The Value of a Registry

Introduction

Intracranial arteriovenous malformations (iAVMs) are rare lesions, with an annual incidence of symptomatic lesions of approximately 1/100,000 population. AVMs are believed to be congenital, often going undetected for decades until found incidentally during workup for other medical issues or for related symptoms such as headaches and seizures. Most commonly, however, the initial presentation of an iAVM is due to rupture and cerebral hemorrhage. The risk of hemorrhage related to iAVMs is approximately 1%–4% per year, and the majority of patients present in the third or fourth decade of life. The heterogeneity of iAVMs makes the management, treatment, and study of these lesions difficult to standardize, with multiple factors, including AVM size, arteriovenous architecture, patient age, and AVM location, all impacting clinical decision-making. Given the absence of strong data regarding management of iAVMs, the American Stroke Association Stroke Council has released recommendations detailing the use of microsurgical, endovascular, and/or radiosurgical interventions to serve as a guideline for treatment. A conservative approach to management of iAVMs (medical treatment and observation) is also frequently employed, especially for lesions deemed to carry a significantly high risk of intervention-associated morbidity and mortality. The need for high-level evidence for the management of iAVMs is a current gap in the practice of neurovascular surgery, one that multiple studies have attempted to address.

Issues With Randomized Control Trials and iAVMs

The lack of scientifically proven benefit of interventional treatment for iAVMs led to the development of the ARUBA study, which examined whether medical management alone vs medical management with interventional therapy would lead to differences in the time to death or symptomatic stroke. The study concluded that medical management alone was superior to medical management with interventional therapy (surgery, endovascular treatment, or stereotactic therapy alone or in combination). The study was stopped early by the National Institute of Neurological Disorders and Stroke safety monitoring board due to a difference in primary outcome of 30.7% in the interventional group vs 10.1% in the medical management alone group. This study has only added to the controversies surrounding the management of iAVMs. Proponents argued that the study was the first to attempt a true randomized trial among this group of varied patients and presentations, with beneficial information gained regarding the merits and risk of intervention as well as the natural history of iAVMs.

Despite these methodological goals, the study has drawn substantial criticism related to the study design itself. ARUBA grouped various treatment modalities as a single arm, which may either mask a harmful intervention or invalidate a beneficial one. Grouping the interventional treatments together without appropriate stratification and subgroup analysis makes it difficult to accurately assess the true differences in the primary outcome between the intervention arm and the medical management arm. This is especially true in the ARUBA study, as only 5 patients were treated with surgery alone, with a further 13 patients treated with surgery and either endovascular intervention or radiosurgery. Notably, endovascular embolization alone is not viewed as standard of care for curative treatment of iAVMs in the United States. With over 30 patients receiving endovascular treatment alone, the interventional management of this arm of the study does not reflect common neurosurgical practice in the United States.

Moreover, the study was not appropriately powered to determine differences between the different treatments within the intervention arm. Additional concerns from the ARUBA study include improper hypothesis generation, inappropriate endpoints, lack of standardization, low enrollment, premature interruption, and overall study design. While the study did not adequately answer the intended questions of whether medical management was preferable to intervention in patients with unruptured iAVMs, it did serve as a valuable example of the need for further research strategies and options in the study of these lesions.

Traditionally, randomized control trials (RCTs) and meta-analyses of RCTs have been the gold standard of evidence in scientific studies. They are designed to be highly controlled experiments that test specific hypotheses with the aim of applying the findings to real-world clinical practice. As such, they offer an excellent basis to evaluate the safety and efficacy of treatments, whether they be new devices, drugs, or interventional techniques. RCTs are designed with stringent criteria to best evaluate the desired scientific question, relying on inclusion and exclusion criteria, appropriate statistical power and modeling, precisely defined study protocols, and clearly delineated outcomes. While this methodological approach has tremendous value, there are inherent limitations to its use and applicability among larger populations. In rare diseases, strict criteria regarding patient inclusion or exclusion can lead to severe underpowering of RCTs, making it difficult to draw any definitive conclusions about a study’s outcomes.

In the case of trials that focus on surgical diseases, the design and execution of RCTs are even more arduous, as there can be many treatment options and interventions for the same disease state. Patient-centered decision-making can also create challenges for RCTs of surgical management, as patients may be less agreeable to randomization due to concern of not receiving all possible treatments, especially in a trial comparing medical treatment alone vs surgical intervention plus medical treatment. Additionally, many physicians outside of larger academic centers lack the resources and infrastructure necessary to participate in well-designed RCTs, and there are many patients whose clinical and treatment data are simply unavailable to RCTs. This can be a particular problem for studies of rare diseases.

In the context of iAVMs, the lack of well-conducted and designed RCTs in the literature is due to many of these and other factors: relatively low incidence of disease, overall heterogeneous patient population/disease characteristics, and lack of standardized treatment paradigms in the clinical community. The inherent limitations of RCTs hinder research into diseases like iAVMs, necessitating the development of national and international patient registries to increase the information available to clinicians and researchers.