Nutritional status and survival of maintenance hemodialysis patients receiving lanthanum carbonate Free

Abstract

INTRODUCTION

The mortality risk is unacceptably high in patients with end-stage renal disease on dialysis [1]. Hyperphosphatemia [2–4] and poor nutritional status [5–7] are among the major factors accounting for the high mortality in this population. Because dietary phosphate restriction is not sufficiently effective at controlling hyperphosphatemia and may lead to an inadequate protein intake [8, 9], the administration of oral phosphate binders is the cornerstone of therapy for hyperphosphatemia [10, 11]. Until recently, the predominant pharmacological therapies for hyperphosphatemia were calcium-based phosphate binders and sevelamer hydrochloride. However, the efficacy of these binders is often limited due to concerns regarding excessive calcium loading from calcium-based binders and gastrointestinal adverse effects of sevelamer. This situation poses additional challenges for clinicians, increasing the severity of dietary phosphate restriction, which can further exacerbate protein malnutrition and lead to an increased mortality risk [12].

Lanthanum carbonate is a nonaluminum, calcium-free phosphate binder that effectively lowers serum phosphorus levels [13, 14], even in patients with uncontrolled hyperphosphatemia with previous phosphorus-lowering therapy [15]. We previously conducted a 3-year historical cohort study in maintenance hemodialysis patients and found that the prescription of lanthanum was associated with improved survival, but the difference was significant only in the subgroup of patients with uncontrolled hyperphosphatemia and not in the overall population [16]. We recognized that our findings were limited by low statistical power, mainly due to the relatively short follow-up period following the initiation of lanthanum. Therefore we extended the follow-up period of the cohort to retest the hypothesis that the addition of lanthanum was associated with improved survival. We also explored the possible mechanisms through which lanthanum was associated with improved survival. We hypothesized that the survival benefit associated with lanthanum was mediated by improved control of hyperphosphatemia, a reduction in the calcium load and/or changes in the nutritional status.

MATERIALS AND METHODS

Study design

The design of the historical cohort study was described previously [16]. Briefly, the study population included 2292 patients from 22 facilities who were receiving maintenance hemodialysis for >3 months as of 31 December 2008. Demographic, weight and height, hemodialysis prescription, vascular access, comorbid conditions and censoring data were collected retrospectively via medical record abstraction. Laboratory results and records of drug administration were collected prospectively and entered uniformly into a central database. The serum calcium levels were corrected for the albumin concentration using Payne’s formula [17]. The parathyroid hormone (PTH) levels were measured using the Elecsys intact PTH assay (F. Hoffmann-La Roche, Basel, Switzerland) at 21 facilities and the whole PTH assay (Scantibodies Laboratories, Santee, CA, USA) at one facility. The whole PTH levels were converted to intact PTH levels with the following equation: intact PTH = whole PTH × 1.7 [18]. For the present analysis, we extended the follow-up period until 31 December 2015. This study was approved by the Institutional Review Board of Tokai University School of Medicine, which waived the need for written informed consent.

Outcomes and exposures

The primary outcome was 7-year all-cause mortality. We also examined cardiovascular mortality, defined as deaths caused by acute myocardial infarction, heart failure, cerebrovascular disease, cardiac arrhythmia, cardiac arrest, aortic disease and other cardiac disease. The primary exposure was the initiation of lanthanum. We compared the survival of patients who started lanthanum with the survival of patients who were not treated with lanthanum during the study period.

To minimize potential confounding and selection biases in this observational study, we performed a propensity score-matched analysis [19]. The propensity for lanthanum prescription was determined by logistic regression analysis using the following variables: age, gender, dialysis duration, primary cause of renal failure, mean blood pressure, body mass index (BMI), vascular access, coronary artery disease, stroke, peripheral artery disease, history of fracture, history of parathyroidectomy, dialysis adequacy (Kt/V), normalized protein catabolic rate (nPCR), albumin, hemoglobin, creatinine, calcium, phosphorus, intact PTH, alkaline phosphatase, total cholesterol, erythropoiesis-stimulating agent (ESA) use, calcium carbonate use, sevelamer hydrochloride use, vitamin D receptor activator (VDRA) use and cinacalcet hydrochloride use. To adjust for covariates measured at the time that the decision was made to initiate therapy, we estimated propensity scores for the lanthanum group using variables 3 months prior to the first lanthanum prescription. For the control group, we calculated propensity scores using variables at the time of cohort entry (i.e. 31 December 2008). We created propensity score–matched pairs of patients who were or were not treated with lanthanum with a caliper of 0.1 of the SD of the logit of the propensity score. For both groups, follow-up started at the time of propensity score calculation and ended at the time of death, loss to follow-up or on 31 December 2015, whichever came first. Because all patients in the lanthanum group survived at least 3 months of follow-up, we excluded those who were censored during the first 3 months in the control group to match the conditions between the two groups. In an effort to mimic an intention-to-treat analysis, patients who initiated lanthanum conservatively remained in the lanthanum group for all further analyses.

