Aspirin and Risk of Subarachnoid Hemorrhage
Systematic Review and Meta-Analysis
Background and Purpose—Recent studies have suggested that the use of low-dose aspirin may reduce the risk of aneurysmal subarachnoid hemorrhage (aSAH). We aimed to evaluate any association between aspirin use and risk of aSAH based on the literature, and whether this is influenced by duration or frequency of aspirin use.
Methods—A search of electronic databases was done from inception to September 2016. For each study, data on risk of aSAH in aspirin versus nonaspirin users were used to generate odds ratios and 95% confidence intervals, and combined using inverse variance–weighted averages of logarithmic odds ratios in a random-effects models.
Results—From 7 included studies, no significant difference was noted between aspirin use of any duration or frequency and nonaspirin users (odds ratio, 1.00; 95% confidence interval, 0.81–1.24; P=0.99). We found a significant association between short-term use of aspirin (<3 months) and the risk of aSAH (odds ratio, 1.61; 95% confidence interval, 1.20–2.18; P=0.002). No significant difference was found in terms of risk of aSAH for 3 to 12 months, 1 to 3 years, and >3 years of durations of use. No significant association was found between infrequent aspirin use (≤2× per week) or frequent use (≥3× per week) with risk of aSAH.
Conclusions—Current evidence suggests that short-term (<3 months) use of aspirin is associated with increased risk of aSAH. Limitations include substantial heterogenity of the included studies. The role of long-term aspirin in reducing risk of aSAH remains unclear and ideally should be addressed by an appropriately designed randomized controlled trial.
Aneurysmal subarachnoid hemorrhage (aSAH) remains a morbid and lethal condition.1–4 Despite an increase in the number of unruptured intracranial aneurysms treated with endovascular techniques or microsurgical clipping, the incidence of aSAH is relatively unchanged in many countries including the United States.5 It seems as if risk factor modification such as smoking cessation has more impact on prevention of aSAH than treatment of unruptured aneurysms. For many patients with an unruptured aneurysm, the decision whether to pursue invasive therapy or continued observation with risk factor modification is challenging. Although antiplatelets and antithrombotic agents have been widely used in the prevention of thromboembolism and stroke, their effects on prevention of aSAH are not well established.
There have been several recent studies that have suggested that the use of low-dose aspirin may reduce the risk of aSAH. Hasan et al6 first reported the protective role of aspirin against the risk of aSAH, finding that aspirin use 3× per week was associated with significantly reduced adjusted odds of aSAH. The authors hypothesized that the protective effect was secondary to modulation of the inflammatory pathways that have been implicated in the pathogenesis of aneurysm development and rupture.7–10 As a widely used and commonly available medication with an acceptable safety profile and dosing regimen, aspirin may be a promising treatment option for this indication.
However, this evidence is conflicting with some studies suggesting that aspirin may be associated with increased risk of aSAH after short-term use.11,12 It seems that the beneficial effects of aspirin, if any, on the risk of aSAH is dependent on the duration of aspirin use and the prescribed dosing regimen. Given that aspirin use has the potential to increase the amount of subarachnoid blood after an aneurysm rupture, it is absolutely critical to establish the nature of the relationship between regular aspirin use and risk of aSAH. The evidence base on the role of aspirin in preventing aSAH remains unclear. As such, we performed a systematic review and meta-analysis of all the available literature to assess the potential benefits and risks of aspirin use with regard to the risk of aSAH.
Literature Search Strategy
Electronic searches were performed using Ovid Medline, PubMed, Cochrane Central Register of Controlled Trials, Cochrane Database of Systematic Reviews, ACP Journal Club, Database of Abstracts of Review of Effectiveness, and Google Scholar from their dates of inception to September 2016. To achieve the maximum sensitivity of the search strategy, we combined the terms: aspirin, acetylsalicylic acid, aneurysm, subarachnoid hemorrhage, SAH as either key words or MeSH terms. The reference lists of all retrieved articles were reviewed for further identification of potentially relevant studies using the inclusion and exclusion criteria.
