Sex and Age Interactions and Differences in Outcomes After Intracerebral Hemorrhage

Abstract

Background:

Compared to ischemic stroke, sex differences in response to intracerebral hemorrhage (ICH) are largely unexplored, and their potential interactions with patient age have not been examined. This study hypothesized that risk for poor outcome is greater in women with increasing age.

Methods and Results:

The Get With The Guidelines^®–Stroke database was used to assess differences between men and women with ICH. Data from 192,826 ICH patients admitted from January 1, 2009 through March 31, 2014 to 1,728 fully participating sites were analyzed using logistic regression to test interactions between age/sex and outcome.

Results:

In the total study population, 48.9% were women (median age 75; male median age 67). On admission, women over 65 years were less likely to have atrial fibrillation or dyslipidemia, or use antiplatelet therapy or cholesterol reducers, but more likely to suffer worse neurological deficit than men over 65. As age increased, odds of in-hospital mortality increased in both men and women, although the relationship was stronger in men. Odds of combined mortality and discharge to hospice were similar in men and women with increasing age, but odds for discharge to home and independent ambulation at discharge decreased more in women with increasing age.

Conclusions:

After adjusting for covariates, modest sex differences in early outcomes after ICH were linked to age. While statistically significant, detected interactions should be considered in context. Future study may examine whether sex-based interactions represent biologic or treatment differences, unmeasured covariates, or some combination thereof.

Introduction

Intracerebral hemorrhage (ICH) is a prevalent stroke subtype with a high degree of morbidity and mortality. While several patient-specific factors have been repeatedly shown to influence outcome after ICH,¹ strong association between sex and outcome has not been demonstrated. An apparent lack of sex differences in outcome is in sharp contrast to ischemic stroke.² While a review of studies on sex differences in ICH was recently published,³ the lack of obvious sex differences in ICH outcome may be due to at least two factors. First, few studies have directly assessed sex differences in meaningful clinical outcomes using large ICH patient populations. Second, outcome studies that include sex as a covariate often treat sex as a constant without assessing effects of aging. This may not be appropriate.

There are many explanations for potential sex differences in outcomes after ICH as people age. The presence or absence of gonadal hormones is a major biological difference between men and women, and is well known to affect brain responses to injury.⁴ Hormonal changes during menopause have also been associated with incidence and outcome of ischemic stroke⁵ and ICH.⁶ Furthermore, exogenous administration of estrogen and progesterone improves outcomes in models of ICH.^7,8 In addition, nonbiological explanations are also possible. These include sex differences in symptom recognition, timing of and appropriate options for treatment, delivery of care, and premorbid psychosocial environment. However, possible age/sex interactions that affect these factors have not been investigated in ICH.

To test our hypothesis that age affects ICH outcomes differently in men and women, the Get With The Guidelines^® (GWTG)-Stroke database was used to assess the effect of interactions between age and sex on quality-of-care metrics and in-hospital outcomes.

Methods

Present data were generated using the GWTG-Stroke program, which has been previously described.⁹ Briefly, participating hospitals used an Internet-based Patient Management Tool (Outcome Sciences, Inc., Cambridge, MA) to enter data, receive decision support, and obtain feedback through on-demand reports of performance on quality measures. Hospitals were instructed to record data from consecutive stroke admissions, and could also choose to record data from consecutive ICH admissions. Case ascertainment was based on clinical findings during hospitalization, or retrospectively, on diagnosis-related group codes, or both. Eligibility of each case was confirmed at chart review before abstraction.

Trained hospital personnel abstracted data using the Internet-based Patient Management Tool with standardized data definitions and detailed coding instructions. Data included demographics, insurance status, medical history, initial findings from head computerized tomography (CT), in-hospital treatment and events, discharge treatment and counseling, discharge destination, and mortality. Because the internet-based system performed checks to ensure that the reported data were complete and internally consistent, and data quality was monitored for completeness and accuracy, GWTG-Stroke data quality is high.^10,11

Each participating hospital received human research approval to enroll cases without individual patient consent under the common rule, or a waiver of authorization and exemption from subsequent review by their Institutional Review Board. Outcome Sciences, Inc. serves as the data collection and coordination center for GWTG-Stroke. The Duke Clinical Research Institute serves as the data analysis center, and has Institutional Review Board approval to analyze aggregate deidentified data for research purposes.

Patient population

Of the 200,780 patients admitted for ICH from January 1, 2009 through March 31, 2014 at 1,745 hospitals in the GWTG-Stroke program, 7,954 were excluded because they were discharged/transferred to another acute care facility, left against medical advice, or lacked discharge destination data. The final analysis sample consisted of 192,826 ICH admissions from 1,728 hospitals.

