Assessment of the likelihood of hypothyroidism in dogs diagnosed with and treated for hypothyroidism at primary care practices: 102 cases (2016‐2021)

Abstract

Background

There is a possibility that an incorrect diagnosis of hypothyroidism could be made in euthyroid dogs, and the prevalence of hypothyroidism in the dog population remains unknown.

Objectives

To retrospectively assess the percentage of dogs diagnosed with, and treated for, hypothyroidism at first opinion practice which are likely to be hypothyroid and require levothyroxine supplementation.

Animals

One hundred two client‐owned dogs were included in this study.

Materials and Methods

The computerized databases of 7 first opinion practices were searched to identify dogs treated with levothyroxine supplementation. Three European College of Veterinary Internal Medicine—Companian Animals (ECVIM‐CA) diplomates independently assigned 1 of 4 clinical assessments to each case as follows: confirmed or likely hypothyroid, hypothyroidism suspected but not confirmed, hypothyroidism considered unlikely, and no reason to suspect hypothyroidism. They commented as to whether or not they thought levothyroxine supplementation was appropriate.

Results

The clinical assessments of “confirmed or likely hypothyroid”; “Hypothyroidism suspected but not confirmed”; “Hypothyroidism considered unlikely”; and “No reason to suspect hypothyroidism” was assigned respectively by Clinician 1 to 38.2%, 5.9%, 3.9%, and 52% of cases, by Clinician 2 to 48%, 22.6%, 22.6%, 6.9% of cases, and by Clinician 3 to 55.9%, 11.8%, 13.7% and 18.6%. Clinician 1, Clinician 2, and Clinician 3 considered levothyroxine supplementation not indicated in 58.8%, 52.9%, and 45.1% of cases, respectively.

Conclusion

These results support the concern that hypothyroidism might be overly and incorrectly diagnosed in first opinion practice, and that thyroid function testing should be performed only in those dogs with a high pretest probability of the disease.

Abbreviations

ECVIM‐CA

European College of Veterinary Internal Medicine—Companian Animals

ED

equilibrium dialysis

fT4

free thyroxine

RIA

radioimmunoassay

TSH

thyroid stimulating hormone

TT4

total thyroxine

1. INTRODUCTION

Hypothyroidism is one of the most commonly diagnosed endocrinopathies in dogs. ¹ , ² Primary hypothyroidism accounts for about 95% of diagnosed cases, with lymphocytic thyroiditis and idiopathic follicular atrophy accounting for approximately 50% of diagnosed cases, each. ³ , ⁴ Hypothyroidism is commonly diagnosed in first opinion practice, and diagnosed dogs are typically treated with lifelong levothyroxine supplementation.

The “gold‐standard” test for the diagnosis of primary hypothyroidism is considered to be the thyroid stimulating hormone (TSH) stimulation test for several years, ⁵ , ⁶ , ⁷ , ⁸ scintigraphy is able to distinguish between hypothyroidism and nonthyroid illness in dogs. ⁹ , ¹⁰ Because of the lack of readily‐available TSH preparations and scintigraphy around the world, these tests are infrequently performed. Measurement of serum total thyroxine (TT4), TSH, and free thyroxine (fT4) concentrations are most commonly performed. These tests have important limitations with regards to sensitivity and specificity for the diagnosis of primary hypothyroidism. Test sensitivities for TT4, TSH, and fT4 are 89%‐100%, 58%‐87%, and 80%‐98% respectively, with test specificities 73%‐82%, 82%‐100%, and 78%‐94% respectively. ² , ¹¹ , ¹² , ¹³ , ¹⁴ , ¹⁵ , ¹⁶ , ¹⁷ , ¹⁸

The clinical signs of primary hypothyroidism are often vague and nonspecific, and dogs with signs of dermatologic and metabolic diseases are most commonly encountered. ¹ , ⁴ , ¹⁹ , ²⁰ , ²¹ , ²² , ²³ , ²⁴ , ²⁵ , ²⁶ As there are no highly specific or pathognomonic clinical signs associated with primary hypothyroidism, the possibility of nonthyroidal illness as a cause for the presenting signs must always be considered.

In the absence of a readily‐available gold standard test for the diagnosis of primary hypothyroidism, and considering the aforementioned limitations with regards to the sensitivity and specificity of the commonly used tests, thyroid function testing should only be performed in dogs with compatible clinical signs, compatible changes on routine clinicopathological tests, and after considering the potential confounding effects of any nonthyroidal factors. In order to improve the diagnostic performance of any utilized test, and to increase the positive and negative predictive values of the test being performed, it is important that only dogs with a high pretest probability of the disease in question are tested. The true prevalence of primary hypothyroidism in dogs remains unknown, ⁴ and there is a possibility that an incorrect diagnosis of primary hypothyroidism be made in euthyroid dogs. ²⁷

The purpose of our study is to evaluate the likely percentage of dogs diagnosed with and treated for primary hypothyroidism in first opinion practice, which are truly hypothyroid and where levothyroxine supplementation is considered indicated. We hypothesize that a significant percentage of dogs are erroneously diagnosed with hypothyroidism and unnecessarily treated with levothyroxine supplementation.

2. MATERIALS AND METHODS

This study was approved by the Royal College of Veterinary Surgeons Ethics Review Panel in accordance with current guidelines on good clinical practice.

