To assess the influence of thyroid function on natriuretic peptide concentration in hyperthyroid cats before and after treatment.
To assess the influence of thyroid function on natriuretic peptide concentration in hyperthyroid cats before and after treatment.
Serum natriuretic peptide concentration was measured in 61 hyperthyroid cats recruited from first-opinion clinics before and after treatment.
Following successful treatment, total thyroxine, heart rate, systolic blood pressure and packed cell volume all decreased and bodyweight and creatinine concentrations increased. Furthermore, a significant (P < 0·001) decline in NT-proBNP concentration but not NT-proANP was identified.
Thyroid function has a modest but significant effect on NT-proBNP concentration. Thyroid status should be taken into account when interpreting NT-proBNP concentrations in cats.
Natriuretic peptides (NPs) are a family of neurohormones with potent natriuretic and vasorelaxant properties. They are involved in the body's defence against hypertension and plasma volume expansion (Levin and others 1998). The two most studied NPs are atrial natriuretic peptide (ANP) and brain natriuretic peptide (BNP). Under physiological conditions, ANP and BNP are stored in granules within atrial cardiomyocytes and released in response to atrial stretch. NP production is chronically up-regulated in hypertrophied and failing hearts (Yoshimura and others 1993, Luchner and others 2001). Studies across a number of species indicate that in the presence of cardiac disease BNP synthesis is up-regulated in response to chronic increases in atrial and ventricular pressure, and the major site of production switches from the atria to the ventricles (Yasue and others 1994, Luchner and others 2001, Liu and others 2002).
These observations have encouraged the use of NP as biomarkers to detect and assess the severity of heart disease in humans (Mueller and others 2004, Januzzi and others 2005), dogs (Boswood and others 2003, MacDonald and others 2003, DeFrancesco and others 2007, Prosek and others 2007, Boswood and others 2008, Fine and others 2008, Oyama and others 2008) and cats (MacLean and others 2006, Connolly and others 2008, 2009). Because of their longer half-life and better stability, the biologically inactive N-terminal fragments, NT-proBNP and NT-proANP, are now more widely used as biomarkers than the biologically active C-terminal fragments. Studies performed in cats to date have shown that NP can be used to distinguish healthy cats from those with occult heart disease or failure (Connolly and others 2008, Fox and others 2011) and to differentiate cardiac from non-cardiac causes of dyspnoea in the emergency situation (Connolly and others 2009, Fox and others 2009).
Hyperthyroidism is the most common endocrine disease of cats (Peterson and Ward 2007), and thyroid hormones have profound effects on the cardiovascular system (Kahaly and Dillmann 2005, Klein and Danzi 2007). In hyperthyroid cats, cardiovascular signs, including tachycardia, tachypnoea and heart murmurs, are common, and occasionally heart failure is identified (Broussard and others 1995, Syme 2007). Hyperthyroidism causes tachycardia as a result of adrenergic hyper-responsiveness (Carvalho-Bianco and others 2004) and decreased systemic vascular resistance (SVR) (Muenster and others 1959) resulting in increased cardiac output (Biondi and others 2002). One stimulus for decreased SVR is nitric-oxide-induced peripheral vasodilation facilitating adequate oxygen supply to tissues with thyroid-driven increased metabolic demand. In addition, a direct relaxant effect of T3 on vascular smooth muscle cells has been demonstrated in isolated skeletal muscle arteries (Park and others 1997) and in cell culture preparations (Ojamaa and others 1996). These persistent stimuli result in arterial underfilling, which in turn chronically stimulates the renin-angiotensin-aldosterone system (RAAS) leading to increased vascular volume, myocardial stretch and NP release. A number of studies have demonstrated elevated NP concentrations in human hyperthyroid patients (Mori and others 1990, Diekman and others 2001, Schultz and others 2004, Arikan and others 2007, Ertugrul and others 2009 ). The influence of thyroid function on NP concentration in cats is unknown. Because NP measurement is increasingly being used for the diagnosis, prognosis and management of feline cardiac disease (Connolly 2010), information about factors that may affect NP concentration, including hyperthyroidism, is essential.
The purpose of the present study was to test the hypothesis that circulating NP concentration would be increased in hyperthyroid cats and that NP concentration would decrease following restoration of the euthyroid state.
The protocol used in this study was approved by the Ethics and Welfare Committee of The Royal Veterinary College. Informed consent was obtained from all owners of cats involved in the study. Hyperthyroid cats were recruited from three first-opinion clinics from 1999 to 2008, as part of ongoing research on feline geriatric diseases.
A full physical examination, plasma biochemistry, total thyroxine (T4) measurement, packed cell volume (PCV) and urine analysis were carried out in all cats. Systolic blood pressure (SBP) was measured using a Doppler technique as described previously (Jepson and others 2007).
