Abstract
Hypertension is a common problem in older cats, particularly associated with chronic kidney disease (CKD). Reduced activity of 11β-hydroxysteroid dehydrogenase type 2 predisposes to hypertension in human patients by allowing excessive stimulation of the mineralocorticoid receptor by cortisol. This study was designed to test the hypothesis that reduced conversion of cortisol to cortisone contributes to the development of systemic hypertension in some cats with CKD and idiopathic hypertension (iHT). The study included 60 client-owned cats: 21 clinically normal, 16 normotensive cats with CKD (NTCKD), 14 hypertensive cats with CKD (HTCKD) and nine iHTs. Urine cortisol and cortisone were extracted into dichloromethane and chloroform, respectively, prior to analysis by radioimmunoassay. Data are reported as median and range. The Kruskall–Wallis test was used to compare cortisol:cortisone ratios between groups with post-hoc testing using the Mann–Whitney
Hypertension is being increasingly recognised in feline patients, presumably as a result of increased awareness of the problem. It is most frequently identified in cats with chronic kidney disease (CKD). Reduced plasma potassium is a significant risk factor for hypertension in cats with CKD. 1 It has been suggested that hypertension in cats may be caused by relative or absolute hyperaldosteronism, 2 a suggestion supported by the association of low plasma potassium concentration and hypertension. Indeed, there are increasing numbers of case reports of cats with primary hyperaldosteronism due to functional adrenal tumours. 3–9
Excessive secretion of the mineralocorticoid aldosterone has long been known to result in hypertension in humans. 10 Hyporeninaemic hyperaldosteronism was recently reported in a population of hypertensive cats and adrenal gland histopathology in a proportion of these cats revealed extensive micronodular hyperplasia. 11 The majority of these cats had CKD at the time of diagnosis. An earlier study of cats with hypertension and CKD also showed mean aldosterone concentrations to be significantly higher in hypertensive cats with renal disease when compared to normotensive non-azotaemic controls. 2 However, a different study failed to identify any difference in aldosterone concentrations between normal and hypertensive cats. 12
In human patients, conditions have been recognised where hypertension and hypokalaemia feature but mineralocorticoid concentrations are normal or low. One such syndrome is that of apparent mineralocorticoid excess (AME) which occurs when there are mutations in the 11β-hydroxysteroid dehydrogenase type 2 (11β-HSD2) encoding gene. 13 The enzyme, 11β-HSD2, catalyses the conversion of hormonally active cortisol to inactive cortisone. This enzyme ‘protects’ the mineralocorticoid receptor (MR, with which it co-localises) from activation by cortisol and is therefore vital in dictating the specificity of the MR for aldosterone. 14 Inhibition of 11β-HSD2 or mutations in the genes encoding 11β-HSD2 would result in decreased conversion of cortisol to cortisone and excessive stimulation of the mineralocorticoid receptor by cortisol binding to it. Agonistic effects at the MR would be expected to result in hypokalaemia, predominantly via potassium diuresis; hypertension, through sodium retention and extracellular fluid volume expansion; and physiological suppression of renin and aldosterone concentrations. 11β-HSD2 has been shown to be markedly impaired in human patients with CKD, as assessed by both the measurement of urinary metabolite ratios 15 and 11β-HSD2 mRNA expression 16 and if the same is true in cats this could account, at least in part, for the hypertension that is seen in a proportion of cats with CKD. The conversion of cortisol to cortisone can be bi-directional dependent upon the enzyme involved and has been termed the ‘cortisol–cortisone shuttle’. 17 In common with other species, two isoforms of 11β-HSD have been identified in the cat (11β-HSD1 and 11β-HSD2). The enzyme 11β-HSD1 is mainly present in the liver and kidneys. In the feline liver it appears to act as a reductase enzyme, converting cortisone into cortisol, although its efficacy is reportedly poor and its exact function has not yet been elucidated. 18 In the renal parenchyma, 11β-HSD1 has bi-directional activity with predominantly dehydrogenase activity, converting active cortisol to its inactive form cortisone and, therefore, like 11β-HSD2 it ‘protects’ the mineralocorticoid receptor. The enzyme 11β-HSD2 is present in mineralocorticoid target tissues such as the kidney 18 and is uni-directional converting active cortisol to inactive cortisone.
The aim of the present study was to evaluate the efficacy of the cortisol–cortisone shuttle by measurement of cortisol and cortisone concentrations in the urine of feline patients diagnosed with hypertension with or without biochemical evidence of CKD and compare the results obtained with those found in normotensive cats.
