Introduction
Stroke is a leading cause of death and long-term disability in the USA [Reference Virani, Alonso and Benjamin1]. Intracerebral hemorrhage (ICH), in particular, accounts for 10% of strokes but contributes to 50% of stroke mortality [Reference Woo, Rosand and Kidwell2]. ICH risk factors in the US Stroke Belt population are not well characterized [Reference Howard, Cushman and Howard3, Reference Howard and Howard4]. This region consists of a group of Southeastern states including Alabama, Arkansas, Georgia, Indiana, Kentucky, Louisiana, Mississippi, North Carolina, South Carolina, Tennessee, and Virginia demonstrated in 1980 to have 10% higher stroke death rates than the US average [Reference Howard and Howard4]. Those living in the Stroke Belt have been noted to have more hypertension and diabetes mellitus [Reference Howard and Howard4, Reference Liao, Greenlund, Croft, Keenan and Giles5] along with lower socioeconomic status and less access to quality care [Reference Howard and Howard4]. The Stroke Belt population includes 27% non-Hispanic Blacks as compared to the national average of 10% [Reference Howard and Howard4]. As a result, this region is primarily biracial. Many prior studies focusing on racial disparities in this region have demonstrated disparities between those of Black and White race [Reference Howard and Howard4]. Improving ICH outcomes and reducing disparities in this region requires a clear understanding of the underlying associations with risk factors and outcomes.
In addition to geographic disparities, racial disparities in risk factors [Reference Kissela, Schneider and Kleindorfer6, Reference Walsh, Woo and Sekar7] and outcomes [Reference Kissela, Schneider and Kleindorfer6] have been previously demonstrated in stroke patients [Reference Cruz-Flores, Rabinstein and Biller8]. A limitation of some previous studies examining disparities is that they often focus on an all-stroke prospective without specifically examining ICH [Reference Howard and Howard4]. Other studies have shown racial disparities in comorbidities [Reference Howard and Howard4], access to health care [Reference Cohen and Villarroel9], and socioeconomic status [Reference Howard and Howard4] among ICH patients with limited emphasis on outcomes. Finally, while some studies have shown racial differences in both risk factors [Reference Howard, Cushman and Howard3] and outcomes [Reference Viswanathan and Greenberg10, Reference Miyares, Falcone and Leasure11] for ICH, these prior studies may be difficult to interpret due to variability in measurements of functional outcome. A study within the Stroke Belt is necessary to address the breadth and complexity of ICH and how it affects various outcome measurements.
Lastly, the influence of age, which carries a significant association with ICH etiology [Reference Forti, Maioli and Domenico Spampinato12, Reference Iwuchukwu, Mahale and Ryder13], further compounds these considerations regarding ICH risk factors and outcomes. For example, while recent studies indicated poorer outcomes for White patients after ICH, these disparities were specific for younger adults and from a national perspective [Reference Boan, Feng and Ovbiagele14]. Within the Stroke Belt, greater incidence for ICH has been observed in Black patients for ages below 65, and White patients for ages above 65 [Reference Howard, Cushman and Howard3]. As such, also examining ICH racial disparities within the context of age is necessary.
As age, geographic, and racial disparities have all been shown to affect risk factors, etiology, and outcomes, examining the influence of race on the complex nature of ICH requires a nuanced approach. The purpose of this biracial study is to examine potential race and age disparities for spontaneous ICH, as well as potential differences in the predictors for poor in-hospital outcomes based on race, within a Stroke Belt population.
Methods
Study Population
This is a secondary analysis of prospectively screened patients from a case–control study based on consecutive hospital admissions for bleeding events at a tertiary stroke center, the University of Alabama at Birmingham (UAB), between January 2017 and December 2019 (excluding a 5-month pause in enrollment from January to May 2018). Potential study participants included all bleeding-related admissions noted on the prospectively screened hospital admission list daily. Patients (≥18 years of age) with a major hemorrhage as defined using the International Society on Thrombosis and Hemostasis (ISTH) criteria were included [Reference Schulman and Kearon15]. From these screened major hemorrhages, intracranial hemorrhage events were selected in February 2020 for investigation.
A board-certified vascular neurologist (MJL) reviewed all patients admitted with intracranial hemorrhages in April and May 2020. We excluded patients with subarachnoid, subdural, and epidural hemorrhages, those with secondary intracranial hemorrhage (trauma, tumor, vascular lesion, secondary conversion of infarct, venous thrombosis, vasculopathy/vasculitis), or with major trauma. We included hemorrhages secondary to anticoagulant use unless a secondary etiology was identified. All cases not involving patients of the Black or White race were excluded to allow for direct comparisons between these two groups.