Quality-of-care measures

Prespecified stroke quality-of-care measures have been published previously.⁹ Only quality measures that were applicable to the care of ICH patients were analyzed, that is, four process measures and two secondary prevention measures. The process measures were deep venous thrombosis (DVT) screening and prophylaxis started by the second hospital day for nonambulatory patients (DVT Prophylaxis), in-hospital rehabilitation (Rehabilitation), dysphagia screening before oral intake (Dysphagia Screen), and door-to-CT time ≤25 minutes in patients with symptoms lasting <3 hours (Door-to-Image). The secondary prevention measures were counseling on tobacco abstention (Smoking Counsel) and stroke risk factors, warning signs, and treatment (Stroke Education). Acceptable options for DVT Prophylaxis included pneumatic compression devices and/or anticoagulants. Patients with documented reasons for nontreatment related to an individual measure were identified by the local sites and excluded from analysis for that measure. For example, nonsmokers were not eligible for Smoking Cessation, and patients who died were not eligible for Stroke Education.

Outcome measures

Prespecified stroke outcome measures have been published previously.⁹ Outcome measures specific to ICH included in-hospital mortality, combined in-hospital mortality or discharge to hospice, discharge to home, independent ambulation at discharge, and total length of hospital stay. Each outcome was modeled separately.

Statistical analysis

Demographic and clinical characteristics, comorbidities, and hospital characteristics are described for ICH patients overall and by sex/age groups using proportions for categorical variables and medians with 25th and 75th percentiles for continuous variables. Differences in these characteristics were compared using chi-square tests for categorical variables and Wilcoxon rank-sum tests for continuous variables. Results are reported as means with standard deviations, or minima and maxima of continuous variables where appropriate.

Logistic regression was used to assess the interaction between age (as a continuous variable) and sex with respect to outcomes. The adjusted model included race (white vs. other), medical history of atrial fibrillation/flutter (AF), previous stroke/TIA, coronary artery disease (CAD)/prior myocardial infarction (MI), carotid stenosis, insulin- and noninsulin-treated diabetes mellitus (DM), peripheral vascular disease, hypertension (HTN), dyslipidemia, smoking, international normalized ratio (INR), glucose, systolic blood pressure, creatinine, prior antiplatelet use, prior anticoagulant use, region, hospital type (teaching/nonteaching), number of beds, annual ICH volume, rural location, and The Joint Commission primary stroke center status. To assess severity of initial injury, the National Institutes of Health Stroke Scale (NIHSS) score was incorporated into models when available. Linear relationships between continuous variables and outcomes were assessed by comparing the fit of a restricted cubic spline and transformations made as needed. No violation of this assumption was identified for age.

Multiple imputation was used for missing data in regression models. However, medical history variables were imputed to “no,” since it was assumed that hospital personnel did not fill in collection form sections when “no” answers applied. NIHSS and hospital characteristics were not imputed. Twenty-five imputations were used for all other variables, and results reflect the final estimates accounting for variation due to “missingness.” Regression models were assessed in the subset with nonmissing NIHSS by further adjustment for NIH stroke scale. A p-value less than 0.05 was considered significant. All statistical analyses were performed using SAS Version 9.3 software (SAS Institute, Cary, NC).

Results

To compare baseline characteristics by categories of age and sex in the total study cohort (n = 192,826), age was dichotomized at 65 years (43,105 male patients <65 years; 55,652 male patients ≥65 years; 29,419 female patients <65 years; 64,650 female patients ≥65 years; Table 1). While the total study cohort was nearly evenly split between men and women (48.9% women), there were obvious sample size differences between men and women above or below 65 years of age. As expected, nearly all measured factors varied by age, but sex differences in the same age group were few. Notably, Hispanic race/ethnicity and self-pay or no insurance was highest in younger males; Medicaid payment was highest in younger females; and arrival by EMS was highest in older women and lowest in younger women. Regarding patient-specific factors, history of HTN was lowest in younger women, and history of smoking was lowest in older women. However, history of AF, CAD/prior MI, DM, and dyslipidemia was highest in older males. Accordingly, anticoagulant, antiplatelet, cholesterol reducer, and diabetic medication use were all highest in older men. Interestingly, admission serum creatinine concentration was highest in younger men, and admission NIHSS scores were highest in older women.

Table 1.

