Article Text

Download PDFPDF
Original research
Prevalence of unruptured intracranial aneurysms: impact of different definitions – the Tromsø Study
  1. Liv-Hege Johnsen1,2,
  2. Marit Herder1,2,
  3. Torgil Vangberg2,3,
  4. Roar Kloster2,4,
  5. Tor Ingebrigtsen2,4,
  6. Jørgen Gjernes Isaksen2,4,
  7. Ellisiv B Mathiesen2,5
  1. 1Department of Radiology, University Hospital of North Norway, Tromso, Norway
  2. 2Department of Clinical Medicine, UiT The Arctic University, Tromso, Norway
  3. 3PET Imaging Center, University Hospital of North Norway, Tromso, Norway
  4. 4Department of Neurosurgery, Ophthalmology, and Otorhinolaryngology, University Hospital of North Norway, Tromso, Norway
  5. 5Department of Neurology, University Hospital of North Norway, Tromso, Norway
  1. Correspondence to Liv-Hege Johnsen, Department of Radiology, University Hospital of North Norway, Tromso, Norway; liv.hege.johnsen{at}


Background Management of incidental unruptured intracranial aneurysms (UIAs) remains challenging and depends on their risk of rupture, estimated from the assumed prevalence of aneurysms and the incidence of aneurysmal subarachnoid haemorrhage. Reported prevalence varies, and consistent criteria for definition of UIAs are lacking. We aimed to study the prevalence of UIAs in a general population according to different definitions of aneurysm.

Methods Cross-sectional population-based study using 3-dimensional time-of-flight 3 Tesla MR angiography to identify size, type and location of UIAs in 1862 adults aged 40–84 years. Size was measured as the maximal distance between any two points in the aneurysm sac. Prevalence was estimated for different diameter cutoffs (≥1, 2 and 3 mm) with and without inclusion of extradural aneurysms.

Results The overall prevalence of intradural saccular aneurysms ≥2 mm was 6.6% (95% CI 5.4% to 7.6%), 7.5% (95% CI 5.9% to 9.2%) in women and 5.5% (95% CI 4.1% to 7.2%) in men. Depending on the definition of an aneurysm, the overall prevalence ranged from 3.8% (95% CI 3.0% to 4.8%) for intradural aneurysms ≥3 mm to 8.3% (95% CI 7.1% to 9.7%) when both intradural and extradural aneurysms ≥1 mm were included.

Conclusion Prevalence in this study was higher than previously observed in other Western populations and was substantially influenced by definitions according to size and extradural or intradural location. The high prevalence of UIAs sized <5 mm may suggest lower rupture risk than previously estimated. Consensus on more robust and consistent radiological definitions of UIAs is warranted.


Data availability statement

Data are available on reasonable request. Data may be obtained from a third party and are not publicly available. Data availability are restricted due to their sensitive nature. Deidentified data can be obtained by application to the Tromsø Study, please contact for details.

Statistics from


  • The prevalence of unruptured intracranial aneurysms (UIAs) in previous population-based studies varied from 1.9% to 7.0%. Definition of aneurysm was not consistent between studies.


  • The prevalence of intradural UIAs ≥2 mm was 6.6 %, which is higher than in previous studies of Western populations. The prevalence ranged from 3.8% to 8.3%, depending on definition according to size and extradural or intradural location.


  • A uniform definition of UIAs is necessary for comparison of prevalence across studies. The high prevalence of UIAs sized <5 mm may suggest lower rupture risk than previously estimated. This supports increased restraint with prophylactic repair of such lesions.


Aneurysmal subarachnoid haemorrhage (aSAH) is a life-threatening subtype of stroke which affects relatively young people and has a high case fatality and morbidity.1 2 With improved availability and utilisation of advanced neuroimaging such as CT angiography and MR angiography (MRA), unruptured intracranial aneurysms (UIAs) are found more frequently. Counselling and shared decision-making for patients with an incidentally detected UIA depend on the aneurysm’s rupture risk, estimated from the prevalence of UIAs and the incidence of aSAH in the population. In addition, individual risk factors for aSAH and treatment related complications must be considered.

