Stroke, Physical Function, and Death Over a 15-Year Period in Older Australian Women
Background and Purpose—As populations age, an increasing number of older women are living with stroke. This study looks at long-term outcomes for women with stroke, comparing mortality rates for women with poor physical function (PF) and those with higher levels of function. The purpose is to understand not only how long women might live after a stroke, but also how long they live with physical disability.
Methods—The study uses 15 years of data on women from the Australian Longitudinal Study on Women’s Health 1921 to 1926 cohort. The risk of stroke and the risk of stroke and poor PF were estimated using Cox proportional hazard model. Among women who reported a stroke during the study period, mortality risk was compared according to their physical functioning level after that stroke.
Results—Almost half of the women who had a stroke and poor PF survived past 10 years. The 10-year mortality rate was 37% for women with stroke and adequate PF and 51% for women with stroke and poor PF at the time of the stroke (hazard rate ratio, 1.52; 95% CI, 1.18–1.95; P=0.0015 adjusting for demographic and health covariates).
Conclusions—This study provides evidence of the long-term outcomes of stroke among older women, with women living for many years with poor PF. This outcome has important implications for the women’s quality of life during their later years and in understanding the burden of disability associated with stroke.
Global incidence of stroke reported in 2010 was almost 17 million, with a prevalence of 33 million and nearly 7 million deaths attributed to stroke.1 As populations age, the rate of stroke and its associated morbidity and mortality will increase, especially if risk factors such as tobacco smoking and hypertension are not controlled.2,3 Because women live longer than men, stroke is a particular risk for older women, with 60% of all strokes occurring in women.4 Women also tend to be older when they experience stroke and are more likely to die or experience physical disability from stroke than men.4
Although many studies have examined short-term survival after stroke, fewer long-term studies have been performed. In Denmark, a 5-year study of people with first stroke found that 28% died within 28 days, 41% within 1 year, and 60% within 5 years5; a 10-year study found a mortality rate of 81% and an association between stroke severity and survival.6 In a 12-year study, Byles et al7 found that two third of older Australian women with no previous history of stroke were still alive compared with only half with a history of stroke. A 21-year New Zealand study found that the mortality rate in those with stroke was double that of the general population. These studies indicate the long-term mortality risk associated with stroke.
Women who survive stroke are likely to experience poor physical function (PF),4 and many live with these limitations for many years. Few studies have examined PF and long-term survival among older women with stroke. This study investigates long-term outcomes for women with stroke, comparing mortality rates for those with poor and adequate PF, to understand how long women might live with stroke and with physical disability.
Data are from the 1921 to 1926 cohort of the Australian Longitudinal Study on Women’s Health (ALSWH; http://www.alswh.org.au/).8 Participants were randomly selected from the national health insurance database, with oversampling of women living in rural or remote areas. The study recruited and surveyed 12 432 women aged from 70 to 75 years in 1996 (born in 1921–1926). These women were followed up every 3 years between 1999 and 2011. Ethical approval was obtained from the Universities of Newcastle and Queensland.
Stroke status was based on the yes/no answers to the following questions: “Have you ever been told by a doctor that you have: Stroke” (Survey 1, 1996); “In the last 3 years have you been told by a doctor that you have: Stroke” (Survey 2, 1999); “In the past 3 years, have you been diagnosed with or treated for: Stroke” (Survey 3–6, 2002–2011). History of stroke was only classified as “yes” for participants who reported stroke at Survey 1 (baseline).
PF was measured using the Medical Outcome Short form 36 questionnaire (SF36).9 The PF subscale is a valid assessment of long-term physical health in stroke survivors.10 Questions ask respondents to assess how health conditions limit their ability to participate in activities such as vigorous activities, lifting or carrying, and bathing and dressing as “limited a lot” (score 1), “limited a little” (2), or “not limited at all” (3). Final scores range from 0 to 100, with a higher score indicating a higher level of PF. In this analysis, SF36 PF scores <40 were classified as poor PF, and scores ≥40 were classified as adequate PF. This cut point demonstrated good discrimination for various daily activities, and once a woman scored <40, they were unlikely to have substantial improvement in their PF score on later surveys, indicating probable permanent physical impairment.
Demographic covariates included in the analyses included age at the baseline survey; area of residence (urban versus nonurban) using the Accessibility/Remoteness Index of Australia11; level of education achieved (no qualification; school qualification; trade, certificate or diploma; and university degree or higher), and partner status (partnered versus not partnered). The Duke Social Support Index was measured at Surveys 1 and 2 only. Health covariates included body mass index as classified by the World Health Organization,12 smoking status (never smoked; ex-smoker, or current smoker), and self-reported diagnosis of diabetes mellitus, heart disease, hypertension, low iron, and cancer. Date and cause of death were obtained from the National Death Index (http://www.aihw.gov.au/national-death-index/). Depression and comorbidities, such as renal disease, cancer, and dementia, although included in the initial modeling, did not yield significance and were, therefore, excluded from further analysis.