Material and Methods
Patient Selection
To determine whether urinary cortisol and cortisone concentrations are altered in systemic hypertension, urine samples were collected from clinically normal cats, normotensive cats with chronic kidney disease (NTCKD), hypertensive cats with chronic kidney disease (HTCKD) and cats with idiopathic hypertension (iHT). Samples were also collected in a small number of the HTCKD and iHT cats following the implementation of anti-hypertensive therapy. The cats were recruited from the general, client-owned population that presented to two first opinion veterinary clinics between May 1995 and October 2005: The People's Dispensary for Sick Animals, Bow, and the Beaumont Animals Hospital, Camden Town, both in London. The clinical criteria used to categorise cats into these groups were as follows:
To be considered clinically normal, cats had to have no clinically significant abnormalities detected on history, physical examination, or routine plasma biochemistry and have a urine specific gravity>1.035. In an attempt to age-match these cats with those in the diseased groups; cats were actively recruited into this group if they were at least 12 years old. However, younger cats were not excluded.
To be classified as having CKD, cats had to have plasma creatinine concentrations persistently greater than the upper limit of the laboratory reference range (>177 μmol/l [>2.0 mg/dl] on at least two occasions 8–12 weeks apart) and urine specific gravity≤1.035, together with historical or physical examination findings consistent with the diagnosis, such as polyuria, polydipsia, and palpably small kidneys. Cats with acute renal failure, diagnosed on the basis of clinical history and response to treatment, were excluded.
To be classified as having iHT, cats had to have no clinically significant abnormalities on routine plasma biochemistry (consisting of alkaline phosphatase and alanine transaminase activity, bilirubin, total protein, albumin, sodium, potassium, phosphorus, urea and creatinine) not be anaemic and have total thyroxine measurements <40 nmol/l [<3.12 μg/dl] (reference range, 10–55 nmol/l [0.78–4.29 μg/dl]). Cats that became azotaemic within a 9-month time period, following the diagnosis of iHT, were not included in the iHT group.
Measurements of systolic blood pressure (SBP) were made from the foreleg by using a Doppler flow detector (Parks model 811B, Perimed, Bury St Edmonds, UK) with a 9.5-MHz probe, as described previously. 1 Five measurements of SBP were obtained in a measurement session and the arithmetic mean was used for subsequent data analysis. Cats were considered hypertensive, and treatment was implemented if SBP was ≥175 mmHg on more than one occasion or if SBP was ≥175 mmHg and hypertensive ocular lesions were observed. Anti-hypertensive therapy (amlodipine 0.625–1.25 mg/cat/day; Istin, Pfizer, Sandwich, UK) was instituted as outlined above with the aim of reducing the SBP to less than 165 mmHg.
Cats with hyperthyroidism were excluded from all the groups in the study, as were cats receiving glucocorticoid medication. The collection and storage of blood and urine samples were performed with the informed consent of the cat's owners. Blood samples were obtained by jugular venepuncture, and urine samples were collected by cystocentesis. To prevent any cat being included in the analysis on more than one occasion, in any cat for which urine samples from multiple visits were available for analysis, the first available sample was used. Post-treatment samples were the first available sample following adequate blood pressure control. Only a small subset of cats was examined post treatment. The Ethics and Welfare Committee of the Royal Veterinary College approved the study protocol.
All the azotaemic cats received standard therapy for CKD. This included feeding of a low-protein, low-phosphate diet (Royal Canin Renal Diet, Melton Mowbray, Leicestershire, UK) if this was palatable to the cats and if the owners were willing to feed it. None of the cats received angiotensin converting enzyme inhibitors. Cats also received standard medical care for any concurrent medical problems that developed.
Urinalysis
After collection, urine samples were stored on ice until urinalysis was performed. This was commenced within 6 h of sample collection. Analysis consisted of measurement of specific gravity with a refractometer, use of a standard multi-test urine dipstick, and microscopic examination of urine sediment. After urinalysis, the remaining urine was centrifuged (10 min at 1000 ×
Cortisol and Cortisone Assays
Urine cortisol was extracted into dichloromethane and dried extracts were analysed by radioimmunoassay as described previously. 19 Urine cortisone was extracted into chloroform and also analysed by radioimmunoassay as described previously. 20 Correction of individual samples for extraction efficiency was not performed as previous validation assays have shown recovery to be very high. Recovery of a spike of 50 nmol/l cortisone added to 10 human urine samples with unspiked cortisone values of 50–110 nmol/l was 92% (± 2SD range 89–95%). Recovery of titrated cortisone spiked into six human urine samples was 93% (± 2SD range 90–96%). The anti-cortisone antiserum (N-137) shows very high specificity and cross-reacts less than 0.1% with cortisol. The precision profile for the urine cortisone assay showed between batch (inter-assay) precisions of less than 10% for urine cortisone concentrations over 10 nmol/l. 19,20 A urinary cortisol–cortisone ratio (CC) was calculated. When cortisol and cortisone measurements were considered individually their values were normalised to a urine specific gravity of 1.010 prior to statistical analysis to account for the effect of variable urine volumes produced.
