Grzegorz M. Kubiak M.D., Clinical ward of Cardiology, Medical University of Silesia, Marii Curie-Skłodowskiej St. 10, 41–800 Zabrze, phone/fax: +48 32 271 10 10, e-mail: greg_kubiak@yahoo.com

Abbreviations

ALT – alanine transferase

AST – aspartate aminotransferase

CKD – chronic kidney disease

CAD – coronary artery disease

DCM – dilated cardiomyopathy

eGFR – estimated glomerular filtration rate

ESC – European Society of Cardiology

HF – heart failure

MDRD – Modification of Diet in Renal Disease

NT-proBNP – N-terminal pro-brain natriuretic peptide

PTH – parathyroid hormone

P – phosphorus

SD – standard deviation

Ca2+ – total calcium

Introduction

It is estimated that 1–2% of the population in developed countries suffers from heart failure (HF), and half of these cases are associated with impaired left ventricle contractility mainly due to coronary artery disease (CAD) or hypertension [1]. Patients with heart failure caused by dilated cardiomyopathy (DCM) of non-ischaemic origin remain an interesting observational group in whom the relationship between the serum concentration of parathyroid hormone (PTH) and other biochemical and clinical parameters was not fully elucidated. PTH is a polypeptide that is produced in the structure of the parathyroid gland in response to the diminution of calcium ions (Ca2+) in the extracellular fluid. Its secretion is based on the principle of a negative feedback loop with the serum concentration of physiologically active Ca2+ ions. PTH increases the serum concentration of ionised Ca2+ due to an enhancement of its reabsorption in renal tubules and increase in phosphate secretion and osteoclast activation [2]. On the basis of the regulation of alpha-1 hydroxylation of 25-OH vitamin D, it has a crucial effect on the absorption of Ca2+ in the intestinal tract [3]. PTH has a significant effect on the cardiovascular system via its widespread receptors in the myocardium, endothelial, and smooth muscle cells. Its biologically active form, known as intact PTH, is secreted by the parathyroid gland and consists of eighty-four amino acids. PTH resolves into two components – active N-terminal PTH and inactive C-terminal PTH, which is excreted via the renal route. The active PTH forms (84-PTH and N-PTH) are mainly metabolised in the liver, resulting in a short half-life of intact PTH, which is relatively independent of renal function, although its serum concentration correlates with the absolute amount of the hormone and its activity [3, 4]. The PTH serum concentration is increased in primary hyperparathyroidism, vitamin D deficiency, or chronic kidney disease (CKD). Augmentation of PTH serum concentration in HF patients has a potential negative effect on symptom severity and prognosis.

Material and methods

The aim of this study was to assess the serum concentrations of PTH, total calcium (Ca2+), P, and vitamin D in 65 consecutive patients with non-ischaemic heart failure, who were hospitalised in our centre from March 1, 2012 to September 15, 2014. The control group consisted of sixty-two non-HF volunteers. Exclusion criteria in both groups were: CKD stage four or higher, vitamin D supplementation, diagnosed primary hyperparathyroidism, or malignancy. The blood samples of the HF patients enrolled in the study were collected on admission. The study complies with the Declaration of Helsinki, was approved by the local ethics committee (number of approval KNW/0022/KB1/40/I/12/13), and all patients provided informed consent.

Serum intact-PTH concentration was assessed in both groups using the immunoradiometric assay technique with the DIAsource hPTH-120 min-IRMA Kit (DIAsource Immunoassays S.A., Belgium). The vitamin D serum concentration was assessed using the radioimmunoassay with the DIAsource 25-OH vitamin D total-RIA-CT Kit (DIAsource Immunoassays S.A., Belgium). Serum P and total Ca2+ concentrations were assessed using the enzymatic method as described by Spinreact S.A., Spain and the colorimetric Arsenazo III method as described by Biomaxima S.A., Poland. In the HF patients, routine examinations were performed on admission according to the guidelines of the European Society of Cardiology (ESC), which consisted of peripheral blood morphology, creatinine, aspartate (AST), and alanine (ALT) transferases and N-terminal probrain natriuretic peptide (NT-proBNP) serum concentrations. The estimated glomerular filtration rate (eGFR) was calculated on the basis of the Modification of Diet in Renal Disease (MDRD) formula. The New York Heart Association (NYHA) functional class assessment and echocardiography examination were performed upon admission to the hospital. Left ventricular systolic function impairment was caused by dilated cardiomyopathy in the majority of cases (60 patients – 92.3% of cases), which were mostly postinflammatory, idiopathic, or, in one female, postpartum in origin. All patients were subjected to invasive coronary angiography to exclude the ischaemic background of HF. Patients were requested to characterise their daily profile of sun exposure during the last eight weeks as “above” or “below” seven hours a week, equivalent to one hour per day. Importantly, the population of Poland, which has approximately forty million people, exhibits a temperate climate in Central Europe, and its inhabitants frequently suffer from vitamin D deficiency due to insufficient sun exposure and nutritional supply.