The reported prevalence of UIAs varies with study design, study population and aneurysm characteristics.3 In a systematic review and meta-analysis of 83 studies published before 2011, the overall prevalence in a population without comorbidity, mean age of 50 years and 50% men, was estimated to 3.2% (95% CI 1.9% to 5.2%).4 Since then, three cross-sectional population-based angiography studies have been published. The Nord-Trøndelag Health (HUNT) Study in Norway reported a prevalence of 1.9% in 1006 subjects aged 50–65 years,5 while the Rotterdam Scan Study reported a prevalence of 2.3% in 5800 participants with a mean age of 64.9 years.6 A corresponding study of 4813 participants aged 35 to 75 years in Shanghai, China, found an overall prevalence of 7.0%.7

While intracranial aneurysms are defined as abnormal focal dilations of an artery in the brain, the exact definition based on size and intradural or extradural location may vary between studies. Both differences in definitions and the use of imaging techniques could influence the detection rate. In this study, we aimed to determine the prevalence of UIAs in a general Norwegian population aged 40 years and older and explore how the definition of UIA influences prevalence estimates.

Material and methods

Study participants

The Tromsø Study is a large longitudinal and multipurpose population health study of adult inhabitants in the municipality of Tromsø, Norway, started in 1974.8 9 The study design is repeated cross-sectional surveys to which total birth cohorts and selected samples are invited. The seventh survey was conducted in 2015–2016 and consisted of two visits. The recruitment was based on the official population registry. All citizens aged 40 years and older (n=32 591) were invited to participate in the first visit, and 21 083 (64.7%) attended.

A subsample consisting of 75% who were randomly drawn from 10-year age-bands in men and women and of 25% who had participated in the second visit of the sixth Tromsø Study was invited to a second visit for various extended clinical examinations, and 8346 (90.2%) of the 9253 invited attended. Due to limited capacity for MRI, we chose to invite the 3027 participants who had also been examined with ultrasound of the carotid arteries to participate in an MRI study conducted in 2016–2017. Of these, 975 did not attend, 169 were excluded because of contraindications for MRI and 5 were excluded because they had moved or died before the MRI examination. Of the 1878 subjects who underwent MRA scanning, images were missing in eight participants and image quality was deemed as insufficient for assessment of UIAs in seven participants. One participant later withdrew the consent and was excluded, leaving 1862 participants who were included in the final analysis (online supplemental figure 1).

Supplemental material

Cardiovascular risk factors

Information on smoking and drinking habits, medication use or prevalent cardiovascular disease such as history of stroke, coronary heart disease and diabetes was obtained from questionnaires. Standardised measurements of blood pressure, height and weight were obtained by trained personnel. Serum total cholesterol and high-density lipoprotein (HDL) cholesterol were analysed by standard enzymatic methods. Body mass index (BMI) was calculated as weight divided by height squared (kg/m2). Hypertension was defined as mean systolic blood pressure ≥140 mm Hg and/or mean diastolic blood pressure ≥90 mm Hg and/or use of blood pressure-lowering medication and/or self-reported hypertension.


Participants were scanned at the University Hospital of North Norway, Tromsø, with a 3 Tesla (3T) Siemens Skyra MR scanner (Siemens Healthcare, Erlangen, Germany). We used the three-dimensional time-of-flight MRA sequence (3D-TOF-MRA) for detection of aneurysms and made maximum intensity projection (MIP) reconstructions, volume-rendered (VR) reconstructions and, when needed, additional single arterial segmentation to optimise the evaluation of aneurysm morphology from the source images. Aneurysms were measured on multiplanar reconstructions (MPR). Details about the technique are provided in online supplemental material.

Aneurysm definition

All visible abnormal focal saccular dilatations were measured from centre of the neck plane to the dome apex and then perpendicular to this axis. Additionally, any diameter between any two points in the aneurysm sac that was larger than these two diameters was recorded, and the maximum diameter defined the aneurysm size. Fusiform aneurysms (n=5) were excluded.

