Cardiac troponin I in cats with compromised renal function

Abstract

Objectives

Cardiac troponins are sensitive and specific markers of myocardial injury. However, their reliability in renal disease has been questioned owing to possible renal involvement in troponin elimination. The primary objective of the present study was to examine whether serum cardiac troponin I is elevated in cats with compromised renal function and no clinically relevant structural cardiac disease. A secondary objective was to examine whether cardiac troponin I is measurable in the urine of cats with normal and compromised renal function.

Methods

This prospective case-control study included 52 cats (19 with compromised renal function, 19 with primary cardiac disease and 14 healthy controls). For all cats, clinical examination, echocardiography, electrocardiography, blood pressure, complete blood count, biochemistry, serum thyroxine and urinalysis were performed. Cardiac troponin I was measured in the serum and urine of each cat.

Results

Median (range) serum cardiac troponin I concentrations were 0.052 ng/ml (0.015–0.78 ng/ml) for the renal group, 0.083 ng/ml (0.003–3.27 ng/ml) for the cardiac group and 0.012 ng/ml (0.003–0.14 ng/ml) for the control group. The renal and cardiac groups both had significantly higher serum cardiac troponin I concentrations than the control group, whereas no difference could be detected between the renal and cardiac groups. In the renal group 7/19 cats had measurable urine cardiac troponin I, whereas cardiac troponin I was measurable in the urine of one cat in the cardiac group and two healthy controls. There was no significant correlation between serum and urine cardiac troponin I.

Conclusions and relevance

Elevated serum cardiac troponin I in cats with compromised renal function may occur without evidence of clinically relevant structural cardiac disease. Moreover, detecting cardiac troponin I in urine is most likely in cats with compromised renal function.

Introduction

The significance of myocardial injury in cardiac disease and, importantly, in non-cardiac disease, has become apparent in recent years.^1,2 While it is standard procedure to evaluate hepatic and renal biomarkers as part of an animal’s minimum database, the state of the heart has generally been examined through diagnostic imaging and electrocardiogram (ECG). As cardiac biomarkers are gaining their place in veterinary medicine, correct interpretation of these markers is paramount.

Cardiac troponins play an integral role in myocardial sarcomere contraction. The subunits cardiac troponin I and T (cTnI and cTnT) are cardio-specific. As they are solely intracellular proteins, their presence in circulation signifies cardiomyocyte injury, for which they have become gold-standard biomarkers.^1,3 While much is known about troponins with regard to their function, structure and release from injured cardiomyocytes, the process of their elimination has yet to be clarified. Originally believed to be cleared by the reticulo-endothelial system due to molecular size, renal elimination of fragments has also been suggested.^4,5 Consequently, it has been discussed whether cardiac troponins are reliable cardiac markers in patients with renal disease as declining renal function could lead to plasma accumulation.^4,5

Two studies originally revealed that dogs and cats with renal disease frequently have increased serum cTnI concentrations.^6,7 Nevertheless, this finding has several possible causes. Firstly, patients in these studies did not undergo echocardiography (ECHO) and might have had concurrent cardiac disease, in which case cTnI could be reliably reflecting primary myocardial injury. Secondly, uraemic toxins and altered haemodynamic states might cause secondary myocardial injury (ie, myocardial injury induced by non-cardiac disease), as is seen with severe systemic disease.^2,8 In humans the codependence of the heart and kidney on each other in disease states has been thoroughly described (cardiorenal syndrome), ⁹ which makes this theory likely, but the role of troponins in this context has not been elucidated. Finally, accumulation due to reduced renal elimination could lead to increased serum cTnI concentrations.

Only this last theory would compromise the reliability of cardiac troponins as reflective of myocardial injury in patients with renal dysfunction. A recent study in dogs with chronic kidney disease (CKD) appears to contradict this theory, as glomerular filtration rate (GFR) was not associated with serum cTnI. ¹⁰ However, further exploration of cardiac troponins in animals with compromised renal function is necessary for correct interpretation of the biomarker in this context.

The primary objective of the present study was to examine whether serum cTnI is elevated in cats with compromised renal function and no evidence of clinically relevant structural cardiac disease. A secondary objective was to examine whether cTnI is measurable in the urine of cats with normal and compromised renal function.

