Intraobserver and interobserver agreement of 8 segmental reflexes in healthy dogs

Abstract Background No available literature supports the claim that the patellar and withdrawal (flexor) reflexes are the only reliable segmental reflexes in dogs. Objective Measure intra‐ and interobserver agreement of 8 segmental reflexes in dogs without clinical evidence of orthopedic or neurologic disease. Animals One‐hundred and one client‐ or staff‐owned dogs between 1 and 10 years of age with no clinical evidence of orthopedic disease, myelopathy, or neuromuscular disease. Methods Descriptive study. The intraobserver proportion of agreement (%) of responses to selected segmental reflexes in right versus left limbs by 3 observers was calculated and reported. The interobserver agreement of 2 observers of responses to selected reflexes was estimated by calculating proportions of agreement, kappa values, and 95% confidence intervals. A segmental reflex with an acceptable agreement was defined as that with a proportion of agreement ≥90% and a Kappa value ≥0.61 in both limbs. Results The intraobserver proportion of agreement for all 3 observers was high (≥95%) for the extensor carpi radialis, withdrawal, patellar, and cranial tibial reflexes. Between observers 1 and 3 and observers 2 and 3, the interobserver proportion of agreement was high (≥ 92%) for the extensor carpi radialis (κ 0.66, not determined [ND]), withdrawal (both limbs, κ ND), patellar (κ ND), and cranial tibial reflexes (κ ND). Conclusions and Clinical Importance The extensor carpi radialis, withdrawal, patellar, and cranial tibial reflexes had a higher proportion of agreement and kappa values between 2 observers.

2][3][4][5] Two studies have evaluated the interobserver agreement of segmental reflexes in dogs.Both studies found the level of experience of the observer has an influence on the interobserver agreement in the biceps reflex, thus that reflex was assessed by observers with the highest level of expertise (ie, neurologists). 6,7The second study identified hair coat length as a factor affecting the interpretation of the biceps reflex, as well as level of expertise influencing interobserver agreement for the patellar reflex. 7The classification of the cranial tibial and extensor carpi radialis reflexes as myotatic reflexes also recently has been challenged. 8,9An unpublished report indicated that the extensor carpi radialis and cranial tibial reflexes were still present in dog cadavers after transection of the radial and sciatic nerves. 8This led to a study that found those reflexes persisted after a local block, indicating that the motor response observed for those reflexes may not be dependent solely on the reflex arc, but also an idiomuscular response. 9r objective was to measure the intra-and interobserver agreement of 8 segmental reflexes in the limbs of dogs without clinical evidence of orthopedic or neurologic disease by 3 veterinary clinicians.
We hypothesized that the patellar and withdrawal reflexes in the thoracic and pelvic limbs would show acceptable agreement and reliable results in clinically normal dogs.

| MATERIALS AND METHODS
The study was performed in accordance with the guidelines of the Institutional Animal Care and Use Committee (IACUC) and the Veterinary Hospital Research Review Committee (VHRRC) of the University of Florida.

| Study animals
A convenience sample of 101 client-or staff-owned dogs that were presented to the neurology service at the University of Florida Small Animal Hospital were sequentially included in the study from September 2021 to August 2022.Dogs of any breed, sex, or weight and between 1 and 10 years of age were included.Puppies (<1 year of age), geriatric dogs (>10 years of age), and dogs with clinical evidence of orthopedic disease that inhibited normal range of motion of the limbs, those with myelopathy and neuromuscular diseases, or with suspicion of such diseases, were not included.Any dog that could not tolerate having all of its segmental reflexes tested in lateral recumbency because of temperament also was excluded from the study.

