You are on page 1of 12

ARTICLE

Sex differences in IV thrombolysis treatment for


acute ischemic stroke
A systematic review and meta-analysis
Brent Strong, Lynda D. Lisabeth, PhD, and Mathew Reeves, PhD Correspondence

®
Dr. Reeves
Neurology 2020;95:e11-e22. doi:10.1212/WNL.0000000000009733 reevesm@msu.edu

Abstract
Objective
A prior meta-analysis of reports published between 2000 and 2008 found that women were 30%
less likely to receive IV recombinant tissue plasminogen activator (rtPA) treatment for stroke
than men; we updated this meta-analysis to determine if this sex difference persisted.

Methods
We identified studies that reported sex-specific IV rtPA treatment rates for acute ischemic
stroke published between 2008 and 2018. Eligible studies included representative populations
of patients with ischemic stroke from hospital-based, registry-based, or administrative data.
Random effects odds ratios (ORs) were generated to quantify sex differences.

Results
Downloaded from https://www.neurology.org by King Saud Bin Abdulaziz University for Health on 11 March 2024

Twenty-four eligible studies were identified during this 10-year period. The summary un-
adjusted OR based on 17 studies with data on all ischemic stroke patients was 0.87 (95%
confidence interval [CI], 0.82–0.93), indicating that women had 13% lower odds of receiving
IV rtPA treatment than men. However, substantial between-study variability existed. Lower
treatment odds in women were also observed in 7 studies that provided data on the subgroup of
patients eligible for IV rtPA treatment, although the summary OR of 0.95 (95% CI, 0.88–1.02)
was not statistically significant. Examination of time trends across 33 studies published between
2000 and 2018 found evidence that the sex difference had narrowed in more recent years.

Conclusions
Although there is considerable variability in the findings of individual studies, pooled data from
recent studies show that women with acute stroke are less likely to be treated with IV
thrombolysis compared with men. However, the size of this difference has narrowed compared
to studies published before 2008.

From the Undergraduate Professorial Assistantship Program, Honors College (B.S.), and Department of Epidemiology and Biostatistics, College of Human Medicine (M.R.), Michigan
State University, East Lansing; and Department of Epidemiology, School of Public Health (L.D.L.), University of Michigan, Ann Arbor.

Go to Neurology.org/N for full disclosures. Funding information and disclosures deemed relevant by the authors, if any, are provided at the end of the article.

Copyright © 2020 American Academy of Neurology e11


Copyright © 2020 American Academy of Neurology. Unauthorized reproduction of this article is prohibited.
Glossary
AIS = acute ischemic stroke; CI = confidence interval; OR = odds ratio; rtPA = recombinant tissue plasminogen activator.

Interest in how women are treated following an acute stroke national-level hospital discharge data or Medicare fee-for-
and the impact of sex differences in care on subsequent out- service data—were also eligible. The reference lists of any
comes has grown rapidly in the last 10 years.1–7 The finding of systematic or narrative reviews identified in the search were
a sex difference in clinical care or clinical outcomes following also screened for additional potentially relevant studies. We
an acute stroke raises the concern that such differences rep- excluded conference abstracts, other unpublished gray liter-
resent a bias (or disparity) in either access to or delivery of ature, and non-English articles. We used a standard text for
medical care. However, not all sex differences represent in- guidance on methodologic aspects of conducting the sys-
equitable medical care; sex differences can be a result of a host tematic review and meta-analysis11 and the PRISMA report-
of legitimate factors that influence eligibility for treatment ing guidelines.12
and/or patient preferences for treatment. The failure to
completely account for the full range of relevant factors can Study selection and data abstraction
result in residual confounding and potentially spurious Two authors (B.S., M.R.) screened abstracts and titles for
associations. relevancy and then conducted an independent full-text review
of potentially relevant studies. Studies that met eligibility
Given that IV thrombolysis remains one of the few evidence- criteria then underwent dual data abstraction. Abstracted data
based medical treatments available for acute ischemic stroke included study design, time period of case enrollment, defi-
(AIS), the potential for sex differences in both its utilization nition and size of numerator and denominator populations,
and clinical efficacy has remained of high interest. A prior sex-specific treatment rates (%), odds ratios (ORs) describing
meta-analysis that included 18 studies published between the sex difference (both unadjusted and adjusted), 95% con-
2000 and 2008 found that women had a 30% lower odds of fidence intervals (CIs), and any confounders that were ad-
receiving IV recombinant tissue plasminogen activator (rtPA) justed for. We also abstracted data describing each study’s
for AIS compared to men.8 Since this time, there have been patient population including the mean age and range, sex
Downloaded from https://www.neurology.org by King Saud Bin Abdulaziz University for Health on 11 March 2024

