Challenges to Performance of Clinical Trials on Vascular Access

For the nearly 500,000 patients in the United States relying on hemodialysis, vascular access is a critical component of the care they receive, and identifying and individualizing the optimal access type for this population has been difficult.¹

Vascular access options for hemodialysis include tunneled dialysis catheters (TDCs), arteriovenous fistulas (AVFs), and arteriovenous grafts (AVGs). All approaches have their merits and associated risks in patency, morbidity, mortality, and quality of life.² Furthermore, once a particular access has been established, different approaches to maintenance can also be evaluated. To fairly compare these approaches, both observational trials³ and randomized control trials (RCTs) have been conducted.^4,5

RCTs and meta-analyses of several RCTs provide the highest level of evidence to evaluate a particular therapy, largely because of the belief that these designs can reduce, if not eliminate, investigator bias.⁶ When conducting a meta-analysis, authors may include low-power studies, creating heterogeneity in the analysis.^6,7 To remedy these weaknesses in meta-analyses and generate more generalizable results for the overall patient population requiring hemodialysis, multiple adequately powered, low-bias RCTs are required.

Conducting such trials presents many challenges for vascular access investigators. To date, no multicenter trials comparing AVFs and AVGs have been completed, although a trial is currently in process.⁴ Unlike many disease states in nephrology, vascular access trials require the close cooperation of surgeons and interventionalists, suggesting that the surgical and interventional literature may hold important insights that will assist with designing and executing vascular access trials. Review of the scientific literature has identified several barriers, which include heterogeneous end points with sometimes questionable clinical significance, restrictive inclusion and exclusion criteria, center and investigator variability, difficulty generating trial awareness, provider treatment bias, patient treatment bias, the presence of competing trials, the creation of patient inconvenience, and finally the economic cost to perform the study. Table 1 summarizes the key points and possible solutions.

Table 1.

Challenges and possible solutions to performance of vascular access trials

Issue	Possible Resolution
Heterogeneous end points	Utilize standardized, patient-centered end points
Appropriate inclusion and exclusion criteria	Maintain broad inclusion criteria and educate investigators carefully regarding use of the technique in question
Center and investigator variability	Coach patients and investigators to improve communication and ensure consistent approaches to the clinical issue
Difficulty generating trial awareness	Pursue an interdisciplinary approach to disease and trial awareness
Provider treatment bias	Educate providers about gaps in the literature to justify clinical equipoise involving different treatments
Patient treatment bias	Communicate with patients and provide trustworthy sources to inform them of treatment options, thus making patients more amenable to randomization
Presence of competing trials	Ensure patients are enrolled in the most appropriate trial, on the basis of clinical judgment
Patient inconvenience that reduces follow-up and can be a barrier to enrollment	Recognize challenges faced by patients requiring hemodialysis in adhering to treatment protocols and conscientiously design protocols to minimize them
Cost	Proactively work with funding agencies to identify and apply for appropriate resources

Heterogeneous End Points

Valid comparison between numerous studies relies on low outcome measurement variability. As studied by Tong, et al., most reported trial outcomes are both inconsistent and not patient centered, with the most commonly reported variables being phosphate, biomarkers, and dialysis adequacy. These are reported far more than outcomes most important to patients, which measure quality of life and frequency of complications.⁸

This variability is currently being mitigated by the Standardized Outcomes in Nephrology initiative.⁸ In this initiative, 13 patients/caregivers met with 46 professionals in a workshop to develop a core outcome measure for vascular access trials.⁹ From this workshop, both patients and professionals agreed that the highest measure of function was “uninterrupted use of the access without the need for interventions” and “ability to receive prescribed dialysis.”⁹ This consensus was further studied in a survey of 873 patients and professionals across 58 countries, evaluating their value of 12 defined outcomes with a Likert scale.¹⁰ Although patients and professionals had differences in the exact outcomes they prioritize, thematic analysis of the survey found six highly valued outcomes: the necessity for hemodialysis, applicability across vascular access types, frequency and severity of debilitation, minimizing the risk of hospitalization and death, optimizing technical competence and adherence to best practice, and direct effect on appearance and lifestyle. This precise, extensive measurement of professional and patient values informs the patient-centered approach in vascular access trials.

