Trends in Substance Abuse Preceding Stroke Among Young Adults: a Population-Based Study

Abstract

Background and Purpose

About 5% of strokes occur in adults 18 to 44 years old. Substance abuse is a prevalent risk factor for stroke in young adults. We sought to identify trends in substance abuse detection among stroke patients.

Methods

Using a population-based design, we sought to identify all patients age 18-54 suffering a stroke (ischemic or hemorrhagic) in the Greater Cincinnati/Northern Kentucky Study region during 1993-4, 1999, and 2005. Demographic and clinical characteristics as well as substance use data were obtained retrospectively from chart review and adjudicated by physicians.

Results

The number of young patients identified with a stroke increased from 1993/1994 (297) to 2005 (501). Blacks (61% vs. 51%; p<0.02) and males (61% vs. 47%; p<0.002) reported substance abuse (current smoking, alcohol or street drug use) more frequently than whites and females. Overall use of substances increased across study periods, 45% in 1993 vs. 62% in 2005 (p=0.003). The trend was significant for street drug use (3.8% in 1993 vs. 19.8% in 2005) and ever smoking (49% in 1993 vs. 66% in 2005). Documentation of both cocaine and marijuana use increased over time. In 2005 half of young adults with a stroke were current smokers and 1 out of 5 abused street drugs.

Conclusions

Substance abuse is common in young adults suffering a stroke. The observed increase in substance abuse is contributing to the increased incidence of stroke in young adults”. Patients younger than 55 who suffer a stroke should be routinely screened and counseled regarding substance abuse.

Introduction

In 2007, national estimates from the Nationwide Inpatient Sample reported that 4.9% of all strokes in the US were in adults between the ages of 18 and 44.¹ Stroke in this young age group is associated with higher societal costs compared to older age groups.¹ More worrisome, stroke in this age group appears to be increasing.² In our population-based stroke study of Northern Kentucky and Southern Ohio, the stroke incidence in those aged 20 to 54 years old significantly increased from 1993 to 2005 (manuscript in press, Neurology). The reasons for this increase are unclear, but may involve earlier age of onset and increasing prevalence of traditional risk factors such as diabetes, hypertension, and obesity. It may also be secondary in part to increased use of illicit drugs², a more common cause of stroke among the young.

In the Baltimore – Washington Cooperative Young Stroke Study, which included patients 18 to 44 years old, the use of illicit drugs was found to be the fifth most common etiology for ischemic stroke (9% of the study population).³ However, this study included only ischemic strokes, and likely underestimates the contribution of substance abuse as an etiology for all strokes in this population, as hemorrhagic stroke is also associated with substance abuse. Smoking,⁴ cocaine,⁵ and alcohol,⁶ are associated with both ischemic and hemorrhagic strokes, while amphetamines are associated mostly with hemorrhagic events.² The relationship between marijuana and stroke is less clear.⁷

In the present study we sought to identify trends in substance abuse detection among persons aged 18 to 54 who had a diagnosis of stroke during the periods of 7/1/1993 to 6/30/1994, 1999, and 2005, who live in the Greater Cincinnati / Northern Kentucky area.

Methods

The Greater Cincinnati/Northern Kentucky region includes 2 southern Ohio counties and 3 Northern Kentucky counties that abut the Ohio River. This represents a large 1.3 million, biracial, metropolitan population, with similar proportion of African Americans, and similar demographic and socioeconomic status indicators, to the United States population in general, except that the Hispanic population is relatively under represented. Although residents of nearby counties seek care at the 17 to 19 acute care hospitals in the study region, dependent upon study year, only residents of the five study area counties and aged 18 to 54 are included in this analysis (identified as being from the study area based on their zip code). Study periods are 7/1/93 to 6/30/94, 1/1/99 to 12/31/99, and 1/1/2005 to 12/31/2005. The screening of cases has been described elsewhere.⁸

Briefly, study nurses reviewed the medical records of all patients with ICD-9 codes 430–436 as primary or secondary discharge diagnoses from the acute-care hospitals in the study region. Strokes not found by the latter screening were ascertained by monitoring all stroke-related visits to all local public health clinics, hospital-based outpatient clinics and the five county coroners' offices. Further monitoring was performed by examining the records of potential stroke cases in a random sample of primary care physicians' offices and nursing homes in the GCNK region. Sampling was necessary given the large number of physician offices and nursing homes in the region. Non hospital sites were selected randomly, for each study period, by the study statistician from a list generated from a combination of the local yellow pages and the American Medical Association listing of physicians in the region. All events were cross-checked within and between sources to prevent double counting. Institutional Review Board approval was obtained at each participating study site during all study periods. Abstracted information and all available neuroimages were then reviewed by a study physician, who decided whether the patient met the case definition of stroke. The events were classified as ischemic stroke, transient ischemic attack, intracerebral hemorrhage (ICH) or subarachnoid hemorrhage (SAH) according to definitions adapted from the Classification of Cerebrovascular Diseases III.⁹

All ischemic, hemorrhagic and strokes of unknown subtype during the study period were included in this analysis. Events classified as transient ischemic attack were not included. Results of serum or urine drug testing were noted when performed, but as this is based on retrospective medical record review, drug testing was not mandated.

