Introduction
Dietary phosphorous overload and positive calcium balance depending on calcium-containing phosphate binder are regarded as contributory factors of vascular calcification and increased mortality in dialysis patients (1–5). Coronary artery calcification (CAC) has been found in patients in early stages of CKD, termed nondialysis-dependent CKD patients (NDD-CKD), and portends a poor prognosis (6–10). In this population, however, no study has evaluated the potential role of phosphorus binders on mortality.
In this pilot study, we aimed to test the potential difference in all-cause mortality rate as the primary end point, as well as on dialysis inception and the composite end point (mortality and dialysis inception) as secondary end points, in NDD-CKD patients treated with different phosphate binders.
Materials and Methods
This is a multicenter, prospective, randomized, nonblinded study. The primary end point was all-cause mortality; secondary end points were dialysis inception and the composite end point (all-cause mortality and dialysis inception).
This study should be regarded as a pilot study because no data were available from the medical literature on the potential effect of phosphorus binders on mortality in nondialysis-dependent CKD patients. On the basis of our historical data, we decided to randomize 240 patients to have 80% power to detect a 60% reduction in all-cause mortality at the 5% significance level considering a 7% of all-cause mortality rate in the control group (7). In addition, because of limited resources and because sevelamer was not reimbursed by the national health care system at the time of the study design, the sample size was limited to ≤120 patients being treated with sevelamer for 3 years.
Consecutive asymptomatic outpatients receiving care at 12 nephrology clinics in South Italy were evaluated. Inclusion criteria were being aged ≥18 years, having 6 months of follow-up before the enrollment, and having stage 3–4 CKD. Exclusion criteria were heart failure, coronary artery disease, previous myocardial infarction, coronary by-pass, angioplasty, stroke, arrhythmia, liver dysfunction, nephrotic syndrome, and fast progression of kidney function (defined as 24-hour measured creatinine clearance loss ≥12 ml/min per year). The latter may not allow a long-lasting follow-up because of accelerate dialysis initiation. Patients were informed of the nature of study and that sevelamer was not approved by the Italian Health Department in NDD-CKD patients. Patients entered the study after they had provided written informed consent. This study was conducted in adherence to the Declaration of Helsinki.
Enrollment began in November 2005. Patients were randomized to receive either open-label calcium carbonate or sevelamer. A binder was randomly assigned (in a 1:1 fashion) by the coordinating center. A single simple randomization list was generated by computer and was concealed with the use of numbered, opaque sealed envelopes that were opened in sequence by the administrative personnel of the coordinating center not involved in patient care. All participating centers followed the same procedure via a phone call to the coordinating center.
The initial dose of each binder was established based on data of a previous study in predialysis patients in which 1600 mg/d sevelamer and 2000 mg/d calcium carbonate were found as the minimum dosages able to reduce urinary phosphorus excretion (11). Clinicians were free to increase the initial dose of binder to maintain serum phosphorus concentrations between 2.7 and 4.6 mg/dl for patients with stage 3–4 CKD, and between 3.5 and 5.5 mg/dl for patients with stage 5 CKD, as suggested by Kidney Disease Outcomes Quality Initiative (KDOQI) guidelines available at the time the study was started. Investigators were not blinded to binder assignment. The follow-up continued until death of any cause, dialysis inception, or 36 months from study entry.
Blood chemistry was assessed at enrollment and every 6 months. Laboratory variables recorded closer to the event or at the end of study was regarded as final; intermediate recordings performed after assignment to binder were averaged and reported as on-treatment average.
GFR was obtained by the 24-hour measured creatinine clearance. Intact parathyroid hormone (PTH) was assayed by using the chemiluminescent immunometric method (Diagnostic Products, Los Angeles, CA) and high-sensitivity C-reactive protein (CRP) by the immunoturbidimetric method.
The CAC score was assessed by a multi-slice computed tomography (CT) scan (GE Medical Systems) at study entry as well as scans at 6, 12, 18, and 24 months. CT scans were performed at and analyzed by a single center (Department of Radiology, Ospedale di Solofra, Avellino). Two readers unaware of phosphate binder assignments read the CT scans. CAC score was performed using a semi-automated threshold-dependent algorithm. Calcifications within the coronary artery tree above a threshold of 130 HU were included. CAC score was reported in Agatston units (AUs).