A further issue is the importance of understanding the causal pathway that ultimately leads to a clinical end point of interest. Frequently, a surrogate end point is used instead of a clinically meaningful one because of practicality or ethical concerns. This issue was seen in the Dialysis Access Consortium Study group trial, which evaluated the use of clopidogrel as an adjunct for AVF maturation. Although clopidogrel was associated with improved patency, this did not translate to improved suitability for use on dialysis.¹¹ A similar issue was explored in the A randomized trial Comparing Catheters to fistulas in Elderly patientS Starting HemoDialysis trial, which aimed to randomize patients between AVFs and TDCs. The investigators hypothesized that the observed improvement in survival after AVF creation was due to surgeon selection of healthier patients for AVF surgery, not necessarily the AVF itself.¹²

Restrictive Inclusion and Exclusion Criteria

Inclusion and exclusion criteria for patients must be balanced and leave space for clinical expertise while remaining strict enough to rigorously evaluate the treatments of interest. As an example, a hypothetical device for a stenotic AV graft may be specifically tailored to a particular geometry and material. As such, recruiting a homogeneous population with these favorable characteristics may increase the expected magnitude of treatment effect. As a result, study sponsors may impose restrictive inclusion and exclusion criteria to match this predicted population. This practice will create a significant difference between the study cohort and the population that will receive the treatment in practice, jeopardizing external validity. These policies also unnecessarily restrict opportunities to access promising therapies and slow trial accrual. One solution is to keep the inclusion criteria as broad as possible, but carefully educate investigators about the appropriate use of the study device or intervention.

A policy of making trials as inclusive as possible to protect external validity has been studied in other disease processes.¹³ A similar policy should be pursued in vascular access trials.

Center and Investigator Variability

Previously anecdotal knowledge about center inconsistency was quantified in centers participating in the Best Endovascular vs. Best Surgical Therapy in Patients with Critical Limb Threatening Ischemia (BEST-CLI) trial, which investigated treatments for critical limb ischemia.¹⁴

Sites were separated into two categories on the basis of trial recruitment performance—high-performing sites (HPSs) and low-performing sites (LPSs). The authors found that the investigators recruiting in HPSs were most impeded by the inability to convince patients and families that alternative methods of revascularization may be equivalent, whereas LPS investigators struggled with patients consenting to their treatment being chosen at random. The differences highlighted between BEST-CLI HPSs and LPSs are widely applicable to vascular access studies, which typically involve numerous sites of varying recruitment levels.⁵ In addition to the potential for hindering enrollment of patients, such significant interinvestigator differences can lead to criticism that there is a threat to internal validity because the HPS and LPS patients may be systematically different and the treatments also be systematically different. Careful selection of trial sites can help mitigate this concern.

Difficulty Generating Trial Awareness

When patients receive care, they are often unaware that treatment options include clinical trials.¹⁵ Traditional methods of increasing clinical trial awareness include “piggybacking” on mainstream disease awareness months, broadcasting radio ads with educational messages, and advertising in media sources used by a particular population. Furthermore, building relationships with multidisciplinary teams both inside and outside the hospital who may refer patients is time consuming. These outside teams are critical for generating referrals, and good lines of communication are required to raise awareness and stand out in a crowded field of other competing priorities.

Although certain oncologic disease surged in public awareness with these public awareness campaigns, the role of RCTs among patients requiring hemodialysis treatment is low, despite the large population of patients requiring hemodialysis. Such low awareness manifests in 25 recent RCTs recruiting at a mean of 80 patients total per study, with trials having to end because of recruitment as low as 0.18 patient recruited per center per month.⁵

Provider Treatment Bias

Vascular access surgeons and interventionalists at a particular site are typically veteran clinicians, with a large portfolio of patients and procedures. They are also well versed with the existing literature and may have significant experience with a particular AVF technique or AVG material. However, the ethical basis for an RCT is predicated on the presence of clinical equipoise, as determined by the collective profession, between two treatments. An alternative rationale is that an individual clinician may be uncertain between two treatment options for their patient.¹⁶ Clinicians may individually believe that a certain treatment is superior for a patient's needs. Because of this provider-side preference, investigators may refuse to enroll their patient in an RCT.

One example of this effect can be seen from the “Fistula First” initiative.¹⁷ Since then, evidence has accumulated that patient factors such as advantageous anatomy and fewer comorbidities may be the dominant reason for the observed outcomes in patients with AVF. Furthermore, although antiplatelets, such as clopidogrel, can reduce the frequency of early fistula thrombosis, the functional patency rate for AVF maturation was only 75% at 2 years, with almost 50% of fistulas requiring an intervention to maintain use or treat a complication.^3,11 This leaves a high proportion of patients with AVF surgeries unable to experience the long-term benefit of fistula over graft creation.¹⁸ It was also found that patients were forced to rely on suboptimal dialysis access methods such as TDCs in the several months required for AVF maturation, further eroding the long-term survival advantage of AVFs.¹⁹

The influential “Fistula First” program still affects both care providers and patients when creating a treatment plan for vascular access²⁰ and may lead to challenges in recruitment for studies that suggest that an AVF may not be the best first access.¹²

Patient Treatment Bias

Patients considering dialysis access procedures frequently have multiple sources of information available, including online services, but are also influenced by family members or members of the dialysis care team. As a result, they may have a strong preference toward a particular treatment option. This can extend to the point that they will not participate in randomization between possibly equivalent options as part of their medical care, or even be nonadherent to the randomization if enrolled.