Patient Characteristics

Characteristic	Overall, n = 192,826	Men <65 years of age, n = 43,105	Men ≥65 years of age, n = 55,652	Women <65 years of age, n = 29,419	Women ≥65 years of age, n = 64,650
Age, median (IQR), year	71 (58–81)	54 (48–60)	77 (71–83)	54 (48–60)	80 (74–86)
Race/ethnicity, %
Other	9.5	11.3	8.5	11	8.4
Hispanic	8.3	13	6.5	9.6	6.1
African American	17	27.1	9.5	29.6	11.1
Caucasian	65.2	48.6	75.5	49.8	74.5
Health insurance, %
Self pay/no insurance	9.3	24.2	1.4	20.3	1.4
Medicare	36.9	10.1	52.9	9.7	52.7
Medicaid	11.4	17.1	5.6	20.1	8.5
Private/VA/Champus	42.4	48.6	39.7	49.9	37.4
BMI, median (IQR), kg/m²	26.5 (23–30.9)	28.1 (24.4–32.9)	26.1 (23.2–29.6)	28.2 (23.8–34.2)	25 (21.6–29.3)
Missing, %	32.2	30.6	32.4	30.3	33.9
Medical history, %
AF/flutter	15.8	5	24	3.7	21.4
CAD/prior MI	19.3	11.3	31.2	8.4	19.2
Prosthetic heart valve	1.5	1	2.1	1.1	1.4
Heart failure	6.7	4.3	8	4.2	8.3
PVD	3.3	1.8	4.9	1.6	3.5
Dyslipidemia	32.2	21.3	41.5	19.6	37.1
Diabetes mellitus	25.5	23.5	29.8	22.8	24.4
Hypertension	73.7	69.4	76.8	65	77.9
Smoker	13.4	24.8	8.5	22.4	6
Carotid stenosis	1.9	0.7	3.2	0.9	2.2
Previous stroke/TIA	24.7	17.1	29.2	19.3	28.8
Prior Medications, %
Anticoagulant	18.4	9.1	27	8.9	22.4
Antiplatelet	35.7	21.7	48.5	19.5	41.9
Antihypertensive	59.9	43.9	69	45.4	69.5
Cholesterol reducer	34.1	20.5	46.4	19.5	39.2
Antiglycemic	18.7	16	23.4	15.8	18
Diagnosis and evaluation
EMS arrival	50.3	46.1	51.1	42.9	55.8
Onset to arrival, median (IQR), min	179 (63–435)	173 (60–428)	186 (68–435)	168 (60–415)	180 (63–450)
Missing, %	52.1	51.9	51.4	54.4	51.9
Admission NIHSS, median (IQR)	9 (3–20)	9 (3–18)	8 (3–19)	9 (2–19)	10 (3–21)
Missing, %	52.5	52.5	50.9	54.6	52.9
INR, median (IQR)	1.1 (1–1.2)	1 (1–1.1)	1.1 (1–1.3)	1 (1–1.1)	1.1 (1–1.2)
Missing, %	27.2	25.9	27.3	26	28.5
Glucose, median (IQR), mg/dL	131 (107–169)	130 (107–171)	130 (107–166)	131 (106–174)	132 (108–168)
Missing, %	27.8	27	28.4	26.5	28.5
Systolic blood pressure, median (IQR), mmHg	161 (140–187)	163 (140–193)	159 (140–182)	158 (134–189)	162 (142–187)
Missing, %	21.9	22.7	21.5	22.2	21.6
Creatinine, median (IQR), mg/dL	1 (0.8–1.2)	1 (0.8–1.3)	1.1 (0.9–1.4)	0.8 (0.7–1)	0.9 (0.7–1.1)
Missing, %	21.9	21.2	22.3	21.1	22.5
Hospital characteristics
Academic, %	71	75.2	68.8	75.7	67.9
Rural location, %	2.7	1.8	3.3	1.8	3.3
Region, %
West	20.9	22.3	20.3	21.4	20.3
South	35.9	39	34.1	39.9	33.4
Midwest	19.5	18.2	20.2	19.3	19.9
Northeast	23.7	20.6	25.3	19.4	26.4
Number of beds, median (IQR)	428 (306–623)	462 (331–679)	416 (296–610)	465 (334–679)	408 (280–601)
Annual ICH admissions, median (IQR)	51.7 (29.1–89)	55.9 (33.4–95.5)	49.2 (27.3–85.9)	57.2 (33.6–96.7)	47.5 (26.1–81.3)
Primary stroke center, %	56.3	56.6	56.5	57.1	55.6

All variables demonstrated a p < 0.001 between groups.

AF, atrial fibrillation; BMI, body mass index; CAD, coronary artery disease; EMS, emergency medical service; ICH, intracerebral hemorrhage; INR, international normalized ratio; IQR, interquartile range; MI, myocardial infarction; n, number of subjects; NIHSS, National Institutes of Health Stroke Scale; PVD, peripheral vascular disease; TIA, transient ischemic attack; VA, Veterans Affairs.

Associations between age and quality-of-care measures were generally similar in men and women (Table 2). Door-to-Image was the lone exception. The association between older age and decreased likelihood of “within window” imaging was more pronounced in women than men. The association between age, and combined mortality or discharge to hospice was similar for men and women (Table 3). The association between age and in-hospital mortality was stronger among men than women such that every 10-year increase in age was associated with a 9% increase in odds of death for men, but only 6% increase in odds of death for women (Fig. 1). However, among those discharged alive, women showed stronger associations between older age and lower frequency of independent ambulation and discharge to home compared to men.

FIG. 1.

Mortality rates for men and women as they age. Plots for the observed mortality rates for males and females at each 5-year age group and the predicted probabilities for in-hospital mortality (A, unadjusted and B, adjusted). The plots show data between the 5th and 95th percentiles of age.

Table 2.

