Age is not a limiting factor for preemptive arteriovenous access creation before hemodialysis to improve patient survival

Article information

Korean J Nephrol. 2026;.j.krcp.25.342

Publication date (electronic) : 2026 June 2

doi : https://doi.org/10.23876/j.krcp.25.342

Chang Min Park¹, Jooun Park², Do Hyoung Kim³, Hyangkyoung Kim⁴, Young Soo Kim¹, Hyung Seok Lee⁵, Sang Jun Park⁶, Min-Ho Kim^7,*, Hoon Suk Park^2,*

¹Division of Nephrology, Department of Internal Medicine, Uijeongbu St. Mary’s Hospital, College of Medicine, The Catholic University of Korea, Uijeongbu, Republic of Korea

²Division of Nephrology, Department of Internal Medicine, Seoul St. Mary’s Hospital, College of Medicine, The Catholic University of Korea, Seoul, Republic of Korea

³Division of Nephrology, Department of Internal Medicine, Kangnam Sacred Heart Hospital, Hallym University College of Medicine, Seoul, Republic of Korea

⁴Department of Surgery, Ewha Womans University Mokdong Hospital, Seoul, Republic of Korea

⁵Division of Nephrology, Department of Internal Medicine, Hallym University Sacred Heart Hospital, Anyang, Republic of Korea

⁶Department of Surgery, Ulsan University Hospital, University of Ulsan College of Medicine, Ulsan, Republic of Korea

⁷Ewha Medical Data Organization, Ewha Womans University Seoul Hospital, Seoul, Republic of Korea

Min-Ho Kim Ewha Medical Data Organization, Ewha Womans University Seoul Hospital, Seoul, Republic of Korea 260 Gonghang-daero, Gangseo-gu, Seoul 07804, Republic of Korea Email: mino-kim@naver.com

Hoon Suk Park Division of Nephrology, Department of Internal Medicine, Seoul St. Mary’s Hospital, College of Medicine, The Catholic University of Korea, 222 Banpodaero, Seocho-gu, Seoul 06591, Republic of Korea. E-mail: cttailor@catholic.ac.kr

*Min-Ho Kim and Hoon Suk Park contributed equally to this study as co-corresponding authors.This study was presented in part as an oral presentation at the 2025 Korean Society of Nephrology (KSN) Annual Meeting in Seoul, Korea.

Received 2025 September 26; Revised 2026 January 16; Accepted 2026 February 20.

Abstract

Introduction

Chronic kidney disease (CKD) is a growing global health concern, and the number of patients undergoing dialysis is rapidly increasing. In many countries, including South Korea, hemodialysis is the most common dialysis modality for patients with end-stage renal disease (ESRD). The prevalence of patients undergoing hemodialysis continues to rise, with a corresponding increase in their mean age [1–3]. Therefore, determining the optimal vascular access type and timing of arteriovenous (AV) creation in patients undergoing hemodialysis is becoming challenging and important.

Among AV access types, an AV fistula (AVF) has a high patency rate and a low risk of infection; therefore, the “fistula first” strategy has generally been recommended for patients undergoing hemodialysis [4,5]. In contrast, a central venous catheter (CVC) allows immediate use after placement but is associated with a high risk of infection and increased mortality [6–9]. Recently, the “catheter last” strategy has also been emphasized for patients undergoing hemodialysis [10,11]. Thus, starting hemodialysis with an AVF or AV graft (AVG) in place (initial AVF or AVG) is thought to reduce catheter dependency and improve survival outcomes. However, it remains unclear whether the use of initial AV access in older and very old adult patients, such as octogenarians and nonagenarians, provides a clinically meaningful survival benefit.

We conducted a comprehensive observational study using the Korean National Health Insurance Service (NHIS) database to evaluate whether initiating hemodialysis with AV access in place provides a survival benefit in older and very old adult patients and to determine whether age should be a limiting factor when deciding to create AV access in patients before dialysis.

Methods

This study used data from the Korean NHIS on patients undergoing maintenance hemodialysis between January 2012 and December 2021. Data extraction was based on the diagnostic and procedure codes from the Korean Classification of Diseases (6th edition) and the International Classification of Diseases, 10th revision (ICD-10). Data collection and publication were approved by the Institutional Review Board (IRB) of the Ewha Womans University Medical Center, Republic of Korea (IRB no. EUMC 2023-03-040). The IRB waived the requirement for written informed consent because of the retrospective nature of the study.

