems Associated with Canine Aggression

An attempt has been made to specify definitional criteria for behavioral diagnoses. The advantage to these is that they do not rely on nonspecific signs, so conditions that share signs are not confused, and conditions that show an atypical form are not ruled out because a nonspecific sign appears discordant. Clusters of nonspecific signs within a diagnosis or those shared in comorbid diagnoses may help to identify treatment and pathological groups. While this system is frequently used, it is not universal. It may be helpful to view these diagnostic categories as guidelines that will continue to be redefined, given that behavioral medicine is a quickly growing and evolving field.

Some of these problems, in feline-specific form, also occur in cats; these include fear aggression, idiopathic aggression, interanimal aggression, pain aggression, play aggression, predatory aggression, redirected aggression, and territorial aggression.

Dominance/impulse control aggression has the following necessary condition: abnormal, inappropriate, out-of-context aggression (threat, challenge, or attack) consistently exhibited by dogs toward people under any circumstance involving passive or active control of the dog’s behavior or the dog’s access to the behavior. The following condition is sufficient (eg, if it is met there is little doubt that the diagnosis is valid): intensification of any aggressive response from the dog upon any passive or active correction or interruption of the dog’s behavior or the dog’s access to the behavior. The key is assessment of the dog’s need to control for the sake of maintaining control. The behaviors exhibited by dogs with this diagnosis are actually a rule structure that allows the dogs to learn how to act in situations that are potentially unclear, uncertain, and anxiety-provoking to them.

This definition of dominance/impulse control aggression is discrete and does not couple the challenge to food (food-related aggression), toys (possessive aggression), or space (territorial aggression). These aggressions can all be correlates of dominance/impulse control aggression and, when associated with it, may be indicative of a more severe situation. Control and access are key—most of the problems with diagnoses arise from human misunderstanding of canine social systems, canine signaling, and canine anxieties associated with endogenous uncertainty about contextually appropriate responses. Diagnosis cannot be made on the basis of a single event. The behavior, once it begins, will become more visible and consistent, but data on early signs, patterns of change with experience, and changes in intensity are lacking.

The necessary and sufficient conditions are different from the common descriptions of dominance/impulse control aggression that specify that the dog will often react to being pushed on, to being corrected with a leash, or to being pushed from a sofa by a person. The number of situations in which the dog reacts inappropriately, or the intensity with which he or she reacts, do not affect the necessary and sufficient conditions, although these factors may affect the ability to treat the condition, the prognosis, and the risk to people. These dogs are not fearful, and they exhibit none of the behavioral or physiologic signs associated with fear. All canine aggressions are likely based to some degree in anxiety, and the uncertainty associated with anxiety is not equivalent to fear.

Fear aggression has the following necessary condition: aggression that consistently occurs concomitant with behavioral and physiologic signs of fear as identified by withdrawal, passive, and avoidance behaviors associated with the sympathetic nervous system. The following condition is sufficient: as above and the aggression is accompanied by urination or defecation, or when the aggression is only active (beyond just posturing) when the target of the aggression is not interacting with the subject.

The actual behaviors associated with fear, fear aggression, and any other aggression that is primarily driven by anxiety are poorly qualified and quantified. However, in contrast to dominance/impulse control aggression, the dog relinquishes control and withdraws. Only when the dog can no longer avoid or withdraw does frank aggression occur. Also, unlike dominance/impulse control aggression, these dogs never deliberately provoke a situation or initially and voluntarily contribute to its escalation.

In extreme cases, the sufficient conditions are clear; in less clear situations, which could be due to uncertainty on the animal’s part, caution is urged in ruling out all other aggressions. The most likely diagnosis is the one that is most consistent with all signs and criteria. Fear aggression does not have to occur consistently, although identification of the fearful stimuli will permit assessment of the extent to which the behaviors are consistent and pose a predictable risk. Fear aggression is also seen in cats, in which it is one of the most common types of aggression, both toward other cats and people.

