;on behalf of the Korean Society of Geriatric NephrologyIntroduction
The global older population and proportion of older patients with end-stage kidney disease (ESKD) have been increasing. The prevalence of ESKD in the United States from 2000 to 2020 has been consistently higher in older adults than in young individuals and has increased by 54.6% over the past 10 years, particularly in those aged ≥75 years [1]. Similarly, the mean age of patients undergoing dialysis in Korea has gradually increased. The average age of patients undergoing dialysis in Korea in 2023 was 66.8 years, and the proportion of patients aged ≥65 years undergoing dialysis has been steadily increasing annually, with a prevalence of 57.2% in 2023 [2,3].
Osteoporosis and osteopenia are common in older patients and those undergoing dialysis. Data from the National Health and Nutrition Examination Survey, 2017 to 2018, showed that the age-adjusted prevalence of osteoporosis was higher in those aged 65 years and over (17.7%) than in those aged 50 to 64 years (8.4%). The age-adjusted prevalence of osteopenia was also higher in those aged 65 years and above (47.5%) than in those aged 50 to 64 years (39.3%) [4]. The prevalence of osteoporosis in chronic kidney disease (CKD) has been reported to range from 8% to 43%, while that of osteopenia ranges from 33.3% to 81%. The prevalence of osteopenia is higher in dialysis patients than in predialysis patients with CKD; however, osteoporosis is slightly lower in the dialysis group than in the predialysis CKD group [5,6].
In dialysis patients, osteoporosis and osteopenia are associated with high mortality, cardiovascular disease (CVD) risk, and fracture risk [7]. Medical costs related to osteoporosis pose a significant social burden [8]. Fractures are a significant contributor to mortality, and the resulting disability increases the risk of CVD, negatively impacts psychosocial well-being, and reduces quality of life [9–11].
Osteoporosis, osteopenia, and fractures are common and serious complications in older dialysis patients; however, studies on these issues in older patients newly starting dialysis remain limited. Therefore, we aimed to identify the factors associated with osteoporosis and osteopenia in this population using a multicenter retrospective cohort and to elucidate their clinical impact.
Methods
Study design and populations
The data analyzed in this study were obtained from the Korean Society of Geriatric Nephrology cohort, a multicenter retrospective cohort of incident hemodialysis patients aged ≥70 years, enrolled between January 2010 and December 2017. Participants with no T-score data (n = 2,055) were excluded, and 710 were enrolled (Fig. 1). Incident hemodialysis was defined as a patient diagnosed with end-stage renal disease and who started maintenance dialysis within 3 months. Clinical data were collected within 1 month before the first hemodialysis and included age, sex, body mass index (BMI), comorbidities (diabetes mellitus, hypertension, ischemic heart disease, congestive heart failure, peripheral artery disease, arrythmia, stroke, CVD, liver cirrhosis, rheumatic disease, malignancy, dementia), history of fracture, nursing home residents, medications (renin-angiotensin system [RAS] inhibitors, beta-blockers, diuretics, phosphate binder), laboratory findings, and bone mineral density (BMD). The CVD was defined as having any of the following comorbidities: ischemic heart disease, congestive heart failure, peripheral artery disease, arrhythmia, or stroke [12,13]. The lowest T-score from a BMD test obtained via dual-energy X-ray absorptiometry within 1 year prior to hemodialysis initiation was used. We identified the characteristics and clinical outcomes of incident hemodialysis patients based on their BMD. We divided the participants into three groups according to T-score criteria: osteoporosis, osteopenia, and normal BMD.
Primary outcome and definitions of variables
Osteoporosis was defined as a T-score of –2.50 or lower at the femoral neck, lumbar spine, or both. Osteopenia was defined as a T-score between –1.00 and –2.49 at the femoral neck, lumbar spine, or both [4,14].
The primary outcome was all-cause mortality. Mortality data were collected from the Korean National Statistical Office (Microdata Integrated Service, on-demand; https://mdis.kostat.go.kr).
Statistical analysis
A two-sample t test was conducted for continuous variables based on the homogeneity of variance, while a Pearson chi-square test was used for categorical variables.
