Effect of Paricalcitol on Left Ventricular Mass and Function in CKD—The OPERA Trial

Study Enrollment and Study Participants

In total, 229 eligible subjects were screened between May of 2008 and April of 2010 from the Renal Clinic of the Queen Mary Hospital in Hong Kong; 60 subjects were finally enrolled into the study, with 30 subjects randomly assigned to receive paricalcitol treatment and 30 subjects randomly assigned to receive placebo treatment. Figure 1 shows the recruitment process. Table 1 shows the baseline demographics and clinical characteristics of the two groups. Demographics were comparable between groups, except that there was a trend to more diabetes and higher usage of renin-angiotensin system (RAS) blockers in the group randomized to placebo (P=0.10) and a trend to more patients having background coronary artery disease and heart failure and more aspirin users in the group randomized to paricalcitol. However, these differences were not statistically significant. BP was well controlled in both groups. The majority of patients in both groups received medications that block the RAS. Baseline estimated GFR and 24-hour urine protein excretion were both lower in the group randomized to paricalcitol than the group randomized to placebo, but the difference was not statistically significant. Biochemical parameters of CKD–mineral bone disease, including calcium, phosphorus, intact parathyroid hormone (iPTH), and alkaline phosphatase, were well balanced between groups. Median iPTH levels were approximately two times the upper limit of normal at study baseline.

Baseline cardiac magnetic resonance imaging- (MRI-) and echocardiography-derived functional parameters were well balanced between groups (Table 2). LV ejection fraction was well preserved in both groups, but average peak early diastolic mitral annular velocity (E′) was below seven, indicating diastolic dysfunction in both groups.

Primary End Point

The primary end point was change in LV mass indexed by body surface area or height^2.7 after 52 weeks, which did not differ between groups (Table 3). The primary analysis did not change and remained insignificant (P=0.90), even after adjustment was made for RAS blockers use and baseline heart failure.

Secondary End Points

Change in other prespecified cardiac MRI parameters, such as LV volume index, did not differ between the two groups. Changes in LV ejection fraction and other echocardiographic parameters, including ratio of early to late transmitral inflow velocity (E/A), tissue Doppler-derived measure of E′, late diastolic mitral annular velocity (A′), systolic mitral annular velocity (S′), and ratio of E to E′ after 52 weeks, were not significantly different between the two groups.

Figure 2 and Table 4 detail the changes in key laboratory parameters and BP after 52 weeks. Of 60 patients, only 1 patient in the placebo group had study capsule stepped up to 2 mcg daily. Paricalcitol treatment reduced iPTH by a median of −86 pg/ml (interquartile range, −131 to−-43), whereas placebo treatment increased iPTH by a median of +21 pg/ml (interquartile range, −25 to +134; P<0.001). Paricalcitol treatment reduced alkaline phosphatase by a median of −12 U/L (interquartile range, −21 to −1) in contrast to placebo treatment, which increased alkaline phosphatase by a median of +2 U/L (interquartile range, −6 to +10; P=0.001). Overall, 19 patients (63.3%) treated with paricalcitol versus 1 patient (3.3%) treated with placebo had at least two subsequent iPTH levels reduced ≥50% to baseline levels during study period (P<0.001).²¹ Paricalcitol treatment resulted in a greater increase in serum calcium after 52 weeks than placebo (+0.08 [+0.02 to +0.16] versus +0.01 (−0.06 to +0.05) mmol/L; P=0.03). Change in 24-hour urine protein after 52 weeks did not differ significantly between paricalcitol and placebo groups.

Using mixed effects repeated measures modeling, similar results were observed, with significant overall between-group differences for changes in iPTH (P<0.001) and alkaline phosphatase (P<0.001) from baseline to 52 weeks. iPTH and alkaline phosphatase reduced in the paricalcitol group but increased in the placebo group from baseline to 52 weeks. The paricalcitol group showed greater increase in serum calcium from baseline to 52 weeks than the placebo group (P<0.001). The paricalcitol group showed significantly more decline in estimated GFR (P=0.002) but not serum urea (P=0.80) from baseline to 52 weeks compared with the placebo group. The paricalcitol group showed significantly greater reduction in systolic BP (P=0.03) compared with the placebo group from baseline to 52 weeks. By 52 weeks, systolic BP reduced by −4.49 mmHg (−6.51 to −2.48) in the paricalcitol group versus −3.03 mmHg (−5.04 to −1.01) in the placebo group. Reduction in diastolic BP from baseline to 52 weeks did not differ between paricalcitol and placebo groups (P=0.90) (Supplemental Table 1).