Eligible studies for the present systematic review and meta-analysis included those that reported an odds ratio (OR) and confidence interval (CI) for aspirin use with risk of aSAH. The OR was manually derived from studies that reported the proportion of patients who developed aSAH in aspirin users versus nonaspirin users or from the relative risk reported from the included study. When possible, the adjusted OR and CI were used for the present meta-analysis. When institutions published duplicate studies with accumulating numbers of patients or increased length of follow-up, only the most complete reports were included for quantitative assessment at each time interval. All publications were limited to those involving human subjects and in the English language. Abstracts, case reports, conference presentations, editorials, reviews, and expert opinions were excluded.
All data were extracted from article texts, tables, and figures using a standardized form. Information was collected included (1) study characteristics, (2) number of patients in aspirin versus nonaspirin user groups, (3) effect size for relationship between aspirin use versus nonaspirin use and risk of aSAH, (4) factors/confounders adjusted for when the study reported the relationship between aspirin use and risk of aSAH. Because quality scoring is controversial in meta-analyses of observational studies, each article included in our analysis was appraised based on a critical review checklist of the Dutch Cochrane Center proposed by the Meta-Analysis of Observational Studies in Epidemiology group.13 The key points of this checklist include (1) clear definition of study population; (2) clear definition of outcomes and outcome assessment; (3) independent assessment of outcome parameters; (4) sufficient duration of follow-up; (5) no selective loss during follow-up; and (6) important confounders and prognostic factors identified. The final results were reviewed by senior investigators.
For each study, data on risk of aSAH in aspirin versus nonaspirin users were used to generate OR and 95% CIs. Study-specific estimates were combined using inverse variance–weighted averages of logarithmic ORs in a random-effects models. Between-study heterogeneity was analyzed by means of standard χ2 tests. We also quantified the degree of heterogeneity using the I2 statistic, which represents the percentage of the total variability across studies that is because of heterogeneity. I2 values of 25%, 50%, and 75% corresponded to low, moderate, and high degrees of heterogeneity, respectively. All analyses were conducted using review manager (RevMan version 5.3.5; The Nordic Cochrane Center, The Cochrane Collaboration, Copenhagen). Publication bias was assessed using funnel plot asymmetry, Trim-and-Fill analysis, and Egger test for publication bias.
A total of 5978 references were identified through 6 electronic database searches (Figure 1). After exclusion of duplicate or irrelevant references, 5921 potentially relevant articles were retrieved. After a detailed evaluation of these articles via title and abstract screening, 21 studies remained for assessment. After applying the selection criteria, 7 studies6,11,12,14–17 were selected for further analysis. The study characteristics of these trials are summarized in the Table. All included studies were population based with the exception of the study of Gross et al,14 which was a retrospective institutional review over a 7-year period. Risk of bias assessment for included studies is shown in Table I in the online-only Data Supplement. Specific study population details for each included study is outlined in Table II in the in the online-only Data Supplement.
Risk of aSAH in Patients Currently Using Aspirin of Any Duration or Frequency
The risk of aSAH in aspirin users (of any duration and frequency) was compared with that that in nonaspirin users and is summarized in the Forest plot in Figure 2. There was no significant difference noted between aspirin and nonaspirin users (OR, 1.00; 95% CI, 0.81–1.24; I2=72%; P=0.99), however, there was significant heterogeneity (P<0.0001). To determine potential sources of heterogeneity, we performed subgroup analyses based on the duration of aspirin use and the frequency of aspirin use per week.
Risk of aSAH With Aspirin Use <3 Months in Duration
The relationship between aspirin use in the short-term (<3 months) and the risk of aSAH was reported in 3 studies11,12,16 (Figure 3A). There was a significant association between short-term use of aspirin and the risk of aSAH (OR, 1.61; 95% CI, 1.20–2.18; I2=37%; P=0.002), and no significant heterogeneity was detected (P=0.19).
