Risk of Rupture of an Intracranial Aneurysm Based on Patient Characteristics
A Case–Control Study
Background and Purpose—Knowledge about risk factors contributes to understanding the pathophysiological mechanisms that cause intracranial aneurysm rupture and helps to develop possible treatment strategies. We aimed to study lifestyle and personal characteristics as risk factors for the rupture of intracranial aneurysms.
Methods—We performed a case–control study with 250 patients with an aneurysmal subarachnoid hemorrhage and 206 patients with an unruptured intracranial aneurysm. All patients with an aneurysmal subarachnoid hemorrhage and patients with a unruptured intracranial aneurysm were asked to fill in a structured questionnaire about their lifestyle and medical history. For patients with an unruptured intracranial aneurysm, we also collected data on the indication for imaging. With logistic regression analysis, we identified independent risk factors for aneurysmal rupture.
Results—Reasons for imaging in patients with an unruptured intracranial aneurysm were atherosclerotic disease (23%), positive family history (18%), headache (8%), preventive screening (3%), and other (46%). Factors that increased risk for aneurysmal rupture were smoking (odds ratio, 1.9; 95% confidence interval, 1.2–3.0) and migraine (2.4; 1.1–5.1); hypercholesterolemia decreased this risk (0.4; 0.2–1.0), whereas a history of hypertension did not independently influence the risk.
Conclusions—Smoking, migraine and, inversely, hypercholesterolemia are independent risk factors for aneurysmal rupture. Data from the questionnaire are insufficient to conclude whether hypercholesterolemia or its treatment with statins exerts a risk-reducing effect. The pathophysiological mechanisms through which smoking and migraine increase the risk of aneurysmal rupture should be investigated in further studies. Although a history of hypertension does not increase risk of rupture, a sudden rise in blood pressure might still trigger aneurysmal rupture.
Three percent of the general population harbors an unruptured intracranial aneurysm (UIA).1 Reducing the risk of rupture of an UIA is an important method for reducing the incidence of aneurysmal subarachnoid hemorrhage (aSAH), which has remained relatively stable over the past decades.2
Little is known about modifiable patient characteristics as risk factors for aneurysmal rupture. Smoking is a risk factor in many, but not all, studies.3–7 Data on hypertension remain even more controversial,3,5–9 and only few studies report on other risk factors, such as alcohol, diabetes mellitus, and hypercholesterolemia.6–8
Knowledge about risk factors for aneurysmal rupture contributes to understanding the pathophysiological mechanisms that cause aneurysm rupture and helps to develop possible treatment strategies other than invasive interventions to reduce this risk. We compared the characteristics of patients with an UIA and patients with an aSAH to study data on lifestyle, clinical condition, and family history as risk factors for the rupture of UIAs.
Between September 2006 and September 2009, we included 250 consecutive patients who were admitted for aSAH at the Utrecht Stroke Center of the University Medical Center Utrecht in a case-crossover study.10 Patients were eligible if they were ≥18 years of age and spoke Dutch or English or had a proxy who did so. Informed consent was given by either the patient or the proxy. aSAH was defined as an abrupt onset of severe headache or loss of consciousness with or without focal neurological signs and subarachnoid blood proven by computed tomography or lumbar puncture, and an aneurysm proven with computed tomography angiography, MR angiography, or conventional angiography. The patients included in a previous case-crossover study now served as cases in the current case–control study.10
As comparison group, we included 206 patients with an UIA from the same catchment area, who were treated at or referred to the Utrecht Stroke Center. Eligible were patients who had an UIA, who never had an aSAH, and were ≥18 years of age. The UIA had to be confirmed by computed tomography angiography, MR angiography, or conventional angiography. The study protocol was approved by the Medico-Ethical Review Committee of our hospital.
All patients with an aSAH and patients with an UIA were asked to fill in a structured questionnaire about their lifestyle, clinical condition, and family history before their aSAH or the discovery of their UIA. We collected data on demographics (age, sex), height and weight, medical history (diabetes mellitus, heart disease, hypertension, hypercholesterolemia, migraine), smoking, use of alcohol, and physical exercise.10 We also asked for family history of stroke (other than aSAH). We excluded aSAH in family history of stroke because this was often a reason for screening in patients with an UIA and therefore might lead to selection bias. Smoking was defined as smoking at the time of aSAH or diagnosis of the UIA. Heart disease included myocardial infarction, angina pectoris, coronary artery bypass grafting, and percutaneous transluminal coronary arterioplasty. Diagnoses of diabetes mellitus, hypertension, heart disease, and hypercholesterolemia were checked against medical records to confirm that the diagnosis was made before discovery of the UIA or aSAH. Family history was based on information from the questionnaire. Excessive alcohol use was defined as ≥18 U (ie, ≥150 g) per week.11 Lean body mass was defined as a body mass index <22. Regular physical activity was defined as vigorous physical exercise (metabolic equivalent >6) >3 times a week.12
For patients with an UIA, we also collected data on the indication for imaging from their medical records and predefined 5 categories: atherosclerotic disease (including transient ischemic attack and stroke), positive family history for UIA or aSAH, any type of headache, preventive screening without medical reasons, and other.
