Vitamin D receptor (VDR) gene polymorphism is associated with left ventricular (LV) mass and predicts left ventricular hypertrophy (LVH) progression in end-stage renal disease (ESRD) patients

Abstract

Left ventricular hypertrophy (LVH) is a strong cardiovascular risk marker in end-stage renal disease (ESRD) patients. Vitamin D deficiency and/or disturbed vitamin D signaling has been implicated in LVH in experimental models. Because the BsmI vitamin D receptor VDR gene polymorphism may alter VDR function, we performed a cross-sectional and longitudinal study in a cohort of 182 dialysis patients to investigate (1) the relationship between BsmI VDR gene polymorphism and left ventricular mass index (LVMI) measured by echocardiography and (2) the predictive power of this polymorphism for progression in LVH over a 18 ± 2 months of follow-up. As a reference group, we used 175 healthy subjects matched to the study population as for age and sex. The distribution of BsmI genotypes did not significantly deviate from Hardy-Weinberg equilibrium either in patients or in the control group of healthy subjects. The frequency of the B allele of BsmI polymorphism (40.4%) in dialysis patients was similar to that of healthy control subjects (38.6%), and the number of B alleles was directly related to LVMI (r = 0.20, P = .007). This relationship remained robust (β = 0.19, P = .006) in multivariate analysis adjusting for traditional and nontraditional risk factors and antihypertensive and calcitriol treatment. In the longitudinal study, LVMI rose from 60.1 ± 17.9 to 64.2 ± 19.3 g/m^2.7 (P < .001), and again, the number of B alleles was associated with LVMI changes both in crude and in fully adjusted analyses. These cross-sectional and longitudinal observations coherently support the hypothesis that altered vitamin D signaling is implicated in LVH in ESRD patients. © 2010 American Society for Bone and Mineral Research

Introduction

Left ventricular hypertrophy (LVH) is recognized as one of the strongest risk factors for all-cause and cardiovascular mortality in patients with end-stage renal disease (ESRD).1, 2 The pathogenesis of LVH in ESRD is multifactorial, and arterial hypertension, hyperparathyroidism, severe anemia, hypoalbuminemia, chronic volume overload,3 sympathetic overactivity,4 and accumulation of the endogenous inhibitor of nitric oxide (NO) synthase asymmetric dimethylarginine (ADMA)5 have been implicated as causative mechanisms responsible for LVH in these patients. However, collectively, these risk factors explain only in part the variability in left ventricular (LV) mass in the dialysis population.

Vitamin D insufficiency or deficiency is pervasive in ESRD patients and may have substantial health implications in this population.6 The myocardium is an important target of vitamin D. Vitamin D receptor (VDR) knockout mice show myocardial renin overexpression and marked cardiomyocyte hypertrophy.7 In vitro 1,25-dihydroxyvitamin D₃ [1,25(OH)₂D₃] attenuates cardiomyocyte proliferation8 and hypertrophy,9 and treatment with paricalcitol attenuates the development of LVH and LV dysfunction in Dahl rats,10 a reliable animal model of LVH and vitamin D deficiency. Retrospective analyses in hemodialysis patients and small clinical trials show that treatment with a vitamin D analogue may regress LVH in these patients,11–13 providing further rationale for a full-fledged randomized clinical trial in ESRD. Until now, studies in the VDR knockout mice represent the most convincing evidence of the fundamental role of vitamin D in the regulation of LV mass.

While the clinical trial remains the ultimate test for establishing whether or not vitamin D is implicated in LVH in ESRD, examining the association between VDR gene polymorphisms and LV mass represents an additional relevant step for unraveling the role of vitamin D in cardiac disease in these patients. An important advantage of genetic association studies is that the transmission of genes occurs at random (Mendelian randomization), thus mimicking the design of classical intervention trials.14 Three common polymorphisms (BsmI, ApaI, and TaqI) at the 3' end of the VDR have been intensively investigated, and one of these (BsmI) already has been associated with survival in the dialysis population,15 suggesting that this polymorphism may be implicated in the high risk of this population. However, the association between the BsmI polymorphism and LVH has never been tested. Therefore, we performed an association study investigating the relationship between this polymorphism and LV mass in a cohort of ESRD patients. To minimize the risk that the observed association are merely due to chance, we performed a study including both a cross-sectional analysis of the association between BsmI VDR gene polymorphism and LV mass index (LVMI) and a longitudinal observation aimed at establishing whether or not this polymorphism predicts LVH progression in ESRD. For the first time, we demonstrate that the BsmI VDR gene polymorphism is independently related to LVH and to LVH progression in ESRD patients.