Statistical Methods
All analyses were performed using computerised statistical analysis software (SPSS 14.0 for Windows, Chicago, IL). As the data were not normally distributed, results are reported as median and range. Cortisol/cortisone ratios in the different groups were compared using the Kruskal–Wallis test with post-hoc testing between groups by the Mann–Whitney
Results
A total of 60 cats were selected for inclusion into the study, including 21 clinically normal, 16 NTCKD, 14 HTCKD and nine cats with iHT. The age of three cats was unknown. The remaining cats were 12.6 (range, 3.0–20.0) years old. The age of the cats was well matched between the four groups (Table 1). Forty-five domestic shorthair cats, seven domestic longhair cats, two British Blue, one Havana, one Birman, one Siamese, one Persian and two Burmese cats were included in the study. No breed was overrepresented in any one group.
Demographic data for the 60 cats included in the study
n=indicates number of cats included in the analysis. Rows with groups bearing different letters display a significant difference. †The number of crosses indicates the number of cats of unknown age.
None of the cats in the iHT group had clinical signs or results of routine biochemical tests that were compatible with known causes of systolic hypertension. Urinary cortisol and cortisone concentrations when considered individually were not significantly different between the four groups of cats (Fig 1). Plasma potassium concentration was significantly different between the four groups (
Box and whisker plot showing urinary cortisol and cortisone values normalised to urine specific gravity 1.010. Outliers, defined as cases with values between 1.5 and 3 box lengths from the upper or lower edge of the box, are represented as open circles. Extremes, defined as cases with values more than 3 box lengths from the upper or lower edge of the box, are marked with asterisks. One outlier and seven extremes are not represented on this figure and were above 40 nmol/l. There was no statistical difference between groups in either cortisol (
Urinary cortisol–cortisone ratios were significantly different between the four groups of cats (
Box and whisker plot showing urinary cortisol–cortisone ratios. Outliers, defined as cases with values between 1.5 and 3 box lengths from the upper or lower edge of the box, are represented as open circles. The urinary cortisol:cortisone ratio was significantly higher in clinically normal cats when compared to NTCKD, HTCKD and cats with iHT. Statistical comparisons between groups were by the Kruskal–Wallis test with post-hoc comparisons by the Mann–Whitney
Urinary cortisol–cortisone ratios pre- and post treatment in the combined subset of HTCKD (
Discussion
The results of the present study do not indicate that an elevated urinary cortisol–cortisone ratio is present in hypertensive cats. In fact, the shuttle actually appears more effective in cats with CKD (normotensive and hypertensive) and iHT than in normal cats with a higher proportion of the excreted corticoid being in the form of cortisone. The mechanism for this is not clear.
Twenty-four hour urinary cortisol and cortisone measurement is known to be a sensitive marker of 11β-HSD2 dysfunction in humans. 21 Assuming single urinary cortisol and cortisone measurements accurately reflect 24 h measurements in cats, as they do in people, 22 and that single measurements of cortisol are not significantly affected by stress, the cortisol–cortisone shuttle appears to be working efficiently based on these results. The stress of hospitalisation has been shown to increase the urinary corticoid:creatinine ratio in cats; 23 however, the cats in the present study were not hospitalised and the urine samples taken would, in most cases, be representative of urine collected in the urinary bladder over the previous few hours which should, at least in part, lessen the influence of any acute cortisol increase due to stress at the time of sample collection.
In the absence of 11β-HSD2, hypertension and hypokalaemia result in feedback inhibition of aldosterone production and aldosterone concentrations are low as a result. 17 There are suggestions that aldosterone is elevated in HTCKD, 2 although interpretation of a single aldosterone measurement is problematic and that study included only low numbers of cats. A further study failed to identify any difference in aldosterone concentrations between normal cats, hypertensive cats with chronic kidney and cats with iHT. 12 Although Jensen's observations would argue against a deficiency of 11β-HSD2 as a cause of hypertension in cats with CKD the literature is contradictory and further work is required.
The diagnosis of iHT may be problematic and an American College of Veterinary Internal Medicine (ACVIM) consensus statement has recently tried to establish more rigorous diagnostic criteria 24 for this condition. These include the demonstration of a sustained increase in blood pressure concurrent with a normal complete blood count, serum biochemistry, and urinalysis and depending on the clinical findings, renal ultrasound examination, measurement of glomerular filtration rate, quantitative assessment of proteinuria and thyroid hormone determination. Additional tests that are recommended for consideration include serum and urine aldosterone and catecholamine concentrations and adrenal ultrasound examination. Aldosterone concentrations were available for five of the nine iHT cats in the present study and were normal in all of them. This excluded primary hyperaldosteronism in at least the majority of this group, although the further investigations detailed above would have been optimal to categorically determine that their hypertension was truly idiopathic. It is unlikely that the pathogenesis of hypertension is identical in every cat, and it is also improbable that a single mechanism would explain hypertension in all cats. Therefore, it may also have been optimal if cats in the HTCKD group had undergone some degree of further evaluation following the guidelines set out by the ACVIM consensus statement, although this would have been difficult in the setting of first opinion, charity clinics.