| Study design
All study dogs underwent a general physical examination and full neurological examination by a neurology resident (observer 3, BC).The intraobserver proportion of agreement was calculated by measuring the agreement of segmental reflexes of the right limb versus the left limb by 2 observers (the resident and 1 of 2 board-certified neurologists).
The examinations were conducted independently 1 time by each observer.Observer 1 (SCJ) evaluated 65 dogs, observer 2 (GG) evaluated 36 dogs, and observer 3 evaluated 101 dogs.Once the examinations were performed, results were recorded for later analysis.
The interobserver proportion of agreement was calculated by measuring and comparing the proportion of agreement of segmental reflexes between observers 1 and 3, and observers 2 and 3.The segmental reflexes were independently tested 1 time by the observers.Extensor carpi radialis reflex: The antebrachium is rested in the nontesting hand of the observer.The location of the muscle belly of the extensor carpi radialis muscle then is estimated by drawing an imaginary line from the lateral epicondyle of the humerus to the mid-dorsal aspect of the carpus and divided into thirds.

| Examination of segmental reflexes
The muscle belly of the extensor carpi radialis muscle is located at the most proximal third of the line.The muscle belly is then struck by a pleximeter.A positive response is defined by a sudden extension of the carpal joint.A negative response is defined by the absence of extension of the carpal joint.
Withdrawal (flexor) reflex in the thoracic limb: The limb is held in extension and a noxious stimulus is applied to the third or fourth digit by pinching the bone of the digits or the interdigital skin.A positive response is defined by the presence of flexion of the carpus, elbow, and shoulder.A negative response is defined by an absence of flexion of the carpus, elbow, and shoulder.
Patellar reflex: The patella is palpated to ensure it is not luxated.
If luxated, the patella should be replaced in the patellar groove of the femur.The limb is held with the stifle flexed and the patellar tendon is identified and struck with a pleximeter.A positive response is defined by the presence of a sudden extension of the stifle.A negative response is defined by the absence of extension of the stifle.

Cranial tibial reflex:
The limb is held resting in the nontesting hand of the observer.The muscle belly of the cranial tibial muscle is identified in the proximal region of the craniolateral aspect of the tibia.The largest area of the muscle belly is struck with a pleximeter.A positive response is defined by a sudden flexion of the tarsus.A negative response is defined by the absence of flexion at the tarsus.

Gastrocnemius reflex:
The metatarsus is held with the stifle in extension and the hock is flexed and externally rotated to load the calcaneal tendon.The common calcaneal tendon just proximal to its insertion at the calcaneus is struck with a pleximeter.
A positive response is defined by a sudden extension of the tarsus or contraction of the gastrocnemius muscle or the caudal thigh muscles.A negative response is defined by the absence of a contraction of the gastrocnemius and caudal thigh muscles and absence of extension of the hock.
Withdrawal (flexor) reflex of the pelvic limb: This reflex is tested in similar fashion as the withdrawal reflex in the thoracic limb.A positive response is defined by the presence of flexion of the hock, stifle, and hip.A negative response is defined by the absence of flexion of the hock, stifle, and hip.

| Statistical analysis
For each observer, the intraobserver agreement of dog responses to    Nine of the 14 forebrain-localized dogs had been diagnosed with idiopathic epilepsy or a behavioral disorder.The other 5 dogs were diagnosed with meningoencephalitis of unknown origin (4) or a meningioma (1).The dog with a cerebellar localization had a normal neurological examination and brain magnetic resonance imaging (MRI), and had episodes of cerebellar ataxia.

