Reversible antioxidant depletion is found in hyperthyroid humans, and antioxidant depletion increases the risk of methimazole toxicosis in rats.
Reversible antioxidant depletion is found in hyperthyroid humans, and antioxidant depletion increases the risk of methimazole toxicosis in rats.
To determine whether abnormalities in concentrations of blood antioxidants or urinary isoprostanes were present in hyperthyroid cats, and were reversible after radioiodine treatment. To determine whether or not antioxidant abnormalities were associated with idiosyncratic methimazole toxicosis.
Hyperthyroid cats presented for radioiodine treatment (n = 44) and healthy mature adult control cats (n = 37).
Prospective, controlled, observational study. Red blood cell glutathione (GSH), plasma ascorbate (AA), plasma free retinol (vitamin A), α-tocopherol (vitamin E), and urinary free 8-isoprostanes in hyperthyroid cats were compared to healthy cats and to hyperthyroid cats 2 months after treatment.
Blood antioxidants were not significantly different in hyperthyroid cats (mean GSH 1.6 ± 0.3 mM; AA 12.8 ± 4.9 μM, and vitamin E, 25 ± 14 μg/mL) compared to controls (GSH 1.4 ± 0.4 mM; AA 15.0 ± 6.6 μM, and vitamin E, 25 ± 17 μg/mL). Urinary isoprostanes were increased in hyperthyroid cats (292 ± 211 pg/mg creatinine) compared to controls (169 ± 82 pg/mg; P = .006), particularly in hyperthyroid cats with a USG < 1.035. Plasma free vitamin A was higher in hyperthyroid cats (0.54 ± 0.28 μg/mL versus 0.38 ± 0.21 in controls; P = .007). Both abnormalities normalized after radioiodine treatment. No association was found between oxidative status and prior idiosyncratic methimazole toxicosis.
Increased urinary isoprostane could reflect reversible renal oxidative stress induced by hyperthyroidism, and this requires additional evaluation.
urine specific gravity
University of Wisconsin Veterinary Medical Teaching Hospital
Hyperthyroidism is the most common endocrine disorder in cats over 8 years of age, and shares many clinical features with hyperthyroidism caused by Graves' disease and toxic nodular goiter in humans. Methimazole is the most common drug used to treat hyperthyroidism in cats in the United States, particularly when radioiodine is not readily available or is cost prohibitive. Dose-dependent gastrointestinal upset occurs in about 10% of cats, and idiosyncratic neutropenia, thrombocytopenia, facial excoriation or hepatotoxicosis occurs in 2–7% of treated cats. These latter idiosyncratic reactions require drug discontinuation, can occasionally necessitate hospitalization, and contraindicate future use of methimazole in that individual. The mechanisms of these idiosyncratic adverse reactions are not understood, and there are no known predictors or preventative measures.
Antioxidant depletion and oxidative stress are considered risk factors for several idiosyncratic drug toxicity syndromes in humans and animal models, to include sulfonamide skin rash, azathioprine hepatotoxicosis, and clozapine-induced neutropenia. Most notably, mice that are depleted of GSH are at higher risk of developing methimazole hepatotoxicosis compared to antioxidant-replete mice.
Oxidative stress is well documented in human patients and experimental animals with hyperthyroidism.[8-12] Markers of oxidative stress normalize after treatment and a return to a euthyroid state.[11, 13, 14] For example, patients with toxic multinodular goiter have reduced plasma concentrations of vitamin C and E, and hyperthyroid women have lower serum concentrations of retinol (vitamin A). Patients with Graves' disease have decreased total plasma thiols and increased thiobarbituric-reacting substances (TBARS; a measure of oxidative stress), which resolve with treatment[8, 11] In addition, rats treated with supraphysiologic concentrations of thyroxine develop decreased hepatic GSH concentrations, indicating that high thyroxine concentrations directly contribute to oxidative stress.