Cats were included in the hyperthyroid group if they had appropriate clinical signs and an elevated total T4 concentration (total T4 > 55 nmol/L; laboratory reference interval 10 to 55 nmol/L). Hyperthyroid cats were treated using methimazole (Felimazole®, Dechra, Shropshire, UK) alone or methimazole and subsequent surgical thyroidectomy. Cats achieving euthyroid status (total T4 <40 nmol/L) within 100 days of initiating treatment were selected for inclusion in the study of response to treatment. Previously treated hyperthyroid cats and cats showing overt signs of heart failure were excluded from the study. Cats with systemic hypertension (SBP > 170 mmHg or >160 mmHg in association with hypertensive retinal lesions) were also excluded because of previous reports that this is associated with increased NT-proBNP concentrations (Lalor and others 2009).
Blood samples were collected from the jugular vein and placed into lithium heparin tubes. Blood was kept on ice or in a refrigerator at 4 °C for no longer than 60 minutes, and then centrifuged at 4 °C for 10 minutes at 3000 rpm. The plasma was separated and frozen at −80 °C until the assays were performed. The samples were sent on dry ice to an external commercial laboratory (Idexx Laboratories, West Yorkshire, UK) to determine NT-proBNP and NT-proANP concentrations. The laboratory technician was blinded to the patient's status. The limit of detection of the assay was 10 pmol/L for NT-proBNP and 50 pmol/L for NT-proANP. Assay validation was performed by using plasma samples from cats with expected low (healthy cats), medium (cats with heart disease confirmed by ultrasound but no heart failure) and high NP concentrations (cats in heart failure). These samples were then used to calculate intra- and inter-assay coefficients of variation (CV).
Statistical analyses were performed using commercially available computerised software (SPSS version 18.0 for Windows, SPSS Inc, USA). Quantitative data were assessed graphically for normality and are reported as median (25th, 75th percentile). A Mann–Whitney test was used to compare NT-proBNP concentration between hyperthyroid cats with and without auscultated abnormalities (murmurs, arrhythmia, gallop rhythm) or tachypnoea. Paired data (NPs and clinical variables) before and after treatment of hyperthyroidism were compared with the Wilcoxon signed-rank test. Univariable linear regression was used to evaluate potential associations between heart rate, SBP, age, creatinine, total T4 and bodyweight and the square root of NT-BNP concentration in the hyperthyroid cats at baseline. The assumption of linearity was evaluated by plotting the residuals against each independent variable. Residuals were assessed for normality by visual inspection of histograms and probability plots, and assessed for constant variance by visual inspection of plots of residuals against predicted values. For each variable, Cook's distance was plotted against the centred leveraged values. If individual data points appeared to exert excessive influence, then the analysis was repeated with that value removed from the analysis.
Results were considered statistically significant when the P value was less than 0·05.
Assay validation: The intra-assay coefficient of variation (CV) of the NT-proANP assay (n = 6) was 14·7, 4·2 and 7·3% and the inter-assay CV (n = 3) was 28·8, 10·7 and 11·6% for samples with low, medium and high concentrations, respectively. The intra-assay CV of the NT-proBNP assay (n = 6) was 3·4 and 6·2% and the inter-assay CV (n = 3) was 12·5 and 26·7% for samples with medium and high NP concentrations, respectively. Samples with undetectable NT-proBNP (<10 pmol/L) concentrations were consistently undetectable (n = 6 for intra-assay, n = 3 for inter-assay).
Baseline data: Eighty-five hyperthyroid cats were enrolled in the study with a total T4 concentration of 119·8 (89·6, 172·5) nmol/L. The average age, which was known for 62 cats, was 14·1 (12·0, 16·1) years. Heart rate at the time of diagnosis of hyperthyroidism was 216 (188, 240) beats per minute. Many (38·8%) of the hyperthyroid cats in this study had NT-proBNP concentrations that are considered highly suggestive for the presence of underlying cardiac disease according to the current manufacturer's (Idexx laboratories, Westbrook, USA) recommendations (Table 1). In total, 48 of 85 (56%) hyperthyroid cats had an auscultated cardiac abnormality (murmur in 27, gallop in 10 and arrhythmia in 4, with 6 cats having two abnormalities). However, NT-proBNP and NT-proANP were not significantly different in cats with and without abnormal heart sounds. Five hyperthyroid cats were tachypnoeic on initial examination with a respiratory rate greater than 40 breaths per minute and intermittent open mouth breathing when handled (three of these cats also had a murmur and one had an arrhythmia). These cats had significantly higher NT-proBNP concentrations than the hyperthyroid cats that were eupnoeic [754 (509, 892) versus 182 (51, 321) pmol/L; P = 0·003] although the tachypnoeic cats also had higher total T4 concentrations [200 (139, 288) versus 119 (86, 167) nmol/L; P = 0·024], which may have confounded the analysis. There was no difference in NT-proANP concentrations between tachypnoeic and eupnoeic hyperthyroid cats [0·51 (0·42, 1·61) versus 0·65 (0·34, 1·09) nmol/L; P = 0·218]. Tachypnoeic cats did not receive any additional treatment other than that required for control of the hyperthyroid state and their respiratory rate normalised once they had acclimatised to the hospital environment. In univariable linear regression analysis, only total T4 (R = 0·314, P = 0·003) was associated with the square-root NT proBNP concentration, so multi-variable analysis was not performed. Variables not associated with the square root of NT proBNP concentration were age (R = 0·088, P = 0·433), creatinine (R = 0·205, P = 0·071), heart rate (R = 0·05, P = 0·662), bodyweight (R = 0·203, P = 0·069) and SBP (R = 0·082, P = 0·458). The association between square root of NT proBNP and creatinine was also assessed with the exclusion of one cat because of the latter's extreme creatinine value (394 µmol/L); however, the results remained non-significant (R = 0·145, P = 0·209).