Materials and methods

Cats with compromised renal function or with primary structural cardiac disease were included consecutively to a renal and a cardiac group, respectively, as they presented to the University Hospital for Companion Animals, University of Copenhagen, Denmark, from 2014–2016. The renal group consisted of cats with either CKD stage 2 or higher (stable azotaemic renal disease staged on creatinine measurement according to the International Renal Interest Society [IRIS] guidelines), or cats with ureteral obstruction with azotaemia and evidence of underlying CKD. The cardiac group consisted of cats with hypertrophic cardiomyopathy (HCM) defined as maximal end-diastolic thickness of the interventricular septum or left ventricular free wall >5.5 mm. A healthy control group, age-matched to the renal group, was recruited through the department’s HCM screening programme, as well as from cats belonging to hospital personnel, and consisted of cats deemed healthy based on history, clinical examination, ECHO, ECG, blood pressure and clinical pathology (complete blood count [CBC], biochemistry, serum thyroxine and urinalysis).

For each cat, a clinical examination was carried out, and baseline systolic blood pressure was measured (PetMAP Graphic System; Ramsey Chemical). ECHO was performed and the data analysed (EchoPAC for PC, version 113; GE Healthcare) by an experienced sonographer using a Vivid E9 ultrasonographic system according to a previously described protocol. ¹¹ Blood was collected for CBC, biochemistry and serum thyroxine, and urinalysis was performed on a fresh urine sample obtained through cystocentesis.

For both the renal and cardiac groups, cats with evidence of comorbidities were excluded. Cats with HCM were excluded if history, clinical examination, CBC, biochemistry, serum thyroxine, blood pressure or urinalysis indicated comorbidities. Cats with compromised renal function were excluded if history, clinical examination, CBC or biochemistry (other than those associated with CKD), serum thyroxine or ECHO indicated comorbidities. In case of an active urine sediment, urine culture was performed. If subclinical bacteriuria was identified, the cat was either excluded or was re-examined and re-included after appropriate therapy and unremarkable sediment on a subsequent urinalysis. Additional investigations, such as diagnostic imaging, were performed on individual cats as deemed necessary by the attending clinician.

Serum for cTnI analysis was stored in cryovials at −80°C within 5 h of blood collection. Urine for cTnI analysis was stored in cryovials at −80°C within 1 h of collection. Serum and urine vials were stored for a maximum of 15 months until batch analysis. cTnI was analysed using a commercially available immunoassay (ADVIA Centaur CP TnI-ultra; Siemens Healthcare Diagnostics), which has been validated previously for use in cats. ¹² Its detection limit is 0.003 ng/ml according to this validation study. Analysis took place in four batches (corresponding serum/urine samples were analysed together) during the study period in order to minimise storage time for urine samples for which cTnI stability studies are unavailable (a flow chart is available in Figure 1). At a later time point (1–3 years after urine cTnI measurement) it was decided to analyse urine creatinine for normalising of urine cTnI to urine creatinine. Creatinine was measured in urine aliquots not previously thawed, using the assay Siemens Advia Chemistry Enzymatic Creatinine-2.

Figure 1

Flow chart of urine cardiac troponin I (cTnI) analysis. HCM = hypertrophic cardiomyopathy; CKD = chronic kidney disease

The study was approved by the departmental ethical committee and subject to informed owner consent for each cat.

Statistical analysis

Normality was tested using the D’Agostino–Pearson omnibus test. Logarithmic transformation was applied where this made a Gaussian distribution acceptable. ANOVA was used to compare overall differences of the log-transformed data between cardiac, renal and control groups. Post-hoc analysis of the one-way ANOVA was performed using Tukey’s honest significant difference test. Graphical assessment along with Pearson’s (Gaussian data) or Spearman’s (non-Gaussian data) correlation coefficient were used to investigate correlations between serum cTnI and creatinine, phosphate, haematocrit and systolic blood pressure, and between serum cTnI and urine cTnI (in total, as well as normalised to urine creatinine). All analyses were conducted using commercial statistical software (GraphPad Prism 7.0). Statistical significance was defined as P <0.05.