Oxidative stress has been found in association with various diseases in cats, including liver disease, FIV infection, steatitis, and acetaminophen toxicity. In a survey of cats with heart disease, a small group of 11 cats with concurrent hyperthyroidism showed trends toward decreases in plasma vitamin E and increases in vitamin A concentrations, but this did not reach statistical significance. There was a significant increase in plasma ascorbate concentrations in hospitalized cats with a variety of illnesses. Only 2 cats in that study were hyperthyroid, so a valid evaluation of markers of oxidative stress in hyperthyroid cats is lacking.
The primary aim of this study was to determine whether or not measures of oxidative stress were present in cats with hyperthyroidism compared to mature adult control cats, and if present, whether or not these were reversible with radioiodine treatment. A secondary aim was to determine whether or not individual markers of oxidative stress were associated with idiosyncratic adverse reactions to methimazole in hyperthyroid cats.
Client-owned hyperthyroid cats that presented to the UW VMTH for radioiodine treatment between January 2010 and April 2011 were recruited for the study. Cats with clinical signs of hyperthyroidism (weight loss, polyuria, polydipsia, polyphagia, vomiting, or diarrhea), a palpable goiter, and a serum total T4 >4.8 μg/dL (reference range: 1.9–4.8 μg/dL) were eligible for enrollment. Cats with concurrent heart failure, symptomatic renal failure (IRIS stage 3 or 4, creatinine ≥2.9 mg/dL; www.iris-kidney.com), systemic neoplasia, chronic liver disease, immune-mediated disease, or systemic infection, any of which could influence antioxidant status independently of thyroid status, were excluded. Those hyperthyroid cats that had been treated with methimazole and had previously developed idiosyncratic toxicity (facial excoriation, thrombocytopenia, neutropenia, or hepatopathy) were identified. At the time of evaluation, all cats were required to be clinically and biochemically hyperthyroid, had not received methimazole treatment for at least 1 week, and if relevant, were completely recovered from any methimazole adverse reactions. For the control group, otherwise healthy euthyroid cats ≥5 years of age were recruited from the UW VMTH Primary Care service and from pets belonging to the VMTH staff. Age, breed, sex, neuter status, body weight, body condition score, current diet, and drug treatments were recorded for all cats. Informed consent was obtained from all cat owners, and the study was approved by the University of Wisconsin-Madison Institutional Animal Care and Use Committee.
Both control and hyperthyroid cats were screened with a physical exam, CBC, biochemical panel, urine specific gravity (USG), blood pressure, and serum total T4. Additional blood in heparin (3 mL) was also obtained for antioxidant assays, along with urine (>2 mL) for urinary free 8-isoprostane concentrations. Urine was refrigerated immediately; butylated hydroxytoluene (BHT) in 95% ethanol was added to each 2 mL aliquot of urine, to a final concentration of 0.005% BHT. Urine was frozen in aliquots at −80°C until shipping for isoprostane analyses. Blood tubes were covered in foil, refrigerated immediately, and centrifuged within 3 hours of phlebotomy. Red blood cells (RBC) were harvested for glutathione determinations, and plasma was collected and aliquoted for vitamin E, vitamin A, and ascorbate measurements. Samples were frozen at −80°C under dark conditions to prevent degradation of vitamins E and A. Ascorbate and GSH samples were analyzed within 1 week of sample collection to assure stability.[21, 22] Samples from control cats were recruited and analyzed concurrently with the hyperthyroid cats.
Erythrocyte reduced glutathione was measured using a bromobimane-based HPLC method with fluorescence detection, and plasma reduced ascorbate was measured by HPLC with UV detection, as previously validated for cats in our laboratory. Plasma vitamin E (α-tocopherol) and free vitamin A (retinol) concentrations were measured at the Wisconsin Veterinary Diagnostic Laboratory, using HPLC methods previously validated in cats. Urine free 8-isoprostane concentrations were quantified by a commercial laboratory,1 using affinity column purification followed by a competitive enzyme immunoassay (8-Isoprostane Express EIA kit),1 and were normalized to urine creatinine, as previously reported in cats. Urinary isoprostanes from hyperthyroid cats were run in batches concurrently with control cats.