Data are expressed as mean ± SD, median and interquartile range (IQR), or frequencies. The paired samples t test for continuous variables and chi-squared test were used for comparison of baseline clinical characteristics and to compare final and on-treatment average values versus baseline values for both treatment groups. Kaplan–Meier survival curves were created and a long-rank test was used for comparisons between patients treated with sevelamer and with calcium carbonate. Multivariable cox-regression analysis was performed to assess predictors of events. Cox analysis was adjusted for covariates regarded as traditional predictors of all-cause mortality or dialysis inception such as age, sex, diabetes, hypertension, GFR, serum calcium, PTH, and CRP. Baseline time invariant and time-varying covariates (treated as repeated measures) were included in the models. Baseline time invariant covariates included age, sex, diabetes, hypertension, treatment group assignment after randomization, basal values of creatinine clearance, and CAC score. Time-varying covariates included creatinine clearance, serum concentration of calcium, phosphorus, PTH, CRP, total cholesterol, triglycerides, and LDL cholesterol. The largest possible meaningful considered model included clinical and biochemical variables on the basis of their plausible univariate relation to the outcome, considering the overall model fit and hazard proportionality. Variables to be dropped as non-confounders were eliminated manually (backward elimination), monitoring variations of the exposure regression coefficient and giving validity precedence over precision. The contribution of covariates to explain the dependent variable was assessed using a two-tailed Wald test, with P<0.05 considered as significant. Model specification, proportionality assumption, and overall fit were checked by using re-estimation and formal and graphical tests based on residuals, as well as by testing the interaction with time of the variables in the model. Data are expressed as hazard ratios and 95% confidence intervals.
Results
From the initial cohort of 330 NDD-CKD patients assessed for eligibility, 239 patients were enrolled in this study (Figure 1). Of these, 27 exited the study for various reasons. A total of 212 patients were considered the efficacy analyses. Patients randomized either to sevelamer (n=107) or calcium carbonate (n=105). The mean dose of sevelamer and calcium carbonate was 2184 mg/d (±592 SD) and 2950 mg/d (±703 SD), respectively. No differences in clinical characteristics or use of statins (35%), calcium channel blockers (15%), β-blockers (60%), ACE inhibitors/angiotensin II receptor antagonists (72%), or native or active vitamin D (10%) were found between groups.
;)
Figure 1: Patient flow chart indicating selection and discontinuation.
The changes in biochemistry observed in the two groups during the observation period are reported in Table 1. Final and on-treatment average phosphorous concentration was significantly (P<0.01) lower in patients treated with sevelamer; however, the proportion of patients on CKD stage–specific target for serum phosphorous was similar between the two groups. A better PTH control was also noted in sevelamer-treated patients compared with calcium carbonate–treated patients. Final and on-treatment average serum calcium concentration decreased in the sevelamer group (P<0.01) and increased in calcium carbonate group (P<0.01). Final and on-treatment average total cholesterol and LDL cholesterol were significantly reduced in the sevelamer group. Final triglycerides were higher both in patients treated with sevelamer (P<0.02) and calcium carbonate (P<0.02).
Table 1: Clinical and laboratory characteristics of the whole study population and according to the phosphate binder
The event-free survival curves of patients treated with sevelamer or calcium carbonate are shown in Figures 2, 3, and 4. The rate of all-cause mortality, dialysis inception, and the composite end point was significantly less frequent (log-rank test = 11.46; P<0.01) among patients randomized to sevelamer. Indeed, 12 and 22 patients died, 31 and 42 patents started dialysis, and 43 and 64 patients experienced the composite end point in the sevelamer and calcium carbonate group, respectively.
Figure 2: All-cause mortality in patients randomized either to sevelamer (dotted line) or calcium carbonate (continuous line). The all-cause mortality rate was significantly higher (P<0.05) among patients receiving calcium carbonate. Numbers indicate patients at risk for each time period.