Clinical Characteristics and Data Collection
Demographic information, medical comorbidities, social variables, admission and discharge medications, admission laboratory values, hospital management details, radiographic features, and in-hospital outcomes were gathered in February 2020 through retrospective chart review using an electronic health record system. Transfer status was defined as a transfer from other hospitals or freestanding emergency rooms. A measure of rurality was obtained by converting documented home zip code to Rural–Urban Commuting Area-2 (RUCA-2) codes [Reference Hart, Larson and Lishner16]. RUCA-2 scores were dichotomized. A score of 1–3 was designated as urban, and a score of 4–10 was designated as rural, as micropolitan areas could be alternatively defined as large rural areas according to definition criteria [Reference Larson17]. The stroke etiology/mechanism was determined using neurology/neurosurgery attending documentation and independently reviewed by the board-certified vascular neurologist (MJL).
Clinical severity was defined using the National Institutes of Health Stroke Scale (NIHSS; scores of 0–42 with higher scores indicating higher degrees of stroke severity [18]) and ICH score (a clinical grading scale with scores of 0–6; higher scores reflecting greater stroke severity [Reference Hemphill, Bonovich, Besmertis, Manley and Johnston19]) on admission were obtained from clinical documentation by trained healthcare providers. For clinical documentation of past medical history, missing data were assessed as indicating that the patient did not have the condition. For example, if no admission documentation mentioned a history of hypertension, then the patient was recorded as not having hypertension. In this way, no admission medical history data were defined as missing. Admission laboratory values were defined as the first value, if present, only within 24 h of admission. Information regarding mortality, discharge modified Rankin score (mRS), and discharge disposition were gathered from the inpatient hospital discharge note. The mRS is a 6-point disability scale with higher numbers indicating more severe disability and a score of 6 indicating death [Reference Bonita and Beaglehole20]. If the discharge mRS was not given in the provider discharge note (about 25% of the time), then it was determined based on the discharge physical exam, discharge inpatient physical therapy documentation, discharge inpatient occupational therapy documentation, and discharge nursing note.
For this analysis, we compared data based on the Black and White races. Our primary outcomes of interest were in-hospital mortality (defined as death from any cause during the singular inpatient hospital stay associated with the ICH event) and favorable discharge functional status (defined as an mRS score of 0–2). We also examined favorable discharge disposition which was defined as discharge home or to an inpatient rehabilitation facility, while poor discharge disposition was defined as the discharge to a skilled nursing facility, long-term acute care hospital, or death.
Statistical Analysis
Statistical Analysis Software (SAS) version 9.4 was used. Mean and median values were created from quantitative variables for direct statistical comparison by race. Univariate and bivariate statistical analyses were performed to determine significant differences in the prevalence of risk factors between Black and White race subgroups, as well as rates of poor outcomes. These analyses included two-sided Student’s t-tests for continuous variables and chi-square analyses for categorical variables. To determine risk factors for in-hospital mortality, poor discharge mRS, or poor discharge disposition, logistic regression was performed. To assess confounding, variables significant on univariate analysis were then modeled in a stepwise fashion with other covariates plausibly associated with in-hospital mortality. Statistical significance was defined as a bidirectional alpha level of 0.05.
As the ICH score is frequently used to guide treatment and prognostication decisions, we sought to determine which items were most important in predicting in-hospital mortality from admission data. A multivariable model of admission data for the prediction of in-hospital mortality was created separately for each race. Variables previously shown to be associated with in-hospital mortality based on the ICH score, including Glasgow Coma Scale (GCS), infratentorial location, intraventricular hemorrhage, age greater than 80, and hematoma volume, were included in the model, in addition to the statistically significant variables for in-hospital mortality determined on our univariable analysis. Receiver operating characteristic (ROC) curves for these multivariable models were then created to examine the performance of these models.
Results
A total of 148,741 admissions occurred during the study period, including 3647 bleeding events screened by hospital admission diagnosis, 1234 cases of intracranial bleeding, and 310 unique cases of spontaneous ICH. Cases of spontaneous ICH not involving patients of the Black or White race were excluded (Asian race [n = 6]; missing or unknown race [n = 29]). This resulted in 275 cases suitable for inclusion in the study (see CONSORT diagram, Fig. 1).