Food-related aggression has the following necessary condition: consistent aggression that is exhibited in the presence, and only in the presence, of pet food, bones, rawhides, biscuits, blood, or human food in the absence of abuse or starvation. The following condition is sufficient: as above with the aggression occurring only in the presence of a range or subset of the items listed. This is a restrictive and specific diagnosis. Number or range of items involved, while possibly reflecting danger and risk, do not affect the diagnosis. It is possible that aggressions stimulated by different classes of food indicate varying neurochemical modalities, and these differences may represent subclasses of this diagnosis. This diagnosis serves to highlight that food is not a possession, but rather something very different from a possession. This difference has been noted in neuroanatomic and neurochemical studies of aggression and is probably real. Certainly, a good evolutionary case could be made for aggression related to food being potentially important. While this aggression may be associated with dominance/impulse control aggression (ie, an animal can have both diagnoses, but development of food-related aggression usually precedes development of dominance aggression), it is absolutely, categorically different, based on the necessary and sufficient criteria listed for each. Food-related aggression can be a singular diagnosis, unrelated to any dominance aggression. Furthermore, if an event related to impulse control aggression only incidentally involved food, or food was the vehicle for the aggression only once (ie, there is no pattern of other food-related responses), and the owner has been able to take or interfere with the food item at other times, the issue is one of control (ie, impulse control aggression), not food.

Idiopathic aggression has the following necessary and sufficient condition: aggression that occurs in an unpredictable, toggle-switch manner in contexts not associated with stimuli noted for any other behavioral aggression diagnosis or with any underlying causal physical or physiologic condition. This diagnosis must be distinguished from any neurologic condition. Intensive characterization of attendant behaviors is necessary to differentiate this from the most common condition with which it is confused—undiagnosed or subtle dominance/impulse control aggression. Explosive, unpredictable aggression can be associated with “rage,” a term that should not be used because of the inability to adequately define the analogous emotional conditions in pets that are experienced and described by people. Unpredictability is a function of the quality of observational skills and knowledge. Although idiopathic aggression may be considered a common problem by clients, once a detailed history is collected it becomes clear that this is a rare diagnosis. This form of aggression is also reported to occur in cats. This diagnosis may simply be a misidentified variant of impulse control aggression.

Interanimal or interdog aggression has the following necessary condition: consistent, volitional, proactive aggression that is not contextual given the social signals, threat circumstances, or response received. The following condition is sufficient: as above in the absence of any signal or interaction from the animal that is attacked. It is emphasized that, at some level, the behaviors involved with aggression are normal behaviors. This diagnostic category, while usually associated with changes in social hierarchy that are often related to the development of social maturity in one of the involved animals, does not depend on either hierarchy or social maturity; it depends on the contextual response. This subtle but important distinction supports the contention that social shifts and occasional threats can be normal. A change in behavior is not necessary, although it may be usual, because if this is truly a diagnosis of an abnormal behavior, some animals will respond with aggression regardless of circumstances. More specific categories would be interdog and intercat aggression associated with hierarchical disputes. Interanimal aggression is also seen in cats.

Anthropomorphic terms like sibling rivalry should be avoided. Not only are they likely to be incorrect, they also obscure the difference between truly pathologic aggression and what may be a normal canine redress of mild social conflicts. Caution in generalized and anthropocentric views and terms is urged.

Maternal aggression has the following necessary condition: consistent aggression (threat, challenge, or contest) directed toward puppies in the absence of pain, challenges, or threats to the mother by the puppies. The following condition is sufficient: unprovoked, age-inappropriate attacks on puppies by the mother. When maternal aggression is profound, it is extremely easy to recognize, even though puppies are not necessarily injured or killed. The extent to which discrete aggressive behaviors can be a component of normal maternal behavior has not been well quantified.

Inherited forms of maternal aggression are seen in both cats and dogs. Treatment is important when the condition is familial. The ethics of continued breeding in such circumstances should be considered.

Pain aggression has the following necessary condition: consistent aggressive behavior, in excess of that required to indicate concern and to effect restraint, demonstrated only in a context known or potentially associated with pain but that may not be painful itself. The following condition is sufficient: as above in the absence of any behavioral and physiologic signs of fear as identified by withdrawal, passive, and avoidance behaviors associated with the sympathetic nervous system. Evaluation of pain in animals is very difficult and even more subjective than in humans (see pain management, Pain Management: Introduction). This is a diagnosis of degree and correlation: conditions that are known to cause pain (eg, fractured legs) could render the animal resistant to manipulation. Domestic animals do not have opposable thumbs and so may use their mouths to grasp and restrain. For this diagnosis to be made, fear must not be primary (although anticipation of pain and the attendant anxiety may be involved), and the behaviors must be in excess of those required to indicate the animal’s concern. Pain aggression is also seen in cats. Newer protocols for controlling perceived or anticipated pain should decrease the incidence of this condition.

