Vitamin D receptor knockout mice develop typical signs of congestive heart failure (CHF). In approximately 20% of stable CHF patients, frankly low concentrations of the vitamin D hormone calcitriol are found.
Vitamin D receptor knockout mice develop typical signs of congestive heart failure (CHF). In approximately 20% of stable CHF patients, frankly low concentrations of the vitamin D hormone calcitriol are found.
We investigated whether serum calcitriol concentrations predict clinical outcome in end-stage CHF.
We collected blood samples in 383 end-stage CHF patients who were on a waiting list for cardiac transplantation. We assessed associations of calcitriol with disease severity and freedom from event (death or cardiac transplantation) during 1-year follow-up. In electively listed patients (n=325), 31% had deficient calcitriol levels (<43pmol/l) compared to 47% in urgently/high urgently listed patients (n=58; P<0.001). As determined by multivariable logistic regression, calcitriol was an independent predictor of the listing status ‘urgent/high urgent’ (P<0.001). Calcitriol concentrations were also significantly lower in patients with an event (n=233) compared to those who survived on the waiting list (P<0.001). Cox regression analysis revealed that patients in the highest calcitriol tertile had a hazard ratio (95% CI) for an event of 0.506 (0.334–0.767) compared with patients in the lowest calcitriol tertile (P=0.005), after adjustment for potential confounders.
Data indicate that low serum calcitriol concentrations are independently associated with poor clinical outcome in end-stage CHF.
Approximately 10 million Europeans and 5 million Americans suffer from congestive heart failure (CHF) [1,2]. We have recently hypothesized that low vitamin D status may contribute to the aetiology of CHF . In line with this suggestion, vitamin D receptor knockout mice develop typical signs of CHF such as cardiac hypertrophy, over-stimulation of the renin-angiotensin-aldosterone-system, high blood pressure, and increased levels of atrial natriuretic peptide [4  6]. In experimental animals with hyperaldosteronism, the vitamin D hormone calcitriol and dietary calcium and magnesium supplements can prevent both oxidative stress and an increase in cytosolic free ionized calcium . These pathophysiological alterations are typical findings in CHF [7,8].
In humans, the principal source of vitamin D is synthesis in the skin after radiation with the solar ultraviolet B spectrum. Vitamin D is then metabolized by a hepatic hydroxylase into 25-hydroxyvitamin D (25(OH)D) and by a renal 1α-hydroxylase into calcitriol. This step is under control of parathyroid hormone (PTH). In case of vitamin D deficiency/insufficiency, renal synthesis of calcitriol becomes substrate dependent. i.e. dependent on the circulating 25(OH)D concentration . Moreover, renal calcitriol synthesis is reduced in patients with kidney disease . There is also evidence that pro-inflammatory cytokines such as tumour necrosis factor (TNF)-α can suppress calcitriol concentrations .
Frankly low serum calcitriol concentrations (<37 pmol/l) occur in approximately 20% of stable CHF patients . Several disease-related alterations such as limited solar ultraviolet B exposure, impaired renal function, and elevated TNF-α levels may explain this relatively high prevalence of calcitriol deficiency in CHF patients.
Patients with end-stage CHF have very high 1-year mortality rates of up to 50 % and more . In this particular group, vitamin D status has not been characterized yet. We therefore aimed to study vitamin D status in end-stage CHF patients and to investigate whether serum calcitriol levels are associated with clinical outcome in these patients.
We consecutively recruited all end-stage CHF patients at our institution, who were on the Eurotransplant waiting list (Leiden, The Netherlands) for cardiac transplantation between May 2004 and April 2006. Only patients with cardiac re-transplantation (n=3) or an age of less than 18 years were excluded (n=8). Thus, we could include 383 patients. Eurotransplant has strict listing criteria: Elective candidates must have a cardiac index (CI) below 2.2 l/min/m2 and an LVEF below 35%. In order to be listed ‘urgent’, patients must be hospitalised, their CI must not exceed 2.1 l/min/m2, and their LVEF must not exceed 30% while they are being treated with catecholamines. For the classification ‘high urgent’, patients have to fulfil the same criteria as for urgent listing. In addition, they have to receive dobutamine doses of at least 7.5 μg/kg body weight/min or milrinone doses of at least 0.49 μg/kg body weight/min. Out of the 383 patients that were included, listing status was elective in 325 patients and urgent/high urgent in 58 patients. The majority of the study cohort was male (81.1%). All patients gave written informed consent to the study procedures. The study was approved by the local Ethics Committee.