| DISCUSSION
The observation of segmental reflexes still plays an integral role in neurolocalization for the practitioner despite the subjective interpretation of the findings.Our results support the hypothesis that the patellar and withdrawal reflexes would be reliably present in the study population of dogs.Based on the frequency of positive responses and agreement between observers, additional reflexes that were considered reliably present in our study sample include the extensor carpi radialis reflex and the cranial tibial reflex.
[3][4][5] Several studies have examined the accuracy of these reflexes in localizing a lesion.A retrospective study investigated the accuracy of the withdrawal reflex in localizing a cervical intervertebral disc herniation in dogs compared with the MRI findings. 12This study concluded that the withdrawal reflex was not reliable in its ability to localize a lesion to the C1-C5 vs the C6-T2 segments.A second study assessed the accuracy of neurologic examination findings in localizing the site of thoracolumbar intervertebral disc herniations in dogs. 13A decreased withdrawal reflex correctly predicted a lesion in the L4-S3 region only 37.5% of the time.In that same study evaluating the patellar reflex, a decreased patellar reflex correctly predicted a lesion in the L4-S3 region only 16.7% of the time whereas an increased patellar reflex was more accurate (87%) in predicting a lesion in the T3-L3 region.A third study investigated the accuracy of the patellar reflex in localizing a thoracolumbar disc extrusion. 14All of the dogs in that study with a decreased patellar reflex had lesions within the T3-L3 region.
T A B L E 6 Interobserver agreement and kappa results of selected reflexes in study dogs between Observer 2 and 3: left limb only.On the other hand, only 2 studies have investigated spinal reflexes in normal dogs.Both of these studies identified that the experience of the examiner can influence the interobserver agreement of the biceps reflex. 10,11If the examiner is very experienced, the biceps reflex can be a reliable reflex.This hypothesis was supported by the intraobserver agreements for the biceps reflex in our study.The agreement was higher for the 2 neurologists (observers 1 and 2) compared with a lower agreement for observer 3 who was a neurology resident.In the other study, long-haired dogs had a 2.6-fold increased chance of disagreement for the biceps reflex compared with shorthaired dogs, and interobserver agreement was found to be better with higher levels of expertise for the patellar reflex. 7Sixteen dogs in our study were breeds with long hair coats (eg, Golden retriever, Cavalier King Charles spaniel, Australian Shepherd, Shetland sheepdog, Flat-coated retriever, Yorkshire terrier, and English setter).
Even with the inclusion of these dogs, our findings still suggest the extensor carpi radialis reflex, patellar reflex, cranial tibial reflex, withdrawal reflex, and possibly the biceps reflex are reliably present in dogs.Our findings also support the hypothesis that a more experienced observer can influence the observed agreement on the biceps reflex.The intraobserver agreement for the biceps reflex for both of the board-certified neurologists (observers 1 and 2) was 94% whereas the intraobserver agreement for the biceps reflex for the neurology resident (observer 3) was 88%.Additionally, the interobserver agreement was reliable only for the left limb but not the right which may be explained by the technique used to isolate the tendon of insertion of the biceps brachii and how the animal was positioned in reference to where the observer was in testing the reflex, which could affect how well the reflex could be observed.Video analysis may have helped in the interpretation of subtle responses that could be easily missed during clinical examination.