Results

The study included 52 cats (demographic data are presented in Table 1). Nineteen cats were included in the renal group (10 cats with CKD IRIS stage 2, four cats with CKD IRIS stage 3, one cat with CKD IRIS stage 4, one cat with previously known CKD presenting with acute-on-chronic disease [unilateral ureteral obstruction] and three cats with unilateral ureteral obstruction and evidence of underlying CKD [low urine specific gravity (3/3) and ultrasonographic evidence of chronic renal parenchymal changes (2/3)]). Of cats in the renal group, nine had mild, insignificant ECHO findings (five had false tendons [without secondary hypertrophy], two had a trivial–mild aortic insufficiency and two had mild diastolic dysfunction [no left atrium enlargement, E/A ratio (ratio of early to late ventricular filling velocities) <1]). Nineteen cats with HCM were included in the cardiac group (five of which had advanced disease, with two of these five in congestive heart failure based on clinical and ECHO findings). Finally, 14 healthy cats were included in the control group.

Table 1

Demographic data of included cats

Group	n	Breeds	Sex	Median (range) age (years)
Renal	19	12 DSH, 2 DLH, 2 NFC, 1 RB*, 2 MB	1 FE, 7 FN, 11 MN	11 (2–17)
Cardiac	19	9 BSH, 4 MC † , 4 NFC, 1 DSH, 1 Ragdoll	1 FE, 4 FN, 14 MN	5 (1–11)
Control	14	6 DSH, 5 BSH, 1 Somali, 1 NFC, 1 MB	5 FE, 5 FN, 1 ME, 3 MN	7.5 (3–16)

DSH = domestic shorthair; DLH = domestic longhair; NFC = Norwegian Forest Cat; RB = Russian blue; MB = mixed breed; FE = female entire; ME = male entire; FN = female neutered; MN = male neutered; BSH = British Shorthair; MC = Maine Coon

*

This cat had received fluid therapy at its referring practice prior to admission to the hospital and inclusion in the study. This had caused its urine specific gravity (USG) to fall from 1.025 (at the referring practice) to 1.017 (at admission). Nevertheless, it had detectable urine cardiac troponin I.

†

One Maine Coon had received furosemide for congestive heart failure for 4 days prior to study inclusion. Therefore, on the day of study inclusion this cat had a serum creatinine concentration of 165 µmol/l and a USG of 1.018, but we decided to include the cat because its creatinine had been measured at 89 µmol/l and its USG at 1.050 on the day of initiation of furosemide therapy

Median (range) serum cTnI concentrations were 0.052 ng/ml (range 0.015–0.78 ng/ml) for the renal group, 0.083 ng/ml (range 0.003–3.27 ng/ml) for the cardiac group and 0.012 ng/ml (range 0.003–0.14 ng/ml) for the control group. Cats in the cardiac and renal groups both had significantly higher serum cTnI concentrations than the control group (P = 0.002; P = 0.004), whereas no difference could be detected between the renal and cardiac groups (P = 0.96) (Figure 2). A significant but weak correlation between serum cTnI and creatinine was found for the entire study population (Spearman’s r = 0.43, P = 0.0015) but not between serum cTnI and phosphate (Spearman’s r = −0.06, P = 0.66), haematocrit (Pearson’s r = 0.005, P = 0.97) or systolic blood pressure (Spearman’s r = 0.03, P = 0.86).

Figure 2

Serum cardiac troponin I (cTnI) concentrations of a healthy control group (n = 14), a cardiac group (cats with hypertrophic cardiomyopathy, n = 19) and a renal group (cats with compromised renal function, n = 19). Geometrical mean concentrations are shown as horizontal lines. Statistically significant differences between groups are indicated with an asterisk (ANOVA with post-hoc analysis using Tukey’s honest significant difference)

Seven of the 19 cats in the renal group had measurable urine cTnI of 0.008 ng/ml (range 0.005–0.026 ng/ml), corresponding to 0.014 ng/mg (range 0.0045–0.052 ng/mg) creatinine for these cats, whereas cTnI was measurable in the urine of only one cat (0.003 ng/ml, corresponding to 0.001 ng/mg creatinine) in the cardiac group and two healthy controls (0.005 and 0.015 ng/ml, corresponding to 0.0012 and 0.0062 ng/mg creatinine, respectively) (Figure 3). Cats included in each analysed batch of samples, sample storage time, reagent batch used and urine troponin results are presented in the flow chart in Figure 1.

Figure 3

(a) Urine cardiac troponin I (cTnI) concentrations and (b) urine cTnI:creatinine ratio of a healthy control group (n = 14), a cardiac group (cats with hypertrophic cardiomyopathy, n = 19) and a renal group (cats with compromised renal function, n = 19)

There was no significant correlation between serum and urine cTnI (Spearman’s r = 0.26, P = 0.06) or between serum cTnI and urine cTnI:creatinine ratio (Spearman’s r = 0.26, P = 0.06).