The study design includes a cross-sectional and a prospective part. In the cross-sectional part, we divided the study cohort into two groups. Group 1 (designated EL) consists of the electively listed patients for cardiac transplantation, whereas group 2 (designated UL) consists of the patients who were listed as ‘urgent/high urgent’. We examined whether calcitriol is a predictor of the listing status UL. In the prospective part of the study, we followed the patients on the waiting list for up to 1 year after blood sampling. Thereafter, we examined whether serum calcitriol predicted survival during follow-up.
We used the medical records of the patients to assess age, anthropometric data, diagnosis, concomitant diagnosis such as diabetes mellitus and renal insufficiency, concomitant procedures such as previous thoracic surgery and ventricular assist device implantation, relevant medications, and haemodynamic parameters such as LVEF, CI, left ventricular end-diastolic diameter, and pulmonary vascular resistance. From each patient we collected fasting blood samples between 7:00 AM and 9:00 AM after a 10–12 h overnight fast. In the EL group, all blood samples were drawn during a regular outpatient visit. These visits are generally necessary for status re-evaluation every 6 months. In the UL group, we collected the blood samples within the first 3 days of hospital admission. Blood samples were centrifuged at 1500 g immediately after being drawn and aliquots were then frozen at −80 °C until analysis. We categorized serum concentrations of 25(OH)D, which is the hallmark for determining vitamin D status, on the basis of published cut-offs: <25 nmol/l for deficiency, 25–49.9 nmol/l for insufficiency, 50–74.9 nmol/l for borderline status, and ≥75 nmol/l for normal status . We considered a calcitriol concentration below the reference range of the assay (<43 pmol/l, see below) as deficient. However, because no generally accepted cut-offs for calcitriol categories exist, we also divided calcitriol concentrations into tertiles for statistical analysis (see Results section).
We recorded all deaths that occurred during follow-up. Causes of death were classified as sudden cardiac death, cardiac death, and non-cardiac death such as infection, multiple organ failure, stroke, renal failure, and intestinal failure. An event was considered if a patient died or had to be transplanted during follow-up. Patients with an event are designated non-survivors.
We measured 25-hydroxyvitamin D [(25OH)D] by radioimmunoassay (DiaSorin, Stillwater, MN, USA) and the vitamin D-hormone calcitriol using a competitive enzyme-linked immunosorbent assay (ELISA) (Immundiagnostik, Bensheim, Germany) after solid-phase extraction (reference range: 43–132 pmol/l). Intra- and interassay coefficients of variation (CV) for the two vitamin D metabolites were below 7.0% and below 9.0%, respectively. We used highly sensitive ELISA kits (R&D, Minneapolis, MN, USA) to analyze plasma concentrations of interleukin 6 (IL-6), interleukin 10 (IL-10), human active and pro-Matrix Metalloproteinase 2 (total MMP-2) and 9 (total MMP-9), and human tissue inhibitor of metalloproteinases 1 (TIMP-1). Intra- and interassay CVs were below 10.0%. We measured sodium (Na), magnesium (Mg), calcium (Ca), creatinine (Crea), and high sensitive C-reactive peptide (hsCRP) using the Architect autoanalyzer (Abbott, Wiesbaden, Germany), N-terminal propeptide of brain natriuretic peptide (NTproBNP) was measured by Elecsys1010 (Roche, Mannheim, Germany), and tumor necrosis factor(TNF)-α and intact parathyroid hormone (PTH) by Immulite (DPC, Bad Nauheim, Germany).
Categorical variables are reported using the percentage of observations. Continuous variables are expressed as mean and SEM. We tested the normal distribution of the data using the Kolmogorov-Smirnov test. Normal distribution was considered if P-values were above 0.05. Data were then evaluated using the unpaired t-test (normal distributed data). Non-normal distributed data (hsCRP, NTproBNP, 25(OH)D, PTH) were log transformed before analysis. Fisher's exact test was used to test differences in categorical variables. We used Spearman's rank correlation coefficient (rs) and multiple linear regression analysis to assess interrelationships between continuous variables.
