Summary
Background and objectives Arteriovenous fistula (AVF) failure remains an important cause of morbidity in hemodialysis patients. The exact underlying mechanisms responsible for AVF failure are unknown but processes like proliferation, inflammation, vascular remodeling, and thrombosis are thought to be involved. The current objective was to investigate the association between AVF failure and single nucleotide polymorphisms (SNPs) in genes related to these pathophysiologic processes in a large population of incident hemodialysis patients.
Design, setting, participants, & measurements A total of 479 incident hemodialysis patients were included between January 1997 and April 2004. Follow-up lasted 2 years or until AVF failure, defined as surgery, percutaneous endovascular intervention, or abandonment of the vascular access. Forty-three SNPs in 26 genes, related to proliferation, inflammation, endothelial function, vascular remodeling, coagulation, and calcium/phosphate metabolism, were genotyped. Relations were analyzed using Cox regression analysis.
Results In total, 207 (43.2%) patients developed AVF failure. After adjustment, two SNPs were significantly associated with an increased risk of AVF failure. The hazard ratio (95% confidence interval) of LRP1 rs1466535 was 1.75 (1.15 to 2.66) and patients with factor V Leiden had a hazard ratio of 2.54 (1.41 to 4.56) to develop AVF failure. The other SNPs were not associated with AVF failure.
Conclusions In this large cohort of hemodialysis patients, only 2 of the 43 candidate SNPs were associated with an increased risk of AVF failure. Whether other factors, like local hemodynamic circumstances, are more important or other SNPs play a role in AVF failure remains to be elucidated.
Introduction
A durable vascular access to the bloodstream is of vital importance for patients undergoing chronic hemodialysis. However, vascular access dysfunction is currently the Achilles’ heel of hemodialysis therapy, accounting for 20% of all hospitalizations in hemodialysis patients leading to unacceptable high morbidity and economic burden (1,2). For chronic hemodialysis, arteriovenous fistula (AVF) is the preferred modality in view of the superior patency rates compared with arteriovenous synthetic grafts. Nonetheless, the durability of AVFs is far from optimal, with 1-year primary patency rates ranging from 60% to 65% (3,4).
The vast majority of arteriovenous (AV) access failure is caused by thrombosis, secondary to disproportionate intimal hyperplasia and impaired outward remodeling of the venous outflow tract (5–8). The stimuli responsible for the localized intimal hyperplastic response in the venous outflow tract are multifactorial and include hemodynamic factors such as turbulent flow, the prothrombotic environment that results from endothelial damage, as well as vascular inflammation (9). The stenotic vascular lesions that arise from this intimal hyperplastic response mainly consist of vascular smooth muscle cells, myofibroblasts, and extracellular matrix proteins (10). Excessive accumulation of extracellular matrix is mediated by several growth factors (5,9,11). Morphologically, these stenotic lesions closely resemble restenotic lesions after percutaneous coronary intervention (12,13). Additional specific pathophysiologic stimuli for intimal hyperplasia in vascular access stenosis include the abnormal calcium/phosphate metabolism in patients with CKD that result in arterial as well as venous calcification of the tunica media (14). Therefore, processes related to vascular function and remodeling, growth factors for extracellular matrix formation, inflammation, coagulation, and calcium/phosphate metabolism probably play an important role in AVF failure. Currently, it is unknown why AVF failure occurs in some individuals but not in others. It has been suggested that genetic factors could play a role in the development of AVF failure (15). However, limited studies have investigated the effects of genetic risk factors that play a role in these processes on AVF failure.
The aim of this study was to investigate the association between AVF failure and single nucleotide polymorphism (SNPs) involved in processes related to endothelial function and vascular remodeling, growth factors, inflammation, coagulation, and calcium/phosphate metabolism in a large population of incident hemodialysis patients.
Materials and Methods
Patients
The Netherlands Cooperative Study on the Adequacy of Dialysis (NECOSAD) is a prospective multicenter cohort study in which incident ESRD patients from 38 dialysis centers in The Netherlands were included. The study was in accordance with the Declaration of Helsinki and was approved by all local medical ethics committees. All patients gave informed consent. We followed patients until AVF failure (defined as surgery, percutaneous endovascular intervention, or abandonment of the vascular access in the first 2 years on dialysis), death, or censoring, that is, transfer to a nonparticipating dialysis center, withdrawal from the study, switch to peritoneal dialysis, transplantation, or end of the follow-up period (April 2006).