Eligible patients were adults (aged ≥18 years) who received maintenance hemodialysis, identified by either a specific ESRD code (V001) or a CKD diagnostic code (N18 or N19). Patients were included if they had procedure codes for hemodialysis (O7020, O7021, and O9991) recorded at least once per week for more than three months. The index date was defined as the earliest date on which a patient underwent one of the following procedures to start hemodialysis: AVF creation (codes: O2011, O2012, and O2081), AVG creation (code: O2082), or hemodialysis catheter insertion (codes: O7011, O7012, O7013, O7014, O7015, O7016, O7017, and O7018). We classified patients according to whether they underwent AVF or AVG creation, or hemodialysis catheter insertion on the index date.

Patients with an AVF or AVG creation code on the index date and no subsequent catheter insertion were considered to have undergone hemodialysis with AV access and were classified into group AV (AVF and AVG). Patients with a catheter insertion code on the index date were considered to have started hemodialysis with a CVC in place. The patients who subsequently underwent AVF or AVG creation were classified into the CVC-to-AV group. Patients who continued hemodialysis with CVCs only and those without AV access creation were excluded. In addition, patients with a history of hemodialysis catheter insertion before the index date were excluded. Patients who had undergone peritoneal dialysis or kidney transplantation were also excluded (Fig. 1).

Figure 1.

Selection of patients.

AVG, arteriovenous graft; AVF, arteriovenous fistula; ESRD, end-stage renal disease; HD, hemodialysis; KT, kidney transplantation; PD, peritoneal dialysis.

We evaluated the all-cause mortality based on the type of initial vascular access during the follow-up period (group AV vs. group CVC-to-AV). Both groups were followed from the initiation of maintenance hemodialysis until death. The incident date of maintenance hemodialysis was defined as the first date on which the patient received four or more hemodialysis codes within a consecutive 4-day period. Patient mortality was identified using the assumed date of death (variable: DTH_ASSMD_DT) from the NHIS database. We also compared the survival outcomes between the AVF and CVC-to-AVF groups, as well as between the AVG and CVC-to-AVG groups. Additionally, we conducted subgroup analyses stratified by age, sex, and underlying diseases. Age was categorized into three subgroups focusing on older patients: 70–79, 80–89, and ≥90 years. Comorbidities included diabetes mellitus (DM), hypertension (HTN), coronary artery disease (CAD), cerebrovascular accident (CVA), and peripheral vascular disease (PVD), identified by ICD-10 codes from the NHIS database. Each comorbidity was defined based on the presence of the corresponding ICD-10 code recorded before the index date (Supplementary Table 1, available online).

The baseline characteristics of patients are presented as frequencies and percentages for categorical variables and as means and standard deviations for continuous variables. Comparisons of baseline characteristics between groups were performed using the chi-square tests for categorical variables and the Student t tests or one-way analysis of variance for continuous variables. Patient survival was analyzed using the Kaplan-Meier method, and groups were compared using the log-rank test. Multivariate analyses of the clinical variables associated with patient survival outcomes were performed using Cox proportional hazards models. Findings with a p-value of <0.05 were considered statistically significant. Statistical analyses were performed using IBM SPSS version 21.0 (IBM Corp.).

Results

Comparison of patient survival based on initial vascular access

A total of 68,393 patients underwent hemodialysis during the study period. Among the total cohort of patients, 29,381 (43.0%) started hemodialysis with an AVF or AVG (AV group), while 39,012 (57.0%) started hemodialysis with a CVC (CVC-to-AV group). The baseline characteristics of each group are presented in Table 1. The mean patient age was 65.20 ± 13.33 years, with males accounting for 61.5% of the study population. When comparing the AV group and CVC-to-AV group, the mean age in the CVC-to-AV group was slightly higher (65.85 ± 13.48 years vs. 64.33 ± 13.06 years, p < 0.001). Hypertension was more prevalent in the AV group, whereas DM and vascular diseases, including CAD, CVAs, and PVD, were more common in the CVC-to-AV group.

Table 1.

Baseline characteristics by initial vascular access type

During an average follow-up period of 3.50 ± 2.25 years, 25,332 patients (37.0%) died (Table 2). Kaplan-Meier survival analysis comparing the survival rates of the AV group and the CVC-to-AV group showed that the AV group had a significantly higher survival rate than the CVC-to-AV group (log-rank p < 0.001) (Fig. 2).

Table 2.

Follow-up duration and mortality outcomes

Figure 2.