To identify clinical factors associated with osteoporosis and osteopenia, we performed multivariable logistic regression analysis. Variables that showed a significant association (p < 0.10) in the univariable analysis (sex, history of fracture, nursing home residence, RAS inhibitors, and phosphate binders) or were of considerable significance (age, BMI, diabetes mellitus, and CVD) were entered together in the multivariable logistic regression models. In addition, we examined the collinearity between the variables by calculating the generalized variance inflation factor (GVIF). If the GVIF[1/(2 × degree of freedom)] was equal to or greater than 2, the variable was considered significant for collinearity.
To evaluate the impact of BMD categories and clinical variables on all-cause mortality, we employed Kaplan-Meier survival analysis and Cox proportional hazards models. Survival differences among the three BMD groups were assessed using the log-rank test. For Cox regression, we included variables with p-value <0.10 in the univariable analysis (e.g., BMI, hemoglobin, albumin, CVD, liver cirrhosis, fracture history, nursing home residence, RAS inhibitors, and beta-blockers), or were of considerable significance (age, sex, intact parathyroid hormone [i-PTH], diabetes mellitus) entered together. Statistical analyses were performed using IBM SPSS version 27.0 (IBM Corp.). Statistical significance was defined as a p < 0.05.
Ethical considerations
This study complied with the Declaration of Helsinki, and the Institutional Review Board of each center approved the study protocol (Supplementary Table 1, available online). The requirement for informed consent was waived due to the retrospective nature of the study.
Results
Baseline characteristics according to the bone mineral density group
Table 1 summarizes the baseline characteristics of the older patients undergoing incident hemodialysis, categorized by the BMD groups. At the time of hemodialysis initiation, the mean age of all participants was 78.4 ± 5.6 years; the mean BMI was 23.6 ± 4.4 kg/m2; and the average T-score was –1.7 ± 1.9. The proportion of males was 50.7%; 58.8% had diabetes mellitus; 86.2% had hypertension; 53.0% had CVD; 11.1% had a history of fracture; and 7.4% were nursing home residents, indicating that many participants had comorbidities. The proportions of the normal BMD, osteopenia, and osteoporosis groups were 41.4%, 19.6%, and 39.0%, respectively.
Compared with the normal BMD group at the time of hemodialysis initiation, the osteoporosis and osteopenia groups were significantly younger, had a higher proportion of males. The serum hemoglobin and i-PTH levels were the highest in the osteopenia group. In contrast, the incidences of peripheral artery disease and phosphate binder use were significantly lower in the osteoporosis and osteopenia groups than in the normal BMD group. Rheumatic disease and diuretic use were more common in the osteoporosis and osteopenia groups than in the normal BMD group. The history of fracture was the most frequent in the osteoporosis group.
In addition, we compared the baseline characteristics of the BMD measurement group (included group) and the non-measurement group (excluded group) (Supplementary Table 2, available online). The included group was older and had higher BMI and serum albumin levels than the excluded group, but the difference was minimal. However, the included group had a higher proportion of males, fracture history, and stroke than the excluded group, and the use of RAS inhibitors, beta-blockers, and phosphate binders was significantly higher. Although the proportions of hypertension, peripheral artery disease, and rheumatic disease were lower in the included group than in the excluded group, there were no statistically significant differences in other important underlying diseases such as diabetes mellitus, ischemic heart disease, congestive heart failure, CVD, liver cirrhosis, and malignancy between the two groups.
Factors associated with osteoporosis and osteopenia
We sought to identify the factors associated with osteoporosis (Table 2). In the logistic regression analyses, sex (odds ratio [OR], 4.70; 95% confidence interval [CI], 3.33–6.68; p < 0.001), history of fracture (OR, 1.86; 95% CI, 1.11–3.13; p = 0.02), and phosphate binder use (reference group: no phosphate binder; calcium-based phosphate binder: OR, 0.76 [95% CI, 0.53–1.09]; p = 0.14; non-calcium-based phosphate binder: OR, 0.35 [95% CI, 0.18–0.67]; p = 0.002; p for trend = 0.007) were significantly associated with osteoporosis even after adjustment of covariates. We examined GVIF to check for collinearity among covariates used in logistic regression analysis, and no significant collinearity was found among covariates (Supplementary Table 3, available online).