Risk of aSAH in Aspirin Use of 3- to 12-Month Duration
These subgroups were reported in 3 studies11,15,16 (Figure 3B). No significant association was found between aspirin use in the short to medium term and the risk of aSAH (OR, 1.12; 95% CI, 0.79–1.58; I2=78%; P=0.53). Significant heterogeneity was noted (P=0.01).
Risk of aSAH in Aspirin Use of 1- to 3-Year Duration
These subgroups were reported in 3 studies11,12,16 (Figure 3C). No significant association was found between aspirin use in the long term and the risk of aSAH (OR, 1.01; 95% CI, 0.87–1.17; I2=0%; P=0.86). No significant heterogeneity was noted (P=0.95).
Risk of aSAH in Aspirin Use >3 Years in Duration
These subgroups were reported in 4 studies6,11,16,17 (Figure 3D). No significant association was found between aspirin use in the very long term and the risk of aSAH (OR, 0.92; 95% CI, 0.68–1.25; I2=64%; P=0.60). Significant heterogeneity was noted (P=0.004).
Risk of aSAH in Current Aspirin Use ≤2× Per Week
These subgroups were reported in 2 studies6,17 (Figure 4A). No significant association was found between infrequent aspirin use (≤2× per week) and the risk of aSAH (OR, 0.85; 95% CI, 0.55–1.33; I2=0%; P=0.48). No significant heterogeneity was noted (P=0.79).
Risk of aSAH in the Current Aspirin Use ≥3× Per Week
These subgroups were reported in 2 studies6,17 (Figure 4B). No significant association was found between infrequent aspirin use (≥3× per week) and the risk of aSAH (OR, 1.00; 95% CI, 0.49–2.05; I2=76%; P=1.00). Significant heterogeneity was noted (P=0.005).
Potential publication bias in the present study was tested using several methods. Egger test for publication bias was also nonsignificant (P=0.38, 2 tailed). Publication bias was assessed using funnel plot methodology, shown in Figure 5. Trim-and-fill analysis was performed to determine whether there were any missing studies in the literature that could contribute as a source of publication bias (Figure 5). One study was noted to be missing according to trim-and-fill analysis, but when this was imputed theoretically, there was no significant change to the pooled effect size. As such, publication bias was not a significant influencing factor in the pooled effect size of aspirin use and the risk of aSAH.
As we continue to expand our understanding of aneurysm pathophysiology and their complications, there is increasing importance placed on understanding and potentially mitigating the risk factors that give rise to aSAH. Given the widespread use of aspirin and antiplatelet agents, establishing the association, if any, and direction of association with regard to the risk of aSAH is of utmost importance from both clinical and public health perspectives. To address this question, we performed a systematic review and meta-analysis of available studies exploring the association between aspirin use and the risk of aSAH. Our analysis included several patient subgroups. We showed that when considered collectively, aspirin use, regardless of frequency and duration, was not associated with a significantly increased risk of aSAH when compared with no therapy. In subgroup analysis, patients on aspirin of <3-month duration had significantly higher risk of aSAH. For patients on aspirin for intermediate-term (3–12 months; 1–3 years) and longer-term (>3 years) duration, no significant difference in the risk of aSAH was found between aspirin users and nonaspirin users. Subgroup analysis according to the frequency of aspirin use per week (<2× per week) or (≥3× per week) demonstrated no significant difference between aspirin and nonaspirin users.
