Effects of calcium-containing phosphate binders on cardiovascular events and mortality in predialysis CKD stage 5 patients.

BACKGROUND: Hyperphosphatemia and calcium load were associated with vascular calcification. The role of calcium-containing phosphate binders (CCPBs) use as important determinants of death and cardiovascular events in predialysis hyperphosphatemic chronic kidney disease (CKD) patients remain unclear due to the absence of evidence for reduced mortality with CCPB use compared with placebo. We aimed to investigate the effect of using CCPBs or nonuse in all-cause mortality rates and cardiovascular events in CKD stage 5 patients between 2000 and 2005 in the Taiwanese National Health Insurance Research Database.
METHODS: Patients with known coronary heart disease and those who had undergone dialysis or renal transplantation were excluded. The CCPB users were matched with nonusers by the propensity score model. Multivariable Cox proportional hazards model was used to estimate hazard ratios (HRs) of all-cause mortality and cardiovascular events.
RESULTS: During a mean follow-up of 4.58 years, 879 CCPB users were matched with 3516 nonusers. CCPB use was an independent risk factor for cardiovascular events [adjusted hazard ratio (HR) 1.583, 95% confidence interval (CI) 1.393-1.799]. The increased cardiovascular risk was dose-dependent and consistent across all subgroup analyses. Compared with no use, CCPB use was associated with no significant all-cause mortality risk (1.74 vs. 1.75 events per 100 person-years, adjusted HR 0.964, 95% CI 0.692-1.310).
CONCLUSIONS: CCPB use in CKD stage 5 patients was associated with a significantly increased cardiovascular event risk compared with the nonusers, whereas the all-cause mortality risk was similar between the two groups. Whether these relationships are causal require further randomized controlled trials.

Introduction

Chronic kidney disease (CKD) is a worldwide public health issue that caused premature mortality or 20 million disability-adjusted life years in 2010 [1]. Accelerated and progressive vascular calcification might play major roles in the development of cardiovascular disease, the leading cause of death in CKD [2]. During the evolution of CKD, adaptive responses fail to maintain the calcium–phosphate homeostasis. Hyperphosphatemia develops gradually, with the prevalence increasing up to 40% in those with estimated glomerular filtration < 20 ml/min/1.73m [3, 4], which is associated with the progression of secondary hyperparathyroidism and renal bone disease. Observational studies showed a significant association between hyperphosphatemia and high mortality among patients on hemodialysis [2]. Furthermore, high serum phosphorus levels were shown to be associated with increased cardiovascular mortality and hospitalization risks among patients on hemodialysis [3]. As an inevitable clinical result of stage 5 CKD, controlling elevated serum phosphate levels by dietary restriction and intestinal phosphate binders is an important issue in clinical practice. Phosphate binders are suggested by the Kidney Disease: Improving Global Outcomes (KDIGO) clinical practice guidelines to treat hyperphosphatemia in patients with stage 3–5 nondialysis CKD [5].

Calcium-containing phosphate binders (CCPBs) reduce serum phosphorus levels in advanced CKD. However, the positive calcium balance observed in several studies suggested soft-tissue deposition after calcium exposure, implicating the potentially harmful effects of generous calcium supplementation in patients with stage 3–4 CKD and normal phosphate levels [6, 7]. CCPBs is an important cause of hypercalcemia in CKD population, raising the risk of vascular calcification and leading to cardiovascular disease [8]. The prevalence of hypercalcemia in CKD stage 4 and 5 patients was 13.4%, which was lower than dialysis patients. The cause of higher rate of hypercalcemia in dialysis patients might be due to greater usage of vitamin D analogues and high-dose CCPBs [9]. In hyperphosphatemic patients with predialysis CKD, CCPBs were reported to be associated with worse adverse effects on vascular calcification and significantly higher all-cause mortality compared with calcium-free binders [10, 11]. These results were reinforced by several systematic reviews [12-15], altogether suggesting that calcium-free phosphate binders might be either beneficial or otherwise not harmful compared with CCPBs in treating hyperphosphatemia in CKD. Based on the cumulative outcomes suggesting potential damage with CCPBs, the KDIGO CKD mineral bone disorder (MBD) guidelines recommend restricting the CCPB dose in predialysis patients with CKD [16].

It remains uncertain whether preventing the development of hyperphosphatemia with CCPBs might prevent or enhance the development of vascular calcification, cardiovascular risks and mortality and in patients with stage 5 CKD. Although CCPBs are used widely in patients with high cardiovascular risk, there are no adequately powered, well-designed, and long-term randomized controlled trials to evaluate whether regular CCPB use in stage 5 CKD affects serious outcomes such as mortality or cardiovascular events compared with placebo or no treatment. However, there is inadequate motivation for the pharmaceutical industry to support clinical trials for medications that are already considered necessary by practicing physicians and have attained high utility in the CKD population. We, therefore, conducted a nationwide population-based retrospective observational cohort study using the National Health Insurance (NHI) Research Database (NHIRD) in Taiwan, which provided an exceptional opportunity to evaluate the effects of CCPBs on mortality and coronary heart disease risks in stage 5 CKD.

Materials and methods

Data source

This nationwide population-based cohort study uses Taiwan's NHIRD, comprising the standard health care data submitted by medical institutions seeking reimbursement through the NHI, which covers the medical needs of 99.19% of the 23 million individuals in the population of Taiwan. The identified information derived from the NHIRD included birth date, sex, residence area, drug prescriptions, medical procedures, and diagnostic codes according to the International Classification of Diseases 9th Revision-Clinical Modification (ICD-9-CM). The study was approved by the Institutional Review Board of Tri-Service General Hospital (TSGH IRB No. 2–105–05–082.).

Study design and participants

This study included patients with a primary diagnosis of CKD who received erythropoiesis-stimulating agents (ESAs) between January 1, 2000, and June 30, 2005, and were followed up until December 31, 2005, as potential study subjects who were identified in the NHIRD.

The NHI reimbursement regulations in Taiwan [17] state that the ESA treatment can be initiated in predialysis patients with CKD accompanied with a serum creatinine concentration > 530 μmol/L (approximately equivalent to stage 5 CKD) and a hematocrit level < 28% to maintain a packed-cell volume of no more than 36%. The selected cohort, which was validated previously, was a representative of the predialysis patients with stage 5 CKD in Taiwan [18, 19]. The current study included patients with a primary diagnosis of CKD (ICD-9 codes, including 016.0, 042, 095.4, 189, 223, 236.9, 250.4, 271.4, 274.1, 440.1, 442.1, 446.21, 447.3, 572.4, 580–589, 590–591, 593, 642.1, 646.2, 753, and 984) and those who were receiving ESAs covered by the NHI (identified by a serum creatinine concentration > 530 μmol/L). Patients younger than 18 years, those with coronary heart disease diagnosed before the study period and those who had undergone dialysis or renal transplantation before the ESA treatment, were excluded. The prescription information of the 90 days after the first ESA treatment was used to ascertain CCPB use, and the 91st day after the first ESA prescription was defined as the index date for study entry. The patients who commenced renal replacement therapy or died and those who were prescribed a CCPB between the first ESA treatment and the index date were also excluded.