After evaluation and sample collection, hyperthyroid cats were treated routinely with a single SC dose of radioiodine, with dosing based on serum T4, estimated goiter size, and severity of clinical signs. Cats were returned 2 months post treatment for routine re-evaluation and antioxidant measurements. Body weight, body condition score, status of clinical signs, a limited biochemical panel, USG, blood pressure, and serum T4, as well as erythrocyte GSH, plasma ascorbate, vitamins E and A, and urinary isoprostane concentrations, were again evaluated.
A sample size calculation predicted that 35 cats in each group would provide >95% power to detect a 2.5-fold difference in plasma ascorbate between healthy and hyperthyroid cats, as we had observed in cats with various illnesses compared to healthy controls, and 92% power to detect, as significant, a 30% decrease in vitamin E between groups, based on data in healthy cats and those with heart disease. All parameters were checked for normal distribution using a scatter plot and comparison of means and medians, before applying parametric tests. Parameters were compared between healthy and hyperthyroid cats at baseline using an unpaired t-test; parameters among healthy, hyperthyroid cats at baseline, and hyperthyroid cats after radioiodine treatment were compared using ANOVA with Tukey's posthoc test, with P < .05. ANOVA was also used to compare antioxidant parameters among cats without adverse reactions to methimazole, cats with simple GI upset, and cats with idiosyncratic toxicity. Correlations between variables were assessed using a Pearson's correlation test.
Fifty-one hyperthyroid cats were evaluated; 7 of these cats were not eligible for enrollment (5 had serum T4 values within the normal range, and 2 had disqualifying concurrent disease, including heart failure and stage 3 IRIS renal disease). Therefore, 44 hyperthyroid cats were enrolled. Post-treatment data were obtained for 35 hyperthyroid cats (5 cats did not receive radioiodine treatment, 2 cats died before the 2-month recheck, and 2 cats were lost to follow-up). All but 2 cats were euthyroid at follow-up; both cats had a reduction in serum T4 and were included in analyses. A total of 38 healthy mature control cats were screened, and 37 were eligible based on normal blood work and physical exam. Domestic short- and long-hair breeds made up more than 95% of both groups of cats. Although mature adult cats were purposefully recruited as controls, the healthy cats were significantly younger (mean age of 8.8 years [range: 5–17] than the hyperthyroid cats (11.8 years [range: 6–17], P < .0001; Table 1).
|Control (n = 37)||HT before Treatment (n = 44)||HT after Treatment (n = 35)|
|BW (kg)||5.0 ± 1.5 (3.1–9.8)||4.1 ± 0.9a (2.6–6.1)||4.4 ± 1.0 (2.8–6.6)|
|BCS (scale of 1–9)||5.9 ± 1.0 (5–8)||4.7 ± 1.3a (1–7)||5.2 ± 1.3b (2–7)|
|Age (yrs)||8.8 ± 2.7 (5–17)||11.8 ± 2.0a (6–17)||NA|
|T4 (μg/dL)||3.2 ± 0.6 (2.1–4.5)||11.5 ± 6.3a (4.9–25.8)||2.6 ± 1.1b (0.9–6.7)|
|BUN (mg/dL)||26.5 ± 5.1 (18–39)||25.1 ± 6.9 (12–42)||27.8 ± 9.2 (16–63)|
|Creatinine (mg/dL)||1.4 ± 0.4 (0.9–2.2)||1.3 ± 0.5 (0.5–2.9)||1.7 ± 0.5c (1.1–3.2)|
|USG||1.041 ± 0.003 (1.035–1.050)||1.035 ± 0.010a (1.011–1.053)||1.031 ± 0.010a (1.012–1.040)|
Hyperthyroid cats had significantly lower body weights and body condition scores compared to controls (Table 1). Hyperthyroid cats also had more dilute urine (mean: 1.036) compared to mature healthy cats (mean: 1.041, P = .001), but showed no differences in serum urea nitrogen or creatinine concentrations. After radioiodine treatment, USG did not decrease significantly in hyperthyroid cats (mean: 1.031); however, serum creatinine concentrations rose modestly (mean: 1.7 mg/dL, range: 1.1–3.2) compared to before treatment (1.3 mg/dL, range: 0.5–2.9; P = .001; Table 1).