The computerized databases of 7 first opinion practices in the UK were searched to identify dogs treated with levothyroxine between January 1, 2016, and August 1, 2021. The search term “levothyroxine” was used to interrogate the systems, in order to identify all dogs which were prescribed this medication. Dogs were included if diagnosed with hypothyroidism and subsequently prescribed levothyroxine supplementation, regardless of the method of diagnosis or tests performed and regardless of the degree of confidence in the diagnosis by the veterinarian. Dogs were excluded if data potentially pertinent to making the diagnosis of hypothyroidism were available to the case veterinarian at the time of diagnosis, but not subsequently available for collection and review at the time of data collection for this study. All case data were collected by 1 representative from each first opinion practice and anonymized before provision to the authors who reviewed the case data.

Data collected included breed, sex, neuter status, age at diagnosis, copies of the complete pertinent clinical notes, routine clinicopathological test results from within 6 weeks of diagnosis, results of performed thyroid function tests, thyroid function test methodology (eg, in‐house or external lab, and immunoassay or equilibrium dialysis if free T4 was measured), information regarding concurrent illnesses, and medications administered within the 4 weeks before thyroid function testing. For each case, it was also noted whether measurement of TT4 was specifically requested by the veterinarian, or was included as part of the routine clinicopathological tests without being specifically requested by the veterinarian.

Data from all cases were reviewed by 3 ECVIM‐CA (internal medicine) diplomates. Each diplomate assessed the case data independently and blindly. Each clinician assigned 1 of 4 clinical assessments to each case. The 4 clinical assessments were as follows; confirmed or very likely hypothyroid, hypothyroidism suspected but not confirmed, hypothyroidism considered unlikely, no reason to suspect hypothyroidism or thyroid function testing not considered indicated. Clinical assessment was further recategorized into 2 categories: (1) hypothyroidism suspected or confirmed and (0) hypothyroidism unlikely or not suspected for ease of further descriptive analysis. After case assessment, each clinician also commented as to whether or not they considered levothyroxine supplementation indicated for that case. If levothyroxine supplementation was considered indicated or not by at least 2 of the clinicians, that was considered the consensus assessment for that case.

Data was collated and stored in a Microsoft Excel (Office 365) spreadsheet and imported into statistical software Stata (IC v.13.0) for subsequent statistical analyses. Categorical data were summarized as proportions and expressed as percentages. Dog age was formally assessed for normality using the Shapiro‐Wilk test. Because of a nonnormal distribution (Shapiro‐Wilk P < .05), dog age was described using median and range.

Chi‐squared tests were used for each clinician's assessment, as well as the consensus between all clinicians, to assess the relationship between whether levothyroxine supplementation was indicated and whether an in‐house or external laboratory service was used to measure total TT4 and whether levothyroxine supplementation was indicated and whether a serum TT4 measurement was specifically requested or hypothyroidism was mentioned in the initial differential list of the veterinarian. A Fisher's exact test was used for each clinician's assessment, as well as the consensus between all clinicians, to assess the relationship between whether levothyroxine supplementation was indicated and whether a complete thyroid profile had been performed. A chi‐squared test was additionally used to assess the relationship between whether a serum TT4 measurement was specifically requested or hypothyroidism was mentioned in the initial differential list of the veterinarian and whether an in‐house or external laboratory service was used to measure serum TT4 concentrations. A Venn diagram was used to further describe cases that had a complete thyroid profile performed, where serum TT4 measurement was specifically requested or hypothyroidism was mentioned in the initial differential list of the veterinarian and where serum TT4 was measured using an in‐house machine. Significance was set at P ≤ .05 and P‐values were not adjusted for multiple comparisons.

Inter‐rater agreement between the 3 clinicians was assessed by estimating intraclass correlations (ICCs), which alongside measuring the agreement between individual ratings, also took into account the magnitude of disagreement between raters. ²⁸ A 2‐way random‐effects, absolute agreement ICC coefficient (95% CI) was calculated for decisions related to assignment of clinical assessments and ultimately whether levothyroxine supplementation was indicated. The level of agreement was evaluated based on guidelines provided by Koo and Li in 2016, with ICC <0.5 indicating poor, 0.5‐0.75 moderate, 0.75‐0.9 good and >0.9 excellent reliability. ²⁹ Venn diagrams were used to further describe the relationships between the 3 clinician's assessments.

3. RESULTS

3.1. Signalment

A total of 105 dogs met the inclusion criteria; however, because of the incomplete clinical records, 3 cases were excluded resulting in 102 dogs being retained in the study. There were 48 neutered females (47.1%), 37 neutered males (36.3%), 9 entire females (8.8%), and 8 entire males (7.8%). The median age at diagnosis was 9 years (range, 2‐15 years).