The analyses excluded people who responded to Survey 1 only. The PF cutoff was examined in terms of the proportion of women who answered limited a lot to each of the 10 SF36 PF items and evaluated over several surveys. The incidence of reporting new stroke (reported surveys 2–6) according to prior stroke status (past stroke reported at Survey 1) was evaluated, as well as the incidence of reporting new stroke with concomitant poor PF recorded at the same survey as the stroke (new stroke with poor PF). Cox proportional hazards models were used to compare the risk of new stroke with poor PF and prior stroke status, adjusting for demographic factors and health covariates at Survey 1. Finally, the cumulative incidence of mortality after the occurrence of new stroke was examined according to PF status classified on the same survey as the new stroke was reported. Mortality risk was also assessed using proportional hazards modeling, accounting for prior stroke, PF status, demographic characteristics, and health covariates at the time of the index stroke during the study period (the new stroke). Death data were censored at March 15, 2012, allowing for a maximum follow-up time of 14.4 years. All death data were collected by linking survey data to the Australian National Death Index data. These records are obtained from the Registrars of Births, Deaths and Marriages. All analyses were conducted using SAS 9.4.13
Of the 12 432 women in the ALSWH 1921 to 1926 cohort, 1671 (13.4%) responded to Survey 1 only and were, therefore, excluded from analyses, leaving 10 761 women. Of these, 7766 (72.2%) were alive at Survey 6. At Survey 1, 518 (4.8%) women reported prestudy history of stroke and 1745 (16.2%) were classified as having poor PF. Table 1 shows the extent to which women with poor PF were limited on the SF36 PF activities, with most women limited a lot on most activities. Moreover, 82.8% of women who had poor PF at Survey 1 still had poor PF at Survey 6, indicating little recovery using the <40 cut point.
The cumulative incidence of respondents (Figure I in the online-only Data Supplement) reporting new stroke during the study period indicated that the risk of new stroke during the study period was 5.22× (95% CI, 4.42–6.16) higher in those with prior stroke than those with no prior stroke (P<0.0001). Figure 1 shows that women with a prior history of stroke were also more likely to have a stroke with concomitant poor PF (P<0.0001).
When compared with women with no history of stroke and after adjusting for demographic covariates (Table 2), women with a prior history of stroke were 6× (95% CI, 4.79–7.51) more likely to experience a recurrent stroke with poor PF (P<0.0001). After adjusting for demographic and health covariates, this association attenuated to 4.60× the risk but remained significant (95% CI, 3.64–5.97; P<0.0001).
For women who reported a new stroke during the study period (Survey 2–6), those who also had poor PF were more likely to be not partnered, overweight, or obese and diagnosed with diabetes mellitus, heart disease, and hypertension (Table 3).
Ten-year mortality rate was 51% for women with poor PF and 37% for women with adequate PF. Over 14 years (end of follow-up), mortality rate was 73% and 58.4%, respectively. Median survival time was 10 years for poor PF and 12.3 years for adequate PF. Among women who reported a stroke during the study period, those with poor PF had 67% increased mortality risk (P<0.0001) compared with those with adequate PF. The proportion of deaths attributed to cerebrovascular diseases was similar for women with stroke and poor PF (34.7%) and women with stroke and adequate PF (34.4%). In contrast, among women with no report of stroke, 15.4% died with cerebrovascular disease as the cause of death.
After adjusting for demographic and health covariates, PF remained significantly associated with mortality risk (hazard ratio [HR], 1.52; 95% CI, 1.18–1.95) among women with new stroke (P=0.0015), but a prior history of stroke was not a significant factor for mortality rates in all models (P>0.05; Figure 2). The hazard rate of mortality increased with age (HR, 1.22) and was higher among women who were underweight (HR, 1.76) or diagnosed with diabetes mellitus (HR, 1.61) or heart disease (HR, 1.35; Table 4).
This study shows the incidence of self-reported stroke among a large cohort of Australian women, the high incidence of recurrent stroke among those who survived a past stroke, and the risk of stroke with concomitant poor PF. Moreover, the study shows that many women who experience stroke survive many years with poor PF. This finding is important in estimating the burden of stroke across women’s later life and in planning for their health and aged care needs.