|Cut-point NT-proBNP (pmol/L)||Manufacturer's interpretation||n (%) All cats at baseline (n = 85)||n (%) Cats with follow-up (n = 61)|
|<100||Clinically significant cardiomyopathy is highly unlikely||29 (34·1)||20 (32·8)||36 (59·0)|
|100 to 270||Clinically significant cardiomyopathy is unlikely but early disease may be present||23 (27·1)||17 (27·6)||17 (27·6)|
|>270||Clinically significant cardiomyopathy is highly likely||33 (38·8)||24 (39·3)||8 (13·1)|
Response to treatment: Data from 61 cats were available for evaluation of the response to treatment. The length of time between pre- and post-treatment evaluations was 35 (28, 70) days and the length of time that the cats were on treatment before the post-treatment sample was 28 (22, 42) days. Total T4, heart rate, SBP and PCV all decreased and bodyweight and creatinine concentrations increased with treatment as detailed in Table 2. Establishment of euthyroidism was associated with a statistically significant decline in NT-proBNP (P < 0·001), but no change in NT-proANP (P = 0·136) concentration as shown in Fig 1a and b.
|Total thyroxine (nmol/L)||133·0 (86·7, 174·0)||11·9 (4·4, 19·7)||61||<0·001|
|Creatinine (µmol/L)||99·8 (83·3, 126·7)||136·7 (108·0, 162·0)||54||<0·001|
|Heart rate (min−1)||216 (192, 240)||180 (156, 192)||52||<0·001|
|Bodyweight (kg)||3·27 (2·92, 3·80)||3·42 (3·02, 3·94)||58||<0·001|
|SBP (mmHg)||155 (138, 167)||146 (137, 159)||60||0·006|
This is the first study to demonstrate the influence of thyroid function on circulating NT-proBNP concentration in cats. Restoration of the euthyroid state in a large cohort of hyperthyroid cats resulted in a significant decrease in NT-proBNP concentration. These findings reflect three previous studies in human patients which identified a correlation between thyroid function and serum NT-proBNP concentration (Schultz and others 2004, Arikan and others 2007, Ertugrul and others 2009). Additionally, successful treatment of hyperthyroidism in these patients resulted in a significant decrease in serum NT-proBNP concentration (Schultz and others 2004, Ertugrul and others 2009), and, importantly, it was established that this decrease was a direct result of the normalisation of thyroid status and could not be explained by other confounders that could influence NP concentration.
Two hypotheses have been proposed to explain the relationship between NP concentration and thyroid status. Evidence from rat models indicates that thyroid hormones influence NP concentration by directly modulating transcription of the peptides (Liang and others 2003). Alternatively, thyroid function may indirectly affect NP concentrations by altering cardiovascular haemodynamics, specifically by decreasing peripheral vascular resistance and increasing plasma volume and cardiac output (Klein and Danzi 2007). The present study was not designed to distinguish between these two hypotheses and it is possible that NP concentration is influenced both directly and indirectly by thyroid function.
The findings from this study have important clinical implications. The effect of thyroid function on NP concentration should be considered when using NT-proBNP concentration in the assessment of heart disease in elderly cats. For instance, a raised NT-proBNP concentration in an elderly cat with consistent tachycardia might be partly or completely due to the direct and indirect manifestations of thyroid hyperactivity on NP concentration rather than because of primary myocardial disease. Therefore, verification of thyroid status would be indicated especially if NT-proBNP was being used as the sole method of diagnosing cardiac disease in such a case.
