Article Text

Download PDFPDF
Original research
Cigarette smoking and risk of intracranial aneurysms in middle-aged women
  1. Christopher S Ogilvy1,
  2. Santiago Gomez-Paz1,
  3. Kimberly P Kicielinski1,
  4. Mohamed M Salem1,
  5. Yosuke Akamatsu1,
  6. Muhammad Waqas2,3,
  7. Hamid Hussain Rai2,3,
  8. Joshua S Catapano4,
  9. Sandeep Muram5,
  10. Mohammed Elghareeb6,
  11. Adnan H Siddiqui3,7,
  12. Elad I Levy3,7,
  13. Michael T Lawton4,
  14. Alim P Mitha5,
  15. Brian L Hoh6,
  16. Adam Polifka6,
  17. W Christopher Fox6,
  18. Justin M Moore1,
  19. Ajith J Thomas1
  1. 1 Division of Neurosurgery, Beth Israel Deaconess Medical Center, Harvard Medical School, Boston, Massachusetts, USA
  2. 2 Department of Neurosurgery, University at Buffalo School of Medicine and Biomedical Sciences, Buffalo, New York, United States
  3. 3 Department of Neurosurgery, Gates Vascular Institute, Buffalo, New York, United States
  4. 4 Department of Neurosurgery, Barrow Neurological Institute, St. Joseph's Hospital and Medical Center, Phoenix, Arizona, United States
  5. 5 Division of Neurosurgery, University of Calgary, Calgary, Alberta, Canada
  6. 6 Lillian S. Wells Department of Neurosurgery, University of Florida, Gainesville, Florida, United States
  7. 7 Department of Neurosurgery and Radiology, Canon Strole and Vascular Research Center, University at Buffalo, State University of New York, Buffalo, New York, United States
  1. Correspondence to Dr Christopher S Ogilvy, Neurosurgical Service, Beth Israel Deaconess Medical Center, Harvard Medical School, Boston 02215, Massachusetts, USA; cogilvy{at}


Background and purpose We previously reported a single-centre study demonstrating that smoking confers a six-fold increased risk for having an unruptured intracranial aneurysm (UIA) in women aged between 30 and 60 years and this risk was higher if the patient had chronic hypertension. There are no data with greater generalisability evaluating this association. We aimed to validate our previous findings in women from a multicentre study.

Methods A multicentre case-control study on women aged between 30 and 60 years, that had magnetic resonance angiography (MRA) during the period 2016–2018. Cases were those with an incidental UIA, and these were matched to controls based on age and ethnicity. A multivariable conditional logistic regression was conducted to evaluate smoking status and hypertension differences between cases and controls.

Results From 545 eligible patients, 113 aneurysm patients were matched to 113 controls. The most common reason for imaging was due to chronic headaches in 62.5% of cases and 44.3% of controls. A positive smoking history was encountered in 57.5% of cases and in 37.2% of controls. A multivariable analysis demonstrated a significant association between positive smoking history (OR 3.7, 95%CI 1.61 to 8.50), hypertension (OR 3.16, 95% CI 1.17 to 8.52) and both factors combined with a diagnosis of an incidental UIA (OR 6.9, 95% CI 2.49 to 19.24).

Conclusions Women aged between 30 and 60 years with a positive smoking history have a four-fold increased risk for having an UIA, and a seven-fold increased risk if they have underlying chronic hypertension. These findings indicate that women aged between 30 and 60 years with a positive smoking history might benefit from a screening recommendation.

View Full Text

Statistics from


Background and objectives

There are an estimated 6.5 million people harbouring an unruptured intracranial aneurysm (UIA) in the United States.1–3 There are risk factors that influence the prevalence of UIAs as well as those associated with aneurysm rupture/aneurysmal subarachnoid hemorrhage (aSAH).4 5 There is a higher likelihood of developing UIAs in patients with past family history of UIA, previously diagnosed UIAs, polycystic kidney disease and connective tissue disorders. The incidence of UIA diagnosis has increased substantially over the past 30 years with improved neuroimaging.6 Currently, there is thought to be no benefit in screening for UIAs in the general population.7 Screening for an UIA is recommended in patients with two or more first-degree family members with history of intracranial aneurysms and in patients with autosomal-dominant polycystic kidney disease.7

