Does the selective serotonin reuptake inhibitor (SSRI) fluoxetine modify canine anxiety related behaviour?

[...]

Appraisal, application and reflection

The importance of the human-animal bond was accentuated in combating stay-at-home isolation in people due to the COVID-19 pandemic, resulting in increasing dog adoption rates and ownership. As people return to their pre-pandemic lives, such as returning to the work-place and external commitments, anxiety related conditions in dogs may become more prevalent: attributed to  overattachment to the owner, traumatic events while alone, and genetic predisposition (Flannigan & Dodman, 2001). Options for minimising problem behaviours in dogs are crucial to minimise pet relinquishment and/or euthanasia (Chutter et al., 2019). There is a long history of the use of various human anxiolytic drugs in companion animals, and fluoxetine is a widely prescribed SSRI in people and has been used in canine practice (Simpson et al., 2007; Pineda et al., 2018; Chutter et al. 2019; and Papich, 2020). It has also been suggested that anxiolytic therapy in conjunction with a behaviour modification program is the best strategy to minimise anxiety disorders in dogs (Landsberg, 2001; and Overall, 2013). An anxiolytic pharmacological intervention reportedly promotes the animal’s response to a behavioural modification program (Pineda et al., 2018).

Administration of tricyclic antidepressants (TCAs), SSRIs, and benzodiazepines are the more commonly used therapeutic drugs of choice to control anxiety, fear, and/or aggression with varying success in veterinary practice (Gruen & Sherman 2008). Fluoxetine, a SSRI registered for dogs is capable of increasing central nervous system (CNS) serotonin synapse concentrations with few adverse effects (Pineda et al., 2018). Fluoxetine delays the reuptake of serotonin resulting in an increase and prolongation of serotonin in those brain synapses where serotonin acts as a neurotransmitter (Papich, 2020).

This PICO question focused on dogs at least 8 months of age, when most dogs have passed the early and late socialisation period (Harvey et al., 2016). Dogs’ behavioural and social maturity are naturally developed between 12 and 24 months of age, therefore, the period before individuals reaching maturity is important in the development of their future behaviour (Harvey et al., 2016).

Irimajiri et al. (2009)  focused on a reduction in the severity of canine compulsive disorders (Irimajiri et al., 2009). Compulsive disorders are repetitive behaviours. In dogs this is characterised by the patient engaging in a repetitive, stereotyped behaviour for a significant amount of time (Crowell-Davis, 2006). Behaviours of compulsive disorders are derived from normal behaviours and may arise when the patient is exposed to an under-stimulating, such as prolonged separation anxiety, or from an overstimulating environment (Crowell-Davis, 2006).    

There were numerous similarities between both studies’ designs: they were both run out of multiple veterinary hospitals (multi-centre), placebo controlled, parallel-arm studies, used  board-certified veterinary behaviourists to confirm diagnosis of anxiety behaviours; used  veterinary behaviourists maintaining contact with owners every 2 weeks by telephone, the owners were blinded to the treatment, similar pretreatment conditions, dogs medicated with the same fluoxetine dosage for the same duration, and were reliant on owners reporting their perceptions of their dog’s level of anxiety to the chief investigators. Both studies were financially supported by Lilly Animal Health, Greenfield, Indiana, USA. Irimajiri et al. (2009) also analysed the owners’ daily diaries of their pets’ behaviour after 6 weeks of treatment, however, it is not reported if this also occurred in Landsberg et al. (2008). There were similarities in the outcomes, both studies reported some significant reduction in some of the target behaviours between the respective placebo treated dogs. Landsberg et al. (2008) also reported that fluoxetine treated dogs had a significantly faster mean rate of change in destructive / rearranging behaviour (P = 0.028), but not in the rate of change in the other three SA behaviours. Irimajiri et al. (2009) reported that the mean number and duration of compulsive episodes did not differ significantly between groups.

Possible explanations for the differences in outcomes reported between the two studies could be attributable to both studies being focused on different targeted behaviours. For example; Irimajiri et al. (2009) reported 52 of the 62 (84% ) subjects were purebreds with 19% of subjects being German Shepherd Dogs, while Landsberg et al. (2008) reported that > 60% of the dogs in both groups were purebreds, however no breed was over-represented. As Irimajiri et al. (2009) did ‘block’ subjects based on clinical diagnosis e.g. ‘German Shepherd Dogs with tail-chasing’ and distributed such blocks evenly into the treatment and placebo groups, it is difficult to comment whether the proportion of purebreds was a confounding factor in one or both studies.