a growing number of publications that address the issue of sex proportion, and proportion of minority subjects. If sex-
differences in access to and use of IV thrombolysis. Some of specific treatment rates or the unadjusted OR of the sex dif-
these studies reported a significant sex difference,6,9 while ference in IV rtPA use was not provided in the publication, we
others did not.10 Given the volume of new studies published calculated them when data were available to do so. When
on the topic and the potential that the increased attention provided, we also abstracted data on patient subgroups eli-
given to the care and treatment of stroke in women might gible for thrombolysis treatment as defined by the combina-
have changed treatment patterns over the last decade, we tion of time of arrival, time of treatment, and
conducted an updated systematic review and meta-analysis to contraindications. However, the definition of these eligible
determine if the use of IV thrombolysis still differed by sex. treatment subgroups varied among studies—for example,
“arrive by 2 hours, treat by 3 with no contraindications”13,14 or
“arrive by 4 hours.”9 Disagreements were resolved by con-
Methods ference. In several cases, authors were contacted to ascertain
further information about their data.
Eligibility criteria and search
We followed the same search strategy to identify potentially Quality scoring
relevant studies that was used in the prior meta-analysis, We developed a quality assessment instrument by adapting
which included studies published up to March 2008.8 We items from the Newcastle–Ottawa Scale.15 All eligible studies
searched MEDLINE, EMBASE, and the ISI Web of Science were scored (0, 1, or 2, where 2 is best and 0 worst) on the
databases for relevant articles published between April 2008 following 4 criteria: (1) the representativeness of the overall
and December 2018. We used the same combination of study population, (2) the number and impact of exclusions
search terms, which included the following: (1) cerebrovas- applied to the initial patient cohort, (3) adjustment for po-
cular accident (MeSH) or “stroke”; (2) “r-tPA” or “tPA” or tential confounding variables that might affect sex differences
“thrombolysis” or “thrombolytics”; and (3) “sex” or “sex in IV rtPA utilization, (4) the method by which the outcome
factors” or “sex ratio” or “sex distribution.”8 Relevant studies (IV rtPA treatment or not) was ascertained. Studies that
were those that reported sex-specific treatment rates of IV provided data specific to the subgroup of patients eligible for
rtPA (thrombolysis) for AIS in typical hospital settings. IV rtPA treatment were not scored for criterion 3 (adjust-
Typical hospital settings could include individual community- ment) because further adjustment among patients already
based or stroke-specialist centers, regional hospital networks, eligible for thrombolysis is unnecessary. To assess overall
hospital systems, or hospital-based stroke registries. Studies study quality, we aggregated the scores to generate a 0–8 scale
that used hospital billing (administrative) data—such as (or 0–6 in the case of studies of eligible subgroups). A more

e12 Neurology | Volume 95, Number 1 | July 7, 2020 Neurology.org/N


Copyright © 2020 American Academy of Neurology. Unauthorized reproduction of this article is prohibited.
detailed description of the quality assessment instrument Data availability
along with examples of the scoring is provided in table e-1 The protocol for this systematic review as well as the data used
(doi.org/10.5061/dryad.1zcrjdfnc). in the meta-analysis are available upon request from the
authors. The data supplement is available from dryad (doi.
In addition to the above quality scoring, we also classified org/10.5061/dryad.1zcrjdfnc).
studies into 1 of 4 groups based on their primary focus.
Studies classified as A were those that had a clear objective to
examine sex differences in IV rtPA utilization. Studies classi- Results
fied as B examined sex differences in IV rtPA effectiveness but
had no clear objective to study IV rtPA utilization per se. The search identified 1,097 initial hits, of which 66 studies
Studies classified as C focused on IV rtPA treatment more passed the initial relevancy screen and underwent full review
broadly but had no particular objective to study sex differ- (figure 1). Of these, 31 met eligibility criteria and provided
ences. Studies classified as D examined sex differences in acute sufficient data to either abstract or calculate an OR comparing
stroke care but had no particular objective to focus on IV rtPA IV rtPA treatment rates in women with those in men. Seven
per se. This classification system was created because studies studies were excluded because they used the same data source
examining a hypothesis that is directly related to sex differ- from a similar time period as another publication (see data
ences in IV rtPA utilization (i.e., group A) might be expected supplement, doi.org/10.5061/dryad.1zcrjdfnc). Thus a total
to have the most validity, whereas those studies that did not of 24 studies were included in the final systematic review and
have any specific objective to study IV rtPA (i.e., group D) meta-analysis; study characteristics and IV rtPA treatment
might have the least. rates by sex for the 24 studies are described in tables 1 and 2.
There were 11 hospital-based, 5 registry-based, and 8 ad-
Statistical analysis ministrative data studies. Seventeen studies provided data on
The primary outcome measure of interest was the unadjusted treatment rates among all patients with AIS (table 1), 2
OR comparing the IV rtPA treatment rate in women with that studies only provided data for an adjusted OR estimate for all
of men, among all AIS admissions. We conducted all meta- patients with AIS (table 1), and 7 studies provided data spe-
analyses using the metan command in Stata (StataCorp, cific to eligible treatment subgroups (table 2). Individual
College Station, TX). Due to the variability in study design, studies differed widely in terms of sample size (range
Downloaded from https://www.neurology.org by King Saud Bin Abdulaziz University for Health on 11 March 2024

we only report random effects–based results. Heterogeneity 602–605,960 for all AIS) and treatment rates (range
was summarized using the Cochran Q statistic and the I2 1.5%–27.8% for all patients with AIS and 10.0%–88.0% for
statistic. The metaninf command was used to conduct an eligible treatment subgroups). The results of the quality as-
influence analysis to investigate the effect of individual studies sessment scoring and the classification system that reflected
on the summary OR. When available, we also analyzed ad- the primary study focus (i.e., groups A–D) are summarized in
justed ORs and ORs calculated among the subgroups of table e-2 (doi.org/10.5061/dryad.1zcrjdfnc). Studies differed
patients eligible for treatment. If we were only provided a 95% widely in terms of quality, with total scores ranging from 1 to 8
CI of the OR, we used both the upper and lower bounds of the with a median of 5. Only 5 (21%) of the 24 studies were
CI to estimate the standard error.16 Other prespecified sub- classified as group A, indicating that they had the specific goal
group analyses included geographic region (North America, of examining sex differences in IV rtPA utilization. Of the
Europe, Asia), study design (hospital-based studies and reg- remaining studies, a total of 3, 9, and 7 studies were classified
istries vs administrative data), quality score (0–8), primary as group B, C, or D, respectively.
study focus (groups A–D), and the time period of publication
(2009–2013, 2014–2018). Pairwise comparisons of summary Seventeen studies provided data on the unadjusted sex dif-
ORs between subgroups were made using a Z test. To ex- ference among all patients with AIS; the forest plot of the
amine longer-term secular trends, we combined the studies unadjusted ORs from the 17 studies is shown in figure 2. The
found in the current search with the 16 studies identified in random effects summary unadjusted OR was 0.87 (95% CI,
the prior meta-analysis (published between 2000 and 2008)8 0.82–0.93), indicating that women had a statistically signifi-
that provided unadjusted OR estimates of sex differences in IV cant 13% lower odds of receiving IV rtPA treatment than men.
thrombolysis use among all AIS admissions. These 33 studies However, the Q statistic was highly significant (p < 0.001) and
were then categorized according to date of publication the I2 statistic was 87.8%, indicating substantial between-
(i.e., 2000–2004, 2005–2008, 2009–2013, 2014–2018) and study variability. The influence analysis revealed that none of
differences were tested using a linear test for trend and pair- the 17 individual studies had a disproportionate effect on the
wise Z tests. overall unadjusted OR.