Statistical analysis

We used the χ² test, Student’s t-test and Wilcoxon rank-sum test to compare the baseline characteristics and laboratory results between the lanthanum and control groups. Longitudinal changes in continuous variables were examined using mixed-effects models. Nonnormally distributed variables were natural log transformed for the analysis.

We examined the risk of death associated with the serum phosphorus levels and nutritional indicators in the entire cohort using multivariate Cox regression. The association of serum phosphorus with mortality was examined in a model adjusted for case-mix covariates (age, gender, dialysis duration, primary cause of renal failure, mean blood pressure, vascular access, cardiovascular comorbidities, history of fracture and Kt/V) and nutritional indicators (BMI, nPCR, albumin, creatinine and total cholesterol). The association of each nutritional indicator with mortality was examined with adjustments for the case-mix covariates and phosphorus.

In the propensity score–matched analysis, we compared the mortality rates between the lanthanum group and the control group using the Kaplan–Meier method and univariate Cox regression, because the matching strategy eliminated all differences in baseline characteristics between the groups. We performed subgroup analyses stratified by age, gender, dialysis duration, primary cause of renal failure, BMI, cardiovascular comorbidities, albumin, calcium, phosphorus, intact PTH, calcium carbonate use, sevelamer hydrochloride use and VDRA use.

To explore whether the association between lanthanum and survival was mediated by changes in mineral metabolism and the nutritional status over time, we examined the change in the point estimate for lanthanum in models that adjusted for time-varying levels of mineral metabolism parameters (calcium, phosphorus and intact PTH) and nutritional parameters (BMI, nPCR, albumin, creatinine and total cholesterol).

For the survival analysis, we replaced missing data using multiple imputation with five imputed datasets. Statistical analyses were performed on each imputed dataset and were finally pooled to achieve single parameter estimates. P < 0.05 was considered statistically significant. All analyses were performed using SPSS Statistics 24 (IBM, Tokyo, Japan) and R 3.2.5 (R Foundation for Statistical Computing, Vienna, Austria).

RESULTS

Hyperphosphatemia, nutritional status and survival in the entire cohort

Among the 2292 patients in the entire cohort, the mean age was 65 years, 64% were male, 30% had serum phosphorus  >6.0 mg/dL and 87% were treated with phosphate binders (calcium carbonate and/or sevelamer hydrochloride). A detailed description of the baseline characteristics is provided in Supplementary data, Table S1. During a mean ± SD follow-up of 4.9 ± 2.3 years, 679 patients died (61.0/1000 person-years) in the entire cohort. After multivariable adjustment, higher serum phosphorus levels and lower nutritional indicators (BMI, albumin and creatinine) were each independently associated with an increased risk of death (Supplementary data, Table S2).

Lanthanum and other phosphate binder use in the entire cohort

After the market introduction of lanthanum in March 2009, the percentage of patients receiving this binder gradually increased to >40% by June 2015. A total of 924 patients received lanthanum at least once during the study period. In contrast, prescriptions of calcium carbonate and sevelamer hydrochloride gradually decreased during the study period (Figure 1). Among the patients who initiated lanthanum, the proportion of patients remaining on the drug decreased gradually, with a plateau of ∼70%. The initial daily median dose of lanthanum was 750 mg (10th–90th percentile 500–1500) and increased gradually to reach a plateau of 1500 mg (10th–90th percentile 750–2250).

FIGURE 1

Prescription (percentage of patients) of calcium carbonate, sevelamer hydrochloride and lanthanum carbonate in the overall cohort.

