[Skip to Navigation]
Sign In
Figure.  Time to Rupture After Observation of Intracranial Aneurysm Growth
Time to Rupture After Observation of Intracranial Aneurysm Growth
Table 1.  Patient and Aneurysm Characteristics
Patient and Aneurysm Characteristics
Table 2.  Characteristics of Patients and Aneurysms During Follow-up After Growth Detection
Characteristics of Patients and Aneurysms During Follow-up After Growth Detection
Table 3.  Comparison Between Ruptured and Unruptured Aneurysms After Detection of Growth
Comparison Between Ruptured and Unruptured Aneurysms After Detection of Growth
Table 4.  Triple-S Prediction Model for Estimating Rupture Risk After Aneurysm Growtha
Triple-S Prediction Model for Estimating Rupture Risk After Aneurysm Growtha
1.
Vlak  MHM, Algra  A, Brandenburg  R, Rinkel  GJE.  Prevalence of unruptured intracranial aneurysms, with emphasis on sex, age, comorbidity, country, and time period: a systematic review and meta-analysis.   Lancet Neurol. 2011;10(7):626-636. doi:10.1016/S1474-4422(11)70109-0PubMedGoogle ScholarCrossref
2.
Gabriel  RA, Kim  H, Sidney  S,  et al.  Ten-year detection rate of brain arteriovenous malformations in a large, multiethnic, defined population.   Stroke. 2010;41(1):21-26. doi:10.1161/STROKEAHA.109.566018PubMedGoogle ScholarCrossref
3.
Greving  JP, Wermer  MJH, Brown  RD  Jr,  et al.  Development of the PHASES score for prediction of risk of rupture of intracranial aneurysms: a pooled analysis of six prospective cohort studies.   Lancet Neurol. 2014;13(1):59-66. doi:10.1016/S1474-4422(13)70263-1PubMedGoogle ScholarCrossref
4.
Algra  AM, Lindgren  A, Vergouwen  MDI,  et al.  Procedural clinical complications, case-fatality risks, and risk factors in endovascular and neurosurgical treatment of unruptured intracranial aneurysms: a systematic review and meta-analysis.   JAMA Neurol. 2019;76(3):282-293. doi:10.1001/jamaneurol.2018.4165PubMedGoogle ScholarCrossref
5.
Villablanca  JP, Duckwiler  GR, Jahan  R,  et al.  Natural history of asymptomatic unruptured cerebral aneurysms evaluated at CT angiography: growth and rupture incidence and correlation with epidemiologic risk factors.   Radiology. 2013;269(1):258-265. doi:10.1148/radiol.13121188PubMedGoogle ScholarCrossref
6.
Mehan  WA  Jr, Romero  JM, Hirsch  JA,  et al.  Unruptured intracranial aneurysms conservatively followed with serial CT angiography: could morphology and growth predict rupture?   J Neurointerv Surg. 2014;6(10):761-766. doi:10.1136/neurintsurg-2013-010944PubMedGoogle ScholarCrossref
7.
Inoue  T, Shimizu  H, Fujimura  M, Saito  A, Tominaga  T.  Annual rupture risk of growing unruptured cerebral aneurysms detected by magnetic resonance angiography.   J Neurosurg. 2012;117(1):20-25. doi:10.3171/2012.4.JNS112225PubMedGoogle ScholarCrossref
8.
Chien  A, Callender  RA, Yokota  H,  et al.  Unruptured intracranial aneurysm growth trajectory: occurrence and rate of enlargement in 520 longitudinally followed cases.   J Neurosurg. 2019;132(4):1077-1087. doi:10.3171/2018.11.JNS181814PubMedGoogle ScholarCrossref
9.
Korja  M, Lehto  H, Juvela  S.  Lifelong rupture risk of intracranial aneurysms depends on risk factors: a prospective Finnish cohort study.   Stroke. 2014;45(7):1958-1963. doi:10.1161/STROKEAHA.114.005318PubMedGoogle ScholarCrossref
10.
Brinjikji  W, Zhu  YQ, Lanzino  G,  et al.  Risk factors for growth of intracranial aneurysms: a systematic review and meta-analysis.   AJNR Am J Neuroradiol. 2016;37(4):615-620. doi:10.3174/ajnr.A4575PubMedGoogle ScholarCrossref
11.
Hackenberg  KAM, Algra  A, Salman  RAS,  et al; Unruptured Aneurysms and SAH CDE Project Investigators.  Definition and prioritization of data elements for cohort studies and clinical trials on patients with unruptured intracranial aneurysms: proposal of a multidisciplinary research group.   Neurocrit Care. 2019;30(suppl 1):87-101. doi:10.1007/s12028-019-00729-0PubMedGoogle ScholarCrossref
12.
Backes  D, Vergouwen  MDI, Velthuis  BK,  et al.  Difference in aneurysm characteristics between ruptured and unruptured aneurysms in patients with multiple intracranial aneurysms.   Stroke. 2014;45(5):1299-1303. doi:10.1161/STROKEAHA.113.004421PubMedGoogle ScholarCrossref
13.
Harrell  FE,.  Regression Modeling Strategies. Springer International Publishing; 2015. doi:10.1007/978-3-319-19425-7
14.
Jou  LD, Mawad  ME.  Growth rate and rupture rate of unruptured intracranial aneurysms: a population approach.   Biomed Eng Online. 2009;8:11. doi:10.1186/1475-925X-8-11PubMedGoogle ScholarCrossref
15.
Chang  HS.  Simulation of the natural history of cerebral aneurysms based on data from the International Study of Unruptured Intracranial Aneurysms.   J Neurosurg. 2006;104(2):188-194. doi:10.3171/jns.2006.104.2.188PubMedGoogle ScholarCrossref
16.
Koffijberg  H, Buskens  E, Algra  A, Wermer  MJH, Rinkel  GJE.  Growth rates of intracranial aneurysms: exploring constancy.   J Neurosurg. 2008;109(2):176-185. doi:10.3171/JNS/2008/109/8/0176PubMedGoogle ScholarCrossref
17.
Etminan  N, Rinkel  GJ.  Unruptured intracranial aneurysms: development, rupture and preventive management.   Nat Rev Neurol. 2016;12(12):699-713. doi:10.1038/nrneurol.2016.150PubMedGoogle ScholarCrossref
Limit 200 characters
Limit 25 characters
Conflicts of Interest Disclosure

Identify all potential conflicts of interest that might be relevant to your comment.