Adverse Events

All patients completed the final visit at 52 weeks without any dropouts or deaths. One patient was not analyzed in the primary end point because of loss of cardiac MR images. Adverse events and hospitalizations of the two groups are summarized in Table 5. One patient in the paricalcitol group progressed to ESRD and required dialysis treatment at 48 weeks. Two patients in the paricalcitol group had hospitalizations versus 10 patients in the placebo group (P=0.02). None of the patients in the paricalcitol group had cardiovascular-related hospitalizations, whereas five patients in the placebo group had six cardiovascular-related hospitalizations. A significant difference in total hospitalization days within the study period was observed between paricalcitol and placebo groups (P=0.02). Using Cox regression analysis, the presence of baseline cardiovascular disease and heart failure showed no significant relationship with subsequent hospitalizations (hazard ratio, 0.88; 95% confidence interval [95% CI], 0.47 to 1.63; P=0.70). Except hypercalcemia, all other adverse events were judged to be unrelated to the study drug. Hypercalcemia (defined as serum calcium>2.55 mmol/L), which was judged to be possibly drug-related, occurred in 13 patients (43.3%) in the paricalcitol group and 1 patient (3.3%) in the placebo group (P<0.001); 9 of 13 patients (69.2%) in the paricalcitol group versus 1 patient in the placebo group who developed hypercalcemia received concomitant calcium-based phosphate binder.

Study Protocol

This prospective, double blind, randomized, placebo-controlled trial was conducted in a university teaching hospital and a major regional tertiary care hospital in Hong Kong. The study protocol was approved by the Institutional Review Board and Ethics Committee. All patients provided written informed consent before study entry.

Study Population

In total, 60 subjects with stages 3–5 nondialysis CKD were recruited. Eligibility criteria included subjects with age between 18 and 75 years, GFR estimated using the four-variable Modification of Diet in Renal Disease⁴¹ equation<60 ml/min per 1.73 m² diagnosed for more than 3 months who are not expected to start dialysis within the next 12 months, screening iPTH≥55 pg/ml, serum calcium<10.2 mg/dl (2.55 mmol/L), serum phosphorus≤5.2 mg/dl (1.68 mmol/L), calcium×phosphorus product<54 mg²/dl² (4.36 mmol²/L²), and no form of vitamin D therapy in the previous 4 weeks. For women, the subject was either not of childbearing potential, which was defined as postmenopausal for at least 1 year or surgically sterile (bilateral tubal ligation, bilateral oophorectomy, or hysterectomy), or if of childbearing potential, had practiced birth control measures. Subjects taking medications that block the RAS must have drug dosage unaltered for at least 3 months before study entry and throughout the study period.

Exclusion criteria include history of an allergic reaction to vitamin D or related compounds, history of renal stones, current malignancy, clinically significant gastrointestinal disease or liver disease, acute renal failure in the recent 3 months, a history of drug or alcohol abuse within 6 months before the screening phase, known human immunodeficiency virus-positive status, evidence of poor compliance with diet and medication, active granulomatous disease, currently receiving medications that may affect calcium or phosphorus metabolism (such as calcitonin, cinacalcet, bisphophonates, or vitamin D compounds), currently receiving or had received glucocorticoid or other immunosuppressive treatment for more than 14 days within the recent 6 months before screening, use of stable estrogen and/or progestogen therapy, pregnancy, and contraindications for MRI examination.

Study Design

Patients who fulfilled all of the above eligibility criteria provided informed consent and underwent a screening phase, during which time screening echocardiography was performed to estimate LV mass index. Estimated GFR, serum calcium, and phosphorus were repeated to confirm final eligibility for study inclusion. In essence, screening echocardiography showing evidence of LV hypertrophy as defined by an LV mass index>115 g/m² in men and >95 g/m² in women in accordance with the updated guideline from the American Society of Echocardiography⁴² would be considered eligible for study inclusion.

Eligible subjects were randomized in a double blind fashion in a 1:1 ratio to receive either oral paricalcitol capsules (active treatment group) or placebo (control group). The randomization schedule was computer-generated by the sponsor using the WebRando System (Abbvie), and study capsules were supplied by the sponsor. The stratification factor was iPTH level≥500 or <500 pg/ml. The block sizes were either two or four.