Methods

The study protocol was in conformity with the local ethical guidelines of our institution and informed consent was obtained from each participant.

Patients

One-hundred and eighty-two hemodialysis patients (104 males and 78 females, all Caucasians) who had been on regular dialysis treatment (RDT) for at least 6 months and who were free of overt infections (i.e., fever, infected vascular access, or peritonitis or exit-site infection) were recruited for the study. Patients were being treated thrice weekly with standard bicarbonate dialysis (Na 138 mmol/L, HCO₃ 5 mmol/L, K 1.5 mmol/L, Ca 1.25 mmol/L, Mg 0.75 mmol/L) with either Cuprophan or semisynthetic membranes.

Control subjects

As a control group, we enrolled a series of 175 age- and sex-matched healthy blood donors from the same geographic area of patients.

Genotyping of the BsmI VDR gene polymorphism

The participants were genotyped for the G → A change identified as BsmI polymorphism (bB alleles) in intron 8 of the VDR gene on chromosome 12. This polymorphism, described under identification number rs1544410, was studied by high-throughput TaqMan allelic discrimination assay. Genomic DNA was extracted from peripheral blood leukocytes by the salting-out technique.16 The assay mix (including unlabeled PCR primers FAM and VIC dye-labeled TaqMan MGB probes) of the Assays-on-Demand was designed and provided by Applied Biosystems (Applied Biosystems, Inc., Foster City, CA, USA). The reaction system contained 1 to 5 ng of genomic DNA, 12.5 µL of TaqMan Universal PCR Master Mix, 2 × No AmpErase UNG, and 1.25 µL 40 × Assay Mix and was adjusted with H₂O for a total volume of 25 µL. The genotyping was performed on an ABI Prism 7300, and a random 5% of samples were repeated independently to confirm genotyping results. The genotype results for these samples were completely consistent.

Laboratory measurements

A fasting blood sampling was performed during the midweek nondialysis day. Serum cholesterol, albumin, phosphate, parathormone (PTH), and hemoglobin measurements were made using standard methods in a routine clinical laboratory. Serum C-reactive protein (CRP), plasma total homocysteine, and asymmetric dimethylarginine (ADMA) were measured as previously reported.17 The plasma concentration of norepinephrine was measured by a commercially available radioimmunoassay (RIA) kit (Amicyl-test TM, Immunological Laboratories, Hamburg, Germany). The intraassay coefficient of variation was 7% to 15%.

Echocardiography and blood pressure

These studies were performed on a nondialysis day within 2 hours of blood sampling. All echocardiographic measurements were carried out according to the recommendations of the American Society of Echocardiography by an observer unaware of biochemical results. Left ventricular mass (LVM) was calculated according to the Devereux formula and indexed to height^2.7 (LVMI).18 LVH was defined by an LVMI of over 47 g/m^2.7 in women or over 50 g/m^2.7 in men. The height-based indexing of LVM was specifically chosen to minimize any potential distortion attributable to extracellular volume expansion (surface area indexing being weight sensitive).2

Blood pressure (BP) was measured before dialysis and during all dialysis sessions of the month (i.e., 12 measurements) preceding the baseline echocardiographic study. Average values were considered for statistical analysis. Average monthly BP is associated with LV mass equally well as 24 hour ambulatory BP in dialysis patients.19

Patients who repeated the echocardiographic study

Thirty-four of the 182 patients who entered into this study did not repeat the second echocardiography because of death. Therefore, 148 patients were left for the longitudinal part of this study aimed at examining the relationship between the BsmI polymorphism and changes in LVMI. These patients did not materially differ from the original study population (data not shown).

Statistical analysis

Power analysis

We assumed additivity of the effects of the B allele on LV mass index. Such an assumption is preferred when allele dominance cannot be assumed.20 Nonrandomized parallel studies indicate that the lack of 1,25(OH)₂D₃ treatment may be responsible for a 15% increase in LV mass in ESRD patients.13 We hypothesised that homozygotes for the protective allele of the BmsI polymorphism gene (bb) have a LV mass that is 15% lower than homozygotes for the risk allele (BB) and that LV mass in heterozygotes (Bb) takes an intermediate value (7.5% higher than in bb homozygotes). We estimated that LVMI in Bb patients would be 61 g/m^2.7 (± SD 15), that is the average value of LVMI in hemodialysis patients in our center. On the basis of these data, we calculated that the average value of LVMI would be 56 g/m^2.7 in bb (7.5% lower than in Bb) and 66 g/m^2.7 in BB (7.5% higher than in Bb) genotypes. Power analysis based on these data indicated that a sample size of 180 dialysis patients had an 80% probability to detect as statistically significant (α error = 0.05) a 5 g/m^2.7 average increase in LVMI for each unitary increase in the number of B alleles.