Conflicts of interest comprise financial interests, activities, and relationships within the past 3 years including but not limited to employment, affiliation, grants or funding, consultancies, honoraria or payment, speaker's bureaus, stock ownership or options, expert testimony, royalties, donation of medical equipment, or patents planned, pending, or issued.

Err on the side of full disclosure.

If you have no conflicts of interest, check "No potential conflicts of interest" in the box below. The information will be posted with your response.

Not all submitted comments are published. Please see our commenting policy for details.

Limit 140 characters
Limit 3600 characters or approximately 600 words
    4 Comments for this article
    EXPAND ALL
    Triple S model may underestimate the rupture risk of enlarged aneurysms.
    Toshikazu Kimura, DMSc | Japanese Red Cross Medical Center
    We read with great interest the article by Laura T van der Kamp et al. “Risk of Rupture After Intracranial Aneurysm Growth”1 The authors collected a large number of unruptured cerebral aneurysm enlargements (n=312) from multicenter data and performed a survival analysis based on the time from enlargement to the onset of subarachnoid hemorrhage (SAH) in 128 patients who were not treated radically. Twenty-five cases of hemorrhage were observed, and the relatively high number of hemorrhagic events makes their conclusion that "UCAs have a high risk of SAH in the immediate period of enlargement, and this risk decreases over time" compelling. It has been speculated that many small aneurysms may develop SAH within a few months of aneurysm formation, and the same may be true for enlargement. Considering the prevalence of UCAs, it is likely that the default biological response to aneurysm formation or enlargement is repair.

    However, there are some aspects of their results that are questionable. Aneurysm enlargement is a relatively rare event, so it is reasonable to need to collect cases from multiple centers. The differences between Japanese/Finnish and European/North American populations, which are believed to have different bleeding rates (which probably also affect the rate of enlargement), have been analyzed. In this study, the Japanese/Finns were 30% of the total population. The risk of bleeding in the Chinese population is not clear at present, but the adequacy of the sample composition may be an issue when considering the prediction of bleeding risk due to enlargement.

    Second, although they observed two cases of bleeding from enlarged ICA aneurysms, analysis of actual aneurysms that caused SAH showed significantly less bleeding from ICA aneurysms compared to the major sites (MCA, Acom, Pcom, BA)2 ICA aneurysms should have an extremely low risk of bleeding compared to the rest of the sites, which may underestimate the overall bleeding rate.

    Lastly, they excluded cases that were treated after enlargement (128/312), which, based on previous reports, likely had a high bleeding risk and acceptable treatment risk. As mentioned in Limitation, this is likely to have led to a low estimate of the calculated absolute risk. Therefore, the triple S model that they advocate should be used with great caution.

    Despite the inevitable bias that aneurysms with unignorable rupture risk will be treated, the authors' observational study of many enlarged aneurysms should contribute to better risk communication.

    References

    1. Laura T van der Kamp, GJE R, et al. Risk of Rupture After Intracranial Aneurysm Growth. JAMA Neurol. Published online August 30, 2021. doi:10.1001/JAMANEUROL.2021.2915
    2. Molyneux A, Kerr R, et al, International Subarachnoid Aneurysm Trial (ISAT) Collaborative Group. International Subarachnoid Aneurysm Trial (ISAT) of neurosurgical clipping versus endovascular coiling in 2143 patients with ruptured intracranial aneurysms: a randomised trial. Lancet (London, England). 2002;360(9342):1267-1274. doi:10.1016/S0140-6736(02)11314-6
    CONFLICT OF INTEREST: None Reported
    READ MORE
    RE: Risk of Rupture After Intracranial Aneurysm Growth
    Tomoyuki Kawada, MD | Nippon Medical School
    van der Kamp et al. conducted a prospective study by compiling 15 international cohorts (1). The authors handled 312 patients with unruptured cerebral aneurysm enlargements. During 864 aneurysm-years of follow-up, 25 (7.6%) of these aneurysms ruptured. The hazard ratios (HRs) (95% confidence intervals [CIs]) of size (>=7mm) and irregular shape for rupture event were 3.1 (1.4-7.2) and 2.9 (1.3-6.5), respectively. In contrast, HR of each site for rupture event did not reach the level of significance. Peduzzi et al. simulated the effect of events per independent variable (EPV) in proportional hazards regression analysis (2,3). As a conclusion, EPV value less than 10 has some problems to keep validity of the statistical model. Although EPV value less than 10 is also acceptable to investigate the association (4), the enough number of EVP is preferable in prediction model for the valid estimation. I suppose that the number of events was small for stable estimates, and wide ranges of 95% confidence interval may reflect the lack of statistical power.

    In addition, the authors did not cite a paper by Hostettler et al (5), which was a case-control study. The adjusted odds ratios of Black ethnicity compared to White and some aneurysm locations compared to middle cerebral artery for rupture significantly increased. In addition, antihypertensive medication, hypercholesterolemia, aspirin use, internal carotid artery location, and aneurysm size were inversely associated with rupture. Although I cannot understand the inverse relationship between aneurysm size and rupture event, some protective factors should be specified by further prospective studies. Especially in the ethnicity difference in risk of rupture, I suppose that there is a limitation of international cohorts, and ethnicity-specific risk assessment might also be needed.

