Article Text

Download PDFPDF

Original research
Methamphetamine use is an independent predictor of poor outcome after aneurysmal subarachnoid hemorrhage
  1. Karam Moon1,
  2. Felipe C Albuquerque1,
  3. Mario Mitkov2,
  4. Andrew F Ducruet1,
  5. David A Wilson1,
  6. R Webster Crowley1,
  7. Peter Nakaji1,
  8. Cameron G McDougall1
  1. 1Division of Neurological Surgery, Barrow Neurological Institute, St Joseph's Hospital and Medical Center, Phoenix, Arizona, USA
  2. 2School of Medicine, Creighton University, Omaha, Nebraska, USA
  1. Correspondence to Dr Felipe C Albuquerque, c/o Neuroscience Publications, Division of Neurological Surgery, Barrow Neurological Institute, St Joseph's Hospital and Medical Center, 350 W Thomas Road, Phoenix, AZ 85013, USA; Neuropub{at}


Background Clinical outcomes of methamphetamine users with aneurysmal subarachnoid hemorrhage (aSAH) are unknown.

Objective To analyze differences in presentation, in-hospital morbidity, and outcomes between methamphetamine users and non-users.

Methods All 472 patients included in the Barrow Ruptured Aneurysm Trial from 2003 to 2007 were reviewed. Patients with 1- and 3-year follow-up were included in this analysis (n=398). Methamphetamine users were identified as patients who provided a history of methamphetamine use on admission or tested positive on urine toxicology testing. Methamphetamine users were compared with non-users using univariate analysis. Outcomes were then analyzed using multivariate logistic regression models for demographic characteristics, medical comorbidities, radiographic and clinical presentation, and vasospasm.

Results Thirty-one patients (7.8%) were identified as methamphetamine users in this cohort. Methamphetamine users were younger than non-users (mean age 42.8 vs 55 years, p<0.001). In multivariate logistic regression models, methamphetamine use was an independent predictor of poor Glasgow Outcome Scale score at both 1 year (OR=5.02; 95% CI 1.03 to 24.48; p<0.05) and 3 years (OR=7.18; 95% CI 1.73 to 29.87; p=0.007). Other independent predictors in this model included older age, clinical vasospasm, diabetes, and aneurysm size. Cocaine and tobacco use were not significantly associated with poor outcome in our cohort. Methamphetamine use was not significantly associated with vasospasm, higher Fisher or Hunt and Hess grade, or intraparenchymal hemorrhage/intraventricular hemorrhage.

Conclusions Methamphetamine users have significantly worse outcomes at 1 and 3 years following aSAH. Further analysis is necessary to understand the pathological response associated with methamphetamine use in this setting.

  • Aneurysm
  • Flow Diverter
  • Angiography
  • Catheter
  • Device

Statistics from


Methamphetamine is a psychoactive amphetamine with highly potent stimulant properties. Although the sale and possession of methamphetamine is restricted in the USA, its use increased by 500% from 1992 to 2002, and an estimated 1.2 million people reported its use in 2012.1 ,2 Methamphetamine abuse has been associated with cerebrovascular complications such as ischemic stroke, subarachnoid hemorrhage, intraparenchymal hemorrhage (IPH), and vasculitis, although its mechanistic effects on the cerebrovascular architecture are not clear.1 ,3–8 Furthermore, recent data have shown an association between aneurysmal subarachnoid hemorrhage (aSAH) and methamphetamine use, but long-term outcomes in follow-up have not yet been demonstrated in this cohort.9

The incidence of aSAH and its sequelae are well described, with the most serious in-hospital complications, including rebleeds, vasospasm, ischemic infarcts, and hydrocephalus. The prevalence of aSAH has also been reported in groups with a history of substance use, including tobacco and illicit drugs such as cocaine, with some series reporting higher rates of in-hospital morbidity and worse outcomes.10–16 Interestingly, while younger patients have traditionally been thought to have better outcomes after aSAH, they are also more likely to have a history of substance use.1 ,17

Given that younger age is a predictor of better outcome after aSAH, the clinical prognoses for methamphetamine users in this setting are uncertain. To better understand the effects of this agent on clinical outcome, we compared a large series of methamphetamine users with non-users in the setting of aSAH to determine differences in patient characteristics, clinical presentation, complications such as vasospasm, and outcomes at follow-up.