The association between calcitriol, other biochemical parameters, and clinical variables with listing status was assessed by means of logistic regression. Variables with univariable probability values of P<0.1 were included in a stepwise logistic regression model. Kaplan-Meier curves were generated to investigate the association of calcitriol tertiles with the probability of an event during follow-up as a function of time after blood sampling. The log-rank test was used to test for differences in event rates between tertiles. We examined the associations between serum calcitriol and the risk of an event using univariable and multivariable Cox proportional hazard analysis. The number of variables that can be included for multivariable testing is limited and equals the square root of the number of events . Therefore, we included only factors in the multivariable analysis that differed significantly between survivors and non-survivors (see below). Results are presented as hazard ratios (HRs) with 95% confidence interval (CI). We used the statistical software package SPSS, version 14 (Chicago, Illinois, USA), to perform the analyses.
Characteristics of the EL group and the UL group are presented in Table 1. The hemodynamic parameters mirror CHF severity of the two groups. The use of inotropics was significantly higher, whereas the use of digitalis glycosides was significantly lower in the UL group compared with the EL group. Generally, both study groups had a high percentage of 25(OH)D and calcitriol values lying in the deficiency range. However, concentrations of vitamin D metabolites were more affected in the UL group than in the EL group. In detail, 25(OH)D concentrations were below 25 nmol/l, 50 nmol/l and 75 nmol/l in 50.2%, 86.2%, and 95.4%, respectively, in the EL group. The corresponding values in the UL group were 56.9%, 94.8%, and 100%, respectively. In the EL group and the UL group, 30.8% and 46.6% of patients respectively, had calcitriol concentrations in the deficiency range. Serum calcitriol levels did not differ in patients with dilated cardiomyopathy or ischaemic heart disease (data not shown). The percentage of previous thoracic surgery was lower and time on the waiting list was shorter in the UL group than in the EL group. In addition, serum concentrations of PTH, creatinine, sodium, and MMP-2 were lower in the UL group than in the EL group. In the UL group only 15.5% of patients had biochemical signs of hyperparathyroidism (PTH levels>65 pg/ml), whereas in the EL group 46.9% had elevated PTH levels.
|Parameter||Elective (n=325)||Urgent/high urgent (n=58)||P-value|
|Sex distribution (%males)||81.0||81.6||0.921|
|Time on the waiting list until blood collection (days)||234±36||103±36||0.009|
|Dilated cardiomyopathy (%)||39.7||50.0||0.191|
|Coronary heart disease (%)||47.9||41.4||0.392|
|Diabetes mellitus (%)||27.5||15.5||0.341|
|Renal insufficiency (%)||32.1||22.4||0.319|
|Previous thoracic surgery (%)||84.7||67.2||0.005|
|Ventricular assist device implantation (%)||26.1||27.8||0.871|
|25-hydroxyvitamin D (nmol/)||35.0±3.0||23.3±2.0||0.020|
|Parathyroid hormone (pg/ml)||78.8±3.7||40.2±3.5||<0.001|
|C-reactive peptide (mg/l)||2.03±0.21||3.02±0.64||0.068|
|Tumour necrosis factor-α (pg/ml)||11.9±0.5||12.6±0.87||0.166|
|N-terminal propeptide of brain natriuretic peptide (pg/ml)||4093±352||3452±582||0.404|
|Matrix metalloproteinase 2 (ng/ml)||294±5||262±9||0.017|
|Matrix metalloproteinase 9 (ng/ml)||583±22||535±44||0.334|
|Tissue inhibitor of matrix metalloproteinases-1 (ng/ml)||181±5||209±27||0.604|
|Cardiac index (l/min/m2)||1.77±0.04||1.81±0.09||0.573|
|Left ventricular ejection fraction (%)||29.4±0.7||28.0±1.4||0.480|
|Left ventricular end-diastolic diameter (mm)||70.0±0.8||67.9±2.0||0.495|
|Pulmonary vascular resistance (dyn*sec/cm5)||252±11||224±20||0.775|