    References
    1. van der Kamp LT, Rinkel GJE, Verbaan D, et al. Risk of Rupture After Intracranial Aneurysm Growth. JAMA Neurol 2021 Aug 30. doi: 10.1001/jamaneurol.2021.2915. [Epub ahead of print]
    2. Concato J, Peduzzi P, Holford TR, et al. Importance of events per independent variable in proportional hazards analysis. I. Background, goals, and general strategy. J Clin Epidemiol 1995;48(12):1495-1501.
    3. Peduzzi P, Concato J, Feinstein AR, et al. Importance of events per independent variable in proportional hazards regression analysis. II. Accuracy and precision of regression estimates. J Clin Epidemiol 1995;48(12):1503-1510.
    4. Vittinghoff E, McCulloch CE. Relaxing the rule of ten events per variable in logistic and Cox regression. Am J Epidemiol 2007;165(6):710-718.
    5. Hostettler IC, Alg VS, Shahi N, et al. Characteristics of Unruptured Compared to Ruptured Intracranial Aneurysms: A Multicenter Case-Control Study. Neurosurgery 2018;83(1):43-52.
    CONFLICT OF INTEREST: None Reported
    READ MORE
    Accurately Assessing the Risk of Rupture After Intracranial Aneurysm Growth
    Prateek Agarwal, MD, MBA | University of Pittsburgh Medical Center
    We read with great interest the recently published investigation entitled “Risk of Rupture After Intracranial Aneurysm Growth”.[1] We applaud the authors’ attempt to determine the risk of intracranial aneurysm rupture after growth, as this can inform decision calculus with data-driven evidence. While some prior studies have already tackled this question [2,3], this investigation identifies several risk factors for aneurysm rupture after growth detection and presents an accessible Triple-S Prediction model for physicians to employ in risk-benefit conversations with patients.

    However, we have significant concerns regarding data analysis and assumptions. As the authors note, their study is observational and
    vulnerable to selection bias, resulting in underestimation of aneurysm rupture risk after growth detection. For instance, 41% of aneurysms followed after growth detection were treated. Physicians likely considered these aneurysms more prone to rupture, leaving behind a relatively benign population of untreated aneurysms for follow-up. This is supported by larger median growth (1.8 mm vs. 1.5 mm), greater number with irregular shape (58% vs 51%), and higher proportion arising from the MCA (37% vs. 29%) for treated versus untreated aneurysms, factors which all translate to an increased rupture risk in this study.

    Even more impactful, 27 patients with less than 1 day follow-up after growth detection were excluded. If one includes the 6 patients who ruptured the day growth was detected, the absolute risk of rupture after growth detection is 9.3%, compared to 7.6% quoted in the study, and the absolute cumulative risk of rupture at 6 months is approximately 4.6%, compared to 2.9% quoted in the study. Assuming that some of the 13 patients with no clinical follow-up or 8 patients who were acutely treated would have had single aneurysm rupture after growth, the absolute cumulative risk of rupture at 6 months would increase by approximately 0.3 percentage points per ruptured aneurysm. This yields a 6-month cumulative rupture risk ranging from 4.6% to 10.3%, beyond the 6 patients who had rupture on the same day as growth.
    It is imperative that we do not provide a false sense of security to physicians and patients with cerebral aneurysm growth who must decide between intervention and observation. The fact that 6 aneurysms ruptured the day growth was detected underscores the urgent importance of careful patient evaluation. Consequently, it is our responsibility to caution readers that this study, despite its merits, likely significantly underestimates cerebral aneurysm rupture risk after growth detection.
    Prateek Agarwal, MD, MBA, Robert M. Friedlander, MD, MA
    References
    1. van der Kamp LT, Rinkel GJE, Verbaan D, et al. Risk of Rupture After Intracranial Aneurysm Growth. JAMA Neurol. 2021.
    2. Brinjikji W, Zhu YQ, Lanzino G, et al. Risk Factors for Growth of Intracranial Aneurysms: A Systematic Review and Meta-Analysis. AJNR Am J Neuroradiol. 2016;37(4):615-620.
    3. Inoue T, Shimizu H, Fujimura M, Saito A, Tominaga T. Annual rupture risk of growing unruptured cerebral aneurysms detected by magnetic resonance angiography. J Neurosurg. 2012;117(1):20-25.
    CONFLICT OF INTEREST: None Reported
    READ MORE
    Questions on Utilising the Triple-S Risk Prediction Model to estimate Risk of Rupture After Intracranial Aneurysm Growth
    Ronil Chandra, MBBS, MMed, Leon T. Lai, PhD, Thanh G. Phan, PhD | Monash Medical Centre, Monash Health, Monash University, Melbourne, Australia
    We read with great interest the article by van der Kamp and colleagues1 in which they investigated the absolute risk of intracranial aneurysm rupture after detection of growth and developed a rupture risk prediction model using aneurysm size, site, and shape. This is an important clinical problem to address, as the absolute rupture risk is not clear in the months and years following detection of aneurysm growth, and the findings in this study provide potential guidance for management of this cohort.
    However, utilisation of the triple-S model underestimates the risk of rupture, in part due to the inevitable treatment selection
    bias where 41% of the cohort (128 of the 312 patients) had aneurysms that were selected for repair. More importantly, this perpetuates the flawed assumption that small aneurysms (<7mm) have negligible risk of rupture.2 In the current study, aneurysm repair was not performed at random, but judgement was used to select patients with aneurysms perceived to be at higher rupture risk or with lower treatment risk associated with repair.
    Of the remaining patients with growing aneurysms that were not treated, a large proportion (38%) that subsequently ruptured were <7mm after growth was detected. In addition, whilst it’s known that the risk of rupture increases with age,3 a smaller proportion of elderly patients were treated compared to younger patients (19% of patients aged ≥70 years treated compared to 48% with age <70 years), presumably related to perceived reduced benefit over the remaining healthy lifespan, and treatment complication risk.4
    Therefore, the risk predictions, and perhaps the manuscript title, would be more accurately interpreted by the JAMA Neurology readership as the Risk of Rupture after Intracranial Aneurysm Growth in patients selected for management without repair.