Kaplan-Meier survival curves by initial vascular access type.

AV, arteriovenous access; CVC, central venous catheter.

In the Cox proportional hazards analysis, the CVC-to-AV group had a significantly higher mortality rate than the initial AV group (hazard ratio [HR], 1.388; 95% confidence interval [CI], 1.353–1.424; p < 0.001). Older age, DM, CAD, CVA, and PVD were significantly associated with increased mortality (p < 0.001). Female sex (HR, 0.808; 95% CI, 0.788–0.829; p < 0.001) and HTN (HR, 0.759; 95% CI, 0.703–0.820; p < 0.001) were associated with a reduced risk of mortality (Table 3).

Table 3.

Risk factors associated with mortality in the overall study population

Comparison of patient survival by initial vascular access in older age groups

We compared the survival rates and HRs between the AV and CVC-to-AV groups in older patients aged ≥70 years (Table 4). Among patients aged 70–79 and 80–89 years, the CVC-to-AV group was associated with higher mortality compared with the AV group (p < 0.001). However, among patients aged ≥90 years, there was no statistically significant difference in mortality between the two groups (p = 0.376). Kaplan-Meier survival curves for each age group are presented in Fig. 3. Consistent with these results, among patients aged 70–79 and 80–89 years, the AV group had a significantly higher survival rate than the CVC-to-AV group (log-rank p < 0.01), whereas no significant difference was observed among patients aged ≥90 years (log-rank p = 0.483).

Table 4.

Cox regression for patient survival across older age groups

Figure 3.

Kaplan-Meier survival curves by age group: initial AV vs. CVC-to-AV.

(A) 70–79 years; (B) 80–89 years; (C) ≥90 years.

AV, arteriovenous access; CVC, central venous catheter.

For patients with permanent AVFs, Table 5 presents the HRs comparing survival outcomes between the initial AVF and CVC-to-AVF groups in older age groups. Initial AVFs were significantly associated with better survival in patients younger than 90 years (p < 0.001); however, this benefit was not observed in patients aged ≥90 years (p = 0.648). The Kaplan-Meier curves showed the same pattern (Supplementary Fig. 2, available online). Similarly, Table 6 presents the HRs comparing the initial AVG group with the CVC-to-AVG group. Initial AVGs were consistently associated with better survival across the older age groups, including patients aged ≥90 years (p = 0.024). Kaplan-Meier survival analyses are shown in Supplementary Fig. 3 (available online). Similar patterns were observed in the 70–79 and 80–89 years groups, but no statistically significant difference was found among patients aged ≥90 years (log-rank p = 0.123).

Table 5.

Cox regression for patient survival by age group: CVC-to-AVF vs. AVF (reference)

Table 6.

Cox regression for patient survival by age group: CVC-to-AVG vs. AVG (reference)

When patients were stratified by sex and DM status, initial AVF was associated with a statistically significant survival advantage over CVC-to-AVF in patients aged ≤80 years, regardless of sex or DM status (p < 0.001). Among octogenarians (aged 80–89 years), this benefit was observed only in patients with DM (p < 0.001). No significant survival advantage was observed in nonagenarians (aged ≥90 years, p > 0.05) (Supplementary Fig. 4, available online).

Among patients with AVGs, initial AVGs showed a statistically significant association with better survival in octogenarian male patients regardless of DM status (p < 0.001 for DM, p = 0.033 for non-DM) and in females with DM (p = 0.026). In nonagenarians, this benefit was observed only in males without DM (p = 0.020) (Supplementary Fig. 5, available online).

Discussion

We conducted an observational study using the Korean NHIS database to compare the survival rates according to the initial vascular access type in patients undergoing hemodialysis. Specifically, we assessed whether starting hemodialysis with AV access (either AVF or AVG) provided a survival advantage, particularly in older and very old adult patients. Our findings showed that patients who initiated hemodialysis with AV access in place were associated with better survival rates than those who started hemodialysis with a CVC and later underwent AV access creation. When stratified by age, patients aged <80 years who began hemodialysis with an initial AVF or AVG were associated with better survival outcomes. Among those aged 80–89 years, an initial AVF or AVG was also associated with improved survival, especially in diabetic patients. In patients aged ≥90 years, having initial AVFs in place was not associated with a survival benefit; however, having initial AVGs in place was associated with a survival benefit compared with CVC-to-AVG transition, particularly in males and patients without DM.