Can et al demonstrated that both current and past cigarette smokers were more likely to have an aSAH.8 Additionally, there is a clear association between smoking and the likelihood of having an aneurysmal growth and subsequent rupture.9 Therefore, patients with an UIA are advised against consumption of tobacco to avoid devastating consequences, including death.7 We previously reported a single-centre study that showed a prevalence of 19% of UIA among women aged between 30 and 60 years who smoke cigarettes. Additionally, there was a significant relationship between tobacco consumption and incidental UIA diagnosis.10 Previous studies have demonstrated that UIAs are more common in women.11 12 The purpose of this study was to assess the relationship between first-hand cigarette smoking and an incidental diagnosis of an UIA in females aged between 30 and 60 years using a multicentre matched case-control analysis.


Study design, setting and participants

We performed a multicentre matched case-control study of patients with an incidental diagnosis of a UIA in the neuroradiology database of all the magnetic resonance angiographies (MRAs) performed in five large academic hospitals from the United States and Canada from 2016 through 2018. Data-sharing agreements and institutional review board (IRB) approval at each institution were obtained prior to the commencement of this study. Patient consent was not required due to the retrospective nature of the study. Eligible case participants included female patients, aged between 30 and 60 years, with an incidental diagnosis of a UIA on MRA. Patients were considered ineligible if they had any past medical history with relation to an UIA, as well as those with family history of UIA, aSAH, connective tissue disorders or any cerebrovascular anomaly. Patients diagnosed in the setting of clinical suspicion for a UIA were also deemed ineligible. Eligible control patients were those that had an MRA for any workup reason, but with a normal imaging study. These were matched to each aneurysm patient based on age and race.

Variables, data source and measurements

The presence of a UIA was recorded from the MRA report performed by a neuroradiologist from each institution, and confirmed independently by the corresponding authors at each site. Cases were identified from either a prospective repository of UIA patients, or from the neuroradiology data repository from each institution. A preliminary screen using an advanced search with keywords such as 'intracranial aneurysm' or 'unruptured aneurysm' was performed at each institution. Thereafter, a manual screening was conducted to identify cases diagnosed incidentally. Then, each aneurysm patient was screened for exclusion characteristics to ensure absence of any characteristic considered part of the exclusion criteria, such as family history of UIAs, connective tissue disorders, and other previously mentioned exclusion characteristics. Simultaneously, the control subjects were obtained from the MRA neuroradiological repositories from each institution, by utilising keywords such as 'normal study'. Both cases and controls were obtained from the prespecified timeframe of 2016 to 2018. A pre-hoc sample size calculation was performed to arrive at the minimum number that each centre should be able to contribute to have a minimum power of 80% to reject the null hypothesis. This assessment was based on a prevalence of UIAs of 2% and a prevalence of smoking in females between the ages of 30 and 60 years of 15% (Massachusetts).13 14 Assuming an encountered OR of 2, the hypothesised rate in a case-control design would be 30%. Therefore, with a power at 80% and an alpha of 5% level two-sided test, the minimum sample size required was 132 aneurysm patients and 396 controls. Each centre would contribute both aneurysm patients and control subjects in a ratio of 1:3.

Due to the large likelihood ratio degrees of freeedom in the standard (ie, unmatched) multivariable logistic regression model relative to the total number of observations, there was a raising concern of a biased coefficient estimates. A final decision was made to opt for a conditional logistic regression analysis, matching on two important literature confounders (ie, age and race). The results from the unmatched dataset are shown in the online supplementary material (online supplementary table 1,2). The data source for all the variables used for this study were the patients’ medical record, self-reported questionnaire, outpatient clinical notes and the MRA report.

Supplemental material

The exposure of interest was a positive history of first-hand cigarette smoking. This was recorded as positive if the patient described any previous use of tobacco cigarettes, including both past exposure for 1 year or more, or if the person was a current cigarette smoker. A diagnosis of chronic hypertension (with or without adequate control/medication) was also utilised as part of the primary analysis. We further characterised the smoking habit, if available, along with details on years of smoking and years since last cigarette consumed. Baseline demographic data, such as age, ethnicity, race, past medical history, reason for imaging, date of aneurysm diagnosis, aneurysm characteristics (eg, maximal diameter, parent artery), treatment, follow-up data and occlusion status at last follow-up, were also recorded.