Fig. 1. CONSORT diagram. ICH, intracerebral hemorrhage; UAB, University of Alabama at Birmingham.
Patient characteristics are presented in Table 1. For the overall cohort, the median age was 66.5 years, 51% female, 41% Black, and 71% originating from an urban residence. Sixty-one percent of patients were transferred from other hospitals/freestanding emergency rooms. The majority of patients (52%) were diagnosed with a hypertensive hemorrhage, 10% were felt to be related to cerebral amyloid angiopathy (CAA), and 14% were related to anticoagulants. We observed a 31% in-hospital mortality rate, overall. Median discharge mRS was 4 with only 27% having a favorable discharge mRS of 0–2. Nearly half of the patients had a favorable discharge disposition to home or an inpatient rehabilitation facility, however. Our study also did not find differences in markers of ICH severity including initial hematoma volumes (P = 0.97) and ICH scores (P = 0.92) based on sex.
Table 1. Patient characteristics
CT, computed tomography; DVT, deep venous thrombosis; EVD, external ventricular drain; ICH, intracerebral hemorrhage; INR, international normalized ratio; mRS, modified Rankin scale; NIHSS, National Institutes of Health Stroke Scale; PE, pulmonary embolism; RUCA, rural urban commuting area.
Differences in ICH Characteristics and Risk Factors by Race
Racial differences in demographics, risk factor prevalence, admission lab values, radiologic data, and discharge disposition were observed among ICH patients (Table 1). Black patients were younger (60.6 vs. 71.4 years, P < 0.0001), more likely to reside in urban areas (81% vs. 64%, P < 0.01), and were less likely to have an interhospital transfer (44% vs. 74%, P < 0.01). Black patients more often had a previous diagnosis of hypertension. They also were admitted with lower median leukocyte count, as well as smaller median initial hematoma volume (9.1 vs. 12.6 mL, P = 0.041). No significant differences in admission NIHSS score or ICH score were observed based on race. Furthermore, no significant differences were observed with in-hospital mortality or poor discharge mRS, although there were significantly greater rates of favorable discharge disposition among Black patients (57% vs. 43%, P = 0.002). However, adjustment for initial hematoma volume attenuated this association.
Patients were further stratified to those below 65 (“young”) and those 65 and older (“old”) to describe the effect of age (Table 2). Across age groups, Black patients had a lower likelihood of interhospital transfer compared to White patients. Among the older population, Black patients more often originated from an urban residence and had higher rates of previous hypertension (98% vs. 74%, P < 0.01). We also observed that young Black patients had smaller hematoma volumes (8.5 vs. 11.1 mL, P = 0.03), but volumes did not significantly differ among older age groups. A lower rate of code status change to Do-Not-Resuscitate (DNR) for Black patients was also observed among the young cohort (7% vs. 23%, P = 0.01), while there was no significant difference in the older cohort.
Table 2. Patient characteristics stratified by race and age
EVD, external ventricular drain; ICH, intracerebral hemorrhage; mRS, modified Rankin scale; NIHSS, National Institutes of Health Stroke Score; RUCA, rural urban commuting area.
When compared solely on the basis of age (not race), young patients had higher rates of deep bleeds involving the basal ganglia, internal capsule, thalamus (P = 0.01), and hypertensive or cryptogenic etiologies (P < 0.01). When comparing young Black patients to old Black patients or young White patients to old White patients, both comparisons also showed higher rates of deep bleeds (P < 0.01 and P = 0.02, respectively) and hypertensive etiology (P = 0.02 and P < 0.01, respectively) among the young patients. Yet, further stratification by both race and age did not reveal significant differences in ICH location or etiology across races and ages. Finally, in-hospital mortality and functional outcome did not differ across races and ages.
Predictors of In-Hospital Mortality
The results for univariate risk for in-hospital mortality based on logistic regression are shown in Table 3. For both Black and White patients, in-hospital mortality was predicted by admission leukocyte count, admission glucose level, admission NIHSS, admission ICH score, need for intubation, initial hematoma volume, intraventricular hemorrhage, spot sign on CT angiogram, and hematoma expansion on follow-up imaging. Black patients requiring intubation were at a 22-fold increased risk of in-hospital mortality, whereas White patients were at a 20-fold increased risk. When adjusted for initial hematoma volume, intraventricular extension, and ICH score, intubation remained a highly significant predictor of mortality. For only Black patients, in-hospital mortality was also predicted by the previous history of hyperlipidemia (admission total cholesterol), documented antiplatelet use, and neurosurgical intervention (including placement of external ventricular drain [EVD] for intracranial pressure monitoring). There were no variables that predicted in-hospital mortality for only White patients.