Statistical tests

The relationship between the number of B alleles of BsmI VDR polymorphism and LV mass at baseline and with LV mass changes during the follow-up was investigated by univariate and multiple linear regression analyses (Models 1 to 4 in Tables 2 and 3). As potential confounders we considered Framingham risk factors (e.g., age, sex, smoking, diabetes, systolic pressure and antihypertensive treatment, cholesterol, and previous cardiovascular events), risk factors peculiar to ESRD (e.g., duration of dialysis treatment, haemoglobin, albumin, phosphate, PTH, and fractional urea clearance [Kt/V]), and emerging risk factors (e.g., CRP, homocysteine, ADMA, and norepinephrine). To control for the potential confounding effect of current calcitriol therapy, we also included this factor in the multiple regression analysis. Data are expressed as standardized regression coefficient β and P value. The deviation of Hardy-Weinberg equilibrium was investigated by calculating the chi-square test between observed and expected frequencies. Since we did not test other VDR single-nucleotide polymorphisms (SNPs), no adjustment for multiple testing was done.

Data are expressed as mean ± SD (normally distributed data), median and interquartile range (nonnormally distributed data), or percent frequency, as appropriate, and by trend analysis. Positively skewed distributed data (e.g., CRP, homocysteine, and ADMA) were log transformed (lg₁₀) before data analysis.

All calculations were done using two standard statistical packages (SPSS for Windows, Version 9.0.1, © SPSS Inc., Chicago, IL, USA, and NCSS PASS (2004), Number Cruncher Statistical Systems, Kaisville, UT, USA).

Results

In the entire study population (see Table 1), the median duration of RDT was 45 months (interquartile range 20 to 110 months). The average urea Kt/V in these patients was 1.22 ± 0.28. Twenty-six patients were diabetics, and 68 were habitual smokers (22 ± 16 cigarettes/day). Sixty-nine patients were on antihypertensive drugs [48 on monotherapy with angiotensin-converting enzyme (ACE) inhibitors, AT-1 antagonists, calcium channel blockers, and alpha and beta blockers and the remaining 21 on double or triple therapy with various combinations of these drugs]. One hundred patients were on treatment with erythropoietin and 95 with calcitriol. At baseline, LVMI was on average 61 ± 18 g/m^2.7. The distribution of LVMI in the study population is presented graphically in Fig. 1.

Table 1. Main Demographic, Clinical, Biochemical, and Echocardiographic Data in Hemodialysis Patients

	Vitamin D receptor (BsmI polymorphism)			P for trend
	bb (n = 61)	Bb (n = 95)	BB (n = 26)
Data are expressed as mean ± SD, median and interquartile range, or as percent frequency, as appropriate.
Age (years)	60 ± 13	57 ± 16	60 ± 14	.63
Duration of RDT (months)	47 (21–121)	43 (20–92)	46 (14–121)	.72
Male sex, n (%)	34 (56%)	54 (57%)	16 (62%)	.65
Smoking, n (%)	22 (36%)	38 (40%)	8 (31%)	.83
Diabetes, n (%)	10 (16%)	13 (14%)	3 (11%)	.52
Patients with previous cardiovascular events, n (%)	33 (54%)	44 (46%)	14 (54%)	.74
On antihypertensive treatment, n (%)	23 (38%)	38 (40%)	8 (31%)	.69
On EPO treatment, n (%)	29 (47%)	59 (62%)	12 (46%)	.62
On calcitriol treatment, n (%)	28 (46%)	55 (58%)	12 (46%)	.61
Systolic pressure (mmHg)	137 ± 25	143 ± 24	147 ± 22	.05
Diastolic pressure (mmHg)	74 ± 14	78 ± 13	78 ± 12	.12
Heart rate (b/min)	80 ± 10	80 ± 10	76 ± 9	.27
Cholesterol (mmol/L)	5.26 ± 1.65	5.38 ± 1.37	5.75 ± 1.47	.19
Haemoglobin (g/L)	107 ± 22	106 ± 17	107 ± 20	.96
Albumin (g/L)	41 ± 6	42 ± 4	41 ± 6	.47
Calcium phosphate (mmol^2/L^2)	4.5 ± 1.3	4.6 ± 1.2	4.4 ± 1.8	.95
Phosphate (mmol/L)	1.96 ± 0.50	2.02 ± 0.44	1.99 ± 0.49	.73
PTH (pg/mL)	161 (64–345)	171 (70–405)	145 (53–436)	.51
Norepinephrine (nmol/L)	3.3 (2.0–6.4)	3.1 (1.6–5.3)	3.7 (1.9–5.6)	.60
CRP (mg/L)	9.5 (3.4–20.1)	8.1 (3.4–15.0)	8.0 (3.4–21.1)	.66
ADMA (µmol/L)	23.2 (17.4–35.2)	28.4 (21.0–44.5)	25.2 (19.3–36.7)	.34
Homocysteine (µmol/L)	2.42 (1.55–3.88)	2.88 (1.68–4.08)	3.33 (1.38–4.15)	.23
Kt/V	1.23 ± 0.29	1.20 ± 0.25	1.27 ± 0.32	.74
LVMI (g/m^2.7)	56.4 ± 17.7	62.7 ± 19.1	66.9 ± 15.4	.007