As determined by logistic regression, only calcitriol (P<0.001) and PTH (P<0.001) were significantly associated with listing status. Higher serum concentrations of both parameters predicted a lower risk of urgent/high urgent listing. The HR per pmol/l increase in calcitriol was 0.983 (CI 0.979; 0.993) and the HR per pg/ml increase in PTH was 0.981 (CI 0.972;0.990). Calcitriol was inversely associated with TIMP-1 (rs=−0.322; P<0.001), NT-proBNP (rs=−0.133; P=0.012), CRP (rs=−0.218; P<0.001), IL-6 (rs=−0.213; P=0.003), and TNF-α (rs=−0.188; P<0.001). In addition, calcitriol was directly related to 25(OH)D (rs=0.206; P<0.001) and PTH (rs=0.155; P=0.003). Calcitriol concentrations did not differ according to drug use and were unrelated to haemodynamic parameters (data not shown). As determined by multiple regression analysis with serum calcitriol as independent variable, calcitriol was inversely associated with creatinine (P<0.030), IL-6 (P<0.001), and TIMP-1 (P<0.001), and directly associated with 25(OH)D (P<0.028) and PTH (P<0.045) (multiple r=0.346; P<0.001). PTH concentrations were more closely related to NT-proBNP (rs=0.273; P<0.001) than to 25(OH)D (rs=0.030; n.s.) or calcitriol (see before).
Two-hundred and thirty three patients were considered non-survivors during follow-up. Of these 233 patients, 78 patients (20.4%) died, whereas 155 had to be transplanted (40.5%). Causes of death were sudden cardiac death (N=10), cardiac related death (N=23), and non-cardiac related deaths (N=45) such as multiple organ failure (N=26), sepsis or infection (N=9), renal failure (N=8), intestinal failure (N=1), and stroke (N=1). Characteristics of survivors and non-survivors are given in Table 2. Non-survivors had lower concentrations of calcitriol, magnesium, and sodium, and higher concentrations of CRP, IL-6, TNF-α, NT-proBNP, and TIMP-1. Out of the 233 non-survivors, 42.1% had calcitriol concentrations below 43 pmol/l.
|Parameter||Survivors (n=150)||Non-survivors (n=233)||P-value|
|Sex distribution (%males)||80.9||81.2||0.667|
|Cardiac index (l/min/m2)||1.81±0.06||1.75±0.07||0.335|
|Left ventricular ejection fraction (%)||32.8±0.9||29.8±1.3||0.380|
|Left ventricular end-diastolic diameter (mm)||70.1±0.9||69.2±1.8||0.197|
|Pulmonary vascular resistance (dyn*sec/cm5)||263±18||285±29||0.476|
|25-hydroxyvitamin D (nmol/l)||33.3±2.8||31.8±4.7||0.602|
|Parathyroid hormone (pg/ml)||81.1±5.3||70.0±4.2||0.071|
|C-reactive peptide (mg/l)||0.91±0.11||3.08±0.32||<0.001|
|Tumour necrosis factor-α (pg/ml)||11.1±0.6||12.7±0.6||0.019|
|N-terminal propeptide of brain natriuretic peptide (pg/ml)||3352±326||4549±481||0.019|
|Matrix metalloproteinase 2 (ng/ml)||291±7||286±6||0.736|
|Matrix metalloproteinase 9 (ng/ml)||525±26||612±29||0.185|
|Tissue inhibitor of matrix metalloproteinases 1 (ng/ml)||159±8||204±9||<0.001|
In Fig. 1, survival rates are presented according to calcitriol tertiles. The Kaplan-Meier curves illustrate a lower 1-year survival in those patients with lower calcitriol concentrations (P<0.001; log-rank test). In detail, survival rates were 61.4%, 40.3%, and 25.7% in patients in the highest, intermediate, and lowest calcitriol tertile, respectively. Thus, the unadjusted event rate was 2.4 times higher in patients in the lowest calcitriol tertile compared with patients in the highest tertile. Univariable Cox proportional hazard analysis showed that higher calcitriol concentrations and higher magnesium and sodium concentrations were associated with higher survival rates, whereas serum concentrations of some other biochemical parameters were inversely associated with survival (Table 3). We then investigated the association between calcitriol tertiles (independent variable) and survival (dependent variable) in multivariable analyses (Table 4). In the first model, unadjusted data are presented. The second model added adjustment for renal function (serum creatinine concentration). The third model further added adjustment for the inflammation markers CRP, IL-6, and TNFα. In a final model, we also adjusted for magnesium, sodium, TIMP-1 and NTproBNP. The relation between calcitriol tertile and risk of an event remained significant after all adjustments. Adjusted risk was 1.98 times higher in patients in the lowest calcitriol tertile compared with patients in the highest tertile. When calcitriol was used as continuous variable, the HR per pmol/l increase in calcitriol was 0.972 (95% CI: 0.962;0.982) for model 1, 0.972 (95% CI: 0.962;0.982) for model 2, 0.978 (95% CI: 0.968;0.989) for model 3, and 0.982 (95% CI: 0.971;0.994) for model 4.