The patients who were using a CCPB within the 90 days after the first ESA prescription were defined as CCPB users and the remaining patients were designated as nonusers. The ICD-9-CM codes were used to identify comorbidities that were diagnosed within the three years before the index date and included the following diagnoses (ICD-9-CM codes) that are considered to increase the mortality risk: diabetes mellitus (250), hypertensive cardiovascular disease (401–405), heart failure (428), cerebrovascular disease (430–438), cirrhosis (571), peripheral arterial occlusive disease (440.0, 440.2, 440.3, 440.8, 440.9, 443, 444.0, 444.22, 444.8, 447.8, and 447.9), cancer (140–208), and dyslipidemia (272). The Charlson comorbidity index was used to quantify the comorbidity profiles. To balance the two groups based on known confounding factors, the propensity score for the likelihood of receiving CCPBs was calculated using the baseline covariates listed in Table 1. The CCPB users were matched to the nonusers using a ratio of 1:4 based on age, sex, and the propensity score.

Table 1

Baseline characteristics of the study patients by calcium-containing phosphate binder use before and after propensity score matching.

	Before propensity score matching			After propensity score matching
	User (n = 879)	Nonuser (n = 7245)	P value	User (n = 879)	Nonuser (n = 3516)	P value*
Age (years)	64.80 ± 13.72	60.75 ± 14.96	<0.001	64.80 ± 13.72	64.63 ± 14.39	0.751
Age group (years)			0.025			0.999
18–44	87 (9.90)	806 (11.12)		87 (9.90)	348 (9.90)
45–64	308 (35.04)	2,801 (38.66)		308 (35.04)	1,232 (35.04)
≥65	484 (55.06)	3,638 (50.21)		484 (55.06)	1,936 (55.06)
Sex			0.150			0.999
Male	446 (50.74)	3,862 (53.31)		446 (50.74)	1,784 (50.74)
Female	433 (49.26)	3,383 (46.69)		433 (49.26)	1,732 (49.26)
Comorbidities
Diabetes mellitus	322 (36.63)	2,624 (36.22)	0.809	322 (36.63)	1,295 (36.83)	0.913
Hypertensive cardiovascular disease	326 (37.09)	1,949 (26.90)	<0.001	326 (37.09)	1,277 (36.32)	0.672
Dyslipidemia	24 (2.73)	297 (4.10)	0.049	24 (2.73)	89 (2.53)	0.739
Cirrhosis	29 (3.30)	582 (8.03)	<0.001	29 (3.30)	125 (3.56)	0.712
Cancer	53 (6.03)	487 (6.72)	0.436	53 (6.03)	215 (6.11)	0.925
Cerebrovascular disease	53 (6.03)	377 (5.20)	0.302	53 (6.03)	217 (6.17)	0.875
Heart failure	100 (11.38)	894 (12.34)	0.441	100 (11.38)	397 (11.29)	0.943
Charlson comorbidity index score	0.18 ± 0.50	0.28 ± 0.63	<0.001	0.18 ± 0.50	0.21 ± 0.55	0.123

* The chi-square or Fisher’s exact test was used to evaluate categorical variables, and Student’s t-test was used for continuous variables.

Study outcomes and follow-up

The primary outcome was all-cause mortality. The primary diagnosis of in-hospital death or the first-listed discharge diagnosis at the last hospitalization within the three months before death for out-of-hospital deaths was defined as the cause of death, as indicated previously [20]. To identify the association between CCPBs and coronary artery calcification in patients with CKD, the number of admissions for coronary heart disease (ICD-9 codes 410–414, excluding 412 and 414.1) was also determined.

Statistical analysis

The clinical characteristics of the CCPB users and nonusers were described using percentages and means ± standard deviation for categorical and continuous variables, respectively. Differences between the CCPB users and nonusers were compared by independent Student’s t-test or the χ2 test, where appropriate. The cumulative incidences of all-cause mortality and coronary heart disease over time were compared between the CCPB users and nonusers using the Kaplan–Meier method.

The multivariable Cox proportional hazards model was used before and after propensity score matching to estimate hazard ratios (HRs) of outcomes after controlling for age, sex, diabetes mellitus, hypertensive cardiovascular disease, heart failure, cerebrovascular disease, cirrhosis, cancer, dyslipidemia, and Charlson comorbidity index score. Subgroup analyses were also performed to determine the HRs of the outcomes in the CCPB users and nonusers.

To assess cumulative dose-related effects, the risks of mortality and coronary heart disease were evaluated according to the defined daily dose (DDD) during the 90-day exposure period (≤15, 16–40, and >40 DDD) and the prescribed daily dose (≤500, 501–1000, and >1000 mg) relative to no CCPB use. DDD represents as a technical unit of measurement which defines the assumed usual maintenance dose per day for a drug used for its main indication in adults. Though only give an estimate of consumption and not an exact picture of actual use, DDD is useful in assessing trends in drug consumption and performing comparisons between population groups. A two-tailed P value < 0.05 was considered statistically significant. All statistical analyses were performed using SAS (version 9.3; SAS Institute) and Stata SE (version 11.0; StataCorp).

Results

Patient characteristics

A total of 8124 patients diagnosed with stage 5 CKD between January 1, 2000, and June 30, 2005, including 879 CCPB users and 7245 nonusers, comprised the study cohort, as depicted in Fig 1. Table 1 shows the baseline characteristics of the two groups categorized by CCPB use before and after propensity score matching. Before matching, the users were older and more likely to have hypertensive cardiovascular disease. The rates of those with cirrhosis and higher Charlson comorbidity index scores were greater among the nonusers compared with the CCPB users. After matching with a 1:4 ratio according to the propensity score, 879 CCPB users and 3516 nonusers were included in the primary outcome; the baseline characteristics of the two groups after matching were comparable.

Fig 1

Study profile.

aNumbers for exclusions may not sum because of patients fulfilling more than one criterion. bPatients in the treated cohort (received CCPB for ≥ 90 days) were matched at a ratio of 1:4 with those in the untreated cohort (never received CCPB) by means of propensity scores. Abbreviations: CKD, chronic kidney disease; ESA, erythropoiesis-stimulating agent; CCPB, calcium-containing phosphate binder.

