Introduction
The Korean Society of Nephrology (KSN) has managed a nationwide registry of end-stage renal disease (ESRD) patients since 1985 [1,2]. Leveraging this extensive dataset, our study aims to discern mortality trends spanning from 2001 to 2022. Through this investigation, we approach mortality from two distinct angles: an analysis of annual mortality rates and an examination of survival rates following the initiation of dialysis.
To further comprehend the impact of the global coronavirus disease 2019 (COVID-19) pandemic on mortality among dialysis patients, we investigated the fluctuations in short-term mortality rates following dialysis initiation during the pandemic period. Lastly, we sought to further examine which disease groups exerted the greatest influence on mortality among dialysis patients and whether these disease groups changed over time as time elapsed.
By scrutinizing these dimensions, we aim to gain comprehensive insights into the dynamics of mortality among ESRD patients over the specified period.
Methods
Data collection
The study included patients aged 19 years or older who initiated dialysis treatment between 2001 and 2022, with the event end date set as December 31, 2022. Data collection was conducted through the online dialysis patient registry program on the KSN website. Baseline characteristics were retrieved, including age, sex, comorbidities, and primary causes of renal failure (diabetes mellitus [DM], hypertension, glomerular disease, cystic kidney disease, unknown, and other causes). Dialysis-related data included information on dialysis modality, initiation timing, and maintenance duration. Patients receiving both hemodialysis (HD) and peritoneal dialysis (PD) were excluded from the outcome analysis.
Mortality analysis encompassed two primary dimensions: the annual mortality rate and survival outcomes over time post-dialysis commencement. To evaluate trends in annual all-cause mortality rates, patients were stratified into groups based on sex and dialysis modality, as well as into four age groups (19 to 44 years, 45 to 64 years, 65 to 74 years, and 75 years and older) for subsequent analysis [3]. Additionally, the impact of DM on mortality was assessed by categorizing patients into four groups based on DM presence and dialysis modality. Furthermore, the effect of hemodiafiltration (HDF) on mortality was examined by analyzing mortality rates among patients receiving HDF at least once a week.
To analyze survivability, cumulative survival rates at 3 months, 1 years, 2 years, 3 years, and 5 years post-dialysis initiation were examined. To assess the variation in survivability trends, the cumulative survival rates were analyzed using the 5 years mark post-dialysis initiation as the representative period. These results were collected separately for each dialysis modality, and data collection spanned from 2008 to 2017.
Additionally, to assess short-term survivability during the COVID-19 pandemic from 2020 to 2022, we analyzed “2-year mortality after dialysis initiation” for patients who started dialysis from 2017 to 2020. For the 2-year post-dialysis mortality analysis, patients were divided into four age groups based on the same criteria as for the overall mortality analysis.
Causes of death were compared from 2001 to 2022 and categorized into cardiac event, vascular event, infection, liver disease, social event, and other causes. Social event was defined as the patient refusing further treatment, suicide, or therapy ceased for other cause.
Given the large sample size, anonymity of the patients studied and non-intrusive nature of the research, the requirement for written consent was waived.
Statistical analysis
Continuous variables were presented as means ± standard deviations and were compared using the Student t test. Categorical variables were expressed as frequencies and proportions and were compared using the chi-square test. Annual all-cause mortality rates and 2-year mortality rates during the COVID-19 pandemic were presented as “deaths per 1,000 person-years.” Poisson regression was conducted to analyze trends in mortality. To evaluate the probability of 5-year survival after dialysis initiation, Cox regression analysis was performed.
Differences were considered statistically significant at p < 0.05. Regression analysis and model building were performed using STATA MP software (version 14.2; Stata Corp).
Sensitivity analysis
In cases where acute renal replacement therapy is needed, HD is preferred over PD. Therefore, to assess the potential impact of acute cases on outcomes, a sensitivity analysis was conducted. Acute cases were defined as those occurring within 90 days after dialysis initiation, and all patients who died during this period were excluded from the mortality analysis. These results were adjusted for age and sex.
Moreover, recognizing DM as a significant factor known to influence the prognosis of dialysis patients, we performed additional adjustments for DM status to further validate the results.
