Correlation of thyroid hormone measurements with thyroid stimulating hormone stimulation test results in radioiodine‐treated cats

Abstract Background Iatrogenic hypothyroidism can develop after radioiodine‐I131 (RAI) treatment of hyperthyroid cats and can be diagnosed using the thyroid stimulating hormone (TSH) stimulation test. Objectives To assess the effect of noncritical illness on TSH stimulation test results in euthyroid and RAI‐treated cats. To assess the correlation of low total‐thyroxine (tT4), low free‐thyroxine (fT4), and high TSH concentrations with TSH stimulation test results. Animals Thirty‐three euthyroid adult cats and 118 client‐owned cats previously treated with RAI. Methods Total‐thyroxine, fT4, and TSH were measured, and a TSH stimulation test was performed in all cats. Euthyroid control cats were divided into apparently healthy and noncritical illness groups. RAI‐treated cats were divided into RAI‐hypothyroid (after‐stimulation tT4 ≤ 1.5 μg/dL), RAI‐euthyroid (after‐stimulation tT4 ≥ 2.3 μg/dL OR after‐stimulation tT4 1.5‐2.3 μg/dL and before : after tT4 ratio > 1.5), and RAI‐equivocal (after stimulation tT4 1.5‐2.3 μg/dL and tT4 ratio < 1.5) groups. Results Noncritical illness did not significantly affect the tT4 following TSH stimulation in euthyroid (P = .38) or RAI‐treated cats (P = .54). There were 21 cats in the RAI‐equivocal group. Twenty‐two (85%) RAI‐hypothyroid cats (n = 26) and 10/71 (14%) of RAI‐euthyroid cats had high TSH (≥0.3 ng/mL). Twenty‐three (88%) RAI‐hypothyroid cats had low fT4 (<0.70 ng/dL). Of the 5 (7%) RAI‐euthyroid cats with low fT4, only one also had high TSH. Only 5/26 (19%) RAI‐hypothyroid cats had tT4 below the laboratory reference interval (<0.78 μg/dL). Conclusions and Clinical Relevance The veterinary‐specific chemiluminescent fT4 immunoassay and canine‐specific TSH immunoassay can be used to aid in the diagnosis of iatrogenic hypothyroidism in cats.

Conclusions and Clinical Relevance: The veterinary-specific chemiluminescent fT4 immunoassay and canine-specific TSH immunoassay can be used to aid in the diagnosis of iatrogenic hypothyroidism in cats.

K E Y W O R D S
blood pressure, chronic kidney disease, feline, free thyroxine, hypertension, triiodothyronine., TSH