Open in new tabDownload slide

Lanthanum and survival in the propensity score–matched cohort

Of the 2292 patients in the entire cohort, we excluded 10 patients (0.4%) who died, 11 patients (0.5%) who transferred to other facilities during the first 3 months of follow-up and two patients (0.1%) with missing data on lanthanum prescription. After the exclusions, 2269 patients were available for the analysis. A comparison of the baseline characteristics between the patients who subsequently received lanthanum versus those who did not, is shown in Table 1. The patients who initiated lanthanum were younger; had a longer dialysis duration; were less likely to have diabetes as a cause of renal failure; had a lower prevalence of stroke; were more likely to have a history of parathyroidectomy; had higher BMIs; had higher serum creatinine, calcium and phosphorus levels; had higher intact PTH levels and were more likely to have a fistula for vascular access and receive sevelamer hydrochloride and cinacalcet hydrochloride. These differences in the baseline characteristics between the groups were eliminated in the propensity score–matched cohort.

The Kaplan–Meier survival curves in the propensity score–matched cohort are shown in Figure 2. The lanthanum group had a significantly lower risk of mortality than the control group {54.3 versus 70.5 deaths/1000 patient-years; hazard ratio [HR] 0.77 [95% confidence interval (CI) 0.61–0.97]}. The results were qualitatively unchanged after multivariable adjustment for the baseline characteristics [HR 0.74 (95% CI 0.56–0.98)]. In the stratified analyses, patients in the lanthanum group had a significantly decreased risk of mortality in many but not all strata, whereas in no stratum was favored for lack of lanthanum (Figure 3). Other propensity score–based approaches and conventional multivariable Cox regression yielded similar effects of lanthanum on survival (Supplementary data, Table S3). Similar findings were noted when the analysis was restricted to cardiovascular mortality [HR 0.49 (95% CI 0.29–0.82)].

FIGURE 2

Kaplan–Meier analysis of survival comparing the lanthanum group and the propensity score–matched controls.

Open in new tabDownload slide

FIGURE 3

Stratified HRs (95% CIs) for all-cause mortality between the lanthanum group and the propensity score–matched controls. DM, diabetes mellitus.

Open in new tabDownload slide

The median time from cohort entry to lanthanum initiation was 24 months (10th–90th percentile 6–60), thus the survival follow-up for the lanthanum group started at a median of 21 months (10th–90th percentile 3–57) after cohort entry. To account for the difference in the starting point of survival follow-up between the two groups, we performed an additional sensitivity analysis in which follow-up for the control group started 21 months after cohort entry, with propensity scores calculated at this time point. In this analysis, the survival benefit associated with lanthanum remained materially unchanged [HR 0.67 (95% CI 0.51–0.88)].

To further explore the potential mediators of the association between lanthanum and survival, we examined longitudinal trends in indices of mineral metabolism and nutritional status in the propensity score–matched cohort (Table 2). The serum phosphorus levels tended to decrease comparably in both groups, with no significant between-group differences. The serum calcium levels tended to decrease and the intact PTH levels tended to increase in the lanthanum group, whereas no consistent changes were observed in these values over time in the control group. Nutritional indicators, including nPCR, creatinine and total cholesterol, tended to decrease in the control group, whereas these changes were less pronounced in the lanthanum group. When the survival analysis was adjusted for time-varying phosphorus or other mineral metabolism parameters, the benefit associated with lanthanum was qualitatively unchanged (Table 3). However, the survival benefit was substantially attenuated and no longer significant when adjusted for time-varying indicators of nutritional status.

DISCUSSION

In this 7-year historical cohort study of maintenance hemodialysis patients, treatment with lanthanum was associated with a significant survival advantage compared with treatment without lanthanum. The benefit of lanthanum was independent of established risk factors and robust to different analysis strategies, including propensity score methods. Interestingly, patients treated with lanthanum maintained a better nutritional status during the follow-up period than the propensity score-matched controls, whereas the serum phosphorus levels tended to decrease comparably in both groups. Furthermore, adjustments for time-varying indicators of nutritional status but not time-varying phosphorus attenuated the association of lanthanum with survival. These data collectively suggest a potential benefit of lanthanum for maintenance of nutritional status and provide a new paradigm for the treatment of hyperphosphatemia in patients on hemodialysis, which is a population with a high rate of malnutrition.