Standard protocol approvals, registrations, A total of 10 studies provided data on the adjusted OR for the
and patient consents treatment sex difference among all patients with AIS; 8 of
Review board approval and informed consent were not these came from the pool of 17 studies that reported an
obtained because this research makes use of only published, unadjusted OR, while 2 other studies provided data only on
de-identified data. the adjusted OR (table 1). The summary adjusted OR from

Neurology.org/N Neurology | Volume 95, Number 1 | July 7, 2020 e13


Copyright © 2020 American Academy of Neurology. Unauthorized reproduction of this article is prohibited.
the 10 studies was 0.95 (95% CI, 0.89–1.01) but with con- noted, descriptions of the eligible subgroups differed among
siderable between-study heterogeneity (Q p < 0.001, I2 = studies (see footnotes, table 2).
74%) (figure 3). Individual factors that were adjusted for
varied among the studies, but most often included age (n = Prespecified subgroup and sensitivity analyses
10), comorbidities (n = 8), stroke severity (n = 7), race/ The subgroup analysis that examined studies from different
ethnicity (n = 4), and hospital-related factors (n = 4). A list geographical regions found that sex differences in IV rtPA use
of the specific factors adjusted for in each of the 10 studies among all patients with AIS were observed both in European
is provided in table e-3 (doi.org/10.5061/dryad.1zcrjdfnc). and United States–based studies but not in studies from Asia
To further elucidate the effect of adjustment on the ob- (figure e-3, doi.org/10.5061/dryad.1zcrjdfnc). The summary
served sex differences among all AIS, we compared the unadjusted OR for the 6 European studies was 0.82 (95%
results of the 8 studies that included both unadjusted and CI, 0.78–0.85) with no significant heterogeneity (Q p =
adjusted data; the unadjusted summary OR of 0.88 (95% 0.350, I2 = 10.3%), and for the 7 United States–based
CI, 0.81–0.97, Q p < 0.001, I2 = 90.6%) (figure e-1, doi.org/ studies it was 0.85 (95% CI, 0.75–0.96) but with substantial
10.5061/dryad.1zcrjdfnc) was similar to the adjusted between study heterogeneity (Q p < 0.001, I2 = 92.9%).
summary OR of 0.91 (95% CI, 0.87–0.96, Q p = 0.003, I2 = In contrast, the summary OR for the 4 Asian studies was
67.3%) (figure e-2, doi.org/10.5061/dryad.1zcrjdfnc), in- 0.98 (95% CI, 0.94–1.03) with no significant heterogeneity
dicating that adjustment for these various factors made (Q p = 0.889, I2 = 0%). The summary OR for the Asian
relatively little difference in the magnitude of the sex studies was statistically significantly different from the
difference. European (Z test p < 0.001) and United States–based
(Z test p = 0.03) studies.
There were 7 studies that provided data on the subgroup of
patients who were eligible for IV rtPA treatment; the summary When the results of the 11 hospital-based and registry-based
unadjusted OR was 0.95 (95% CI, 0.88–1.02), indicating that the studies were compared to the 6 administrative data studies,
odds of treatment were still lower for women but that the estimate the summary unadjusted ORs among all patients with AIS
was no longer statistically significant within these selected patient were similar (0.90 [95% CI, 0.83–0.97] and 0.86 [95% CI,
subpopulations. There was only modest between-study hetero- 0.78–0.95], respectively; Z test p = 0.522) (figure e-4, doi.org/
geneity (Q test p = 0.186, I2 = 31.7%) (figure 4). As previously 10.5061/dryad.1zcrjdfnc). Both subgroups had either modest
Downloaded from https://www.neurology.org by King Saud Bin Abdulaziz University for Health on 11 March 2024

Figure 1 PRISMA flow diagram12 including reasons for exclusion of full-text articles

e14 Neurology | Volume 95, Number 1 | July 7, 2020 Neurology.org/N


Copyright © 2020 American Academy of Neurology. Unauthorized reproduction of this article is prohibited.
Neurology.org/N

Table 1 Characteristics of 19 included studies providing sex-specific utilization data among all patients with acute ischemic stroke (AIS) classified according to the type of
data provided: all AIS with unadjusted estimates (n = 17 studies), AIS with only an adjusted estimate provided (n = 2 studies)
Copyright © 2020 American Academy of Neurology. Unauthorized reproduction of this article is prohibited.