Figure 3: Inception to dialysis in patients randomized either to sevelamer (dotted line) or calcium carbonate (continuous line). Numbers indicate patients at risk for each time period.
Figure 4: Event-free survival from the composite end point of all-cause mortality and dialysis inception among patients treated either with sevelamer (dotted line) or calcium carbonate (continuous line). Survival was significantly (log-rank test = 11.46; P<0.01) worse among patients receiving calcium carbonate. Numbers indicate patients at risk for each time period.
Results of unadjusted, adjusted for baseline covariates, and adjusted for baseline covariates plus time-varying covariates models are shown in Table 2. In all models, all-cause mortality and the composite end point were lower in patients receiving sevelamer. In unadjusted models and in adjusted models for baseline covariates, dialysis inception was lower in patients receiving sevelamer. In adjusted models for baseline covariates plus time-varying covariates, the protective effect of sevelamer was no longer observed.
Table 2: Hazard ratios and 95% confidence intervals for all-cause mortality, dialysis inception, and composite outcome for patients receiving sevelamer compared with patients receiving calcium carbonate
At baseline, the percentage of calcified patients (62.6% vs 47.6%; P=0.02) and the median CAC score (122 AU [IQR, 0–180] versus 0 AU [IQR, 0–215]; P<0.01] were higher in the sevelamer group.
Considering only patients with CAC scores >0, a significant regression of CAC score was observed in 24 patients treated with sevelamer and in 2 patients treated with calcium carbonate. During the 24-month observation period, de novo onset of CAC was recorded in 5 and in 45 patients assigned to sevelamer and calcium carbonate, respectively. The final cumulative percentage of de novo onset of CAC was 12.8% in sevelamer-treated patients and 81.8% in calcium carbonate–treated patients; in the latter group, the increase in CAC score was also greater.
Episodes of hypercalcemia were more frequent in patients taking calcium carbonate than in those taking sevelamer (n=82 [78%] versus n=6 [5%]; P<0.01).
Discussion
The main result of this pilot study is that a significant difference in all-cause mortality was achieved by treating NDD-CKD patients with different phosphate binders. To the best of our knowledge, no similar interventional study has been performed in NDD-CKD patients to date. Current data show that sevelamer distinctly improves survival in NDD-CKD patients; therefore, these data are in keeping with current guidelines that caution against the indiscriminate use of calcium-containing phosphate binders in CKD patients (5).
Final and on-treatment average phosphorus concentrations were higher among patients assigned to calcium carbonate as opposed to sevelamer. This could increase the suspicion that calcium carbonate was underdosed. Likely, the tendency toward a more rapid decline of renal function among calcium-treated patients may explain this finding. Indeed, more patients in the calcium arm progressed to stage 5 CKD during follow-up. In this stage, higher levels of serum phosphorus and PTH were allowed by the Kidney Disease Outcomes Quality Initiative guidelines; therefore, final and on-treatment average concentrations moved toward high levels. In addition, the proportion of patients on target for serum phosphorus during the follow-up was not different between the two groups. Finally, higher serum calcium concentrations and greater occurrence of episodes of hypercalcemia among patients treated with calcium carbonate further challenge the hypothesis that calcium carbonate was underdosed. Any further increase of calcium carbonate dosage would have exposed patients to greater risk of drug-related side effects with relevant consequences on outcomes.
The different all-cause mortality rates noted in the treatment groups cannot be fully explained by the current data. Despite the fact that the majority of available studies have shown a strong association between serum phosphorous and mortality, the all-cause mortality in our pilot study was not related to any parameter of mineral metabolism. Nevertheless, this is not the first study that failed to demonstrate such an association; the authors of a few previous studies have also reached a similar conclusion (11,12). Aside from statistical reasons, a hypothesis can be advanced. Serum phosphorus represents a minimal part of the body phosphorus pool and positive phosphorus balance can be achieved without a significant change of its serum level. This may be particularly important in early stages of CKD, in which compensatory mechanisms such as fibroblast growth factor 23 (FGF-23) are operative. Thus, the all-cause mortality might be associated with phosphorous overload rather than with its serum level. However, due to the fact that markers of phosphorous balance such as FGF-23 or phosphaturia were not measured in this study, this hypothesis cannot be validated.