Patients and methods

Patient selection and characteristics

The Barrow Ruptured Aneurysm Trial (BRAT) was a randomized prospective trial comparing microsurgical clip occlusion and coil embolization for the treatment of ruptured cerebral aneurysms, with intention-to-treat analysis. This study and its methodology was published previously.18 All 472 patients included in BRAT, spanning a period of 2003 to 2007, were analyzed in this subset study. The study database and patient records were reviewed for patient characteristics, radiographic data, and outcomes at 1 and 3 years. Methamphetamine users were identified on admission by a history of methamphetamine use given on interview by the patient/surrogate or a positive urine toxicology test. Vasospasm was classified as radiographic, clinical, or both. Radiographic vasospasm was defined as narrowing of major intracranial vessels on CT angiography or DSA. Clinical vasospasm was defined as a combination of radiographic vasospasm and attributable neurological deficit in the absence of other obvious organic causes such as new intracranial hemorrhage. Vasospasm was not defined or diagnosed solely by transcranial Doppler ultrasonography at any time. BRAT study patients were excluded from this subset analysis if they had a subarachnoid hemorrhage that was determined to be angiographically negative for an intracranial aneurysm. Furthermore, patients were excluded if an adequate social history could not be determined or obtained.

The demographic characteristics of methamphetamine users were compared with those of non-users using univariate analysis. Multivariate logistic regression models were then used to predict independent predictors of poor outcome, dichotomized by Glasgow Outcome Scale (GOS) outcomes at 1 and 3 years (poor scores=1–3, good scores=4 or 5), including demographic data, clinical presentation grades, and presence or absence of vasospasm during hospitalization.

Statistical analysis

χ2 Tests were used to investigate association between categorical variables. Independent sample t tests were used to compare proportion of patients with demographic characteristics presenting with, versus without, clinical vasospasm. Wilcoxon rank sum tests were used to compare distributions of variables with ordinal scales. Multivariate logistic regression models were then used to calculate ORs and predict dichotomous outcome variables. All analyses were done using SPSS V.20.


Patient characteristics and univariate comparison

In the trial database, 398 patients met criteria for inclusion in this study, of whom 31 patients (7.8%) were identified as methamphetamine users. Fifty-seven angiographically negative patients were excluded, of whom four were methamphetamine users. Clinical characteristics and presentation of the included cohorts were analyzed using univariate analysis (table 1). Of methamphetamine users, 13 (41.9%) were male and 18 (58.1%) were female, although there was no significant difference in gender makeup between the two cohorts (p=0.15). Methamphetamine users were significantly younger than non-users, with a mean age of 42.8 vs 55.0 years (p<0.001). Patients with a history of methamphetamine use were also more likely to have a history indicative of polysubstance use, with significantly higher rates of usage of cocaine (p<0.001) and tobacco (p<0.001). Non-users were more likely to have a history of hypertension than methamphetamine users (p=0.01), while there was no difference in the history of diabetes. In addition, there was no significant difference in the distribution of anterior versus posterior circulation aneurysms between the two cohorts (p=0.96). There were also no significant differences in measures of radiographic presentation, including presence of intraparenchymal hemorrhage (IPH) (p=0.50), intraventricular hemorrhage (IVH) (p=0.45), or blood in the fourth ventricle (p=0.68). Mean aneurysm size was 7.04±3.92 mm for methamphetamine users and 6.48±3.67 mm for non-users (p=0.27).

Table 1

Univariable comparison of demographics and radiographic presentation between patients with aneurysmal subarachnoid hemorrhage with and without a history of methamphetamine use

Rates of radiographic and clinical vasospasm were not significantly different between the two cohorts (p=0.35 and p=0.31, respectively). No correlation was seen between the location of aneurysms, divided into anterior and posterior circulation, and the presence of either clinical or radiographic vasospasm (p=0.54 and p=0.99, respectively). In addition, treatment modality (clip vs coil) was not significantly associated with either radiographic or clinical vasospasm (p=0.29 and p=0.27, respectively). Patients with clinical vasospasm were more likely to be male (p=0.009) and have radiographic evidence of vasospasm (p<0.001). Patients with radiographic vasospasm were more likely to have an IPH (p=0.02), and suffer clinical vasospasm (p<0.001). In addition, rates of radiographic vasospasm were seen to increase significantly with worsening of Fisher grades (p=0.01).

Distributions of admission Hunt and Hess grades, Fisher grades, and GOS scores at 1 and 3 years were analyzed using a Wilcoxon rank sum test (figures 1 and 2). Distributions of Hunt and Hess grades (p=0.29) or Fisher grades (p=0.33) upon admission did not vary significantly with methamphetamine use. However, GOS scores at 1 year, using the same model, demonstrated significantly worse outcomes for methamphetamine users (p=0.04). At 3 years, GOS scores showed a trend towards worse outcomes for methamphetamine users (p=0.06).

