Meta-Analysis: The Efficacy and Safety of Paricalcitol for the Treatment of Secondary Hyperparathyroidism and Proteinuria in Chronic Kidney Disease

Introduction. Previous studies have demonstrated the safety and efficacy of using Paricalcitol for the treatment of secondary hyperparathyroidism (SHPT) in patients on dialysis. The aim of the current meta-analysis was to assess the safety and efficacy of Paricalcitol for the management of SHPT in patients with chronic kidney disease (CKD) not yet on dialysis. A secondary aim was to determine if sufficient data was available to assess the effect of Paricalcitol for the management of proteinuria. Methods. A meta-analysis was conducted using the Cochrane Collaboration's RevMan 4.2 software. Results. Paricalcitol is effective in lowering PTH in patients with CKD not yet on dialysis and is also effective in lowering proteinuria in diabetic CKD patients. However, we uncovered a safety signal identifying an elevated calcium phosphate product and a trend towards the development of hypercalcemia. A phosphate elevation was not demonstrated because the target used in the clinical studies was a P > 5.5 mg/dl, a value appropriate for dialysis patients and not CKD patients. Conclusion. Although Paricalcitol is effective in lowering PTH, we advise caution in the use of any active Vitamin D analogues in patients with CKD because of the potential risk of exacerbating vascular calcification.


Introduction
One of the greatest therapeutic challenges in the chronic kidney disease population is the management of bone and mineral metabolic parameters in order to preserve bone integrity, minimize cardiovascular calci�cation, and manage serum levels of parathyroid hormone (PTH), calcium and phosphorus. e cornerstone of this condition is characterized by the CKD-MBD (chronic kidney disease-mineral bone disorder) syndrome where there is secondary hyperparathyroidism (SHPT), manifested by parathyroid hyperplasia and upregulated synthesis and secretion of PTH [1,2]. In addition, there is an elevation of the serum phosphate, a reduction in serum calcium, and an absolute reduction of active vitamin D (calcitriol) levels caused by a reduction of the synthetic 1a-hydroxylase encoded by the CYP27B1 gene and an increase in the catabolic 24a-hydroxylase encoded by the CYP 24 gene. Both of these enzymatic changes are characteristically present in CKD and are very likely mediated by the high levels of FGF23 also characteristically present in CKD [3]. e other features of this syndrome include renal osteodystrophy where an abnormality of bone anabolism causes high bone turnover disease, fractures, vascular calci-�cation, and cardiovascular complications. Slowing the rate of progression towards end-stage renal disease is one of the key goals of medical intervention in this patient group.
Secondary hyperparathyroidism (SHPT) is a common and early complication of CKD. Targeting SHPT in patients with CKD and end stage renal disease on dialysis with active vitamin D analogues such as Paricalcitol has been the subject of multiple research studies in patients. Numerous studies of mixed quality, targeting various surrogate outcome measures have been published clearly demonstrating the biologic importance of the therapy [4][5][6][7]10]. Active vitamin D analogues including Paricalcitol have shown demonstrably favorable effects on SHPT [4,[6][7][8][9][10][11] and proteinuria [4][5][6][12][13][14]. In these studies, no clinically important or statistically evident change in eGFR has been shown [4-6, 8, 9, 12, 14]. e active vitamin D analogues have also shown clear evidence for decreases in cardiovascular events [15], and improved survival in hemodialysis patients. [16,17]. e majority of the published research data, however, has been obtained in patients on dialysis [7,10,11,[15][16][17]. e role of treatment with Paricalcitol in CKD targeting early SHPT is less clear [4-6, 8, 9]. A previously published meta-analysis summarized the efficacy of Paricalcitol therapy for chronic kidney disease combining the data for patients receiving and not yet on dialysis and concluded that Paricalcitol suppresses iPTH and lowers proteinuria in patients with stages 2-5 CKD without an increased risk of adverse events [14]. At least one study has presented evidence for the use of vitamin D analogues in the prevention of vascular calci�cation [18].
One of the many biologic actions of active vitamin D is to cause an increase in the amount of intestinal calcium and phosphorus absorption, resulting in hypercalcemia and hyperphosphatemia. At higher than physiologic dosages, active vitamin D may actually increase bone resorption. Paricalcitol, a synthetic vitamin D analogue engineered to effectively suppress secretion of PTH with fewer hypercalcemic and hyperphosphatemic side effects, has been shown to effectively reduce PTH and also reduce proteinuria in recent studies in patients with CKD [4-6, 12, 14, 16, 17]. e goal of the present meta-analysis was to evaluate the efficacy and safety of treatment with Paricalcitol in the management of SHPT, proteinuria, and preservation of renal function in patients with CKD. In particular, we wanted to evaluate whether there is sufficient published data to recommend treatment with Paricalcitol to patients with CKD and SHPT not yet on dialysis.

