Introduction

Methamphetamine abuse has emerged as a risk factor for both hemorrhagic and ischemic stroke in recent years1,2,3,4,5,6,7,8,9,10. Although population-based study and forensic analysis of fatal strokes showed significant predominance of hemorrhagic stroke in methamphetamine users3,6,7, methamphetamine abuse has also been increasingly reported to be associated with acute ischemic stroke (AIS)1,2,4,5,6.

The mechanisms by which methamphetamine causes stroke are still unknown. Case studies and forensic analysis showed atherosclerotic stenoses, arterial dissection, and berry aneurysms in patients with methamphetamine-associated stroke2,4,6,7. Methamphetamine abuse was also shown to produce a dose-dependent elevation of blood pressure and chronic hypertension11,12,13.

Chronic hypertension is a major risk factor for cerebral small vessel disease (cSVD) and stroke13,14,15,16. cSVD refers to a group of pathological processes that affect the small perforating vessels and capillaries in the brain16,17. Radiographically, it is characterized by deep white matter hyperdensities (WMHs), lacunar infarcts, microbleeds, and enlarged perivascular spaces (PVS) on magnetic resonance imaging (MRI) of the brain17. cSVD is a common cause of stroke, cognitive impairment, and vascular dementia15,16,17.

The aim of this study is to investigate whether methamphetamine abuse increases the risk of cSVD in patients with acute ischemic stroke.

Methods

This retrospective study was approved by the University of California Irvine Institutional Review Board (IRB) and the Ethics Committee. Informed consents were waived due to retrospective study design and minimal harm to the patients. All methods in the study were performed in accordance with the relevant guidelines and regulations.

Study population

Consecutive AIS patients admitted at the University of California Irvine Medical Center from January 1, 2013 to December 30, 2018 were included. The patient list was generated from the prospectively maintained American Heart Association (AHA)'s Get-With-The-Guideline stroke data registry at our medical center. The registry uses a web-based patient management tool to collect clinical data on consecutively admitted patients, to provide decision support, and to enable real-time online reporting18. Patients with TIA, stroke mimics, subacute stroke, inpatient stroke, and primary brain hemorrhage were excluded. Stroke transfers from outside facilities were also excluded.

All patients underwent standard diagnostic evaluation and treatment per AHA guidelines19. Based on the history of methamphetamine abuse and/or urine drug screen (UDS), patients were divided into Meth and Non-Meth group. The UDS was performed using EMIT II Plus Amphetamines Assay (Beckman Coulter, Inc) with a sensitivity and specificity of 94.3% and 93.3%, respectively20.

Study parameters

MRI images of the brain were reviewed by an experienced neurologist (Zhu Z) to assess cSVD using modified Fazekas scale13,14,15,21. Both deep and periventricular WMHs were rated from 0 to 3 on fluid-attenuated inversion recovery (FLAIR)- and T2-weighted sequences15. Lacunes were defined as small (< 15 mm) subcortical infarcts13,21. Cerebral microbleeds (CMBs) were rated on susceptibility-weighted imaging (SWI) based on their numbers (< 5, 5–10, or > 10). Enlarged perivascular spaces (PVS) were defined as small (< 3 mm) punctate hyperintensities on FLAIR or T2-weighted images and rated from 0 to 3: 0 (absent), 1 (< 10), 2 (10–25), and 3 (> 25).

Total burdens of cSVD were rated from 0 to 413. One point was added for each of the following findings: confluent deep or periventricular WMHs (Fazekas grade 2 and 3), ≥ 1 lacune, ≥ 1 CMBs, and > 10 enlarged PVS in the basal ganglia on at least one side of the brain.

Propensity score matching

Propensity score matching with a 1:1 ratio was performed to select patients from the Non-Meth group as control. The propensity score was estimated using a logistic regression model based on age, sex, hypertension, diabetes mellitus, hyperlipidemia, obesity, and initial NIHSS score as described (1:1 match, nearest neighbor approach)22.

