Effects of Phosphate Binder Therapy on Vascular Stiffness in Early Stage Chronic Kidney Disease

Abstract

Background/Aims

Cardiovascular disease (CVD) is increased in chronic kidney disease (CKD), and contributed to by the CKD-mineral bone disorder (CKD-MBD). The CKD-MBD begins in early CKD and its vascular manifestations begin with vascular stiffness proceeding to increased carotid artery intima-media thickness (cIMT) and vascular calcification (VC). Phosphorus is associated with this progression and is considered a CVD risk factor in CKD. We hypothesized that modifying phosphorus balance with lanthanum carbonate (LaCO3) in early CKD would not produce hypophosphatemia and may affect vascular manifestations of the CKD-MBD.

Methods

We randomized 38 subjects with normophosphatemic stage 3 CKD to a fixed dose of LaCO3 or matching placebo without adjusting dietary phosphorus in a 12-month randomized, double-blind, pilot and feasibility study. The primary outcome was the change in serum phosphorus. Secondary outcomes were changes in measures of phosphate homeostasis and vascular stiffness assessed by carotid-femoral pulse wave velocity (PWV), cIMT and VC over 12 months.

Results

There were no statistically significant differences between LaCO3 and placebo with respect to the change in serum phosphorus, urinary phosphorus, tubular reabsorption of phosphorus, PWV, cIMT, or VC. Biomarkers of the early CKD-MBD such as plasma fibroblast growth factor-23 (FGF23), Dickkopf-related protein 1 (DKK1), and sclerostin were increased 2–3-fold at baseline but were not affected by LaCO3.

Conclusion

12 months of LaCO3 had no effect on serum phosphorus and did not alter phosphate homeostasis, PWV, cIMT, VC, or biomarkers of the CKD-MBD.

INTRODUCTION

Chronic kidney disease (CKD) is associated with an increased risk of cardiovascular disease (CVD) compared with the general population [1,2]. CVD risk is inversely related to glomerular filtration rate (GFR), and begins to rise in early stages of CKD [1–3]. The global burden of CKD-associated CVD is immense, as stage 3 CKD (GFR 30–59 ml/min/1.73m²) is prevalent in approximately 10% of the general population [4]. Furthermore, the mortality rate due to kidney disease-associated CVD eclipses progressive CKD [5], thus highlighting its public health importance. Classic CVD risk factors do not fully explain the increased CVD risk in CKD [5]. Vascular calcification (VC) [6,7] and hyperphosphatemia [8–12] may be important CKD-- specific risk factors for CVD, and human and animal studies suggest that hyperphosphatemia stimulates VC [13–17]. Multiple studies have demonstrated an association between serum phosphorus levels and risk factors for CVD, including vascular stiffness [18–20] and VC [19]. In recognition of the roles that disorders of mineral homeostasis and skeletal remodeling play in kidney disease-associated CVD, a syndrome was named in 2006, the chronic kidney disease – mineral bone disorder (CKD-MBD) [22].

In early CKD, the serum phosphorus concentration remains normal until late stage 3 or stage 4, but the CKD-MBD begins in early CKD [23–25] suggesting that vascular perturbations of CKD occur prior to hyperphosphatemia. Furthermore, studies have demonstrated that the pathogenesis of the CKD-MBD and associated VC includes other factors along with changes in phosphate homeostasis. Thus, early intervention in the CKD-MBD may affect vascular stiffness before structural abnormalities such as vascular intima-media thickness or calcification are altered.

We hypothesized that in normophosphatemic patients with stage 3 CKD, intestinal phosphate binding with lanthanum carbonate (LaCO3) would not produce hypophosphatemia but may improve vascular manifestations of the CKD-MBD such as vascular stiffness, carotid intima-media thickness (cIMT) and VC. Secondarily, we sought to produce preliminary data that can be used for powering cardiovascular outcomes in future clinical trials of phosphate binders.