Finally, the survival benefit in the sevelamer group could be related to the several pleiotropic effects of the binder beyond its hypophosphatemic effects (13–15). For instance, reductions in CRP, total cholesterol, and LDL cholesterol were observed in our patients and may have accounted for at least part of the difference in all-cause mortality.
Despite the fact that we did not aim specifically to examine CKD progression, the different effects of study phosphate binders on dialysis inception are worth mentioning. At study completion, the percentage of patients starting dialysis was lower in the sevelamer arm compared with the calcium carbonate arm. However, whereas in some statistical models sevelamer was the sole protective factor from inception of dialysis, its protective effect was lost in the model adjusted for time-varying covariates. This was likely due to the small size of the study cohort.
The major limitation of this study is the sample size. At the time the study began, there were no data available from the medical literature on the effects of phosphate binders on all-cause mortality in NDD-CKD patients. Hence, this study was designed as a pilot study in which we aimed to randomize 240 patients to have adequate power to detect a 60% reduction in all-cause mortality in the experimental arm. In addition, because of limited resources and because sevelamer was not reimbursed by the national health care at the time of the study was designed, the sample size was limited to ≤120 patients being treated with sevelamer for 3 years.
Caution should be used in interpreting the results of regression models due to the number of covariates with respect to the occurred events. Nevertheless, the randomized design may ensure the results attained on the estimate of treatment effect.
One other potential limitation may be the fact that there was no specific target of serum phosphorus to be reached in all patients. However, clinicians were free to increase the initial dose of binder at their discretion to maintain serum phosphorus concentrations in the ranges suggested by KDOQI guidelines according to the CKD stage of each patient. Rather than a limitation, this should be regarded as a strength because the study is representative of real life.
Finally, urinary phosphorus excretion or serum FGF-23 levels were not measured. These assessments could have shed some light on phosphate balance as well as compliance to binder therapy and should likely be included in future studies. Despite these limitations, the study might represent a valid aid for designing a properly powered RCT.
Sevelamer provided benefits in all-cause mortality and the composite end point but had less evident advantages in the progression of CKD. Despite the study’s many strengths such as hard end points, prospective design, long follow-up, and repeated measurements of blood chemistry, larger studies are needed to confirm these results.
Disclosures
A.B. has received speaking honoraria from Genzyme and Amgen. D.R. has received speaking honoraria from Abbott, Genzyme, Amgen, Roche, Fresenius Medical Care, Novartis, and Bristol, as well as a research grant from Amgen.
Acknowledgments
The following members of the INDEPENDENT-CKD Study Group collaborated with the authors of this study: Serena Torraca (Salerno); Maria Luisa Sirico (Caserta); Lucia Di Micco (Napoli); Vincenzo Bellizzi (Ospedale Ruggi D’aragona, Salerno); Filippo Aucella (Ospedale S. Giovanni Rotondo, Foggia); Pasquale Guastaferro and Angela Di Gianni (S. Angelo dè Lombardi, Avellino); Luigi Chiuchiulo (Avellino, Dialysis); Vincenzo Tedesco (Montella, Avellino); Mario Migliorati (S. Giorgio A. Cremano, Napoli); Walter De Simone and Bruno Zito (Ospedale Avellino); Ernesto D’avanzo, Sara Bortone, Paola Nargi, and Francesco Saverio Iannaccone (Solofra, Radiologia, Avellino); Patrizia Veniero, Maria Capuano, and Raffaele Genualdo (Ospedale dei Pellegrini, Napoli); Magda Lorenzo, Domenico Santoro, and Ferdinando Avella (Ospedale Nola, Napoli); Luigi Morrone and Vinicio Martignetti (Ospedale Benevento); Carmine Piscopo and Lorella Berardino (Ospedale Solofra, Cardiologia, Avellino); Andrea Pota (Maddaloni); Domenico Matarese and Domenico Vigilante (Ospedale Lucera, Foggia); and Assunta Aquino (Avellino, Malzoni), Rosa Martino, Struzziero Giuseppe, Alfonso Frallicciardi, and Raffaele Tortoriello (ASL Avellino).