Results
Study population
The total data included 173,216 HD patients and 29,154 PD patients, excluding 16,957 patients with missing or erroneous data, 21,055 patients who started dialysis before 2001 or started dialysis after 2022 and were included in the collection data, and 242 patients under the age of 18. We also excluded 3,683 patients receiving both HD and PD, leaving a final sample of 160,433 patients for analysis (Fig. 1). Women comprised 39.9% (n = 64,007) of all patients, with a significantly higher proportion of women in the PD group (42.1%) than in the HD group (39.6%). Age averaged 68.22 years and was significantly lower in the PD group (63.30 years) than in the HD group (68.97 years). DM was the most common cause of ESRD in both HD and PD groups (Table 1).
Annual mortality rate
In the period since 2001, the mortality rate among dialysis patients has shown a consistent annual decline (Fig. 2A). Notably, the mortality rate among HD patients exhibited an increase from 46.2/1,000 person-years in 2019 to 49.7/1,000 person-years in 2020, followed by a further increase to 50.4/1,000 person-years in 2021, before returning to 34.9/1,000 person-years in 2022. PD patients demonstrated the lowest mortality rate of 33.4/1,000 person-years in 2021, which then rose to 36.6/1,000 person-years in 2022. Upon adjusting for age and sex, the mortality rate for all dialysis patients decreased from 74.2/1,000 person-years in 2001 to 56.2/1,000 person-years in 2012, further decreasing to 42.3/1,000 person-years by 2022 (Fig. 2B). Specifically, HD patients exhibited a gradual decline from 63.2/1,000 person-years in 2001 to 53.6/1,000 person-years in 2012, reaching 44.5/1,000 person-years in 2022. The decrease in mortality among PD patients was particularly pronounced, with rates dropping from 128.0/1,000 person-years in 2001 to 71.6/1,000 person-years in 2012, and further to 39.7/1,000 person-years by 2022.
Mortality rates were higher for males than females, which increased for both sexes in 2020 and 2021 compared to 2019, and then decreased again in 2022 (Supplementary Fig. 1, available online). When age-adjusted mortality rates for both sexes were adjusted for age, mortality rates were higher for men by about 4 to 5 per 1,000 person-years (Fig. 3). Although mortality rates trended downward in both sexes, the difference in mortality rates between the two groups did not change significantly from 2001 to 2022.
When patients undergoing dialysis were categorized into four age groups: 19 to 44 years, 45 to 64 years, 65 to 74 years, and 75 years and older; they constituted 7,595 (4.7%), 52,218 (32.5%), 42,673 (26.6%), and 57,950 (36.1%) of the total patients, respectively. Except for individuals aged 75 years and older, the mortality rate among dialysis patients in each group has shown a gradual decline over time (Fig. 4; Supplementary Fig. 2, available online). The pattern of increasing mortality with age remained consistent across different dialysis methods.
Forty-nine percent of the dialysis patients included in the analysis had DM as an underlying condition. Dialysis patients with DM exhibited a higher mortality rate compared to those without DM, although this difference tended to diminish over time (Fig. 5; Supplementary Fig. 3, available online). When stratified into four groups based on DM status and dialysis method, PD patients with DM consistently displayed the highest mortality rate since 2001, while PD patients without DM demonstrated the lowest mortality rate since 2013 (Fig. 6; Supplementary Fig. 4, available online). Until 2019, the mortality rate for HD patients without DM was lower than that for patients with DM. However, in 2020, the mortality rate for HD patients without DM (45.9/1,000 person-years) surpassed that of patients with DM (45.2/1,000 person-years).
Out of 139,324 patients undergoing HD, 13,855 (10%) received HDF at least once a week, and their mortality rates were compared with those of patients on HD alone (Supplementary Fig. 5, available online). Initially, the mortality rate for HD alone was observed to be higher than that for HD plus HDF. However, after adjusting for age and sex, the mortality rate for HD alone exhibited a decreasing trend, while the mortality rate for HD plus HDF showed a slight increase. It is important to note that our analysis did not account for factors such as the specific reasons for HDF, the duration of the procedure, and its frequency. Consequently, further research is warranted to elucidate the underlying causes contributing to the comparatively higher mortality rate observed in HDF patients.