The association of CCPB use with treatment outcomes

During the follow-up period, coronary heart disease was diagnosed in 462/879 (52.55%) CCPB users and 1367/3516 (38.87%) nonusers. The incidence of coronary heart disease was higher among the CCPB users compared with the nonusers (11.99 vs. 8.27 per 100 person-years, crude HR 1.607, 95% confidence interval [CI] 1.442–1.891). After adjustment for age, sex, baseline comorbidities, and Charlson comorbidity index score, the risk for coronary heart disease remained higher in the CCPB users (adjusted HR 1.583, 95% CI 1.393–1.799) (Table 2). However, the risk for all-cause mortality was not significantly different between the CCPB users and the nonusers. In the matched cohort, 111/879 (12.6%) CCPB users and 333/3516 (9.47%) nonusers died during the follow-up period (incidence, 1.74 and 1.75 events per 100 person-years for the CCPB users and the nonusers, respectively; crude HR 0.896, 95% CI 0.521–1.112) (Table 2). The association did not change after adjusting for the baseline covariates (adjusted HR 0.964, 95% CI 0.692–1.310).

Table 2

Risks of all-cause mortality and coronary heart disease in predialysis patients with chronic kidney disease after propensity score matching (n = 4395).

	All-cause mortality				Coronary heart disease
	Events	Incidence	Crude HR	Adjusted HR	Events	Incidence	Crude HR	Adjusted HR
	(n/N)	(per 100 PYs)	(95% CI)	(95% CI)	(n/N)	(per 100 PYs)	(95% CI)	(95% CI)
Matched cohort
Nonusers	333/3516	1.75	1	1	1367/3516	8.27	1	1
Users	111/879	1.74	0.896 (0.521–1.112)	0.964 (0.692–1.310)	462/879	11.99	1.607 (1.442–1.891)	1.583 (1.393–1.799)
Defined daily dose
≤15 DDD	33/276	1.82	0.938 (0.542–1.163)	1.006 (0.721–1.368)	121/276	11.31	1.516 (1.358–1.790)	1.492 (1.311–1.699)
16–40 DDD	39/291	1.79	0.924 (0.531–1.145)	0.995 (0.709–1.352)	153/291	11.98	1.608 (1.440–1.892)	1.585 (1.390–1.801)
≥41 DDD	39/312	1.64	0.845 (0.488–1.049)	0.903 (0.648–1.234)	188/312	12.47	1.673 (1.498–1.964)	1.649 (1.444–1.878)
P for trend				0.279				0.064
Prescribed daily dose
≤500 mg/day	19/183	1.96	1.009 (0.583–1.251)	1.084 (0.777–1.483)	89/183	10.24	1.372 (1.231–1.631)	1.352 (1.184–1.538)
501–1000 mg/day	34/279	1.83	0.937 (0.542–1.167)	1.011 (0.718–1.386)	124/279	11.45	1.531 (1.376–1.813)	1.511 (1.326–1.724)
≥1000 mg/day	58/417	1.64	0.848 (0.489–1.052)	0.91 (0.651–1.240)	249/417	13.09	1.759 (1.574–2.084)	1.73(1.518–1.962)
P for trend				0.156				0.038

Abbreviations: CI, confidence interval; DDD, defined daily dose; HR, hazard ratio; PYs, person-years.

The apparent adverse effects of CCPBs increased substantially with increasing doses. Compared with the nonusers, those using CCPBs had an increased risk of coronary heart disease, and those who were using CCPBs at more than 1000 mg/day had the greatest risk (adjusted HR 1.73, 95%CI 1.518–1.962; P = 0.038 for trend). An increased risk of coronary heart disease was noted with the cumulative CCPB doses greater than 40 DDD (adjusted HR 1.649, 95% CI 1.444–1.878; P = 0.064 for trend). However, this dose-response relationship was not noted for all-cause mortality (Table 2).

For coronary heart disease and all-cause mortality, Kaplan–Meier method was used. Events during the first 3 months were excluded. The cumulative incidence for coronary heart disease was significantly higher among the CCPB users compared with the nonusers (Fig 2A); however, a similar difference was not observed for all-cause mortality (Fig 2B).

Fig 2

Kaplan–Meier Curves for the Cumulative Incidence of (A) Coronary Heart Disease and (B) All-cause Mortality with Calcium-containing Phosphate Binder Use. Comparison of the cumulative incidence of coronary heart disease and all-cause mortality were shown. P values<0.05 were considered statistically significant.

Subgroup and sensitivity analyses

The subgroup and sensitivity analyses were consistent with the analyses of the two main groups, with a comparison of the treated (received CCPB for ≥ 90 days) and untreated (never received CCPB) cohorts. Among patients with advanced CKD, CCPB use was found to be associated with increased risk of coronary heart disease across all clinically relevant subgroups (stratified by age, sex, diabetes mellitus, hypertensive disease, dyslipidemia, liver cirrhosis, cancer, cerebrovascular disease and heart failure) (Fig 3A), whereas the all-cause mortality risk was not significant in any of the subgroup analyses (Fig 3B).

Fig 3

Multivariable Stratified Analyses of the Association Between Calcium-containing Phosphate Binder Use and (A) Coronary Heart Disease Development and (B) All-cause Mortality. Hazard ratios from the subgroup analysis for coronary heart disease and all-cause mortality between CCPBs user and nonuser were shown.

Discussion

This national population-based observational cohort study showed that CCPB use was associated with a 58% increased risk of coronary heart disease among patients with CKD stage 5 after the propensity score matching. A higher risk was noted in those on higher CCPB doses; this result was supported by the observed positive dose-response relationship. Furthermore, these findings were confirmed with sensitivity and subgroup analyses. Although the outcomes of observational studies should be interpreted carefully, the observed dose-dependent association provides strong support for a true relationship between CCPBs and coronary heart disease in advanced CKD. Another major study finding was the lack of an association between CCPBs and higher all-cause mortality risk in predialysis patients with CKD.