|Parameter||Analysis for continuous variables||Top tertile versus lower 2 tertiles|
|HR (95% CI)||P-value||Cut-off value||HR (95% CI)||P-value|
|Calcitriol (pmol/l)||0.974 (0.964–0.93)||<0.001||73 pmol/l||0.449 (0.318–0.634)||<0.001|
|C-reactive protein (mg/dl)||1.094 (1.067–1.121)||<0.001||1.54 mg/dl||2.631 (2.001–3.459)||<0.001|
|Interleukin-6 (pg/ml)||1.024 (1.015–1.024)||<0.001||34 pg/ml||1.743 (1.320–2.295)||<0.001|
|Tumour necrosis factor-α (pg/ml)||1.026 (1.009–1.043)||0.003||12.2 pg/ml||1.324 (1.004–1.747)||0.047|
|Magnesium (mmol/l)||0.579 (0.391–0.857)||0.006||2.21 mmol/l||0.656 (0.487–0.882)||0.005|
|Sodium (mmol/l)||0.945 (0.915–0.976)||0.001||140 mmol/l||0.609 (0.423–0.875)||0.007|
|Tissue inhibitor of matrix metalloproteinases 1 (ng/ml)||1.002 (1.001–1.003)||<0.001||195 ng/ml||3.580 (2.292–5.592)||<0.001|
|n||1-year survival (%)||Model 1 HR (95% CI)||Model 2 HR (95% CI)||Model 3 HR (95% CI)||Model 4 HR (95% CI)|
|43–73 pmol/l||128||40.3||0.653 (0.491–0.895)||0.655 (0.484–0.887)||0.775 (0.553–1.023)||0.837 (0.603–1.163)|
|>73 pmol/l||128||61.4||0.362 (0.250–0.526)||0.356 (0.244–0.518)||0.445 (0.301–0660)||0.506 (0.334–0.767)|
|P for trend||<0.001||<0.001||<0.001||0.005|
To our knowledge, this is the first study investigating the association of calcitriol concentrations with clinical outcome in end-stage CHF patients. The study demonstrates that low circulating calcitriol concentrations are more often found in urgent/high urgent candidates for cardiac transplantation than in elective candidates. An association between lower calcitriol concentrations and higher risk of adverse events such as death and the need for cardiac transplantation was also observed. This association remained significant after adjustment for other potential risk factors for survival. The strength of our study is the prospective nature of the event analysis and the adjustment for various potential confounders. Moreover, we were able to investigate a large and homogenous cohort of end-stage CHF patients.
There is increasing evidence that calcitriol deficiency contributes to CHF severity. First, vitamin D receptor knockout mice develop typical signs of CHF [4  6]. Second, calcitriol levels are lowest in those CHF patients with early onset of the disease . Third, calcitriol therapy significantly reduces left ventricular mass in haemodialysis patients . Finally, calcitriol administration can normalize the impaired contractility of the myocardium that is observed under experimental vitamin D deficiency . Our results of an inverse association between calcitriol and TIMP-1 may offer insights into mechanisms of cardiac hypertrophy under calcitriol deficiency. Alterations in the balance of MMPs and TIMPs are involved in impaired left ventricular remodelling . The RNA expression of TIMP-1 is related to the degree of cardiac fibrosis . In vitamin D deficient subjects, vitamin D supplementation significantly reduces TIMP-1 concentrations .