Exclusion Criteria.
Patients were excluded if they failed to meet the inclusion criteria or if they failed to complete the study protocol. We also excluded animal studies.

Efficacy Indices
Change in iPTH. De�ned as achieving a greater than or equal to 30% decrease in iPTH from baseline for two consecutive measures. Proteinuria. De�ned as a statistically signi�cant decrease in urinary protein-creatinine ratio or urinary albumincreatinine ratio. ΔeGFR. e mean change in eGFR from baseline to �nal visit. Hypercalcemia. De�ned as two consecutive calcium measurements of greater than 2.62 mmol/L or 10.5 mg/dL. Hyperphosphatemia. De�ned as two consecutive phosphorus measurements of greater than 5.5 mg/dL. Elevation in Calcium × Phosphorus Product. De�ned as two consecutive calcium × phosphate product values of greater than 55 mg 2 /dL 2 .

Data Extraction and Quality
Assessment. ree independent authors extracted relevant data from eligible studies. Discrepancies were resolved by discussion and by referencing the original report. Two independent authors assessed each trial using the Jadad rating scale [19] and referred to the Cochrane Reviewers' Handbook 4.2.6 about the quality of randomized controlled trials (randomization, blinding, withdrawal and loss, allocation concealment, and intentional Analysis-A: adequate, B: unclear, C: inadequate, and D: not used) [20].

Data Synthesis and Analysis.
A meta-analysis was conducted using the Cochrane Collaboration's RevMan 4.2 soware. A test of heterogeneity was assessed by the chisquare test ( value and 2 ), which describes the percentage of variability in the effect and estimates the contribution of heterogeneity rather than by chance [21,22]. We summarized treatment effects as relative risks (RRs) for categorical variables and weighted mean differences for continuous variables, with 95% CIs. If no heterogeneity existed among studies ( and 2 > ), the �xed effect model was used. An 2 = , indicated that the variation was caused by sampling error; A 2 < 2 , indicated a slight degree of heterogeneity; A 2 > 0.25-<0.5, indicated a moderate degree of heterogeneity; A 2 > , indicated a high degree of heterogeneity [23]. If 2 > 0.5 or 0.05, the heterogeneity among these studies was considered statistically signi�cant and a descriptive analysis was employed.

Results
A total of 25 articles were retrieved in the initial search. We found 1 ongoing study (VITAL study) [23] and fortunately obtained results about the published VITAL study from the corresponding author [12]. Examination of the abstracts and full texts allowed us to exclude non case-control studies or studies where the participants did not have CKD, leaving 9 articles that form the basis of this meta-analysis.
3.1. Trial Characteristics. e 9 studies included a total of 1113 participants; 20 participants did not complete the protocol and are excluded leaving 1093 participants included in this meta-analysis. 58.2% had diabetic kidney disease, 20.6% had nondiabetic kidney disease, and the remainders were not characterized. e characteristics of the nine studies and the efficacy parameters are summarized in Tables 1(a) and 1(b), respectively.

Trial Quality.
We assessed the quality of included studies using the Jadad rating scale [19] and referred to the Cochrane Reviewer's Handbook 4.2.6 for guidelines used to rate the quality of randomized controlled trials [20] (Table 2). e main factors in�uencing quality were allocation concealment, intentional analysis, withdrawal, and dropout. e primary reasons described for premature withdrawal of these patients were kidney transplantation, increase in iPTH levels, unblinding, and failure to complete a scheduled protocol visit. Each study received a grade of A or B. e Jadad rating score was assigned from 2 to 5 points.