We categorised aneurysm locations on the internal carotid artery (ICA) according to the Bouthiller classification.10 Aneurysms located in and distal to the clinoid segment (C5) were defined as intradural. The C7-segment of the ICA was further subdivided into the posterior communicating artery, the anterior choroid artery and terminus. We categorised UIAs located to the anterior cerebral artery (ACA) as A1-segment, anterior communicating artery and distal ACA, the middle cerebral artery (MCA) as M1-segment, bifurcation of M1 and M2-segment, the posterior cerebral artery as P1-segment and P2-segment and vertebrobasilar arteries as basilar artery, superior cerebellar artery, posterior inferior cerebellar artery and anterior inferior cerebellar artery. The sizes were categorised as ≥1–1.9 mm, ≥2–2.9 mm, ≥3–4.9 mm, ≥5–6.9 mm, ≥7 mm.

Intraobserver and interobserver agreement

Details about the assessment of aneurysms and intraobserver and interobserver agreement are provided in the Supplemental material. All images and reconstructions were evaluated twice by one experienced neuroradiologist (L-HJ) and all aneurysms were measured by two experienced neuroradiologists (L-HJ and MH). Interobserver discrepancies in measurements of UIAs were resolved by consensus between the two observers.

The Cohen’s kappa for interobserver reliability on detecting aneurysms was 0.79 (SE 0.10). The interobserver intraclass correlation coefficient (ICC) for measurement of the size of the aneurysms was 0.93. The intraobserver ICC for the first reader was 0.98.

Clinical follow-up

All participants with vascular pathological findings judged to be clinically relevant were referred to the vascular neurosurgical outpatient clinic for clinical follow-up and management according to department policies. Similarly, incidental clinically relevant non-vascular pathological findings were referred to the other specialist clinics at the University Hospital of North Norway, Tromsø.

Statistical analysis

The data were analysed with Stata for Mac (V.17: StataCorp). Continuous variables are presented as means with 95% CI and categorical variables as percentages (95% CI). For the 1862 participants baseline variables were missing for BMI (n=1), systolic blood pressure (n=6), smoking status (n=18), total cholesterol and HDL cholesterol (n=7), use of lipid-lowering medication (n=44), diabetes mellitus (n=33), history of stroke (n=73), history of coronary heart disease (n=1) and alcohol intake (n=12).

Differences between groups were evaluated using a two-sample t-test for summary data, and z-test was used to compare proportions in one sample.

In the primary analysis of prevalence in the population, we defined an aneurysm as an abnormal focal saccular dilatation of an intradural cerebral artery with a diameter ≥2 mm. We further explored how altered definitions for size (≥1.0 mm, ≥3 mm and ≥5 mm) and/or inclusion of UIAs located extradurally influenced the estimated prevalence.

We followed the STrengthening the Reporting of OBservational studies in Epidemiology statement for reporting.


Characteristics of participants with and without aneurysms are shown in table 1. The mean age of participants was 63.8 years and 52.6% were women. The proportion of participants with hypertension was larger in participants with UIAs than in those without, although not significant. The results were similar when the threshold for systolic blood pressure was set at ≥130 mm Hg. The proportions who reported alcohol intake more than four times per week, had diabetes and were current smokers were larger in participants with UIAs than in those without, but not significant. This did not change after age adjustment and sex adjustment.

Table 1

Characteristics of participants with and without unruptured intradural saccular aneurysms ≥2 mm

A total of 131 intradural saccular aneurysms≥2 mm was detected in 122 of the 1862 participants (table 1), which gives a prevalence of intradural UIAs ≥2 mm of 6.6% (95% CI 5.4% to 7.6%). Three participants had three UIAs and three had two UIAs.

The overall and age-specific and sex-specific prevalences of UIAs are shown in table 2. The prevalence was 7.5% (95% CI 5.9% to 9.2%) in women and 5.5% (95% CI 4.1% to 7.2%) in men. In women, prevalence was 4.5% (95% CI 2.1% to 8.1%) in the 40–54 years age band, 8.2% (95% CI 5.2% to 12.1%) in the 55–64 years age band, 8.3% (95% CI 5.7% to 11.7%) in the 65–74 years age band and 8.6% (95% CI 4.5% to 15.6%) in the 75–84 years age band. In men, prevalence was 5.1% (95% CI 2.2% to 9.8%) in the youngest age band, 5.5% (95% CI 3.0% to 9.2%) in the 55–64 years age band, 6.5% (95% CI 4.0% to 9.7%) in the 65–74 years age band and 3.9% (95% CI 1.4% to 8.3%) in those aged 75–84 years. There was no consistent difference in size of aneurysms between men and women (online supplemental table 1).