As for blood antioxidants, erythrocyte-reduced GSH, plasma ascorbate, and plasma free α-tocopherol concentrations were not different between healthy and hyperthyroid cats, or before and after treatment in hyperthyroid cats (Table 2). However, a significant increase in plasma vitamin A (free retinol) concentrations was observed in hyperthyroid cats (mean: 0.54 μg/mL, range: 0.14–1.30) compared to that in healthy cats (0.38 μg/mL, range: 0.14–1.30; P = .006) (Fig 1). In addition, free retinol concentrations were correlated modestly but significantly with serum T4 at baseline in hyperthyroid cats (r = 0.32, P = .04). Free retinol concentrations decreased after radioiodine treatment, when they were no longer significantly different from controls (0.43 μg/mL, range: 0.15–0.96; P = .04; Fig 1).
|Control (n = 37)||HT Pretreatment (n = 44)||HT Post-treatment (n = 35)|
|Reduced GSH (mM)||1.4 ± 0.4 (0.8–2.0)||1.6 ± 0.3 (1.0–2.3)||1.7 ± 0.3 (1.1–2.7)|
|Plasma ascorbate (μM)||15.0 ± 6.6 (4.2–32.1)||12.8 ± 4.9 (5.5–27.7)||16.3 ± 5.8 (8.3–30.8)|
|Vitamin E (μg/mL)||25 ± 17 (5–79)||25 ± 14 (5–86)||26 ± 16 (6–66)|
Urinary free 8-isoprostanes (normalized to urine creatinine) were significantly elevated in hyperthyroid cats (292 pg/mg, range: 64–1,055) compared to healthy mature cats (169 pg/mg, range: 56–391; P = .003), and these were not significantly different from those in normal cats after radioiodine treatment (220 pg/mg, range: 48–581; P = .07; Fig 2A). When hyperthyroid cats were categorized by USG before treatment, urinary isoprostanes were significantly higher in hyperthyroid cats with dilute urine (USG < 1.035; mean isoprostanes 509 pg/mg, range: 108–1,055) compared to hyperthyroid cats with a USG ≥ 1.035 (mean isoprostanes 231 pg/mg, range: 64–567; P < .001, Fig 2B). In addition, USG was significantly and inversely correlated with urinary isoprostanes across hyperthyroid cats (r = −0.46, P = .006; Fig 2C). In this population, however, urinary isoprostanes were not significantly different at baseline among 19 cats that showed an increase in IRIS stage after treatment (356 pg/mg, range: 93–1,055) compared to 16 cats with stable IRIS stage 1 or 2 renal disease (259 pg/mg, range: 76–843; P = .21).
Thirty of the 44 hyperthyroid cats had been treated previously with methimazole, and had discontinued the drug a median of 12.5 days before evaluation (range: 7–365 days). Of the methimazole-treated cats, only 3 cats had been off methimazole for more than 30 days, whereas 19 cats had been off of methimazole for <14 days, and 8 cats had methimazole discontinued 15–30 days before referral. Because duration of methimazole discontinuation could affect the development of antioxidant abnormalities, GSH, vitamin C, vitamin E, and isoprostanes were also reanalyzed in hyperthyroid cats based on duration of methimazole discontinuation, using the above categories; however, there were still no significant differences among groups or between any 2 groups compared to healthy cats (data not shown).