All variables listed in Table 1 were assessed as possible risk factors. Univariable analysis was performed to calculate crude odds ratios (ORs) with 95% confidence intervals (CI). Subsequently, we performed a multivariable backward stepwise logistic regression analysis to identify independent risk factors for rupture. Risk factors with P<0.05 were considered statistically significant. Only family history of aSAH and a medical history of stroke were not used for multivariable analysis because this was often the reason for screening and therefore may cause selection bias.
The initial analysis for migraine was based on self-report on the questionnaire. We did a sensitivity analysis, including only patients for whom migraine was documented as such in the medical records.
Because many patients with an UIA had imaging because of atherosclerotic disease, we performed 2 additional sensitivity analyses to investigate the role of selection bias. First, we did an analysis excluding patients with an UIA who were screened for atherosclerosis. Second, we did an analysis excluding all patients with an UIA and patients with an aSAH with a history of stroke.
The baseline characteristics of patients with an aSAH and patients with an UIA are summarized in Table 1. Reasons for imaging in patients with an UIA were atherosclerotic disease (23%), positive family history (18%), headache (8%), preventive screening (3%), and other (46%), including seizures, diplopia, or probably unrelated symptoms, such as dizziness.
Risk Factors for Rupture
In the univariable analysis, we found an increased risk of aneurysm rupture with smoking and a history of migraine, and a decreased risk with hypercholesterolemia, hypertension, diabetes mellitus, and a history of heart disease. Only current smoking (OR, 1.9; 95% CI, 1.2–3.0), hypercholesterolemia (OR, 0.4; 95% CI, 0.2–1.0), and history of migraine (OR, 2.4; 95% CI, 1.1–5.1) were independent risk factors (Table 2). Family history of aSAH also decreased the risk of aSAH in univariable analysis, but this factor was not considered in multivariable analysis because it was often the indication for imaging.
In the sensitivity analysis on migraine, as documented in the medical records, the OR of migraine in univariable analysis was 3.7 (95% CI, 1.7–10). In multivariable analysis, migraine was no longer an independent risk factor.
We studied the influence of atherosclerosis by doing 2 additional analyses: one excluding patients with an UIA who were screened for atherosclerosis and another excluding all patients with an UIA and patients with an aSAH with a history of stroke. In both analyses, we found that point estimates for smoking, a history of migraine, and hypercholesterolemia remained virtually the same, although the OR for hypercholesterolemia was no longer statistically significant. (Table I in the online-only Data Supplement).
Current smoking and a history of migraine increase the risk of rupture of an intracranial aneurysm. Hypercholesterolemia decreases this risk, whereas hypertension does not independently affect the risk of rupture.
In previous studies from the same catchment area, we reported that smoking is an independent risk factor for aneurysmal formation as well as for aSAH.13,14 Although smoking increased the risk of rupture, we also previously reported that smoking is not a trigger factor for aneurysmal rupture.5,10 This suggests that smoking weakens the wall of the UIAs, making them more vulnerable to trigger factors and thus rupture. The weakening of the vessel wall in smokers may be caused by an increased aneurysm growth,4,8 or by inflammation of the vessel wall, rendering it more prone to rupture.15,16
Migraine was also an independent risk factor for aneurysmal rupture, but we could not find other studies reporting on migraine as risk factor for aneurysmal rupture. So far, there is no reliable information about a possible pathway through which migraine might increase the risk of rupture. Migraine is not a trigger factor for aneurysmal rupture,10 so, like smoking, migraine may lead to weakening of the aneurysmal wall. The pathway through which migraine may weaken the aneurysmal wall is unknown. It might be a migraine-related process, but also a process influenced by antimigraine drugs.
Hypercholesterolemia independently decreased the risk of rupture. Recently, a Japanese case–control study reported similar results,3 whereas others found no effect on risk of rupture.6,7 It is unknown whether the reduction in risk of rupture is caused by the hypercholesterolemia itself or by the use of statins.
Although on one hand, hypertension is a risk factor for developing UIAs and aSAH, and on the other, circumstances that lead to a sudden and short-lasting rise in blood pressure are a trigger factor for aSAH, we found no long-term effect of hypertension on the risk of aneurysmal rupture in patients with a known UIA.3,4,7,10,13,14 A possible explanation for these findings is that hypertension (when it is not yet known by the patient) leads to aneurysm formation, but that treatment of hypertension reduces the further growth of aneurysms and thereby the risk of rupture in the long term, whereas sudden bouts of hypertension still may provoke aneurysm rupture. In patients known to have an UIA, hypertension might be more thoroughly treated than in patients with hypertension who are not aware of having an UIA. From this point of view, aneurysmal rupture may be seen as a failure of treatment of hypertension. Also, our comparison group of patients with an UIA often had a history of atherosclerotic disease, which was frequently the reason for imaging, and may again lead to better treatment of hypertension. In further studies, the time relationship between diagnosis of hypertension, installment of treatment, success of treatment, and aneurysmal rupture should be further investigated.