In approximately 70% of human patients with Grave's disease NT-proBNP concentration normalises following radioactive iodine treatment (Kato and others 2009). Likewise in this study, NT-proBNP concentration decreased significantly after treatment; however, in many cats (41%) it did not decline to the concentration (<100 pmol/L) that the manufacturer's guidelines associate with a low risk of heart disease. One explanation is that a proportion of the hyperthyroid cats may also have had concurrent primary cardiac disease. Alternatively, despite successful restoration of the euthyroid state, there may have been insufficient time between pre- and post-treatment sampling for complete reverse remodelling to occur, or in chronically affected cats complete remodelling may not be possible.
The results of this present study indicate that thyrotoxicosis has little effect on NT-proANP concentration, and comparable findings have also been reported in human patients (Kato and others 2009). Similarly, previous studies in cats have also shown that NT-proANP is less sensitive than NT-proBNP for the assessment of primary myocardial disease (Connolly and others 2008) or in its ability to distinguish cardiac from non-cardiac causes of respiratory distress (Connolly and others 2009).
Other clinical variables such as heart rate, creatinine concentration and blood pressure were significantly affected by treatment for thyrotoxicosis and mirrored changes previously described in other feline studies (Broussard and others 1995, Syme and Elliott 2003, Syme 2007). Creatinine concentrations increased significantly following treatment, and this finding is in agreement with several other reports that have demonstrated a decrease in glomerular filtration rate following successful treatment for hyperthyroidism (Adams and others 1997, Connolly and others 2005, Boag and others 2007). It is interesting to note that despite an increase in serum creatinine concentration following successful treatment for hyperthyroidism, the NT-proBNP concentration decreased significantly. This suggests that renal function exerts minimal effect on NT-proBNP concentration in cats, which is in contrast to that which has been described in human and canine patients (Mark and others 2006, Boswood and others 2008, Raffan and others 2009, Schmidt and others 2009). A recent study in cats with normotensive or hypertensive chronic kidney disease (CKD) indicated that for cats with mild and moderate CKD, hypertension had a markedly greater influence on NT-proBNP concentration than creatinine concentration (Lalor and others 2009). In the present study SBP was not associated with NT-proBNP concentration, probably because very few of the cats that were included had significant hypertension (none of the cats had ocular damage). In experimental models, hyperthyroidism tends to cause a reduction in diastolic blood pressure attributable to vasodilation and increases in SBP due to increases in stroke volume (Fazio and others 2004). The increase in heart rate that occurs in hyperthyroidism may also contribute to the observed increases in SBP because of reduced dynamic compliance of the arterial tree. This is because, in humans at least, when the heart rate is elevated, the reflected pressure wave from the peripheral arterial tree may summate with the forward pressure wave from a subsequent cardiac contraction, increasing systolic pressure (Biondi and others 2002). Whether this theory holds for the higher heart rates observed in cats is unknown.
An important limitation of the present study is the lack of echocardiographic examinations in the hyperthyroid cats before and after treatment. This limitation is a consequence of the retrospective nature of the study that recruited cats from first-opinion clinics. The population could therefore have included cats with primary heart disease, which could have influenced circulating NP concentrations. Despite this limitation, the results are considered meaningful as the hyperthyroid cats acted as their own controls, and it is unlikely that another factor apart from thyroid disease either directly or indirectly (as a result of cardiac remodelling) could be responsible for the changes in NP concentrations observed in a short period following successful treatment. Another limitation is the long storage period of up to 10 years for some samples. It cannot be excluded that this prolonged period could have affected the results; however, in both this and a -previous study using samples stored at −80 °C over a similar time period, a significant drop in NP concentration was not seen (Lalor and others 2009). The effect of storage would also be evident in both the baseline and post-treatment samples. An ideal follow-up study would prospectively evaluate circulating NT-proBNP in newly diagnosed hyperthyroid cats before and at fixed time points after successful treatment and would also include serial echocardiographic assessment.
In conclusion, this study demonstrates that hyperthyroidism has a modest but significant effect on NP concentration. NT-proBNP appears more susceptible than NT-proANP to changes in thyroid function. Thyroid status should be taken into account when interpreting NT-proBNP results.
We would like to thank the owners of the cats who willingly participated in this study. We would like to thank the veterinarians and staff at the Beaumont Sainsbury Animal's Hospital, People's Dispensary for Sick Animals, the Blue Cross Hospital, Victoria and the Queen Mother Hospital for Animals. We would like to thank the PetSavers foundation for funding this study, and Idexx laboratories, especially Andrew Beardow, for their support and performing the analysis of the samples.
None of the authors of this article has a financial or personal relationship with other people or organisations that could inappropriately influence or bias the content of the paper.
The authors confirm that all individuals personally acknowledged have given their permission to be listed.
The authors confirm that all co-authors have given their permission to be listed.