Impact of Cocaine Use on Acute Ischemic Stroke Patients: Insights from Nationwide Inpatient Sample in the United States

Cocaine is the third most common substance of abuse after cannabis and alcohol. The use of cocaine as an illicit substance is implicated as a causative factor for multisystem derangements ranging from an acute crisis to chronic complications. Vasospasm is the proposed mechanism behind adverse events resulting from cocaine abuse, acute ischemic strokes (AIS) being one of the few. Our study looked into in-hospital outcomes owing to cocaine use in the large population based study of AIS patients. Using the national inpatient sample (NIS) database from 2014 of United States of America, we identified AIS patients with cocaine use using International Classification of Disease, Ninth Revision (ICD-9) codes. We compared demographics, mortality, in-hospital outcomes and comorbidities between AIS with cocaine use cohort versus AIS without cocaine use cohort. Acute ischemic strokes (AIS) with cocaine group consisted of higher number of older patients (> 85 years) (25.6% versus 18.7%, p <0.001) and females (52.4% versus 51.0%, p <0.001). Cocaine cohort had higher incidence of valvular disorders (13.2% versus 9.7%, p <0.001), venous thromboembolism (3.5% versus 2.6%, p<0.03), vasculitis (0.9% versus 0.4%, p <0.003), sudden cardiac death (0.4% versus 0.2%, p<0.02), epilepsy (10.1% versus 7.4%, p <0.001) and major depression (13.2% versus 10.7%, p<0.007). The multivariate logistic regression analysis found cocaine use to be the major risk factor for hospitalization in AIS cohort. In-hospital mortality (odds ratio (OR)= 1.4, 95% confidence interval= 1.1-1.9, p <0.003) and the disposition to short-term hospitals (odds ratio (OR)= 2.6, 95% confidence interval = 2.1-3.3, p <0.001) were also higher in cocaine cohort. Venous thromboembolism was observed to be linked with cocaine use (OR= 1.5, 95% confidence interval= 1.0-2.1, p < 0.01) but less severely than vasculitis (OR= 3.0, 95% confidence interval= 1.6-5.8, p <0.001). Further prospective research is warranted in this direction to improve the outcomes for AIS and lessen the financial burden on the healthcare system of the United States.


Introduction
The use of cocaine as an illicit drug surged in the United States of America between 2002 and 1 2 3 4 5 6 7 8 2007 and currently, it is the second most abused drug after cannabis and alcohol [1][2][3]. Among the adults, the annual prevalence of cocaine use was 1.5%, while some states reported the prevalence of 5.5% to 5.8% in the age group of 18 to 25 years. The United Nations Office on Drugs and Crime and the European Monitoring Centre for Drugs and Drug Addiction has reported an increase of cocaine use in particular parts of the world [4]. The burden of health care due to cocaine dependence was reported high in the recent studies [4]. Around 23.9 million people aged 12 years and above were reported using illicit drugs in 2012 according to data from the National Survey on Drug Use and Health [5]. A recent study from a community hospital found 2.3% people being cocaine positive during drug screening in the age group of 65 years and older. In the USA, the areas of primary concern are the one rampaged with poverty and poor education [6] One report from the emergency department notes in the Detroit area in 2002 showed cocaine use of around 182/100,000 of the population [7]. It is not merely a problem of one country; rather it has turned into a global issue [1]. The cocaine use leads to the spectrum of multisystem derangements ranging from mild intoxication to severe complications like acute myocardial infarction, seizures and acute ischemic stroke [8]. Compared to the corresponding peer groups in the general population, cocaine users tend to have four to eight times higher mortality [9][10]. Use of cocaine is presumed to be one of the major risk factors for cerebrovascular disease, including stroke. Acute ischemic stroke (AIS) is labeled as the third leading cause of disability-adjusted life years [10].
Previous studies have reported a 19% increase in the incidence of strokes due to cocaine use in the last two decades [11]. The rise in cocaine-associated morbidity and mortality posed it as a major public health concern [11][12]. Impacts on health care economies due to stroke-related disability is devastating owing to medical cost, rehabilitation cost and cost due to loss of workforce. A direct or indirect burden of around $68.9 billion was imposed on the US healthcare owing to strokes in 2009, a major part of which was comprised of strokes as a result of illicit drugs use [2]. We aim to evaluate various factors associated with acute ischemic strokes (AIS) risk and to develop the management strategies to mitigate mortality rates within cocaineinduced stroke population.