Associations of Sex/Age with Quality-of-Care Measures After Intracerebral Hemorrhage

				Unadjusted model yearly increase in age			Adjusted model per yearly increase in age
Quality of care measures	Group	<65 years of age, n (%)	≥65 years of age, n (%)	OR	95% CI	p	OR	95% CI	p
DVT prophylaxis				Interaction		0.047	Interaction		0.0639
	Male	24635 (97.05)	27853 (96.94)	0.987	0.96–1.015	0.354	0.987	0.95–1.023	0.478
	Female	16647 (97.4)	29732 (96.91)	0.95	0.925–0.976	<0.001	0.944	0.91–0.977	0.001
Smoking cessation				Interaction		0.623	Interaction		0.876
	Male	7591 (96.26)	2840 (94.32)	0.905	0.84–0.974	0.008	0.918	0.841–0.995	0.045
	Female	4646 (95.68)	2260 (95.16)	0.929	0.859–1.005	0.065	0.926	0.854–1.007	0.085
Rehabilitation				Interaction		0.002	Interaction		0.06
	Male	30584 (96.47)	33477 (97.73)	1.232	1.202–1.264	<0.001	1.263	1.219–1.308	<0.001
	Female	20700 (95.93)	36313 (98.03)	1.295	1.262–1.329	<0.001	1.309	1.266–1.353	<0.001
Dysphagia screening				Interaction		<0.001	Interaction		0.118
	Male	23721 (81.29)	27849 (81.26)	1.046	1.034–1.058	<0.001	1.061	1.046–1.077	<0.001
	Female	15552 (79.42)	30793 (81.74)	1.077	1.063–1.09	<0.001	1.078	1.061–1.095	<0.001
Stroke education				Interaction		0.093	Interaction		0.559
	Male	12605 (82.99)	9100 (81.69)	1.028	1.009–1.047	0.003	1.031	1.007–1.055	0.011
	Female	8474 (80.61)	8547 (81.12)	1.052	1.031–1.073	<0.001	1.04	1.016–1.064	0.001
Door-to-image				Interaction		0.029	Interaction		0.016
	Male	8360 (32.21)	10475 (30.27)	0.966	0.954–0.978	<0.001	0.952	0.939–0.965	<0.001
	Female	4935 (29.1)	11817 (28.14)	0.983	0.972–0.955	0.004	0.933	0.921–0.945	<0.001

Descriptive statistics for men and women above and below 65 years of age, and ORs of quality-of-care measures for increasing age per year, taken from logistic regression models, including the interaction between age (as a continuous variable) and sex. Stated 95% CIs that do not contain 1.00 suggest trends in age for the group of interest. An OR of X indicates that for each yearly increase in age, the odds of the outcome increase by X times. An interaction p < 0.05 indicates that age trends differ between male and female patients.

CI, confidence interval; DVT, deep venous thrombosis; OR, odds ratio.

Table 3.

Associations of Sex/Age with Outcomes After Intracerebral Hemorrhage

				Unadjusted model per yearly increase in age			Adjusted model per yearly increase in age
Outcomes	Group	<65 years of age, n (%)	≥65 years of age, n (%)	OR	95% CI	p	OR	95% CI	p
In-hospital mortality				Interaction		0.0245	Interaction		0.0003
	Male	9543 (22.14)	14837 (26.66)	1.116	1.104–1.128	<0.0001	1.092	1.079–1.105	<0.0001
	Female	6505 (22.11)	17489 (27.05)	1.099	1.086–1.111	<0.0001	1.064	1.050–1.077	<0.0001
In-hospital mortality or discharged to hospice				Interaction		0.0314	Interaction		0.1270
	Male	10635 (24.67)	19837 (35.64)	1.255	1.242–1.269	<0.0001	1.238	1.223–1.253	<0.0001
	Female	7255 (24.66)	25058 (38.76)	1.280	1.267–1.294	<0.0001	1.252	1.237–1.267	<0.0001
Discharged to home				Interaction		<0.0001	Interaction		<0.0001
	Male	15407 (35.74)	11388 (20.46)	0.732	0.724, 0.740	<0.0001	0.734	0.725–0.743	<0.0001
	Female	10692 (36.34)	10886 (16.84)	0.688	0.681–0.696	<0.0001	0.702	0.693–0.712	<0.0001
Independent ambulation at discharge				Interaction		<0.0001	Interaction		<0.0001
	Male	11678 (32.67)	9319 (20.11)	0.757	0.748–0.766	<0.0001	0.748	0.737–0.758	<0.0001
	Female	7766 (31.84)	8611 (16.00)	0.712	0.704–0.721	<0.0001	0.717	0.707–0.727	<0.0001
Length of stay >5 days				Interaction		0.0144	Interaction		<0.0001
	Male	14593 (52.51)	15425 (40.77)	0.865	0.855–0.875	<0.0001	0.884	0.874–0.895	<0.0001
	Female	9296 (51.00)	17394 (37.97)	0.850	0.840–0.860	<0.0001	0.856	0.845–0.867	<0.0001

Descriptive statistics for men and women above and below 65 years of age, and odds ratios of outcomes for increasing age per year, taken from logistic regression models, including the interaction between age (as a continuous variable) and sex. Stated 95% CIs that do not contain 1.00 suggest trends in age for the group of interest. An odds ratio of X indicates that for each 10-year increase in age, the odds of the outcome increase by X times. An interaction p < 0.05 indicates that age trends differ between male and female patients.