Of the 30 cats that had been treated with methimazole, 13 cats were tolerant of methimazole, 7 cats had simple gastrointestinal upset, and 10 cats had documented prior idiosyncratic adverse reactions (7 with facial pruritus, 2 with blood dyscrasias, and 1 with a hepatopathy; Table 3). Antioxidant parameters were not significantly different among these groups of cats, with no apparent trends noted.
|Methimazole Tolerant (n = 13)||Idiosyncratic Reaction (n = 10)||P Value|
|Reduced GSH (mM)||1.5 ± 0.3 (1.2–2.2)||1.6 ± 0.3 (1.2–2.1)||.67|
|Plasma ascorbate (μM)||11.9 ± 4.4 (6.8–22.7)||12.9 ± 4.8 (5.5–21.0)||.62|
|Vitamin E (μg/mL)||20 ± 14 (5–52)||25 ± 11 (15–47)||.35|
|Vitamin A (μg/mL)||0.37 ± 0.19 (0.17–0.72)||0.55 ± 0.21 (0.14–0.87)||.05|
|Urinary isoprostanes (pg/mg)||251 ± 224 (93–843)||366 ± 192 (98–621)||.26|
Unlike what is observed in humans with hyperthyroidism, we did not find a comparable GSH deficiency in hyperthyroid cats in the current study. This apparent species disparity could be attributed to differences in the pathogenesis of hyperthyroidism in cats and humans. However, we have previously found that cats with a wide variety of illnesses have normal blood GSH concentrations, which is distinct from blood GSH deficiencies seen in hospitalized dogs and humans.[21, 26, 27] Therefore, there might be a broader species difference in the way that cats respond to diseases that typically produce oxidative stress in blood cells. None of these studies have examined GSH concentrations in hepatic or other tissues, which might help to understand this observed difference. In addition, measurement of oxidized glutathione (GSSG) might have been helpful in our study. However, GSSG can be falsely elevated because of ex vivo oxidation, which leads to spuriously low GSH : GSSG ratios.[28, 29] Ex vivo GSH oxidation was prevented by bromobimane in our assay; however, bromobimane also independently interfered with recovery of GSSG as GSH in our initial assay validation experiments in feline erythrocytes (data not shown), so we were not comfortable reporting GSSG in our analyses.
We also compared plasma reduced ascorbate (vitamin C), free α-tocopherol (vitamin E), and free retinol (vitamin A) concentrations in hyperthyroid and mature healthy cats. Ascorbate is an important intracellular and extracellular antioxidant that, among other important roles, recycles glutathione and vitamin E to their reduced and functional states.[30, 31] Hyperthyroid humans have decreased plasma ascorbate concentrations, and this partially normalizes with treatment. Conversely, we previously found unexpectedly increased plasma ascorbate concentrations in a heterogeneous group of hospitalized cats. However, in the current study of hyperthyroid cats, plasma ascorbate concentrations were not different compared to healthy mature cats. These findings suggest that ascorbate supplementation is not indicated in this population. We also attempted to measure oxidized ascorbate (dehydroascorbate; DHA) as a measure of oxidative stress. However, we found that, unlike our experience in human plasma, DHA was unstable in feline plasma under the conditions of our assay.
We also examined plasma vitamin E concentrations in our study population. Plasma vitamin E deficiency has also been documented in hyperthyroid humans, with a trend toward lower concentrations in 11 cats with hyperthyroidism and accompanying heart disease. We found no evidence for plasma vitamin E deficiency in our larger group of cats with hyperthyroidism. However, in the 1 cat with hyperthyroidism that was found to be in heart failure and was censored from eligibility, the plasma vitamin E concentration was very low (9 μg/mL), and was outside the 95% confidence interval for vitamin E concentrations in both hyperthyroid (21–29 μg/mL) and healthy mature cats (19–30 μg/mL). Therefore, additional evaluation of plasma vitamin E status may be warranted in hyperthyroid cats with overt heart failure.