Hyperphosphatemia is associated with a higher risk of vascular calcification and worse cardiovascular outcomes [21-25]. CCPBs are widely used to control serum phosphate levels; however, concerns remain regarding their effect on the calcium balance. The risk of hypercalcemia was higher in the CCPB users compared with nonusers [14, 26]. In the CKD population, the harmful effects of calcium are consistent with the role of calcium in the pathophysiology of vascular calcification [27, 28], which may contribute to the initiation or progression of vascular stiffness [29] and coronary artery calcification [30]. In the current study, we found an increase in coronary heart disease in the patients on CCPBs compared with the nonusers, which lends further evidence for the contribution of the positive calcium balance to vascular calcification. Similarly, we found an increased coronary heart disease risk with higher cumulative CCPB doses. The progression of coronary artery calcification was demonstrated to be a reliable method to predict mortality [31, 32], which support the theoretical risk of inducing a positive calcium balance in patients with CKD, in whom an abnormal mineral metabolism might result in dystrophic calcification and higher mortality. To date, two observational studies have examined the association of phosphate binders with a survival benefit in predialysis CKD patients [33, 34], with different results of phosphate binder use. In addition to selection bias, sex, and different CKD stages, the heterogeneity of the phosphate binders use might have also contributed to the conflicting results regarding the survival benefit. In the current study, we focused on the effect of CCPBs on vascular events and mortality and therefore enrolled patients during a period of time when phosphate binders were generally restricted to CCPBs in Taiwan, with consideration of confounding factors such as marketing efforts in promoting non-CCPBs [35]. However, we were unable to demonstrate whether the increase in coronary heart disease in CCPB users was due to an increase in the positive calcium balance or phosphate toxicity because the relevant laboratory values, including baseline calcium, phosphorus, and intact parathyroid hormone concentrations, were not available for adjustments or comparisons. Therefore, well-designed placebo-controlled trials are warranted to determine whether CCPB use in predialysis CKD patients obtain more benefits from lowering phosphate levels or harm in increasing calcium levels.

Winkelmayer et al. evaluated whether CCPBs were associated with increased mortality risk in dialysis-dependent patients with CKD and found no association between CCPB use and 1-year all-cause mortality in dialysis patients (adjusted HR for all-cause mortality in users vs. nonusers 0.85, 95% CI 0.72–1.10) [36]. Although no significant reduction in mortality was observed in the study by Winkelmayer et al. and our study, several differences between these two studies merit discussion. While Winkelmayer et al. enrolled patients on dialysis, we included only predialysis patients. To determine whether lowering serum phosphate with CCPBs achieves net benefits or risks, an association of lower serum phosphate levels with reduced mortality and cardiovascular event rates should be demonstrated. However, dialysis removes phosphorus more efficiently than CCPBs. Thus, we could not distinguish whether the benefit of lowering phosphate was attributable to the phosphate binders or phosphorus removal via dialysis. In the current study, we limited the entire cohort to patients with advanced CKD not on dialysis to remove the confounding effect of dialysis on phosphate levels. Furthermore, we investigated the effect of CCPBs on cardiovascular events, which was not included in the study by Winkelmayer et al. Another difference between the studies is the observation time, which was longer in the current study.

Several important strengths of our study merit discussion. Although clinical trials are the gold standard for evaluating the comparative efficacy of therapeutic interventions, observational studies are useful to evaluate the benefits and adverse effects of CCPBs in the absence of well-designed randomized controlled trials evaluating CCPB versus placebo in predialysis patients with CKD. The study cohort was a large national representation of CCPB use to assess its effect on the development of coronary heart events in predialysis patients with CKD, which is a clinically significant issue that, to our knowledge, has not yet been addressed. Furthermore, the current study examined the effect of CCPB use on all-cause mortality in these patients. Our findings were robust across sensitivity and subgroup analyses, indicating that these findings are generalizable to the predialysis CKD populations with multiple concomitant comorbidities. However, several limitations of the current study should also be acknowledged. First, use and nonuse of CCPB were determined within 90 days after the index date; however, the drug status was not ascertained during the follow-up period and misclassification of drug exposure of interest is possible, which could have resulted in the observed null association between CCPB use and all-cause mortality. And the use of DDD did not permit a more meaningful comparison of calcium exposure as compared to the accurate computation of daily elemental calcium received from phosphate binders. Second, the unmatched CCPB nonusers were generally younger and the rate of hypertension, a known cardiovascular risk factor, was lower, compared with the CCPB users matched by the propensity score. Although the analyses were adjusted for demographic and patient-level clinical variables, this was an observational study with potential unmeasured confounders, such as alcohol or tobacco use, other medication use including lanthanum, sevalamer, warfarin or active vitamin D treatment, and echocardiographic abnormalities. Besides, the laboratory data were not included, and whether baseline phosphorus, calcium, CRP, and parathyroid hormone levels were comparable between the two groups remains uncertain. Furthermore, data on baseline creatinine levels and estimated glomerular filtration rates were not available. The intrinsic limitation of the NHIRD did not allow us to explore a larger CKD cohort of patients with creatinine concentrations < 530 μmol/L who could be identified due to the NHI regulations on ESA prescriptions. Third, based on the different phosphorus content of each meal, a distinctive approach to CCPB use allows for patient authorization with self-adjustment of phosphate binders. Moreover, adherence to phosphate binders confounds the relationship between the medication prescription and the clinical outcomes. Fourth, for most patients, coronary artery calcification is not clinically apparent for at least ten years of dialysis treatment [37]. The duration of the current cohort study might be relatively short to provide solid evidence on the effects of CCPB treatments on vascular calcification or cardiovascular events. Finally, the current study included Taiwanese patients with advanced CKD not on dialysis, and the conclusions may not be extrapolated to other ethnicities or populations.

Conclusions

In summary, compared with the nonusers matched by propensity score, CCPB users had a significantly higher coronary heart disease risk in the current national registry study of patients with stage 5 CKD. The increased coronary heart disease risk was dose-dependent and consistent across all subgroups of interest. However, CCPB use was not associated with higher all-cause mortality risk. Although insufficient in providing a useful cost-effectiveness model, the current study findings have important therapeutic implications, supporting the current KDIGO CKD-MBD recommendations and providing evidence that liberal CCPB use should be restricted in predialysis patients with CKD.

(XLS)

Click here for additional data file.

9 Jul 2020

PONE-D-20-04452

Effects of Calcium-Containing Phosphate Binders on Cardiovascular Events and Mortality in Predialysis CKD Stage 5 Patients

PLOS ONE

Dear Dr. Chu,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

SPECIFIC ACADEMIC EDITOR COMMENTS: Four expert reviewers in the field handled your manuscript. We thank them for their time and efforts. Although interest was found in your study, there were major concerns that arose during review that overshadowed this enthusiasm. These concerns include the need to better explain several vague comments, including the rationale for conducting this study; questions about the experimental design, including dose and duration of CCPB versus non-calcium based phosphate binder usage; more specifics about the patients cohort need to be provided and additional outcomes are requested; and there are comments about the limitations of this study. All reviewers' comments must be addressed in the revised manuscript.

Please submit your revised manuscript by Aug 23 2020 11:59PM. If you will need more time than this to complete your revisions, please reply to this message or contact the journal office at plosone@plos.org. When you're ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

Please include the following items when submitting your revised manuscript:

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.