In the subset of ICH patients with NIHSS scores (Table 4), sex differences between age and in-hospital mortality were more pronounced, such that for every 10-year increase in age, odds of death increased 13% for men, but only 6% for women. Age/sex interactions found for discharge to home and independent ambulation were nonsignificant when NIHSS scores were included. Similarly, age/sex interactions were no longer significant for quality-of-care measures, including Door-to-Image times, when NIHSS scores were included.

Table 4.

Associations of Sex/Age with Outcomes After Intracerebral Hemorrhage with National Institutes of Health Stroke Scale Scores as a Covariate

				Unadjusted model per yearly increase in age			Adjusted model per yearly increase in age
Outcomes	Group	<65 years of age, n (%)	≥65 years of age, n (%)	OR	95% CI	p	OR	95% CI	p
In-hospital mortality				Interaction		0.144	Interaction		<0.001
	Male	3447 (16.85)	5936 (21.71)	1.154	1.132–1.176	<0.001	1.128	1.103–1.154	<0.001
	Female	2199 (16.48)	6652 (21.84)	1.135	1.116–1.154	<0.001	1.06	1.038–1.082	<0.001
In-hospital mortality or discharged to hospice				Interaction		0.049	Interaction		0.874
	Male	3898 (19.05)	8402 (30.75)	1.329	1.304–1.354	<0.001	1.391	1.357–1.425	<0.001
	Female	2496 (18.7)	10251 (33.66)	1.36	1.336–1.385	<0.001	1.388	1.357–1.42	<0.001
Discharged to home				Interaction		<0.001	Interaction		0.29
	Male	7478 (36.54)	5730 (20.97)	0.732	0.717–0.739	<0.001	0.661	0.646–0.677	<0.001
	Female	5039 (37.76)	5330 (17.5)	0.688	0.677–0.689	<0.001	0.65	0.635–0.666	<0.001
Independent ambulation at discharge				Interaction		<0.001	Interaction		0.869
	Male	5905 (34.66)	4979 (21.91)	0.756	0.745–0.768	<0.001	0.695	0.635–0.71	<0.001
	Female	3791 (34.13)	4420 (17.49)	0.71	0.698–0.722	<0.001	0.693	0.676–0.71	<0.001
Length of stay >5 days				Interaction		0.712	Interaction		0.043
	Male	7987 (54.87)	8526 (43.03)	0.858	0.845–0.871	<0.001	0.886	0.871–0.901	<0.001
	Female	4840 (53.38)	9442 (41.25)	0.854	0.84–0.869	<0.001	0.867	0.851–0.882	<0.001

Odds ratios of outcomes for increasing age per year, taken from logistic regression models, including the interaction between age (as a continuous variable) and sex. Stated 95% CIs that do not contain 1.00 suggest trends in age for the group of interest. An odds ratio of X indicates that for each 10-year increase in age, the odds of the outcome increase by X times. An interaction p < 0.05 indicates that age trends differ between male and female patients.

Discussion

Based on analyses of this large ICH sample, odds of mortality increase to a modestly greater degree with increasing age among men compared to women, after adjusting for covariates, including NIHSS score. However, when a combined outcome of both mortality and discharge to hospice is analyzed, the increase in odds with advancing age is no longer significantly different between men and women. Furthermore, the odds of independent ambulation or discharge to home decrease to a lesser degree with increasing age among men compared to women. Finally, sex does not appear to be associated with differences in odds of meeting quality-of-care measures after ICH among GWTG-Stroke hospitals.

At older ages, increasing odds of mortality seemingly run counter to increased odds of discharge to home for men. However, the overall relationship between age and early mortality may be similar in men and women, since no age/sex interaction effect was found for the composite endpoint of death or discharge to hospice. One possible explanation for this finding is that men have a greater tendency to die in the hospital, and women tend to be discharged to hospice. Interestingly, this explanation is supported by an earlier hospital-based study of 270 patients with ICH, which found that new, early, do-not-resuscitate orders were more likely in men, after adjusting for covariates.¹² However, sex differences in effects of early comfort measures were not analyzed in this study. Notably, inclusion of NIHSS scores in these models negated age/sex interactions for home discharge and ambulation at discharge, while differences in odds of in-hospital mortality between men and women became more pronounced as they age. Furthermore, it is important to note that, while statistically significant, all differences in odds ratios (ORs) found between men and women as they age were modest.