RESULTS

Baseline Demographics

312 subjects that met inclusion criteria were identified by their primary nephrologist between January 1 and December 31, 2010. 58 (19%) met the study entry criteria (including willingness to participate in the study) and were recruited, 41 were enrolled after a formal review of their medical records to confirm eligibility, and three subjects dropped out prior to the first study visit, leaving 38 subjects for the intention-to-treat analysis (Figure 1). Enrollment of a subject required matching for age, sex, race and diabetes status in a subsequent enrollee producing the large gap between screening and enrollment along with consent. The matching is shown in Table 1, and gender matching was less successful as there were more males enrolled in the placebo group, but the difference was not statistically significant. Other baseline clinical, cardiovascular, and biochemical characteristics were similar between the groups (Tables 1 and 2). Compliance was assessed at each study visit using pill counts, and was defined as the percentage of prescribed doses that were taken by each subject. The overall mean compliance was 84% (range 45–100%); compliance was similar in both groups (LaCO3: 85%, Placebo: 84%).

Figure 1. Enrollment Flowchart and Randomization Strategy.

Table 1.

Baseline Demographic Data

	LaCO3	Placebo
	(n=19)	(n=19)
Demographics
Age (yr)	62 ± 11	61 ± 13
Height (cm)	173 ± 9	174 ± 12
Weight (kg)	97 ± 20	94 ± 23
Body mass index (kg/m2)	33 ± 6	31 ± 5
Gender (M/F)	9/10	14/5
Diabetes Mellitus [n, (%)]	7 (36)	7 (36)
Race
African-American [n, (%)]	4 (21)	5 (26)
Caucasian [n, (%)]	15 (79)	12 (63)
Other [n, (%)]	0 (0)	2 (10)

All values given as mean ± SD or number (percent). There are no significant differences between LaCO3 and Placebo.

Table 2.

Comparison of Changes Phosphorus, Phosphate homeostasis, Biomarkers of CVD, and Biomarkers of CKD-MBD Between Groups.

	LaCO3 (n=19)		Placebo (n=19)
	Baseline	12 months	Baseline	12 months	P
Phosphate homeostasis
Serum phosphorus (mg/dL)	3.5 ± 0.5	3.3 ± 0.5	3.3 ± 0.4	3.2 ± 0.7	0.88
Serum calcium (mg/dL)	9.2 ± 0.5	9.5 ± 0.4	9.3 ± 0.3	9.5 ± 0.4	0.55
Serum creatinine (mg/dL)	1.5 ± 0.3	1.6 ± 0.4	1.8 ± 0.4	1.8 ± 0.5	0.48
24 hour creatinine clearance (ml/min/1.73m2)	47 ± 12	47 ± 16	45 ± 12	46 ± 15	0.97
Intact parathyroid hormone (pg/ml)	76 ± 72	77 ± 53	57 ± 19	65 ± 36	0.48
FGF23 (pg/ml)	69 (24, 117)	55 (33, 367)	55 (25, 201)	55 (33, 299)	0.87
Tubular reabsorption of phosphorus (%)	76 ± 10	74 ± 8	77 ± 11	71 ± 13	0.28
24 hour urine phosphorus (mg)	707 (469, 958)	605 (530, 803)	735 (590, 905)	764 (591, 897)	0.88
Hemodynamics
Systolic BP (mmHg)	130 ± 17	131 ± 14	130 ± 12	128 ± 10	0.65
Diastolic BP (mmHg)	76 (56, 94)	70 (54, 88)	80 (64, 98)	72 (56, 80)	0.30
Cardiovascular Structure/Function
Carotid artery intima-media thickness (mm)	0.8 (0.6, 1.0)	0.9 (0.6, 1.0)	0.8 (0.5, 0.9)	0.8 (0.6, 1.0)	0.98
Left ventricular mass (g/m2.7)	45 ± 13	53 ± 13	50 ± 12	54 ± 13	0.36
Left ventricular ejection fraction (%)	63 ± 11	64 ± 12	62 ± 9	67 ± 6	0.37
Pulse wave velocity (m/s)	10.6 (7.9, 20.0)	10.0 (7.2, 13.1)	10.4 (5.5, 18.4)	10.0 (6.6, 13.3)	0.32
Biomarkers of CKD-MBD
DKK-1 (pg/ml)	1968 ± 834	1985 ± 814	1705 ± 683	1772 ± 655	0.76
Sclerostin (pg/ml)	1156 (710, 2282)	1233 (898, 2361)	1145 (730, 1756)	1320 (573, 2002)	0.40
Bone Mineral Density (SDS Z-score)	2.4 ± 2.1	2.7 ± 2.1	1.3 ± 1.1	1.1 ± 1.1	0.23