Matching datapoints

From the available cohort of patients received from all the participating centres, cases were matched to controls in a 1:1 ratio based on exact age and race. To ensure proper 1:1 matching, additional entries from multiple aneurysms in the same patient were eliminated before running the match algorithm.

Quantitative variables and statistical methods

Continuous datapoints were reported with their mean and SD or median and their respective IQR. These were analysed with either the Student’s t-test or the Mann–Whitney U test depending on data normality. Categorical datapoints are reported as proportions and compared with a Chi-square analysis. An interaction term was added to the primary analysis to evaluate the mixed effect of both cigarette smoking and chronic hypertension. A multivariable model was constructed to explore the association between the exposure of interest (ie, first-hand cigarette smoking) and the proportion of cases (ie, incidental UIA) controlling for potential confounders, selected based on univariable screening and prior literature knowledge. A minimum of 20 observations were included for each variable in the model to avoid overfitting. A secondary analysis was conducted in patients that had additional details on the smoking habit, such as number of cigarettes consumed per day and duration. Statistical significance was set at p<0.05. The analysis was conducted using Stata 15.0 (StataCorp., College Station, Texas, USA).


Patient characteristics and indications for imaging

A total of 152 aneurysm patients (ie, with incidental UIA) harbouring 185 UIAs were identified from the participating centres, with a corresponding 393 controls. Aneurysm patients were subsequently matched to controls in a 1:1 ratio based on exact age and race, yielding 113 cases and 113 corresponding controls (figure 1). Hypertension was more prevalent in cases (46% vs 31%, p=0.02). Similarly, a positive smoking history (either current or past smoker) was more frequent among cases (57.5% vs 37.2%, p=0.002). The most common reason for having an MRA study in cases and controls was due to a workup for chronic headaches (62.5% and 44.3%, respectively). Other reasons for imaging are presented in table 1. For the purpose of the multivariable analysis, the reasons for imaging were categorised as due to chronic headaches or any other reason.

Table 1

Baseline characteristics of case-control subjects

Figure 1

Flow diagram of the study. A total of 545 patient records were available after performing an advanced radiological search with a prespecified age, sex and date of imaging. After eliminating duplicate reports, we assessed both aneurysm patients and controls for eligibility based on inclusion/exclusion criteria. A 1:1 matching technique rendered a total of 113 patient subjects harbouring 128 unruptured intracranial aneurysms and 113 controls. The matching process was based on exact age and self-reported race.

Lesion characteristics

The original cohort of aneurysm cases was comprised of 152 patients harbouring 185 aneurysms. The final population included after matching was 113 patients harbouring 128 aneurysms (100%). The majority of the aneurysms were saccular in morphology (95.3%) and 91.4% were located in the anterior circulation. The internal carotid artery was the most common location within the anterior circulation (53.9%). The median (IQR) aneurysm size was 4 (2–5) mm. Two-thirds of the aneurysms were managed conservatively and 34% were treated (table 2).

Table 2

Per-aneurysm baseline characteristics of matched cases

Comparisons between aneurysm patients and controls

Aneurysm patients and controls had equal demographic characteristics after matching (p>0.99 for age and race). Chronic hypertension was more frequent among aneurysm patients compared with controls (46% vs 31%, respectively; unadjusted p value=0.02). More than half of the aneurysm patients (57%) had a positive smoking history while 37% of the controls were positive for smoking (p=0.002). The median cigarette consumption was 20 cigarettes per day for aneurysm patients and 12 cigarettes per day for controls (p=0.6). While the average duration of smoking was higher in aneurysm patients compared with controls (25 vs 15 years, respectively), this difference was not significant. The multivariable model included the following variables: cigarette smoking history, chronic hypertension and chronic headaches as the reason for imaging (table 3).

Table 3

Conditional logistic multivariable analysis model for the association with a diagnosis of an incidental unruptured intracranial aneurysm between the 1:1 matched case-control patients

Patients with a positive smoking history had an OR of 3.69 for having an UIA (95% CI 1.61 to 8.50; p=0.002). Additionally, patients with chronic hypertension had an OR of 3.16 for having an UIA (95% CI 1.17 to 8.52; p=0.023). There was a significant interaction between a positive smoking history and chronic hypertension with an incidental UIA (OR 6.92; p<0.001).