Statistical analysis

Continuous variables were described by mean ± standard deviation (SD) or median with interquartile range (IQR) based on the results of normality testing. Categorical variables were expressed by counts with percentages. Baseline characteristics and outcome at discharge were compared between Meth and Non-Meth groups by Mann Whitney test for continuous variables and chi-square test for categorical variables. Sensitivity analysis was performed to investigate the effect of methamphetamine abuse and other variables on the development of cSVD. All statistical analyses were performed using SPSS software (version 23.0). A 2-tailed value of p < 0.05 was considered statistically significant. All statistical analyses were reviewed and verified by a biostatistician (Qian T).

Results

From the 1,369 patients with AIS, 63 were found to have a history of methamphetamine abuse or positive UDS. 2 patients were identified to have a positive UDS due to stimulant use for Attention Deficit Hyperactivity Disorders (ADHD) and were excluded from the study. Of the 61 patients in the Meth group, 6 had no MRI study of the brain and were excluded for the investigation of cSVD. In the remaining 55 patients, 7 had concomitant cocaine use and were excluded during the final statistical analysis. In patients with a negative history of meth use or UDS, 9 had cocaine use and was excluded. A 1:1 propensity score-matched control group was selected from the Non-Meth group for sensitivity analysis. The study flowchart is shown in Fig. 1.

Figure 1
figure 1

Study flowchart.

Full size image

The characteristics of the Meth and Non-Meth groups are summarized in Table 1. Compared with Non-Meth group, patients in the Meth group were significantly younger (54.5 ± 9.7 vs. 70.5 ± 12.4, p < 0.001), more likely male (78.7% vs. 54.0%, p < 0.001) and White (78.7% vs. 50.4%, p < 0.001). There appeared to be proportionally less Asians with methamphetamine-associated stroke (3.3% vs. 24.1%, p < 0.001).

Table 1 Demographics and clinical characteristics in the Meth group versus Non-meth group.
Full size table

Non-Meth group had a statistically significant higher rate of hypertension, hyperlipidemia, and statin use than in the Meth group. There was no statistically significant difference between the two groups in the history of diabetes, obesity, antithrombotic use, initial NIHSS score, baseline LDL cholesterol levels, or mechanisms of stroke per TOAST classification (Table 1). However, the patients in the Meth group had a higher rate of smoking (57% vs. 30.1%, p < 0.001) or cocaine use (11.5% vs. 0.7%, p < 0.001) than those in the Non-Meth group.

Sensitivity analysis

Sensitivity analysis was performed to investigate the effect of methamphetamine use as independent variable. Compared with age- and sex-matched control group, there were proportionally more White and less Asian patients in the Meth group (Table 2). There was no significant difference in the history of hypertension, diabetes, hyperlipidemia, obesity, or initial NIHSS score between the Meth group and propensity score matched control group.

Table 2 Comparison of cerebral SVD in Meth versus propensity score-matched control group after removing the patients with stimulant use for ADHD.
Full size table

The representative MRI images of cSVD were shown in Fig. 2. The Meth group were found to have more WMHs, lacunes, and total burdens of cSVD than propensity score-matched control group (Table 2).

Figure 2
figure 2

Representative images of cerebral small vessel disease (cSVD). One point is added for each of the following findings to estimate cSVD score: severe white matter hyperintensities (periventricular Fazekas grade 2 or 3); ≥ 1 lacunar infarcts; ≥ 1 cerebral microbleed; and > 10 perivascular spaces visible in the basal ganglia on at least one side of the brain.

Full size image

There were no significant differences in systolic or diastolic blood pressure (BP) at admission or hospital discharge, DBP reduction during hospitalization, and numbers of antihypertensives required for BP control between the Meth and propensity score-matched control group. The statistically more significant reduction in SBP in the control group was likely due to higher baseline SBP than in the Meth group (Table 2). These results suggested that the increased burdens of cSVD in the Meth group are independent of age, sex, comorbidities, acute hypertension, and stroke severity as measured by initial NIHSS scores.

We performed an additional sensitivity analysis after excluding the 7 patients with concomitant cocaine use in the Meth groups. As showed in Table 3, the results were consistent with what were shown in Table 2. Methamphetamine use remains an independent predictor for cSVD.

