Long-term Prognosis of Vascular Access in Haemodialysis Patients with Systemic Lupus Erythematosus: A Retrospective Cohort Study


 Patients with systemic lupus erythematosus (SLE) have a higher risk of vascular complications. The objective of the study is to analyse the differences in the risk of arteriovenous fistula or graft (AVF/AVG) dysfunction in haemodialysis patients with and without SLE over a 10-year period. A final sample of 1366 SLE and 4098 non-SLE patients remained after propensity score matching. The outcome measure primarily evaluated is the incidence rate of AVF/AVG dysfunction. SLE patients had higher incidence rates of AVF/AVG dysfunction than non-SLE patients in all time periods. Three specific time periods that reached significant difference were the following: (1) after 1 year (incidence rates = 15.21, 13.01, respectively; subdistribution hazard ratio (SHR) = 1.15; P = 0.007), (2) 1st-to-10th-year period (incidence rates = 15.36 and 13.25, respectively; SHR = 1.16; P = 0.007), and (3) overall period (incidence rates = 23.53 and 21.66, respectively; SHR = 1.09; P = 0.027). In conclusion, there were significantly higher incidence rates of AVF/AVG dysfunction in SLE patients during the long-term follow-up period in this study.


Introduction
Systemic lupus erythematosus (SLE) is a systemic autoimmune disease that has a worldwide prevalence ranging from 0.3 to 23 per 100,000 person-years, affecting many of different age, racial, and ethnic groups [1]. Asian SLE patients manifest higher rates of renal involvement (50-60%) compared to Caucasian patients (30-38%) and are often associated with greater risk of severe renal disease [2]. There were approximately 612.8 SLE cases per 100,000 patient-years that progressed to end-stage renal disease (ESRD) and received haemodialysis (HD) treatment based on the National Health Insurance Research Database (NHIRD) in Taiwan between 2000 to 2008 [3].
Vascular diseases are commonly observed in SLE patients. Vascular access dysfunction, involving either arteriovenous stulas (AVF) or arteriovenous grafts (AVG), is an important factor that not only determines the quality of HD, but also has crucial impact on morbidity and mortality. Prolonging access patency and limiting the complications of a functioning access require interprofessional collaborative practice.
Previous research pointed out that SLE patients on HD are more probable to develop vascular access thrombosis [4]. Theoretically, SLE patients may have increased risk of vascular access patency loss even from the time of AVF/AVG creation. However, it is still uncertain whether or not the SLE disease itself has an impact on vascular access patency. There are studies conducted that were focused only on the short-term outcome (within 1 year) of vascular access patency in SLE-ESRD patients, and the conclusion is still very controversial [4,5]. So far little is known whether or not there is a difference in the rate of AVF/AVG dysfunction between SLE patients and non-SLE patients during the long-term follow-up period (after 1 year and onwards). Therefore, the aim of the study is to investigate the long-term dysfunction rate of AVF/AVG in HD patients with and without SLE.

Study population and data source
In this retrospective cohort study, data were obtained from Taiwan's NHIRD. Since 1995, all citizens and residents in Taiwan are provided with compulsory universal health insurance. The program provides full coverage for renal replacement therapy for patients with ESRD. Healthcare institutions are then required to submit computerized claim documents for renal replacement therapy to the National Health Insurance Administration. Taiwan's NHIRD is a population-based data source for producing realworld data to help make diagnostic decisions and health care policies, which covers almost all of the inpatient and outpatient medical records for Taiwan's 23 million residents. Information such as patient identi cation number, birthday, gender, dates of hospital admission and discharge, healthcare institutions providing services, ICD-9-CM/ICD-10-CM diagnostic and procedure codes, and outcomes, among many other data, are stored in this database.
The study was carried out in accordance with the Helsinki Declaration (edition 6, revised 2000) and was approved by the Institutional Review Board of Taipei Veterans General Hospital (2020-09-018BC). The need of informed consent was waived by the Institutional Review Board of Taipei Veterans General Hospital since the dataset was encrypted and de-identi ed.
The primary outcome in this study is the cumulative incidence rate of AVF/AVG dysfunction, which measures the occurrence of an intervention (angioplasty, thrombectomy, or new AVF/AVG creation) from the time of vascular access creation to the rst episode of dysfunction within 3 months, 6 months, 1 year, 5 years, and 10 years. Secondary outcomes include the occurrence of major adverse cardiovascular events (MACE) (de ned as the rst occurrence of death from cardiovascular causes, non-fatal myocardial infarction, or non-fatal stroke), myocardial infarction, and ischaemic stroke.