Race was identified by the patient's self assessment in the medical record. Traditional risk factors were assigned based on historical diagnosis. Patient's who did not report prior diagnosis of hypertension but were taking blood pressure medications were considered hypertensive. Alcohol abuse was considered present if subjects consumed more than 3 servings per day (where 1 serving equals 1 oz of hard liquor, 4 oz of wine, or 12 oz of beer), this threshold level has been identified to be the minimum intake required to increase the risk of ischemic stroke.¹⁰ Alcohol abuse was also considered present if documented as heavy in the chart. Smoking was defined as current if present within the last 3 months. Street drugs (marijuana, cocaine /crack, others) were recorded as present if found on routine urine or blood drug tests or if self reported in the medical chart. Information regarding street drug use within 24 hours of stroke onset was also abstracted if noted in the medical chart.

Statistical methods

Data management and descriptive and comparative analyses were performed using SAS® versions 8.02 and 9.3 respectively (SAS Institute, Cary, NC). Proportions are reported using the weighted estimates which were obtained by including the sampling weights in all analyses as dictated by the study design. A more detailed description of the sample weights has been described elsewhere.¹¹ Generalized estimating equations methods (GEE)¹² were used to examine the differences over time. This methodology also accounted for those patients with more than one event within a study time period. The working correlation structure giving the best model fit was obtained. A binary or multinomial distribution was specified for categorical variables, as appropriate.

Results

In 1993/4, 1999, and 2005, there were a total of 2735, 2875, and 2697 patients aged 18 or older suffering a stroke (ischemic or hemorrhagic). The number of patients aged 18 to 54 suffering a stroke increased from 1993/1994 to 2005; there were 297, 376, and 501 strokes for each of the periods (1993/1994, 1999, and 2005, respectively). Baseline patient and stroke characteristics, as well as an overview of the percentage of substance abuse across study periods are given in Table 1. There was a decrease in the percentage of ischemic strokes over time, but this did not achieve statistical significance. Blacks reported substance abuse (current smoking, alcohol or street drug use) more frequently than whites (61% vs. 51%, p<0.02), and males more than females (61% vs. 47%, p<0.002).

Table 1. Patient Characteristics Across Study Periods.

	No. of patients (weighted %)
	1993-1994 N=297	1999 N=376	2005 N=501
Race
White	189 (69%)	239 (66%)	321 (63%)
Black	103 (30%)	134(34%)	176 (37%)
Other	5 (1.4%)	3 (0.7%)	4 (0.7%)
*Substance Abuse1
Present	138 (45%)	206 (52%)	286 (62%)
* Alcohol or drugs within 24 hours of Stroke	5 (1.4%)	25 (6.3%)	72 (12.8%)
Stroke type
Ischemic	219 (78%)	260 (71%)	374 (69%)
Hemorrhagic	78 (22%)	114 (29%)	121 (22%)
Unknown	0 (0%)	2 (0.5%)	6 (9.2%)

¹Substance abuse combined (Current smoking, alcohol, and street drug use).

Data presented as raw n (weighted %)

^*change over time p<0.05

We showed an increase in the overall use of substances (current smoking, alcohol, and street drug use combined) from 1993 (45%) and 1999 (52%) to 2005 (62%) (p=0.003). This paralleled an increase in the documented use of substances within 24 hours of stroke (1.4% in 1993/1994, 6.3% in 1999 and 12.8% in 2005 (P < 0.0001)). Across all study periods, 28 patients had laboratory evidence of abusing street drugs; the percentage was lower in 1993/4 (8%), increasing in 1999 (15%) and 2005 (19%); this was not statistically significant. The absolute number of patients detected only by self report on the medical chart increased across study periods (9 in 1993/1994, to 22 in 1999 and 34 in 2005) but, the percentage of patients detected by the combination of self report and testing, although increasing, showed no statistically significant change over time (23% for 1993, 28% for 1999, and 37% for 2005). When each substance is analyzed separately, the trend is statistically significant for both street drug use and ever smoking (Table 2), with street drug use showing the most marked increase over time (3.8%, 9.8%, and 19.8%, for 1993/1994, 1999, and 2005 respectively, p=0.001). Documented cocaine use increased from 2.0% in 1993, 4.5% in 1999 to 6.6% in 2005, (p<0.01) and documented marijuana increased from 0.6% in 1993, 5.5% in 1999, to 11.4% in 2005, (p<0.01).