Studies Finding an Association Between Aspirin and a Reduction in aSAH
The possible protective effect of aspirin on risk of aSAH was first reported by Hasan et al,6, in their nested case-controlled study from the ISUIA study (International Study of Unruptured Intracranial Aneurysms). Fifty-eight patients were retrospectively matched to 4 controls each, all of which had been followed up for 5 years as part of the ISUIA study. Interestingly, they showed that aspirin use at least 3× per week was associated with a lower odds of aSAH (OR, 0.27; 95% CI, 0.11–0.67; P=0.03). García-Rodríguez et al16 also reported a large study of 1340 aSAH cases compared with 10 000 controls and demonstrated a significant association between aspirin use for >3-year duration with reduced risk of aSAH (OR, 0.82), whereas no increased risk of intracerebral hemorrhage was identified. Although both studies demonstrated protective effects for aspirin use, it must be noted that this effect was only seen in a subset of patients taking aspirin: patients taking aspirin ≥3× weekly for the ISUIA trial and patients using aspirin for ≥3 years for the study by García-Rodríguez et al.16 Both of these studies are also limited by bias inherent to retrospective data analysis. For the ISUIA trial, no benefit was found with regard to the risk of aSAH when aspirin was used up to twice a week or when used monthly. With regard to the European study, there was no benefit observed for aspirin users in the short and intermediate term (<3 years).
Studies Finding an Association Between Aspirin and an Increase in aSAH
Any potential benefits of aspirin need to be carefully considered and balanced with the possible risk of increased bleeding events or more severe hemorrhage in the event of rupture. Indeed, there have been conflict reports on the influence of aspirin on patients and their subsequent risk of aSAH. Pottegard et al11 performed a large study of Danish neurosurgery units, matching 5834 patients presenting with their first aSAH diagnosis with 40 age-, sex-, and period-matched population controls. The authors reported no reduction in aSAH risk with the long-term aspirin use (OR, 1.11). However, they reported that short-term aspirin use <3 months was significantly associated with an increased risk of aSAH. Similar conclusions were attained by Schmidt et al12 in patients with new use of aspirin. The results of our own meta-analysis demonstrated a significant association between aspirin use of <3 months and an increased risk of aSAH, with OR of 1.61. As such, the prophylactic use of aspirin for aSAH, particularly with known unruptured aneurysms, is not without its own inherent bleeding risks, particularly in the short term.
Aspirin and aSAH: Systematic Review and Meta-Analysis
In the present meta-analysis, we performed a subgroup analyses on the literature based on the duration of aspirin use, as well as the frequency of aspirin use per week. For long-term aspirin use, defined as >3-year duration, pooled data from 4 cohorts did not demonstrate any significant decrease or increase of aSAH risk with the aspirin use (OR, 0.92). This was similarly the case for the intermediate duration use of aspirin for 3 to 12 months (OR, 0.96) and 1 to 3 years (OR, 1.01). As such, based on the limited evidence published to date on this topic, there is no clear consistent evidence linking long-term aspirin use to either a decreased or increased risk of aSAH. What the current data suggest is that if aspirin does provide any beneficial effect in the context of aSAH risk, this effect will be dependent on the duration of aspirin use, and further studies are required to elucidate the optimal patient population who are likely to benefit from such recommendations. The optimal way to address the question whether aspirin can protect against aSAH would be a large, multicenter, randomized trial. Such a trial may be challenging as to what duration and frequency of aspirin to choose.
It has been previously proposed that aspirin may exert a protective affect on aSAH risk via its antiplatelet activity, which can impede thrombus formation, endothelial injury, irritation, and inflammation of the aneurysm wall. However, Pottegard et al11 in a Danish study did not demonstrate any benefit for either aspirin or clopidogrel in terms of the risk of aSAH, which brings into question this hypothesis. Another theory suggests that aspirin may have an inhibitory effect on the inflammatory cascade that is implicated in intracranial aneurysm formation and rupture.8,18–21 Numerous cells and molecular pathways have been implicated in this inflammatory cascade including macrophages, vascular smooth muscle cells, nuclear factor-κB, interleukin-1β, matrix metalloproteinases, and tumor necrosis factor-α.6,18,22 Hasan et al19 demonstrated, in a small randomized sample of patients with unruptured aneurysm who underwent microsurgical clipping after 3 months of aspirin treatment, a decreased expression of inflammatory cells and markers such as cyclooxygenase-2. Ferumoxytol-enhanced magnetic resonance imaging, a marker of macrophage activity, showed a reduction in signal intensity after aspirin treatment.23 Certainly, such findings, in addition to the current lack of affordable and noninvasive options for patients at risk of aSAH, make aspirin a promising candidate. Our results, however, demonstrate a distinct lack of evidence to strongly recommend aspirin as a potential medical therapy to reduce aSAH in the presence of unruptured aneurysms.