| INTRODUCTION
Biochemical evidence of iatrogenic hypothyroidism in radioiodine I 131 (RAI) treated cats has been documented for decades. Historically, iatrogenic hypothyroidism was not thought to be a clinically important problem in most cats, and treatment was not recommended unless clinical signs consistent with hypothyroidism were present. 1 Chronic kidney disease (CKD) is a common problem in the senior cat population, 2,3 and azotemia can be worsened by treatment of hyperthyroidism, because of the decline in glomerular filtration rate that occurs after normalization of thyroid hormone concentrations. 4 Iatrogenic hypothyroidism in cats treated with antithyroid medication and RAI increases the risk of azotemia and decreases survival time in azotemic-treated cats 5,6 and therefore the importance of diagnosis and treatment of iatrogenic hypothyroidism is increasingly recognized. 7 The diagnosis of iatrogenic hypothyroidism can be challenging due to the effect of concurrent illness on serum free thyroxine (fT4) and serum total thyroxine (tT4) concentrations, 8,9 the poor sensitivity and specificity of commercially available thyroid stimulating hormone (TSH) assays, none of which are specific to cats, 10 and the variability in performance of various fT4 assay methodologies. [11][12][13] Few studies have looked at the sensitivity and specificity of the available thyroid hormone analyses for the diagnosis of iatrogenic hypothyroidism in cats. 6,13 Traditionally tT4 has been used alone in cats to diagnose thyroid disease; however, tT4 has poor sensitivity and specificity for the diagnosis of iatrogenic hypothyroidism in this species, 6 especially when laboratory reference intervals (RIs) are used. In humans and dogs, combinations of fT4, tT4, and TSH are used to confirm a diagnosis of hypothyroidism, or, if the diagnosis cannot be confirmed by measurement of single thyroid hormones, thyroid scintigraphy, TSH stimulation test or thyroid biopsy are performed. 14 Thyroid scintigraphy is currently used as the reference test for the diagnosis of hypothyroidism in RAI-treated cats. 6,13 Increased serum TSH concentrations (using canine-specific chemiluminescent TSH immunoassay) are common in hypothyroid cats, whereas fT4 (using an equilibrium dialysis [ED] method) and tT4 concentrations are within laboratory RI in 75% and 46% of hypothyroid cats, respectively. 6 Subclinical hypothyroidism, defined as cats with elevated serum TSH concentration and fT4 or tT4 within the RI, is therefore common in RAI-treated cats. This condition is clinically important due to the potential harmful effects on renal function and consequent decrease in survival time. 6 The TSH stimulation test is considered a reliable reference test for hypothyroidism in healthy dogs that are not on thyroid suppressive medications (eg, glucocorticoids). 15,16 Test criteria for the TSH stimulation test in cats using recombinant human TSH (rhTSH) were derived using 7 to 8 month old, specific pathogen free, research cats which are not representative of the typically senior population of client-owned cats with iatrogenic hypothyroidism. 17 There are limited data published looking at the response of mature adult, or senior cats to the exogenous administration of rhTSH. 18,19 In this prospective study, mature adult cats were recruited as a control group and to assess the effect of noncritical illness on the response of mature adult euthyroid cats to TSH administration. RAI cats were assessed for concurrent illness using the same criteria as used for the control cats (see above) and divided into two groups: no clinically important illness detected and noncritically ill, to analyze the effect of illness on the thyroid hormone results.
RAI cats were also divided into 3 groups (RAI-hypothyroid, RAIeuthyroid, and RAI-equivocal groups) according to their TSH stimulation test results.

| Systolic blood pressure measurement
A noninvasive Doppler technique was used to measure SBP after a period of acclimatization, and using previously published methods. 20 Average SBP was calculated from a minimum of 3 consecutive readings (within 5 mm Hg). If average SBP was ≥160 mm Hg, indirect fundoscopy was performed after applying one drop of tropicamide 1% to both eyes.

| Sample collection and handling
All baseline blood samples for thyroid hormone analysis and biochemistry were collected in the morning and placed into serum separator tubes. After the initial blood collection, 0.05 mg of rhTSH was administered IV and a second blood sample was collected 6 hours later for repeat measurement of tT4. Blood was allowed to clot for a minimum of 20 minutes, centrifuged within 1 hour, stored at 4 C to 6 C. Urine was collected by cystocentesis. All samples collected were analyzed within 24 hours at a single local commercial laboratory.

| Data presentation and statistical analysis
All statistical analyses were performed using proprietary statistical software (IBM SPSS Statistics 24). All analyses used nonparametric tests due to small group sizes and inconsistent normal distribution of data. All data are presented as median (interquartile range). Statistical significance was defined as P < .05. The results of the TSH stimulation test are reported as the serum tT4 concentration 6 hours after IV TSH administration (after-stimulation tT4) and as the ratio of the tT4 concentration before and after TSH administration (T4-ratio).

| Control cats
Using the control group data, a RI for after-stimulation tT4 and tT4 ratio was generated using the robust method with Box-Cox transformation of the data by proprietary software. 24 The after-stimulation-tT4 and tT4 ratio were compared between the two control cat groups using the Mann-Whitney U test with a P value <.05 used to indicate significance.