Subjects assigned to the active arm received a 1-µg oral paricalcitol capsule one time daily if iPTH was <500 pg/ml or a 2-µg oral paricalcitol capsule one time daily if iPTH was ≥500 pg/ml. Thereafter, dose titration was done based on safety reasons for high calcium level. The treatment duration was 52 weeks. Placebo and active treatments were identical in appearance. Study investigators and study subjects were masked to treatment for the duration of the trial. Study subjects returned for follow-up at baseline and weeks 6, 12, 24, 36, 48, and 52, during which time vital signs, including BP and body weight, adverse events, concomitant medications, and compliance to study capsules were recorded. Systolic and diastolic BP was measured two times at least 3 minutes apart on either arm on every follow-up visit after the study subject had rested for at least 15 minutes, and the results were then averaged to give the final systolic and diastolic BP.

Study End Points

The primary end point was change in LV mass index by plain cardiac MRI over 52 weeks. Prespecified secondary end points included changes in LV end systolic and end diastolic volume index and ejection fraction, E/A ratio and deceleration time by flow Doppler, and S′, E′, A′, and E/E′ by tissue Doppler imaging over 52 weeks. Other prespecified secondary biochemical end points included change in iPTH, serum calcium, phosphorus, alkaline phosphatase, estimated GFR, and 24-hour urine protein over 52 weeks.

Plain Cardiac MRI

Consented subjects underwent plain cardiac MRI at baseline and 52 weeks after treatment to estimate LV mass index, LV volumes, and ejection fraction. Subjects were examined in the supine position with a 1.5 T system (CVi; GE Medical Systems), and an eight-element phased-array radiofrequency coil was used for signal reception. All images were obtained during repeated breath holds and gated to the electrocardiogram. Double oblique long-axis scouts were taken to obtain true short- and long-axis references with two-dimensional fast imaging employing steady state acquisition (2D-FIESTA) sequence. Cine images were acquired in short- and long-axis views. Short-axis views (8-mm thickness and 0-mm gap) were obtained throughout the whole heart from left atrium to LV. MRI analysis was performed in a semiautomated fashion with a commercially independent workstation (Advantage Windows; GE Medical Systems). LV volumes, mass, and ejection fraction were analyzed from MRI cine short-axis views using commercially available software (MASS program; Medis) by a single experienced radiologist. The intraobserver tests performed based on repeat analysis 1 month later of the same MR images of 10 study subjects by the same radiologist showed excellent reproducibility, with intraclass correlation coefficients being 0.95 (95% CI, 0.81 to 0.99), 1.00 (95% CI, 0.99 to 1.00), 0.99 (95% CI, 0.96 to 1.00), and 0.95 (95% CI, 0.82 to 0.99) for LV mass, LV end diastolic volume, LV end systolic volume, and ejection fraction, respectively.

Echocardiography

Echocardiography was performed at baseline and 52 weeks after treatment using a Vivid-7 Ultrasonographic System (GE Healthcare) with a multifrequency transducer by a single experienced cardiologist blinded to all clinical details of patients. Data were analyzed offline to assess systolic and diastolic function of the heart. All reported echocardiographic measurements were averaged from three consecutive cycles and analyzed by a single experienced cardiologist. Mitral inflow velocities were recorded using pulsed-wave Doppler with the sample volume placed at the tip of the mitral valve tips from the apical four-chamber view as previously described for peak E-wave velocity, peak A-wave velocity, and deceleration time.⁴³ Myocardial velocities were recorded using tissue Doppler technique as described previously.^44,45 In brief, pulsed-wave tissue Doppler images were acquired over a predetermined two consecutive cardiac cycles for each of the four mitral segments and transferred to a workstation composed of a personal computer with a software package that provides customized image visualization, processing, and analysis (EchoPac; GE-VingMed). The sample volume was placed in the mitral annulus of septal and lateral myocardial segments from the four-chamber view. Mean velocities during S′, E′, and A′ were measured. The final value represented the average of four sites. E/E′ ratio was used as a noninvasive marker of LV filling pressure and diastolic function.⁴⁶