Eligibility included age >18 years, no previous renal replacement therapy, and survival of the initial 3 months of dialysis. For these analyses, we used data from patients included between January 1997 and April 2004 with a functional fistula within 3 months after the first dialysis session from whom DNA was available. Patients that were dialyzed using a central venous catheter and patients on peritoneal dialysis were excluded.
Demographic and Clinical Data
Data on age, sex, and primary kidney disease were collected at the start of dialysis treatment. Primary kidney disease was classified according to the codes of the European Renal Association–European Dialysis and Transplant Association (ERA-EDTA) (16). We grouped patients into four classes of primary kidney disease: GN, diabetes mellitus, renal vascular disease, and other kidney diseases. Other kidney diseases consisted of patients with interstitial nephritis, polycystic kidney diseases, other multisystem diseases, and unknown diseases.
SNP Selection and Genotyping
SNPs previously associated with AV access failure were selected after a systematic search of the literature. Furthermore, we also selected SNPs that were associated with coronary restenosis (13,17,18) and vascular aneurysm formation (19,20), because underlying mechanisms of these diseases are thought to overlap. A MEDLINE search using keywords including “hemodialysis,” “arteriovenous access failure,” and “single nucleotide polymorphism” identified six genes previously associated with AV access failure: TNF-α (21), klotho (22), fibrinogen-β (FGB) (23), factor V (24), and matrix metallopeptidase 1 (MMP1) (25). To broaden the search to coronary restenosis and vascular aneurysm formation, the keywords “coronary restenosis,” “percutaneous coronary intervention,” and “aortic aneurysm” were added, identifying another 20 candidate genes. Only SNPs with a minor allele frequency >1% were included. Two multiplex assays were designed using Assay designer software. The final set included 43 SNPs in 26 genes, involved in processes related to endothelial function and vascular remodeling, growth factors, inflammation, coagulation, and calcium/phosphate metabolism (Supplemental Table 1). All SNPs were genotyped by matrix-assisted laser desorption/ionization-time of flight mass spectrometry, using the MassARRAY methodology (Sequenom Inc., San Diego, CA), following the manufacturer's instructions. As quality control, 5% of the samples were genotyped in duplicate. No inconsistencies were observed. All of the negative controls (2%) were negative. All SNPs were in Hardy-Weinberg equilibrium (P≥0.01) and had a call rate >90% (Supplemental Table 2).
Statistical Analyses
Continuous variables are presented as medians and interquartile ranges (IQRs). Categorical variables are presented as numbers with percentages. We calculated hazard ratios (HRs) with 95% confidence intervals (95% CIs) using Cox regression analysis for heterozygote and mutant genotypes compared with wild-type genotypes for AVF failure (first event) within 2 years of follow-up for the 43 SNPs. For the purposes of epidemiologic comparison, we used the false discovery rate (FDR) to adjust for multiple testing (26). Although no universal FDR significance threshold has been defined, a cut-point of 0.20 has been suggested for candidate gene association studies, meaning that one should expect at most 20% of declared discoveries to be false (27). We therefore used a cut-point of 0.20, which resulted in a corrected level of significance of P=0.01 instead of P=0.05. We also investigated the association between clinical factors and AVF failure and the interaction between these clinical factors and SNPs associated with AVF failure using the relative excess risk due to interaction method. All analyses were performed using SPSS statistical software (version 20.0; SPSS, Chicago, IL).
Results
A total of 479 incident hemodialysis patients with an AVF from the NECOSAD cohort were genotyped. Of the 479 patients, 207 (43.2%) reached the endpoint of AVF failure during follow-up. The absolute incidence of AVF failure was 340 per 1000 person-years. Baseline characteristics of the patients are shown in Table 1. The median age was 65.3 years, 36.3% were women, and 15.2% had diabetes mellitus as their primary kidney disease.
Baseline characteristics
Genes Implicated in Endothelial Function and Vascular Remodeling
The results of the analyses of this set of SNPs are summarized in Table 2. Carriers of the AA genotype of the LDL receptor–related protein 1 (LRP1) rs1466535 had a significant 1.75-fold (95% CI, 1.15 to 2.66) higher risk of AVF failure, with a P value of 0.01 and a FDR of 0.009. The annexin A5 SNPs, MMP1 rs11292517, nitric oxide synthesis 3 (NOS3) rs1799983, elastin rs2071307, and quaking rs3763197 were not associated with AVF failure. The two intergenic SNPs, identified in a genome-wide association study for coronary restenosis, were also not associated with AVF failure, although they both demonstrated a nonsignificant trend toward a higher risk in the variant allele carriers.