Statistical analysis

The distributions of the examined parameters were analysed using the Shapiro-Wilk test. Values were presented as the mean and standard deviation (SD) or as the median in the 25th and 75th percentiles. Nominal and categorical values were expressed in percentages or proportional rates. Linear variables with normal distribution were compared using Student’s t-test. Variables with abnormal distribution were compared using the Kolmogorov-Smirnov, Mann-Whitney U, and Kruskall-Wallis tests. Categorical variables of abnormal distribution were compared using Chi-square test with Yates correction or Fischer ’s exact test. Quantified Spearman’s rho correlation coefficients were used to assess the linear correlations between variables. Multiple regression analysis with PTH as a dependent variable led to a construction of two PTH serum concentration models. The basic model consisted of arbitrarily included parameters while the secondary model was constructed on the base of the backward stepwise method. Differences between the values were medical package (Statsoft Inc.).

Results

The heart failure group (65 participants) and the control group (62 participants) did not differ in age (51.6 vs. 53.8, P = ns) and gender (males 80% and 70%, respectively, P = ns). Group characteristics are presented in Table I.

The mean PTH concentration in the non-ischaemic HF group was 79.5 pg/mL, and the range was 3.9–198.8 pg/mL, SD 46.4 pg/mL. In the control group, the mean PTH concentration was 59.6 pg/mL, and the range was 18.3–124.8 pg/mL, SD 23.8 pg/mL. The mean values of serum PTH concentrations were significantly higher in the heart failure group (P = 0.009) (Fig. 1).

The mean PTH values in specific groups of patients divided according to the NYHA functional class were as follows: NYHA I (mean 58.8 pg/mL, range 6.8–77.5 pg/mL), NYHA II (mean 74.4 pg/mL, range 3.9–198.8 pg/mL), NYHA III (mean 87.0 pg/mL, range 8.3–168.3 pg/mL), and NYHA IV (mean 101.2 pg/mL, range 47.3–162.8 pg/mL), which trended towards statistical significance (P = 0.082) (Fig. 2).

The mean Vitamin D concentration was significantly lower in the HF group (23.4 ng/mL vs. 28.1 ng/mL, P = 0,006), although in both groups the mean concentration was below optimal range (30 ng/mL). In our group Vitamin D deficiency was not significantly correlated to the disease severity (NYHA class) nor another assessed marker of poor prognosis.

Patients with HF, who declared a sun exposure time greater than seven hours a week, had significantly higher serum Ca2+ concentrations (10.0 vs. 9.5 ng/mL, P < 0.025) and were significantly younger (43.6 vs. 53.4 years, P < 0.01) compared to those who declared the lesser sun exposure values. Differences between other analysed parameters (PTH, P, vitamin D) were statistically insignificant. Non-HF volunteers declaring a sun exposure time greater than seven hours a week had higher serum concentrations of vitamin D than those with lower values (30.3 vs. 23.8 ng/mL, P < 0.05), unlike the HF patients (24.1 ng/mL vs. 23.2 ng/mL, P = ns). Differences among other values (PTH, P, Ca2+) were statistically insignificant (Table II).

The Spearmann coefficient analysis revealed a statistically significant correlation between the serum concentration of PTH in HF patients and established factors of worse prognosis, such as age (r = 0.26, P = 0.035), creatinine concentration (r = 0.26, P = 0.037), AST (r = 0.32, P = 0.009), NT-proBNP (r = 0.37, P = 0.003), and eGFR (r = –0.36, P = 0.003). Moreover, statistically significant correlations were observed between PTH and Ca2+ (r = –0.29, P = 0.02) and PTH and P (r = –0.35, P = 0.004). A correlation between PTH and haemoglobin (HGB) showed a trend towards statistical significance (r = –0.21, P = 0.093). All values are presented in Table III.