In our study, the extensor carpi radialis reflex and the cranial tibial reflex were classified as reliable for neurologic examinations in dogs.
Compared with the extensor carpi radialis reflex, the cranial tibial reflex is more easily tested.The patellar tendon is easily palpated.The A previous study comparing the cranial tibial and extensor carpi radialis reflexes in cats concluded that these 2 reflexes are not strictly myotatic reflexes and may represent an idiomuscular response. 9Idiomuscular responses were first described in 1858 15 and are a result of direct activation of the sarcolemma after direct percussion of the muscle. 16This method is the way in which both the extensor carpi radialis and cranial tibial reflexes are tested in dogs and cats.Idiomuscular responses are not commonly used in human medicine because of their uncertain mechanism and clinical relevance. 15However, differences in the character of the response in testing the cranial tibial reflex in which it demonstrated more of a brisk dorsiflexion of the ankle in patients with lower motor neuron signs versus an inversion of the ankle in patients with upper motor neuron lesions seemed helpful in the assessment of patients with lower limb weakness. 15The correlation to veterinary medicine is likely not helpful because these subtle differences in response would be difficult to appreciate.4][5]17 Although idiomuscular responses may be easy to perform, the presence or absence of these responses may not correlate with neurolocalization of the lesion.
Our study had several limitations.First, the number of animals examined by each of the observers was not equal.This factor could have skewed the results and the observed agreements.Each observer made his or her own interpretations of the reflexes examined, which could introduce bias.Video-taped responses by the most experienced observer with subsequent blinded analysis of the responses by observers of various levels of experience could have removed this bias.
Second, some animals included in our study had abnormalities on their neurologic examinations.Advanced imaging was not performed to exclude the possibility of spinal lesions in all of the animals included.
Therefore, some of the dogs may have had spinal lesions.However, all of these dogs were neuro-localized to the brain and were suspected not to have any concurrent myelopathies.Third, although the calculated interobserver agreement and kappa statistic value were high for several investigated reflexes, it is important to examine all calculated outcomes for agreement.For example, for the biceps reflex (observer 1 vs 3, left limb), the observed agreement was 60/65 or 92%, the calculated kappa value was 0.66, and the 95% CI was 0.44 to 0.90.The width of the CI is strongly dependent on sample size. 18The calculated kappa value (0.66) represents the maximum point estimate in the distribution of the CI, and the lower and upper limits are very unlikely to be values in that distribution.If the true kappa value were 0.66, a sample size 2 times higher (ie, 130 dogs) than that used for the biceps reflex would have resulted in a narrower 95% CI (ie, 0.48-0.65),decreasing the magnitude of uncertainty associated with the calculated agreement for the biceps reflex.In our study, 2 crucial investigated reflexes were the patellar and withdrawal reflexes.In both reflexes, the calculated observed agreements were 64/65 or 98% and 65/65 or 100% but kappa values were not calculated because the frequency of dogs in selected categories was 1 or zero.Finally, the external validity in our study was inherently low because the study sample was limited to a convenience sample of up to 101 client-or staff-owned dogs that were presented to a single university hospital for veterinary care.Thus, study results apply to the study sample of investigated dogs and cannot be extrapolated to other dog populations.