| RAI-treated cats
Continuous variables were compared between groups by the Kruskal-Wallis test, followed by Dunn's multiple comparisons test. Hormone values below the limit of detection, or above the upper threshold, of the assay were assigned arbitrary values for the analysis as follows: TSH <0.03 = 0.02 ng/mL, TSH >12 = 12 ng/mL, fT4 < 0.29 ng/ dL = 0.27 ng/dL. All categorical variables were compared using the Pearson Chi-squared test, including an assessment of the proportion of cats with and without illness, between the three thyroid function groups.

| Control cats
A total of 36 control cats were recruited, 12 from private homes and 24 group shelter-housed. Three cats were diagnosed with hyperthyroidism and excluded, therefore 33 cats were included in study. Many shelter cats were of unknown age but all were estimated to be at least 6 years old. Of the 21 cats whose age was known, the median age was 10 years old. [7][8][9][10][11][12][13][14] The majority of control cats included in the study were Domestic short-or long-haired cats (n = 31) and two purebred cats were  included cats with mild to moderate dental calculus or isolated suspect feline odontoclastic resorptive lesions (n = 10), cats with a grade 2 to 3 heart murmur but no other evidence of heart disease (n = 2), and an apparently healthy FIV positive cat (n = 1).
Control cats in group 1 (noncritically ill: n = 18) included a mildly

| TSH stimulation test results in control group cats
The T4 ratio (P = .46) and after-stimulation-tT4 (P = .38) were not significantly different between the two control cat groups ( Table 1).
The control cat data were used to derive a RI minimum afterstimulation tT4 concentration (2.3 μg/dL) and a reference minimum tT4 ratio (1.5) after rhTSH administration. These values were used to determine the euthyroid cut-off points for the RAI-treated cats.

| TSH stimulation test criteria for RAItreated cats
After administration of rhTSH, the control group cats had a higher median after-stimulation tT4 (

| RAI-treated cats
A total of 118 RAI-treated cats were recruited including 106 domestic short-or long-haired cats and 12 purebred cats: all cats were neutered (female n = 64, male n = 54). Twenty cats were receiving medication F I G U R E 1 Boxplot demonstrating the concentration of total thyroxine (tT4), 6 hours after intravenous rhTSH administration, in 4 groups of cats (after-stimulation tT4). The 4 groups are RAI-Euthyroid (RAI-treated cats with good response to rhTSH administration n = 71), RAI-Equivocal (RAI-treated cats with poor response to rhTSH administration n = 21), RAI-Hypothyroid (RAItreated cats with low or low normal tT4 concentration and no, or inadequate, response to rhTSH administration n = 26) and Controls (euthyroid mature adult cats n = 33). The box in the boxplot represents the interquartile (IQ) range, the horizontal bar represents the median value. The whiskers represent the data outside the IQ range but excluding individual values >1.5 boxlengths from the upper and lower edges of the box. Outliers >1.5 box-lengths from the edges of the box are represented by stars. RAI, radioiodine-I 131 ; rhTSH, recombinant human thyroid stimulating hormone F I G U R E 2 Boxplot demonstrating the ratio of total thyroxine concentration (tT4 ratio), before and 6 hours after intravenous rhTSH administration, in 4 groups of cats. Control cats had a significantly higher (P < .001) T4 ratio compared to the radioiodine treated cats. The groups are as described in Figure 1 legend. See Figure 1