Polymorphisms related to endothelial function and vascular remodeling and association with AVF failure
Growth Factors and Related Genes
We genotyped 12 SNPs in eight growth factor–related genes (Table 3). None of the SNPs showed a significant association with AVF failure.
Polymorphisms in growth factor related genes and association with AVF failure
Genes Implicated in Inflammation
In total, eight SNPs were genotyped in several genes implicated in inflammatory pathways (Table 4). We did not find an association between AVF failure and the SNPs in IL-6, IL-10, lymphotoxin-α, CD180, and toll-like receptor 4. The TNF rs1800629 AA genotype was associated with a 1.97-fold higher AVF failure risk compared with the GG genotype (95% CI, 1.00 to 3.87; P=0.05; FDR, 0.01), which was not significant considering the threshold of FDR <0.01. Both IL-10 rs1800896 and rs3024498 GG genotypes compared with AA genotypes were associated with a nonsignificant lower risk of AVF failure (HR, 0.73; 95% CI, 0.49 to 1.10; P=0.14; and HR, 0.73, 95% CI, 0.42 to 1.25; P=0.25).
Polymorphisms related to inflammatory genes and association with AVF failure
Genes Implicated in Calcium/Phosphate Metabolism
Eight SNPs were genotyped in the klotho, vitamin D receptor (VDR), and fetuin-A gene (Table 5). We did not find an association between AVF failure and the SNPs in the klotho gene (rs9527025, rs564481, rs397703, and rs577912), the VDR (rs11574027, rs2238135, and rs4516035), or the fetuin-A gene (rs4918).
Polymorphisms related to calcium/phosphate metabolism and association with AVF failure
Genes Implicated in Coagulation
Factor V rs6025, also known as factor V Leiden, was associated with a 2.54-fold (95% CI, 1.41 to 4.56; P=0.002; FDR 0.005) higher risk of AVF failure. rs1044291 and rs1800787 in the FGB gene were not associated with AVF failure (Table 6).
Polymorphisms related to coagulation and association with AVF failure
Our results did not change when we calculated HRs in another model by combining heterozygote genotypes and mutant genotypes or heterozygote genotypes and wild-type genotypes. The 41 SNPs that were not associated with AVF failure remained unassociated in other models.
Interactions and Combined Effects
Female sex was associated with a 1.48-fold (95% CI, 1.12 to 1.95) higher risk of AVF failure (P=0.01). Furthermore, diabetes mellitus as primary kidney disease, compared with other causes, was associated with a 2.01-fold (95% CI, 1.42 to 2.85) higher risk of AVF failure (P<0.001). The combination of factor V Leiden with diabetes mellitus or female sex and the combination of LRP1 rs1466535 AA genotype and female sex did not result in a higher risk on an additive scale using the relative excess risk due to interaction method (Supplemental Table 3). Patients with LRP1 rs1466535 AA genotype with diabetes mellitus had a 2.97-fold higher risk (95% CI, 1.10 to 8.05; P=0.03) of AVF failure compared with patients without diabetes mellitus and without LRP1 rs1466535 AA genotype (reference), whereas patients without LRP1 rs1466535 AA genotype with diabetes mellitus had a 2.11-fold (95% CI, 1.48 to 3.01; P<0.001) higher risk of AVF failure and patients with LRP1 rs1466535 AA genotype without diabetes mellitus had a 1.63-fold (95% CI, 1.07 to 2.47, P=0.02) higher risk of AVF failure (Supplemental Table 3).
When analyzing the two significantly associated SNPs (LRP1 rs1466535 and factor V Leiden) together, 269 patients were carrier of at least one risk allele of these two SNPs. Of these carriers, 48.3% developed AVF failure. In comparison, 36.7% of the patients with the wild-type genotype of these SNPs (n=210) developed AVF failure (P=0.01).
Discussion
We investigated the association between AVF failure and 43 SNPs in 26 genes involved in processes related to endothelial function and vascular remodeling, growth factors, inflammation, coagulation, and calcium/phosphate metabolism. To our knowledge, this study is the largest study that investigated genetic risk factors of AVF failure. We showed that, after adjustment for multiple comparisons by using FDR, LRP1 rs1466535 and factor V rs6025 (factor V Leiden) were significantly associated with a higher risk of AVF failure. TNF rs1800629 was not significantly associated with AVF after adjustment. No significant associations of AVF failure were observed with the other SNPs.