Play aggression has the following necessary condition: consistent aggression that occurs in contexts in which play behaviors (eg, play bows, yips, shoulder blocks [dogs]; chases, charges, pounces [cats]) would normally occur. The following condition is sufficient: out-of-context, consistent aggression in circumstances in which play is relevant, or in-context aggression consistent with the solicitation of play but that involves actions that would discourage play (eg, biting, pain). The normal, accepted, or in-context range of social play behaviors are relatively well defined when compared with abnormal, unacceptable, or out-of-context behaviors. The difficulty is to distinguish rough play (learned in interactions from other animals or people) from truly abnormal behavior. Analysis of discrete behaviors should also distinguish this from attention-seeking behavior. Play aggression is also seen in cats.

In contrast to previously held beliefs, new research has shown that energetic play by humans with dogs (eg, tug-of-war) does not necessarily produce play aggressions. Humans who play energetically and forcefully with their dogs do so best by understanding and using signals that are like those dogs use themselves. The same pattern is likely to be true for cats, especially if appropriate inhibition of damaging play is discouraged through quick signaling and withdrawal of attention.

Possessive aggression has the following necessary condition: aggression that is consistently directed toward another individual that approaches or attempts to obtain a nonfood object or toy that the aggressor possesses or to which the aggressor controls access. The following condition is sufficient: as above, and, in the absence of the object associated with the contentious behavior, aggression does not occur. This diagnosis includes only nonfood items, although there may be some overlap with food-related aggression (eg, rawhide chew toys). The stimuli and neurochemical changes associated with aggression toward objects and toward food are likely very different. While this aggression may be correlated with the occurrence of canine dominance/impulse control aggression ( Behavioral Problems Associated with Canine Aggression) or feline status-related aggression ( Behavioral Problems Associated with Feline Aggression), the latter behaviors are about control of activity or access—not about control of objects—and the diagnosis of impulse control or status-related aggression should not be made on the basis of a response to an object. For a diagnosis of possessive aggression to be made, the response to the object must be consistent, restrictive, and repeatable. If the animal also fulfills the criteria for dominance or status-related aggression, those diagnoses should be made in addition to, not instead of, the diagnosis of possessive aggression.

Predatory aggression has the following necessary condition: quiet aggression, or behaviors congruent with subsequent predatory behavior (staring, salivating, stalking, body lowering, tail twitching, etc) consistently exhibited in either circumstances associated with predation or toward victims such as human infants or young or ill animals; death is not a necessary sequela, nor is ingestion should death ensue. The following condition is sufficient: quiet, unheralded attacks, generally involving at least one fierce bite and shake, that include staring, salivating, stalking, body lowering, and tail twitching, consistently exhibited toward species-contextual prey items (eg, cats and birds), or toward individuals that exhibit uncoordinated movements and sudden sleep and wake cycles (eg, human infants, young or ill animals, geriatric people); death is not a necessary sequela, nor is ingestion should death ensue. Discrete analyses of the behaviors involved should elucidate different forms of this behavior and the role that the behavior of the victims plays in determining the form that the aggression will take. The latter is important because predatory aggression can also be used to describe aggression to joggers and bicyclists. In the latter case, territorial aggression must be considered, but when predatory aggression involves sentient adult people, it is likely to be categorically different from that described above. Normal, species-typical predatory behavior in the appropriate context is not to be confused with predatory aggression. Predatory aggression also occurs in cats.