Of the 102 included dogs, there were 17 mixed breeds (16.7%), 10 Labrador retrievers (9.8%), 9 Border collies (8.8%), 8 English springer spaniels (7.8%), 6 German shepherd dogs (5.8%), 5 Cocker spaniels (4.9%), 4 Golden retrievers (3.9%), 4 Boxers (3.9%), 4 Staffordshire bull terriers (3.9%), 3 Schnauzers (2.9%), 3 Cavalier King Charles spaniels (2.9%), 3 Bearded collies (2.9%), 3 Jack Russel terriers (2.9%), 3 Doberman pinchers (2.9%), 2 Lurchers (2%), 2 West Highland white terriers (2%), and 2 Poodles (2%). There was 1 each (1%) of the following breeds; Chihuahua, Alaskan Malamute, Australian shepherd dog, French bulldog, Dalmatian, Shih‐tzu, Siberian husky, German spitz, Tibetan terrier, Shetland sheepdog, Dachshund, Border terrier, Airedale terrier, and Beagle.

3.2. Concerns reported by the owners, abnormalities found on physical examination, concurrent illnesses, and ongoing medical treatment

Concerns reported by the owners, abnormalities found on physical examination, concurrent illnesses, and ongoing medical treatment are reported in Table 1 Data S1. Clinical assessment by the Clinician 1, Clinician 2, Clinician 3 after review of the clinical history of 102 dogs treated with levothyroxine supplementation in first opinion practice is reported in Table 2.

TABLE 1.

Summary of the concerns reported by the owners, abnormalities found on physical examination by the veterinarians, concurrent illnesses and ongoing medical treatment reported in the clinical history of the 102 dogs.

Concerns reported by the owners
Skin disorder	37/102 (36.3%)
Pruritus	16 (15.7%)
Poor quality of the coat	11 (10.8%)
Alopecia	7 (6.8%)
Molting	3 (2.9%)
Gastrointestinal disorder	18/102 (17.6%)
Diarrhea	12 (11.7%)
Vomiting	4 (3.9%)
Melena	1 (1%)
Constipation	1 (1%)
Respiratory signs	6/102 (5.9%)
Panting	5 (4.9%)
Cough	3 (2.9%)
Exercise intolerance	2 (1.9%)
Upper respiratory noises and tachypnoea	1 (1%)
Increased respiratory effort	1 (1%)
Orthopedic disorders	11/102 (10.8%)
Stiffness	9 (8.8%)
Lameness	3 (2.9%)
Reluctance to jump and concurrent weakness of the hindlimbs	2 (1.9%)
Collapsing episodes	1 (1%)
Neurological signs	6/102 (5.9%)
Changes in their behavior	6 (5.9%)
Seizure episodes	3 (2.9%)
Facial twitching	2 (1.9%)
Horner's syndrome	1 (1%)
Urinary signs	3/102 (2.9%)
Urinary incontinence	3 (2.9%)
Others
Lethargy	17 (16.6%)
No owner's concerns reported (presented of annual check‐ups)	16 (15.6%)
Gained weight	12 (11.7%)
Increased thirst	6 (5.9%)
Anorexia	5 (4.9%)
Distended abdomen	3 (2.9%)
Loss of weight	3 (2.9%)
Polyphagia	3 (2.9%)
Abnormal growth	1 (1%)
Facial swelling	1 (1%)

Concurrent illnesses
Osteoarthritis	13 (12.7%)
Idiopathic epilepsy	5 (4.9%)
Severe dental disease	4 (3.9%)
Urinary incontinence	2 (1.9%)
Heart disease	2 (1.9%)
Histiocytoma	2 (1.9%)
Pyoderma	2 (1.9%)
Anal gland infection	1 (1%)
Nail infection	1 (1%)
Deaf	1 (1%)
Sudden acquired retinal degeneration syndrome	1 (1%)
Keratoconjunctivitis sicca	1 (1%)
Hepatic neuroendocrine carcinoma	1 (1%)
Hypertension	1 (1%)
Meibomian gland adenoma and a sebaceous hyperplasia (surgically removed 10 days prior thyroid function testing)	1 (1%)

Ongoing medical treatment
Nonsteroidal anti‐inflammatory
Meloxicam	5 (4.9%)
Carprodyl	4 (3.9%)
Previcox	3 (2.9%)
Robenacoxib	2 (1.9%)
Trocoxil	1 (1%)
Painkiller
Gabapentin	2 (1.9%)
Amantadine	2 (1.9%)
Amitriptyline	1 (1%)
Paracetamol	1 (1%)
Immunomodulators
Prednisolone	11 (10.8%)
Chlorphenamine	4 (3.9%)
Oclacitinib	2 (1.9%)
Hydrocortisone acetonate topical spray	1 (1%)
Antibiotics
Amoxicillin clavulanic acid	5 (4.9%)
Cephalosporin	5 (4.9%)
Metronidazole	3 (2.9%)
Enrofloxacin	1 (1%)
Antiepileptic
Phenobarbital	5 (4.9%)
Levetiracetam	1 (1%)
Potassium bromide	1 (1%)
Cardiovascular medications
Pimobendan, furosemide, and benazepril concurrently	2 (1.9%)
Amlodipine	1 (1%)
Others
Phenylpropanolamine	3 (2.9%)
S‐adenosylmethionine and Silybin concurrently	1 (1%)
Ferrous, and cobalamin supplementation	1 (1%)