A study of 20 essential hypertensive human patients showed demonstrably reduced 11β-HSD2 activity in 35% of them, as measured by prolonged half-life of cortisol, although no significant changes in plasma or urinary ratios of cortisol to cortisone were found. 25 Despite negative results in the present study a role for the cortisol:cortisone shuttle in feline hypertension cannot be entirely discounted until more dynamic function tests on this shuttle have been devised for the cat and/or kinetic parameters of the dehydrogenase enzyme have been assessed on renal biopsies. The feline renal cortex has been shown to constitute high enough expression to establish measurable enzyme activity 18 and renal 11β-HSD2 mRNA expression has been successfully measured in human patients with CKD. 16 However, as most cats with systemic hypertension are successfully treated, renal tissue from hypertensive cats will be difficult to obtain.
Given that another isoform of 11β-hydroxysteroid dehydrogenase has been identified in the cat, 11β-HSD1, it could be argued that deficient activity of 11β-HSD2 may not lead to mineralocorticoid excess and hypertension because 11β-HSD1 is present to inactivate cortisol. The bi-directional enzyme, 11β-HSD1 is mainly present in the liver and kidneys 18 and it might also be informative to examine kinetic parameters of this enzyme on hepatic and renal biopsies in the hypertensive feline population.
The measurement of urinary cortisol and cortisone is validated in a number of species but currently not for the cat. Methods are validated for human urine in terms of detection limits, precision, recovery, dilutional parallelism and correlation with high performance liquid chromatography results. 19,20 Given that free steroid is extracted prior to immunoassay and the feline urine matrix removed, it was considered highly unlikely that the assay performance would be any different than from when used to measure cortisol and cortisone in human urine.
Plasma potassium concentrations were significantly different between the normotensive CKD cats and the hypertensive cats (both with CKD and idiopathic) in the present study. Reduced plasma potassium has previously been observed to be a significant risk factor for hypertension in cats 1 and the results from the present study would corroborate this. It is possible that the low potassium concentrations in the hypertensive CKD cats in this study were attributable to hyperaldosteronism, although this would argue against this study's hypothesis as feedback inhibition of aldosterone occurs in patients with a congenital absence of 11β-HSD2, although this may not hold true for patients with CKD. As mentioned previously, aldosterone has been shown to be elevated in hypertensive cats with CKD. 2 The decrease in activity of 11β-HSD2 in human renal disease is less marked than in the syndrome of AME, 16 and other factors are thought to stimulate aldosterone secretion in these patients. Recent evidence suggests that α-adrenergic mediated vasoconstriction may also contribute to sustained hypertension in the syndrome of AME. 26
Urine cortisol levels were not significantly different between groups in the present study. It might be assumed that cats with CKD have increased cortisol secondary to reduced renal clearance. Human patients with CKD have been observed to have prolonged plasma cortisol half-life, 27 although negative feedback mechanisms lead to a concomitant fall in cortisol secretion rate and, therefore, plasma cortisol concentrations remain unchanged. The same mechanism could be postulated to occur in this population of cats. It appears that in cats urinary excretion of steroid hormones, most noticeably cortisol and aldosterone, is less than in other species and that a greater proportion is excreted in the faeces. 28,29 This would only be a problem if either cortisol or cortisone was differentially excreted to a greater degree and this would seem highly unlikely. As long as the cortisol:cortisone ratio remains unchanged it is insignificant if urinary cortisol and cortisone concentrations are low. In this study urinary cortisol and cortisone were well above the limits of detection for the assay so quantification was not thought to be an issue.
The principal limitation of this trial is the small number of cats (60), especially in the iHT group. Normalising urinary cortisol and cortisone values to urine creatinine concentration instead of urine specific gravity may have increased the precision of the individual hormone measurements (cortisol and cortisone) but this would not have influenced the resulting cortisol–cortisone ratio. Measurement of plasma aldosterone in the HTCKD group and iHT group would have been optimal to fully exclude primary hyperaldosteronism. In conclusion, the cortisol–cortisone shuttle appears to be efficient in hypertensive cats and it does not appear to be involved in the pathogenesis of feline hypertension. Additional studies looking at 11β-HSD2 function on renal biopsies may be useful.
Footnotes
Acknowledgements
We are grateful to the clinicians and nurses at the Beaumont Animal Hospital and the People's Dispensary for Sick Animals for their help and cooperation with this study. We would especially like to thank Shanta Cariese for her help with sample processing and Peter Wood at Southampton General Hospital for analysing the samples. The study was supported by a grant from the PetPlan Charitable Trust.