Sensitivity analysis

To further investigate the effects of cigarette smoking duration and intensity of consumption with an incidental UIA diagnosis we performed a sensitivity analysis from the initial cohort prior to excluding the aneurysm patients that were not matched. The median number of cigarettes consumed per day was 10 (IQR 10–20) in both aneurysm patients and controls. However, further analysis demonstrated a significant correlation between smoking duration and an incidental UIA. Specifically, the median duration of cigarette consumption was 29 (IQR 16–40) years in aneurysm patients and 20 (IQR 11–26) years in control subjects (OR 1.03, 95% CI 1.02 to 1.08; p=0.001) (table 4).

Table 4

Sensitivity analysis: logistic multivariable analysis model for the association with a diagnosis of an incidental unruptured intracranial aneurysm between aneurysm patients and control patients with available data on both cigarette smoking duration and quantity of cigarettes per day


Half a decade has elapsed since the American Heart Association released the updated guidelines for the management of UIAs.7 Part of the objectives of establishing standardised recommendations is to provide a summary of the best evidence-based practices for physicians to guide their decisions when taking care of patients. Strategies such as primary prevention, early detection, and improvement in treatment technologies are efforts intended to decrease the incidence of aSAH, considering that such an event carries a high early mortality risk (up to 50%) as well as a mortality risk of 17% in patients that survive the aSAH.15 16 Therefore, screening recommendations for UIAs is considered a beneficial strategy given there is a population at higher risk for both having an UIA and for experiencing an aSAH, compared with the risks associated with current treatments, which can be as low as 1%–3% for having a permanent neurological deficit.17

Women aged 30–60 years and intracranial aneurysms

The decision to focus on patients within this age range was based on the prevalence of cigarette smoking in the United States, which has been reported to be the highest in patients aged between 30 and 60 years.14 In large natural history studies of UIAs,11 12 the majority of patients are women; therefore, we felt that women who smoke would be the most likely to harbour an incidental UIA.

Cigarette smoking sensitivity analysis – duration versus intensity

Can et al performed a case-control study analysing the relationship between aneurysm rupture and smoking intensity and duration in 4701 patients with intracranial aneurysms.8 Interestingly, the authors report that in addition to a positive smoking history, both smoking intensity and duration were significantly associated with ruptured intracranial aneurysms, and this risk persisted after cigarette smoking cessation. While the outcome of the present study is that of unruptured aneurysms, our findings from the sensitivity analysis are similar to those of Can et al showing that the duration of cigarette consumption is an independent risk factor for an incidental UIA. Therefore, these findings, together with those of Can et al, reflect a potential long-standing effect of cigarette consumption on both aneurysm formation and rupture. Furthermore, our current report provides data demonstrating that the duration of cigarette consumption rather than the number of cigarettes consumed has a significantly greater impact on the baseline risk for having an UIA. Each additional year of cigarette consumption confers a 3% increase in the baseline risk for having an incidental UIA diagnosis.


Chronic hypertension has been shown to be a risk factor for aneurysm development and rupture;18 19 however, other studies have found no relationship between hypertension and aneurysm rupture.20 We chose to investigate chronic hypertension in a population without any underlying increased risk for having an UIA. Our findings demonstrate that chronic hypertension is an independent risk factor for incidental UIAs.

Cigarette smoking and hypertension

Our findings demonstrate that women with a positive smoking history have a four-fold increased risk for having an incidental UIA compared with women non-smokers without hypertension. Moreover, women aged between 30 and 60 years with a positive smoking history and chronic hypertension have a seven-fold higher risk for having an incidental UIA when compared with women non-smokers without hypertension. This implies that there is a synergistic effect between these two conditions and the presence of an UIA.

Preventive management

The rationale behind treating an UIA is based on the concept that treatment risk is lower than the lifetime risk of haemorrhage. In this study we found that 66% of the aneurysms underwent conservative management with long-term follow-up. Current literature suggests that both hypertension and smoking increases the overall likelihood for an aneurysm to grow.9 Patients with these underlying characteristics (cigarette smoking and/or hypertension) should undergo long-term follow-up imaging studies due to a higher likelihood of experiencing aneurysm changes, compared with non-smokers. Thirty-four percent of aneurysms in the current study met the criteria for treatment. These data suggests a need to evaluate the potential benefits of a screening recommendation in female smokers aged 30–60 years.