Figure 1. Distribution of left ventricular mass index (LVMI) in the study cohort.

The distribution of BsmI genotypes did not significantly deviate from Hardy-Weinberg equilibrium either in patients (bb genotype 33.5%, Bb genotype 52.2%, and BB genotype 14.3%; χ² = 1.29, P = NS) or in controls (bb genotype 37.8%, Bb genotype 47.4%, and BB genotype 14.9%; χ² = 0.01, P = NS). The frequency of the B allele (38.6%) in healthy subjects was almost identical to that of dialysis patients (40.4%). Data analysis according to BsmI genotypes (see Table 1) revealed that there was a dose-dependent relationship between the number of BsmIB alleles and systolic arterial pressure (bb genotype 137 ± 25 mmHg, Bb genotype 143 ± 24 mmHg, BB genotype 147 ± 22 mmHg; P for trend = 0.048) (see Table 1). No significant difference among the three groups was observed for the remaining variables listed in Table 1.

BsmI genotypes and baseline LVMI: linear regression analyses

On univariate analysis, the number of B alleles was directly and significantly related to LVMI (r = 0.20, P = .007) (Fig. 2, left panel). The independent association between the BsmI polymorphism and LVMI was tested in multiple linear regression models of increasing complexity (Table 2). Data adjustment for age, sex, and duration of RDT did not modify the strength of the B allele–LVMI link (β = 0.21, P = .004) (Model 1), and this also was true when we added into the model other Framingham risk factors (e.g., smoking, systolic pressure, diabetes, cholesterol, and previous cardiovascular events), as well as antihypertensive and calcitriol treatment (β = 0.20, P = .004) (Model 2). Further data adjustment for factors peculiar to ESRD (e.g., hemoglobin, albumin, phosphate, PTH, and Kt/V) (Model 3) and nontraditional risk factors (e.g., CRP, homocysteine, norepinephrine, and ADMA) (Model 4) did not change the strength of the association. In the fully adjusted multiple regression model (Model 4), besides the BsmI polymorphism, serum albumin (β = −0.24, P = .003), systolic arterial pressure (β = 0.24, P = .002), and plasma ADMA (β = 0.17, P = .02) also were significantly and independently related to LVMI.

Figure 2. Relationship between BsmI polymorphism and LVMI and with LVMI changes. Data are mean ± SD.