A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.

An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter.

If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see: http://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols

We look forward to receiving your revised manuscript.

Kind regards,

Frank T. Spradley

Academic Editor

PLOS ONE

Journal requirements:

When submitting your revision, we need you to address these additional requirements.

1. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at

https://journals.plos.org/plosone/s/file?id=wjVg/PLOSOne_formatting_sample_main_body.pdf and

https://journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf

2. Please upload a new copy of Figures 1 & 3 as the detail is not clear. Please follow the link for more information: https://blogs.plos.org/plos/2019/06/looking-good-tips-for-creating-your-plos-figures-graphics/" https://blogs.plos.org/plos/2019/06/looking-good-tips-for-creating-your-plos-figures-graphics/

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Yes

Reviewer #2: Yes

Reviewer #3: Yes

Reviewer #4: Yes

**********

2. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: Yes

Reviewer #2: Yes

Reviewer #3: Yes

Reviewer #4: Yes

**********

3. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

Reviewer #2: Yes

Reviewer #3: Yes

Reviewer #4: Yes

**********

4. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

Reviewer #2: Yes

Reviewer #3: Yes

Reviewer #4: Yes

**********

5. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: The authors did an excellent job analyzing the effect of calcium containing phosphate binders on cardiovascular mortality.

However it is important that the authors identify with more detail the limitations of the study. The authors have mentioned the difficulty to separate Phosphate levels vs calcium load from the binder. There are other factors that should be commented: anticuoagulants-warfarin, treatment with active vitamin D, patients adherence to treatment not only relative to phosphtae control but also interdialytic weight gain, blood pressure medication, PTH levels , CRP levels and others

Reviewer #2: Dear Prof Spradley,

Re: Manuscript ("Effects of Calcium-Containing Phosphate Binders on Cardiovascular Events and Mortality in Predialysis CKD Stage 5 Patients" (PONE-D-20-04452))

Thank you for the kind invitation to review this manuscript. This study is one of the largest study to evaluate the relationship between calcium based phosphate binder and cardiovascular risk among pre-dialysis patients which has not been adequately examined. The study results are interesting and the manuscript is generally well-written.

Attached below are my comments and clarifications regarding the manuscript for the authors’ consideration.

Major comments

1) It will be good for the authors to comment on their focus on Stage 5 predialysis patients and also why patients in stage 4 were not considered. [Noted this was only brought up in the discussion briefly but may be good to describe in the methodology]

2) An important issue of note to evaluate the dose dependent relationship of calcium based phosphate binders is that the dose of elemental calcium within the CCB should be computed. This will facilitate more accurate and meaningful evaluation of this relationship as different CCPB have varying contents of elemental calcium. From the described methodology, it seems that it would be possible to compute the elemental calcium as the dose is available and the name of the CCPB.

3) How was the segregation of the subgroups of daily defined dose (≤15, 16–40, and >40 DDD) determined?

4) An important factor for consideration that would substantially strengthen the relationship between use of CCPB and adverse outcomes would be the use of non-calcium based phosphate binders e.g. lanthanum, and sevalamer as well as vitamin D analogues. It will be good to report these information among users of CCPB and non-users and adjust for theses confounders (especially non-calcium based phosphate binders) in the multivariate analyses.

5) Discussion: beyond the limits of study designs (comparisons made between winkelmayer et al), are there any potential reasons for the finding that there were no relationship between use of calcium based phosphate binders and mortality? In general most renal physicians will endeavour to avoid hypercalcemia among pre-dialysis patients and dialysis patients on CCPB. Given that the calcium balance play a role in vascular calcification, could the titration of CCPB based on calcium-phosphate levels and the usage of non-calcium based phosphate binders play a role? This dynamic titration of phosphate binder doses and granular information may not be adequately captured in large electronic database studies.

6) Another important limitation for this study would be that for changes in prescribed doses of CCPB and the assurance of its continued use throughout the study period could not be assessed (some patients may be potentially converted to non-calcium based phosphate binder during the study due to hypercalcemia possibly?).

Minor comments

1) Abstract (results): it was written that 8124 patients were included but only 879 CCPB users were matched with 3516 nonusers. I think the first sentence may not be required. Clarifications need to be stipulated what were the actual patient population included in the 8124 patients in the abstract (methodology section) as it is not intuitive.

2) Abstract (results): The results regarding the mortality risk between CCPB users vs non-users should be described as no significant differences instead of lower but non-statistically significant.

3) Introduction: The authors have described the relationship between hyperphosphatemia and adverse outcomes extensively. It is also worthwhile to perform a short discussion of the role of hypercalcemia (given the study’s focus on calcium based phosphate binder) e.g. prevalence and related outcomes to better substantiate the need for this study. (consider citing: Int Urol Nephrol. 2018 Oct;50(10):1871-1877; J Clin Endocrinol Metab. 2016 Jun; 101(6): 2440–2449.)

4) What were the types of calcium based phosphate binder included in this study as they carry inherently different amount of elemental calcium?

5) Do consider defining “daily defined doses” in the methodology to help readers to understand this concept and its inherent limitations

Reviewer #3: The paper on Effects of Calcium-Containing Phosphate Binders on Cardiovascular Events and

Mortality in Predialysis CKD Stage 5 Patients support other observations that calcium load increase vascular calcification and in fact is non-traditional risk factor of cardiovascular event and death.

The manuscript is properly designed, conducted and analysed. The only points are listed below:

- many CKD4/5 patients take vitamin D analogs or supplement what may increase calcium absorpition. What percentage of patients did it in both groups?

- other drugs like Vitamin K antagonist increase independetly vascular calcification. How many patients took it?

Such a data are necesary for interpretation and is lacking must be commented as limitation of the study

Reviewer #4: This interesting study by Chu et al uses the Taiwanese national health insurance database to examine the association between use of calcium-containing phosphate binders (CCPB) and all-cause mortality and cardiovascular outcomes for CKD G5 patients. A few comments:

Abstract

-The methods section could include more details about the analysis - What kind of models were used in the analysis. How was cardiovascular events defined?

Introduction

-The focus and study population of the study could be clearer. There seems to be some inconsistency between the Introduction and the abstract as to whether this current study focuses on CKD stage 5 patients with or without hyperphosphatemia.

-Similarly, other studies have explored the link between calcium carbonate and adverse outcomes for pre-dialysis CKD stage 3 and 4. Why would the results be different for CKD stage 5. Please comment on the justification for using a CKD stage 5 population.

Methods

-How representative is the sample used in the study?

-It is not clear if number of admissions or any admissions (binary measure) was used as the outcome in models for coronary heart disease (line 134)?