The present study is constrained by several limitations. Firstly, none of the included studies were randomized trials and as such, all included studies were susceptible to risk of selection bias and study population heterogeneity. We made every effort to include adjusted odds in our pooled analyses although each study adjusted for a different set of confounders. There was only 1 study14 we included where the odds were derived from the data provided in the published report and were thus unadjusted. Second, a major limitation in pooling available results is the heterogeneity in patient population, the duration of use of aspirin, and the frequency of aspirin use. One of the main limitations of using OR in studies with long-term follow-up is that time is not taken into account. Differences in short- and long-term findings may be because of selection bias. In some cases, patients who were followed up for long term and who were benefited or did not show harm may be in fact because of their underlying baseline characteristics—such as being younger and healthier. The population size and thus statistical power for each studied varied and is a limitation of the present analysis. We attempted to mitigate the above heterogeneity by performing subgroup analysis according to duration and frequency of aspirin use, as well as using adjusted OR whenever reported. We also used a random-effects model to incorporate heterogeneity into the study. However, we do acknowledge that the current results still have substantial heterogeneity and thus should be interpreted with caution. Future studies should ensure that their study design accounts for such factors.
Based on the limited quality evidence in the literature, we found that patients on aspirin for <3 months had a significantly higher risk of aSAH (OR 1.6). For patients on aspirin in the intermediate and long term (3–12 months, 1–3 years, and >3 years), no significant difference in the risk of aSAH was found when compared with nonaspirin users. Subgroup analysis according to the frequency of aspirin use per week (<2× per week or ≥3× per week) demonstrated no significant difference between aspirin and nonaspirin users. Ideally, a well-designed, statistically powered, randomized controlled trial is required to determine whether long-term aspirin use can provide any benefit in terms of risk of aSAH.
K. Phan, Dr Moore, Dr Griessenauer, Dr Ogilvy, and Dr Thomas helped in study design and proposal. K. Phan, Dr Moore, Dr Griessenauer, and Dr Thomas helped in data collection, analysis, and interpretation. K. Phan, Dr Moore, Dr Griessenauer, and Dr Thomas helped in drafting the article. Dr Ogilvy and Dr Thomas helped in revision of the article and provided senior expert input. All authors approved the final submitted version of this article.
Guest Editor for this article was Seemant Chaturvedi, MD.
The online-only Data Supplement is available with this article at http://stroke.ahajournals.org/lookup/suppl/doi:10.1161/STROKEAHA.116.015674/-/DC1.
- Received October 6, 2016.
- Revision received January 15, 2017.
- Accepted February 1, 2017.
- © 2017 American Heart Association, Inc.
- Jalbert JJ,
- Isaacs AJ,
- Kamel H,
- Sedrakyan A
- Hasan DM,
- Mahaney KB,
- Brown RD Jr,
- Meissner I,
- Piepgras DG,
- Huston J,
- et al
- Chalouhi N,
- Points L,
- Pierce GL,
- Ballas Z,
- Jabbour P,
- Hasan D
- Garbe E,
- Kreisel SH,
- Behr S
- Iso H,
- Hennekens CH,
- Stampfer MJ,
- Rexrode KM,
- Colditz GA,
- Speizer FE,
- et al
- Hasan DM,
- Chalouhi N,
- Jabbour P,
- Dumont AS,
- Kung DK,
- Magnotta VA,
- et al
- Loch Macdonald R
- Chalouhi N,
- Jabbour P,
- Hasan D,
- Starke RM