Table 3. Predictors of in-hospital mortality among patients with intracerebral hemorrhage (ICH)
CT, computed tomography; EVD, external ventricular drain; INR, international normalized ratio; NIHSS, National Institutes of Health Stroke Scale.
The unique risk factors for in-hospital mortality for Black patients were then examined further by multivariable analysis. After adjustment for initial hematoma volume, antiplatelet use remained significantly associated with mortality (OR 3.40, 95% CI 1.19–9.67, P = 0.02). When further adjusted for history of atrial fibrillation, congestive heart failure, hypertension, age greater than 65, diabetes mellitus, history of stroke or TIA, and peripheral vascular disease, the risk still remained significant (OR 3.60, 95% CI 1.03–12.6, P = 0.047). The influence of neurosurgical intervention was attenuated on adjustment for markers of severity including ICH score and initial hematoma volume.
Multivariable models for in-hospital mortality were created based on ICH score components and variables significant for mortality for each race from univariable analysis (antiplatelet use, hyperlipidemia, leukocyte count, and glucose). For all patients, incremental increases in initial ICH volumes were associated with in-hospital mortality (Table 4). The admission GCS was a significant predictor among White patients (OR 11.8 for each category jump within the ICH score, 98% CI 3.95–35.13, P < 0.0001), but not among Black patients. Presence of intraventricular hemorrhage, age greater than or equal to 80, or infratentorial location did not influence in-hospital mortality for either race based on this multivariable model. Notably, the influence of age on mortality among Black patients was attenuated in the multivariable model after adjustment for hyperlipidemia and antiplatelet use. The ROC curves demonstrated an area under the curve (AUC) of about 0.9, indicating that these models performed well overall. Comparing a model of the original ICH score components (Fig. 2) to our new model including the additional components (Fig. 3), the AUC increased by 0.04 for Black ICH patients and 0.011 for White ICH patients.
Table 4. Multivariable analysis of in-hospital mortality by race
GCS, Glasgow Coma Scale.
Fig. 2. Receiver operating characteristic for in-hospital mortality based on intracerebral hemorrhage (ICH) score. AUC, area under the curve.
Fig. 3. Receiver operating characteristic for in-hospital mortality based on multivariable model. AUC, area under the curve.
Predictors of Poor Functional Outcome
For both Black and White patients, poor functional outcome was predicted by admission NIHSS, ICH score, and initial hematoma volume (Supplemental Table 1). For Black patients only, history of hypertension, antiplatelet use, and intubation status was significantly associated with functional outcomes. The association of hypertension with functional outcomes remained significant after controlling for initial hematoma volume. Also, for Black patients only, interhospital transfer was predictive of good functional outcomes. For White patients only, prior stroke or TIA, admission leukocyte count, admission glucose, neurosurgical intervention (EVD placement), intraventricular extension, and presence of microhemorrhages were each independently significantly associated with poor functional outcome.
Discussion
The purpose of this study was to determine risk factors, ICH characteristics, and outcomes by race for ICH patients within the US Stroke Belt. The overall mortality was determined to be 30% with no differences in ICH severity, mortality, or discharge mRS between Black and White patients. This was true despite smaller initial hematoma volumes among Black patients. Age was associated with ICH etiology and location, and this association was present for both Black and White ICH patients. The ICH score, initial hematoma volume, intraventricular extension, intubation, and higher initial glucose or leukocyte count were consistently predictive of poor outcomes for both races. Intubation was a particularly strong predictor of in-hospital mortality. We noted that clinical documentation of hyperlipidemia (admission total cholesterol) and admission antiplatelet use were predictors of in-hospital mortality for only Black ICH patients. We also observed an increase in code status changes to DNR among young Black patients compared to young White patients. Finally, in examining a race-specific ICH score, we found hyperlipidemia to be a significant predictor for Black patients, but not White patients.