As in prior studies that used the GWTG-Stroke database,¹⁰ adherence to quality-of-care measures was consistently high in this cohort of subjects. However, after adjusting for covariates, many of these measures did not vary significantly between groups. While the latency from patient presentation to initial imaging was the lone measure affected by the interaction of age and sex, this association was lost when NIHSS scores were included. These findings illustrate the complexities in associating quality measures with outcomes such as mortality. Because adherence to the quality measures was so high overall, small differences in adherence between men and women were not likely to have been a major factor in the differences in outcomes. Finally, since many of the GWTG-Stroke quality-of-care metrics were primarily created with ischemic stroke in mind, relevant ICH-specific metrics may not have been incorporated into these models.

These findings highlight complexities of evaluating sex differences after ICH. Regarding mortality, many studies have failed to find sex differences after ICH after covariate adjustment, including studies from Finland,¹³ Sweden,¹⁴ the Netherlands,¹⁵ Spain,¹⁶ Greece,¹⁷ Turkey,¹⁸ Japan,¹⁹ and China.²⁰ Furthermore, meta-analysis found no sex difference in ICH mortality, although data were largely pre-2000.²¹ However, none of these studies assessed potential interactions between sex and age.

In contrast, other studies have found that American women have higher overall in-hospital mortality,²² while men may have higher age-adjusted mortality at 4 weeks after ICH (relative risk [RR] = 1.13; 95% confidence interval [CI] = 1.04–1.22).²³ According to 4-year national death certificate data, American women under 65 years have lower mortality after ICH than men (RR = 0.82; 95% CI = 0.81–0.83).²⁴ Zia et al. found that Swedish men over 75 years have increased mortality at 28 days (OR = 1.5; 95% CI = 1.1–2.2) and at 3 years (OR = 1.7; 95% CI = 1.3–2.3) after ICH compared with age-matched women.²⁵ In seeming contradiction, a separate Swedish study found that women had higher mortality at 3 months (p = 0.048) and at 1 year (p = 0.014) after ICH compared with men, when controlled for age.¹⁴ Regardless, none of these studies assessed an interaction between age and sex for effect on ICH mortality.

While mortality differences have been studied, potential sex differences in neurological outcomes and recovery after ICH are woefully understudied. Women may have worse early^22,26 and better late neurological outcomes.²⁷ However, other studies find no sex difference in neurological outcome at 3 and 6 months after ICH.^28,29 However, the investigators did not assess the interaction between age and sex. Further complicating these mixed results, sex differences may be influenced regionally or by racial/ethnic backgrounds with neurological outcomes differing in Asian populations.^30,31 However, study of Of course, any differences, or lack thereof, may be related to outcome metric and timing of measurement, geographic location, race/ethnicity, age, and gonadal hormone status (e.g., pre- or postmenopausal, hormone replacement therapy, or oophorectomy) of the study population.

This dataset represents the largest, most comprehensive study to date of sex differences in North American patients with ICH. Furthermore, a wide range of settings, from a large number of community hospitals to academic training hospitals, was included. In addition, these findings lend further support to the notion that age is a strong modifier of ICH outcome. However, there are several limitations worth noting. (1) Hospitals that participate in the GWTG-Stroke program may be different from other hospitals. For example, these hospitals are more likely to be Joint Commission-Certified Primary Stroke Centers in urban academic settings. (2) The GWTG-Stroke database was not initially set up to determine associations with long-term outcomes. It is clear that early limitations in care influence the outcome.¹² Given the disparities in short-term outcomes based on the present dataset, it would be of interest to determine whether ambulation at discharge plays a significant role in long-term recovery. (3) Association with socioeconomic status or race/ethnicity was not directly assessed, which may influence outcomes, as demonstrated in countries with universal healthcare.^32,33 (4) Controlling for gonadal hormone concentration at the time of ICH was not possible. Serum hormone concentration can be affected by oophorectomy, comorbid disease, perimenopause/menopausal status, and hormone replacement therapy. (5) The most relevant missing variable was NIHSS data, which limit the ability to assess ICH severity. To this end, hemorrhage volumes were not captured in this database, and thus, could not be incorporated in our models despite clear association with outcome.^34
–36 (6) Only some of the analyzed care measures are recommended by American Heart Association guidelines for ICH. Published after the GWTG Patient Management Tool was in place, the 2010 guidelines recommend DVT screening and prophylaxis, as well as rapid imaging with angiography to discriminate ICH from other stroke subtypes.³⁷ Earlier guidelines from 2007 recommended smoking cessation to prevent recurrent ICH and acknowledged high prevalence of oral dysphagia without specific recommendations for screening.³⁸ (7) Finally, as with all retrospective analyses, the variables and outcome measures in our models cannot establish causation.

In conclusion, as men and women age, the rate of increasing odds of in-hospital mortality is only modestly higher in men compared to women, after adjusting for covariates. However, odds of independent ambulation at discharge and discharge to home decrease at a slower rate in men compared to women as they age. In addition, sex does not appear to be associated with age differences in odds of meeting quality-of-care measures. These findings highlight the complexity of sex differences in outcomes after ICH. The causes of these associations remain unclear and require further in-depth study.

Footnotes

Acknowledgments

The authors gratefully acknowledge the assistance of Kathy Gage, Research Associate in the Department of Anesthesiology at Duke University, for her critical review and revision of the article.