-Was mortality assessed from the data in the Taiwan NHIRD or was there any data linkage to other sources (e.g. death registry)? If not, please comment on the completeness and accuracy of this data.

- CCPB users and non-users could also differ on medications and blood pressure and lab data. Were data on any of these available/looked at and why was this not included in the propensity score?

-Did the authors test for violations of the proportionality assumptions of the Cox regression models?

Results

-The authors found that adverse effects of CCPBs increased with increasing doses and cumulative high doses was also linked to higher risk of coronary heart disease. Were changes in the use of CCPB over the duration of the study period accounted for or explored in any way?

Discussion

-Good consideration of the limitations of the study.

**********

6. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #2: No

Reviewer #3: No

Reviewer #4: Yes: Hilda Hounkpatin

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email PLOS at figures@plos.org. Please note that Supporting Information files do not need this step.

4 Sep 2020

Response to Reviewers:

We thank you for your time and effort to improve this study. The manuscript has been revised according to your valuable comments, and the changes in the revised manuscript of the marked-up version have been highlighted in red font. The itemized responses to four reviewers were listed as below.

Response to comments of Reviewer #1:

1. The authors did an excellent job analyzing the effect of calcium containing phosphate binders on cardiovascular mortality. However it is important that the authors identify with more detail the limitations of the study. The authors have mentioned the difficulty to separate Phosphate levels vs calcium load from the binder. There are other factors that should be commented: anticuoagulants-warfarin, treatment with active vitamin D, patients adherence to treatment not only relative to phosphtae control but also interdialytic weight gain, blood pressure medication, PTH levels , CRP levels and others.

Answer: Thank you very much for your comment. We have addressed other medication including warfarin or active vitamin D treatment as unmeasured confounders in the study limitations (line 307-308, page 18, marked-up version). Because of the limitation of NHIRD, the raw laboratory data were not included. We added CRP as an additional confounder (line 309, page 18, marked-up version). In this study, we enrolled predialysis CKD patients only, so we did not include interdialytic weight gain as a confounder.

Response to comments of Reviewer #2:

Major comments

1. It will be good for the authors to comment on their focus on Stage 5 predialysis patients and also why patients in stage 4 were not considered. [Noted this was only brought up in the discussion briefly but may be good to describe in the methodology]

Answer: Thank you very much for your comment. In this study, the identified information derived from the NHIRD included diagnostic codes according to the ICD9-CM; however, ICD-9-CM did not have a code define stage 1 to 4 CKD patients exactly. We enrolled the clinical condition of stage 5 CKD patients in the analyzed claims database by restricting patients to those receiving erythropoiesis-stimulating agents (ESAs), the medications that are reimbursed under the Taiwan universal health insurance program for patients whose serum creatinine concentrations are > 530 μmol/L. We described this section between line 105-110, page 5.

2. An important issue of note to evaluate the dose dependent relationship of calcium based phosphate binders is that the dose of elemental calcium within the CCB should be computed. This will facilitate more accurate and meaningful evaluation of this relationship as different CCPB have varying contents of elemental calcium. From the described methodology, it seems that it would be possible to compute the elemental calcium as the dose is available and the name of the CCPB.

Answer: We enrolled in calcium carbonate, calcium citrate, and calcium acetate in this study because these medications were provided by the national health insurance program in Taiwan. As the reviewer’s comment, different CCPB have varying contents of elemental calcium. In this study, we did not make a comparison among different CCPB directly; instead, we wanted to compare the effect of accumulated calcium content in CCPB on cardiovascular events. We used defined daily dose (DDD) as a technical unit of measurement defined as the assumed usual maintenance dose per day for a drug used for its main indication in adults. However, DDDs only give an estimate of consumption and not an exact picture of actual use. DDDs provide a fixed unit to assess trends in drug consumption and to perform comparisons between population groups.

3. How was the segregation of the subgroups of daily defined dose (≤15, 16–40, and >40 DDD) determined?

Answer: Thank you very much for your comment. DDD system allows standardization of drug groups and represents a stable drug utilization metric to enable comparisons of drug use and to examine trends in drug use over time. We decided to determine the subgroups of daily defined dose as tertile.

4. An important factor for consideration that would substantially strengthen the relationship between use of CCPB and adverse outcomes would be the use of non-calcium based phosphate binders e.g. lanthanum, and sevalamer as well as vitamin D analogues. It will be good to report these information among users of CCPB and non-users and adjust for theses confounders (especially non-calcium based phosphate binders) in the multivariate analyses.

Answer: Under the Taiwan universal health insurance program, lanthanum and sevelamer were not covered by the insurance program, and these drugs are purchased out of patients’ pockets. Moreover, the medication at one’s own expense could not be validated in NHIRD. However, sevelamer was approved in 2003 and lanthanum was approved in 2006. Both of them were not popular phosphate binders in Taiwan during this study period (January 1, 2000, and June 30, 2005), which might mitigate the confounder effect of non-calcium phosphate binders. We addressed these medications including non-CCPB, warfarin, vitamin D as unmeasured variables in the study limitation (line 307-308, page 18, marked-up version).

5. Discussion: beyond the limits of study designs (comparisons made between winkelmayer et al), are there any potential reasons for the finding that there were no relationship between use of calcium based phosphate binders and mortality? In general most renal physicians will endeavour to avoid hypercalcemia among pre-dialysis patients and dialysis patients on CCPB. Given that the calcium balance play a role in vascular calcification, could the titration of CCPB based on calcium-phosphate levels and the usage of non-calcium based phosphate binders play a role? This dynamic titration of phosphate binder doses and granular information may not be adequately captured in large electronic database studies.

Answer: Thank you very much for your comment. Mineral metabolisms including calcium and phosphorus homeostasis were the promoter of vascular calcification among the CKD population. It is hard to maintain homeostasis of calcium and phosphorus levels among the advanced CKD population. Most renal physicians will prescribe CCPB first in Taiwan if hyperphosphatemia develops because CCPB is provided by the national health insurance program. Besides, during this study period (January 1, 2000, and June 30, 2005), lanthanum and sevelamer were not popular phosphate binders in Taiwan. We admitted that the titration of CCPB and usage of non-CCPB could have made the magnitude of the association of all-cause mortality toward the null. However, there is still no well-designed placebo-controlled trial demonstrating whether vascular calcification could be prevented or reversed with therapies aimed at maintaining calcium and phosphorus homeostasis in predialysis CKD patients. We think that this observational study still provides real-world data about the effect of CCPB use among predialysis CKD patients.