Biochemical Parameters and Analysis

Fasting venous blood samples were collected at baseline and weeks 12, 24, 36, and 52 for assessment of complete blood picture, renal function, liver function, serum calcium and phosphorus, alkaline phosphatase, glucose, iPTH, and lipid profile. Additionally, blood samples were collected at weeks 6, 18, 30, 42, and 48 for renal function test and serum calcium and phosphorus; 24-hour urine was collected at baseline and weeks 24 and 52 for measurement of protein excretion. Plasma iPTH was analyzed by chemiluminescence immunoassay using the IMMULITE 1000 Analyzer (Siemens Healthcare Diagnostics, Deerfield, IL). The other laboratory tests were performed on the Beckman Coulter GEN-S Blood Cell Counter (Beckman Coulter Inc., Miami, FL) or the Roche DP Modular Analyzer (Roche Diagnostics Corp., Indianapolis, IN).

Sample Size Calculation

At the time of study planning, there was only scarce data on changes of LV mass in CKD. Based on previous studies,^31,47 the common SD of LV mass was estimated to be ∼10 g. To give the study a 90% power to detect a significant 10-g difference (α=0.05, two-tailed test) in absolute LV mass or a 2.7 g/m^2.7 difference in LV mass indexed by height^2.7 between the two groups, a minimum of 21 patients was required in each arm. A 10-g difference in absolute LV mass is a clinically important and statistically detectable difference given the use of cardiac MRI.³¹ Taking into account a 20% dropout rate, we recruited a total sample size of 60 subjects, with 30 subjects in each treatment arm.

Safety Analysis and Adverse Events

Safety was assessed through adverse events monitoring, changes from baseline laboratory parameters, especially serum calcium and phosphorus and iPTH, and changes from baseline in vital signs and physical examinations. If albumin-adjusted serum calcium increased above the upper limit of laboratory reference range=10.2 mg/dl (2.55 mmol/L) any time during study period, calcium-based binder, if taking any, was first stopped. Study drug was discontinued temporarily if albumin-adjusted serum calcium increased ≥11 mg/dl (2.74 mmol/L), and it was resumed at a lower dose when serum calcium normalized within the laboratory reference range. Subjects with onset of hypercalcemia, defined as albumin-adjusted serum calcium>10.2 mg/dl (2.55 mmol/L) any time during the study period, were recorded. All hospitalizations and days hospitalized during the study period were captured from the Hong Kong Hospital Authority Centralized Medical Record System. The nature of hospitalizations was reviewed and adjudicated by an independent committee blinded to the treatment arm allocation. Cardiovascular-related hospitalizations included hospitalizations because of acute coronary syndrome, acute myocardial infarction, or unstable angina with electrocardiographically documented changes of myocardial ischemia, electrocardiographically documented bradyarrhythmia or tachyarrhythmia, transient ischemic attack, thromboembolic or hemorrhagic stroke, heart failure/fluid overload, peripheral vascular disease, or sudden cardiac death, all of which were defined previously.⁴⁸ There was 100% agreement in adjudicating the hospitalizations to be of cardiovascular nature by an independent committee.

Statistical Analyses

All statistical analyses were performed using SAS software, version 9.2 (SAS Institute Inc.). Mean (SD) was used to summarize data distributions. Within-group differences were summarized using median (interquartile range). For the primary end point of changes in LV mass index and other secondary end points, the intention to treat approach was used. The analysis of the primary end point was based on the difference between the two groups in the change of LV mass index during the study period. This analysis was performed by the two-sample Wilcoxon rank sum test for data not normally distributed. A similar approach was used for all secondary end points. For those secondary end points that were measured more than two times, including serum calcium and phosphorus, alkaline phosphatase, iPTH, estimated GFR, and BP, we also used the maximum likelihood, mixed effects, repeated measures model and included all longitudinal observations in the intention-to-treat population. The mixed effects model included terms of treatment, visit, and treatment × visit interaction with baseline values considered as covariates. The analyses for the mixed effects repeated measures model were performed using PROC MIXED, with denominator degrees of freedom estimated by the Satterthwaite approximation. The final statistical significance levels for the outcomes were not adjusted for multiple comparisons. Within-participant errors were estimated using exchangeable covariance unless otherwise specified. Overall P values represent the significance level for the overall treatment group effect with all the follow-up times combined.

Safety analysis was conducted by comparing the incidence of adverse events, including hypercalcemia, as well as hospitalizations between the two groups using the Fisher exact test. A P value<0.05 was considered statistically significant.