In our sample, Black ICH patients were on average more than 10 years younger than White ICH patients. This finding is consistent with prior literature [Reference Howard, Cushman and Howard3, Reference Xian, Holloway and Smith21]. The generally younger age of Black patients with ICH may be related to differences in hemorrhage etiology suggested by the more frequent history of hypertension among our Black patients. Previously, the REGARDS study demonstrated that Black participants were more aware of their hypertension than White participants but less likely to have their blood pressure controlled [Reference Howard, Prineas and Moy22]. Prior studies have also shown that hypertension is a powerful risk factor for spontaneous ICH for Black and White patients [Reference Walsh, Woo and Sekar7], with a higher risk for Black patients. Taken together, these two studies emphasize the importance of addressing this disparity. Furthermore, although the history of hypertension was not associated with in-hospital mortality, it was independently associated with poor functional outcomes at discharge for Black patients.
Hypertension seemed to play a similar role for both Black and White ICH patients, consistent with the known associations of hypertensive etiology and deep location with young age [Reference Poon, Fonville and Al-Shahi Salman23]. When first compared solely on the basis of age (not race), young patients had higher rates of deep bleeds and etiologies determined to be hypertensive or cryptogenic. Yet, further stratification by both race and age did not reveal significant differences in ICH location, etiology, outcome, or rates of intraventricular hemorrhage across races and ages. Nevertheless, smaller initial hematoma volumes were seen among young Black ICH patients compared to young White ICH patients. One potential confounder is the increased interhospital transfer rate among young White patients compared to young Black patients. It is possible that interhospital transfer may be associated with early hematoma growth due to delays in the provision of acute hemorrhage care including blood pressure management, neurosurgical intervention, and neurocritical care.
Within this sample, we failed to show that age differences significantly affected outcomes by race. Older age has been associated with ICH etiology [Reference Forti, Maioli and Domenico Spampinato12] and lobar location [Reference Forti, Maioli and Domenico Spampinato12, Reference Iwuchukwu, Mahale and Ryder13], increased hematoma volume and intraventricular hemorrhage [Reference Iwuchukwu, Mahale and Ryder13], as well as poor outcome [Reference Venema, Marini and Brouwers24]. However, it has also been shown that advanced age and cerebral atrophy may offer some protection against in-hospital mortality after ICH [Reference Flaherty, Woo and Haverbusch25], perhaps due to a greater room for hematoma expansion. In our study, as only some of these associations were seen by age stratification alone, the further stratification by both race and age may have limited our statistical power in examining this topic.
Our study did not identify differences in poor outcomes based on sex, which contrasts with the Ethnic and Racial Variations in Intracerebral Hemorrhage (ERICH) study’s report that women had a poorer functional outcome based on discharge mRS [Reference James, Langefeld and Sekar26]. Our study also did not find differences in markers of ICH severity (initial hematoma volumes, ICH scores) based on sex. This further supports the notion that our ICH cases were similar among both sexes. Nevertheless, as clear differences based on sex in both mortality and poor functional outcome were noted at 90 days in ERICH, our study’s inability to support this finding may be due to the limited follow-up in our study.
We also noted an increase in the rates of code status change to DNR for young Black patients compared to young White patients, though no difference was noted in the older age groups. This finding supports results from ERICH suggesting that differences in goals of care decision-making may exist among ICH patients by race [Reference Ormseth, Falcone and Jasak27], as well as other studies showing Black patients were less likely to have DNR changes after subarachnoid hemorrhage [Reference Choi, Fernandez and Jeon28] and younger patients were less likely to have DNR changes [Reference Wenger, Pearson and Desmond29]. As no differences in in-hospital mortality or poor discharge functional outcome were seen here based on race (similar to ERICH), the difference in changes to DNR status between races warrants further investigation. While our finding potentially indicates differences in the quality of end-of-life care, the specific use of certain comfort measures was not examined here. Future studies should examine this topic in greater detail among ICH patients by race.
The effect of antiplatelet therapy on ICH outcomes continues to be a contested topic. In our study, documented antiplatelet use on admission was an independent predictor of in-hospital mortality for Black ICH patients, but not White ICH patients. Some prior studies have clearly shown an association of documented antithrombotic use with increased hematoma growth [Reference Roquer, Vivanco-Hidalgo and Capellades30] and ICH mortality [Reference Toyoda, Yasaka and Nagata31], and others have only shown this association with pre-ICH use of combination antiplatelet therapy [Reference Khan, Siddiqui and Goldstein32]. Still, others have continued to demonstrate that the use of antiplatelets after ICH did not increase the risk of recurrent ICH [Reference Murthy, Biffi and Falcone33], higher hematoma volumes, subsequent expansion, or 90-day clinical outcomes [Reference Sansing, Messe and Cucchiara34].
