11 May 2022
Mark Lowrie
Job Title
Paroxysmal events are notoriously difficult to characterise, as they are rarely witnessed by anyone other than the owner and in-between times impacted dogs are completely normal.
Add to this the fact many of these transient events are caused by a chemical interference within the CNS that upsets the balance between neuronal excitation and inhibition – leaving no trace of their activity – it becomes clear that definitive diagnosis is extremely difficult. One such event would be those reported to occur during intense exercise, often in extremes of heat.
Unless you are feeling particularly active and wish to exercise the dog to the point of precipitating an episode, it is advisable to consider video footage from the owner to avoid ambiguous descriptions and to actually observe the problem for which the owners present the dog.
This article describes the syndrome of exercise-induced collapse in dogs, including when to suspect it, what other conditions may mimic it, and how to diagnose and manage the disorder.
Young Labrador retrievers aged between seven months old to two years old may present with a syndrome of exercise intolerance and collapse (EIC). All colours of Labrador retriever and both sexes may be affected.
Those used in field trials seem particularly prone, but this is probably more related to their activity levels than any particular predisposition.
EIC is characterised by episodic weakness of the limbs followed by ataxia and collapse occurring 5 to 20 minutes after intense exercise. Affected dogs are normal at rest and during normal activity, with well-defined musculature and a keenness to exercise.
As with all conditions, a wide spectrum of severity of clinical signs exists. During the early stages of an episode (that is, within the first five minutes), dogs may develop a rocking or forced gait, carpal hyperextension, a wide-based pelvic limb stance and excessive pelvic limb abduction when turning (Steiss at al, 2004). This may then quickly progress to obvious ataxia of the pelvic limbs, with severe panting and potentially distress.
Collapse is not inevitable in every dog, but will occur in many cases. Most dogs remain conscious and alert during the episodes, although 10% to 25% appear disorientated or have some mentation changes. Death is reported in a small percentage of dogs during the exercise period.
Core body temperature has found to be raised to above 41.5°C (greater than 106°F) in many dogs, although even healthy Labrador retrievers can have similar elevations in temperature without collapsing, meaning this is not a diagnostic sign (Taylor at al, 2008). Recovery can take around 10 to 20 minutes, by which time they are usually back to normal.
Recent genetic advancement has identified the causative mutation for this condition in the Labrador retriever (Paterson at al, 2008). The mutation is found in the dynamin 1 gene (DNM1), and as well as the Labrador retriever, it has also been found in Chesapeake Bay retrievers, curly-coated retrievers, Welsh Pembroke corgis, Boykin spaniels and mixed-breed dogs thought to be Labrador retriever-crosses (Minor et al, 2011).
This mutation has autosomal recessive inheritance, with clinical signs being observed in dogs homozygous for the mutation. It is estimated that 30% to 40% of purebred Labrador retrievers are carriers of this mutation and these individuals are all asymptomatic for EIC.
It is of note that the DNM1 mutation has not been identified in golden retrievers, flat-coated retrievers or Nova Scotia duck tolling retrievers, or 15 other non-retrieving breeds tested.
Dynamin 1 is the protein responsible for an essential molecule in synaptic vesicle endocytosis in the brain and spinal cord during high-level neuronal activity. An interesting observation is that the DNM1 mutation associated with EIC has its most profound effect on DNM1 function when body temperature is elevated, as normally occurs with exercise (Furrow et al, 2013).
This temperature-dependent reversible loss of motor function explains why clinical signs are only seen in affected dogs that undergo strenuous exercise (Taylor et al, 2009).
EIC has historically been a clinical diagnosis of exclusion, ruling out all other causes for exercise intolerance in a young adult, athletic dog. General clinical and neurological examinations of an affected dog are normal.
If a dog is examined during a period of collapse, the hindlimbs may be flaccid, with loss of the patellar reflex. In contrast, the forelimbs may have increased extensor tone.
Assessment of conscious proprioception of all four legs as assessed by response to knuckling and hopping is invariably normal during an episode despite obvious ataxia. The body temperature is often severely elevated and a severe alkalosis is present on blood gas analysis immediately following exercise, although these changes are non-specific.