Table 2

Prevalence (%) of unruptured intradural saccular aneurysms ≥2 mm stratified by age group and sex

Fifty-six of the 131 UIAs (42.7%) were in the ICA, 14 (10.7%) in the ACA, 50 (38.2%) in the MCA and 11 in the posterior circulation (8.4%) (table 3).

Table 3

Location of unruptured intradural saccular aneurysms ≥2 mm in men and women

The overall prevalence according to different definitions of UIAs was 3.8% (95% CI 3.0% to 4.8%) for intradural aneurysms ≥3 mm, 7.4% (95% CI 6.2% to 8.6%) when the cut-off was set at ≥1 mm and 8.3% (95% CI 7.1% to 9.7%) when both intradural and extradural aneurysms were included (table 4, figure 1). Online supplemental figure 2 shows the distribution of all intradural UIAs by size.

Figure 1

Prevalence of intradural unruptured intracranial aneurysms (UIAs) according to different definitions by size, stratified by age group. The Tromsø study.

Table 4

Observed unruptured aneurysms (UIAs) by size and location (intradural only and intradural and extradural)


In this study, the overall prevalence of intradural saccular UIAs ≥2 mm was 6.6%. The prevalence changed substantially depending on definition, both concerning threshold for size and intradural or extradural location.

The prevalence of UIAs in our study is significantly higher than those reported in previous population-based European studies. The prevalence in the HUNT study5 and the Rotterdam Scan study6 was 1.9% (95% CI 1.2% to 2.9%) and 2.3% (95% CI 2.0% to 2.7%), respectively. Our results are closer to the prevalence of 7.0% (95% CI 6.3% to 7.7%) found in the Shanghai study.7 It is likely that the differences between the studies, at least partly, are explained by differences in imaging techniques. Both the HUNT study and the Rotterdam Scan study used 1.5T MRI, while the Shanghai study and this study used 3T (table 5). Sailer et al reviewed studies of the accuracy of MRA in diagnosing intracranial aneurysms and found that 3T performed better than imaging with lower field strengths.11 The higher resolution and improved signal/noise ratio provided by 3T MRA improve the detection rate, particularly for UIAs smaller than 3 mm and UIAs located in the carotid siphon.12–14

Table 5

Comparison of the four population-based studies on prevalence of unruptured intracranial aneurysms (UIAs) according to the number of participants, age range and age mean, MRI field strength, sequence and postprocessing techniques, definition of UIAs and prevalence of UIAs

Postprocessing techniques also influence the detection rate. VR-reconstruction of 3D-TOF-MRA enables dynamic viewing from all directions instead of fixed reconstructions in standard planes, improving accuracy and reducing false-positive detections.11 15 In the HUNT study,5 3D-TOF-MRA with evaluation of source images and fixed MIP was used, while proton density-weighted images without reconstruction of the images with volume rendering technique was used in the Rotterdam Scan study6 (table 5). This study and the Shanghai study7 used 3D-TOF-MRA with reconstructions in MIP and VR and used single-artery highlighting technique when needed.15

Differences in the age distribution may also have contributed, but cannot fully explain the variation in prevalence between studies. The HUNT study5 included participants aged 50–65 years (mean age 58.5 years) (table 5). The age distribution in our study was similar to that of the Rotterdam Scan study,6 which included participants aged 45 years and older (mean age 64.9 years). Despite a lower mean age (52.1 years) of participants in the Shanghai study,7 the prevalence of UIAs was similar to our study.