We previously analyzed the same cohort used in the present study to explore the association of lanthanum with survival over 3 years of observation [16]. In that study, we found a large difference in survival between patients treated with lanthanum and those not treated with lanthanum, but the difference was not significant [HR 0.71 (95% CI 0.47–1.09)]. The major limitation of that study was the relatively short follow-up period in the lanthanum group. In the years after the market introduction of lanthanum in Japan, there have been substantial differences between clinicians in their tendencies to use this phosphate binder, partially due to the absence of long-term safety data. Although this situation was advantageous for performing propensity score matching, the slow growth of the use of lanthanum resulted in the short follow-up period following the initiation of lanthanum use, leading to insufficient statistical power. In the present study, we overcame this limitation by extending the follow-up period and demonstrated a significant survival benefit associated with lanthanum prescription in the overall cohort.

Our findings on the survival benefit associated with lanthanum initiation are consistent with recent studies showing a survival benefit of treatment with phosphate binders versus no treatment [20–22] and a benefit of sevelamer initiation among patients treated with calcium-based binders [23], thus providing a more compelling rationale for controlling serum phosphorus levels with phosphate binders. However, equally or even more important is the finding that patients who were treated with lanthanum maintained a better nutritional status than the propensity score–matched controls. This finding is particularly significant because the baseline nutritional status was comparable between the propensity score–matched groups, thus decreasing the possibility that lanthanum-treated patients were preselected based on their better nutritional status. Although it is well accepted that aggressive dietary phosphate restriction may contribute to protein energy malnutrition [9], patients are often advised to increase the severity of dietary phosphate restriction with the use of phosphate binders when they have uncontrolled hyperphosphatemia. Limiting phosphate-containing food additives may result in less nutritional impairment [8], but achieving control of serum phosphorus while simultaneously maintaining an appropriate nutritional status is still difficult in current practice, as highlighted by the present results showing a progressive decline in nutritional indicators in the control group. Therefore, our findings that patients who initiated lanthanum maintained a better nutritional status have important clinical implications. Our results suggest that increasing the use of phosphate binders may lead to relaxation of dietary phosphate restriction, increased protein intake and improved nutritional status. This possibility was also suggested by the Dialysis Outcomes and Practice Patterns Study (DOPPS), which showed higher levels of nutritional status in facilities prescribing phosphate binders to a larger proportion of patients [21]. In clinical practice, patients occasionally experience insufficient reductions in serum phosphorus after increasing the dose of phosphate binders, presumably due to a more liberal dietary phosphate intake. Our results suggest that this situation would be acceptable given that the improved nutritional status may lead to improved survival. Future research should focus on whether treatment with phosphate binders improves the nutritional status and, if so, whether the improved nutritional status translates into better survival.

There are several possible mechanisms for the association of lanthanum initiation with improved survival. Our original hypothesis was that improved control of serum phosphorus by initiating lanthanum would attenuate phosphate toxicity and thereby improve cardiovascular outcomes. However, the survival benefit associated with lanthanum was not attenuated by adjustments for time-varying phosphorus. A similar finding was also reported by Isakova et al. [20]. One possible explanation for our findings is the comparable reductions in serum phosphorus between the groups, which are presumably due to liberal dietary phosphate intake in the lanthanum group and aggressive dietary phosphate restriction in the control group. In contrast, patients who were prescribed lanthanum maintained a better nutritional status and the survival benefit of lanthanum was attenuated by adjustments in time-varying indicators of the nutritional status, suggesting that the benefit of lanthanum might be partially mediated by a better nutritional status after the initiation of lanthanum. However, it should be recognized that simply adjusting for potential mediators in standard regression models could introduce new sources of bias [24]. Further work is required to fully elucidate the mechanisms linking lanthanum prescription and improved survival. It is also important to note that more aggressive dose escalation of phosphate binders than current clinical practice would be required to sufficiently decrease the serum phosphorus levels in cases with increased phosphate intake. Whether such aggressive use of phosphate binders leads to improved survival also requires further study.