The association of calcitriol with survival in our study is another aspect of clinical interest. There is already evidence from retrospective studies in chronic kidney disease that calcitriol can improve survival [21,22]. Patients with end-stage kidney disease often have calcitriol concentrations that are similar to the levels we observed in our patients in the lowest calcitriol tertile . In haemodialysis patients, Shoji et al.  demonstrated an approximately 3.5 times higher cardiovascular mortality in 1-α vitamin D non-users than in 1-α vitamin D users during a mean follow-up of 61 months. In the very large study of Teng et al. , all-cause mortality was approximately 2.3 times higher in dialysis patients not treated with vitamin D compared to patients treated with active vitamin D during the first year of follow-up (1-year mortality rates of approximately 26% and 11%, respectively). Although non-cardiac related death contributed more than 50% to 1-year mortality in our study, it is likely that CHF-related pathophysiological alterations have contributed significantly to some causes of death such as multiple organ failure, sepsis, and renal and intestinal failure. The role of calcitriol deficiency in cardiac-related and non-cardiac related causes of death is not completely clear. Note that calcitriol receptors are present in most tissues and cells in the body. Calcitriol is able to elicit a wide variety of biologic responses including the modulation of the immune system. Therefore, calcitriol is generally regarded as an important protective factor for cellular health . In line with these considerations, a recent meta-analysis of controlled trials revealed that vitamin D supplementation is able to reduce total mortality in humans .
In patients with kidney disease, serum calcitriol is closely related to creatinine clearance . In the present study, calcitriol concentrations were also correlated with serum creatinine, but were additionally related to the inflammation marker IL-6, and to 25(OH)D and PTH. Inflammation markers such as IL-6, CRP, and TNF-α are well-known risk factors for survival in CHF [25,26]. Since calcitriol can suppress pro-inflammatory cytokines such as IL-6 and TNF-α in vitro  and vitamin D supplements can suppress CRP and pro-inflammatory cytokines in vivo [28,29], low calcitriol levels may contribute to the inflammatory processes in CHF. As mentioned before, there is also evidence that inflammatory cytokines can suppress calcitriol concentrations . Thus, CHF patients may enter a circulus vitiosus of low calcitriol levels and high inflammatory cytokines if calcitriol deficiency persists.
A very high percentage of the entire study cohort had deficient serum 25(OH)D concentrations and some had also secondary hyperparathyroidism. Both, 25(OH)D and PTH were predictors of serum calcitriol. As mentioned before, the deficient 25(OH)D levels may at least in part be due to disease-related limited mobility leading to low ultraviolet-B induced skin synthesis of vitamin D . Despite the generally high percentage of patients with secondary hyperparathyroidism, PTH levels were surprisingly low in some patients with low calcitriol levels, e.g. those patients in the UL group. There is evidence that PTH can increase heart rate, myocardial blood flow, and cardiac output . Therefore, we suggest that higher PTH concentrations in some patients with low vitamin D status were at least in part the result of an adaptation process to the severity of the disease. In line with this assumption, PTH concentrations were more closely correlated with NT-proBNP (Results section), which is an indicator for CHF severity , than with circulating 25(OH)D and calcitriol concentrations. Low PTH levels in combination with low calcitriol concentrations may thus further impair cardiac function. In this context, some may argue that the missing association of calcitriol with haemodynamic parameters is surprising. However, note that end-stage CHF patients receive various cardio-protective drugs, which may have masked an effect of endogenous factors such as calcitriol on haemodynamics. Moreover, in end-stage CHF patients who do not survive or have to be transplanted because of low cardiac output, haemodynamics often worsen markedly only a few days before an event occurs. Therefore, we believe that listing status and freedom from event are more appropriate parameters than haemodynamic parameters to investigate calcitriol effects on clinical outcome in CHF.
In summary, we could demonstrate a high prevalence of low vitamin D metabolite levels in end-stage CHF patients. Our data indicate that deficient circulating concentrations of the vitamin D hormone calcitriol are an independent predictor for poor outcome in these patients.