Patient Selection
Data were collected retrospectively from the NHIRD for HD patients in Taiwan between 2000 and 2011. The patients were divided into two groups (SLE and non-SLE group). The exclusion criteria of the study are as follows: (1) under 20 years old, (2) undergoing peritoneal dialysis, (3) pregnant, (4) kidney transplant recipients, (5) had never initiated HD via AVF/AVG or had a permanent double-lumen catheter placed after AVF/AVG creation, and (6) ineligible for the National Health Insurance catastrophic illness card (given to HD patients who require life-long renal replacement therapy).
Statistical analysis SAS version 8.0 (SAS Institute, Cary, North Carolina, USA) was used to conduct statistical analysis. Continuous variables were expressed in mean ± SD and analysed using t-test. All data were normally distributed. Categorical variables were expressed in number and percentage and analysed using chi-square test. The propensity scores of the probability of SLE diagnosis were determined using multivariate logistic regression analysis, conditional on the baseline covariates (Supplementary Table 1). Three non-SLE patients were matched with each patient in the SLE cohort with a similar propensity score based on the nearest neighbor matching without replacement using calipers of width equal to 0.1 of the standard deviation of the logit of the propensity score. The survival curves for the cumulative incidence rate of AVF/AVG dysfunction were analysed using Cox regression model, Kaplan-Meier method, and log-rank test. All reported tests were two-sided and the statistically signi cant value was set at P < 0.05.

Results
A total of 146818 HD patients were enrolled. However, 65308 patients were excluded from the study for the following reasons: 276 were under 20 years old, 14111 underwent peritoneal dialysis, 0 was pregnant, 1523 were kidney transplant recipients, 37149 had never initiated HD via AVF/AVG, and 9153 had a permanent double-lumen catheter placed after AVF/AVG creation. A total of 81510 patients were selected, which comprised of 1366 SLE and 80144 non-SLE patients. After implementing the catastrophic illness card exclusion criteria, 1366 SLE and 4098 non-SLE patients remained in the study after propensity score matching with a ratio of 1 to 3.
The baseline characteristics of SLE and non-SLE patients shown in Table 1 were not found to be signi cantly different. The distribution of patients in the SLE and non-SLE group have similar values in terms of mean age (51 and 51.2 years old, respectively), gender percentage (23% males and 77% females; 25% males and 75% females, respectively), mean Charlson Comorbidity Index scores (6.0 for both groups), and number of patients with AVF (1189 and 3577, respectively). The use of concomitant medications and presence of comorbidities were also similar in both groups.  Fig. 1 con rmed these results where there was a statistically signi cant difference in the cumulative incidence of AVF/AVG dysfunction between SLE and non-SLE patients (P = 0.048).  The incidence rates (per 100 person-years) and risks of MACE, myocardial infarction, and ischaemic stroke in SLE and non-SLE patients are shown in Table 4 and the differences between groups were not found to be statistically signi cant. However, SLE patients have a lower incidence rate of ischaemic stroke than non-SLE patients (0.84 vs. 1.09, respectively; SHR = 0.77; P = 0.074) where it may seem that SLE patients may seem to have a lower risk of developing ischaemic stroke.

