Young Women Had More Strokes Than Young Men in a Large, United States Claims Sample

Abstract

Background and Purpose:

Cardiovascular risk factors, which are overall more prevalent in men, are considered the major risk factors for strokes amongst young adults. However, recent European data found the incidence of strokes to be higher in young women. Using a large U.S. claims sample, we examined sex differences in the index stroke rate of young adults.

Methods:

We performed a retrospective cohort study of enrollees in a 10% random sample of PharMetrics, a nationally representative claims database of insured Americans from 2001 to 2014. Outcomes were index ischemic stroke events, based on inpatient admissions using ICD-9 codes. The index stroke rate was estimated from Poisson rate models with time varying covariates for two-year periods, stratified by sex and age groups.

Results:

We identified 20,554 index strokes(50.4% women; mean age 63) including 5,198 in young adults ages 15-54. There was no difference by sex in the index stroke rate in the extremes of age groups 15-24 and ≥75 years old. However, in the 25-34 and 35-44 year age groups, more women had strokes than men (Incidence Rate Ratio(IRR) men:women 0.70, 95%CI:0.57-0.86; 0.87, 95%CI:0.78-0.98, respectively). In contrast, in the 45-54, 55-64 ,and 65-74 year age groups, more men had strokes(IRR 1.25 95%CI:1.16-1.33; 1.41 95%CI:1.18-1.34; 1.18 95%CI:1.12-125, respectively).

Conclusions:

More young women than men have strokes, suggesting possible importance of sex-mediated etiologies of stroke. Understanding these drivers is critical to stroke treatment and prevention efforts in young adults.

INTRODUCTION

While stroke incidence in older adults has declined dramatically, stroke in young adults continues to rise.¹ The social and economic burden of strokes in young adults are substantial due to loss of prime productive years, longer time spent with disability, and increased mortality.² Despite the severity of these long-term consequences, etiologies of strokes in the young remain poorly understood.

Cardiovascular risk factors are considered the major drivers of strokes amongst young adults,^{3, 4} and prior studies found cardiovascular risk factors to be more prevalent among young adults with strokes than their peers in the general population.⁵ However, data from Europe suggests that despite having fewer cardiovascular risk factors, young women have more strokes than men.^{6, 7} More recently, a Dutch cohort of young adults found that women had strokes at a younger age with fewer vascular risk factors.⁶ Studying stroke in young adults is challenging because the incidence is relatively low and additional studies are necessary to verify these novel findings.

Using a large U.S. longitudinal administrative sample, we examine whether sex-based differences in young adult stroke incidence also exist in the U.S. and whether these differences changed over time.

METHODS

This retrospective cohort study includes a 10% randomly selected sample of a large commercial health insurance database, PharMetrics, including all paid medical and pharmacy claims for more than 70 million members across the U.S. Subjects were included in this study if they were (1)at least 15 years old and (2)enrolled continuously for at least six consecutive months between January 1, 2001 and December 31, 2014. This study protocol, which relied on a pre-existing, de-identified database, was ‘exempt’ by the University of Colorado Institutional Review Board and patient consent was waived. The data that support the findings of this study are available from the corresponding author upon reasonable request.

Outcomes and Variables

Stroke was defined as an inpatient admission with acute ischemic stroke as the primary diagnosis(using ICD-9 codes 433.x1, 434.x1, and 436).⁶ The first inpatient admission for stroke was considered the index stroke event. Patients with a previous outpatient or inpatient diagnosis of any cerebrovascular disease were excluded. We were only interested in the index stroke event to estimate the incidence of stroke, which reduces confounding from stroke recurrence by sex.

Statistical Approach

The index stroke rate was estimated from Poisson rate models with time varying covariates for two-year periods by age group. Comparisons between index stroke rates were estimated using incident rate ratios(IRR) of men:women. Interactions between sex and calendar year were tested for changes overtime. Statistical significance was determined as p<0.05. All analyses were performed with SAS 14.2(SAS Institute, Cary NC).

RESULTS

Approximately 5.8 million enrollees, including 20,554 index strokes(5,198 in young adults <55) were identified(Table). Half of all strokes(50.4%) occurred in women. The average age for both women and men with index stroke was approximately 63 years(SD 13.8). The index stroke rate increased with age in both men and women.