Table 2. Multiple Linear Regression Models for Baseline LVMI

	Dependent variable: LVMI
	Unadjusted r (P)	Model 1 β (P)	Model 2 B (P)	Model 3 β (P)	Model 4 β (P)
Data are expressed as beta regression coefficients (β) and P values.
BsmI polymorphism	0.20 (.007)	0.21 (.004)	0.20 (.004)	0.19 (.004)	0.18 (.006)
Age		0.27 (<.001)	0.21 (.06)	0.13 (.10)	0.12 (.12)
Sex		0.07 (.37)	0.07 (.40)	0.13 (.10)	0.09 (.28)
Duration of RDT		0.04 (.58)	0.08 (.28)	0.07 (.35)	0.03 (.67)
Smoking			0.01 (.85)	0.04 (.60)	0.07 (.35)
Diabetes			0.05 (.48)	0.10 (.16)	0.09 (.21)
Cholesterol			−0.12 (.09)	0.02 (.80)	−0.04 (.58)
Systolic pressure			0.23 (.003)	0.22 (.003)	0.24 (.002)
Previous cardiovascular events			0.18 (.02)	0.16 (.04)	0.14 (.07)
Antihypertensive treatment			0.12(.12)	0.08 (.31)	0.05 (.55)
On treatment with calcitriol			−0.03(.62)	−0.01 (.83)	−0.05 (.47)
Hemoglobin				−0.12 (.09)	−0.11 (.12)
Albumin				−0.28 (.001)	−0.24 (.003)
Phosphate				0.10 (.15)	0.08 (.22)
PTH				0.12 (.08)	0.07 (.31)
Kt/V				−0.001 (.98)	−0.02 (.75)
CRP					−0.008 (.91)
Homocysteine					−0.008 (.91)
ADMA					0.17 (.02)
Norepinephrine					0.06 (.43)

BsmI genotypes and changes in LVMI

Since baseline LVMI was inversely associated with LVMI changes (r = −0.18, P = .03), indicating regression to the mean, LVMI changes therefore were adjusted for the corresponding baseline values before analyses. As shown in Table 3, the number of B alleles was associated with LVMI changes both in crude (see Fig. 2, right panel) and in adjusted analyses, the correlation coefficient of this relationship remaining virtually unchanged after full data adjustment across models of increasing complexity (β = 0.19, P = .03).

Table 3. Multiple Linear Regression Models for Changes in LVMI

	Dependent variable: changes in LVMI (g/m^2.7/month) adjusted for LVMI at baseline
	Unadjusted β (P)	Model 1 β (P)	Model 2 β (P)	Model 3 β (P)	Model 4 β (P)
Data are expressed as beta regression coefficients (β) and P values.
BsmI polymorphism	0.19 (.02)	0.19 (.02)	0.18 (.04)	0.19 (.03)	0.19 (.03)
Age		0.01 (.89)	0.04 (.68)	0.05 (.62)	0.07 (.51)
Sex		0.003 (.96)	−0.05 (.62)	−0.10 (.32)	−0.12 (.30)
Duration of RDT		−0.03 (.69)	−0.05 (.58)	−0.04 (.69)	−0.06 (.57)
Smoking			0.06 (.53)	0.03 (.76)	0.04 (.70)
Diabetes			−0.11 (.19)	−0.12 (.17)	−0.10 (.31)
Cholesterol			0.08 (.33)	−0.12 (.21)	−0.15 (.15)
Systolic pressure			0.03 (.74)	0.04 (.70)	0.03 (.74)
Previous cardiovascular events			−0.02 (.82)	0.004 (.97)	−0.02 (.80)
Antihypertensive treatment			0.007(.94)	0.04 (.68)	0.03 (.74)
On treatment with calcitriol			0.15(.07)	0.16 (.07)	0.16 (.07)
Hemoglobin				0.12 (.20)	0.11 (.25)
Albumin				0.10 (.34)	0.08 (.47)
Phosphate				−0.08 (.35)	−0.07 (.48)
PTH				0.17 (.05)	0.12 (.19)
Kt/V				−0.06 (.57)	−0.06 (.56)
CRP					−0.003 (.97)
Homocysteine					0.13 (.19)
ADMA					0.02 (.85)
Norepinephrine					0.07 (.49)

Discussion

This study shows that in chronic dialysis patients the frequency of the B allele of the BsmI VDR gene polymorphism is independently related to LVH and associated with higher progression rate of LVH. Overall, these genetic associations are compatible with the hypothesis that altered vitamin D metabolism exerts an important influence on the control of LV mass in this population.

LVH is considered to be a clinical indicator integrating the long-term exposure not only to pressure overload but also to several hemodynamic and nonhemodynamic factors. Regression of LVH is associated with a 59% lower risk of subsequent adverse events compared with the persistence or new development of LVH.21 LVH is exceedingly frequent in dialysis patients, with a prevalence ranging from 60% to 78%. In line with studies in other conditions, cohort studies in dialysis patients solidly link LVH progression and mortality22 and improved survival when LVH regresses.23 Even though several causative factors are implicated in high LV mass in the dialysis population, regression of LVH in this population is difficult to achieve, and even a most intensive treatment strategy including long nocturnal dialysis failed to normalize LV mass in several patients.24

Even though experimental evidence linking vitamin D deficiency and alterations in LV mass and function appears coherent and convincing,7–9 the issue received only scant attention in dialysis patients. In an uncontrolled series of 12 dialysis patients,11 treatment with 1α-hydroxycholecalciferol for 6 weeks elicited an increase in LV contractility, as measured by fractional fiber shortening and mean velocity of fiber shortening. In a nonrandomized parallel study, intravenous calcitriol caused a marked reduction in LVH, whereas LVH did not regress in untreated patients. Apart from these small nonrandomized studies, until now, there has been no study exploring the link between the vitamin D system and LV mass in this population.