Table 2. Trends in Substance Use Over Time Among Stroke Patients.

Strokes only	1993/1994 (n=297	1999 (n=376)	2005 (n=501)
Smoking
Current §	129 (42%)	187 (47%)	245 (52%)
Past	24 (7%)	29 (7%)	61 (14%)
* Never	144 (51%)	160 (46%)	195 (35%)
Heavy Alcohol use
Yes	28 (8%)	48 (12%)	52 (15%)
**Street drug use
Yes	13 (3.8%)	39 (9.8%)	96 (19.8%)

Data presented as raw n (weighted %)

^*change over time p=0.01

^**change over time p<0.001

^§Current if present within the last 3 months

Across study periods, the highest percentage of patients who had documented abuse of street drugs were aged less than 35, 19% compared to 12% for those 35 to 54 years of age (p=0.06). Statistically significant increased documented use over time was only evident for those aged 35 to 54 (p < 0.01). In contrast to street drug use, there was a trend toward smoking being more common in those aged 35 – 54 (49%) versus those aged 20 - 34 (39%) (p=0.06). Current smoking did not significantly change over time (Table 3). Substance abuse was found in both ischemic and hemorrhagic events. Compared to ICH, patients with SAH and ischemic stroke had higher rates of documented current tobacco use (33% vs. 52% and 49% respectively; P<0.01) and SAH patients reported less alcohol abuse than ICH (13% vs. 6% and 10%, for ICH, SAH and ischemic stroke respectively; P=0.02). There was no significant difference between type of event and street drugs use.

Table 3. Substance Use by Age In Stroke Patients Across Study Periods.

	Current smoker			Street drug abuse
Age (years)	1993-94	1999	2005	1993-94	1999	2005
18 - 35	15 (38%)	17 (44%)	17 (36%)	5 (12%)	6 (15%)	13 (28%)
35-44	26 (41%)	45 (48%)	66 (55%)	2 (3%)	15 (16%)	24 (20%)*
45-54	88 (43%)	125 (47%)	162 (53%)	6 (2%)	18 (7%)	59 (19%)**

Data presented as raw n (weighted %)

^*change from 1993/1994 to 2005 p<0.01

^**Change from 1993/1994 and 1999 to 2005 p<0.01

Discussion

Street drug use and ever smoking among young adults with a diagnosis of acute stroke is common and is being more frequently detected within our study population across study periods. It is worrisome that in 2005 half of young adults with stroke were current smokers and 1 out of 5 abused street drugs. Tobacco is a well known risk factor for stroke, the relative risk of stroke associated with cigarette smoking has been calculated at 2.9 for ischemic events (in those less than 55 years old)¹³ and 1.31 for intracerebral hemorrhage;¹⁴ in the period of 2000 to 2004 cigarette smoking resulted in an estimated yearly 443,000 premature deaths.¹⁵ Smoking increases cardiovascular risk by inducing a pro-inflammatory state, arterial endothelial dysfunction, oxidative stress, increased platelet aggregability, blood viscocity, and coagulability, as well as impaired insulin and lipid metabolism.¹⁶ Similarly, many street drugs have been linked to increased stroke risk (cocaine, amphetamines, opiates, phencyclidine, and marijuana). In our study marijuana was the most commonly detected street drug followed by cocaine. Marijuana may increase stroke risk^{7, 17} but the evidence is more robust for cocaine.^18-22 Petitti et al reported an odds ratio of 13.9 for cocaine use associated with stroke in a case control study of women 15 to 44 years old in the Kaiser Permanente Medical Care Program.²³ Postulated stroke mechanisms include cocaine induced vasospasm, cerebral vasculitis, enhanced platelet aggregation, hypertensive crisis associated with altered cerebral autoregulation, and cardio embolism. ²⁴ Amphetamines have also been associated with acute intracranial hypertension and possibly vasculitis.^{2, 23}

Our findings build on those reported by George et al.²⁵ Using data from 1995 to 2008 from the Nationwide Inpatient Sample of the Healthcare Cost and Utilization Project she found increasing use of substances, alcohol, and tobacco, as well as other traditional stroke risk factors, in adolescents and young adults hospitalized with an acute ischemic stroke. We however, found a high prevalence, but no increasing trends, for current tobacco and heavy alcohol use. This likely reflects state specific trends as our findings are similar to what has been reported for the states of Ohio and Kentucky. From 1995 to 2005, the Behavioral Risk Factor Surveillance System (BRFSS) did not report an increase in adult current smokers or heavy drinkers.²⁶