Protective aggression has the following necessary condition: aggression that is consistently demonstrated when an individual or class of individuals is approached by a third party, in the absence of an actual, contextual threat from that third party. The following condition is sufficient: as above when the aggression intensifies with decreasing distance, or with vocal or physical cues that could indicate excitement or threat, despite attempts at intervention, correction, or the desire to interact on the part of the individual being “protected.” Protective and territorial aggression are often lumped, making it unlikely that the impetus will exist to learn if they are behaviorally discrete. It is important to acknowledge that some degree of in-context, innate protectiveness is desired in most pet dogs. Diagnosis of protective aggression must be made only after the relevance of the context in which it occurs has been evaluated. Many dogs appear to react aggressively when confined to cars. If this is the only circumstance in which the dog is aggressive, the behavior should be considered a variant of normal.

Redirected aggression has the following necessary condition: aggression that is consistently directed toward a third party when the dog is thwarted or interrupted from exhibiting aggressive behaviors to its primary target. The following condition is sufficient: aggression that is instantly and consistently directed toward a third party when the dog is thwarted or interrupted from exhibiting aggressive behaviors to its primary target; the aggression is not accidental, and the dog will actively pursue the third party, particularly if they were associated directly with the interruption of the animal’s previous behaviors. Redirected aggression must be differentiated from displacement activity in which both target and behavior have been altered as a result of a thwarted, interrupted, or corrected behavior. This diagnosis is specific and unassailable when identified by discrete behavioral descriptions. The most common diagnostic error is calling a behavior redirected aggression when the aggression was actually accidental and the result of direct intervention in the absence of sufficient time for the aggressors to stop their activity (eg, reaching between 2 fighting animals and being bitten because one animal was already in the process of biting the other and could not stop). Redirected aggression is also seen in cats.

Territorial aggression has the following necessary condition: aggression that is consistently demonstrated in the vicinity of a mobile (eg, car) or stationary (eg, yard) circumscribed area when that area is approached by another individual in the absence of an actual, contextual threat from that individual. The following condition is sufficient: as above in which the aggression intensifies with decreasing distance despite attempts at intervention, correction, or the desire to interact on the part of the approaching individual. If the only context in which the dog protects a territory is when confined in a vehicle, this can be considered a variant of normal behavior. Territorial aggression is also seen in cats, although new research has shown that most aggression exhibited by cats to other cats is about social relationships and not space.

Other Causes of Aggression:

In addition to sensory system impairment and changes in an animal’s physical and mental ability to negotiate their environment, sudden aggression can also be the result of infection, toxicity, or an atypical drug response. Sudden atypical aggression has been reported for dogs treated with acepromazine, and for dogs and cats treated with diazepam. While the benzodiazepine events are truly paradoxical, those associated with dissociative neuroleptics may not be. Because of their ability to alter perception while rendering the animal relatively immobile, such agents may worsen anxiety-based conditions, not improve them. This is one rationale for advising against the use of acepromazine for the treatment of noise phobias.

Treatment of Behavioral Problems Associated with Canine Aggression:

The treatment of aggression can be complex and may best be performed by a specialist. The first step in treating any aggression is to obtain an accurate diagnosis. Then, any provocative circumstances must be avoided—repetition of aggression teaches the dog to be better at executing it and reinforces the association between context and behavior. Most behavior modification focuses on counterconditioning and desensitization using food treats or rewards. Owners must understand the difference between treats and bribes; the latter will guarantee treatment failure. The dog should be taught to defer to the owner for everything it wants. The goal is not to make the dog “submissive,” but rather to teach the dog to attend to the owner for cues as to the appropriateness of its behavior. Dogs defer to other dogs by sitting or lying down and waiting for cues that tell them when they can proceed with their next behavior. Sitting or lying down acts as a “stop” command and allows owners to regain control of the situation. These dogs must earn all of their attention—even maintenance attention (eg, feeding, walking). Head collars (head halters, eg, Gentle Leader^® or Halti^®) can be a great help in treating or preventing problem canine behaviors involving aggression and can render any aggressive dog safer. Almost without exception, physical punishment, including the use of prong collars and electric shock collars, alpha rolls, and dominance downs can make an already aggressive dog worse. Owners should be discouraged from using these techniques, especially in the absence of supervision. Antidepressants (eg, amitriptyline, clomipramine, fluoxetine) have been helpful in treating anxiety associated with the aggression and in facilitating behavior modification. Currently, all behavioral treatment with psychotropic agents is extra-label, requires informed consent, and should be performed only after a thorough physical examination and laboratory evaluation (CBC and serum biochemistry profile).