| CONCLUSION
The patellar and the withdrawal reflexes in both the thoracic and pelvic limbs, as well as the extensor carpi radialis and the cranial tibial reflexes, were classified as reliable in a study sample of 101 dogs.
Our findings support inclusion of the patellar and withdrawal reflexes in all neurological examinations.

ACKNOWLEDGMENT
No funding was received for this study.

CONFLICT OF INTEREST DECLARATION
Authors declare no conflict of interest.

OFF-LABEL ANTIMICROBIAL DECLARATION
Authors declare no off-label use of antimicrobials.
Eight segmental reflexes (4 in the thoracic limb and 4 in the pelvic limb) were tested in every dog on both sides.All of the segmental reflexes were tested on the nondependent limb with the dog in lateral recumbency.The segmental reflexes tested in the thoracic limb were the biceps reflex, triceps reflex, extensor carpi radialis reflex, and the withdrawal reflex.The segmental reflexes tested in the pelvic limb were the patellar reflex, cranial tibial reflex, gastrocnemius reflex, and the withdrawal reflex.All of the reflex responses were classified as positive or negative.The description of how the stimulus is applied and the definition of a positive versus a negative response for each segmental reflex are described as follows (Video S1 available): Biceps reflex: The antebrachium is supported with the limb externally rotated.A finger is placed directly over the tendon of insertion of the biceps brachii and a stimulus is provided by striking the dorsal aspect of the finger overlying the tendon with a pleximeter.A positive response is defined by a contraction of the biceps brachii muscle, or a sudden flexion of the cubital (elbow) joint.A negative response is defined by the absence of contraction of the biceps brachii muscle and absence of flexion of the cubital joint.Triceps reflex: The elbow is flexed to at least 90 and externally rotated by approximately 45 .The tendon of insertion of the triceps brachii is struck with a pleximeter just above its attachment to the olecranon.A positive response is defined by a contraction of the triceps brachii muscle or a sudden extension of the cubital joint.A negative response is defined by the absence of contraction of the triceps brachii muscle and absence of extension of the cubital joint.
selected segmental reflexes in right versus left limbs was calculated by dividing the sum of the total number of dogs with negative responses in left and right limbs and the total number of dogs with positive responses in both limbs by the total number of dogs tested.Interobserver agreement was measured by calculating the observed proportion of agreement.In this study, kappa values (k) and 95% confidence intervals (CI) were calculated and reported as measures of reliability.Kappa values were interpreted as follows: excellent agreement (0.93-1.00); very good agreement (0.81-0.92); good agreement (0.61-0.80); fair agreement (0.41-0.60); slight agreement (0.21-0.40); poor agreement (0.01-0.20); no agreement (≤0.00).
DogsOne-hundred and ten dogs (3 intact males, 51 neutered males, 4 intact females, and 52 spayed females) were recruited to participate in the study.Eight dogs (6 neutered males and 2 spayed females) were excluded from the study because their temperaments did not allow for testing of the reflexes in lateral recumbency.One additional dog (6.5-year-old, intact male Golden retriever) was excluded because the dog developed neck pain the day before examination.A total of 101 dogs met the inclusion criteria.Breeds represented included mixed breed (n = 46), Golden retriever (n = 6), Labrador retriever (n = 6), American Staffordshire terrier (n = 4), Australian shepherd (n = 3), French bulldog (n = 3), Boston terrier (n = 2), bull terrier (n = 2), Cavalier King Charles spaniel (n = 2), Shetland sheepdog (n = 2), Weimaraner (n = 2), and 1 of each of the following breeds: American cranial tibialis muscle has a large muscle belly at the cranio-proximal region of the tibia that is easily palpated.The withdrawal reflex is tested in the form of a noxious stimulus by pinching the toes.The responses generated for these 3 reflexes are not subtle or difficult to observe.The extensor carpi radialis muscle on the other hand is a smaller muscle and difficult to palpate externally.It lies in proximity to the common digital extensor and lateral digital extensor muscles that also serve to extend the carpus.If the examiner mistakenly strikes one of the other carpal extensors, the response is likely still observed as an extension of the carpus.
Intraobserver agreement results of selected reflexes in study dogs by Observer 1.
Intraobserver agreement results of selected reflexes in study dogs by Observer 2.
10,11Kappa analysis was conducted using commercial software (MedCalc Statistical Software version 20.111, MedCalc Software Ltd, Ostend, Belgium; https://www.medcalc.org;2022).For selected comparisons with perfect (100%) or near perfect (98%) observed agreement, kappa was not determined (ND) because the frequency of dogs inT A B L E 2 Intraobserver agreement results of selected reflexes in study dogs by Observer 3. Interobserver agreement and kappa results of selected reflexes in study dogs between Observer 1 and 3: left limb only.Interobserver agreement and kappa results of selected reflexes in study dogs between Observer 1 and 3: right limb only.
bulldog, Anatolian shepherd, Australian cattle dog, Basset hound, Belgian Malinois, Boxer, Coonhound, Dachshund, English setter, Flat-coated retriever, German shepherd, German shorthaired pointer, Giant Schnauzer, Great Dane, Greyhound, Jack Russell terrier, Miniature Poodle, Neapolitan mastiff, Pembroke Welsh corgi, Pug, Rottweiler, Whippet, and Yorkshire terrier.There were 54 females (4 intact, T A B L E 3 Abbreviation: ND, not determined because the frequency of dogs in selected categories was zero.T A B L E 5 T A B L E 7 Interobserver agreement and kappa results of selected reflexes in study dogs between Observer 2 and 3: right limb only. ANIMAL CARE AND USE COMMITTEE (IACUC) OR OTHER APPROVAL DECLARATION Approved by the College of Veterinary Medicine Hospital Research Review Committee and the University of Florida IACUC, study number 202111318.