Funding Sources

The GWTG-Stroke program is provided by the American Heart Association/American Stroke Association. The GWTG-Stroke program is currently supported, in part, by a charitable contribution from Ortho-McNeil. GWTG-Stroke has been funded in the past through support from Boeringher-Ingelheim and Merck, Bristol-Myers Squibb/Sanofi Pharmaceutical Partnership, and the AHA Pharmaceutical Roundtable.

Author Disclosure Statement

Dr. Deepak L. Bhatt discloses the following relationships: Advisory Board—Cardax, Elsevier Practice Update Cardiology, Medscape Cardiology, Regado Biosciences; Board of Directors—Boston VA Research Institute, Society of Cardiovascular Patient Care; Chair—American Heart Association GWTG Steering Committee; Data Monitoring Committees—Duke Clinical Research Institute, Harvard Clinical Research Institute, Mayo Clinic, Population Health Research Institute; Honoraria—American College of Cardiology (Senior Associate Editor, Clinical Trials and News, ACC.org), Belvoir Publications (Editor in Chief, Harvard Heart Letter), Duke Clinical Research Institute (clinical trial steering committees), Harvard Clinical Research Institute (clinical trial steering committee), HMP Communications (Editor in Chief, Journal of Invasive Cardiology), Journal of the American College of Cardiology (Associate Editor; Section Editor, Pharmacology), Population Health Research Institute (clinical trial steering committee), Slack Publications (Chief Medical Editor, Cardiology Today's Intervention), and WebMD (CME steering committees); Other—Clinical Cardiology (Deputy Editor); Research Funding–Amarin, AstraZeneca, Bristol-Myers Squibb, Eisai, Ethicon, Forest Laboratories, Ischemix, Medtronic, Pfizer, Roche, Sanofi Aventis, The Medicines Company; Unfunded Research–FlowCo, PLx Pharma, Takeda. Dr. Fonarow reports funding from PCORI. All other authors have no conflicts of interest to report.

References

Hemphill

3rd , Bonovich

, Besmertis

, Manley

, Johnston

. The ich score: A simple, reliable grading scale for intracerebral hemorrhage. Stroke, 2001; 32:891–897.

Bushnell

, Reeves

, Zhao

, et al. Sex differences in quality of life after ischemic stroke. Neurology, 2014; 82:922–931.

Gokhale

, Caplan

, James

. Sex differences in incidence, pathophysiology, and outcome of primary intracerebral hemorrhage. Stroke, 2015; 46:886–892.

Liu

, Kelley

, Herson

, Hurn

. Neuroprotection of sex steroids. Minerva Endocrinol, 2010; 35:127–143.

Lisabeth

, Bushnell

. Stroke risk in women: The role of menopause and hormone therapy. Lancet Neurol, 2012; 11:82–91.

Feldmann

, Broderick

, Kernan

, et al. Major risk factors for intracerebral hemorrhage in the young are modifiable. Stroke, 2005; 36:1881–1885.

Auriat

, Plahta

, McGie

, Yan

, Colbourne

. 17beta-estradiol pretreatment reduces bleeding and brain injury after intracerebral hemorrhagic stroke in male rats. J Cereb Blood Flow Metab, 2005; 25:247–256.

Nakamura

, Hua

, Keep

, Park

, Xi

, Hoff

. Estrogen therapy for experimental intracerebral hemorrhage in rats. J Neurosurg, 2005; 103:97–103.

LaBresh

, Reeves

, Frankel

, Albright

, Schwamm

. Hospital treatment of patients with ischemic stroke or transient ischemic attack using the “get with the guidelines” program. Arch Intern Med, 2008; 168:411–417.

10.

Fonarow

, Reeves

, Smith

, et al. Characteristics, performance measures, and in-hospital outcomes of the first one million stroke and transient ischemic attack admissions in get with the guidelines-stroke. Circ Cardiovasc Qual Outcomes, 2010; 3:291–302.

11.

Xian

, Fonarow

, Reeves

, et al. Data quality in the American heart association get with the guidelines-stroke (GWTG-stroke): Results from a national data validation audit. Am Heart J, 2012; 163:392–398, e391.

12.

Zahuranec

, Brown

, Lisabeth

, et al. Early care limitations independently predict mortality after intracerebral hemorrhage. Neurology, 2007; 68:1651–1657.

13.

Fogelholm

, Murros

, Rissanen

, Avikainen

. Long term survival after primary intracerebral haemorrhage: A retrospective population based study. J Neurol Neurosurg Psychiatry, 2005; 76:1534–1538.

14.

Nilsson

, Lindgren

, Brandt

, Saveland

. Prediction of death in patients with primary intracerebral hemorrhage: A prospective study of a defined population. J Neurosurg, 2002; 97:531–536.

15.

Franke

, van Swieten

, Algra

, van Gijn

. Prognostic factors in patients with intracerebral haematoma. J Neurol Neurosurg Psychiatry, 1992; 55:653–657.