Survival probability of incident dialysis
The cumulative survival rate at 3 months post-dialysis initiation for HD patients remained around 99% since 2008, with no significant changes (Table 2). Similarly, there were no substantial variations in the 1-year survival rate, although a slight decrease was observed starting from 2020, encompassing the pandemic period. Survival rates at 2 years, 3 years, and 5 years post-dialysis initiation exhibited a pattern of slight increase until 2014, followed by a subsequent decline. However, when adjusting the 5-year cumulative survival rate for age and sex, it was noted that the survival rate gradually improved over time.
For PD patients, both short-term and long-term cumulative survival rates improved post-dialysis initiation (Table 3). However, similar to HD, there was a slight decrease in survival rates during the period affected by the COVID-19 pandemic, particularly notable in the 5 years post-dialysis initiation mortality rate, which decreased significantly from 80.3% in 2016 to 73.8% in 2017. When adjusted for age and sex, it was observed that survival rates showed a greater improvement over time compared to HD, indicating a more substantial enhancement.
From 2008 to 2017, the 5-year survival rate for patients who started HD has been gradually increasing. Patients who started HD in 2008 had a 5-year survival rate of 76%, while patients who started HD in 2017, 10 years later, had a 5-year survival rate of 79% (Supplementary Fig. 6, available online). There has been a significant increase in the 5-year survival rate for patients initiated on PD between 2008 and 2017. Patients initiated on PD in 2008 had a 5-year survival rate of 68%, while patients initiated on PD in 2017 had a 5-year survival rate of 75%, representing a 5-year survival increase of approximately 7% over 10 years (Supplementary Fig. 7, available online).
Short-term mortality during COVID-19 pandemic
The 2-year mortality rate in HD patients, when adjusted for sex, showed a gradual increase for patients initiating dialysis in 2018, 2019, and 2020 compared to those initiating dialysis in 2017 across age groups of 19 to 44, 45 to 64, and over 75 years. The increase was minimal for individuals under 65 years old, but notably larger for those aged 75 years and above. Interestingly, for individuals aged 65 to 74 years, the mortality rate actually decreased since 2017 (Supplementary Fig. 8A, available online). In PD patients, when adjusted for sex, the 2-year mortality rates demonstrated a gradual increase for ages 19 to 44 and those over 75 years old, while showing a gradual decrease for ages 45 to 75 years (Supplementary Fig. 8B, available online).
Cause of death
Since 2001, the primary causes of death among all dialysis patients have remained relatively unchanged, primarily consisting of heart disease, infection, and vascular issues (Table 4). In 2022, heart disease accounted for 34.1% of deaths, infections for 25.8%, and vascular problems for 10.1%. Cardiac causes encompassed acute myocardial infarction and cardiac arrest, while infections included sepsis and pneumonia. Vascular causes comprised cerebrovascular disease, pulmonary embolism, and gastrointestinal bleeding.
The leading causes of death vary slightly between HD and PD patients (Supplementary Fig. 9, available online). Among HD patients, the top three causes of death continue to be heart disease, infection, and vascular issues, showing minimal change over the past two decades. In PD patients, deaths attributed to heart disease surpassed those from infections except in the early 2000s, but infections surpassed heart disease in 2018 and 2020. Deaths resulting from vascular problems were less frequent compared to HD patients.
Sensitivity analysis
Of the total 160,433 patients, 15,198 (9.5%) died within 90 days. Even when excluding acute deaths, mortality in PD patients decreased more rapidly than mortality in HD patients, and mortality in PD patients became lower than mortality in HD patients between 2019 and 2020 (Supplementary Fig. 10, available online). Annual mortality rates of overall patients dropped from 71.0/1,000 person-years in 2001 to 54.3/1,000 person-years in 2012, and further to 40.3/1,000 person-years by 2022.