All values given as mean ± SD or median (range). Calcium volume scores have been square root transformed. Within each group, there were no significant differences from baseline to the 12 month visit. P-values represent the comparison of the change from baseline to the 12 month visit between LaCO3 and Placebo.

LaCO3 was well tolerated during the study period. The most commonly reported adverse effect was nausea, which occurred in 5 subjects (26%) compared to 2 (11%) in the placebo group. Three subjects with nausea left the study before completing all 5 visits. Headache, constipation, and hyperglycemia each occurred in 1 subject. The subject with hyperglycemia was later diagnosed with type 2 diabetes mellitus. There were no deaths or serious adverse events reported during the study period.

Primary Outcome

Mean serum phosphorus levels were 3.3–3.5 mg/dl at baseline in all subjects. There were no instances of hypophosphatemia over the 12 months, and LaCO3 had no significant effect on the primary outcome measure, change in mean fasting serum phosphorus from baseline to month 12 (Table 2).

Secondary Outcomes

Phosphate Homeostasis and Biomarkers of the Early CKD-MBD

The tubular reabsorption of phosphorus (TRP) was decreased in both groups at baseline (76% and 77% in LaCO3 and placebo, respectively) and was not affected by LaCO3 (Table 2). Baseline median urinary phosphorus excretion was 707 and 735 mg/day in LaCO3 and placebo, respectively (Table 2). Urinary phosphorus excretion decreased to 605 mg/day in LaCO3 and increased to 764 mg/day in placebo by month 12, but these changes were not significant (Table 2, Figure 3A). Baseline mean intact PTH levels were 76 and 57 pg/ml in LaCO3 and placebo, respectively (upper limit of normal in our local assay is 72 pg/ml) (Table 2). PTH levels did not change significantly over the 12 months in either group(Table 2). There were no significant differences in calcium, creatinine, or creatinine clearance between groups at baseline or at month 12 (Table 2).

Figure 3. Urinary phosphate and FGF23 levels in the LaCO3 and placebo groups at baseline and at 12 months.

Boxplot representation of (A), Urinary Phosphate and (B), FGF23 levels at baseline and 12 months. Median values are the horizontal lines within the shaded areas. The shaded areas represent the 75^th (top) to 25^th (bottom) percentiles, and the vertical lines with whiskers represent the full range.

Median FGF23 levels in the cohort at baseline were 58 pg/ml (range 24–201), which were elevated compared to a reference group of 450 patients in the Diabetes Heart Study with normal glomerular filtration rates that were measured simultaneously with our cohort (37 pg/ml (range 0–539)). LaCO3 did not significantly affect FGF23 levels, which were decreased from 69 pg/ml at baseline to 55 pg/ml at month 12 in the LaCO3 group, compared to no change from 55 pg/ml in placebo (Table 2, Figure 3B).

Baseline levels of DKK1 and sclerostin were 2–3-fold higher in both groups than reference values determined simultaneously in the Diabetes Heart Study cohort that had normal kidney function. After 12 months of treatment there were no differences in plasma DKK1 or sclerostin between or within groups (Table 2). Bone mineral density was normal and remained stable or slightly improved in each group after 12 months (Table 2).