Abnormalities found on physical examination by the veterinarians
Alopecia	22 (21.5%)
Otitis	9 (8.8%)
Poor quality of coat	7 (6.8%)
Erythema of the skin	4 (3.9%)
Muscle atrophy	4 (3.9%)
Pain found on orthopedic examination	4 (3.9%)
Distended abdomen	3 (2.9%)
Abdominal discomfort	3 (2.9%)
Pododermatitis	3 (2.9%)
Corneal lipid deposit	2 (1.9%)
Hyperpigmentation of the skin	2 (1.9%)
Heat murmur	2 (1.9%)
Peripheral enlargement of the lymph nodes	2 (1.9%) (1 [1%] of these dogs had cytology performed revealing metastatic mast cell tumor)
Tachypnoea	1 (1%)
Tachycardia	1 (1%)
Bradycardia	1 (1%)
Pyrexia	1 (1%)
Hypothermia	1 (1%)
Thin skin	1 (1%)
Hyperkeratosis	1 (1%)
Enlarged testicle (histology revealed Sertoli cells tumor)	1 (1%)
Mammary mass (cytology revealed a mammary carcinoma)	1 (1%)
Mass on one of his digits (cytology revealed a soft tissue sarcoma)	1 (1%)
Abdominal mass noticed on palpation (unknown origin)	1 (1%)
Large perianal mass (unknown origin)	1 (1%)
Vulva swelling and vaginal discharge	1 (1%)
Fleas found on examination	1 (1%)

TABLE 2.

Clinical assessment by the Clinician 1, Clinician 2, Clinician 3 following review of the clinical history of 102 dogs treated with levothyroxine supplementation in first opinion practice.

	Clinical assessment of the 102 client‐owner dogs
	“No reason to suspect hypothyroidism and/or thyroid function testing not considered indicated”	“Confirmed or very likely hypothyroid”	“Hypothyroidism suspected but not confirmed”	“Hypothyroidism considered unlikely”	“Levothyroxine supplementation was considered indicated”
Clinician 1	53 (52%)	39 (38.2%)	6 (5.9%)	4 (3.9%)	42 (41.1%)
Clinician 2	7 (6.9%)	49 (48%)	23 (22.6%)	23 (22.6%)	48 (47%)
Clinician 3	19 (18.6%)	57 (55.9%)	12 (11.8%)	14 (13.7%)	56 (54.9%)

	Levothyroxine supplementation was considered not indicated
Serum TT4 concentration was measured by	Clinician 1	Clinician 2	Clinician 3
External laboratory 78/102 (76.5%)	44 (56.4%)	37 (47.4%)	31 (39.7%)
In‐house machine 24/102 (23.5%)	16 (66.7%)	17 (70.8%)	15 (62.5%)

	Levothyroxine supplementation was considered not indicated
Serum TT4 measurement was specifically requested, or hypothyroidism was mentioned in the initial differential list of the veterinarian or was included as part of the routine clinicopathological results	Clinician 1	Clinician 2	Clinician 3
Serum TT4 measurement was specifically requested 73/102 (71.6%)	40 (54.8%)	33 (45.2%)	31 (42.4%)
Serum TT4 measurement was not specifically requested and was included as part of the routine clinicopathological results 29/102 (28.4%)	20 (69%)	21 (72.4%)	15 (51.7%)

3.3. Thyroid function tests

Measurement of serum TT4 concentration was performed in all 102 cases and was below reference range in 91/102 (89.2%) cases and within reference range in 11/102 (10.8%) cases. Total T4 was measured using an external commercial laboratory test in 78/102 (76.5%) cases, and an in‐house machine for 24/102 (23.5%) cases. Clinician 1, Clinician 2, and Clinician 3 considered levothyroxine supplementation not indicated in 44/78 (56%), 37/78 (47%), and 31/78 (40%) of cases respectively, where TT4 was measured by an external commercial laboratory. Clinician 1, Clinician 2, and Clinician 3 considered levothyroxine supplementation not indicated in 16/24 (67%), 17/24 (71%), and 15/24 (63%) of cases respectively, where an in‐house machine was used to measure TT4. There was a significant association between laboratory service and whether levothyroxine supplementation was indicated in assessments by Clinician 2 (χ ² = 4.0, P = .04) and Clinician 3 (χ ² = 3.8, P = .05) but not by Clinician 1 (χ ² = 0.8, P = .37). In assessments by Clinicians 2 and 3, levothyroxine supplementation was not indicated in a higher proportion of cases where serum TT4 was measured using an in‐house machine which was consistent with the consensus assessment.

Serum TT4 measurement was specifically requested, or hypothyroidism was mentioned in the initial differential list of the veterinarian in 73/102 (71.6%) cases. Clinician 1, Clinician 2, and Clinician 3 considered levothyroxine supplementation not indicated in 40/73 (55%), 33/73 (45%), and 31/73 (42%) of these 73 cases, respectively. Total T4 measurement was included as part of the routine clinicopathological results, and not specifically requested or hypothyroidism not mentioned in the initial differential list by the veterinarian, in 29/102 (28.4%) cases. In these cases, Clinician 1, Clinician 2, and Clinician 3 considered levothyroxine supplementation not indicated in 20/29 (69%), 21/29 (72%) and 15/29 (52%), respectively. There was a significant association between whether serum TT4 measurement was specifically requested or hypothyroidism was mentioned in the initial differential list of the veterinarian and whether levothyroxine supplementation was indicated in assessments by Clinician 2 (χ ² = 6.2, P = .01) but not by Clinician 1 (χ ² = 1.7, P = .19) and Clinician 3 (χ ² = 0.7, P = .4). In assessments by Clinicians 2, levothyroxine supplementation was not indicated in a higher proportion of cases where serum TT4 measurement was not specifically requested or hypothyroidism was not mentioned in the initial differential list of the veterinarian. An association between whether serum TT4 measurement was specifically requested or hypothyroidism was mentioned in the initial differential list of the veterinarian and laboratory service (external or in‐house measurement of serum TT4) was not identified (χ ² = 1.3, P = .26).