6. Another important limitation for this study would be that for changes in prescribed doses of CCPB and the assurance of its continued use throughout the study period could not be assessed (some patients may be potentially converted to non-calcium based phosphate binder during the study due to hypercalcemia possibly?).

Answer: Thank you very much for your comment. We agree that the misclassification of drug exposure of interest is possible, which could have resulted in the observed null association between CCPB use and all-cause mortality. We added a section to describe this important limitation (line 299-302, page 17-18, marked-up version).

Minor comments

1. Abstract (results): it was written that 8124 patients were included but only 879 CCPB users were matched with 3516 nonusers. I think the first sentence may not be required. Clarifications need to be stipulated what were the actual patient population included in the 8124 patients in the abstract (methodology section) as it is not intuitive.

Answer: Thank you very much for your comment. We had deleted the first sentence in line 29, page 2.

2. Abstract (results): The results regarding the mortality risk between CCPB users vs non-users should be described as no significant differences instead of lower but non-statistically significant.

Answer: Thank you very much for your comment. We had corrected as reviewer’s suggestion (line 33-34, page 2, marked-up version).

3. Introduction: The authors have described the relationship between hyperphosphatemia and adverse outcomes extensively. It is also worthwhile to perform a short discussion of the role of hypercalcemia (given the study’s focus on calcium based phosphate binder) e.g. prevalence and related outcomes to better substantiate the need for this study. (consider citing: Int Urol Nephrol. 2018 Oct;50(10):1871-1877; J Clin Endocrinol Metab. 2016 Jun; 101(6): 2440–2449.)

Answer: Thanks for your suggestion and we added a section to discuss the prevalence and adverse effect of hypercalcemia in predialysis CKD patients. Please see line 63-67, page 4, marked-up version.

4. What were the types of calcium based phosphate binder included in this study as they carry inherently different amount of elemental calcium?

Answer: Thank you very much for your comment. We enrolled in calcium carbonate, calcium citrate, and calcium acetate in this study because these medications were provided by the national health insurance program in Taiwan. These medication usages could be tracked from the registry for drug prescriptions in NHIRD. Considering the different amounts of elemental calcium, we used DDD to estimate and perform comparisons between population groups.

5. Do consider defining “daily defined doses” in the methodology to help readers to understand this concept and its inherent limitations.

Answer: Thank you very much for your comment. We added two sentences to describe the definition and inherent limitation of DDD between lines 158-162, page 9-10, marked-up version.

Response to comments of Reviewer #3:

1. many CKD4/5 patients take vitamin D analogs or supplement what may increase calcium absorpition. What percentage of patients did it in both groups?

Answer: Thank you very much for your comment. Use of vitamin D analogs or supplement is not provided by the national health insurance program in Taiwan. We could not track these medication usage data from the registry for drug prescriptions in NHIRD because these medications were at one’s own expense. We addressed vitamin D as unmeasured variables and one of the limitations (line 307-308, page 18, marked-up version).

2. other drugs like Vitamin K antagonist increase independetly vascular calcification. How many patients took it?

Answer: Thank you very much for your comment. We addressed Vitamin K antagonist and other drugs as unmeasured variables and one of the limitations (line 307-308, page 18, marked-up version).

Response to comments of Reviewer #4:

1. The methods section could include more details about the analysis - What kind of models were used in the analysis. How was cardiovascular events defined?

Answer: Thanks for your suggestion. We added a sentence to describe the statistical analysis in more detail (line 26-27, page 2, marked-up version). We described the statistical models for the main analysis in line 150-153, statistical analysis section, page 9. Considering the brief overview of the investigation, we decided to put the definition of cardiovascular events in line 141-142, page 9, marked-up version.

2. The focus and study population of the study could be clearer. There seems to be some inconsistency between the Introduction and the abstract as to whether this current study focuses on CKD stage 5 patients with or without hyperphosphatemia.

Answer: Thank you very much for your comment. Detailed laboratory test results and medical notes were not included in NHIRD, and baseline phosphorus, calcium, CRP, and parathyroid hormone levels were uncertain. Considering this limitation, we defined the inclusion of stage 5 CKD patients in the analyzed claims database by restricting patients to those receiving erythropoiesis-stimulating agents (ESAs), the medications that are reimbursed under the Taiwan universal health insurance program for patients whose serum creatinine concentrations are > 530 μmol/L.

3. Similarly, other studies have explored the link between calcium carbonate and adverse outcomes for pre-dialysis CKD stage 3 and 4. Why would the results be different for CKD stage 5. Please comment on the justification for using a CKD stage 5 population.

Answer: The different results of calcium use and adverse outcomes might be due to the diverse stage of the CKD population. Here are the reasons for using CKD stage 5 patients as the study population. First, hyperphosphatemia develops gradually in CKD patients with estimated glomerular filtration < 20 ml/min/1.73m. Second, advanced CKD patients have worse mineral metabolisms including calcium and phosphorus homeostasis, which were the promoter of vascular calcification among the CKD population.

4. How representative is the sample used in the study?

Answer: National health insurance covers the medical needs of 99.19% of the 23 million individuals in the population of Taiwan and the NHIRD comprises the standard health care data submitted by medical institutions, which is ideal for nationwide population-based cohort study.

5. It is not clear if number of admissions or any admissions (binary measure) was used as the outcome in models for coronary heart disease (line 134)?

Answer: Thanks for your comment. We used Cox model and the outcome was time to first event. If patients had multiple episodes of coronary heart disease, we analyzed the first event. NHIRD is a valid resource for population research on cardiovascular diseases, as the validity of acute myocardial infarction (AMI) diagnosis coding in the NHIRD has been demonstrated (J Epidemiol. 2014;24(6):500-7.). As published previously (J Clin Oncol. 2017 Nov 10;35(32):3697-3705.; Int J Cardiol. 2013 Oct 3;168(3):2616-21.), we estimated the incident coronary heart disease (ICD-9 codes 410–414, excluding 412 and 414.1) by employing ICD-9 diagnosis codes in inpatient, outpatient, and emergency department records.

6. Was mortality assessed from the data in the Taiwan NHIRD or was there any data linkage to other sources (e.g. death registry)? If not, please comment on the completeness and accuracy of this data.

Answer: Thanks for your comment. We assessed the cause of death as the primary diagnosis of in-hospital death or the first-listed discharge diagnosis at the last hospitalization within the three months before death for out-of-hospital deaths.

7. CCPB users and non-users could also differ on medications and blood pressure and lab data. Were data on any of these available/looked at and why was this not included in the propensity score?

Answer: Thanks for your comment. One of the major limitations of NHIRD is that detailed laboratory test results and medical notes were not included, and whether baseline laboratory tests or other non-calcium content phosphate binders were comparable between the two groups remains uncertain. We addressed this limitation to line 305-310, page 18, marked-up version.