Our study has some limitations. First, selection bias by indication for screening should be considered, in particular when studying atherosclerotic risk factors and family history of aSAH or UIA. About 23% of the patients with an UIA were investigated because of atherosclerotic disease. The large proportion of patients with atherosclerosis in the group with an UIA may have resulted in a higher frequency of atherosclerotic risk factors in the group with UIAs compared with the group with aSAH. The low risk or rupture for history of diabetes mellitus, hypertension, heart disease, and family history of stroke in the univariable analysis probably reflects the selection bias. Also, the risk of rupture for smoking might be potentially lowered by the selection bias. In the multivariable analysis, hypercholesterolemia and smoking were the only statistically significant atherosclerosis-associated risk factors for rupture. To study the possible effect of selection bias on our results for smoking and hypercholesterolemia, we did 2 sensitivity analyses, excluding patients screened for atherosclerotic disease or with a history of stroke, and found that point estimates for smoking and hypercholesterolemia remained virtually the same. This shows that smoking and hypercholesterolemia are indeed associated with risk of aneurysmal rupture, and that this finding is not explained to a considerable extent by selection bias. Also, 18% of all patients with an UIA were screened because of a positive family history. It is therefore very likely that the protective effect of family history of aSAH on the risk of rupture in the univariable analysis is the result of inclusion bias. Thus, a family history of aSAH could not be studied in the current case–control study and was not considered in the multivariable analysis. Second, headaches might have been misclassified as migraine because we based the diagnosis on self-reporting from the questionnaire, and not on a formal interview using the International Classification of Headaches to establish the diagnosis of migraine. Other types of headache, previous warning leaks, and even aSAH might be mistaken for a migraine attack. Also, recall bias may have played a role because patients with aSAH or their proxies might remember bouts of headaches more often than patients with UIAs. Our sensitivity analyses, however, showed that misclassification and recall bias do not explain our results. Finally, recall bias for factors other than headache may have influenced our results, because the time lag between diagnosis and questionnaire in patients with an aSAH is much shorter than in patients with an UIA.
Our study also has several strong points. First, because both patients with an aSAH and patients with an UIA came from the same catchment area, we were able to study risk factors for aneurysmal rupture in a rather defined population. Second, the inclusion of only first-ever cases of aSAH reduced selection and misclassification bias. Third, we had access to individual patient data, which allowed multivariable analysis, including a large number of potential risk factors with specific attention for patient characteristics instead of aneurysm characteristics. Fourth, there was a uniform way of classification of characteristics because all participants were seen at the same hospital.
Currently, small aneurysms (<7 mm) are often left untreated and monitored over time. However, some rupture during follow-up. Based on our results, patients who harbor an aneurysm should be even more strongly advised to quit smoking. Also, our results support the need for further pathophysiological studies on how migraine and hypercholesterolemia exert their effects on risk of rupture of aneurysms. This knowledge may lead to new treatment strategies to reduce the risk of rupture of UIAs. If the risk decreasing effect of hypercholesterolemia is mediated by statin use, statins might be an additional treatment option for decreasing the prevalence of UIAs, and thereby the incidence of aSAH, and should also be further investigated. In contrast, if the risk attributable to migraine is caused by the use of antimigraine drugs, the use of antimigraine drugs should be discussed with the patients. However, a study on the influence of drugs on the risk of rupture of UIAs will be very challenging because such a study would need to have a prospective design with very long-term follow-up.
Sources of Funding
This study was funded by the Julius Center for Health Sciences and Primary Care and the Department of Neurology of the University Medical Center Utrecht.
The sponsor had no role in the study design, data collection, data analysis, data interpretation, or writing of the report. The corresponding author had full access to all data in the study and had the final responsibility for the decision to submit for publication.
The online-only Data Supplement is available with this article at http://stroke.ahajournals.org/lookup/suppl/doi:10.1161/STROKEAHA.111.000679/-/DC1.
- Received January 2, 2013.
- Accepted February 22, 2013.
- © 2013 American Heart Association, Inc.
- de Rooij NK,
- Linn FH,
- van der Plas JA,
- Algra A,
- Rinkel GJ
- Wermer MJ,
- van der Schaaf IC,
- Algra A,
- Rinkel GJ
- Juvela S,
- Poussa K,
- Porras M
- Vlak MH,
- Rinkel GJ,
- Greebe P,
- van der Bom JG,
- Algra A
- Feigin VL,
- Rinkel GJ,
- Lawes CM,
- Algra A,
- Bennett DA,
- van Gijn J,
- et al
- Mittleman MA,
- Maclure M,
- Tofler GH,
- Sherwood JB,
- Goldberg RJ,
- Muller JE
- Vlak MH,
- Rinkel GJ,
- Greebe P,
- Greving JP,
- Algra A
- Vlak MH,
- Rinkel GJ,
- Greebe P,
- Algra A