Cardiovascular Endpoints

Blood pressure was well controlled in both groups, with 32 of 38 subjects (84%) taking antihypertensive medications or ACEI for proteinuria throughout the study (Table 2). The majority of our cohort demonstrated vascular stiffness at the baseline visit, with PWV greater than the 50^th percentile for age using data from “The Reference Values for Arterial Stiffness” Collaboration (Table 2) [26]. The decrease in PWV from baseline to month 12 in the LaCO3 group [10.6 (7.9–20.0) to 10.0 (7.2–13.1) m/s, Figure 2], was not significant when compared to placebo (Table 2). Baseline cIMT was normal in both groups. With our observed effect on pulse wave velocity, a statistical power of 0.8 and a probability level of 0.05, the sample size per group with a two tailed hypothesis would need to be 45 patients per group.

Figure 2. Effect of LaCO3 on Pulse Wave Velocity.

Boxplot representation of carotid-femoral pulse wave velocity data at baseline and the 12-month visit in each group. The line in the shaded area is the median value, the top of the shaded area is the 75^th percentile, the bottom of the shaded area is the 25^th percentile and the whiskers represent the range of the values.

There was no change in cIMT from baseline to month 12 within LaCO3 or placebo (Table 2). Low grade VC was common in the carotid arteries, coronary arteries, and aorta in both groups at the baseline visit (Table 3) as determined by Agatston score and calcium volume. After 12 months, the progression of the Agatston score or calcium volume was minimal, and there were no differences in VC in the carotid arteries, coronary arteries, or aorta between LaCO3 and placebo (Table 3, supplemental Figure 1). Left ventricular ejection fraction (LVEF) remained stable in both groups. LVM/Ht^2.7 increased within the LaCO3 group after 12 months of treatment, but this trend was not statistically significant (Table 2).

Table 3.

Vascular Calcification Between Groups

	LaCO3 (n=19)		Placebo (n=19)
	Baseline	12 months	Baseline	12 months	P
Agatston Score
Right Carotid Artery	0 (0, 674)	27 (0, 734)	0 (0, 513)	0 (0, 400)	0.48
Left Carotid Artery	23 (0, 795)	111 (0, 902)	16 (0, 282)	21 (0, 397)	0.69
Coronary Arteries	170 (0, 4357)	248 (0, 5835)	246 (0, 2146)	386 (0, 2443)	0.86
Aorta	778 (0, 25269)	243 (0, 35071)	1675 (0, 10384)	2730 (0, 10994)	0.58
Calcium Volume (mm3)
Right Carotid Artery	7 ± 9	7 ± 9	4 ± 6	5 ± 6	0.57
Left Carotid Artery	8 ± 10	9 ± 10	5 ± 5	5 ± 6	0.88
Coronary Arteries	20 ± 21	23 ± 23	17 ± 16	20 ± 17	0.70
Aorta	42 ± 46	46 ± 53	39 ± 32	43 ± 30	0.70

All values given as mean ± SD or median (range). Calcium volume scores have been square root transformed. Within each group, there were no significant differences from baseline to the 12 month visit. P-values represent the comparison of the change from baseline to the 12 month visit between LaCO3 and Placebo.

DISCUSSION

This randomized, placebo-controlled, pilot and feasibility study demonstrates that 12 months of treatment with the phosphate binder LaCO3 had no significant effect on the serum phosphorus level in normophosphatemic patients with CKD stage 3a [27]. These results were compatible with our preclinical studies of phosphate binders in models of normophosphatemic early CKD. LaCO3 had no significant effect on urine phosphorus excretion, TRP or plasma levels of the phosphaturic hormone FGF23. Furthermore, LaCO3 did not alter PWV, cIMT or VC compared to placebo.