Published online ahead of print. Publication date available at www.cjasn.org.
References
1. Raggi P, Boulay A, Chasan-Taber S, Amin N, Dillon M, Burke SK, Chertow GM: Cardiac calcification in adult hemodialysis patients. A link between end-stage renal disease and cardiovascular disease? J Am Coll Cardiol 39: 695–701, 2002
2. Sarnak MJ: Cardiovascular complications in chronic kidney disease. Am J Kidney Dis 41[Suppl]: 11–17, 2003
3. London GM: Cardiovascular calcifications in uremic patients: Clinical impact on cardiovascular function. J Am Soc Nephrol 14[Suppl 4]: S305–S309, 2003
4. National Kidney Foundation: K/DOQI clinical practice guidelines for bone metabolism and disease in chronic kidney disease. Am J Kidney Dis 42[Suppl 3]: S1–S201, 2003
5. Kidney Disease: Improving Global Outcomes (KDIGO) CKD-MBD Work GroupKDIGO clinical practice guideline for the diagnosis, evaluation, prevention, and treatment of chronic kidney disease-mineral and bone disorder (CKD-MBD). Kidney Int 76: S1–S130, 2009
6. Russo D, Palmiero G, De Blasio AP, Balletta MM, Andreucci VE: Coronary artery calcification in patients with CRF not undergoing dialysis. Am J Kidney Dis 44: 1024–1030, 2004
7. Russo D, Corrao S, Miranda I, Ruocco C, Manzi S, Elefante R, Brancaccio D, Cozzolino M, Biondi ML, Andreucci VE: Progression of coronary artery calcification in predialysis patients. Am J Nephrol 27: 152–158, 2007
8. Kramer H, Toto R, Peshock R, Cooper R, Victor R: Association between chronic kidney disease and coronary artery calcification: The Dallas Heart Study. J Am Soc Nephrol 16: 507–513, 2005
9. Chiu YW, Adler SG, Budoff MJ, Takasu J, Ashai J, Mehrotra R: Coronary artery calcification and mortality in diabetic patients with proteinuria. Kidney Int 77: 1107–1114, 2010
10. Merjanian R, Budoff M, Adler S, Berman N, Mehrotra R: Coronary artery, aortic wall, and valvular calcification in nondialyzed individuals with type 2 diabetes and renal disease. Kidney Int 64: 263–271, 2003
11. Russo D, Miranda I, Ruocco C, Battaglia Y, Buonanno E, Manzi S, Russo L, Scafarto A, Andreucci VE: The progression of coronary artery calcification in predialysis patients on calcium carbonate or sevelamer. Kidney Int 72: 1255–1261, 2007
12. Russo D, Morrone LF, Brancaccio S, Napolitano P, Salvatore E, Spadola R, Imbriaco M, Russo CV, Andreucci VE: Pulse pressure and presence of coronary artery calcification. Clin J Am Soc Nephrol 4: 316–322, 2009
13. Chertow GM, Burke SK, Dillon MA, Slatopolsky E: Long-term effects of sevelamer hydrochloride on the calcium x phosphate product and lipid profile of haemodialysis patients. Nephrol Dial Transplant 14: 2907–2914, 1999
14. Chertow GM, Burke SK, Raggi PTreat to Goal Working Group: Sevelamer attenuates the progression of coronary and aortic calcification in hemodialysis patients. Kidney Int 62: 245–252, 2002
15. Brandenburg VM, Schlieper G, Heussen N, Holzmann S, Busch B, Evenepoel P, Vanholder R, Meijers B, Meert N, Fassbender WJ, Floege J, Jahnen-Dechent W, Ketteler M: Serological cardiovascular and mortality risk predictors in dialysis patients receiving sevelamer: A prospective study. Nephrol Dial Transplant 25: 2672–2679, 2010