This latter effect has been seen in the Asymptomatic Carotid Atherosclerosis Study vascular surgery trial, which assessed carotid stenosis treatment.²¹ Surgical studies, like many hemodialysis vascular access RCTs, have a significant nonadherence to assigned randomization, although the actual reasons for this is ultimately unclear. The Asymptomatic Carotid Atherosclerosis Study was such a study, wherein control group patients received best medical therapy, as defined at that time, and treatment group patients also underwent a carotid endarterectomy. In this trial, 101 of the 825 assigned surgical patients declined carotid endarterectomy despite previous agreement to accept either treatment.²¹

This nonadherence dilutes the observed treatment effect and creates an additional task for the investigator to educate patients about the trial and the randomization process. It is especially important because vascular access studies are primarily surgical, thus requiring the greatest degree of investigator communication to ensure randomization adherence.

Presence of Competing Trials

Investigators often run clinical trials serving patient populations with multiple conditions. This is especially the case among the patient population requiring hemodialysis vascular access. Many patients have numerous comorbidities, qualifying them for multiple trials to treat any one of their conditions.²² In this situation, more than one study can fail from lack of recruitment or low-powered results. Another issue is the reality of limited investigator time. Investigators may be overwhelmed by the requirements associated with trial recruitment when they already have busy schedules and other trials to consider, as well. Once again, another burden is on investigators, who need to increase specific trial awareness, while ensuring that patients are enrolled in appropriate trials that are consistent with prevailing clinical judgment.

Creation of Patient Inconvenience

Patients often face “arduous, unpaid work” while navigating the health system to attend appointments, financial treatment, and remain compliant with in-home care.²³ These barrier points are essentially unpaid work that can be defined as “hassles.” As measured by a study of a multisite telephone survey about women veteran's health care at veteran's affairs, 39% of patients reported four or more hassles in obtaining their care, which was associated with a five-fold increase in the predicted probability of delaying or forgoing care.²³

Patients receiving hemodialysis are uniquely challenged with the need to attend typically thrice weekly dialysis treatments. Consequently, investigators must ask patients to participate in appointments, follow-up care, and surveillance studies.

Economic Cost

The economic expense of conducting any clinical trial can be considerable. Research coordinators and nurses are required to ensure adherence to protocols. Information technology experts are needed for data management and security. The time of the physician investigators must also be accounted. For vascular access trials in particular, the cost of the device or pharmaceutical being evaluated can also be substantial and, in many circumstances, can only be borne by a commercial entity with an interest in the product. Although such an arrangement can lead to productive science, there is also the concern that this may lead to bias.²⁴ Some innovative techniques do not lend themselves easily to this mechanism, and in that case government funding is an option, but can also be quite difficult to obtain.²⁵ Despite these challenges, very successful trials have been proposed and executed with varied funding mechanisms.^3,11,26,27

Conclusions

In the vascular access space, there is a need for high-quality evidence that can help guide practice. RCTs are a critical part of this effort but are fraught with nuance and challenge for the investigators designing and conducting them. Investigators can glean important “lessons learned” from trials in other parts of medicine, including the surgery and radiology literature. Recurring themes include balancing the need for adequate study enrollment and methodological imperatives, especially appropriate inclusion and exclusion criteria. Trials must also be designed to enhance patient and investigator ability to adhere not only with randomization but also with follow-up and treatment procedures. Vascular access trialists are charged with “threading the needle,” ensuring appropriate scientific rigor in trial design, and then executing well, as they advance their field of study.

Disclosures

T. H. Yuo reports consultancy agreement with Becton Dickinson, Merit Medical, and WL Gore; research funding from Merit Medical; honoraria from Merit Medical; advisory or leadership role in BD and Medtronic; and site principal investigator for an investigational device for Merit Medical. The remaining author has nothing to disclose.

Funding

None.

Author Contributions

Conceptualization: Theodore H. Yuo.

Formal analysis: Theodore H. Yuo.

Investigation: Nicole G. Alindogan.

Writing – original draft: Nicole G. Alindogan, Theodore H. Yuo.

Writing – review & editing: Nicole G. Alindogan, H. Theodore Yuo.