All remaining clinicopathological and electrodiagnostic tests, and histological examination of muscle biopsies, are normal. Therefore, diagnosis is by inspection, with typically clinical signs in an otherwise healthy dog increasing suspicion of the condition. A commercially available DNA test for the DNM1 gene mutation is now available for diagnosing EIC-affected Labrador retrievers (Paterson et al, 2008) and related retrievers (Minor et al, 2011).
Demonstration that a dog has two copies of the DNM1 mutation is the best means to reach a diagnosis, but this does not necessarily rule out other causes of exercise intolerance or collapse.
No curative treatment for this disease exists (Taylor et al, 2008). Most affected dogs make suitable pets and should refrain from intense exercises such as running, chasing balls or toys, and hunting and field trial events.
If the dogs survive, the condition is not progressive, therefore they can live a normal lifespan if not heavily exercised. It is considered that the episodes of collapse decrease in frequency with age – possibly as a result of being less excitable. Most dogs can live normal active lives if specific trigger activities are avoided or done in moderation.
An exercise-induced collapse syndrome similar to the EIC in Labrador retrievers has been recognised in border collies. It has coined a number of different names, including border collie collapse (BCC) or exercise-induced hyperthermia.
Similar to Labrador retrievers, it is most common in working farm dogs or those involved in strenuous activities such as agility, flyball or those that retrieve tennis balls over a long period of time. The key differences between BCC and EIC in Labrador retrievers are that the majority of border collies (up to 85%) exhibit changes in mentation such as being disorientated, distracted or dull (Taylor et al, 2016a). The ataxia is different in that it often has a vestibular quality (that is, dogs show a loss of balance to one side) and tends to affect all four legs rather than just the pelvic limbs.
Some border collies are described as having abnormal compulsive activity with aimless wandering, pacing and circling. Some affected dogs will exhibit squinting of the eyes or rapid repetitive eyelid blinking, and photosensitivity that may be difficult to distinguish from partial seizure activity (Taylor et al, 2016b).
Labrador retrievers with EIC tend to continue to try to exercise when affected as opposed to border collies that tend to wander aimlessly, or stagger and fall. Finally, border collies with BCC are negative for the DNM1 mutation causing EIC in the Labrador retriever with the causative mutation yet to be identified, although testing for this mutation when suspecting BCC is advisable. Aside from these discrepancies, the conditions in the border collie and Labrador retriever are otherwise relatively similar.
Although first described in border collies, a similar collapse condition has been observed in other breeds, including the Australian cattle dog, Australian kelpie, Australian shepherd dog, bearded collie and Shetland sheepdog.
Myasthenia gravis is a disorder of the neuromuscular junction, and, as such, is considered a neuromuscular condition. Therefore, the main presenting sign is severe weakness or paresis that worsens with exercise. This is different to EIC/BCC, in which incoordination (that is, ataxia) is seen with exercise.
Ataxia is not a feature of myasthenia gravis and, therefore, the presence of this single clinical sign is sufficient to rule out the condition. If ataxia is not present and the weakness is exercise induced, then serological testing for acquired myasthenia gravis is advisable by testing serological presence of acetylcholine receptor antibodies.
Metabolic myopathies are exceptionally rare and tend to result in weakness rather than ataxia. Lactate and pyruvate testing (resting and post-exercise) can be done to screen for these conditions, but practically this is very difficult and pragmatically would not lead to a cure for the condition. In practice this is rarely, if ever, performed.
Polymyositis and centronuclear myopathy (also known as hereditary Labrador retriever myopathy or type two myofibre deficiency) refer to disease of the muscle, and so, are considered neuromuscular conditions. In this sense they do not result in ataxia and usually cause persistent rather than the episodic exercise-induced signs seen with the exercise-induced collapse disorders.
It is always hoped that heart disease severe enough to result in collapse will be clinically evident in between the syncopal episodes. In practice, however, this is not always the case.