Different definitions of aneurysms could also explain the differences in prevalence estimates between studies. The Shanghai study used the same threshold for size (≥2 mm) as we did, but also included extradural UIAs located in the C3- and C4-segments of the ICA (table 5). The HUNT Study5 and the Rotterdam Scan Study6 did not provide exact definitions by size or intradural or extradural location. We chose not to include extradural UIAs in our main prevalence estimation because these aneurysms usually will not cause aSAH.16 UIAs in the C5-segment, on the other hand, were defined as intradural. This is because the location of the dural rings, and most importantly the distal dural ring, cannot be reliably determined with MRI; hence UIAs in the C5-segment of the ICA are potentially intradural.17

The mean maximum size of the UIAs was 3.8 mm in our study, compared with 3.5 mm in the Shanghai study,7 5.6 mm in HUNT5 and 4.5 mm in the Rotterdam Scan study.6 This range can probably be explained by the use of 3.0 T 3D-TOF-MRA and improved post-processing abilities. Approximately 80% of the UIAs in this study were smaller than 5 mm compared with 90% in the Shanghai study, and the proportions of UIAs with diameters below 10 mm were 99.2% and 99.5%, respectively.

This study indicates that aneurysms as small as 1.0 mm can be detected with 3T MRA.15 18 On the other hand, a very low size threshold increases the likelihood of false-positive findings due to limitations in the method regarding visualization of small vessels as they get narrower. Therefore, detection and categorisation of UIAs sized 1.0–1.9 mm should be interpreted with caution. In this study, we detected 19 intradural aneurysms that were 1.0–1.9 mm. 43% (56/131) of aneurysms were as small as 2.0–2.9 mm (online supplemental figure 2).

The prevalence increased with age in both women and men, except for the men in the oldest age group (75–84 years). It is unlikely that prevalent aneurysms regress in old age. Accordingly, the observed prevalence could be biased by the low sample size in this age-band, or by selection bias in the recruitment of old participants.

The PHASES (Population, Hypertension, Age, Size of aneurysm, Earlier SAH from another aneurysm and Site of aneurysm) score for prediction of rupture assumes that aneurysms sized <7.0 mm carries a very low rupture risk.19 Nevertheless, controversy exists about the management of UIAs <7.0 mm because they are a common cause of aSAH despite their low rupture risk.20 Studies that stratify the risk by location and size indicate that small aneurysms are not a homogeneous group. Aneurysms located in the posterior circulation and the posterior communicating artery,21 as well as aneurysms sized 4–6 mm located in the anterior communicating artery or the distal anterior artery, have higher rupture risk than small aneurysms in general.22 The assumed elevated risk of rupture according to these studies19–22 applies to 29 (22%) of the 131 UIAs observed in this study, 9 UIAs≥7 mm and 20 UIAs <7 mm (7 UIAs in the posterior communicating artery, 9 UIAs in the posterior circulation and 4 UIAs 4–6 mm in the anterior communicating artery).

In a retrospective study of aSAH from the northern Norway region23 the annual frequency of aSAH in the Tromsø population was 1.15 from aneurysms <5 mm and 1.84 from aneurysms≥5 mm. Based on this and the observed prevalence of UIA in this study, we estimated an annual risk of rupture of 0.07% for aneurysms <5 mm and 0.41% for aneurysms ≥5 mm. The low rupture risk of UIAs <5 mm supports restraint with prophylactic repair of such lesions.

Longitudinal follow-up studies and further analysis of aneurysm size and localisation in aSAH patients compared with the MRA screening group from the same general population might provide additional knowledge on size-specific and location-specific rupture risk.

In accordance with the Shanghai study,7 we found that most aneurysms were located to the ICA but with a significantly lower proportion (42.8% compared with 81% in the Shanghai study). The HUNT study5 and the Rotterdam Scan study6 reported 31.5% and 30% of the UIAS in the ICA, respectively. The inclusion of extradural UIAs in the Shanghai study and the lower field strength MRI technology in the HUNT and Rotterdam study probably contribute to this discrepancy. The proportion of UIAs in the ACA with branches was also comparable to the Shanghai study and lower than in the HUNT and Rotterdam studies.