Previous studies have shown that CVC use in older adult patients with hemodialysis is associated with higher rates of infection and increased mortality [12–16]. Also, studies have shown that CVC use is associated with higher systemic inflammation (e.g., elevated C-reactive protein and interleukin-6 levels) and worse nutritional profiles than AVF use, which may contribute to adverse survival outcomes [17,18]. Therefore, starting hemodialysis with AV access can reduce CVC dependency and improve survival outcomes [4,5,10,11]. However, whether this approach provides survival benefits to older and very old adult patients remains unclear. Arhuidese et al. [19] reported that initiating hemodialysis with AVFs in patients aged ≥75 years resulted in the highest patency rates and greatest survival benefits. Other studies have shown mixed results in very old patients, such as those aged >80 years. Ko et al. [20] reported no significant survival difference between initial AVF and CVC-to-AVF among octogenarians, whereas other studies have demonstrated a survival advantage of initial AVFs, even in octogenarians [21,22].

In our study, we found that initiating hemodialysis with AVFs or AVGs was significantly associated with better survival outcomes among diabetic octogenarians. Patients with DM undergoing hemodialysis with a CVC are at an increased risk of catheter-related infections, which substantially increases the risk of mortality. The risk of infection tends to be even higher in older patients [23–27]. Thus, CVC use may have had a more pronounced impact on mortality in octogenarians with DM. Accordingly, starting hemodialysis with an AVF or AVG in this population may reduce CVC use and infection-related complications, leading to better survival, as observed in our cohort. Recent studies have also reported that age itself does not significantly affect AVF’s primary or secondary patency [28–30], and mature AVFs can be successfully maintained, even in patients over 80 years of age [31–33]. These findings suggest that initial AVF can be considered in very old patients, particularly octogenarians, and initiating hemodialysis with a mature AVF can reduce CVC-related complications and improve survival outcomes.

Furthermore, we observed that initial AVGs were associated with better survival in nonagenarians, particularly in male patients without DM, likely because of reduced CVC dependency at the initiation of hemodialysis. Nonagenarian males often have multiple comorbidities and poor vascular conditions, which may increase the risk of AVF maturation failure and limit its use. However, AVGs have a lower risk of maturation failure than AVFs and can be used earlier, thereby reducing the need to start hemodialysis with CVCs [34,35]. The absence of a survival benefit in nonagenarians with DM, compared to those without DM, may be explained by their higher burden of comorbid conditions and shorter life expectancies. These factors could attenuate the survival impact of reducing CVC dependency through initial AVG use. Although studies in patients aged ≥90 years are limited, our findings suggest that preemptive AVG creation and initial use at hemodialysis initiation can be considered in selected nonagenarians with reasonable life expectancies.

In our study, 57.04% of Korean patients started hemodialysis with CVCs and later underwent AV access creation, whereas 42.96% initiated dialysis directly with AV access. According to the 2024 United States Renal Data System (USRDS) report, among patients who initiated hemodialysis in 2022, 84.7% of patients in the United States began hemodialysis with CVCs, with only 12.5% and 2.8% starting with AVFs or AVGs, respectively [2]. Thus, the proportion of patients initiating hemodialysis with AV access is relatively higher in South Korea. Given that our study confirmed a survival benefit associated with initial AV access, this trend can be considered favorable, and efforts should be made to maintain and further promote early AV access.

This study had several limitations. First, because this study used Korean data, the results may not be applicable globally. The results may differ among populations owing to racial differences and variations in healthcare systems. Second, this was a retrospective observational study using healthcare insurance data, making it difficult to fully control for confounding variables that may have affected the results. Additionally, because the data were obtained using diagnostic and procedure codes in the NHIS database, they may not fully reflect the actual clinical status of the patients, such as frailty, functional capacity, and life expectancy. In particular, because healthier patients are generally more likely to undergo AV access creation before starting hemodialysis, this may have introduced selection bias in the survival comparison. Also, important clinical variables that could affect patient survival, such as laboratory data, BMI, insurance type, and medication use, were not included in this study.

In conclusion, initial AV access was associated with improved survival in both older and very old adult patients undergoing hemodialysis. Among very old individuals, particularly octogenarians with DM, the initial use of AVFs or AVGs was associated with improved survival outcomes. In addition, initial AVG use showed a survival benefit in males and nonagenarians without DM. These findings suggest that AV access creation can be reasonably considered in very old patients—including octogenarians and selected nonagenarians—before hemodialysis.

Supplementary Materials

Article information Continued