Data source
We utilized the discharge data from the National Inpatient Sample (NIS) of the Healthcare Cost and Utilization Project (HCUP) as a source. The NIS is an all-payer dataset that includes around eight million (around 20% of the stratified sample) inpatient admissions and discharges from almost 1050 USA hospitals, excluding long term care and rehabilitation facilities. The NIS data set is unweighted and it results in the weighted estimate of the total discharge number of the US population when we apply the discharge weight to the unweighted data. We excluded the data of missing information such as age, gender, discharge condition or primary diagnosis.

Baseline characteristics of acute ischemic strokes cohort
We identified the total of 584,115 patients with AIS by using the discharge data from NIS of the year 2014. We further stratified cohorts into a cocaine use (N=1135) and non-cocaine use (N=582980   Table 2 shows different variables that were used in the multivariate logistic regression model to identify AIS risk factors requiring hospitalization. We found cocaine use to be the major risk factor for hospitalization. In-hospital mortality was also observed to be higher in cocaine cohort with the 95% confidence interval (CI) 1.1-1.9 (OR= 1.4, 95% CI= 1.1-1.9, p <0.003). A disposition to short-term hospitals (OR= 2.6, 95% CI= 2.1-3.3, p <0.001) and home healthcare (OR= 1.5, 95% CI= 1.2-1.9, p <0.001) was also significantly higher after adjusting for confounders. Personal history of sudden cardiac arrests (OR= 7.9, 95% CI= 3.1-20.1, p <0.001) were significantly associated with cocaine use which could be another manifestation due to potential vasospasm [8]. Venous thromboembolism was observed to be linked with cocaine use (OR= 1.5, 95% CI= 1.0-2.1, p < 0.01), but less severely than vasculitis (OR= 3.0, 95% CI= 1.6-5.8, p<0.001) (   Table 3 shows the gender comparison in co-morbidities associated mortality in the cocaine cohort. Higher overall mortality due to cardiac (except arrhythmia and supraventricular premature beats) causes and acute renal failure was observed in males, whereas females had increased overall mortality owing to elevated cholesterol, heart valve disorders, vasculitis, epilepsy, arrhythmia, supraventricular premature beats (SVPB), peripheral arterial thromboembolism and heart valve disorders. A similar rate of mortalities between males and females was found due to events of elevated blood pressure without hypertension (which could be owing to incidental cocaine intake) and acute cerebrovascular disease.   Table 4 shows the comparison of various comorbidity related mortality odds between cocaine and non-cocaine cohorts. Mortality odds owing to liver disease, metastatic cancer, cardiomyopathy, acute myocardial infarction, and epilepsy were increased in both non-cocaine and cocaine cohort. Whereas, increased odds of mortality in the non-cocaine cohort were observed due to coagulopathy, disorders of fluid and electrolyte, obesity, weight loss, solid tumor without metastasis, elevated cholesterol, pneumothorax and pleurisy and congestive heart failure. Effect on mortality due to other variables is shown in Table 4.