Serum TSH concentration was measured in 87/102 (85.3%) cases and was within reference range in 21/87 (24%) cases, and above reference range in 66/87 (76%) cases. Serum fT4 concentration was measured in 25/102 (25%) cases, 13/25 (52%) of which were performed by equilibrium dialysis. Serum fT4 concentrations measured without equilibrium dialysis were below reference range in 9/12 (75%) cases, and within reference range in 3/12 (25%) cases. Serum fT4 concentrations measured by equilibrium dialysis were below reference range in 9/13 (69%) cases, and within reference range in 4/13 (31%) cases. Serum antithyroglobulin autoantibody concentrations were measured in 11/102 (10.8%) cases, and were positive in 3 cases, and negative in 8 cases. Of the cases where TT4 was measured using an in‐house machine, serum TSH concentration was measured in 19/24 (79%) cases and was within reference range in 4/19 cases (21%), and above reference range in 15/19 (79%) cases. Serum fT4 concentration was measured in 2/24 (8%) cases, and was performed by equilibrium dialysis for each of these cases (100%). Serum fT4 concentrations measured by equilibrium dialysis were below reference range in 1 (50%) case, and within reference range in 1 (50%) case. Serum antithyroglobulin autoantibody concentration was measured in 1/24 (4%) dog, and was negative. Serum TSH, fT4, and antithyroglobulin autoantibody concentrations were all performed by an external laboratory. Of the 11 (10.8%) dogs where serum antithyroglobulin autoantibody concentrations was measured, serum TT4, TSH and fT4 concentrations were also performed in 10/11 (91%) dogs, and so these dogs had a complete thyroid profile performed. Of the 25 dogs (24.5%) that had serum fT4 concentration measured, 24/25 (96%) dogs had serum TT4, and TSH concentrations performed concurrently. One of the dogs (4%) had serum TT4 and fT4 concentrations performed; however, serum TSH concentration was not measured. An association between whether levothyroxine supplementation was indicated and whether a complete thyroid profile was performed was not identified in any of Clinician's assessments, nor in the consensus. There was only 1 case that had a full panel of tests, where TT4 test was specifically requested and where the test was performed in‐house while 48% of the cases where the TT4 test was specifically requested by the veterinarian did not have a full panel and testing was performed externally (Figure 1).

FIGURE 1.

A Venn diagram showing the overlap of cases of dogs diagnosed with and treated for hypothyroidism at primary care practices that had a complete thyroid profile performed, whether serum TT4 measurement was specifically requested or hypothyroidism was mentioned in the initial differential list of the veterinarian and whether TT4 was measured using an in‐house machine.

3.4. Agreement between the 3 clinicians

Each case was independently and blindly reviewed by 3 internal medicine diplomates, designated Clinician 1, Clinician 2, and Clinician 3. Inter‐rater agreement between the 3 clinicians based on individual ratings was found to be poor to moderate (ICC 0.46, 95% CI 0.29, 0.61) when it came to assigning clinical assessments. When recategorizing clinical assessment into 2 categories (hypothyroidism suspected or confirmed and hypothyroidism unlikely or not suspected) consensus between the 3 clinicians as to whether hypothyroidism was suspected or confirmed was reached for 38 (37%) cases while consensus on the likely absence of hypothyroidism was reached in 18 (18%) cases (Figure 2). Clinicians 2 and 3 appeared to be more similar in their clinical assessments than Clinician 1.

FIGURE 2.

A Venn diagram showing agreement as to whether hypothyroidism was suspected or confirmed based on independent and blinded review of clinical records of dogs diagnosed with and treated for hypothyroidism at primary care practices as assessed by 3 internal medicine diplomates (Clinicians 1, 2, and 3).

Clinician 1 gave the clinical assessment of “No reason to suspect hypothyroidism or thyroid function testing not considered indicated” to 53/102 (52%) cases, and levothyroxine supplementation was considered not indicated in all of these cases. “Confirmed or likely hypothyroid” was assigned to 39/102 (38.2%) cases, with levothyroxine supplementation considered indicated in all of these cases. “Hypothyroidism suspected but not confirmed” was assigned to 6/102 (5.9%) cases, with levothyroxine supplementation considered indicated in 3 of them. “Hypothyroidism considered unlikely” was assigned to 4/102 (3.9%) cases and levothyroxine was not considered indicated in all of these cases. In total, Clinician 1 considered levothyroxine supplementation indicated in 42/102 (41.1%) cases and not indicated in 60/102 (58.8%) cases. Of the 60 cases where levothyroxine supplementation was considered not indicated, 19/60 (32%) of these cases were suspected to have concurrent disease potentially resulting in euthyroid‐sick syndrome, 11/60 dogs (18%) were receiving medication where interference with thyroid function testing was considered possible, and 2/60 dogs (3%) were considered to have breed‐related alterations in their thyroid function test results. The 28/60 (47%) remaining cases were considered as not showing enough clinical evidence to support hypothyroidism. Of the 24 cases where total T4 was measured using an in‐house machine, 6/24 (25%) of these cases were suspected to have concurrent disease potentially resulting in euthyroid‐sick syndrome, 3/24 dogs (13%) were receiving medication where interference with thyroid function testing was considered possible, and 7/24 (29%) dogs were considered not showing enough clinical evidence to support hypothyroidism.