Female rtPA- Female Male rtPA- Male Adjusted


treated Female treatment treated Male treatment estimate
a
Authors Year Country Time Study design cases, n denominator, n rate, % cases, n denominator, n rate, % provided

All AIS unadjusted


estimates (n = 17)

Allen et al.32 2009 United 2004 Multiple- 39 604 6.5 47 630 7.5 Yes
States hospital
Downloaded from https://www.neurology.org by King Saud Bin Abdulaziz University for Health on 11 March 2024

Asdaghi et al.13 2016 United 2010–2014 Registry 2,397 25,059 9.6 2,606 25,543 10.2 Yes
States

Boehme et al.30 2014 United 2004–2011 Multiple- 679 2,372 28.6 706 2,605 27.1 No
States hospital

Clua-Espuny 2015 Spain 2006–2014 Multiple- 53 597 8.9 60 675 8.9 No


et al.10 hospital

de Ridder et al.9 2013 Holland 2003–2005 Multiple- 314 2,778 11.3 382 2,737 14.0 No
hospital

Eriksson et al.22 2010 Sweden 2003–2008 Registry 957 29,506 3.2 1,578 41,199 3.8 Yes

6
Faigle et al. 2017 United 2007–2011 Administrative 7,905 179,664 4.4 7,877 157,537 5.0 Yes
States

Huang et al.34 2010 China 2006 Registry 178 1810 9.8 307 2,972 10.3 Yes

Kunisawa 2014 Japan 2010–2011 Administrative 236 4,389 5.4 321 6,226 5.2 Yes
et al.31
Neurology | Volume 95, Number 1 | July 7, 2020

Łabuz-Roszak 2018 Poland 2009–2015 Administrative 1,565 35,880 4.4 1,690 33,107 5.1 No
et al.39

Lee et al.17 2018 United 2012–2015 Single- 71 456 15.6 113 470 24 No
States hospital

Nagaraja et al.40 2012 United 2006 Multiple- 8 320 2.5 10 282 3.5 No
States hospital

Nardetto et al.41 2017 Italy 2007–2015 Administrative 1,672 31,414 5.3 2012 29,648 6.8 No

Park et al.25 2013 South 2008 Multiple- 190 2,830 6.7 261 3,805 6.9 Yes
Korea hospital

Santalucia 2013 Italy 2011 Registry 41 410 10 62 527 11.8 No


et al.42

Towfighi et al.43 2013 United 1997–2006 Administrative 4,219 332,225 1.3 4,763 273,735 1.7 Yes
States

Continued
e15
or substantial between-study heterogeneity (I2 = 56.6% and

Studies using primarily administrative data were classified as retrospective designs; studies that used data from medical records or registries (i.e., single-hospital, multiple-hospital, and registry) were classified as prospective
Table 1 Characteristics of 19 included studies providing sex-specific utilization data among all patients with acute ischemic stroke (AIS) classified according to the type of data
I2 = 94.8%, respectively).

Adjusted
estimate
provided

For some studies, the number of treated cases and/or denominators were estimated from the reported treatment rates. Abbreviations: NA = data not available; rtPA = recombinant tissue plasminogen activator.
After combining data from the 16 studies included in the previous

Yes

Yes
No
meta-analysis,8 there were a total of 33 studies that provided data
treatment on the unadjusted OR among all patients with AIS. After cate-
gorizing these studies into 4 time periods according to the date of
rate, %

publication, the pooled results from more recently published


Male

2.7

NA

NA
studies tended to show a smaller sex difference (figure 5). Al-
though the linear test for trend for the summary ORs across the 4
time periods was not statistically significant (p = 0.15), when the
denominator, n

summary ORs for the earlier time periods were compared to the
provided: all AIS with unadjusted estimates (n = 17 studies), AIS with only an adjusted estimate provided (n = 2 studies) (continued)

2014–2018 period (OR, 0.90 [95% CI, 0.84–0.97]), the


2000–2004 estimate (OR, 0.73 [95% CI, 0.67–0.80]) was sig-
107,915

nificantly different (Z test p < 0.001).


2,570

5,862
Male

Finally, among the 17 studies that provided unadjusted data


Male rtPA-

on all patients with AIS, when studies with a quality assessment


cases, n
treated

score of 5 or lower (n = 9) were compared to studies with a score


2,940

of 6 or higher (n = 8), the summary unadjusted ORs were the same


NA

NA

(0.88 [95% CI, 0.81–0.95] and 0.88 [95% CI, 0.78–0.98], re-
spectively) (figure e-5, doi.org/10.5061/dryad.1zcrjdfnc). Both
treatment

subgroups had significant between-study heterogeneity. With


Female

rate, %

respect to the primary study focus, studies classified under


2.7

NA

NA

group A (i.e., specific objective to look at sex differences in IV


rtPA use) found a slightly greater sex difference (summary
unadjusted OR, 0.80 [95% CI, 0.69–0.92]) than studies
Downloaded from https://www.neurology.org by King Saud Bin Abdulaziz University for Health on 11 March 2024

denominator, n

classified as either B or C (summary unadjusted OR, 0.87


[95% CI, 0.78–0.96]) or group D (summary unadjusted OR,
0.96 [95% CI, 0.91–1.02]) (figure e-6, doi.org/10.5061/
Female

91,323

2,707

5,768

dryad.1zcrjdfnc). Only the pairwise difference between group


A and group D was statistically significant (Z test p = 0.02).
Female rtPA-

Discussion
cases, n
treated

2,444

NA

NA

This updated meta-analysis found 24 studies published since


2008 that provided data to address the issue of sex differences
a

Administrative

Administrative
Study design

in IV rtPA treatment among patients with AIS. Although the


results of individual studies varied substantially, the pooled
Multiple-
hospital

unadjusted random effects OR among all patients with AIS,


which was based on data from 17 studies and over 1 million
patients with stroke, indicated that women were less likely to
2011–2014

2000–2012

2009–2013

receive IV thrombolysis treatment compared to men. It is im-


Time

portant to note that the 13% lower relative odds of treatment


observed in these 17 studies reflects modest absolute differences
in treatment rates between women and men; most of the dif-
Country

Thailand

United

United
States

States

ferences were in the range of 0.5%–1.0%, with the largest being


8.4%.17 However, even small absolute differences could translate
into a large number of untreated women, given the high in-
2018

2016

2016
Year

cidence of AIS. Moreover, missed opportunities to deliver IV


rtPA could have greater consequences in women because they
estimate only (n =

have poorer stroke outcomes than men18 but derive at least


Domino et al.33
All AIS: adjusted
Vongmongkol

Zachrison

equivalent treatment benefit.1


et al.44

et al.37
Sauser-
Authors

designs.