A resting ECG is always sensible in cases of collapse to assess for arrhythmias, and Holter monitoring is now more widely available and should be considered if a suspicion exists of a cardiac problem, alongside echocardiography.
Hypoglycaemia is most commonly secondary to an insulinoma, although other causes are possible. Consideration once existed that this commonly occurred in farm dogs that had intense exercise throughout the day. However, as our diagnostic abilities have improved, this condition of farm dogs has been less commonly identified and although it may well be the cause for some cases of collapse, the true hypoglycaemia of working dogs is exceptionally rare.
Ideally, blood glucose measurement during a collapse episode is the best diagnostic test for this condition, but in practice, this is very difficult as inevitably the dog is a long way from a veterinary practice when the episode is encountered.
The author’s preference is for serial blood glucose measurements to be collected every 2 hours following a 12-hour fast for 6 hours. If blood glucose concentrations are persistently above 3.5mmol/L, the author is usually comfortable with ruling out hypoglycaemia as the cause for collapse.
Paroxysmal movement disorders are relatively common conditions in dogs, and are being identified more frequently now we have access to smartphone technology that allows an owner to film any abnormal episode their dog has in the home setting (Lowrie and Garosi, 2017). Several categories of movement disorder exist, although the most common type in dogs is called a non-kinesigenic paroxysmal dyskinesia.
As the name implies, these disorders are not associated with movement and so are easily distinguished from exercise-induced collapsing conditions by their association with rest. However, a less common condition referred to as kinesigenic dyskinesia has been identified in German short-haired pointers (Harcourt-Brown, 2008), and this is associated with movement or light exercise.
Again, this is relatively easily distinguished from exercise-induced collapse. However, in people, an exercise-induced paroxysmal dyskinesia exists. Clearly this would be much harder to distinguish from EIC/BCC in that it shares many of the same clinical features, with only some subtle differences.
Many features of these exercise-induced collapsing disorders share similarities with heat stroke. Initial work into EIC/BCC focused on core body temperature during the episode and it was found that, although it was elevated, it was never higher than unaffected dogs undergoing similar levels of exercise.
It is important to recognise that the collapse episodes we see in dogs with exercise-induced collapse are very different from those associated with actual heat stroke. Heat stroke severe enough to cause mentation changes, gait abnormalities and collapse will be life-threatening, and often fatal (Bruchim at al, 2006).
Recovery, if it does occur, is slow and prolonged (hours to days), even with intensive treatment. Measurement of creatine kinase activity is markedly increased in heat stroke as opposed to EIC/BCC, where it is normal or only mildly elevated. Dogs with heat stroke can also develop other problems, such as acute renal failure.
The majority of dogs that collapse due to heat stroke will also show mentation changes and gait abnormalities that are persistent and that last hours to days, as opposed to BCC or EIC in which the disturbance is only transient with a relatively rapid recovery.
In addition, the recurrent nature of EIC or BCC in moderate or cool temperatures distinguishes it from heat stroke, which would usually only occur on a single occasion in particularly hot conditions.
Similar to heat stroke, malignant hyperthermia – when left untreated – is a progressive fatal condition causing multisystem organ failure (Brunson and Hogan, 2004). It is also most commonly seen in dogs in the perioperative period, but is reported to occur on rare occasions with exercise.
The paroxysmal and transient nature of EIC/BCC is a feature shared with the more common presentation of an epileptic seizure. Episodic collapse could be easily mistaken for a partial seizure. Features that differentiate EIC/BCC from an epileptic seizure include:
In summary, the syndrome of exercise-induced collapse should be considered when a dog is presented for ataxia and collapse following a period of strenuous exercise (regardless of the ambient temperature).
The condition is most commonly recognised in the Labrador retriever, border collie and related breeds, although it can, in theory, affect any breed.
The ataxia may be restricted to the pelvic limbs or affect all four limbs, mentation changes may be present with a degree of vestibular ataxia in collie dogs, and dogs should be completely normal in between episodes, with good activity levels.
If the condition is suspected, it is strongly recommended that the dog is tested for the DNM1 mutation, although this has so far only been found to be present in retrievers.