8. Did the authors test for violations of the proportionality assumptions of the Cox regression models?

Answer: Thank you very much for your comment. We used Omnibus Test for the Cox Model. Test event of coronary heart disease, -2 Log likelihood = 26,786.101，Omnibus test: P<0.001; test event of all-cause mortality, -2 Log likelihood = 6,195.208，Omnibus test: P<0.001

9. The authors found that adverse effects of CCPBs increased with increasing doses and cumulative high doses was also linked to higher risk of coronary heart disease. Were changes in the use of CCPB over the duration of the study period accounted for or explored in any way?

Answer: Thank you very much for your comment. Use and nonuse of CCPB were determined within 90 days after the index date (the first date of using ESA); however, the drug status was not ascertained during the follow-up period. We admitted that the misclassification of drug exposure of interest is possible, which could have resulted in the observed null association between CCPB use and all-cause mortality. We added a section to describe this important limitation (line 299-302, page 17-18, marked-up version).

10. Good consideration of the limitations of the study.

Answer: Thank you very much for your comment.

Submitted filename: Response to Reviewers.docx

Click here for additional data file.

16 Sep 2020

PONE-D-20-04452R1

Effects of Calcium-Containing Phosphate Binders on Cardiovascular Events and Mortality in Predialysis CKD Stage 5 Patients

PLOS ONE

Dear Dr. Chu,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

There are still some comments that require the authors' attention. Please respond to ALL of reviewer #2's comments in your revised manuscript.

Please submit your revised manuscript by Oct 31 2020 11:59PM. If you will need more time than this to complete your revisions, please reply to this message or contact the journal office at plosone@plos.org. When you're ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

Please include the following items when submitting your revised manuscript:

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.

A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.

An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter.

If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see: http://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols

We look forward to receiving your revised manuscript.

Kind regards,

Frank T. Spradley

Academic Editor

PLOS ONE

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. If the authors have adequately addressed your comments raised in a previous round of review and you feel that this manuscript is now acceptable for publication, you may indicate that here to bypass the “Comments to the Author” section, enter your conflict of interest statement in the “Confidential to Editor” section, and submit your "Accept" recommendation.

Reviewer #2: (No Response)

Reviewer #4: All comments have been addressed

**********

2. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #2: Yes

Reviewer #4: Yes

**********

3. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #2: Yes

Reviewer #4: Yes

**********

4. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #2: No

Reviewer #4: Yes

**********

5. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #2: Yes

Reviewer #4: Yes

**********

6. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #2: Dear editor,

The authors have performed significant efforts to revise the manuscript based on the 4 reviewers' comments. One issue remained however inadequately addressed.

1. The computation of daily elemental calcium received from phosphate binders would actually permit a more meaningful comparison of calcium exposure compared to ddd as it reflects a more accurate assessment of calcium exposure to patient and a more standardised unit of measure. There is probably no need to compare between different phosphate binders as highlighted by the authors if they were to compute these as the unit of measure would be standardised across patients.

The authors may wish to comment on the above.

Thank you

Reviewer #4: My earlier comments and queries have been addressed. The methodology is appropriate and sound and the paper is presented clearly.

**********

7. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #2: No

Reviewer #4: No

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email PLOS at figures@plos.org. Please note that Supporting Information files do not need this step.

12 Oct 2020

Response to Reviewer:

We thank you for your time and effort to improve this study. The manuscript has been revised according to your valuable comments, and the changes in the revised manuscript of the marked-up version have been highlighted in red font. The response to reviewer #2 was listed as below.

Response to comments of Reviewer #2:

1. The authors have performed significant efforts to revise the manuscript based on the 4 reviewers' comments. One issue remained however inadequately addressed. The computation of daily elemental calcium received from phosphate binders would actually permit a more meaningful comparison of calcium exposure compared to ddd as it reflects a more accurate assessment of calcium exposure to patient and a more standardised unit of measure. There is probably no need to compare between different phosphate binders as highlighted by the authors if they were to compute these as the unit of measure would be standardised across patients.

Answer: Thank you very much for your comment. We agreed with your comment that the computation of daily phosphate binders elemental calcium would permit more a meaningful comparison of calcium exposure compared to the defined daily dose because of a more accurate assessment of calcium exposure to the patient and a more standardized unit of measure. However, data in the National Health Insurance Research Database were de-identified by scrambling the identification codes of both patients and medical facilities, we could not track each patients’ accurate element calcium exposure or query the data at any level using this database. We added a section to describe this important limitation (line 301-303, page 18, marked-up version). To compare the effect of accumulated calcium content in CCPB on cardiovascular events, we used the defined daily dose (DDD) as a technical unit of measurement defined as the assumed usual maintenance dose per day for a drug used for its main indication in adults.

Submitted filename: Response to Reviewers.docx

Click here for additional data file.

15 Oct 2020

Effects of Calcium-Containing Phosphate Binders on Cardiovascular Events and Mortality in Predialysis CKD Stage 5 Patients

PONE-D-20-04452R2

Dear Dr. Chu,

We’re pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it meets all outstanding technical requirements.

Within one week, you’ll receive an e-mail detailing the required amendments. When these have been addressed, you’ll receive a formal acceptance letter and your manuscript will be scheduled for publication.

An invoice for payment will follow shortly after the formal acceptance. To ensure an efficient process, please log into Editorial Manager at http://www.editorialmanager.com/pone/, click the 'Update My Information' link at the top of the page, and double check that your user information is up-to-date. If you have any billing related questions, please contact our Author Billing department directly at authorbilling@plos.org.

If your institution or institutions have a press office, please notify them about your upcoming paper to help maximize its impact. If they’ll be preparing press materials, please inform our press team as soon as possible -- no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org.

Kind regards,

Frank T. Spradley

Academic Editor

PLOS ONE

20 Oct 2020

PONE-D-20-04452R2

Effects of Calcium-Containing Phosphate Binders on Cardiovascular Events and Mortality in Predialysis CKD Stage 5 Patients

Dear Dr. Chu:

I'm pleased to inform you that your manuscript has been deemed suitable for publication in PLOS ONE. Congratulations! Your manuscript is now with our production department.

If your institution or institutions have a press office, please let them know about your upcoming paper now to help maximize its impact. If they'll be preparing press materials, please inform our press team within the next 48 hours. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information please contact onepress@plos.org.

If we can help with anything else, please email us at plosone@plos.org.

Thank you for submitting your work to PLOS ONE and supporting open access.

Kind regards,

PLOS ONE Editorial Office Staff

on behalf of

Dr. Frank T. Spradley

Academic Editor

PLOS ONE