Clinician 2 gave the clinical assessment of “Confirmed or likely hypothyroid” to 49/102 (48%) cases, with levothyroxine supplementation considered indicated in 46/49 (94%), and not indicated in 3/49 (6%) of these cases. “Hypothyroidism suspected but not confirmed” was assigned to 23/102 (22.6%) cases, with levothyroxine supplementation considered indicated in 2/23 of these cases, and not indicated in 21/23 of these cases. “Hypothyroidism considered unlikely” was assigned to 23/102 (22.6%) of cases, and levothyroxine supplementation was considered not indicated in all of these cases. “No reason to suspect hypothyroidism or thyroid function testing not considered indicated” was assigned to 7/102 (6.9%) cases, and levothyroxine supplementation was considered not indicated in all of these cases. In total, Clinician 2 considered levothyroxine supplementation indicated in 48/102 (47%) of cases and not indicated in 54/102 (52.9%) cases. Of the 54 cases where levothyroxine supplementation was considered not indicated, 13/54 (24%) were suspected to have concurrent disease potentially resulting in euthyroid‐sick syndrome. In 13/54 dogs (24%), medications where interference with thyroid function testing was considered possible were being administered, and 2/54 (4%) dogs were considered to have breed‐related alterations in their thyroid function test results. The 26/54 (48%) remaining cases were considered not showing enough clinical evidence to support hypothyroidism. Of the 24 cases where total T4 was measured using an in‐house machine, 3/24 (13%) of these cases were suspected to have concurrent disease potentially resulting in euthyroid‐sick syndrome, 4/24 dogs (17%) were receiving medication where interference with thyroid function testing was considered possible, and 10/24 (42%) dogs were considered not showing enough clinical evidence to support hypothyroidism.

Clinician 3 gave the clinical assessment of “Confirmed or likely hypothyroid” to 57/102 (55.9%) cases, with levothyroxine supplementation considered indicated in 56/57, and not indicated in 1/57 of these cases. “No reason to suspect hypothyroidism or thyroid function testing not considered indicated” was assigned to 19/102 (18.6%) cases, and levothyroxine supplementation was considered not indicated in all of these cases. “Hypothyroidism considered unlikely” was assigned to 14/102 (13.7%) cases, and levothyroxine supplementation was considered not indicated in all of these cases. “Hypothyroidism suspected but not confirmed” was assigned to 12/102 (11.8%) cases, and levothyroxine supplementation was considered not indicated in all of these cases. In total, Clinician 3 considered levothyroxine supplementation indicated in 56/102 (54.9%) of cases and not indicated in 46/102 (45.1%) cases. Of the 46 cases where levothyroxine supplementation was considered not indicated, 10/46 (22%) of these cases were suspected to have concurrent disease potentially resulting in euthyroid‐sick syndrome, 3/46 dogs (7%) were receiving medication where interference with thyroid function testing was considered possible, and 2/46 (5%) dogs were considered to have breed‐related alterations in their thyroid function test results. The remaining 31/46 (67%) cases were considered not showing enough clinical evidence to support hypothyroidism. Of the 24 cases where total T4 was measured using an in‐house machine, 2/24 (8%) of these cases were suspected to have concurrent disease potentially resulting in euthyroid‐sick syndrome, 1/24 dogs (4%) were receiving medication where interference with thyroid function testing was considered possible, and 12/24 (50%) dogs were considered not showing enough clinical evidence to support hypothyroidism.

With regards to whether or not levothyroxine supplementation was indicated for each individual case, the inter‐rater agreement between the 3 clinicians was poor to moderate (ICC 0.55, 95% CI 0.44, 0.65). The 3 clinicians were unanimous in 67/102 (65.7%) cases. Of these 67 cases, the assessment was that levothyroxine supplementation was not indicated in 36/67 (54%) cases and indicated in 31/67 (46%) cases (Figure 3). Two of the 3 clinicians agreed upon whether or not levothyroxine supplementation was indicated in the other 35/102 (34.3%) cases and of these cases, levothyroxine supplementation was considered indicated by 2 clinicians in 18/35 (51%) cases and considered not indicated by 2 clinicians in 17/35 (49%) cases. In total, of the 102 cases reviewed, the consensus was that levothyroxine supplementation was considered not indicated in 53/102 (52%) cases and indicated in 49/102 (48%) cases.

FIGURE 3.

A Venn diagram showing the agreement as to whether levothyroxine supplementation was indicated based on independent and blinded review of clinical records of dogs diagnosed with and treated for hypothyroidism at primary care practices as assessed by 3 internal medicine diplomates (Clinicians 1, 2, and 3).