Two Consecutive Decreases of Greater an or Equal
to 30% in iPTH. e six studies that compared this efficacy index included a total of 720 participants; 369 and 351 treated with Paricalcitol and placebo, respectively ( Figure 1). All six studies had homogeneity (heterozygosity test, 2 = 3.28, = 0. , 2 = 0%). When the �xed-effect model was used to merge RR values, the pooled RR was 6.97 (95% CI 5.27-9.23, = 3.5 , 0.0000 ; Table 3). is indicated that the Paricalcitol treated patients had a statistically signi�cant sustained reduction in serum iPTH levels during the observation period.

ΔeGFR.
Among the three studies that reported this efficacy index, a total of 468 patients were included; 221 and 247 in the Paricalcitol and placebo groups, respectively (Figure 2). All three studies had heterogeneity (heterozygosity test, 2 = 240.0 , 0.0000 , 2 = 98.8%; Table 3). Because of this high heterogeneity, the effect of Paricalcitol on the ΔeGFR using formal meta-analysis technique is not certain. Using a more conventional descriptive analysis, the data from each study (Table 4) show no statistically signi�cant difference between the Paricalcitol-treated and placebo groups implying that Paricalcitol had no negative impact on renal function.

Hypercalcemia.
Among the six studies where this efficacy parameter is reported, 875 participants were evaluated for the incidence of hypercalcemia; 495 and 380 in Paricalcitol and placebo groups, respectively ( Figure 5). All six studies had homogeneity (heterozygosity test, 2 = 0. 4, = 0.9 , 2 = 0%). When the �xed-effect model was used to merge RR values, the pooled RR was 2.91 (95% CI 0.86-9.90, = . , = 0.09; Table 3). is indicated that there was no statistically signi�cant difference in the incidence of hypercalcemia between the Paricalcitol and placebo groups though a trend towards hypercalcemia was evident in the Paricalcitol-treated groups, where 10 of 495 in the Paricalcitol group and 1 of 380 in the placebo group developed hypercalcemia. ere was insufficient data to determine a dose-response effect comparing 1 ug versus 2 ug dosing.

Hyperphosphatemia.
Among the three studies reporting this efficacy parameter, 478 participants were evaluated for the incidence of hyperphosphatemia; 233 and 245 in the Paricalcitol and placebo groups, respectively ( Figure 6). All three studies had homogeneity (heterozygosity test, 2 = 0. 0, = 0.90, 2 = 0%). When the �xed-effect model was used to merge RR values, the pooled RR was 0.94 (95% CI 0.56-1.58, = 0.22, = 0.82; Table 3). is indicated that there was no statistically signi�cant difference in the incidence of hyperphosphatemia between the Paricalcitol and placebo groups. e studies included, used 5.5 mg/dL, as the upper limit of the acceptable range for the serum phosphate. e current KDIGO guidelines recommend targeting the serum phosphate to the normal range, in this case a level 4.0 mg/dL [24].     Coyne et al. [9] Greenbaum et al. [10] Ross et al. [11] Total (95% CI) Total events: 322 (paricalcitol), 44  Coyne et al. [9] Total (95% CI)

Elevation in Calcium × Phosphorus (Ca × P) Product.
In the two studies reporting this efficacy parameter, 423 participants were evaluated for a change in calcium × phosphorus product levels, 205 and 218 in the Paricalcitol and placebo groups, respectively (Figure 7). Both studies had homogeneity (heterozygosity test, 2 = 0.30, = 0. , 2 = 0%). When the �xed-e�ect model was used to merge RR values, the pooled RR was 1.97 (95% CI 1.06-3.67, = 2. , = 0.03; Table 3). While, the data shows that there was no statistically signi�cant di�erence in the incidence of an elevation in Ca × P product between the Paricalcitol and placebo groups for each individual study (Table 5), the pooled data in the meta-analysis do show a statistically signi�cant increase in the incidence of an elevated Ca × P product between the Paricalcitol-and placebo-treated groups ( = 0.03).