| Relationship between thyroid status and radioiodine treatment
The median [IQ range] dose of RAI received by the RAI-treated cats was 3.5 [2.5-4.1] mCi (Table 2) and the majority of cats were included in the study at least 12 months (422.5 [269-697] days) after RAI treatment. In cases where two RAI treatments were given, the second dose is included in the data set and the date of the second treatment used to calculate time after treatment. The median tT4 concentration at diagnosis (before RAI treatment) in the RAI-treated cats was 9.1 (6.1-12.7) μg/dL. There were no significant differences in the radioiodine dose received (P = .41), highest before-treatment tT4 (P = .83), or time interval between RAI treatment and TSH stimulation test (P = .22), between the three groups ( Table 2).
F I G U R E 3 Scatterplot of baseline total thyroxine and endogenous TSH concentrations in RAI-treated cats divided into three groups based on response to exogenous rhTSH (groups described in Figure 1.). The horizontal dotted line represents the lower end of the laboratory reference range for tT4 and the vertical dotted line represents a previously published 6 reference cut-off for TSH in cats (0.3 ng/mL). The gray box depicts the diagnostic area for hypothyroidism if both tT4 and TSH concentrations were to be used for diagnosis, and using the published reference cut-offs. RAI, radioiodine-I 131 ; rhTSH, recombinant human thyroid stimulating hormone; TSH, thyroid stimulating hormone T A B L E 4 Proportion (number) of RAI-treated cats with fT4, and tT4 below the laboratory reference range, and TSH > 0.3 ng/mL when divided into RAIhypothyroid, RAI-equivocal, and RAIeuthyroid groups using the TSH stimulation test results F I G U R E 4 Scatterplot of baseline free thyroxine and endogenous TSH concentrations in RAI-treated cats divided into three groups based on response to exogenous rhTSH (groups described in Figure 1).

| Relationship between thyroid status and SBP
Cats receiving amlodipine treatment at the time of examination were not included in the analysis of blood pressure (n = 4; all in the RAIeuthyroid group). There was no significant difference in SBP between the three RAI-treated groups; however, SBP was higher in the RAIhypothyroid group compared to the control group (P = .02). stimulation test is a test of thyroid reserve and it is apparent from our data set that the majority of RAI-treated cats had a decreased thyroid reserve compared to untreated euthyroid cats. As with all endocrine diseases, it is hard to place defined cut-points into a system where many individuals fall into a gray zone between clearly normal and clearly abnormal, and we therefore included a 3rd group of cats with "equivocal" function. This approach has been used previously, for example, in studies of thyroid function in dogs 15 and in a recent study of diagnostic tests for pancreatitis in cats. 28 The inclusion of this third group of cats meant that a calculation of sensitivity and specificity for thyroid hormones as diagnostic tests for hypothyroidism could not be performed in our study.