The multiple regression analysis with PTH as a dependent variable enabled the construction of a primary model, which consisted of subsequent arbitrarily included parameters: age, LVEF, eGFR and serum blood concentrations of vitamin D, phosphorus, and NT-proBNP. Further calculations according to the backward stepwise method led to the construction of the secondary model, which included age, phosphorus, and NT-proBNP serum blood concentrations. P values were 0.004, 0.012, and 0.022, respectively. All data are presented in Table IV.

Discussion

According to our best knowledge the issue concerning the dependency between serum PTH concentration in HF patients of non-ischaemic origin and sun exposure in comparison to non-HF volunteers has not been elucidated. The data considering vitamin D and PTH concentration in the Caucasian population of HF patients living in Central Europe remain scarce and unclear. PTH is not routinely evaluated in HF patients, although it may be a potentially important factor in the pathogenesis of HF and is an important clinical factor in anticipating worse prognosis of the disease [6, 7]. It is proposed that maintaining proper PTH serum blood concentration, by normalising serum Ca2+ and vitamin D concentrations, effective treatment of CKD, or potentially usage of calcimimetics, may play a positive role in HF patients treatment and decision-making processes [8]. Elevation of serum PTH is induced by several mechanisms, such as activation of the renin-angiotensinaldosterone system (RAAS), which results in an increase in renal Ca2+ ion depletion, a high rate of diuretic intake in this group of patients, progressive renal degeneration due to chronic hypoperfusion resulting in excessive PTH retention, and vitamin D deficiency due to a decrease in physical activity with a subsequent decrease in sun exposure and induction of PTH secretion [9]. Serum PTH concentration is increased by neurohormonal system activation, which primarily plays a compensatory role in the early stage of HF; however, subsequently, it has a growing decompensating impact on the course of the disease [10]. Myocardial cells express PTH-specific receptors, whose stimulation might be responsible for hypertrophy, apoptosis, and fibrosis, which is directly induced by either the activation of protein kinases A and C or arterial hypertension [11]. Numerous biological processes are considered to be crucial in improving or worsening HF [12, 13]. It is widely speculated that elevated PTH contributes to augmentation of catabolic cellular activity and oxidative stress. Pilz et al. proved the impact of increased PTH for cardio-vascular and overall mortality in a large cohort of patients referred to coronary angiography [14]. The hazard ratio (HR) for occurrence of death during a nearly eight-year follow-up period between the third and fourth PTH quartiles was 1.85 for sudden cardiac death and 1.94 for HF cause. Pro-atherosclerotic impact of PTH, although often hypothesised, remains not fully elucidated. Despite the fact that LURIC study investigators failed to assess a clear link between PTH and angiographic stage of coronary heart disease (CAD), they assumed the association between NT-proBNP, NYHA class, and PTH being an independent predictor of hospitalisation for heart failure [15]. This was consistent also with our study in which we found the relation between PTH and age, NT-proBNP, and phosphorus serum blood concentrations. CAD as a cause of HF was in these cases meticulously excluded by the coronary angiography without significant lesions performed within twelve months prior to the examination. PTH concentration could have been affected by a high intake percentage of diuretics and angiotensin-converting enzyme inhibitors (ACE-I) (88% and 100% of patients, respectively), as well as by the occurrence rate of CKD in the third stage with eGFR < 60 (mL/min/1.73 m2) in 14% of the patients. It is worth mentioning that in the publication of Schierbeck et al. the percentage of diuretics and ACE-I intake were similar and varied between 76–97% and 57–81%, respectively, in patients divided into quartiles according to PTH serum concentration [16]. Moreover, investigators found a positive correlation of PTH with age, NTproBNP, and alkaline phosphatase – Pearson correlation coefficients were 0.233, 0.365, and 0.322 (P = 0.004, P < 0.001, and P < 0.001), respectively, and were coherent with our results. Reported negative correlation of PTH with ionised calcium (–0.354, P < 0.001), eGFR (–0.487, P < 0.001), and haemoglobin (–0.208, P = 0.01) support our findings, which are similar although performed in a different group of patients hospitalised due to HF. Correlation