In a previous study, women with hyperthyroidism were found to have decreased concentrations of plasma retinol. We found unexpectedly higher concentrations of plasma vitamin A (as free retinol) in hyperthyroid cats, compared to those in controls. A similar trend was also reported in cats with cardiomyopathy and hyperthyroidism. These higher free retinol concentrations could be caused by competition between retinol and thyroxine for protein binding, which is consistent with the modest positive correlation that we observed between serum T4 and free retinol concentrations. Retinol and T4 are both carried by a complex composed of retinol binding protein (RBP) and transthyretin (a.k.a. thyroxine binding prealbumin) in humans. Both these proteins are also found in cat sera, and have been shown to interact in a complex with thyroxine in vitro.[33, 34] Serum RBP concentrations are similar in hyperthyroid and healthy cats. Therefore, limited availability of RBP for retinol binding in the presence of high serum T4 could explain higher free retinol concentrations in hyperthyroid cats, which resolved with a return to normal serum T4 concentrations.
In addition to blood antioxidants, we measured urinary concentrations of 8-isoprostanes (primarily 8-isoPGF2α), which are eicosanoids that are generated from nonenzymatic lipid peroxidation. The 8-isoprostanes are used as biomarkers of oxidative stress in humans,[35-38] are stable in urine, and have been measured in cats, dogs, horses, and cattle. Urinary isoprostanes have been shown to be elevated in cats that are obese, and in a small group of clinically ill cats, compared to those in healthy controls. In our study, urinary 8-isoprostanes were significantly increased in hyperthyroid cats, even though circulating blood antioxidants were not depleted. More specifically, urinary 8-isoprostanes were significantly higher in hyperthyroid cats with relatively dilute urine, compared to hyperthyroid cats with a USG ≥ 1.035, and urine isoprostanes were correlated inversely with USG. However, urinary 8-isoprostanes returned to control levels in these cats after radioiodine treatment. Although USG is a crude marker of nephron loss, these findings suggest that reversible oxidative stress may be more substantial in hyperthyroid cats with preexisting renal disease. It is not known what proportions of urinary isoprostanes are from renal, versus systemic, origin. Our findings may reflect an increase in renal isoprostane generation in these cats, or could represent systemic isoprostane production from generalized oxidative stress, which was not detected with our blood antioxidant measurements. Future studies examining both urinary and plasma isoprostanes are indicated to address this question.
The major isoprostane measured by our assay, 8-isoPGF2α, is a mediator of renal arteriolar constriction, and could contribute to the pathogenesis of renal damage as well as act as a marker of lipid peroxidation. Although urinary isoprostanes were not significantly higher at baseline between cats that showed an increase in IRIS stage after treatment (356 pg/mg) compared to those with a stable IRIS stage (259 pg/mg), a larger study with greater power is warranted to further evaluate whether or not there is a relationship between urinary isoprostanes and renal outcomes in hyperthyroid cats.
No association was found between antioxidant status and prior idiosyncratic toxicity from methimazole in the hyperthyroid cats. The study design may have been more sensitive to such a risk factor if we had sampled antioxidant status in cats before methimazole treatment, rather than subsequent to the adverse reaction. In addition, because we evaluated cats at the time of referral for radioiodine, many cats had only recently had methimazole discontinued. Although serum T4 concentrations had returned to previous high levels, this may have been an inadequate amount of time for antioxidant deficiencies to recur, had they been present before methimazole treatment. Although we did not find differences in any of the antioxidant measures in cats based on duration of time that methimazole had been discontinued, we cannot completely rule out an effect from duration of methimazole withdrawal.
In conclusion, the results of this study do not support supplementation with GSH precursors, ascorbate, vitamin E, or vitamin A in hyperthyroid cats. However, reversible elevations in urinary 8-isoprostanes were found in hyperthyroid cats, particularly those with minimally concentrated urine at presentation. Additional evaluation of urinary 8-isoprostanes as a potential marker of renal injury in hyperthyroid cats is warranted.
The authors thank Helen Schultz, Terri Gregson, and Angela Daugherty for invaluable assistance with sample collection. Research funded by the Waltham Foundation.
Cayman Chemical Company, Ann Arbor, MI