4. DISCUSSION

In those cases where levothyroxine supplementation was considered not indicated, the most common reason given by each of the 3 case assessors for such an assessment was an apparent lack of clinical signs and clinicopathologic findings compatible with hypothyroidism. This is important to highlight as a high index of suspicion for hypothyroidism should exist before testing thyroid function in order to improve its predictive value. Concurrent nonthyroidal illness and concurrent medications were also commonly listed factors which resulted in an assessment of levothyroxine supplementation not being indicated. Some dogs were diagnosed with malignant neoplasia or pyometra at the time of the consultation. We can also notice that several dogs have been administered medications such as glucocorticoids or phenobarbital. The dosage, and duration of the treatment were not reviewed and so conclusion regarding how likely these medications could affect the results of the dogs that were tested for hypothyroidism could not be drawn. However, it raises the concern that a lack of knowledge regarding the fact that some factors could reduce the sensitivity and specificity of thyroid function tests. Multiple nonthyroidal factors including breed, ³⁰ , ³¹ , ³² , ³³ , ³⁴ , ³⁵ , ³⁶ , ³⁷ , ³⁸ concurrent illness, ³⁹ , ⁴⁰ , ⁴¹ certain medications, ⁴² , ⁴³ , ⁴⁴ , ⁴⁵ , ⁴⁶ age, ³⁴ and sex ⁴ can significantly affect the results of measured TT4, TSH and fT4 concentrations. Drugs such as glucocorticoids, phenobarbital, aspirin, ketoprofen, carprofen, clomipramide, and sulfonamides affect the results of thyroid function tests. ⁴² , ⁴³ , ⁴⁴ , ⁴⁵ , ⁴⁶ Decreased serum TT4 and increased serum TSH concentrations occur in dogs treated with phenobarbital. ⁴⁷ , ⁴⁸ Prednisolone decreases serum TT3, TT4, and fT4 concentrations, depending on the dose administered ⁴⁹ , ⁵⁰ , ⁵¹ and the duration of the therapy ⁴⁹ , ⁵¹ ; however, it does not affect the serum TSH concentration. ⁵⁰ , ⁵¹ Sulphonamides can induce hypothyroidism, an adverse effect which can be reversible upon stopping administration of this medication. ⁵² , ⁵³ , ⁵⁴ , ⁵⁵ As thyroid function testing can be affected by several factors, review of the dog history, thorough clinical assessment to ensure that the dog has clinical signs compatible with hypothyroidism, and where appropriate, diagnostic testing to exclude concurrent illness should be performed before performing endocrine testing and making a diagnosis of hypothyroidism. ⁵⁶ , ⁵⁷ In dogs with clinical signs compatible with hypothyroidism but suffering from nonthyroidal illness (NTI), scintigraphy could be considered in ordered to distinguish hypothyroidism from NTI. ⁹ , ¹⁰ Otherwise, repeating the thyroid function testing could be performed several weeks after resolution of the illness or discontinuation of the drugs affecting the thyroid function tests. While there seems to be a perception that levothyroxine supplementation is “safe,” this is not necessarily the case and excess supplementation can cause clinically relevant side‐effects, or thyroid atrophy in treated euthyroid dogs. ⁵⁸ Dogs on treatment are also typically subjected to repeat and lifelong blood sampling for monitoring purposes and this has direct dog implications along with financial implications for the owner. Furthermore, the misdiagnosis of hypothyroidism could delay the diagnosis of any other concurrent disease. ⁵⁸

In this study, almost half of the dogs were neutered females, and less than 8% of the dogs were entire male. Intact males and neutered females are believed to be more likely to be diagnosed with hypothyroidism. ¹ Furthermore, breeds reported at higher risk of developing the disease such as Cocker spaniels, Golden retriever, Boxer, Schnauzer, Doberman pinscher, and Beagle were included in our study, and represented about 19.6% of our study sample. ³ , ⁵⁹ , ⁶⁰ , ⁶¹ , ⁶² , ⁶³ The mean age of the dogs included in this study is 9 years old, and could represent a population of elderly dogs suffering with concurrent illness.

There was a significant association between whether serum TT4 measurement was specifically requested or hypothyroidism was mentioned in the initial differential list of the veterinarian and whether levothyroxine supplementation was indicated in assessments by 1 of the Clinician but not by 2 of them. Because of the retrospective nature of this study, we cannot exclude that hypothyroidism was considered as part of the diagnostic differential of the veterinarian, but not mentioned in the clinical record of the dogs. However, it is becoming increasingly common for serum TT4 concentration to be measured as part of “routine” clinicopathological testing, with some commercial laboratories including TT4 measurement as part of their canine “general health” profiles, regardless of whether or not the veterinary surgeon has a clinical suspicion for hypothyroidism. It is generally widely accepted that only dogs with a clinical suspicion for hypothyroidism should be screened for the disease by measuring TT4. ⁵⁸ At the time of manuscript preparation (August 2022), 1 large, commercial laboratory states on their website that they have data on file showing “1 in 7 profiles run on seemingly healthy adult dogs show a low T4 result …,” and they suggest “always” measuring T4 when “running blood work on a senior dog ….” The authors of this manuscript strongly disagree with the suggestion that TT4 should be measured in every senior dog as part of general wellness screens. With studies showing the prevalence of hypothyroidism to be approximately 1% in the canine population, ¹ , ⁴ the authors of this study feel that the report of 1 in 7 (14%) “seemingly healthy dogs” having a low TT4 is a reason not to measure TT4 unless there is a clinical suspicion for hypothyroidism. As discussed, the diagnostic performance and positive and negative predicative values of any test are influenced by the prevalence of the disease in the population being tested. Testing only those dogs with a high pretest probability of the disease in question results in increased reliability when interpreting test results and making a diagnosis, and this is particularly important for tests with low specificity, such as TT4 for diagnosing hypothyroidism.