Discussion
e CKD-MBD syndrome characteristic of chronic kidney disease (CKD) of virtually any etiology imposes the burden of excess mineral retention enhancing cardiovascular risk by promoting the development of vascular calci�cation [13,[25][26][27]. e hallmark biochemical abnormalities identi�ed in CKD are a reduced level of active vitamin D which results in an elevated blood level of PTH by upregulation of the synthesis and secretion of parathyroid hormone [27]. Targeting PTH synthesis by treating the active vitamin D insufficiency is the generally accepted standard of care [28,29]. However, the treatment with active vitamin D analogues may promote further retention of calcium and phosphate and potentially worsen the cardiovascular risk pro�le of the patients with CKD being treated. Paricalcitol is a synthetic active vitamin D analog chemically designed to limit the absorption of calcium and phosphate by the intestine [30]. In low doses, Paricalcitol results in a 10-fold reduction of calcium absorption compared to calcitriol [31]. Paricalcitol acts as an active agonist for the vitamin D receptor and in the parathyroid gland negatively regulates the gene transcription for PTH thus lowering the blood parathyroid hormone level [32][33][34][35]. However, Paricalcitol like all currently available active vitamin D analogues directly binds to the VDR in  [9] 13/101 7/108 0.161 many tissues [36]. Intestinal activation of the VDR can cause hypercalcemia and hyperphosphatemia by enhanced intestinal absorption. Hyperphosphatemia by itself has been associated with increased mortality in patients on dialysis as well as those with CKD not yet on dialysis [36,37]. Paricalcitol has more selective activation of VDR in the parathyroid gland over that in the intestine and bone and offers the possibility of minimizing the risk of hypercalcemia and hyperphosphatemia, while still signi�cantly reducing PTH [30,38,39]. e goal of the current meta-analysis was to assess whether there is sufficient evidence-based data to recommend Paricalcitol for the management of SHPT in chronic kidney disease patients not yet on dialysis. A second goal of this meta-analysis was to assess whether there is sufficient evidence-based data to recommend Paricalcitol in the management of proteinuric renal disease.
e results of the current meta-analysis indicate that Paricalcitol can decrease iPTH levels signi�cantly with a relatively low incidence of hypercalcemia and hyperphosphatemia. However, we noted that there was a statistically signi�cant difference in elevated Ca × P product levels, which was not found in previous studies [8,9,14]. A recently published meta-analysis including patients with CKD and patients with ESRD on dialysis did not identify any safety issues [14]. In contrast, our study looked at patients with CKD only and this safety issue was readily identi�ed. Although there is no statistically signi�cant difference in the incidences of hypercalcemia and hyperphosphatemia between Paricalcitol compared to placebo, an element of caution is needed in interpreting this data. Active vitamin D analogues, including Paricalcitol, are very likely to have a dose-dependent effect, with higher doses possibly resulting in both hypercalcemia and hyperphosphatemia potentially contributing to an increase in cardiovascular complications. In the studies included in this review, there was insufficient power to identify a dose-response effect. Furthermore, the studies included used 5.5 mg/dL as the upper limit of the acceptable range for the serum phosphate. is level is recommended in the KDOQI guidelines for patients on dialysis [40]. However, patients with chronic kidney disease usually do not reach such high levels and the current KDIGO guidelines recommend targeting the serum phosphate to the normal range, in this case a level (<4.0 mg/dL) [24]. us, the currently available data may grossly underestimate the effect of Paricalcitol on the serum phosphate in this patient group with CKD not on dialysis.
Paricalcitol has recently been shown to reduce proteinuria in patients with diabetic kidney disease [4-6, 12, 14]. De Zeeuw et al. [12] reported changes in urinary albuminto-creatinine ratio (UACR) and did not note a dose-response relation with Paricalcitol. Due to the limited number of studies included in this meta-analysis about this component, we also did not observe a signi�cant difference between the 1 microgram Paricalcitol treated groups compared with the 2 microgram Paricalcitol treated groups in reducing proteinuria. e current meta-analysis included three studies [4,5,12] that found a signi�cant decrease in proteinuria with Paricalcitol therapy compared with placebo. Agarwal et al. [6] analyzed data from a randomized controlled trial comparing