Landsberg et al. (2008) showed a significant improvement specifically of destructive / rearranging behaviour within 3 weeks of treatment compared to those dogs treated with a placebo. It was noteworthy that the onset of improvement in overall and individual severity scores was rapid in the fluoxetine treatment group, occurring within a week after treatment was implemented. This is a significant finding as standard literature states that psychotropic drugs can take up to 3 to 5 weeks to take effect (Overall, 2013). Additionally, a longer treatment period may have been warranted as it has been suggested that at least 8 weeks of treatment  are required to observe a change behaviour when SSRIs are administer to human patients with obsessive compulsive disease (OCD) (Liebowitz et al., 2002).

Landsberg et al. (2008) also reported that there was a reduction in some of the behaviours such as destructive / rearranging behaviour from baseline in the placebo dogs. Landsberg et al. (2008) suggests that the high placebo response study may have been a result of investigators feeling ethically obligated to provide some incompliant guidance to owners on their dog’s behaviour when questioned.

There are numerous limitations inherent within these studies, that include:

Possible sources of selection bias.

It is possible that pets with other concurrent behaviour problems might have been inadvertently enrolled in either study (Landsberg et al., 2008).

Possible sources of information bias. 

Misclassification bias: reliance of owner information and inaccurate observations may have led to inconsistent and varied treatment success in both studies. For example Irimajiri et al. (2009) reported that some owners did not complete all entries on a daily basis and some participants had multiple family members completing diary entries (Irimajiri et al., 2009).

There is no indication in either study whether the baseline data between the treatment and control groups were comparable prior to commencing either the fluoxetine or placebo dosing.

Owners were asked to respond to a previously validated (Hewson et al., 1998) 5-point Likert scale to identify changes in severity of compulsive disorders (Irimajiri et al., 2009) whereas Landsberg et al. (2008) used a four point scale for the owners to assess the overall level of separation anxiety during the preceding week (0, absent; 1, mild; 2, moderate; 3, severe). This inconsistency in Likert score items may have had some effect on determining the outcomes.

While both studies use the word ‘frequency’, they determined frequency differently. Landsberg et al. (2008) used relative frequency of each specific SA behaviour calculated for each dog by dividing the specific SA behaviour by the total number of owner departures, while Irimajiri et al. (2009) used owner observation diary data summarised for each 14 day period as the mean number of episodes per day.

It is possible across the 42 treatment days that there was some owner ‘drift’ in their perception of severity of targeted behaviours. No ‘behavioural anchors’ are mentioned as part of the ‘severity’ rating in either study which may have influenced the reliability and validity of each ‘severity rating’.

Possible sources of confounding bias.

Effect of home stimuli such as frequency and duration of interactions with humans (such as owners, family members etc.) or the frequency / and duration of daily exercise on outcomes, questions the accuracy of comparative treatment effects and were not addressed in either study.

In both studies (Landsberg et al. 2008; and Irimajiri et al. 2009) there was no mention the time of day the fluoxetine was administered, nor whether it was administered with or without food. However, the Reconcile^Ò chewable tablets label states that the medication can be administered with, or without food.  The most common adverse experiences associated with fluoxetine medication were lethargy (18.2–39%), anorexia (24.2–23%) as well as other adverse experiences that affected only a few subjects (Landsberg et al., 2008; and Irimajiri et al., 2009). Both studies reported in most cases signs resolved within a few weeks or with a dose reduction. 

In conclusion, the SSRI fluoxetine is a palatable and well-tolerated psychotherapeutic medicine for dogs at recommended doses. Based on the outcomes of the two studies provided in this Knowledge Summary there is moderate evidence that fluoxetine reduces the severity of anxiety related conditions. However, both studies also recommend that behavioural and environmental modifications are important adjuncts to pharmacologic treatment for dogs with separation anxiety (Landsberg et al., 2008) and / or compulsive disorders (Irimajiri et al., 2009).

[…]