There was a striking difference between our study and the Shanghai study with regards to the proportion of aneurysms located in the MCA (38.2% in our study compared with 4.1% in the Shanghai study) and, to a lesser extent, the proportion of aneurysms located in the posterior circulation (8.5% compared with 2.4%). The HUNT and Rotterdam studies reported proportions of MCA aneurysms comparable to those observed in our study, but a lower proportion of posterior circulation aneurysms. Other MRI-based prevalence studies from Asia report proportions of UIAs in the MCA ranging from 9% to 28.6%.12 24–28 The different location-distribution of UIAs between European and Asian population-based studies is interesting, especially the difference in the proportion of MCA aneurysms. This suggests that the distribution of UIAs is probably not the same in different populations, which should be considered when estimating rupture risk based on aneurysm location.

Strengths and limitations

The main strength of this study is the population-based study design, with a balanced age distribution. Another strength is the use of 3D-TOF-MRA performed on 3T MRI combined with postprocessing abilities, including VR-3D reconstruction, to detect UIAs.

Measurements of aneurysms on MRI can be both underestimated and overestimated. In an in vitro study from Japan, Takao et al found that measurement of aneurysm height was lower than in the actual model,29 and in VR 3D-TOF-MRA, there was an overestimation of the aneurysm neck.30 In our study, all measurements were done at MPR and not VR, as VR is more prone to errors. Furthermore, as the threshold for window width and window level was allowed to be changed when measuring UIAs, this could also lead to measurement errors, particularly for small aneurysms. Another limitation of 3D-TOF-MRA is the possibility of missing aneurysms with slow or turbulent flow due to loss of signal, particularly in large-sized aneurysms. Also, the analyses in this study were limited to aneurysm size and did not include other aneurysm related risk factors for rupture such as aneurysm morphology or growth.

The number of participants in this study was limited, resulting in relatively wide CIs for the age-specific and sex-specific prevalence estimates, especially for the oldest age bands. Twenty-five per cent of the participants were invited because of previous participation in the Tromsø Study, which may have led to some degree of selection bias. The mean age and the proportion of women were somewhat higher in non-attendees (mean age 65.0 years, 55.0% women) than in attendees of the MRI study (mean age 63.8 years, 53.2% women). Lower participation rate in the oldest age groups due to disease or disability may have led to underestimation of UIA prevalence in the elderly.

Another limitation is that we did not have information on family history or genetic conditions associated with increased prevalence of UIA. Finally, the findings may not apply to other populations and ethnic groups.


The overall prevalence of UIAs ≥2 mm in adults aged 40–84 years in our study was 6.6%. This is higher than previously observed in Western populations. Depending on definition by size and intradural or extradural location, the prevalence ranged from 3.8% for intradural aneurysm ≥3 mm to 8.3% for intradural and extradural aneurysms ≥1 mm. The high prevalence of UIAs sized <5 mm may suggest lower rupture risk than previously estimated. More robust and consistent radiological definitions of intracranial aneurysms are needed to compare results from different populations and to improve shared decision-making in management of individuals with incidental UIAs.

Data availability statement

Data are available on reasonable request. Data may be obtained from a third party and are not publicly available. Data availability are restricted due to their sensitive nature. Deidentified data can be obtained by application to the Tromsø Study, please contact for details.

Ethics statements

Patient consent for publication

Ethics approval

This study involves human participants and was approved by Regional Committee of Medical and Health Research Ethics North Norway ID 9702 (file no. 2014/1665) Participants gave informed consent to participate in the study before taking part.


We thank Heidi Johansen, general manager of the seventh wave of the Tromsø Study, and the MRI technicians at the Department of Radiology, University Hospital of North Norway, for their contribution to the study. We also would like to thank all participants of the study.



  • JGI and EBM contributed equally.

  • Contributors Study conception and design: EBM, JGI; data collection: L-HJ; Imaging analysis: L-HJ, MH; analysis and interpretation of the results: L-HJ, EBM, JGI, TI, TV; manuscript draft: L-HJ. All authors provided critical review, edited and approved the final manuscript. EBM is guarantor of the study.

  • Funding L-HJ received funding from the Northern Norway Regional Health Authority (Grant SFP 1283-16). The MRI/MRA study was funded by the Northern Norway Regional Health Authority (Grant SFP 1271-16).

  • Competing interests None declared.

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.