Discussion
The main hypothesis of the study is that there may be a difference in the risk of AVF/AVG dysfunction between SLE and non-SLE patients due to endothelial activation, apoptosis, and atherogenesis [6][7][8]. To the best of our knowledge, previous research addressing the risk of vascular access dysfunction in SLE patients receiving HD had not analysed the rate of dysfunction in SLE patients after 1 year [4,5]. The observation period in this study was over a 10-year duration. The results have shown that SLE patients on HD had a signi cantly higher risk of developing AVF/AVG dysfunction during the long-term period follow up, especially after the rst year, 1st-year-to-10th-year, and overall periods. The possible pathogenetic mechanisms associated with a higher risk of vascular access thrombosis (VAT) include the Virchow triad, which consists of stasis, hypercoagulability, and endothelial injury [10]. Stasis is the condition of reduced blood ow within the vascular access. Hypoalbuminemia is a predisposing factor for stasis and is usually attributed to nephrotic syndrome or disease exacerbation in SLE, both of which may lead to vascular access dysfunction [11,12]. Hypercoagulability in SLE may be attributed to lupus-speci c antibodies (aPLs), which can lead to VAT through possible mechanisms such as atherogenesis, and endothelial activation [4]. In ammation in SLE may also increase certain procoagulant factors that may increase the risk of developing VAT [13]. Endothelial activation and damage is commonly observed in SLE patients. Different mechanisms have been proposed to explain the prevalence of endothelial dysfunction in SLE [14]. Atehortúa et al. pointed out that different components of the immune system seem to participate in endothelial injury, such as autoantibody production and immune complex formation, which is characterized by an increase in the expression of adhesion molecules, production of pro-in ammatory cytokines and prothrombotic factors, oxidative stress upregulation, and abnormal vascular tone modulation [15]. The structural damage and attenuation of endothelial function in vascular access may lead to their loss of viability and integrity, which may eventually result in possible long-term vascular access failure.
Numerous studies have shown that SLE patients have increased risks of developing MACE, acute myocardial infarction, and stroke [33][34][35]. However, this study did not demonstrate SLE patients having a higher risk of developing MACE, AMI, and stroke than non-SLE patients. Differences in demographic characteristics in the SLE population of this study may account for the different outcomes and further studies may be needed to reevaluate the relationship between SLE, HD, and the aforementioned adverse events.
There are several limitations of this study that should be noted. This is a retrospective study and utilized a database where laboratory markers as potential prognostic variables cannot be analysed. It was also conducted in a single country (Taiwan) where all of the participants were of Chinese ethnicity. The prognosis and outcomes between SLE and non-SLE patients with different ethnicities is unknown, and may limit the generalization of results. The number of patients with AVG listed in the database were also too few and was combined with the number of patients with AVF for the nal analysis. In spite of the aforementioned limitations, this study enrolled the largest number of SLE patients in analysing vascular access and has the longest observation period of 10 years.
AVF/AVG dysfunction in SLE patients is of crucial clinical relevance since it worsens the quality of life and is a clinical challenge for the healthcare professionals in HD units. Additional randomized large-scale prospective studies are needed in the future to con rm the results in this study and to also address the following important issues: (1) the roles of autoantibodies and other additional factors contributing to pathogenesis of AVF/AVG dysfunction, (2) the role of antiplatelet or anticoagulation in preventive strategy against VAT, and (3) the interaction between SLE, hemostasis, and immunological system in the pathogenesis of thromboembolism in SLE patients under maintenance HD.
In conclusion, there were signi cantly higher incidence rates of AVF/AVG dysfunction in SLE patients than non-SLE patients during long-term follow-up period (especially after 1 year and during the 1st-to-10th year period) in this study. A multi-disciplinary approach may be considered to improve vascular access patency in SLE patients. In order to extend the durability of a permanent vascular access and prevent further complications, such as VAT, early diagnostic evaluation of aCL and LAC activity in ESRD patients is recommended.

Declarations
Competing interests writing -original draft: F.Y.C and C.F.C.; Manuscript writing -review and editing: A.C.T. and C.C.L. All authors have approved the submitted version of the manuscript.