Figure 1

Distributions of admission Hunt and Hess grades for patients with aneurysmal subarachnoid hemorrhage and without a history of methamphetamine use. There was no significant difference in the distributions (p=0.29). Reproduced with permission from the Barrow Neurological Institute.

Figure 2

Distributions of Glasgow Outcome Scale (GOS) scores for patients with aneurysmal subarachnoid hemorrhage and without a history of methamphetamine use. (A) Distributions at 1 year, demonstrating significantly worse GOS scores for methamphetamine users (p=0.04). (B) Distributions at 3 years, demonstrating a trend towards worse GOS scores for methamphetamine users (p=0.06). Reproduced with permission from the Barrow Neurological Institute.

Multivariate comparison

Multivariate logistic regression models were used to calculate ORs and predict dichotomous outcome variables for patients at 1- and 3-year follow-up (table 2). The 1-year logistic regression model predicted 80% of cases accurately and was significant, χ2 (6, N=341)=41.56, p<0.001. Independent predictors of poor outcome at 1 year included age (p<0.001), aneurysm size (p=0.001), methamphetamine use (p=0.048), and clinical vasospasm (p=0.003; table 3). The 3-year logistic regression model predicted 76% of cases accurately and was significant, χ2 (6, N=337)=42.26, p<0.001. Results were similar, with history of diabetes becoming significant (p=0.05), in addition to age (p<0.001), aneurysm size (p=0.009), methamphetamine use (p=0.007), and clinical vasospasm (p=0.034). Interestingly, cocaine use was not seen to be an independent predictor of poor outcome in either model.

Table 2

Multivariate logistic regression model predicting poor GOS score in patients with aneurysmal subarachnoid hemorrhage at 3-year follow-up

Table 3

Multivariate logistic regression model predicting poor GOS score in patients with aneurysmal subarachnoid hemorrhage at 1-year follow-up


The association between aSAH and substance abuse has been well-studied. Previous reports have indicated an increased incidence of aSAH in groups with a history of cocaine, alcohol, and tobacco use.11–13 ,19 ,20 Cocaine, in particular, has a dopaminergic mechanism of action similar to that of methamphetamines, and has been shown to be an independent predictor of vasospasm, although its effects on clinical outcome are less clear.11 ,12

Similarly, we sought to study the effects of methamphetamine use on the incidence of aSAH. While a causative relationship cannot be shown by retrospective analysis, it is clear that there is an association between methamphetamine use and aSAH in a younger cohort of patients. While the nationwide prevalence of methamphetamine use is quoted as 0.4%, rates of use in the BRAT study population were a surprising 7.8%.2 This is similar to the rate of 7.0% found in a recent series of patients with aSAH but much higher than the rate of 1.9% found for all cerebrovascular complications seen in methamphetamine users in another series.3 ,9 More importantly, the clinical prognosis of this subset of patients has been largely unknown until now. While Beadell et al9 found no statistical difference in GOS at discharge, a trend towards worse outcomes was seen.9 Methamphetamine users in that series were also significantly younger than non-users. As 1- and 3-year follow-up data from our analyses show, methamphetamine use is an independent predictor of poor outcomes, which represent the only clinical follow-up data for this cohort in the literature.

What remains unclear is the mechanism by which methamphetamines increase the risk of poor outcomes from aSAH. Methamphetamine is a sympathomimetic agent available in oral, inhaled, smoked, or intravenous form that results in the release of norepinephrine and dopamine from synaptic nerve endings, and blocks their reuptake. Its effects on systemic vasculature, including reports of vasculitis, cerebral angiitis, and accelerated atherosclerosis, suggest chronic and repetitive injuries to vessel wall architecture. Several reports have suggested that these injuries might be induced or exacerbated by hypertensive episodes, given the drug's mechanism of action.1 ,3 ,7 ,8 Therefore, it is certainly possible that these effects, both in the acute and chronic timeframe, may facilitate the formation of de novo aneurysms at an earlier age through worsening of flow turbulence and vessel wall inflammation. On the other hand, users of methamphetamine and other illicit substances have been shown to be significantly younger than their counterparts, which should promote relatively improved clinical outcomes following aSAH. Analysis of in-hospital morbidity did not prove to be as revealing or intuitive as expected. No significant differences in clinical or radiographic presentation at the outset of hemorrhage were found, implying that the outcomes of this cohort are independent of Fisher or Hunt and Hess grades. Therefore, it is likely that the detrimental effects of methamphetamine use at follow-up negate any protective benefits of younger age in this cohort.