Other investigators have demonstrated small or no changes in phosphate homeostasis during intervention with phosphate binders in early CKD [28,29]. Hill et al [29] showed that the phosphate binder calcium carbonate produced little change in phosphate homeostasis in normophosphatemic subjects, in agreement with our results using the binder LaCO3. Our cohort size was twice that of Gonzalez-Parra et al [28], who performed an open label study of LaCO3 in hyperphosphatemic patients. They also reported a lack of change in serum phosphorus but significant reductions in urinary phosphate excretion and increased TRP that were not detected in our cohort. Our study was designed as a pilot and feasibility study, and there was insufficient power to detect changes in phosphate homeostasis during the study. Using a two-tailed alpha of 0.05, a post-hoc analysis revealed that our sample size of 38 subjects was adequately powered to detect a 0.5 mg/dl difference in serum phosphorus levels, a 10 percentage point difference in TRP, and 400 mg difference in 24-hour urine phosphorus between groups with 80% power. Alternatively, passive phosphate absorption may be resistant to binders in the presence of normophosphatemia, and therefore the change in phosphate homeostasis in our study was truly insignificant. Several studies of phosphate binders in early CKD have shown lack of change in FGF23 levels but modest changes in phosphate homeostasis [30–32]. These studies were in cohorts containing hyperphosphatemic subjects with more advanced kidney disease than our cohort, which may account for the lack of change in FGF23 levels or phosphate homeostasis that we observed. In contrast, other reports show reductions in FGF23 with phosphate binders [28,33,34]. Closer inspection of these studies reveals higher baseline levels of FGF23 than in our cohort, shorter periods of treatment and changes in levels similar to those reported here (Table 2). Although it is not routine clinical practice, coupling dietary restriction with phosphate binders has recently been shown to decrease FGF23 levels in CKD [35,36]. Sigrist et al suggest that dietary restriction alone may be adequate for reducing FGF23 levels in early CKD, but they demonstrate additivity with aluminum based phosphate binders [35]. Thus, it is possible that phosphate binders plus dietary phosphorus restriction may be necessary to reduce FGF23 levels in early CKD.

Parathyroid hormone levels were normal at baseline in our cohort as a whole (68 pg/ml), and were unchanged by LaCO3 just as in the LaCO3 arm in Block et al [30]. Thus, the changes in phosphate homeostasis in our study were minimal, but consistent across all parameters used to assess phosphate homeostasis.

Recent data demonstrate that the CKD-MBD begins in early CKD. Elevations of FGF23 in the skeleton and the circulation are present in stage 2–3 CKD [24,37,38], along with reductions in the tubular reabsorption of phosphorus (TRP) [38] and the onset of vascular stiffness [18]. Our results are in agreement with these studies, showing that in normophosphatemic subjects with stage 3 CKD, circulating FGF23, DKK1 and sclerostin were increased and TRP was decreased at baseline. All subjects had elevated PWV at baseline, confirming that vascular stiffness was present.

An important contribution of our study is the assessment of vascular stiffness in each treatment arm using carotid-femoral PWV. After 12 months of therapy, LaCO3 did not significantly improve PWV compared to placebo. Recent clinical and translational studies have highlighted the importance of FGF23 signaling in vascular stiffness and overall CVD risk [39,40]. We therefore speculate that the lack of significant improvement in vascular stiffness in our study may be linked to the lack of significant changes in phosphate homeostasis and FGF23 levels. Future studies that combine phosphate binders with dietary phosphorus restriction may produce the necessary changes in phosphate homeostasis and FGF23 to reduce markers of CVD risk such as PWV. For our observed effects on PWV, a statistical power of 0.8 and a probability level of 0.05, the sample size per group with a two tailed hypothesis would need to be 45 patients per group.

Our cohort demonstrated normal ranges of cIMT but significant VC at baseline with no significant change in either parameter after 12 months of LaCO3. The normal cIMT range in our cohort may reflect the early stage of CKD and could explain the observed lack of improvement with LaCO3. Although VC was detectable in the cohort at baseline, the progression over 12 months was minimal in both LaCO3 and Placebo. Our results do not disagree with recent studies from Block et al [30] that reported increased coronary artery calcification in a composite cohort of subjects with stage 3–4 CKD receiving calcium acetate, LaCO3, or sevelamer carbonate. In their study the progression of VC in the LaCO3 arm also did not differ from placebo. The progression of VC in the combined study population appeared to derive from the subgroup treated with calcium acetate. Despite the initial consensus around phosphorus as a CVD risk factor [8–12,41], recent studies have questioned this concept and therefore the impetus to intervene in phosphate homeostasis early in the course of CKD [30,42].