Validation of radioimmunoassay (RIA) and chemiluminescent techniques for measurement of serum TT4 concentration in dogs have been published. ⁶⁴ , ⁶⁵ In‐house measurement of TT4 is commonly performed and there are concerns regarding the accuracy of these in‐house machines. ⁶⁶ Despite being easily accessible, rapid, and economical, and giving the opportunity to have a rapid result and discussion with the dog owner, there are opposing opinions regarding the accuracy of such tests. One previous study found that reliance on an in‐house ELISA for TT4 measurement could result in appropriate clinical decision making in 62% of dogs, as results could be underestimated or overestimated when compared with RIA. ⁶⁶ In our study, there was a significant association between laboratory service and whether levothyroxine supplementation was indicated in assessments by 2 of the clinicians but not by 1 of them. These results might suggest that the use of in‐house machines compared to an external laboratory for the measurement of TT4 could increase the risk of supplementation with levothyroxine to dogs that do not require it.

Thyroxine concentration can be affected by the presence of autoantibodies in hypothyroid dogs, and by concurrent disease or certain medications. ⁶⁵ , ⁶⁷ , ⁶⁸ Radioimmunoassay and nonisotopic methods are commonly used to measure TT4. If the assay is unable to reliably distinguish antibody‐bound hormone, the presence of autoantibodies can falsely increase or decrease the estimated serum TT4 concentration ⁶⁵ , ⁶⁷ , ⁶⁸ and can lead to misdiagnosis of hypothyroidism. ³ , ⁶⁹ The free concentration of thyroxine measured by equilibrium dialysis (ED) is less affected by concurrent factors compared to TT4. ¹³ , ⁵⁸ , ⁶⁷ Other methods have been developed in order to measure fT4, using thyroxine analogue hormone, nonreactive to binding proteins. This method depends on the dominance of hormone binding by circulating thyroxine‐binding globulin, and results to be more affected by concurrent factors such as concurrent disease, presence of autoantibodies, abnormal thyroid hormone binding proteins, and high serum concentration in free fatty acid. The measurement of fT4 concentration by these methods is imprecise compared to equilibrium dialysis. ⁶⁸ , ⁷⁰ , ⁷¹ In our study, fT4 was measured in 25 cases, but only 13 of them were performed by equilibrium dialysis, which may decrease the accuracy of this test making the fT4 results of the cases included in this study questionable.

Poor to moderate inter‐rater agreement was found between the 3 clinicians for both assignment of clinical assessments and whether levothyroxine supplementation was ultimately indicated. While the clinical interpretation between the 3 diplomates assessing the included cases can be considered a limitation of this study, the fact that the clinicians have a different opinion on the cases highlights that a gold standard test for the diagnosis of hypothyroidism is needed. In order to decrease the bias of inherent difference in subjective clinical opinions, 3 clinicians were requested to clinically assess the history, of the dogs enrolled in this project and the final decision (levothyroxine supplementation indicated or not) was based on the prevalent common opinion. Furthermore, 4 clinical assessments were defined and allocated accordingly to each dog. These categories were subjective, and could have led to individual variation regarding the interpretation of the hypothyroidism status of the dog. The quality of the clinical records could as well have affected the assignment given by the diplomates. This could explain why the inter‐rater agreement regarding the clinical assessment between the 3 clinicians based on individual ratings was poor compared to the inter‐rater agreement regarding whether or not levothyroxine supplementation was indicated which was found to be on the lower end of moderate.

The limitations of this study are intrinsically related to its retrospective nature because of incomplete clinical records. Some potentially pertinent clinical notes might have been missing. Moreover, this study reviewed client‐owned dogs treated with levothyroxine following diagnosis of hypothyroidism in first opinion practices, but it does not review how often first‐opinion practitioners correctly decided that a dog with clinical suspicion of hypothyroidism or abnormal thyroid function testing did not have hypothyroidism and did not need l‐thyroxine treatment.

CONFLICT OF INTEREST DECLARATION

Authors declare no conflict of interest.

OFF‐LABEL ANTIMICROBIAL DECLARATION

Authors declare no off‐label use of antimicrobials.

INSTITUTIONAL ANIMAL CARE AND USE COMMITTEE (IACUC) OR OTHER APPROVAL DECLARATION

Approved by the Royal College of Veterinary Surgeons.

HUMAN ETHICS APPROVAL DECLARATION

Authors declare human ethics approval was not needed for this study.

Supporting information

Data S1. Supporting information.

Travail V, Fernandez Sanchez C, Costo JM, et al. Assessment of the likelihood of hypothyroidism in dogs diagnosed with and treated for hypothyroidism at primary care practices: 102 cases (2016‐2021). J Vet Intern Med. 2024;38(2):931‐941. doi: 10.1111/jvim.16993