    References:
    1. van der Kamp LT, Rinkel GJE, Verbaan D, et al. Risk of Rupture After Intracranial Aneurysm Growth. JAMA Neurol. 2021. doi:10.1001/jamaneurol.2021.2915
    2. Wiebers DO. Unruptured intracranial aneurysms: natural history, clinical outcome, and risks of surgical and endovascular treatment. The Lancet. 2003;362(9378):103-110.
    3. Etminan N, Chang HS, Hackenberg K, et al. Worldwide Incidence of Aneurysmal Subarachnoid Hemorrhage According to Region, Time Period, Blood Pressure, and Smoking Prevalence in the Population: A Systematic Review and Meta-analysis. JAMA Neurol. 2019;76(5):588-597.
    4. Algra AM, Lindgren A, Vergouwen MDI, et al. Procedural Clinical Complications, Case-Fatality Risks, and Risk Factors in Endovascular and Neurosurgical Treatment of Unruptured Intracranial Aneurysms: A Systematic Review and Meta-analysis. JAMA Neurol. 2019;76(3):282-293.



    CONFLICT OF INTEREST: RVC reports grants from National Health & Medical Research Council of Australia (APP1143155) and TGP reports grants from National Health & Medical Research Council of Australia (GNT1163282), outside the submitted work.
    READ MORE
    Original Investigation
    August 30, 2021

    Risk of Rupture After Intracranial Aneurysm Growth

    Author Affiliations
    • 1Department of Neurology and Neurosurgery, University Medical Center Utrecht Brain Center, Utrecht University, University Medical Center Utrecht, Utrecht, the Netherlands
    • 2Department of Neurosurgery, Amsterdam University Medical Centers, Amsterdam, the Netherlands
    • 3Department of Radiology and Nuclear Medicine, Amsterdam University Medical Centers, Amsterdam, the Netherlands
    • 4Department of Neurosurgery, the Jikei University School of Medicine, Tokyo, Japan
    • 5Department of Clinical Radiology, Kuopio University Hospital, Kuopio, Finland
    • 6Institute of Clinical Medicine, School of Medicine, Faculty of Health Sciences, University of Eastern Finland, Kuopio, Finland
    • 7Department of Neurosurgery, Kuopio University Hospital, Kuopio, Finland
    • 8Department of Neurosurgery, Institute of Neurological Science, Glasgow, United Kingdom
    • 9Division of Neuroradiology, Joint Department of Medical Imaging and Neurosurgery, Toronto Western Hospital, University Health Network, Toronto, Ontario, Canada
    • 10Department of Surgical Neurology, Akita Cerebrospinal and Cardiovascular Center, Akita, Japan
    • 11Department of Advanced Neurosurgery, Ehime University Graduate School of Medicine, Toon City, Ehime, Japan
    • 12Department of Neurology, Haaglanden Medical Center, the Hague, the Netherlands
    • 13Clinical Neurosciences, University of Turku, Turku, Finland
    • 14Department of Neurosurgery, Neurocenter, Turku University Hospital, Turku, Finland
    • 15Department of Neurosurgery, Radboud University Medical Center, Nijmegen, the Netherlands
    • 16Department of Surgery, Prince of Wales Hospital, Hong Kong, China
    • 17Department of Imaging and Interventional Radiology, Basement, Yue Kong Pao Centre for Cancer and the Lady Pao Children’s Cancer Centre, Prince of Wales Hospital, Hong Kong, China
    • 18Department of Neurological Surgery, Nippon Medical School, Tokyo, Japan
    • 19Department of Neurosurgery, Kyorin University, Tokyo, Japan
    • 20Department of Neurosurgery, University Hospital Mannheim, University of Heidelberg, Mannheim, Germany
    • 21Department of Radiology, University Medical Center Utrecht, Utrecht University, Utrecht, the Netherlands
    • 22Julius Centre for Health Sciences and Primary Care, University Medical Centre Utrecht, Utrecht, the Netherlands
    JAMA Neurol. 2021;78(10):1228-1235. doi:10.1001/jamaneurol.2021.2915
    Key Points

    Question  What is the absolute risk of rupture of an intracranial aneurysm with growth detected during follow-up?

    Findings  In this cohort study including 312 patients with aneurysms with growth, within 1 year after growth detection, the absolute risk of rupture was 4.3%. In the triple-S prediction model based on 3 independent predictors of rupture (size, site, and shape), the 1-year risk of rupture ranged from 2.1% to 10.6%.

    Meaning  In this study, the triple-S prediction model was useful in estimating the risk of rupture after detection of growth, which can serve as a starting point for discussing preventive aneurysm treatment.

    Abstract

    Importance  Unruptured intracranial aneurysms not undergoing preventive endovascular or neurosurgical treatment are often monitored radiologically to detect aneurysm growth, which is associated with an increase in risk of rupture. However, the absolute risk of aneurysm rupture after detection of growth remains unclear.

    Objective  To determine the absolute risk of rupture of an aneurysm after detection of growth during follow-up and to develop a prediction model for rupture.

    Design, Setting, and Participants  Individual patient data were obtained from 15 international cohorts. Patients 18 years and older who had follow-up imaging for at least 1 untreated unruptured intracranial aneurysm with growth detected at follow-up imaging and with 1 day or longer of follow-up after growth were included. Fusiform or arteriovenous malformation-related aneurysms were excluded. Of the 5166 eligible patients who had follow-up imaging for intracranial aneurysms, 4827 were excluded because no aneurysm growth was detected, and 27 were excluded because they had less than 1 day follow-up after detection of growth.

    Exposures  All included aneurysms had growth, defined as 1 mm or greater increase in 1 direction at follow-up imaging.

    Main Outcomes and Measures  The primary outcome was aneurysm rupture. The absolute risk of rupture was measured with the Kaplan-Meier estimate at 3 time points (6 months, 1 year, and 2 years) after initial growth. Cox proportional hazards regression was used to identify predictors of rupture after growth detection.