Discussion
The current study found 96.5% of the AIS cocaine cohort was of the age group above 45 years with age ranging from 18 years to 85 years and above. In the age group of 85 years and above, the prevalence of AIS within the cocaine group surpassed the non-cocaine users. A plausible explanation could be that the cumulative effect of traditional risk factors, along with the longterm accumulation of chronic cocaine effect makes such population more vulnerable towards the risk of stroke [2]. The frequency of hospitalization was high among the urban hospitals set up.
A majority of the AIS patients visited the private, nonprofit hospitals. Among these, the odds of ones with cocaine use visiting the government, non-federal hospital were significantly low. The nature of the admission was nonelective understandable for most of the AIS cohort and within this cohort; it was significantly higher among the cocaine users. It could be because most of the patients are chronic cocaine abusers rather than acute. This finding is also justifiable from the older age group pattern of the study subject which is prone to the cumulative effect of the cocaine rather than acute features [2]. Odds of hospitalization among the whites were higher compared to the blacks and Hispanics in AIS cocaine cohort versus AIS non-cocaine cohort.
The mortality was found higher in blacks as compared to whites and Hispanic (21.4% versus 8.6% versus 11.1%, P=0.004 respectively). Analyzing the disposition of the patients, the shortterm hospitals stay and death was significantly higher among the cocaine users. The associated higher comorbidities could be a possible explanation for such disposition in cocaine users as compared to the non-cocaine group. This finding is a serious concern because such patients could lead to a significant burden on the healthcare infrastructure. With increasing median household income, the frequency of hospitalization significantly increased among the cocaine users suggesting that the ones with low income and living in a poverty have a lower risk of using cocaine. Another reason could be that socioeconomic status is the poor predictor of the stroke among the cocaine users. Despite having high median household income among the cocaine users, their hospitalization was significantly elevated in the small and medium-sized hospitals. This could be due to the acute nature of the condition among these subjects, requiring urgent admission to any of the nearby facility.
The family history of the stroke was higher among the non-cocaine users as compared to the cocaine users suggesting that the usual mechanism of stroke development is not applicable to the cocaine user. Cocaine users quite commonly bear the traditional cardiovascular risk factors [2]. Sudden cardiac death, paroxysmal supraventricular tachycardia (PSVT), vasculitis and venous thromboembolism take higher odds of hospitalization.
Odds of hospitalization due to the paralysis were significantly higher among the non-cocaine users while seizures were high among the cocaine users. The frequency of depression was significantly higher among the cocaine abusers signifying the high morbidity among such populations. The incidence of diabetes and congestive heart failure (CHF) was severely high among the cocaine users, while the frequency of hypertension was quite similar to the other studies [2].
There was a significantly higher rate of valvular heart disease and venous thromboembolism among the cocaine users, suggesting of emboli as the major risk for stroke among these populations as compared to the non-cocaine users [15][16][17]. We reported the higher mortality among the cocaine users as compared to the non-cocaine users [18].
The older age of our study population could be a plausible reason, as studies with the young demographic and mild strokes have reported overall low mortality [19]. When we looked for the multivariate predictors of death in the AIS patients without cocaine use and with cocaine use, we found that epilepsy, peripheral vascular disorders, acute myocardial infarction, cardiomyopathy, tachycardia, metastatic cancer and liver diseases were associated with higher odds of mortality among cocaine users as compared to the non-cocaine users.
Hypertension and diabetes were not found to be associated with the excess mortality in an AIS cocaine cohort compared to the AIS non-cocaine cohort. Several postulated mechanisms for cocaine-induced ischemic stroke has been suggested [20][21][22][23]. Among these, the cardioembolic ischemic strokes and cardiac deaths due to chronic cocaine use have been proposed to be prominent [17,24].

Study limitations
This study has undeniable limitations because of the NIS database which might have coding errors in terms of determining the diagnosis, comorbidities, and complications. Due to the inherent nature of large hospital's database, it may over or underestimate AIS, cocaine use, comorbidities and other clinically relevant variables based on ICD-9 CM codes. This study also lacked variables such as medications and other treatment options related to AIS.
This database does not mention about the cause of death, so we cannot differentiate between in-hospital events and cause of death. It might be possible to have a selection bias in this study because of a retrospective population study. Due to large data size and getting national estimates using discharge weight as provided by NIS database, we could overcome these limitations.

Conclusions
To our knowledge, this is one of the very few studies demonstrating the effects of cocaine use on stroke using the nationally representative data source. Our results displaying the amplitude of the mortality in an AIS-cocaine cohort raised the question whether to consider cocaine as a risk factor in all AIS patients or not. Further research is warranted to evaluate the pathogenesis and health care burden due to cocaine-induced stroke. Table 5 has the list of ICD-9-CM codes for suspected risk factors of AIS, comorbidities, inhospital procedures and outcomes.