When further adjusted for the prevalence of DM, there were no significant fluctuations observed in the annual mortality rate trend (Supplementary Fig. 11, available online). The overall annual mortality rate for all patients decreased from 72.1 in 2001 to 40.97 in 2022. Notably, the rate of decline in mortality was faster for PD patients, resulting in lower mortality rates for PD patients since 2020.
Discussion
All-cause mortality demonstrates a consistent downward trajectory, with intermittent increases noted in 2020 and 2021. These trends appear consistent across different dialysis modalities, although the annualized mortality rate for PD patients exhibits a more rapid decline, resulting in lower mortality rates for PD patients compared to those undergoing HD since 2019. The relatively lower mortality rate among PD patients aligns with recent findings from the United States [4]. However, it is important to note that while the annual mortality rate has remained largely unchanged in the United States, it is declining in Korea. Nonetheless, these findings should not be construed to imply that PD is inherently superior to HD. Potential selection bias stemming from patients ineligible for PD should be considered, necessitating further analysis in settings that control for these confounding factors [5].
The mortality rate among dialysis patients has been declining since 2001, yet it remains notably high among those aged 75 years and older. Although it is presumed that mortality rates may increase among dialysis patients over 75 years old, the impact of the elevated average age needs further confirmation. For instance, referring to the 2018 statistics in Japan, the proportion of patients aged 80 years and over, and even 90 years and over, among those aged 75 years and above is consistently growing [6]. From this perspective, additional discussion on further segmenting the age categories of the elderly population may be warranted in the future.
The 5-year survival rate has shown improvement each year from 2008 to 2017, with a more notable increase observed in PD patients, consistent with previous national reports [7,8]. When examining 2-year mortality rates in HD patients, those who initiated dialysis in 2018, 2019, and 2020 experienced a slight uptick in mortality compared to those who began in 2017. This increase was particularly notable in patients aged 75 years and older, potentially influenced by the pandemic. Further analysis will be necessary to ascertain whether the rises in mortality observed in 2020 and 2021 for HD patients and in 2022 for PD patients are associated with COVID-19.
The primary causes of death among dialysis patients were heart disease and infections, with approximately 20% of cases attributed to unknown causes. Over the past two decades, heart disease and infections have consistently accounted for more than 50% of deaths, while deaths from unknown causes have remained at around 20%. These rates are comparable to those observed in other countries [9]. Efforts to reduce mortality should prioritize specific classification and analysis of the causes of death. Several reports have suggested that patients undergoing renal replacement therapy have a higher mortality rate from COVID-19 infection compared to those who do not undergo such therapy [10]. However, it remains unclear whether this higher mortality reflects susceptibility to the disease itself, inadequate dialysis during the pandemic, or transmission of COVID-19 within dialysis units. Other studies have indicated that HD patients have higher rates of COVID-19 infection than PD patients, leading to a preference for PD [11].
There are a few caveats: First, although the 2-year mortality rate for patients who initiated dialysis in 2017 has been rising during the COVID-19 pandemic, it remains uncertain whether the pandemic directly contributed to this increase in mortality [12,13]. Second, the declining number of PD patients included in the statistics raises concerns about potential selection bias in mortality rates between HD and PD patients from the time of dialysis initiation. Third, while we were able to adjust for age, sex and DM we were unable to account for other variables.
For about 20 years, heart disease and infections have constituted more than 50% of deaths among dialysis patients, while deaths from unknown causes have consistently represented approximately 20%. This pattern is consistent with recent reports from the United States and Japan [3,6]. However, the proportion of deaths attributed to unknown causes was observed to be higher compared to other reports, indicating the need for a more detailed review of the causes of death.
In conclusion, while the mortality and survivability of dialysis patients have shown gradual improvement over time, temporary setbacks were observed during the COVID-19 pandemic. This improvement was notably more significant among diabetic or PD patients compared to HD patients without DM. Additionally, with an increasing number of patients receiving dialysis treatment at an older age, there has been a rise in the mortality rate among those aged over 75 years. Further research and interventions in this regard are warranted to ensure sustained progress in patient care and outcomes.
PDF Links
PubReader
ePub Link
XML Download
Full text via DOI
Download Citation
Supplement figure 1
Print
