More Than Reducing Early Fistula Thrombosis Is Required: Lessons From the Dialysis Access Consortium Clopidogrel Fistula Study

Population 

In the DAC clopidogrel study, 61% of AVFs were not suitable for dialysis, even though the patient risk profile was relatively low: the average age, less than that of the US HD population,4 there were fewer women included, and those with significant vascular disease were often excluded. This failure rate is significantly higher than that reported in other studies7 and highlights the importance of both patient and access selection. Patient criteria for an AVF were not defined in this study and remain undefined in national guidelines and in the wider literature. Preoperative vascular mapping was performed in approximately 75% of patients, yet did not appear to contribute to improved AVF suitability. This confirmation of previous studies that demonstrate limited yield of preoperative vascular mapping20, 21 is disconcerting, especially as it is unclear if preoperative mapping was a standard approach in most centers or if patients with poor vessels on physical exam were sent for mapping. Preoperative mapping typically provides data on vessel size and presence of peripheral and central lesions that are presumably corrected prior to AVF creation. Clearly more than just vessel size is required for eligibility, otherwise preoperative vessel mapping and corrective intervention would have facilitated greater fistula suitability overall.

Outcomes 

The DAC investigators carefully considered and clearly struggled with the selection of outcomes.17 They chose AVF primary patency as the primary outcome since the biological effect of clopidogrel would be more closely related to thrombosis rather than dialysis suitability, which is more influenced by variations in clinical practice. However, the investigators also recognized that “improving fistula patency is of little value if it does not translate into improved suitability… .”17 The investigators indicated that “the completeness and reliability of the ascertainment of this outcome will be a major focus during the conduct of this trial.”17 However, while the definition and ascertainment of the outcome was reliable, was it really valid? Only 2% of patients in each group had surgical or percutaneous intervention to restore patency or promote maturation before 6 weeks. The investigators themselves recognized that “a procedure to restore function of arteriovenous access is required in close to 50% of patients within 1 year of access placement.”17 While this may not specifically relate to early failures, previous studies have demonstrated that approximately one-third (24% to 42%)22, 23, 24 of all AVFs will require intervention to facilitate maturation. In a recent study of the natural history of AVFs,25 only 9% of AVFs matured without need for intervention. Given the low intervention rate, was it really suitability failure that was found or failure of clinicians to intervene early on fistulas that were failing to mature?

While the definition and ascertainment of the suitability outcome was well implemented, careful consideration of aspects that may contribute to fistula dialysis suitability other than preventing early thrombosis, such as stricter protocols for fistula eligibility and facilitative preoperative and postoperative intervention, may have provided a more accurate assessment of the AVF suitability failure outcome. A comparison of outcomes with other studies is difficult due to the lack of standardized definitions for fistula primary failure, failure to mature, and now dialysis suitability. Such standardized definitions are needed.

Considerations for Clinicians 

Clopidogrel cannot be recommended for use to reduce early AVF thrombosis and improve AVF dialysis suitability. Instead, clinicians and surgeons should agree upon and test eligibility criteria for AVF creation to tailor access choice and reduce the high failure rate. A recent, easy-to-administer, 4-variable prediction score has been developed and validated to estimate a patient's risk of fistula failure to mature7 that can be used in conjunction with other eligibility criteria. Given the results of this well designed RCT, a re-evaluation of the Fistula First one-size-fits-all approach may reduce suitability failure. A functioning fistula is the agreed-upon ultimate goal, but placing it first may not always be the correct approach for all patients. For example, if a patient is deemed to be at high risk using eligibility criteria or the prediction score, placing a synthetic graft to mature (dilate and toughen) vessels in preparation for future AVF creation when the graft fails has been demonstrated to be a successful strategy26 and suggested by the National Kidney Foundation Kidney Disease Outcomes Quality Initiative (KDOQI) clinical practice guidelines.27 This study demonstrates that achieving a suitable AVF is not an easy task. The appropriate patient must be identified, with vessel selection, experienced surgical technique, follow up, and facilitative intervention as necessary. When these issues are optimized, one could reconsider a trial of antiplatelet therapy.

Implications for Achieving Performance Measures 

The US CMS developed clinical performance measures (CPMs) based on KDOQI guidelines,1, 5 recommending a national goal of increasing AVF use to 65% with the stipulation that AVF use must be via 2 needles. The panel rationalized that “given the estimation for fistula failure rates, AV fistulas may even need to be placed in up to 85% of patients starting hemodialysis in order to reach the target of 65% for prevalent patients.” Given the results of this study, if 85% of patients received an AVF, only 34% of patients would have useable AVF via 2 needles.

Implications for Researchers 

Choosing the most appropriate primary outcome is always challenging. While thrombosis is a readily quantifiable “hard” outcome, it is only one of many factors that determine AVF suitability, and as such can only be considered a surrogate endpoint. Clinicians need a clinically important outcome that can be interpreted into a meaningful change in practice and that can improve patient care at the bedside. Examining the literature on graft patency highlights this point. Although access flow monitoring with intervention can improve access flow and, in some cases, reduce thrombosis rates, few studies have shown benefit in overall graft survival.28, 29 The limitation of many clinical end points is the need for larger sample sizes, longer follow-up periods, and greater resources that may prohibit the study. While the authors criticize that their own criteria in ascertaining suitability failure were too stringent, they provide an excellent example to researchers on the use of sensitivity analysis and the use of conditional power calculations for the secondary outcome.

While this study was terminated early for a perceived benefit in reducing thrombosis, it was executed following predetermined O'Brien-Fleming stopping rules with 4 interim analyses. The stopping boundary allocated the majority of the total α to the last interim analysis so that a very large treatment effect was required to meet the stopping rule. The careful use of a data and safety monitoring board during the design and interim analysis phases ensured that such termination would provide important information on the secondary clinical endpoint of AVF suitability failure. While the reduced sample size of 877 was 32% smaller than the target sample size for this outcome, the higher than expected failure rate of 61% (versus 40%) provided a 95% CI that suggested clopidogrel might have reduced the risk of suitability failure by a clinically unimportant 6%. The investigators' success in designing and implementing the largest, multicenter double blind RCT of AVF patency to date is a great inspiration to researchers. Such a study could not have been possible without their careful methodological planning, execution, and collaboration.9, 10, 11, 12, 13, 14, 15, 16

In conclusion, while clopidogrel reduced early AVF thrombosis, its use cannot be recommended as it failed to improve AVF suitability for dialysis cannulation. The high AVF suitability failure rate of 61% has significant implications for clinicians, administrators, researchers, and importantly, our patients. One needs to consider a multipronged approach to improving fistula suitability rates. This may include the use of antiplatelet agents, aggressive monitoring with intervention, and specified selection of patients for fistulas and enhanced cannulation skills. Evaluation and validation of such approaches require standardized patient and outcome definitions and rigorous clinical trials testing as exemplified by the DAC.