Screening rationale

Measures aimed to decrease the incidence of aSAH include education, awareness and screening recommendations.7 The primary purpose of early detection and treatment of UIAs arises from the concept that treatment is relatively safer than observation in specific populations that carry a high prevalence and rupture risk. The current guidelines for management of UIAs recommend screening for patients with either autosomal-dominant polycystic kidney disease or those with two or more first-degree family members with a positive history for either aneurysm diagnosis or rupture.7 The prevalence of intracranial aneurysms among patients with a positive family history of UIAs in two or more first-degree relatives is 19.1%,21 and the risk for an aSAH in these patients is up to 7.1% per year.22 In a previous study, it was shown that the prevalence of UIAs in women who smoke was 19.0%.10 Given these numbers, and the data presented in the current study, consideration should be given to screening for UIA in women aged between 30 and 60 years who smoke cigarettes.

Prevalence of intracranial aneurysms in women

It is known that female patients are the most prevalent demographic population among UIA patients.12 A recent study found that women with a positive smoking history had an aneurysm prevalence of 19% while women non-smokers had a prevalence of 1.9%. Interestingly, the prevalence of UIA among female patients with a positive smoking history (19%) was similar to that of patients with a family history of aneurysms (19.1%).21 Our current study was designed to validate the association between cigarettes and incidental UIAs in the absence of a positive familial history in a more representative study sample. The findings from our current report suggest an independent association with cigarette smoking in women in the absence of any known familial influence. The number of aneurysm patients from the Familial Intracranial Aneurysm Study that were treated after screening detection was 10 patients (17%). In the current multicentre study, we observed 37 patients (34%) that were felt to warrant treatment. This suggests that cigarette smoking might pose a greater risk to any given individual than that of a familial component, and even a greater threat to female patients.

Interpretation and limitations

These data provide further evidence that women aged between 30 and 60 years with a positive smoking history have a high risk for harbouring an incidental UIA, which is amplified by the presence of chronic hypertension. Future studies evaluating the role of screening in this population should assess the cost-effectiveness of this strategy. Moreover, as in patients with more than two first-degree family members or autosomal-dominant polycystic kidney disease, a preventive screening recommendation requires considering the psychological strain associated with screening and the potential impact of diagnosing an intracranial aneurysm on quality of life. Ultimately, physicians must seek a joint decision with the patient to weigh any level of anxiety against the available evidence of screening when making any diagnostic or treatment recommendation. Limitations of our data include patient recall-bias, limited data on the smoking habit, constraints inherent in any retrospective study, and lack of data relative to the rupture risk, which is beyond the scope and capabilities of a case-control analysis. To establish the rupture risk in a given population, a prospective follow-up study or a randomised trial would need to be conducted. However, a prospective follow-up study has proven cost-prohibitive and a randomised trial of cigarette smoking in women would be unethical.


Women aged between 30 and 60 years without any underlying risk factor for having an UIA have a four-fold increased risk for having an incidental UIA if they have a positive smoking history. If there is additional underlying chronic hypertension, this risk increases to seven-fold compared with women non-smokers without hypertension. These findings indicate that women aged between 30 and 60 years with a positive smoking history could benefit from a screening recommendation.


View Abstract


  • Twitter @SantiagoGP92, @_AdnanSiddiqui, @EladLevyMD, @mtlawton, @BrianHoh1, @wchrisfox, @DrAjithThomas

  • Contributors SG-P, KPK and CSO conceived and designed the study. SG-P, MMS, YA, MW, HHR, JSC, SM and ME collected the data. SG-P, MMS and YA analysed and interpreted the data under the supervision of JMM, AJT and CSO. SG-P and MMS did the statistical analyses. SG-P and CSO performed the literature search, drafted the paper and all authors reviewed and contributed important intellectual content and edited the manuscript.

  • Funding This work was supported by the Boston Marathon Chair of Research and Fight Like Frank Chair of Research grant from the Brain Aneurysm Foundation to SG-P (BAF2019-7894772232).

  • Competing interests None declared.

  • Patient consent for publication Not required.

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

  • Data availability statement Data are available upon reasonable request.

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.