Discussion

This study documents the presence of increased serum cTnI in cats with compromised renal function and no ECHO evidence of clinically relevant structural cardiac disease. A recent study has reported similar findings in dogs with acute kidney injury. ¹³ Accordingly, while primary cardiac disease should always be considered a potential cause of elevated serum troponins, it should not be thought of as the sole possible cause in dogs or cats with renal disease.

Associations between the degree of azotaemia and serum troponin have been investigated repeatedly with conflicting results,^7,14 with most studies failing to show a significant correlation. A weak correlation, as detected in the present study, might support the idea that haemodynamic changes and uraemic toxins lead to secondary myocardial injury, worsening with more severe renal dysfunction.

Based on the results of this study it is considered highly likely that increased serum cTnI in patients with compromised renal function at least partly reflects secondary myocardial injury (ie, myocardial injury induced by non-cardiac disease). The codependence of the cardio–renal axis is well described, its interplay involving especially shared mechanisms related to volume and blood pressure regulation.^9,13 The so-called cardiorenal syndrome has been defined in humans, referring to pathological changes in one system ultimately affecting the other.^9,15 Chronic renocardiac syndrome is defined as ‘chronic abnormalities in renal function leading to cardiac disease’, ¹⁶ with causes including activation of the renin–angiotensin–aldosterone system and the sympathetic nervous system following a declining GFR, both of which induce cardiac remodelling with left ventricular hypertrophy (LVH) and fibrosis. ¹⁶ Additionally, certain uraemic toxins, as well as hyperphosphataemia and non-regenerative anaemia are believed to contribute to cardiac remodelling.^9,16,17

No significant correlations could be confirmed between serum cTnI and either serum phosphate or haematocrit in the present study, but cats with relevant structural cardiac abnormalities were deliberately excluded from the renal group, and this might have influenced these results. A study in humans, however, reported that LVH did not singularly explain the increased serum troponins, ¹⁸ as is mirrored by the finding of increased serum cTnI in the present cohort of cats without cardiac remodelling. Interestingly, increased cardiac troponins in humans with CKD is associated with a 2–5-fold increased risk of death. ¹⁹ This, in itself, is an argument for occurrence of true cardiac injury rather than falsely increased troponin concentrations. ¹⁸ Therefore, importantly, regardless of the underlying mechanism, increased serum cTnI deserves further attention in cats with CKD as a possible prognostic indicator.

cTnI analysis was also performed on urine samples from cats with normal and decreased renal function in order to investigate renal elimination. If the kidneys are responsible for troponin elimination, and cTnI is detectable in urine, cats with normal renal function and the highest serum cTnI concentrations would be expected to have the highest urinary cTnI concentrations. This was not the case in the present study.

In humans, cardiac troponins have not been detectable in the urine of healthy individuals, although the literature on the subject is sparse. In the one published human study, ²⁰ patients with primary cardiac injury (represented by acute myocardial infarction or recent cardiac surgery), as evidenced by increased serum cardiac troponins, had non-detectable urine cTnI, and only one had detectable urine cTnT. Interestingly, patients with decreased renal function had detectable urinary cTnT, and (although less commonly) cTnI. ²⁰ These results make renal elimination less likely. Firstly, in the above mentioned study, even in cardiac patients with markedly elevated serum troponin concentrations, urinary troponin was not detectable. Secondly, as urinary troponin was found almost exclusively in patients with reduced renal function, its presence more likely reflects increased glomerular loss or decreased tubular reabsorption. The validity of this human study’s results was strengthened by an analytical validation for which spike-and-recovery analysis found no interference of the human urine matrix on troponin results. ²⁰

In the present study urine cTnI was detected in 7/19 cats with compromised kidney function, 1/19 cats with HCM and 2/14 controls. The first two batches of samples analysed (Figure 1) represented the pilot data ²¹ for which it was evaluated whether troponin was at all measurable in feline urine. Of these animals 7/9 cats in the renal group had measurable urine cTnI, whereas this was the case for none of the cats in the cardiac group and only 1/7 controls. This was followed by a longer period of patient inclusion.