Methamphetamine use did not have the expected effect on rates of clinical or radiographic vasospasm. Systemically, the agent has been associated with increased incidence of coronary vasoconstriction and cardiomyopathy.2 ,21 Additionally, methamphetamine-induced cerebral vasospasm has been demonstrated in animal models.5 Beadell et al9 found significantly higher rates of vasospasm among methamphetamine users, although vasospasm was partially defined by transcranial Doppler measurements in their report, which is not typically used as a primary tool for diagnosis of vasospasm at our institution.

The incidence of either type of vasospasm did not differ between the two cohorts in this series, and little was revealed about predictors of vasospasm. The reasons for this are probably several-fold. The effects of methamphetamine on systemic vasculature are difficult to generalize to intracranial vasculature in the presence of aSAH, given the unique environment and inflammatory pathways inherent to the latter. Accordingly, there is a growing body of evidence that cerebral vasospasm in the presence of aSAH is primarily a molecular phenomenon, and thereby difficult to simplify into a retrospective analysis of general risk factors. Many theories have been put forward, including the effects of oxidative stress on smooth muscle cell activation, leading to phenotypic changes and remodeling of vessel wall architecture. The effects of breakdown products from acute clotting, such as ferrous hemoglobin, have been studied, recognizing the imbalance between vasoconstrictors, such as endothelin 1, and vasodilators, such as nitrous oxide.22–24 Needless to say, it is likely that traditional models of vasospasm in neurocritical care represent only the tip of the iceberg. Lastly, our inclusion criteria included patients with a history of methamphetamine use, without necessarily taking into account the temporal proximity of their use to their hemorrhage. Therefore, it is possible that the immediate vasoactive effects of methamphetamines were diluted by the subpopulation of patients who had a more distant history of illicit use, thereby diminishing a correlation with vasospasm. Further subanalysis or prospective studies of methamphetamine users shown to be positive by toxicology screening at the time of admission may show statistically higher rates of vasospasm.

Nevertheless, these data suggest that clinical management of patients with a known history of methamphetamine use should differ from that of their counterparts. Clearly, methamphetamine users present with aneurysmal rupture (and likely aneurysmal development) at an earlier age, which raises the possibility that the natural history of intracranial aneurysms in this cohort may be accelerated, although causation cannot be inferred at this time. The implications of this affect surveillance and management strategy, particularly for unruptured lesions; perhaps more aggressive treatment is reasonable, given their increased propensity for rupture and the subsequently, poorer outcomes in patients with these lesions. Furthermore, methamphetamine users with aSAH have poorer prognoses, and clinical and radiographic follow-up after rupture should be modified accordingly.

The methodology of this study has inherent shortcomings. Data were collected by retrospective review of a previously completed randomized clinical trial at a single institution. Given the regional variability of methamphetamine use, with higher rates in western USA, it is possible that the proportion seen at our institution is higher than at others. Methamphetamine users were identified by a history of abuse or positive urine toxicology test, which does not distinguish between acute and chronic effects of the substance on the burden of cerebrovascular disease. Moreover, continued or repeat use after hospitalization and its effects on clinical outcome at follow-up were not tracked. Methamphetamine users are more likely to be polysubstance users. While cocaine and tobacco use were significantly higher amongst methamphetamine users in this study, other illicit drugs and substances were not included. Similarly, additional medical comorbidities were not tracked, but might be associated with methamphetamine use. Although the effects of methamphetamine on long-term outcomes of other vascular diseases have not been studied, it is likely that socioeconomic differences between the cohorts had an effect on overall patient outcome. Interestingly, while other types of substance use, such as cocaine or tobacco use, often represent a similar demographic, neither was seen to be an independent predictor of poor outcomes. Finally, these outcomes are based on 1- and 3-year analyses of the BRAT data and do not portend longer-term outcomes, although it is reasonable to expect that similar trends will be seen at the BRAT 6- and 10-year analyses.


Methamphetamine use is more prevalent in patients presenting with aSAH than in the general population. Methamphetamine users present with ruptured aneurysms at a younger age and have significantly poorer outcomes at follow-up. Radiographic and clinical presentation of this cohort may not differ significantly from that of non-users and further research is warranted to elucidate mechanistic influences on outcomes following aSAH.


We acknowledge Kristina Chapple, PhD, for statistical analysis, and Clare Prendergast for technical support.



  • Contributors KM, FCA, AFD, RWC, and CGM are responsible for the conception and design of this work. KM, DAW, and MM were the primary data gatherers. Data interpretation and analysis was performed by KM, FCA, AFD, PN, and DAW. The article was drafted by MM, KM, FCA, and AFD. KM, FCA and AFD critically revised the article; all authors approved the final version of the manuscript. The study was supervised by FCA, CGM and PN.

  • Competing interests None.

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

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.