The rs1466535 SNP in the LRP1 gene was the only SNP in the “endothelial function and vascular remodeling” group that was significantly associated with AVF failure. In 2011, this SNP was identified in a genome-wide association study on abdominal aortic aneurysms and that this SNP has a possible functional role in LRP1 expression (20). The mechanism by which LRP1 could influence vascular remodeling leading to aneurysm formation, was suggested to involve regulation of MMP9 expression (20,28). Moreover, LRP1 was shown to be essential for the maintenance of vascular wall integrity mediated through platelet-derived growth factor receptor β and Smad signaling (29). Interestingly, in our study, patients carrying the variant A allele had a higher risk of developing AVF failure, whereas in the above-mentioned study (20), the wild-type allele was the risk allele. Whether this opposite effect indicates that AVF failure is mediated through a lack of vascular expansion, which is necessary for AVF maturation, or that another function LRP1 is involved in AVF failure is unclear. Nevertheless, the observed association of the LRP1 SNP with AVF failure suggests a potential role for LRP1 in vascular remodeling and AVF maturation.
We also showed that factor V Leiden was associated with a 2.54-fold (95% CI, 1.41 to 2.56; P=0.002) higher risk of AVF failure. Factor V Leiden is one of the most common inherited procoagulatory defects and is a well known risk factor for venous thrombosis (30). A previous study also showed that factor V Leiden was associated with vascular access failure (31). In agreement with these observations, another recent study showed that another SNP in the factor V gene was associated with AV access failure in hemodialysis patients (24). Furthermore, thrombophilia (including factor V Leiden) was associated with an increased risk of vascular access dysfunction in yet another study (32). Together, these studies provide growing evidence of a role of factor V Leiden mutation in AV access failure.
Although the -308G>A (rs1800629) SNP in the TNF gene did not remain significantly associated with AVF failure after multiple testing correction, we think it is worth discussing. Because the vascular inflammatory response is likely an important contributor to AVF failure, genetic variation in inflammatory-related genes could be associated with AVF failure. In our study, individuals with the AA genotype of rs1800629 had a nonsignificant higher risk of AVF failure. This SNP has been the focus of several previous studies, focusing on a wide variety of endpoints, including cardiovascular events in rheumatoid arthritis patients (33), irritable bowel syndrome (34), life expectancy (35), tuberculosis susceptibly (36), mortality after ARF (37), and comorbidity and functional status scores in hemodialysis patients (38). The other analyzed SNP in the TNF gene (rs361525) was also not associated with AVF failure in our study, although it was previously associated with coronary restenosis (39). The two SNPs in the anti-inflammatory cytokine IL-10 demonstrated a nonsignificant lower risk of AVF failure development, possibly due to lack of power. Remarkably, these two SNPs were shown to increase the risk of coronary restenosis (18) and cardiovascular events during hemodialysis (40). Whether this discrepancy is caused by the differences in the mechanisms of AVF failure and coronary restenosis remains to be elucidated.
For this study, besides previously reported candidate genes for AV access failure, we also selected SNPs in genes related to coronary restenosis and aortic aneurysm formation. Processes involved in AVF patency, particularly vascular smooth muscle cell proliferation and inflammation, also play a key role in the development of coronary restenosis after percutaneous coronary intervention (13). In contrast with the scarcely available data on genetic determinants of AV access failure, the genetic background of coronary restenosis has been more established (12,13). Although our SNP selection covered a wide range of candidate genes in all involved mechanisms, the obtained results do not indicate an important role for our selected SNPs in the development of AVF failure. One explanation for these observations could be that other SNPs are involved in AVF failure or that genetic susceptibility does not play an important role in the development of AVF failure. Based on other studies that also did not find an association between most candidate SNPs and AV access failure (23–25,41), it might be that local factors, such as hemodynamics and vascular damage, have a more important role in the failure of the vascular access required for hemodialysis than the genetic background.
Diabetes mellitus is a well known risk factor for both renal disease and cardiovascular complications as well as for AVF failure in dialysis patients (42–44). The prevalence of diabetes among dialysis patients differs considerably worldwide. According to the ERA-EDTA Registry, the prevalence of diabetes mellitus among dialysis patients in Europe is 22% (45), in line with the relatively low prevalence in our study population. Although the interaction analysis does show some influence of diabetes on the genetic associations, elaborate subgroup analyses are not appropriate considering the low proportion of diabetic patients in our population.