The ICH score has been tested in specific subpopulations, such as Argentinian or Asian populations [Reference Godoy and Boccio35, Reference Jamora, Kishi-Generao, Bitanga, Gan, Apaga and San Jose36], but further validation specifically for Black patients is still needed. To address this, we analyzed a race-specific ICH score for predicting in-hospital mortality, as has been done in a prior study by Faigle et al. [Reference Faigle, Marsh, Llinas, Urrutia and Gottesman37]. ICH characteristics were similar by race to the prior study, with lower initial hematoma volumes and lower median ICH scores among Black patients. GCS was predictive for only White patients in our study, as opposed to being predictive for either race in the prior study. In our sample, accounting for hyperlipidemia increased the association of this model with eventual in-hospital mortality for Black patients from ICH. Overall, both the study by Faigle et al. and our study demonstrated that the ICH score’s effectiveness may differ by race. As both age and etiology may play a significant role in ICH outcomes, the increased proportion of ICH among young Black ICH patients continues to indicate that race-specific validation of this tool on a larger scale may be needed.
Prior literature regarding hyperlipidemia as a risk factor for an ICH event has been mixed [Reference Woo, Kissela and Khoury38]. Few studies have considered whether hyperlipidemia increases the risk for in-hospital mortality or poor functional outcome after ICH. While it is true that hyperlipidemia may serve here as a surrogate marker for other unknown factors, its continued significance in the multivariable model after controlling for other variables makes this less likely to be the case. Hyperlipidemia has particularly been associated with nonlobar ICH, and Black patients suffer from nonlobar ICH at increased rates [Reference Martini, Flaherty and Brown39]. The significance of hyperlipidemia to ICH outcomes among Black patients needs further examination.
Our study has several strengths including providing details on ICH patients from a representative sample in the US Stroke Belt with a high proportion of Black patients, which allowed for stronger statistical comparisons with White ICH patients. We collected variables to define predictors of in-hospital mortality and discharge functional status with considerations to age and race among a high-risk patient population. Besides these strengths, there are several limitations that must be considered. First, our sample is based on a single-site location at a tertiary stroke center, which may limit generalizability, particularly outside the Stroke Belt and at nonacademic centers. Although the sample was ascertained in a prospective manner to allow for complete capture of all ICH patients presenting to our center, the data itself was ultimately abstracted in a retrospective fashion. Also, the method of defining past medical history as a dichotomous variable may have limited its accuracy. A potential alternative for improving this limitation is using multiple pilot tests to test the study coding manual [Reference Gregory and Radovinsky40]. This may lead to increased internal validity and reproducibility. Yet, even with the use of a coding manual, inaccuracies for this data may be due to variations in clinical documentation methods among healthcare providers, not necessarily as a result of data abstraction methodology. Furthermore, a high proportion of our sample was received in interhospital transfer, and we did not have access to medical records from external facilities. Finally, the primary endpoints of this study involved short-term outcomes precluding differences in outcomes that may become more apparent with longer follow-up. Despite these limitations, this study addresses a number of questions for ICH disparities in the Stroke Belt, as well as topics relevant to in-hospital assessment and prognostication for ICH.
Conclusion
In a sample of ICH patients in a high-risk population in the Southeast USA, we observed a number of biracial differences in risk factor prevalence, patient demographics, and ICH characteristics. We confirmed prior observations of increased prevalence of hypertension, younger age, and increased rates of code status change to DNR among Black stroke patients. We also demonstrated novel disparities involving hyperlipidemia, admission antiplatelet use, and an improved race-specific ICH score. Aside from these observations, in-hospital mortality and immediate functional outcomes were similar among Black and White ICH patients. The observed biracial disparities in ICH related to prevalence of risk factors, differences in the care given to those suffering from non-survivable injuries, and relevance of prognostic tools such as the ICH score point to areas of potential improvement in ICH care.
Supplementary material
To view supplementary material for this article, please visit https://doi.org/10.1017/cts.2021.21.
Acknowledgments
The authors would like to thank Nita A. Limdi, PharmD, PhD, MSPH, FAHA for graciously providing information about the case–control study and supporting data collection, as well as her guidance and recommendations for statistical analysis. We would also like to thank Chrisly Dillon, MD for mentoring support, encouragement, and being an invaluable resource throughout data collection.
Research reported in this publication was partially supported by the National Center for Advancing Translational Sciences of the National Institutes of Health under award number TL1TR003106. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health. NIH R01HL092173; K24HL133373 (NAL).
Disclosures
The authors have no conflicts of interest to declare.
Open access