Why would a sex difference in IV thrombolysis treatment rates


persist even if the absolute differences are small? One
2)

e16 Neurology | Volume 95, Number 1 | July 7, 2020 Neurology.org/N


Copyright © 2020 American Academy of Neurology. Unauthorized reproduction of this article is prohibited.
Neurology.org/N

Table 2 Characteristics of 7 included studies providing sex-specific utilization data among eligible treatment subgroups
Female rtPA-treated Female Female treatment Male rtPA-treated Male Male treatment
Copyright © 2020 American Academy of Neurology. Unauthorized reproduction of this article is prohibited.

Author Year Country Time Study design cases, n denominator, n rate, % cases, n denominator, n rate, %

Asdaghi 2016 United 2010–2014 Registry NA NA 87.8 NA NA 88.1


et al.13,a States

de Ridder 2013 Holland 2003–2005 Multiple- 314 755 41.6 382 902 42.4
et al.9,b hospital

Fredwall 2016 United 2010–2013 Single- 119 354 33.6 122 309 39.5
et al.20,c States hospital
Downloaded from https://www.neurology.org by King Saud Bin Abdulaziz University for Health on 11 March 2024

McDermott 2017 United 2009–2012 Multiple- 35 55 63.6 21 35 60.0


et al.35,d States hospital

Messe et al.14,a 2016 United 2003–2011 Registry 23,254 31,219 74.5 23,162 30,479 76.0
States

Rudd et al.21,e 2011 United 2008 Administrative 61 668 9.1 99 937 10.6
Kingdom

Tafreshi 2010 United 2001–2009 Multiple- 148 390 37.9 146 458 31.9
et al.36,f States hospital

Abbreviations: AIS = acute ischemic stroke; NA = data not available; rtPA = recombinant tissue plasminogen activator.
Two studies9,13 provided data on all patients with AIS and eligible subgroups.
a
Eligible patients with AIS (i.e., absence of contraindications or warnings) with arrival within 2 hours and treatment within 3 hours.
b
Patients with AIS with arrival within 4 hours.
c
Eligible patients with AIS (i.e., absence of contraindications or warnings) treated within 4.5 hours.
d
Patients with AIS with arrival within 2 hours and no contraindications.
e
Eligible patients with AIS (i.e., absence of contraindications or warnings) with age <80 years and arrival within 3 hours.
f
Patients with AIS with a “stroke code” and IV rtPA treatment decision in the emergency department.
Neurology | Volume 95, Number 1 | July 7, 2020
e17
Figure 2 Forest plot of unadjusted odds ratio (OR) of IV recombinant tissue plasminogen activator use in women compared
to men in all acute ischemic stroke admissions
Downloaded from https://www.neurology.org by King Saud Bin Abdulaziz University for Health on 11 March 2024

Random effects model (n = 17 studies). CI = confidence interval.

explanation is that there are differences in the characteristics age (i.e., >80 years), which had been traditionally regarded as
of women and men who present with acute stroke, which for a reason to exclude patients from IV thrombolysis therapy, was
women includes older age, different comorbidity profiles, found to be an explanation for why more women were excluded
higher stroke severity, and a greater chance of being widowed from IV rtPA therapy in 2 of the studies included in this
or living alone, affecting their eligibility for IV rtPA treatment review.20,21 Although there is no strong evidence that women are
and leading to an apparent sex difference in the utilization of more likely to arrive later to the hospital than men following the
the treatment. One approach to test this hypothesis is to onset of stroke symptoms,1 one study included in this review
examine the studies that provide data on eligible treatment found that the later arrival of women to the hospital did explain
subgroups. Although the definition of what constituted an their lower use of IV thrombolysis.9 Among the older pop-
eligible treatment subgroup varied across the 7 studies in- ulation, women are much more likely than men to live alone and
cluded in this meta-analysis, in aggregate, they found only this has been identified as a risk factor for late arrival and failure
a very modest and nonsignificant sex difference (summary to receive IV thrombolysis treatment.22,23
unadjusted OR, 0.95). This estimate was smaller than the
nonsignificant difference reported in the original meta- Another approach to explain sex differences in IV thrombol-
analysis (summary unadjusted OR, 0.81), which was based ysis treatment is to examine the studies that provide adjusted
on only 4 studies.8 estimates among the broad population of all patients with AIS.
Overall, statistical adjustment did attenuate the overall sum-
A detailed accounting of the specific factors that cause the mary OR from 0.87 unadjusted (in 17 studies) to 0.95 ad-
exclusion of women and men from the pool of patients eligible justed (in 10 studies), which suggests that factors such as age
for IV rtPA treatment—most notably contraindications and and severity help explain the origins of some of the observed
warnings for IV rtPA and time of arrival—have been the focus disparity. However, when the effect of adjustment was ex-
of several studies. For example, data from a population-based amined only among the 8 studies that provided both un-
stroke incidence study found no sex differences in eligibility adjusted and adjusted estimates, adjustment had a very
for IV rtPA treatment, except for severe hypertension on modest attenuating effect on the summary ORs (0.88 un-
presentation, which was more common in women.19 Older adjusted vs 0.91 adjusted). Interpretation of the overall effects

e18 Neurology | Volume 95, Number 1 | July 7, 2020 Neurology.org/N


Copyright © 2020 American Academy of Neurology. Unauthorized reproduction of this article is prohibited.
Figure 3 Forest plot of the adjusted odds ratio of IV recombinant tissue plasminogen activator use in women compared to
men in all acute ischemic stroke admissions.