Table:

Index Stroke per 100,000 enrollees by Age and Sex, 2001-2014

	Women, N=3,051,211		Men, N=2,778,041
Age Group	Index Strokes N	Index Stroke Rate Per 100,000(95% CI)	Index Strokes N	Index Stroke Rate Per 100,000(95% CI)	P value†
15-24	103	6.7(5.5-8.1)	82	5.4(4.3-6.7)	0.14
25-34	244	16.9(14.9-19.2)	150	11.9(10.1-14.0)	<0.01
35-44	694	39.9(37.0-43.0)	547	34.9(32.0-43.0)	0.02
45-54	1610	79.8(76.2-83.9)	1788	99.4(94.9-104.1)	<0.01
55-64	2397	146.2(140.4-152.2)	3010	206.6(199.3-214.1)	<0.01
65-74	2434	386.4(371.1-402.4)	2557	456.0(438.3-474.4)	<0.01
75+	2880	841.8(811.1-873.7)	2058	835.3(799.5-872.8)	0.79

CI=confidence intervals.

^†P value comparing index stroke rate in women versus men.

As shown in the Figure, the IRR indicated more strokes among women than men(i.e. <1) in the youngest age group(15-24), but this difference did not reach statistical significance. The IRR indicated more strokes among women in both the 25-34 and 35-44 year age groups. In the next three older age groups(45-54, 55-64, and 65-74), the trend reverses with more strokes among men. The sex differences disappear for patients ≥75 years.

Figure:

Forest Plot of Incidence Rate Ratio(IRR) Men:Women by Age Group. Diamond is IRR of index stroke rate. 95%Confidence Interval(CI) in bars. CI crossing 1 indicates non-significant difference between men and women.

The IRR varied across the age groups(Supplemental Table I). The IRR in the younger age groups for each 2-year interval, were not statistically different. In contrast, the IRRs for the 2-year intervals between the 45-74 year age groups are statistically significant, with men consistently having more strokes. Statistically significant temporal trends in IRR were not detected in any age group.

DISCUSSION

In this large, retrospective cohort of insured Americans, strokes occurred more often in young women ages 25-44 years than in young men. This is the first study to our knowledge to report higher stroke rates among young women in the U.S. Previous studies in the U.S. using the National Inpatient Sample found that young men ages 15-44 had higher stroke hospitalization rates than young women.^{5, 8} One possible explanation for this discrepancy is that the denominator in the prior study was the number of inpatient admissions, which may underestimate the true incidence. Young women, in general, have more hospitalizations than men due to child bearing, thereby diluting stroke hospitalization rates in women. In contrast, the European 15 Cities Young Stroke Study found a female predominance among those younger than age 34 years by a male to female ratio of 0.7, meaning 30% more strokes occurred in young women. Moreover, compared to men, women had strokes at a younger age with fewer vascular risk factors. Unfortunately, this study did not have a reliable population size in which to calculate an incidence and IRR. Most recently, a Dutch study estimated the incidence of strokes among young adults using a nationwide cohort and found more strokes among young women, ages 18-44. The IRR of ischemic stroke in women compared to men declined with age from 2.04 in ages 18-24 to 1.20 in ages 40-44. These results are consistent with our findings.

There are many potential hypotheses regarding the higher incidence of stroke among young women. Endogenous estrogen may serve a preventative role for cardiovascular events in pre-menopausal women.⁹ This is true for acute myocardial infarction, where incidence is much lower in young women.¹⁰ However, we found the opposite for ischemic strokes. With the exception of obesity, the prevalence of cardiovascular risk factors are higher among young men in the U.S.¹¹ Hence, there may be non-cardiovascular risk factors, which are unique or more prevalent in women, leading to strokes. Some possible explanations include pregnancy, oral contraception, and other female-predominant risk factors including migraine headaches and autoimmune disorders.^{12, 13} A better understanding of the etiologies and risk factors of strokes in young women is needed in future research and an important step towards prevention.

There are some limitations to this study. First, our study sample was drawn from an administrative, commercial database. Few Medicaid and Medicare patients were included and uninsured and self-pay patients are not included. Second, we were unable to study race or ethnicity using this database. Previous studies have shown that the incidence of stroke in young adults is higher among Blacks and Hispanics and future studies examining race/ethnicity differences by age and sex are warranted.¹⁴ Last, due to the administrative nature of this dataset, we do not know the etiology of ischemic stroke nor its severity. Future work linking administrative to clinical data would provide a fuller understanding.

CONCLUSION

We found that stroke incidence is higher in young women than young men, suggesting the importance of sex-mediated risk factors. Understanding these factors is critical to stroke treatment and prevention efforts in young adults.

Nonstandard Abbreviations and Acronyms:

ICD

International Classification of Diseases

IRR

Incidence Rate Ratio

CI

Confidence Interval

SAS

Statistical Analysis System