In the next two batches, surprisingly, none of the cats in the renal group had measurable urine cTnI. These cats did not have significantly different serum creatinine concentrations than the initially evaluated cats. Samples were analysed with different batches of cTnI reagent, but, for all, calibration had been unremarkable. It was discussed whether minimal variation in reagent sensitivity could be an issue, but in all cases of measurable urine cTnI, the biomarker had been measured above the assay detection limit for cTnI in serum (0.003 ng/ml) ¹² and within a concentration range also detected in serum samples of the study group. Lack of stability of cTnI in urine was also considered as urine from two cats of the second batch was re-analysed at the time of batches three and four without re-detection of cTnI. But, while samples in the third batch had the longest storage time, the final batch of samples had all been stored for a maximum of 2 months, a time frame similar to that of the two initial batches. Owing to these differences in batch results, it is not possible to conclude with certainty regarding the urine cTnI of cats in this study. Nevertheless, our findings appear to mirror those of the human study, ²⁰ in which mainly patients with decreased renal function had measurable urine troponin.

Intact troponins are considered too large for renal elimination. While post-release proteolysis has been described for cTnI especially, ²² studies have shown conflicting results with regard to fragmentation of cTnT,^5,23 the subunit most commonly increased in the serum of humans with renal disease and most frequently measurable in the urine of humans.^4,18,20 The most thorough investigation of the subject contradicted the occurrence of relevant cTnT fragmentation and supported the theory of reticuloendothelial clearance. ²³ Additionally, the half-life and clearance of serum cTnI in human patients with CKD who develop secondary acute myocardial infarction does not appear to be different from that of patients with normal renal function, a fact that also speaks against decreased troponin elimination in renal disease. ²⁴ Finally, a recent study measured both serum cTnI and GFR in dogs with CKD, and no association between the two was found in a multivariate analysis, concluding that decreased GFR did not appear to result in accumulation of cTnI in circulation. ¹⁰

Lack of validation of the TnI ultra-assay for feline urine is the main study limitation. It cannot be ruled out that the urinary cTnI detected reflected artefactual measurements, although the pilot data appeared to suggest otherwise. Future evaluation of markers of glomerular and tubular function along with urine cTnI in cats with CKD could possibly shed further light on the underlying pathophysiology.

A second limitation is the inclusion of four cats with ureteral obstruction, which made the renal group less uniform. These cats had evidence of underlying CKD, but the contribution of chronic disease and obstruction, respectively, to the cats’ azotaemia is unknown. Nevertheless, as the objective was to evaluate if cTnI could be detected in the urine of cats with normal and compromised kidney function, the cause(s) of compromised function were of less importance, and the cats were considered eligible for inclusion. Also, as not all cats with CKD underwent abdominal ultrasound in their work-up, ureteroliths could not be ruled out in the remaining cats, and hence all were kept in the population. Abdominal ultrasound and radiographs would have been beneficial for further classification of all cats in the renal group. Group classification would also have benefited from measurement of symmetric dimethylarginine in order to rule out early kidney disease in the cardiac and control groups, but this marker was not commercially available at the initiation of the study.

A third limitation is the inclusion of nine cats with trivial cardiac abnormalities (but no primary structural cardiac disease) in the renal group. It has been shown that even mild primary structural heart disease is insufficient to cause detectable myocardial injury. ²⁵ In accordance with this, the serum cTnI concentration of these nine cats all fell within the range of the control group. Hence, their inclusion was considered acceptable.

Fourthly, urine creatinine was measured a considerable time after urine cTnI, but owing to the stability of creatinine in urine, ²⁶ the results are considered reliable. Stability of cTnI in urine, however, is presently unknown, which is, therefore, also a limitation.

Finally, in humans cTnT is the marker most affected by renal disease^18,27 and the one most commonly detected in urine samples, ²⁰ and its measurement in this study could, therefore, have been of interest. Nevertheless, cTnI and cTnT are sources of similar information, with cTnI the most sensitive. ¹ As dogs and cats rarely develop acute myocardial infarction, for which the most severe myocardial injury is often seen, cTnI is believed to be a better marker of the sometimes mild, but still prognostically important, myocardial injury seen in companion animals, ⁸ which is why this study focused on cTnI.

Conclusions

Increased serum cTnI is common in cats with compromised renal function, even in the absence of clinically relevant structural cardiac disease. While an element of decreased troponin elimination cannot be ruled out by the present study, it appears very likely that secondary myocardial injury occurs in these cats.