    Results  A total of 312 patients were included (223 [71%] were women; mean [SD] age, 61 [12] years) with 329 aneurysms with growth. During 864 aneurysm-years of follow-up, 25 (7.6%) of these aneurysms ruptured. The absolute risk of rupture after growth was 2.9% (95% CI, 0.9-4.9) at 6 months, 4.3% (95% CI, 1.9-6.7) at 1 year, and 6.0% (95% CI, 2.9-9.1) at 2 years. In multivariable analyses, predictors of rupture were size (7 mm or larger hazard ratio, 3.1; 95% CI, 1.4-7.2), shape (irregular hazard ratio, 2.9; 95% CI, 1.3-6.5), and site (middle cerebral artery hazard ratio, 3.6; 95% CI, 0.8-16.3; anterior cerebral artery, posterior communicating artery, or posterior circulation hazard ratio, 2.8; 95% CI, 0.6-13.0). In the triple-S (size, site, shape) prediction model, the 1-year risk of rupture ranged from 2.1% to 10.6%.

    Conclusion and Relevance  Within 1 year after growth detection, rupture occurred in approximately 1 of 25 aneurysms. The triple-S risk prediction model can be used to estimate absolute risk of rupture for the initial period after detection of growth.

    Introduction

    Unruptured intracranial aneurysms (UIAs) occur in approximately 3% of the adult population.1 The number of incidentally discovered UIAs has increased owing to the rising availability and use of brain imaging.2 Most UIAs are smaller than 7 mm and have a low risk of rupture.3 Preventive endovascular or neurosurgical aneurysm treatment is often considered, but these treatment options carry a substantial risk of procedure-related complications.4 In many patients with small aneurysms, the aneurysm remains untreated because the risk of treatment complications is estimated to be higher than the risk of rupture. Most of these patients are monitored over time with repeated magnetic resonance angiography or computed tomography angiography to detect aneurysm growth. If aneurysm growth is detected, the aneurysm is often treated. This policy is based on observations that aneurysms with detected growth have a higher risk of rupture than aneurysms without growth.5-10 Although the relative risk of rupture after detection of growth is increased, the absolute risk of rupture after detection of aneurysm growth is unknown. This absolute risk is pivotal for consultation with patients after detection of growth of the aneurysm, because the risk of rupture needs to be weighed against the risk of treatment complications. The purpose of this multicenter cohort study was to determine the absolute risk of rupture of an aneurysm after detection of growth on follow-up imaging and to develop a prediction model for rupture based on predictors of rupture.

    Methods
    Design and Population

    In this retrospective cohort study, individual patient data were obtained for patients with at least 1 UIA with growth detected on follow-up imaging from 15 cohorts in the following countries: Canada (Toronto), China (Hong Kong), Germany (Mannheim), Japan (Akita, Toon City, and Tokyo [3 cohorts]), Finland (Kuopio and Turku), the Netherlands (Amsterdam, Nijmegen, Utrecht, and the Hague), and the United Kingdom (Glasgow). This study followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline. Approval for data collection was obtained from the institutional research ethics board of all participating centers. The need for informed consent was waived and granted by the regulatory authority because data were deidentified.

    We included cohorts with consecutive patients 18 years and older with at least 1 untreated UIA with growth detected at follow-up imaging, and with 1 day or longer follow-up after growth. Growth was defined as 1 mm or greater increase in 1 direction of the UIA on magnetic resonance angiography, computed tomography angiography, or digital subtraction angiography.11 Exclusion criteria were fusiform or arteriovenous malformation-related aneurysm and no indication for follow-up imaging according to the treating physician, for example because of comorbidity, advanced age, or patient preference.

    Measurements

    From medical records, including radiology reports, the following data were collected by local investigators: age, sex, history of subarachnoid hemorrhage, smoking status at time of detection of growth, maximum aneurysm size before growth, date of last imaging before growth, date of imaging that detected growth, maximum aneurysm size at imaging that detected growth, aneurysm shape, aneurysm site, treatment after growth during follow-up, date of treatment after growth, rupture after growth, date of rupture after growth, date of last imaging after growth and date of last follow-up visit after growth. Geographic regions were subdivided into 3 groups: Japan, Finland, and North-America, Europe (other than Finland) and China. These categories were based on a previous pooled analysis of 6 prospective cohort studies, which showed an increase in risk of rupture among Japanese and Finnish patients.3 Irregular aneurysm shape was defined as the presence of blebs, aneurysm wall protrusions, daughter sacs, or multiple lobes.12 Aneurysm size was measured on a 0.1-mm scale by a radiologist or neuroradiologist at the participating center as part of standard clinical care. Measurements were made using electronic calibers on time-of-flight magnetic resonance angiography scans, computed tomography angiography scans, and rotational digital subtraction angiography scans and calibrated calipers on 2-dimensional digital subtraction angiography scans. Next, aneurysm morphology was determined by visual assessment by the radiologist or neuroradiologist of the participating center. The local investigator retrospectively collected these aneurysm measurements from the radiology report for the purpose of this study. Growth factor was calculated by dividing the maximum aneurysm size at imaging that detected growth by the maximum size before growth. Based on the PHASES score, aneurysm site was subdivided into 3 categories: internal carotid artery (excluding posterior communicating artery), middle cerebral artery, and anterior cerebral arteries, posterior communicating artery, or posterior circulation.3

    The primary outcome was rupture of an aneurysm after detection of growth. The period of observation for each UIA after detection of growth was defined as the period between date of imaging that detected growth and date of rupture, date of treatment, or date of last patient contact (last checked in January 2020). Vascular imaging was conducted in the context of standard follow-up in accordance with local institutional guidelines. The decision to recommend aneurysm treatment or watchful waiting to patients with UIAs was determined by the local multidisciplinary team.