Bone mineral density groups and all-cause mortality
During the 36.8 ± 31.1 months of follow-up, 348 (58.8%) died, with the all-cause mortality rate being the highest in the osteoporosis group (79.8%) (osteopenia group, 77.2% and normal BMD group, 35.2%). We aimed to determine the differences in all-cause death by BMD group and the impact of osteoporosis and osteopenia on the risk of death. The Kaplan-Meier curve demonstrated that the normal BMD group had a significantly lower risk of overall death than in the osteoporosis and osteopenia groups (Fig. 2). In the Cox regression analyses (Table 3), osteoporosis (hazard ratio [HR], 1.53; 95% CI, 1.14–2.06; p = 0.005) was significantly associated with all-cause mortality even after adjustment of covariates, but osteopenia was not statistically significant.
Age (HR, 1.02; 95% CI, 1.02–1.07; p = 0.001), BMI (HR, 0.96; 95% CI, 0.93–0.99; p = 0.006), hemoglobin (HR, 1.12; 95% CI, 1.04–1.21; p = 0.004), albumin (HR, 0.69; 95% CI, 0.56–0.85; p = 0.001), diabetes mellitus (HR, 1.41; 95% CI, 1.08–1.83; p = 0.01), liver cirrhosis (HR, 2.52; 95% CI, 1.49–4.25; p = 0.001), nursing home residents (HR, 1.69; 95% CI, 1.09–2.62; p = 0.02), and RAS inhibitor use (HR, 0.61; 95% CI, 0.48–0.77; p < 0.001) were significantly associated with all-cause mortality in older patients undergoing incident hemodialysis (Table 3). In addition, there was no significant collinearity among covariates (Supplementary Table 4, available online).
Discussion
Using a multicenter retrospective cohort, we found that osteoporosis and osteopenia were particularly common in older Korean patients undergoing incident hemodialysis. Osteoporosis was significantly associated with female sex, history of fracture, and absence of phosphate binder use. Additionally, osteoporosis was significantly associated with all-cause mortality. These findings emphasize the need for monitoring and treating osteoporosis to mitigate mortality risk in older hemodialysis patients.
Osteoporosis and osteopenia are associated with an increased risk of death in patients undergoing dialysis. Unlike our study, Jaques et al. [15] conducted a single-center study on patients receiving prevalent dialysis, including those on hemodialysis, peritoneal dialysis, and kidney transplantation. Two other studies focused on patients initiating dialysis; however, both were single-center studies and included relatively young patients, with an average age in their 50s [16,17]. In contrast, our study enrolled older patients (aged ≥70 years) undergoing incident hemodialysis, used a large multicenter cohort, and adjusted for comorbidities, medications, and laboratory findings.
In previous studies, the prevalences of osteoporosis and osteopenia were 5% to 19.3% and 19.3% to 81%, respectively, depending on the BMD measurement site in dialysis patients [6,18]. Notably, osteoporosis prevalence was significantly higher in our study than in previous research (39.0% vs. 5%–19.3%). Unlike our study, previous studies targeted patients undergoing prevalent hemodialysis, had younger participants with an average age in the mid-to-late 50s, and had fewer than 100 participants.
In our study, osteoporosis in older patients undergoing incident hemodialysis was significantly associated with a higher risk of death than that in those with normal BMD. This may be attributed to the combined effects of low BMD and increased risk of fracture in patients with ESKD, as well as the presence of uremic toxins, fibroblast growth factor 23 (FGF23), decreased vitamin D, increased i-PTH, and bone-vascular axis changes associated with CKD. These factors contribute to vascular calcification, a higher risk of cardiovascular events, and overall mortality [7,19,20]. Additionally, our study identified a significant relationship between fracture history and osteoporosis, aligning with previous findings [15]. Given the association between fractures and mortality risk in ESKD, preventing fractures and managing osteoporosis should be prioritized in clinical practice. However, in our study, only 25.7% of the cohort had BMD data before starting dialysis. This suggests that BMD measurement is lacking in real practice, despite the KDIGO (Kidney Disease: Improving Global Outcomes) guidelines recommending a BMD test in patients with CKD G3a-G5D with evidence of CKD-mineral and bone disorder (CKD-MBD) or risk factor for osteoporosis to assess fracture risk. While osteoporosis is a term specific to bone and refers to a low BMD due to altered bone microstructure, CKD-MBD is a broader term that includes abnormalities in bone turnover, mineralization, volume, as well as vascular calcification, abnormalities in calcium, phosphorus, parathyroid hormone, and vitamin D, and is a systemic disorder of mineral and bone metabolism due to CKD [20,21]. Although our study mainly addressed osteoporosis in patients starting dialysis, strictly speaking, it would correspond to osteoporosis in CKD-MBD. Based on this study, it is suggested that BMD measurement should be more actively performed in real-world clinical settings for the prevention, diagnosis, and treatment of CKD-MBD in older dialysis patients.