16.

Rodriguez-Luna

, Rubiera

, Ribo

, et al. Ultraearly hematoma growth predicts poor outcome after acute intracerebral hemorrhage. Neurology, 2011; 77:1599–1604.

17.

Vemmos

, Bots

, Tsibouris

, et al. Stroke incidence and case fatality in Southern Greece: The arcadia stroke registry. Stroke, 1999; 30:363–370.

18.

Yesilot

, Koyuncu

, Coban

, Tuncay

, Bahar

. Gender differences in acute stroke: Istanbul medical school stroke registry. Neurol India, 2011; 59:174–179.

19.

Kimura

, Takishita

, Muratani

, et al. Demographic study of first-ever stroke and acute myocardial infarction in Okinawa, Japan. Intern Med, 1998; 37:736–745.

20.

Zhang

, Yang

, Hong

, et al. Proportion of different subtypes of stroke in china. Stroke, 2003; 34:2091–2096.

21.

van Asch

, Luitse

, Rinkel

, van der Tweel

, Algra

, Klijn

. Incidence, case fatality, and functional outcome of intracerebral haemorrhage over time, according to age, sex, and ethnic origin: A systematic review and meta-analysis. Lancet Neurol, 2010; 9:167–176.

22.

Ganti

, Jain

, Yerragondu

, et al. Female gender remains an independent risk factor for poor outcome after acute nontraumatic intracerebral hemorrhage. Neurol Res Int, 2013; 2013:219097.

23.

Sheikh

, Bullock

. Effect of measurement on sex difference in stroke mortality. Stroke, 2007; 38:1085–1087.

24.

Ayala

, Croft

, Greenlund

, et al. Sex differences in us mortality rates for stroke and stroke subtypes by race/ethnicity and age, 1995–1998. Stroke, 2002; 33:1197–1201.

25.

Zia

, Engstrom

, Svensson

, Norrving

, Pessah-Rasmussen

. Three-year survival and stroke recurrence rates in patients with primary intracerebral hemorrhage. Stroke, 2009; 40:3567–3573.

26.

Umeano

, Phillips-Bute

, Hailey

, et al. Gender and age interact to affect early outcome after intracerebral hemorrhage. PLoS One, 2013; 8:e81664.

27.

Wang

, Lu

, Wang

, et al. Clinical characteristics, management, and functional outcomes in chinese patients within the first year after intracerebral hemorrhage: Analysis from china national stroke registry. CNS Neurosci Ther, 2012; 18:773–780.

28.

Rodriguez-Luna

, Rubiera

, Ribo

, et al. Serum low-density lipoprotein cholesterol level predicts hematoma growth and clinical outcome after acute intracerebral hemorrhage. Stroke, 2011; 42:2447–2452.

29.

Roquer

, Rodriguez-Campello

, Jimenez-Conde

, et al. Sex-related differences in primary intracerebral hemorrhage. Neurology, 2016; 87:257–262.

30.

Hsieh

, Ang

, Ng

, Allen

, King

. Comparison of gender differences in intracerebral hemorrhage in a multi-ethnic Asian population. PLoS One, 2016; 11:e0152945.

31.

Zhou

, Zhang

, Arima

, et al. Sex differences in clinical characteristics and outcomes after intracerebral haemorrhage: Results from a 12-month prospective stroke registry in Nanjing, China. BMC Neurol, 2014; 14:172.

32.

Rey

, Faouzi

, Huchmand-Zadeh

, Michel

. Stroke initial severity and outcome relative to insurance status in a universal health care system in Switzerland. Eur J Neurol, 2011; 18:1094–1097.

33.

Kapral

, Wang

, Mamdani

, Tu

. Effect of socioeconomic status on treatment and mortality after stroke. Stroke, 2002; 33:268–273.

34.

Rost

, Smith

, Chang

, et al. Prediction of functional outcome in patients with primary intracerebral hemorrhage: The FUNC score. Stroke, 2008; 39:2304–2309.

35.

, Zhang

, Qin

. Risk factors of early death in patients with hypertensive intracerebral hemorrhage during hospitalization. Acta Neurochir Suppl, 2011; 111:387–391.

36.

Miyai

, Sonoda

, Nagai

, et al. Results of new policies for inpatient rehabilitation coverage in japan. Neurorehabil Neural Repair, 2011; 25:540–547.

37.

Morgenstern

, Hemphill

3rd , Anderson

, et al. Guidelines for the management of spontaneous intracerebral hemorrhage: A guideline for healthcare professionals from the American Heart Association/American Stroke Association. Stroke, 2010; 41:2108–2129.

38.

Broderick

, Connolly

, Feldmann

, et al. Guidelines for the management of spontaneous intracerebral hemorrhage in adults: 2007 update: A guideline from the American Heart Association/American Stroke Association Stroke Council, High Blood Pressure Research Council, and the Quality of Care and Outcomes in Research Interdisciplinary Working Group. Circulation, 2007; 116:e391–e413.