Introduction

Around 3% of the adult global population has unruptured intracranial aneurysms (UIAs) [1]. The increased usage and improved quality of cranial imaging have resulted in more frequent detection of these lesions. UIAs can remain clinically asymptomatic, present with focal neurological deficits from local mass effect or ischemia, or they may rupture. Rupture of an aneurysm results in aneurysmal subarachnoid hemorrhage (SAH), which has a case fatality rate up to 35% and a high risk for permanent neurological disabilities as well as neuropsychological disorders in survivors [2, 3].

Previous meta-analyses on development or rupture of UIAs have been hampered by varying definitions of risk factors, which sometimes led to inconsistent results [4, 5]. In a pooled analysis of individual patient data from six prospective cohort studies, six easily retrievable predictors enabled the calculation of the 5-year risk of aneurysm rupture: population, hypertension, patient age, aneurysm size, earlier SAH from another aneurysm, and aneurysm site. However, several other potential risk factors, such as smoking status during follow-up or aneurysm morphology, could not be included in the risk score because data were not collected at all or with varying methods [5, 6].

The Common Data Elements (CDE) project for standardizing data for neurological clinical research was initiated by cerebrovascular clinicians and scientists under the auspices of The National Institute of Neurological Disorders and Stroke to facilitate pooling and comparison of such data on cerebrovascular disease.

Process for Selecting CDEs

For a description of the UIA and SAH CDE project, we refer to the main article of this project [7]. For development of CDEs for the WG ‘UIA’ of ‘UIA and SAH,’ a multidisciplinary and international group of nine cerebrovascular specialists on research and treatment of UIAs was assembled by the two work group (WG) co-chairs (NE, GJER) (Fig. 1). Following systematic review and collection of the current data on UIAs by the two co-chairs, existing CDEs on ischemic stroke were integrated in data sheets. These were circulated to all WG members and subsequently reviewed, expanded, or modified by each member. Additionally, the WG developed a proposal for classification of aneurysm morphology based on current data. Further, CDEs that apply to multiple WGs (e.g., aneurysm size or location in the imaging WG) were crosschecked and/or adapted with these subcommittees to exclude heterogeneous definitions of the same CDEs.

Fig. 1
figure 1

Development of CDEs

Full size image

Finally, the CDEs were categorized into four groups: (1) Core CDEs—elements which can be consistently collected across studies and which should be employed in studies concerning the corresponding particular disease or therapeutic area; (2) Supplemental—Highly Recommended CDEs—elements that are essentially based on certain conditions or study types in clinical research studies and that are strongly recommended for the specific disease or therapeutic area; (3) Supplemental CDEs—elements that are commonly collected in clinical research studies, but whose relevance depends on the study design or type of research; and 4. Exploratory CDEs—elements which are reasonable to use, but whose validity is yet limited due to insufficient availability and validation of data. The categorization was proposed by the two co-chairs, and after another round of review and revision of all CDEs, all the WG members agreed on the final proposal of the UIA CDEs. These findings were presented at the 4th Neurocritical Care Research Conference in Houston, Texas, May 2016, for further review and revision within the SAH CDE research group. The final version of the CDEs was once more circulated within the WG, and upon agreement, the final case report forms were developed.

Common Data Elements Overview

The ‘UIA’ WG collected 91 CDEs, 69 new CDEs, and 22 already established CDEs where only minor changes had to be made for our purposes. The CDEs were divided into eight categories: demographics, reason of medical consult and diagnosis, clinical symptoms and assessment at baseline, risk factors, concomitant medications, concomitant diseases, radiological findings, as well as management of unruptured aneurysms. Each CDE was assigned a specific identification number, a CDE name, variable name, definition, classification, permissible values, a code name, a code description and if necessary, a unit of measure and a question text.

The CDEs were classified as 8 Core (see below), 23 Supplemental—Highly Recommended, 25 Supplemental, and 35 Exploratory elements (Tables 1–8).

Table 1 CDEs—demographics
Full size table

Description of Core CDEs

Demographics (Table 1).

  • Patient age (preexisting CDE) [5, 8,9,10].

  • Risk factors (Table 4).

  • Hypertension (defined as systolic blood pressure greater than or equal to 140 mmHg or greater than or equal to 90 mmHg diastolic in adults or systolic or diastolic blood pressure above the 95 percentile in children) (preexisting CDE) [5, 8, 11,12,13].

  • Tobacco smoking status: (a) never, (b) former (including start and end date of smoking), (c) current (including starting date) and (d) unknown (modified CDE) [1, 9,10,11,12,13,14,15,16,17].