Our study has several potential limitations. We had no information about AVF failure before the first successful dialysis session. Whether inclusion of nonmaturating fistula would have affected our results is questionable. It is conceivable that the maturating process of an AVF is mechanistically different compared with AVF failure of a well maturated fistula. Excluding the nonmaturating AVF likely resulted in a more homogeneous study endpoint. However, this hypothesis remains speculative because we do not have data to explore this. In addition, we had no information about the protocols that were used in the various centers for access surveillance to detect stenosis before failure. Furthermore, we had no information about the cause of fistula failure (i.e., thrombosis, stenosis, or other). A previous study suggested that hospital-specific aspects contribute to AVF failure (46). Unfortunately, we do not have information about differences of fistula creation in the involved centers, about the experience of surgeons in fistula creation or about the surgical techniques used to create a fistula. Another limitation is that, despite our large number of patients, we had limited power to find an association between AVF failure and several SNPs. In addition, we investigated the association between multiple SNPs and fistula failure, thereby increasing the chance of false positive findings. However, because we selected only candidate genes based on previous studies, and because we adjusted for multiple comparisons by using FDRs with the threshold set at 20%, previously suggested to be an appropriate threshold for candidate gene association studies (27), we have tried to minimize the chance of false positive findings. The general strength of this study was the large and well defined Dutch cohort of incident hemodialysis patients with available DNA for investigation of genetic risk factors. Another potential strength of our study was that we only included incident hemodialysis patient, thereby limiting survivor bias.
Current guidelines for prevention of vascular access failure recommend uniform surveillance of all patients (47). Identification of genetic risk factors might lead to a more directed approach for surveillance techniques. Risk factors for primary patency loss could be used to focus on specific patient groups for more intensive surveillance. Furthermore, increasing our understanding of the molecular mechanisms of AVF failure could aid the development of better preventive measures or new treatment modalities.
In conclusion, we found that LRP1 rs1466535 and factor V Leiden were associated with an increased risk of AVF failure. Other SNPs in vascular function and remodeling related genes, growth factor genes, inflammation genes, coagulation genes, and calcium metabolism genes were not significantly associated with AVF failure. Further studies on the genetics of AVF failure are needed to unravel the underlying mechanisms of this deleterious condition and thereby help us improve prevention and treatment of AVF failure in this diseased population.
Disclosures
None.
Acknowledgments
We thank the investigators and study nurses of the participating dialysis centers and the data managers of NECOSAD for collection and management of data. The members of the NECOSAD Study Group include: A.J. Apperloo, J.A. Bijlsma, M. Boekhout, W.H. Boer, H.R. Büller, F.Th. de Charro, C.W.H. de Fijter, C.J. Doorenbos, W.J. Fagel, G.W. Feith, L.A.M. Frenken, W. Grave, P.G.G. Gerlag, J.P.M.C. Gorgels, R.M. Huisman, K.J. Jager, K. Jie, W.A.H. Koning-Mulder, M.I. Koolen, T.K. Kremer Hovinga, A.T.J. Lavrijssen, A.J. Luik, K.J. Parlevliet, M.H.M. Raasveld, M.J.M. Schonck, M.M.J. Schuurmans, C.E.H. Siegert, C.A. Stegeman, P. Stevens, J.G.P. Thijssen, R.M. Valentijn, M. van Buren, M.A. van den Dorpel, P.J.M. van der Boog, J. van der Meulen, F.M. van der Sande, A. van Es, J.A.C.A. van Geelen, G.H. Vastenburg, C.A. Verburgh, H.H. Vincent, and P.F. Vos.
We thank the nursing staff of the participating dialysis centers and the staff of the NECOSAD trial office for their invaluable assistance in the collection and management of data for this study. Furthermore, the authors thank Dennis Kremer and Eka Suchiman from the Molecular Epidemiology Section of the Leiden University Medical Center for their expert assistance with the Sequenom Massarray genotyping platform, and Petra Noordijk from the Epidemiology Department for her practical assistance.
J.W.J. was funded by grants from the Interuniversity Cardiology Institute of the Netherlands, the European Community Framework KP7 Programme under grant agreement (HEALTH-F2-2009-223004), the Center for Medical Systems Biology (a center of excellence approved by the Netherlands Genomics Initiative/Netherlands Organization for Scientific Research), and the Netherlands Consortium for Healthy Ageing. The funders had no role in the study design, data collection and analysis, decision to publish, or the preparation of the manuscript.
Footnotes
Published online ahead of print. Publication date available at www.cjasn.org.
This article contains supplemental material online at http://cjasn.asnjournals.org/lookup/suppl/doi:10.2215/CJN.11091012/-/DCSupplemental.
- Received October 28, 2012.
- Accepted February 26, 2013.
- Copyright © 2013 by the American Society of Nephrology
References
In this issue