Random effects model (n = 10 studies). CI = confidence interval; OR = odds ratio.


Downloaded from https://www.neurology.org by King Saud Bin Abdulaziz University for Health on 11 March 2024

of statistical adjustment is also complicated by the fact that household income (n = 1) (table e-3, doi.org/10.5061/
many studies did not adjust for socioeconomic factors that dryad.1zcrjdfnc).
are often different in women with stroke compared to men;
for example, few studies adjusted for health insurance status Another possible explanation for differences in IV throm-
(n = 2), education level (n = 2), living alone (n = 2), or bolysis treatment rates is that the clinical presentation of

Figure 4 Forest plot of the unadjusted odds ratio of IV recombinant tissue plasminogen activator use in women compared
to men among eligible treatment subgroup of patients with acute ischemic stroke

Random effects model (n = 7 studies). See table footnotes for study-specific definitions. CI = confidence interval; OR = odds ratio.

Neurology.org/N Neurology | Volume 95, Number 1 | July 7, 2020 e19


Copyright © 2020 American Academy of Neurology. Unauthorized reproduction of this article is prohibited.
Figure 5 Subgroup analysis by publication date (2000–2004, 2005–2008, 2009–2013, 2014–2018)
Downloaded from https://www.neurology.org by King Saud Bin Abdulaziz University for Health on 11 March 2024

Forest plots of the unadjusted odds ratio (OR) of IV recombinant tissue plasminogen activator use in women compared to men in all acute ischemic stroke
admissions. Random effects model (n = 33 studies). CI = confidence interval.

women with acute stroke is different enough from that of men studies included in our meta-analysis, most did not include
to affect treatment decision-making. An unusual presentation details of presenting symptoms in their patient populations.
of female patients could delay treatment decisions by com- However, one study did find that women were significantly
plicating the confirmation of the diagnosis of AIS or its time of more likely to present with altered mental status than men.25
onset. A recent systematic review of 43 studies that examined The higher prevalence of stroke mimics, especially in younger
sex differences in the clinical presentation of acute stroke women, can also complicate the timely diagnosis of acute
found a tendency for women to present more often with stroke, and result in errors in determining eligibility for IV
nontraditional stroke symptoms (including altered mental rtPA treatment.26,27
status, generalized weakness, and urinary incontinence) and
concluded that these could contribute to delayed recognition Sex differences in patient preference for IV thrombolysis ad-
and treatment.24 Another recent review reached similar con- ministration is another possible explanation for lower utili-
clusions regarding the higher prevalence of nontraditional zation rates in women. A Canadian study of ambulatory
presenting symptoms in women with stroke.1 Among the patients that used a series of case scenarios determined that

e20 Neurology | Volume 95, Number 1 | July 7, 2020 Neurology.org/N


Copyright © 2020 American Academy of Neurology. Unauthorized reproduction of this article is prohibited.
women were less likely to accept thrombolysis treatment and The preponderance of evidence from studies published in
also tended to be less confident and more risk averse in their the last 10 years that compare IV thrombolysis treatment
decisions compared to men.28 However, a single-center study rates in women and men with acute stroke points to lower
from Texas that compared eligible patients who had refused treatment rates in women. However, the absolute differences
treatment with eligible patients who had consented to treat- in treatment rates between women and men were small, and
ment did not find a sex difference in the acceptance of IV when compared to articles published more than 10 years ago
thrombolysis therapy.29 the magnitude of the sex difference is smaller on a relative
basis. This treatment disparity appears to be partially
Another plausible explanation for the sex differences observed explained by sex differences in eligibility for treatment. The
in these studies is residual confounding—perhaps there continued presence of this sex difference calls for further
remains a variable that is either unmeasured or poorly mea- studies designed to explain its origins, as well as ongoing
sured that would explain this sex-based disparity. Studies with surveillance to monitor the magnitude and changes over
more detailed data on the specific particulars of the clinical time. This is especially important as we move towards
presentation of acute stroke cases are needed to help sort out wider implementation of intra-arterial thrombectomy–based
the complicated matrix of issues that might explain differences treatments.38
in eligibility and the use of IV rtPA therapy in women
and men. Study funding
No targeted funding reported.
There are some limitations to our systematic review and meta-
analysis to consider. First, almost all of the meta-analyses we Disclosure
conducted found substantial between-study variability, and The authors report no relevant disclosures. Go to Neurology.
some individual studies documented higher unadjusted or org/N for full disclosures.
adjusted treatment rates in women compared to men,30–36
although in only one was the difference statistically signifi- Publication history
cant.37 This heterogeneity complicates the interpretation of Received by Neurology May 16, 2019. Accepted in final form
the summary (pooled) effect estimates, and emphasizes that December 5, 2019.
the presence and magnitude of any observed sex difference is
Downloaded from https://www.neurology.org by King Saud Bin Abdulaziz University for Health on 11 March 2024