| DISCUSSION
Our study did not find a difference in response to TSH between control cats with noncritical illness and relatively healthy controls, or between the different groups of RAI-treated cats, which concurs with the findings of a previous study. 18 As noted above, sick euthyroidism can lead to a decreased response to exogenous TSH administration, with the potential for a false positive diagnosis of hypothyroidism. 14,29 It was therefore important for us to determine if the suppressed response to TSH in our study was due to concurrent illness, before we attributed the results to a depressed thyroid reserve. There were a wide variety of different types of concurrent illnesses in our study cats, which is common in senior cat populations. Although we were able to apply objective measures of illness in some cases (eg, blood pressure and weight loss), in some cases the categorization of cats as "ill" and "not ill" somewhat subjective (eg, the assessment of dental disease). It is important to note that our recruitment strategy resulted in the inclusion of a relatively healthy population of cats and none of our control cats had subnormal tT4 despite their illnesses. Therefore, our conclusion that noncritical illness had no effect on the results of the TSH stimulation test cannot be extrapolated to sick euthyroid cats with subnormal tT4 concentrations.
High TSH (>0.3 ng/mL) has a 98.5% specificity and a 100% sensitivity for the diagnosis of hypothyroidism 6  The non-ED veterinary free T4 assay used in this study correlated better with a diagnosis of hypothyroidism compared to the ED fT4 assay used in the Peterson study 6 . In the Peterson study, low fT4 (below the laboratory RI) had a sensitivity and specificity of 25% and 98% respectively, whereas in our study, 88% of RAIhypothyroid cats had a fT4 below the laboratory RI. In a separate study, which used the same method and fT4 cut-point (0.7 ng/dL) for diagnosis of hypothyroidism as our study, 41% of euthyroid cats had a fT4 below the laboratory RI, whereas in our study 7% of RAI-euthyroid and 33% of RAI-equivocal had a low fT4. 13 These data suggest that the non-ED veterinary fT4 assay used in our study (and the study by Stameleer et al) gives fewer false negative results (is more sensitive) but more false positive results (is less specific) for iatrogenic hypothyroidism in cats than the assay used in the Peterson study. This could reflect differences in assay calibration or assay sensitivity that leads to lower measured fT4 concentrations with our assay. This supposition is supported by our data set in which >25% of RAI-hypothyroid cats had a fT4 below the limit of detection, whereas most of these cats had detectable tT4 concentrations. Although a high correlation between the non-ED fT4 assay used in our study and tT4 has been reported, 13 the aforementioned study did not perform a Bland Altman analysis to test for systematic differences between these variables. Therefore, it is possible that the non-ED fT4 assay used in both the aforementioned study and our study consistently underestimates the fT4, which could account for the lower false negative and higher false positive rate observed in our study compared to the Peterson study. In addition, only one of the 71 RAI-euthyroid cats had both a low fT4 (0.55 ng/dL) and a high TSH (0.32 ng/mL), which suggests that combining fT4 and TSH measurements might decrease the risk of a false positive hypothyroidism diagnosis compared to using fT4 or TSH alone.
The non-ED fT4 assay is a veterinary specific chemilumines- In our study, baseline tT4 was lower and response to rhTSH administration was decreased in the euthyroid control cats compared to previous studies. [17][18][19] It is possible that these differences in baseline T4 and response to exogenous TSH could be due to the much younger age of the cats in the previous studies (7-8 months and 2 years, respectively), compared with the median age of 9 [7-10.5] years old of the control cats in our study. There is conflicting evidence that tT4 concentrations change with age in cats, with an effect of age being reported in some studies 30 but not in others. 9 In our prospective study, iatrogenic hypothyroidism was common in RAI-treated cats: 22% of recruited cats were diagnosed with hypothyroidism, with a further 18% of cats in the RAI-equivocal group demonstrating a subnormal response to rhTSH administration. The high prevalence of elevated endogenous TSH in our study (37%), alongside within-RI tT4 concentrations, supports the conclusion of a previous study that "subclinical" hypothyroidism appears to be common in RAI-treated cats. 6  In our study, SBP was significantly higher in RAI-hypothyroid cats compared to control cats. In humans, naturally occurring hypothyroidism is associated with an increased prevalence of hypertension, with treatment of hypothyroidism leading to a reduction in systolic and diastolic blood pressure. 32 The pathophysiological mechanism for hypertension in hypothyroidism is unclear; however, it is postulated that hypertension could be secondary to increased total peripheral resistance, increased β-adrenergic responsiveness, or extracellular fluid volume expansion. 32

CONFLICT OF INTEREST DECLARATION
In-kind support of this study was provided by IDEXX Laboratories in the form of no-charge laboratory testing: all blood and urine samples collected during the course of the study were analyzed at IDEXX laboratories at no charge. IDEXX Laboratories staff were not involved in the study design, data collection, interpretation of data, or the writing of the report (other than laboratory methods section).

OFF-LABEL ANTIMICROBIAL DECLARATION
Authors declare no off-label use of antimicrobials.

INSTITUTIONAL ANIMAL CARE AND USE COMMITTEE (IACUC) OR OTHER APPROVAL DECLARATION
This work involved the use of non-experimental (owned) animals only.
Ethical approval for the study was granted by the Douglas College Institutional Animal Care Committee (accredited by the Canadian Council on Animal Care). Informed Consent was obtained in writing from the owner or legal custodian of all animals described in this work for the procedures undertaken. No animals or humans are identifiable within this publication, and therefore additional Informed Consent for publication was not required.