Table 3 Comparison of cerebral SVD in the Meth group versus PS-matched control group after removing the patients with stimulant use for ADHD and cocaine use.
Full size table

Discussion

This single center data demonstrates that methamphetamine abuse is seen in 4.5% of the patients with acute ischemic stroke at our medical center. The patients with methamphetamine use are significantly younger (54.5 ± 9.7 vs. 70.5 ± 12.4, p < 0.001) and more likely male (78.7% vs. 54.0%, p < 0.001). In addition, sensitivity analysis with propensity score-matched controls showed that methamphetamine abuse is associated with increased burdens of cSVD, independent of age, sex, cocaine use, comorbidities, acute hypertension, and stroke severity.

Previous studies showed that elevated blood pressure levels are associated with each of the MRI markers of cSVD14,15. Effective treatment of hypertension may reduce the rates of cSVD and stroke13,15. In this retrospective study, there was no significant difference in blood pressures at admission or hospital discharge, blood pressure reduction during hospitalization, and the numbers of antihypertensives required to control hypertension between the Meth or Meth + group and propensity score-matched Non-Meth controls. It is possible that chronic hypertension from methamphetamine abuse plays an important role in the development of cSVD13.

Advanced age and male sex were reported to be major risk factors for cSVD16,23,24,25. In a recent study, cerebral SVD was seen in 18.9% of age group 70s as compared to 3% in age group 40s23. In a Chinese population-based study, advanced age was shown to be independently associated with the prevalence of cSVD24. In our study, we demonstrated that methamphetamine abuse increases the risk of cSVD in young stroke patients.

In a cohort study of homeless and unstably housed women (n = 30) from San Francisco community, 86% patients had a history of cocaine use and 54% patients had WMHs25. In our cohort, more patients in the Meth group had a history of cocaine abuse than the Non-Meth group (11.5% vs. 0.7%). After excluding patients with cocaine use in both Meth and Non-Meth groups, methamphetamine use remains an independent predictor for cSVD.

Although the Non-Meth group had statistically significant higher rate of hyperlipidemia (40.0% vs. 26.2%, p = 0.028) and statin use (40.5% vs. 15%, p < 0.001) than the Meth group, there was no statistically significant difference in baseline LDL cholesterol levels between the 2 groups. In the sensitivity analysis with propensity-score matched control, Meth use remains an independent predictor of cSVD after adjusting for hyperlipidemia.

There were also significantly more patients with a history of smoking in the Meth group than in the Non-Meth group (57% vs. 30%, p < 0.001). Since previous study demonstrated a dose-dependent association between pack-years of smoking and WMH progression26, smoking is likely a confounding factor in our retrospective study.

There was no statistically significant difference in stroke mechanisms between the Meth and Non-meth group. The Non-meth group had a trend of more cardioembolic stroke, likely due to significantly older ages. There was only one case of possible Reversible Cerebral Vasoconstriction Syndrome (RCVS) as the possible stroke mechanism in the Meth group.

Since cSVD increases the risk of cognitive impairment and vascular dementia, methamphetamine abuse in young adults may have more significant long-term public health concerns than stroke and other detrimental effects.

Our study has a few limitations. First, our data suggested a possible association and cannot prove causal relationship between methamphetamine use and cSVD. Additional studies are required to corroborate our findings. Second, some patients were unable to provide history of drug use due to aphasia or severe neurological deficit. Only 932 patients (71%) in the Non-Meth group had a UDS. The rate of methamphetamine abuse was likely underestimated. Third, there was no information regarding the route, frequency, and duration of methamphetamine abuse. Lastly, smoking, and other confounding factors cannot be ruled out in this retrospective study. Further studies are warranted to adjust for all confounding factors and to investigate the temporal relationship between methamphetamine abuse and cSVD.

Of note, it was challenging to get accurate information on polysubstance abuse, particularly, in patients with neurological deficit. It may be also unethical to conduct randomized controlled studies on methamphetamine abuse. Well-designed prospective registry may be a good option to further investigate the effect of chronic methamphetamine use on cSVD.

In conclusion, our preliminary results demonstrates that methamphetamine abuse is common in young adults with acute ischemic stroke and increases the risk of cSVD. Given increased prevalence of methamphetamine abuse in young adults, additional studies are warranted to investigate the effects of chronic methamphetamine use on the pathogenesis of stroke, cSVD, and vascular dementia.