    Statistical Analysis

    Analyses were aneurysm based. The Kaplan-Meier estimate was used to determine the absolute risk of aneurysm rupture at 6 months, 1 year, and 2 years. Patients were monitored until aneurysm rupture, treatment, or last patient contact. A stratified analysis was performed to investigate potential sex differences in risk of rupture. The impact of the potential predictors on rupture was analyzed with univariable and multivariable Cox proportional hazards models. Potential predictors of rupture were age, sex, previous subarachnoid hemorrhage from another aneurysm, smoking status at time of detection of growth, geographic region, aneurysm size at follow-up imaging that detected growth, growth factor, aneurysm shape at follow-up imaging that detected growth, and aneurysm site. The full model was simplified by exclusion of potential predictors with a backward selection procedure based on a P value less than .20. The log minus log plot for each predictor was visually checked for possible deviations from the assumption of proportional hazards. Results were reported as hazard ratios (HRs) with 95% CIs. Kaplan-Meier curves were used to examine the absolute risk of rupture for each selected predictor. To test the equality of the survival distributions for the different levels of each predictor, the log-rank test pooled over strata was used. A P value less than .05 indicated statistical significance.

    In additional analyses, we built a prediction model. As Cox proportional hazard models are known to overestimate predicted risks when applying the model to new patients, shrinkage was applied using Ridge regression. The amount of shrinkage was based on the full model with all potential predictors to reflect the selection of predictors. Regression coefficients from this analysis were used to estimate 6-month, 1-year, and 2-year risks of rupture. The discriminative ability of this model was estimated with the C statistic, which indicates to what extent the model could distinguish between aneurysms that did and did not rupture during follow-up.13

    Two variables showed missing values: age at time of growth was missing for 1 patient and smoking status for 51 patients. Because data were missing randomly, we used multiple imputation to determine the impact of these variables on the risk of rupture. As we did not include these predictors in the final analysis, all analyses were based on complete data. Statistical analyses were performed in SPSS version 25.0 (IBM) and R version 4.0.2 (The R Foundation).

    Results

    We screened 5166 patients who had follow-up imaging for 6928 untreated UIAs. Of these, 4827 patients were excluded because no aneurysm growth was detected during follow-up and 27 patients were excluded because they had less than 1 day follow-up after detection of growth. The reasons for less than 1 day follow-up for these 27 patients were: 13 patients had no clinical follow-up after growth, 8 patients had treatment on the same day of aneurysm growth detection, and 6 patients had rupture on the same day that growth was detected. As a result, we included 312 patients with 329 aneurysms with detected growth. The total follow-up after growth was 864 aneurysm-years (median, 1.3 years; interquartile range [IQR], 0.4-4.0). Patient and aneurysm characteristics at baseline are shown in Table 1. The mean (SD) age at time of detection of aneurysm growth was 61 (12) years, and 223 patients (71%) were women. Aneurysm growth was 1 to 2 mm in 200 aneurysms (61%), ≥2 to 3 mm in 63 aneurysms (19%), and ≥3 mm in 66 aneurysms (20%). After detection of growth, 134 aneurysms (41%) in 128 patients (41%) were treated after a median (IQR) of 0.4 (0.2-0.8) years. The median (IQR) absolute growth of the treated aneurysms was 1.8 (1.2-3.0) mm compared with 1.5 (1.1-2.1) mm of aneurysms that were not treated after detection of growth. More detailed information on treated and untreated aneurysms is shown in Table 2.

    Aneurysm Rupture

    The number of aneurysms that were monitored and at risk of rupture was 226 at 6 months, 179 at 1 year, and 153 at 2 years. Aneurysm rupture after detection of growth occurred in 25 aneurysms (7.6%) in 24 patients after a median (IQR) of 1.1 (0.1-3.0) years. The absolute cumulative risks of rupture after growth were 2.9% (95% CI, 0.9-4.9) at 6 months, 4.3% (95% CI, 1.9-6.7) at 1 year, and 6.0% (95% CI, 2.9-9.1) at 2 years. The risk of rupture declined over time: 0.48% per month (95% CI, 0.15-0.82) during the first 6 months, 0.23% per month (95% CI, 0.17-0.30) between 7 and 12 months, and 0.14% per month (95% CI, 0.08-0.40) between 12 and 24 months. After 6 years of follow-up, 45 aneurysms were still being monitored and none of these ruptured during the remaining 111 aneurysm-years of follow-up. Stratified analyses in male and female individuals did not show sex differences in risk of rupture.

    In multivariable analysis, predictors of rupture were aneurysm size (7 mm or larger HR, 3.1; 95% CI, 1.4-7.2), aneurysm shape (irregular HR, 2.9; 95% CI, 1.3-6.5), and aneurysm site (middle cerebral artery HR, 3.6; 95% CI, 0.8-16.3; anterior cerebral artery, posterior communicating artery, or posterior circulation HR, 2.8; 95% CI, 0.6-13.0) (Table 3). Kaplan-Meier curves are shown in the Figure. The shrinked HRs of the selected predictors were used for the final prediction model (Table 4). This triple-S (aneurysm size, site, and shape) risk prediction model estimated 6-month, 1-year, and 2-year risks of rupture for an aneurysm after detection of growth. In the triple-S risk prediction model, the 1-year risk of rupture ranged from 2.1% to 10.6%. The C index of this model was 0.72 (95% CI, 0.62-0.82).

    Discussion

    In this cohort study, within 1 year after detection of growth, rupture occurred in approximately 1 of 25 aneurysms. With the triple-S risk prediction model, based on 3 easily retrievable predictors (aneurysm size, site, and shape), absolute risks of rupture could be estimated for the initial period after detection of growth of an aneurysm, with risks in the first year varying from 2.1% to 10.6%.

    The high risk of rupture we found for aneurysms after detection of growth is in line with previous studies in which a higher risk of rupture was found for aneurysms with growth detected at follow-up imaging compared with aneurysms without growth.5-9 These studies provided only relative risks, which are more difficult to interpret in clinical practice than absolute risks. Moreover, all previous studies were single-center studies with a limited number of aneurysms with growth, which precluded identification of predictors of rupture. In addition, there was substantial heterogeneity in the definition of growth across studies. With our study, we tried to overcome the limitations of previous studies. We can now estimate the absolute risks of rupture for the initial 2 years after detection of growth based on 3 easily retrievable predictors.