between PTH and increased risk ratio of cardiovascular death in elderly males and risk of heart failure occurrence in the population of all patients was demonstrated by Hagstrom et al. [17]. Although there have been several studies denying the role of PTH as a factor correlating with the occurrence of cardiovascular disease [18, 19], in the vast majority of publications PTH is recognised as a predictor of hospitalisation due to HF decompensation and cardiovascular, as well as all-cause mortality [5, 14, 16, 20]. The investigation of Gruson et al. revealed that the serum concentration of bioactive PTH in hospitalised HF patients, as assessed using the third-generation test, was elevated in comparison to healthy volunteers and was correlated with NYHA class and neurohormonal activation, which was also present in our study [21]. Interestingly, this dependency characterised predominantly ischaemic compared to non-ischaemic HF patients. The relationship between PTH concentration and the severity of HF in 150 consecutive patients with HF treated in the outpatient clinic shown by Altay et al. was also consistent with our findings [22]. The study of Sugimoto et al. showed that lownormal serum PTH concentration on admission was related to the increased incidence of in-hospital mortality due to acute decompensated heart failure compared to high PTH values (23.1 vs. 11.4, P = 0.013) [23]. Since the deceased patients had significantly higher incidence of diabetes and lower incidence of dyslipidaemia (P = 0.03 and P = 0.045, respectively) they probably represented the group of patients with initially increased risk of not only short- but also long-term mortality. It was reported that impaired secretion of PTH within the physiologic range has a negative impact on contractility of cardiomyocytes [24]. Moreover, the activation of the adrenergic nervous system expressed by the elevation of serum catecholamines results in intracellular calcium overloading and augmentation of oxidative stress. These discrepancies in PTH serum concentrations can be largely explained by the initially different clinical course of acute and chronic heart failure. Polat et al. demonstrated the relationship between 25OHD3 and the echocardiographic findings connected with the negative LV remodelling [25]. We evaluated the possible correlation between PTH and LVEF, which turned out to be statistically insignificant. It might have been caused by the relatively wide interquartile range of LVEF assessments (16–30%) as well as by a frequently observed incoherence between the clinical status and LVEF examination. Serum blood concentrations of vitamin D in HF patients of the Polat et al. study were 24.1 ± 10.4, which was comparable to our findings 23.4 ± 13.0; however, we observed the discrepancy between vitamin D serum blood concentration among the control subjects which were 41.4 ± 20.9 and 28.1 ± 11.3, respectively. At least partially it could be explained by higher solar Direct Normal Irradiation (DNI) in Turkey in comparison to Poland (2000 vs. 1000 kWh/m2+) [26]. Interestingly vitamin D serum blood concentrations in patients with HF did not vary significantly according to declared sun exposure (24.1 vs. 23.2 ng/mL, P = ns), which suggests a more complex restitution background of vitamin D in the HF setting. In this study, intact-PTH serum concentration (second-generation assessment) in non-ischaemic HF patients correlated with well-established markers of unfavourable prognosis (age, NYHA class, eGFR, NT-proBNP, and creatinine). Moreover, the PTH concentration did not correlate with the vitamin D concentration although the prevalence of its deficit was significantly higher among the HF patients compared to the non-HF volunteers (72% vs. 56%, P = 0.046).

Limitations of the study

Certain limitations should be considered when interpreting the findings of our study. Firstly, the analysed data represents a relatively small although highly selected population of consecutive patients with nonischaemic HF. Secondly, the patients were enrolled and evaluated in one centre by the investigators, which might have generated a potential source of selection bias.

Conclusions

PTH concentration is significantly elevated in a Caucasian population of non-ischaemic HF patients compared to non-HF volunteers. This elevation is correlated with established factors of worse prognosis of HF such as age, eGFR, and serum concentrations of phosphorus and NT-proBNP. Declared sun exposure did not affect the serum concentration of vitamin D in the HF group compared to the control group.

Acknowledgements

The research was supported by Medical University of Silesia – grant number KNW-1-103/P/2/0 granted to AK