There are several limitations to this study. The first is that the study was under powered for the cardiovascular outcomes, especially for detection of the modest differences we observed for each outcome between groups. Secondly, the period of observation may have been too short to observe progression in the surrogates of cardiovascular disease selected for study. This will influence design of future studies seeking to intervene in the cardiovascular morbidity associated with CKD.

We conclude that LaCO3 therapy is feasible without hypophosphatemia in stage 3a CKD. In this prospective pilot and feasibility study in early CKD, 12 months of LaCO3 was associated with no significant changes in phosphate homeostasis and no improvement in PWV, a measure of vascular stiffness, cIMT or VC. This study does not lend support for further study of phosphate binders in early stage CKD because of the failure to affect phosphate homeostasis. A practical design that incorporates effective change in phosphate homeostasis is worthy of additional study.

STUDY POPULATION AND METHODS

The study protocol was approved by the Human Research Protection Office at Washington University in St. Louis. Subjects were enrolled after giving informed consent in accordance with guidelines from the Declaration of Helsinki. Subjects were eligible if greater than 18 years of age and had stage 3 CKD (estimated GFR 30–59 ml/min/1.73m²) using the Modification of Diet in Renal Disease study equation [46]. Exclusion criteria included pregnancy, bone disease, myocardial infarction, congestive heart failure, diastolic dysfunction or severe hypertension. Each week members of the research team (M.R., S.C., W.R., or D.W.) identified their clinic patients that satisfied the inclusion criteria and were interested in the study. Those who provided informed consent to our research nurse allowed a review of their medical records to determine final eligibility before enrollment. The cohort size was predetermined by the available funding. As subjects were enrolled, they were stratified for age, gender, race and diabetes status and then randomized into 2 groups, allocated 1:1 to receive either 1000 mg of LaCO3 or a matching placebo with meals three times daily for 12 months. The randomization and double-blind strategy was designed and maintained by our research pharmacist and statistician. Because practitioners often do not prescribe dietary phosphate restriction in normophosphatemic CKD, dietary phosphate intake was not regulated. All participants were encouraged to track phosphate sources in the diet but these data were not collected. The purpose of the study was to determine the cohort sizes needed to adequately power primary and secondary outcomes in definitive prospective trials.

Cardiovascular evaluations

Cardiovascular assessments were performed at baseline and 12 months. Pulse wave velocity (PWV) was determined by use of applanation tonometry of the carotid and femoral arteries (SphygmoCor, AtCor Medical, Australia) as previously described and validated [47–51]. Briefly, subjects were placed in the supine position and after 10 minutes of rest, heart rate and three brachial artery blood pressure measurements were obtained in each arm using a noninvasive manual sphygmomanometer. The high-fidelity transducer was then placed on the subject’s right carotid artery and the recorded pressure waveforms were calibrated using an average of the peripheral brachial artery blood pressure. Then the transducer was applied to the right femoral artery and waveforms were again acquired. PWV was determined as the difference in travel time of the pulse-wave between the heart and each the right carotid and right femoral arteries, divided by the travel distance of the pulse-wave form. The onset of the Q-wave from the surface electrocardiogram (ECG) is used to determine the start of the pulse-wave. Using a tape measure, the distance between the right carotid artery and suprasternal notch was subtracted from the distance between the right femoral artery and suprasternal notch. The applanation tonometry measurements were performed by a research technician who was blinded to clinical data, echocardiographic results and treatment group.

Carotid artery intima-media thickness (cIMT) was measured by a single vascular sonographer from B-mode images of both carotid arteries expressed as the average of the far-- walls of the right and left common carotid arteries; each site represents the average of three separate measurements [52]. The intraclass correlation coefficient for repeated measures of the cIMT is 0.91 and for echocardiographic measurements ranges from 0.85–0.90 at our laboratory.