    Growth of a UIA is not a linear but rather a stochastic process.8,14-17 The high risk of rupture we found after growth may reflect a period of increased instability of the aneurysm wall, which initially leads to growth and then to rupture. However, after growth, the aneurysm may stabilize again; we found that aneurysms can remain stable without rupture up to 12 years after detection of growth. The fact that aneurysm growth is often not followed by rupture explains how small aneurysms can become large aneurysms over time. If a large aneurysm is detected at first-ever brain imaging, it most likely means this aneurysm had multiple episodes of instability with growth but without rupture in the past. This implies that not all growing aneurysms need preventive treatment after growth is observed. The triple-S prediction model can help determine which aneurysms should undergo preventive occlusion and for which aneurysms follow-up imaging can be continued.

    In current clinical practice, the most important factors that are taken into account for treatment decisions in patients with a UIA are the estimated risks of aneurysm rupture and treatment complications, patient anxiety, and life expectancy. For newly detected aneurysms, the PHASES score can be used to estimate the risk of aneurysm rupture.3 For aneurysms showing growth during follow-up imaging, the triple-S prediction model provides estimates of the risk of rupture in the initial period after growth detection. Comparing the results of our study with those of the PHASES score, for most aneurysm categories, the 1-year risk of rupture after growth detection is higher than the 5-year risk of rupture after aneurysm detection. This means that the indication for preventive aneurysm occlusion is stronger for an aneurysm with recently detected growth than for a similar aneurysm without detected growth. The triple-S prediction model also shows that for some aneurysms the absolute risk of rupture remains small after detection of growth, and careful weighing of the risk of rupture vs the risk of treatment complications is needed.

    Strengths and Limitations

    A strength of this study is that it included individual patient data pooled from 15 international cohorts. We thereby had sufficient numbers of patients with aneurysm rupture after detection of growth to assess multiple risk factors for growth. In all included cohorts, growth was defined according to the accepted standards recommended by the National Institutes of Health Common Data Elements Project on Unruptured Intracranial Aneurysms and Subarachnoid Hemorrhage.11 Furthermore, we provided the absolute risks of rupture for 3 time points in the initial 2 years after detection of aneurysm growth. Based on the risk factors we identified, we were able to make a clinically applicable prediction model for rupture risk estimation after detection of growth of an aneurysm. Patients were included from cohorts with low- and high-risk populations from 7 countries and 3 continents, which adds to the external validity of our findings.

    A few study limitations need to be addressed. First, this study cannot be regarded as a natural history study of risk of rupture after growth of an aneurysm because of the inevitable selection that has occurred in the included cohorts. Patients’ decisions whether or not to undergo preventive treatment, follow-up imaging, or neither determines the selection of patients with follow-up imaging. From those with growth detected at follow-up imaging, many would have had preventive aneurysm treatment shortly after detection of aneurysm growth, leading to selection bias toward patients with lower risks of rupture. However, this type of bias is unavoidable in observational studies. Second, the indication for and timing of follow-up imaging were decided by the treating physicians of the local institutions and therefore varied between cohorts and patients. In patients with a longer time interval between 2 scans, potential growth in an early phase of this interval would result in a longer interval prior to the next follow-up scan. Thus the risk of rupture would have already decreased after detection of growth. Third, we did not perform a sample size calculation for this study, as reasonable estimates for a calculation were not known prior to this study. In a post hoc power analysis, we evaluated the number of ruptures needed to detect increasing HRs from 1.3 to 2.0 with 0.80 power. This analysis (eAppendix in the Supplement) suggests that with the 25 ruptures observed in this study, an HR of 1.65 may be detected. As a result, any predictor with a lower HR may have been excluded from the model. Fourth, it is possible that we missed some outcome events, because rupture could have resulted in sudden death without diagnosis of subarachnoid hemorrhage or presentation in another hospital without informing the treating physician of the participating center. Because of these limitations, the presented risks are probably an underestimation of the risk of rupture after growth. Fifth, to our knowledge, no guidelines previously existed for optimal UIA measurement methods, and no optimal threshold value for aneurysms growth based on optimal reliability and agreement measures is available. Therefore, we applied thresholds for aneurysm growth based on the best available evidence in the literature and used the definition of growth of 1 mm as suggested by the National Institutes of Health Common Data Elements Project on Unruptured Intracranial Aneurysms and Subarachnoid Hemorrhage, which was based on consensus between aneurysm experts.11

    Conclusions

    The implications for clinical practice from our study are that preventive endovascular or neurosurgical aneurysm treatment should be reconsidered as soon as aneurysm growth is detected. In such instances of aneurysm growth, our triple-S prediction model can be used by physicians and patients as a starting point for discussing the pros and cons of preventive aneurysm treatment. If it is decided to continue follow-up imaging, it seems reasonable to repeat imaging at a short interval, but actual data on the optimal time interval are lacking and should be gathered in future studies. Future data collection and studies are also needed to validate the prediction model.

    Back to top
    Article Information

    Accepted for Publication: July 9, 2021.

    Published Online: August 30, 2021. doi:10.1001/jamaneurol.2021.2915

    Correction: This article was corrected on January 10, 2022, to fix the labels in the Figure, panel D.

    Corresponding Author: Laura T. van der Kamp, MD, Department of Neurology and Neurosurgery, University Medical Center Utrecht Brain Center, Utrecht University, University Medical Center Utrecht, Room G3-201, Postbox 85500, 3508 GA Utrecht, the Netherlands (l.t.vanderkamp@umcutrecht.nl).

    Author Contributions: Drs van der Kamp and Vergouwen had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

    Concept and design: van der Kamp, Rinkel, Zuithoff, Vergouwen.

    Acquisition, analysis, or interpretation of data: All authors.

    Drafting of the manuscript: van der Kamp, Vergouwen.

    Critical revision of the manuscript for important intellectual content: All authors.

    Statistical analysis: van der Kamp, Zuithoff.

    Obtained funding: Vergouwen.

    Administrative, technical, or material support: van der Kamp, Vandertop, Lindgren, Teo, Igase, Rahi, Abrigo, Etminan.

    Supervision: Rinkel, van der Schaaf, Vergouwen.