Radiological findings (Table 7).

  • Anatomical aneurysm site based on angiography (modified CDE) [5, 9, 18,19,20,21,22,23,24,25,26].

  • Maximum aneurysm diameter (in mm) in any direction (new CDE) [5, 8, 9, 16,17,18,19,20, 22,23,24,25,26,27,28,29].

  • Maximum aneurysm height (in mm) perpendicular to aneurysm neck (new CDE) [30].

  • Maximum aneurysm width (in mm) perpendicular to aneurysm height (new CDE) [30].

  • Aneurysm morphology type: (a) regular, (b) bleb, (c) daughter-sac/multilobed aneurysm (new CDE, Fig. 2) [20,21,22,23, 27, 30,31,32,33].

    Fig. 2
    figure 2

    Classification of aneurysm morphology

    Full size image

Description of UIA CDEs

For ‘reason of medical consult and diagnosis,’ one CDE with eight permissible values was established (Table 2). The subtopic ‘clinical symptoms and assessment at baseline’ contains eight CDEs, of which three were novel and the remaining were edited (Table 3). Twelve CDEs on ‘risk factors’ were established, comprised of six novel and six reutilized CDE. Permissible values and further information were added to the CDE ‘tobacco smoke history status’ as a core item (Table 4). Additionally, a classification concerning the morphology of an aneurysm was established (Fig. 2). There are 10 CDEs in the subtopic ‘concomitant medications’, of which seven were novel and three were modified (Table 5). The subject area ‘concomitant diseases’ contains 21 novel CDEs and three already established CDEs (Table 6). For ‘radiological findings,’ we compiled 27 CDEs, out of these the CDE ‘imaging modality vessel imaging angiography type’ was reused and permissible values were added to the Stroke CDE ‘imaging vessel angiography aneurysm’ (Table 7). Six novel CDEs were established for ‘UIA management’ (Table 8).

Table 2 CDEs—reason of medical consult and diagnosis
Full size table
Table 3 CDEs—clinical symptoms and assessment at baseline
Full size table
Table 4 CDEs—risk factors
Full size table
Table 5 CDEs—concomitant medications
Full size table
Table 6 CDEs—concomitant diseases
Full size table
Table 7 CDEs—radiological findings
Full size table
Table 8 CDEs—Management
Full size table

Limitations

This project has limitations: We identified and defined numerous data elements in the setting of UIAs, based on existing and/or most commonly used definitions. We balanced between very detailed definitions, which would decrease the feasibility of using these, and broad definitions, which are easier to use in clinical practice, but may provide less scientific details. Thus, the definitions and their scientific implication remain uncertain and need to be validated prospectively. Further, for several CDEs, we had to define limits or cut-off values. For example, for hypertension, we used the accepted definitions of the cardiac guidelines, a systolic pressure of 140 mmHg and a diastolic one of 90 mmHg, but it remains uncertain whether these cut-off values for hypertension are clinically relevant in terms of risk factor for growth or rupture. Further, a consensus approach was used to define and rank the individual importance of data elements based on existing literature. Thus, other potentially relevant data elements suggested by experimental or case–control studies could not be included at present because of the lack of validated measurement tools or grading scales for such outcomes (e.g., aneurysm wall inflammation in imaging studies). Additionally, the WG established and agreed on a novel classification system for aneurysm morphology, for which there were no immediate data from the previous cohort or case–control studies to support this exact classification. However, since three-dimensional aneurysm morphologies are difficult to measure or to describe in standardized manner, the WG agreed on a two-dimensional classification as a basis for further research. Lastly, the established CDEs on UIAs may need to be adapted or even expanded, e.g., with neurocognitive outcome measures and grading scales in patients undergoing preventive UIA repair or follow-up imaging in the future, if there are sufficient data to support this. Despite these limitations, standardized collection of the proposed CDEs will at least provide data on whether the CDEs as currently defined are risk factors for the development and rupture of intracranial aneurysms.

Next Steps/Future Work

Future clinical studies need to test and validate the sensitivity and relative importance of the UIA CDEs established. Furthermore, data should be derived from more advanced imaging modalities in the setting of UIAs including their standardized measurement and morphological analysis.

Conclusions

We defined and categorized a total of 91 CDEs, of which 71 were novel for UIAs. These CDEs on UIAs could serve as a basis to standardize future studies and thereby help to harmonize data across studies. However, the CDEs remain to be validated, adapted, and updated in the future based on novel data for optimizing already existing CDEs and for establishing new CDEs.