Consistent with previous studies [22–25], we found that using phosphate binders, particularly non-calcium-based phosphorus binders, was associated with a lower risk of osteoporosis, highlighting the importance of serum phosphate regulation in patients undergoing hemodialysis. This may be because phosphate binders reduce the incidence of CKD-MBD, including changes in FGF23 levels, by reducing the incidence of hyperphosphatemia. Calcium-based phosphorus binders are associated with a higher risk of hypercalcemia and vascular calcification than with non-calcium-based binders, which may adversely affect BMD maintenance [23].
The incidence of osteoporosis is higher in females than in males [26]. Although the same T-score criteria are applied to both females and males, males generally have higher BMD. Females experience a more rapid decline in BMD due to postmenopausal hormonal changes [27]. In this study, the proportion of male patients was higher in the osteoporosis group, and the patients in the osteopenia and osteoporosis groups were significantly younger than those in the normal BMD group. These findings may seem counterintuitive. Previous studies have identified the risk factors that negatively affect bone health in dialysis patients, such as chronic inflammation, malnutrition, and prolonged corticosteroid use, beyond age and sex [28–30]. In this study, BMD testing may not have been consistently applied across all patients, with younger or male patients presenting such risk factors potentially being preferentially selected for assessment. Although the proportions of males in the osteoporosis and osteopenia group were higher than those in the normal BMD group at the initiation of hemodialysis, female patients on dialysis showed a significant association with increased osteoporosis risk, emphasizing that sex differences in bone metabolism remain an important factor even in patients with CKD.
This study has some limitations. First, as a retrospective study, it cannot establish a causal relationship between osteoporosis, osteopenia, and mortality. Additionally, we did not assess the cause of death or clinical outcomes beyond mortality. Osteoporosis is one of the factors significantly associated with overall mortality, and one of the hypotheses supporting it is osteoporosis–vascular calcification–cardiovascular death, but our data did not evaluate the cardiovascular outcomes. Second, while mobility, such as exercise capacity, can affect osteoporosis, it was not included in our analysis. Third, we did not evaluate the response of serial serum phosphate levels to phosphate binder treatment, nor did we investigate the duration or dose of the phosphate binders. Fourth, this study is based on data from patients who initiated dialysis between 2010 and 2017. Since then, clinical practice regarding osteoporosis screening, phosphate binder use, and treatment strategies has evolved considerably. These temporal changes in clinical practice may limit the generalizability of our findings and should be considered when interpreting the results in the context of more recent standards of care. To overcome these limitations, prospective studies are required to investigate the effects of osteoporosis and osteopenia on various important clinical outcomes, such as frailty, fracture, cognitive dysfunction, and CVD in older patients undergoing dialysis. In addition, it is necessary to demonstrate the relationship between osteoporosis and cardiovascular risk and cardiovascular death, and further study how clinical outcomes change with BMD and active osteoporosis treatment.
In conclusion, we demonstrated that osteoporosis and osteopenia are highly prevalent in older patients undergoing incident hemodialysis and that osteoporosis was significantly associated with mortality. These findings underscore the importance of monitoring and treating bone health to reduce the mortality risk in older patients undergoing hemodialysis.
PDF Links
PubReader
ePub Link
XML Download
Full text via DOI
Download Citation
Supplement table 1
Print