Two-dimensional and M-mode echocardiograms and carotid artery ultrasound were performed by use of commercially-available ultrasound equipment (Sequoia-C256, Acuson-- Siemens, Mountain View, CA). Two-dimensional directed echocardiographic measurements included the LV ejection fraction (LVEF) calculated using the biplane method of discs (modified Simpson’s method). LV mass was measured by the M-mode-derived cubed method and indexed to height^2.7 (LVM/Ht^2.7) [53]. All measurements were performed in accordance to published guidelines and represent the average of three consecutive cardiac cycles obtained by a single observer blinded to all clinical parameters and treatment group [54].

Multidetector CT (MDCT) Imaging for Assessment of Vascular Calcification

A 64-slice multidetector CT (MDCT) scanner (Somatom Sensation 64, Siemens Medical Systems, Forchheim, Germany) measured calcium scores and volumes by the Agatston method [55]. After initial scout imaging, the scan fields were set for the neck, chest, and abdomen for measurement of arterial calcification in the carotids (from the arch to 1 mm above the carotid bifurcation), coronary arteries, aortic arch and thoracic aorta (to the top endplate of the T12 vertebral body). The scan parameters included 24 × 1.2 mm collimation, 3 mm slice thickness, 0.37 second rotation time, spiral mode, and 120 kVp at 80 mAs with reconstruction at 60% of the R-R interval. Vascular calcium scores and volumes were measured by use of commercially available software (Vitrea; Vital Images, Inc., Minnetonka, Minnesota) as previously described [55]. All images were evaluated by a single radiologist (A.B.) blinded to all clinical characteristics and the treatment group.

Biomarkers of the chronic kidney disease-mineral bone disorder (CKD-MBD)

Blood samples were obtained from each subject at the baseline visit and after 12 months of treatment. There was intersubject variation in the time of day that blood samples were obtained. Aliquots of plasma were generated from each sample and used immediately for biochemical testing or frozen at 80°C. Plasma levels of FGF23, DKK1 and sclerostin were measured in duplicate using commercially available ELISA kits, according to the manufacturer’s instructions (FGF23: Kainos Laboratories, Tokyo, Japan; DKK1: R&D Systems, Minneapolis, MN; sclerostin: Teco Medical Group, Sissach, Switzerland).

For safety monitoring, plasma phosphorus, calcium, creatinine, and intact parathyroid hormone (iPTH) levels were measured using standard laboratory methods. A random urine sample was obtained concomitantly to calculate the tubular reabsorption of phosphorus. A separate 24 hour urine collection was performed at the baseline and 12 month visits to detect a change in creatinine clearance or phosphorus excretion. We used the 24-hour creatinine excretion to determine the adequacy of each urine collection. Bone mineral density was assessed using DXA scan of the lumbar spine at baseline and 12 months.

Outcome Definitions

The primary outcome was the change in serum phosphorus. Secondary outcomes included change in mean carotid-femoral PWV, a surrogate for vascular stiffness, from baseline to month 12 and changes in: serum phosphorus, 24-hour urine phosphorus, tubular reabsorption of phosphorus (TRP), VC, cIMT, left ventricular mass indexed to height^2.7 (LVM/Ht^2.7), left ventricular ejection fraction (LVEF), and plasma FGF23, DKK1 and sclerostin levels. Given the pilot nature to the design, an empiric number of subjects were enrolled based on available funds.

Statistical Analysis

Statistical analysis was performed by a statistician who remained blinded to the identity of the subjects and their corresponding study groups (J. M.). The data were analyzed using the software package SAS 9.1 (Cary, NC). Calcium volume scores were square-root transformed to limit the likelihood of differences due to variability in MDCT scans, as validated and reported previously [56]. Wilcoxon two-sample test and student’s t test were used to test the differences of continuous variables at baseline and the differences from baseline to the 12 month visit between LaCO3 and Placebo. Signed-rank test (non-normally distributed data) and paired t test (normally distributed data) were used to test differences of continuous variables from baseline to the 12 month visit within LaCO3 or Placebo. Normally distributed data are presented as mean ± standard deviation (SD). Non-normally distributed data are presented as median (range). Categorical variables were compared using Chi-square test. All tests were two-tailed; statistical significance was considered at p < 0.05.