ultimately study- and population-specific. However, all of the


summary OR estimates we generated were less than 1.0, in- Appendix Authors
dicating that overall treatment rates in women were always
Name Location Contribution
lower than those in men, which justifies our conclusion that
the preponderance of evidence still points to lower treatment Brent Michigan State Database searches, protocol
Strong University, East development, abstraction of data,
rates in women. Second, while subgroup analyses can help Lansing statistical analyses, manuscript
identify the source of between-study heterogeneity, almost all preparation
of the prespecified subgroup analyses we conducted still Lynda University of Review of protocol and draft
showed evidence of significant between-study heterogeneity. Lisabeth, Michigan, Ann manuscript
PhD Arbor
Meta-regression can be used to identify factors that contribute
to between-study heterogeneity,11 but it is best used when Mathew Michigan State Design of the study, protocol
there are a large number of studies with a consistent set of Reeves, University, East development, abstraction of data,
PhD Lansing review of statistical analyses,
study-level factors that are plausibly related to sex-specific manuscript preparation
treatment patterns. Such data were lacking among the rele-
vant studies identified in this review. Third, only 5 studies
were conducted with the specific objective of examining sex References
differences in IV thrombolysis treatment. Although we found 1. Bushnell C, Howard VJ, Lisabeth L, et al. Sex differences in the evaluation and
that these studies had a lower summary OR than the others, treatment of acute ischaemic stroke. Lancet Neurol 2018;17:641–650.
2. Phan HT, Blizzard CL, Reeves MJ, et al. Sex differences in long-term mortality after
this difference was not statistically significant. Fourth, al- stroke in the INSTRUCT (International Stroke Outcomes Study): a meta-analysis of
though some studies differentially excluded men or women individual participant data. Circ Cardiovasc Qual Outcomes 2017;10:e003436.
3. Kaplovitch E, Anand SS. Stroke in women: recognizing opportunities for prevention
who did or did not receive IV thrombolysis, the net effect of and treatment. Stroke 2018;49:515–517.
this is small. The effect of patient exclusions was evaluated in 4. Falsetti L, Viticchi G, Buratti L, Balucani C, Marra AM, Silvestrini M. From head to
toe: sex and gender differences in the treatment of ischemic cerebral disease. Phar-
the quality assessment process (criterion 2 [exclusions]; table macol Res 2017;121:240–250.
e-2, doi.org/10.5061/dryad.1zcrjdfnc), where we found that 5. Girijala RL, Sohrabji F, Bush RL. Sex differences in stroke: review of current
knowledge and evidence. Vasc Med 2017;22:135–145.
the majority of studies (75%) made either no or only minor 6. Faigle R, Urrutia VC, Cooper LA, Gottesman RF. Individual and system contributions
subject-level exclusions. Finally, search methods as well as the to race and sex disparities in thrombolysis use for stroke patients in the United States.
Stroke 2017;48:990–997.
screening of titles and abstracts to find studies that provide 7. Reeves MJ, Bushnell CD, Howard G, et al. Sex differences in stroke: epidemiology,
sex-specific data on IV rtPA treatment rates are inexact, and so clinical presentation, medical care, and outcomes. Lancet Neurol 2008;7:915–926.
8. Reeves M, Bhatt A, Jajou P, Brown M, Lisabeth L. Sex differences in the use of
it is likely that there remain unidentified studies that report intravenous rtPA thrombolysis treatment for acute ischemic stroke: a meta-analysis.
relevant data on the topic. Stroke 2009;40:1743–1749.