    Conflict of Interest Disclosures: Dr van den Berg reports fees from Cerenovus paid to his institution outside the submitted work. Dr Boogaarts reports fees for serving as a consultant for Stryker Neurovascular outside the submitted work. No other disclosures were reported.

    Funding/Support: This study was supported by clinical established investigator grant 2018T076 from the Dutch Heart Foundation paid to Dr Vergouwen.

    Role of the Funder/Sponsor: The funder had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

    References
    1.
    Vlak  MHM, Algra  A, Brandenburg  R, Rinkel  GJE.  Prevalence of unruptured intracranial aneurysms, with emphasis on sex, age, comorbidity, country, and time period: a systematic review and meta-analysis.   Lancet Neurol. 2011;10(7):626-636. doi:10.1016/S1474-4422(11)70109-0PubMedGoogle ScholarCrossref
    2.
    Gabriel  RA, Kim  H, Sidney  S,  et al.  Ten-year detection rate of brain arteriovenous malformations in a large, multiethnic, defined population.   Stroke. 2010;41(1):21-26. doi:10.1161/STROKEAHA.109.566018PubMedGoogle ScholarCrossref
    3.
    Greving  JP, Wermer  MJH, Brown  RD  Jr,  et al.  Development of the PHASES score for prediction of risk of rupture of intracranial aneurysms: a pooled analysis of six prospective cohort studies.   Lancet Neurol. 2014;13(1):59-66. doi:10.1016/S1474-4422(13)70263-1PubMedGoogle ScholarCrossref
    4.
    Algra  AM, Lindgren  A, Vergouwen  MDI,  et al.  Procedural clinical complications, case-fatality risks, and risk factors in endovascular and neurosurgical treatment of unruptured intracranial aneurysms: a systematic review and meta-analysis.   JAMA Neurol. 2019;76(3):282-293. doi:10.1001/jamaneurol.2018.4165PubMedGoogle ScholarCrossref
    5.
    Villablanca  JP, Duckwiler  GR, Jahan  R,  et al.  Natural history of asymptomatic unruptured cerebral aneurysms evaluated at CT angiography: growth and rupture incidence and correlation with epidemiologic risk factors.   Radiology. 2013;269(1):258-265. doi:10.1148/radiol.13121188PubMedGoogle ScholarCrossref
    6.
    Mehan  WA  Jr, Romero  JM, Hirsch  JA,  et al.  Unruptured intracranial aneurysms conservatively followed with serial CT angiography: could morphology and growth predict rupture?   J Neurointerv Surg. 2014;6(10):761-766. doi:10.1136/neurintsurg-2013-010944PubMedGoogle ScholarCrossref
    7.
    Inoue  T, Shimizu  H, Fujimura  M, Saito  A, Tominaga  T.  Annual rupture risk of growing unruptured cerebral aneurysms detected by magnetic resonance angiography.   J Neurosurg. 2012;117(1):20-25. doi:10.3171/2012.4.JNS112225PubMedGoogle ScholarCrossref
    8.
    Chien  A, Callender  RA, Yokota  H,  et al.  Unruptured intracranial aneurysm growth trajectory: occurrence and rate of enlargement in 520 longitudinally followed cases.   J Neurosurg. 2019;132(4):1077-1087. doi:10.3171/2018.11.JNS181814PubMedGoogle ScholarCrossref
    9.
    Korja  M, Lehto  H, Juvela  S.  Lifelong rupture risk of intracranial aneurysms depends on risk factors: a prospective Finnish cohort study.   Stroke. 2014;45(7):1958-1963. doi:10.1161/STROKEAHA.114.005318PubMedGoogle ScholarCrossref
    10.
    Brinjikji  W, Zhu  YQ, Lanzino  G,  et al.  Risk factors for growth of intracranial aneurysms: a systematic review and meta-analysis.   AJNR Am J Neuroradiol. 2016;37(4):615-620. doi:10.3174/ajnr.A4575PubMedGoogle ScholarCrossref
    11.
    Hackenberg  KAM, Algra  A, Salman  RAS,  et al; Unruptured Aneurysms and SAH CDE Project Investigators.  Definition and prioritization of data elements for cohort studies and clinical trials on patients with unruptured intracranial aneurysms: proposal of a multidisciplinary research group.   Neurocrit Care. 2019;30(suppl 1):87-101. doi:10.1007/s12028-019-00729-0PubMedGoogle ScholarCrossref
    12.
    Backes  D, Vergouwen  MDI, Velthuis  BK,  et al.  Difference in aneurysm characteristics between ruptured and unruptured aneurysms in patients with multiple intracranial aneurysms.   Stroke. 2014;45(5):1299-1303. doi:10.1161/STROKEAHA.113.004421PubMedGoogle ScholarCrossref
    13.
    Harrell  FE,.  Regression Modeling Strategies. Springer International Publishing; 2015. doi:10.1007/978-3-319-19425-7
    14.
    Jou  LD, Mawad  ME.  Growth rate and rupture rate of unruptured intracranial aneurysms: a population approach.   Biomed Eng Online. 2009;8:11. doi:10.1186/1475-925X-8-11PubMedGoogle ScholarCrossref
    15.
    Chang  HS.  Simulation of the natural history of cerebral aneurysms based on data from the International Study of Unruptured Intracranial Aneurysms.   J Neurosurg. 2006;104(2):188-194. doi:10.3171/jns.2006.104.2.188PubMedGoogle ScholarCrossref
    16.
    Koffijberg  H, Buskens  E, Algra  A, Wermer  MJH, Rinkel  GJE.  Growth rates of intracranial aneurysms: exploring constancy.   J Neurosurg. 2008;109(2):176-185. doi:10.3171/JNS/2008/109/8/0176PubMedGoogle ScholarCrossref
    17.
    Etminan  N, Rinkel  GJ.  Unruptured intracranial aneurysms: development, rupture and preventive management.   Nat Rev Neurol. 2016;12(12):699-713. doi:10.1038/nrneurol.2016.150PubMedGoogle ScholarCrossref
    ×