The response to vitamin D compounds critically depends on the VDR functioning. In humans, three common polymorphisms (BsmI, ApaI, and TaqI) at the 3' end of the VDR have been identified, and these polymorphisms have been linked to a variety of diseases, from low bone mass density (BMD)25 to type 1 diabetes and cancer.26 In ESRD, the B allele of the BsmI polymorphism was associated with a low BMD in patients younger than 65 year of age26 and with inflammation, anemia, and resistance to erythropietin.27 Of note, patients homozygous for the B allele display a substantially shorter survival than heterozygotes or homozygotes for the b allele.15 Given the exceedingly high rate of cardiovascular complications in ESRD, it is plausible that such an association underlies disturbed control of a vitamin D–dependent mechanism affecting the cardiovascular system. To date, there have been no studies investigating the BsmI polymorphism as related to cardiovascular disease neither in the general population nor in the dialysis population.

The consequences of vitamin D deficiency and vitamin D resistance in ESRD depend on a large number of factors. Therefore, confounding is a real possibility in observational studies based on the measurement of vitamin D plasma levels or on vitamin D treatment status (treated versus untreated). Mendelian randomization14 is an interesting option to explore the nature of the LV mass– vitamin D association. Unlike classic epidemiology studies, SNP association studies are unconfounded by behavioral and environmental factors because these factors usually do not alter genotype,14 thus representing a useful approach in studies focusing on the ESRD population.15 As predicated by Mendialian randomization theory, with the exception of blood pressure (see below), risk factors were evenly distributed among patients categorized on the basis of the B allele. We hypothesised that the B allele of the BmsI polymorphism may serve as a marker of altered vitamin D signaling in ESRD patients and assumed that this alteration in BB homozygotes produces an increase in LV mass quantitatively similar to the difference in LV mass between vitamin D–treated and untreated patients.13 At baseline, in our cohort we found an association between the B allele and LVMI. This association was almost unmodified after extensive statistical adjustment for other risk factors, including calcitriol treatment, serum phosphate, and PTH. The independent association of the B allele with systolic blood pressure is of interest in that hypertension is a phenotypic characteristic of the VDR knockout model and because a link between this polymorphism and BP has been noted very recently in another study in ESRD.28 This phenomenon suggests that altered signaling may also contribute to raise BP in ESRD patients. Both 25-hydroxycholecalciferol and 1,25(OH)₂D₃ levels are strongly associated with arterial distensibility, pulse wave velocity, and endothelial function, as measured by the forearm blood flow response to hand warming in ESRD patients,29 indicating a relevant impact of vitamin D status on the control of large and small arterial vessels. Beyond baseline associations, we also found that the B allele predicts LVH progression. This relationship is remarkable in that it also persisted virtually unchanged in a multivariate model considering the regression-to-the-mean phenomenon, which is an important confounder in longitudinal studies of LV mass in ESRD patients.22 Over half the patients in this study were on long-term treatment with calcitriol, but apparently this treatment had no effect on LV mass index nor modified the association between the B allele and this index. Because confounding by indication cannot be eliminated safely by statistical adjustment, we believe that no conclusion can be drawn on the basis of our data on the effect of treatment with activated vitamin D compounds on LVH in ESRD. An ongoing double-blind, placebo-controlled, randomized clinical trial testing the effect of an active form of vitamin D (paracalcitol) on LVMI in ESRD patients (http://clinicaltrials.gov/ct2/show/NCT00616902?term=PRIMO&rank=2) will provide a definitive answer to the question of whether vitamin D administration can reverse LVH in ESRD patients.

In conclusion, in dialysis patients, the B allele of the BsmI VDR gene polymorphism is strongly and independently related to LVH and is associated with a higher progression rate of LVH. Since genes are randomly transmitted (Mendelian randomization), our cross-sectional and longitudinal observations consistently support the hypothesis that altered vitamin D signaling is implicated in LVH in ESRD patients.

Disclosures

The authors state that they have no conflicts of interest.