Neurology.org/N Neurology | Volume 95, Number 1 | July 7, 2020 e21


Copyright © 2020 American Academy of Neurology. Unauthorized reproduction of this article is prohibited.
9. de Ridder I, Dirks M, Niessen L, Dippel D. Unequal access to treatment with in- 27. Merino JG, Luby M, Benson RT, et al. Predictors of acute stroke mimics in 8187
travenous alteplase for women with acute ischemic stroke. Stroke 2013;44: patients referred to a stroke service. J Stroke Cerebrovasc Dis 2013;22:e397–e403.
2610–2612. 28. Kapral MK, Devon J, Winter AL, Wang J, Peters A, Bondy SJ. Gender differences in
10. Clua-Espuny JL, Ripolles-Vicente R, Forcadell-Arenas T, et al. Sex differences in long- stroke care decision-making. Med Care 2006;44:70–80.
term survival after a first stroke with intravenous thrombolysis: ebrictus study. Cer- 29. Vahidy FS, Rahbar MH, Lal AP, Grotta JC, Savitz SI. Patient refusal of thrombolytic
ebrovasc Dis Extra 2015;5:95–102. therapy for suspected acute ischemic stroke. Int J Stroke 2015;10:882–886.
11. Borenstein M, Hedges LV, Higgins JPT, Rothstein HR. Introduction to Meta-Anal- 30. Boehme AK, Siegler JE, Mullen MT, et al. Racial and gender differences in stroke
ysis. Chichester: John Wiley and Sons, Ltd.; 2009. severity, outcomes, and treatment in patients with acute ischemic stroke. J Stroke
12. Moher D, Liberati A, Tetzlaff J, Altman DG, The PG. Preferred reporting items for Cerebrovasc Dis 2014;23:e255–e261.
systematic reviews and meta-analyses: the PRISMA statement. PLoS Med 2009;6: 31. Kunisawa S, Kobayashi D, Lee J, et al. Factors associated with the administration of
e1000097. tissue plasminogen activator for acute ischemic stroke. J Stroke Cerebrovasc Dis 2014;
13. Asdaghi N, Romano JG, Wang K, et al. Sex disparities in ischemic stroke care: FL-PR 23:724–731.
CReSD study (Florida-Puerto Rico Collaboration to Reduce Stroke Disparities). 32. Allen NB, Myers D, Watanabe E, et al. Utilization of intravenous tissue plasminogen
Stroke 2016;47:2618–2626. activator for ischemic stroke: are there sex differences? Cerebrovasc Dis 2009;27:
14. Messe SR, Khatri P, Reeves MJ, et al. Why are acute ischemic stroke patients not 254–258.
receiving IV tPA? Results from a national registry. Neurology 2016;87:1565–1574. 33. Domino JS, Baek J, Meurer WJ, et al. Emerging temporal trends in tissue plasminogen
15. Wells G, Shea B, O’Connell D, et al. The Newcastle-Ottawa Scale (NOS) for assessing activator use: results from the BASIC project. Neurology 2016;87:2184–2191.
the quality of nonrandomised studies in meta-analyses [online]. Available at: ohri.ca/ 34. Huang Y, Wang JG, Wei JW, et al. Age and gender variations in the management of
programs/clinical_epidemiology/oxford.asp. Accessed October 29, 2019. ischaemic stroke in China. Int J Stroke 2010;5:351–359.
16. Woodward M. Rationale and tutorial for analysing and reporting sex differences in 35. McDermott M, Lisabeth LD, Baek J, et al. Sex disparity in stroke quality of care in
cardiovascular associations. Heart 2019;105:1701–1708. a community-based study. J Stroke Cerebrovasc Dis 2017;26:1781–1786.
17. Lee SJ, Heo SH, Ambrosius WT, Bushnell CD. Factors mediating outcome after
36. Tafreshi GM, Raman R, Ernstrom K, Meyer BC, Hemmen TM. Gender differences in
stroke: gender, thrombolysis, and their interaction. Transl Stroke Res 2018;9:
acute stroke treatment: the University of California San Diego experience. Stroke
267–273.
2010;41:1755–1757.
18. Phan HT, Blizzard CL, Reeves MJ, et al. Factors contributing to sex differences in
37. Sauser-Zachrison K, Shen E, Ajani Z, et al. Emergency care of patients with acute
functional outcomes and participation after stroke. Neurology 2018;90:e1945–e1953.
ischemic stroke in the Kaiser Permanente Southern California integrated health
19. Madsen TE, Khoury JC, Alwell KA, et al. Analysis of tissue plasminogen activator
system. Perm J 2016;20:10–13.
eligibility by sex in the Greater Cincinnati/Northern Kentucky stroke study. Stroke
38. Goyal M, Menon BK, van Zwam WH, et al. Endovascular thrombectomy after large-
2015;46:717–721.
vessel ischaemic stroke: a meta-analysis of individual patient data from five rando-
20. Fredwall M, Sternberg S, Blackhurst D, Lee A, Leacock R, Nathaniel TI. Gender
differences in exclusion criteria for recombinant tissue-type plasminogen activator. mised trials. Lancet 2016;387:1723–1731.
J Stroke Cerebrovasc Dis 2016;25:2569–2574. 39. Łabuz-Roszak B, Starostka-Tatar A, Lasek-Bal A, Gierlotka M, Ga̧sior M, Skrzypek M.
21. Rudd AG, Hoffman A, Grant R, Campbell JT, Lowe D. Stroke thrombolysis in Diagnostics, treatment and secondary prevention of ischemic stroke in the Silesian
England, Wales and Northern Ireland: how much do we do and how much do we Province, Poland between 2009 and 2015. Neurol Neurochir Pol 2018;52:235–242.
need? J Neurol Neurosurg Psychiatry 2011;82:14–19. 40. Nagaraja N, Bhattacharya P, Mada F, et al. Gender based differences in acute stroke
22. Eriksson M, Jonsson F, Appelros P, et al. Dissemination of thrombolysis for acute care in Michigan hospitals. J Neurol Sci 2012;314:88–91.
ischemic stroke across a nation: experiences from the Swedish stroke register, 2003 to 41. Nardetto L, Giometto B, Moretto G, Mantoan D, Saia M. Hub-and-spoke stroke
2008. Stroke 2010;41:1115–1122. network in the Veneto Region: a retrospective study investigating the effectiveness of
23. Reeves MJ, Prager M, Fang J, Stamplecoski M, Kapral MK. Impact of living alone on the stroke pathway and trends over time. Neurol Sci 2017;38:2117–2121.
Downloaded from https://www.neurology.org by King Saud Bin Abdulaziz University for Health on 11 March 2024

the care and outcomes of patients with acute stroke. Stroke 2014;45:3083–3085. 42. Santalucia P, Pezzella FR, Sessa M, et al. Sex differences in clinical presentation,
24. Berglund A, Schenck-Gustafsson K, von Euler M. Sex differences in the presentation severity and outcome of stroke: results from a hospital-based registry. Eur J Intern
of stroke. Maturitas 2017;99:47–50. Med 2013;24:167–171.
25. Park SJ, Shin SD, Ro YS, Song KJ, Oh J. Gender differences in emergency stroke care 43. Towfighi A, Markovic D, Ovbiagele B. Sex differences in revascularization inter-
and hospital outcome in acute ischemic stroke: a multicenter observational study. Am ventions after acute ischemic stroke. J Stroke Cerebrovasc Dis 2013;22:e347–e353.
J Emerg Med 2013;31:178–184. 44. Vongmongkol V, Tangcharoensthien V, Greetong T, McNeil E, Chongsuvivatwong
26. Lewandowski C, Mays-Wilson K, Miller J, et al. Safety and outcomes in stroke mimics V. Trend in recombinant tissue plasminogen activator (rtPA) use for ischemic stroke
after intravenous tissue plasminogen activator administration: a single-center expe- in Thailand: geographic inequality, cost of treatment and impact on 30-day case
rience. J Stroke Cerebrovasc Dis 2015;24:48–52. fatality rate. J Med Assoc Thai 2018;101:875–881.

e22 Neurology | Volume 95, Number 1 | July 7, 2020 Neurology.org/N


Copyright © 2020 American Academy of Neurology. Unauthorized reproduction of this article is prohibited.

You might also like