Even though street drug use was more commonly found in those aged 18 – 35, increased documentation (and thus detection) across study periods was evident only in the 35 to 54 year old group. In our study both black race and male gender had increased association with overall use of substances (current smoking, alcohol, and street drug use combined). Alcohol intake at small amounts appears to be protective against cardiovascular events, while consumption at higher amounts is associated with increased risk of atrial fibrillation, ventricular fibrillation, dilated cardiomyopathy, hypertension, dyslipidemia, and a systemic anti-coagulant effect^{27, 28}, all of which increase the risk of stroke. A analysis of a national survey reported an increased use of cocaine among men and blacks²⁹ and the Baltimore – Washington Young Stroke Study found more street drug use among blacks (but no gender differences).³⁰ Our analysis incorporated a retrospective design and cannot correct for testing or diagnosis bias. During all study periods only 28 patients were identified solely on the basis of street drug test results. It is impossible to tell from our data if there were truly higher rates of use, or higher rates of testing, or higher rates of documentation over time. It is possible that black males were singled out in particular for questioning and drug testing, leading to a higher rate of documented use. We are not aware of any guideline or national recommendation that advocated for more aggressive drug questioning / testing in the setting of acute stroke in young adults.

Another finding of our study was the significant increased detection of cocaine and marijuana from 1993 to 2005. Unfortunately, population data regarding trends of recreational substance use from the early 1990's to 2005 is lacking. The National Survey on Drug Use and Health (NSDUH) conducts periodical surveys on the use of substances by those aged 12 years or older but due to changes in methodology starting in 2002 it is not possible to compare reports older than this date with newer ones. Nevertheless, at a national level, no significant increase was reported by the NSDUH in the use of substances (including alcohol, tobacco, and marijuana) from 2002 to 2005, although a slight (not statistically significant) increase in the use of cocaine is mentioned.³¹ National trend analysis of the periods 1992 to 2002 showed a national decrease in cocaine and an increase in marijuana treatment admission rates, but data was incomplete for many states and no trends could be specified for the states of Ohio and Kentucky.³² In a younger population, the National Youth Risk Behavior Survey (YRBS) reported no change in marijuana and an increase of cocaine use between 1993 and 2005 for 9^th through 12^th grade students in public and private schools throughout the United States.³³ Alternatively, an increased intensity of substance use² or adulterated drugs with components that could pose further health hazards³⁴ could also explain our study findings without an overall change in the use of recreational substances at a population level. Ultimately, the design of our study does not allow us to draw conclusions on substance abuse incidence rates given its observational nature and retrospective design. We cannot correct for testing or diagnosis bias, as we did not uniformly record both negative and positive drug tests across study periods, and the testing done was not systematic. It is also possible that those who suffered a stroke and abused substances in 1993-4 were less inclined to seek medical care than their 1999 and 2005 counterparts.

It would have been useful to identify trends for the consumption of amphetamines or amphetamine like substances (mephedrone, pyrovalerone, methylenedioxypyrovalerone), since consumption of these substances is presumed to be on the rise, and in the case of amphetamines, are associated with hemorrhagic stroke.² Another limitation of our study is that we were unable to further classify smoking status by the number of cigarettes consumed, but it is known that the risk of stroke and cardiovascular disease increases even with small exposures.³⁵ Strengths of our investigation include its population based design that is comparable to the United States in general in its proportion of African Americans, demographics and socioeconomic status, the large number of subjects under study, and the availability of data from 3 study periods with consistent methodology allowing for testing trends over time.

Substance abuse is one of many risk factors for stroke. It is likely that no single factor, but rather a mixture of them, is responsible for the increased incidence of stroke observed among young adults in our population. As an example, un-published analysis of our data shows an increasing prevalence of dyslipidemia across study periods (13.1% in 1993-4, 18.8% in 1999, and 24.2% in 2005; P=0.047), but no trend is evident for hypertension, diabetes, or atrial fibrillation.

In summary, substance abuse is a common and important risk factor for stroke in young adults. In our population detection of both street drug use and ever smoking among young adults with a diagnosis of stroke increased from 1993 to 2005, preferentially in those aged 35 to 54. The observed increase in substance abuse is contributing to the increased incidence of stroke in young adults. All patients younger than 55 who suffer a stroke should be routinely screened and counseled regarding substance abuse.