This study has several limitations. First and most importantly, we did not have information on the dietary protein intake. Although the nPCR is widely interpreted as a measure of dietary protein intake, we observed no significant relationship between the nPCR and mortality, which agreed with the findings from the DOPPS [5]. These findings raise concerns about the assumptions behind the use of the nPCR as a surrogate of dietary protein intake. We also lacked information on prescribed phosphate restriction and the measured phosphate intake. The impact of aggressive phosphate binder treatment on dietary phosphate restriction, dietary protein intake and nutritional status should be addressed with dedicated studies. Second, during the propensity score–matching process, patients who were treated with lanthanum for severe hyperphosphatemia but could not be matched with controls were excluded from the analysis. This process was useful for balancing the baseline characteristics between the groups but could have attenuated the effects of lanthanum on serum phosphorus and survival, which might explain why adjustment for time-dependent phosphorus did not attenuate the association between lanthanum and survival. Third, we did not have a measure of adherence to phosphate binders; however, we do not believe that this affected our results, because nonadherence to lanthanum would tend to bias the survival effect of lanthanum toward the null hypothesis. Fourth, the study population was restricted to Japanese hemodialysis patients, which may limit the generalizability of our results to other geographic regions. Finally, as with all observational studies, our results should not be interpreted as causal. We cannot exclude the possibility that a residual confounding or selection bias explained our findings.

In conclusion, the present study demonstrates that the initiation of lanthanum is associated with a better nutritional status and improved survival. Our results support the importance of maintaining an optimal nutritional status during treatment for hyperphosphatemia and suggest the value of lanthanum in this context. Future studies should focus on whether aggressive treatment with phosphate binders improves nutritional status and survival.

ACKNOWLEDGEMENTS

The authors thank the following investigators who participated in the study: Hajime Suzuki (Bousei Hiratsuka Clinic), Mitsunori Yagame (Bousei Oone Clinic), Kayoko Watanabe (Bousei Fujisawa Clinic), Nobuyoshi Takagi (Bousei Kannai Clinic), Hiroshi Kida (Bousei Akabane Clinic), Mitsumine Fukui (Bousei Tanashi Clinic), Ken-ichi Oguchi (Bousei Hospital), Tetsuo Shirai (Bousei Clinic), Mikako Nagaoka (Honatsugi Medical Clinic), Tsuneyoshi Oh (Tsurumi Nishiguchi Hospital), Eiji Nakano (Motomachi Medical Clinic), Takayuki Hashiguchi (Fujisawa Medical Clinic), Hirofumi Ishii (Shonan Seiwa Clinic), Yoshihide Tanaka (Higasiyamato Nangai Clinic), Yasuji Sugano (Kitahachioji Clinic), Toru Furuya (Higashikurume Clinic), Naoto Ishida (Seichi Clinic), Hiroyuki Ogura (Hadano Minamiguchi Clinic), Yoko Omori (Kitasenju Higashiguchi Jin Clinic), Miho Enomoto (Ayase Ekimae Jin Clinic) and Yuichiro Yamaguchi (Adachi Iriya Toneri Clinic).

FUNDING

This work was funded by grants from the Japan Society for the Promotion of Science and Bayer Yakuhin.

AUTHORS’ CONTRIBUTIONS

H.K. and M.F. designed the study with contributions from T.K., T.W., M.H. and T.S. H.K. analyzed data and wrote the draft of the manuscript. All authors contributed to interpretation of the data and critically revised the final draft of the manuscript.

CONFLICT OF INTEREST STATEMENT

H.K. has received honoraria, consulting fees and/or grant/research support from Bayer Yakuhin, Chugai Pharmaceutical, Kyowa Hakko Kirin and Ono Pharmaceutical. T.K. has received honoraria from Bayer Yakuhin, Chugai Pharmaceutical, Kyowa Hakko Kirin, Ono Pharmaceutical and Torii Pharmaceutical. T.W. has received honoraria, consulting fees and/or grant/research support from Chugai Pharmaceutical, Daiichi Sankyo, Kowa Pharmaceutical, Kyowa Hakko Kirin, Novartis, Ono Pharmaceutical, Otsuka Pharmaceutical, Sanofi and Takeda Pharmaceutical. M.F. has received honoraria, consulting fees and/or grant/research support from Astellas Pharma, Bayer Yakuhin, EA Pharma, Kyowa Hakko Kirin, Ono Pharmaceutical and Torii Pharmaceutical. The remaining authors have no conflicts to report.

(See related article by Avesani et al. Liberalizing the diet of patients undergoing dialysis: are we ready? Nephrol Dial Transplant 2019; 34: 180--183)

REFERENCES