Paricalcitol with placebo for the treatment of secondary hyperparathyroidism in chronic kidney disease. Urinalysis dipstick proteinuria (qualitative) was assessed, and a decrease in proteinuria occurred in 51% of Paricalcitol-treated patients compared with 25% of controls. Szeto et al. [13] studied 10 patients with immunoglobulin. A nephropathy-treated with calcitriol, 0.5 ug, twice-week for 12 weeks and found a signi�cant decrease in proteinuria. Moreover, in studies [4-6, 12, 13], ACE inhibitors and/or ARBs were used in the majority of patients, and the interaction with a decrease in proteinuria was not signi�cant in the patients receiving ACE inhibitors or ARBs. ere may be several possible mechanisms of action for this effect on reducing proteinuria, though none are de�nitive. Results of e�perimental studies suggest that the reduction in proteinuria induced by Paricalcitol is caused by inhibition of T-cell proliferation and activation [24,41], reduced cytokine and transforming growth factor production, [42] protection of podocytes [43], and suppression of the renin-angiotensin system [44,45].
e result of changes in eGFR in the current metaanalysis should also be interpreted with caution, because relevant data for this meta-analysis were highly heterogeneous. e studies selected contained descriptive analyses that supported the hypothesis that there was no clinically or statistically signi�cant difference in eGFR between Paricalcitol and placebo groups. is same observation in reference to eGFR was noted in other studies as well [12][13][14]46].
One of the major limitations of this meta-analysis is the inclusion of only a limited number of studies that met the predetermined set of entry criteria. To minimize bias, we thoroughly carried out searches across different databases using explicit criteria for study selection, data analysis, and data abstraction. Not all of the studies used intention to treat analysis, and allocation concealment was adequate in only �ve studies. e absence of both of these components could potentially lead to bias. In addition, we could not assess a funnel plot to reveal possible publication bias. Furthermore, the current KDIGO guideline for the target serum phosphorus is a normal level usually de�ned as <4.0 mg/dL [40]. However, each of the studies included in this metaanalysis used a de�nition of an elevated serum phosphorus as >5.5 mg/dL. e data to recalculate the effect of Paricalcitol using the currently accepted lower serum phosphorus target was not available to us. Lastly, the small number of studies that addressed the use of Paricalcitol for the treatment of proteinuria in patients without diabetic kidney disease contain insufficient data to conclude that Paricalcitol can decrease proteinuria in these conditions.
In conclusion, there is sufficient evidence based data to conclude that Paricalcitol can effectively decrease iPTH levels in patients with CKD not yet on dialysis. We feel that the data evaluating the effects on serum calcium and phosphate are troublesome and our inability to demonstrate adverse events is limited by insufficient power in the analysis. Moreover, the de�nition of hyperphosphatemia masked the effect on the serum phosphorus rendering this data largely uninterpretable. e data evaluating the effect on the serum calcium although not statistically signi�cant with a < , de�nitely showed a trend towards hypercalcemia. is trend was also identi�ed in a previously published meta-analysis which included both patients with CKD and ESRD on dialysis [14]. e statistically signi�cant abnormality highlighting the elevation in Ca × P product supports the notion that clinically signi�cant abnormalities in calcium and phosphate levels may be present with active vitamin D analogue treatment including Paricalcitol. We recommend caution in the use of vitamin D analogues including Paricalcitol in the management of SHPT in CKD patients not on dialysis and advise using the lowest effective dose with careful monitoring for the development of hypercalcemia and hyperphosphatemia.
ere is also evidence-based data to conclude that Paricalcitol can lead to a signi�cant reduction in proteinuria in patients with diabetic CKD with no apparent measurable impact on kidney function, but not in the case of patients with nondiabetic kidney disease where there is insufficient data. Due to inherent limitations of meta-analysis, larger association studies or multicentric case-control studies are needed to con�rm these �ndings, especially the effect on the serum phosphate using the currently accepted lower target blood level and the effect on the serum calcium using a larger database with sufficient power analysis.
�on��ct of �nterests e authors declare that they have no con�ict of interests.