Stroke and Poor Physical Function in Older Women
Among women with stroke, poor PF was associated with diabetes mellitus, heart disease, and hypertension. This indicates that poor PF may not be due exclusively to stroke, but may also be attributed to or exacerbated by other conditions. Management of comorbidity among women with prior stroke is obviously an important component of management of current disability as well as future stroke risk. This raises questions about how to improve the PF in women who are likely to survive many years with significant disability, such as progressive resistance strength training14 or 20 to 30 minutes of moderate activity most days of the week.15 However, the current study found that few women recovered from the low levels of PF reported at the time of experiencing stroke. Is there insufficient opportunity, incentive, and support for older women living with stroke to maintain adequate levels of PF? In Australia, there is currently no national strategy to maintain or improve PF in older stroke survivors and no evidence that consideration has been given to the needs of older women living with stroke for >15 years or to potential benefit of maintaining adequate PF in older women with stroke.
Stroke, Poor Physical Function, and Mortality Risk in Older Women
Although many women lived for over 10 years with poor PF, mortality risk was significantly higher among women with poor PF compared with women with adequate PF. This may reflect stroke severity as well as the effect of other risk factors and comorbidities. Other studies found higher mortality rates in those with stroke and diabetes mellitus16 and heart disease17 and that being underweight can be hazardous in older women with stroke.18 Older women with stroke and body mass index <18.5 may benefit from referral to dietitians.
Stroke, Physical Function, and Long-Term Survival in Older Women
Three years after reporting a new stroke, 97% of women with adequate PF and 89% of those with poor PF survived. Ten years after a new stroke, 63% of women with adequate PF were still alive compared with only 49% with poor PF. After 15 years, this had fallen to 41.6% in those with adequate PF and 27% in those with poor PF, demonstrating older women can survive many years after stroke, even with poor PF. This challenges assumptions that they are probably not going to live long to resource health services designed to their needs.
Strengths, Limitations, and Clinical Implications
This study’s strength is the large sample size, 15-year follow-up and accuracy of death date and cause, and its implications for clinical practice and health promotion targeting older women who have survived stroke for many years.
The study’s limitation is its reliance on self-reported data and that women with stroke who died before they reported stroke or who were too frail to continue could not be included. This would lead to over-representation of women who experienced milder strokes and better PF and recovery, biasing results toward survival and the null association. However, even with this bias, the findings demonstrate an effect on risk of stroke and death, so true association may be stronger. Also, some women who reported stroke may have only experienced stroke-like symptoms; however, one third of deaths among women who reported stroke were because of cerebrovascular disease and when stroke occurred was not confirmed using administrative data, but estimated using a survey’s return date.
Further research is required into how maintaining and sustaining levels of PF can be efficiently achieved in programs that target older women who survive stroke and which effectively meet the needs of older women who are surviving stroke for many years.
This study investigated mortality in older Australian women who reported stroke with or without poor PF. Half of them were still alive 10 years later. This evidence challenges current clinical practice relating to older women living with stroke, particularly those with poor PF, and underscores the need for long-term programs for these women.
The research on which this article is based was conducted as part of the Universities of Newcastle and Queensland. We are grateful to the Australian Government Department of Health for funding and to the women who provided the survey data. We acknowledge the assistance of the Data Linkage Unit at the Australian Institute of Health and Welfare for undertaking the linkage to the National Death Index. This research was supported by infrastructure and staff of the Research Centre for Gender, Health and Ageing and Hunter Medical Research Institute.
Sources of Funding
The Australian Longitudinal Study on Women’s Health study is funded by the Australian Government Department of Health.
The online-only Data Supplement is available with this article at http://stroke.ahajournals.org/lookup/suppl/doi:10.1161/STROKEAHA.115.011456/-/DC1.
- Received September 22, 2015.
- Revision received January 26, 2016.
- Accepted January 28, 2016.
- © 2016 American Heart Association, Inc.
- Feigin VL,
- Forouzanfar MH,
- Krishnamurthi R,
- Mensah GA,
- Connor M,
- Bennett DA,
- et al.,
- Global Burden of Diseases, Injuries, and Risk Factors Study 2010 (GBD 2010) and the GBD Stroke Experts Group
- Roquer J,
- Campello AR,
- Gomis M.
- Brønnum-Hansen H,
- Davidsen M,
- Thorvaldsen P,
- Danish MONICA Study Group
- Brown WJ,
- Bryson L,
- Byles JE,
- Dobson AJ,
- Lee C,
- Mishra G,
- et al
- Anderson C,
- Laubscher S,
- Burns R.
- 11.↵Department of Health and Aged Care. Measuring remoteness: accessibility/remoteness index of Australia (ARIA). Government of Australia Canberra, 2001.
- 12.↵World Health Organization. Physical Status: The Use and Interpretation of Anthropometry. Geneva, Switzerland: World Health Organization; 1995. http://www.world-heart-federation.org/cardiovascular-health/stroke/. Accessed January 8, 2015.
- 13.↵SAS Institute Inc. SAS® 9.4 Guide to Software Updates. Cary, NC; 2013.
- Brach JS,
- Simonsick EM,
- Kritchevsky S,
- Yaffe K,
- Newman AB,
- Health, Aging and Body Composition Study Research Group
- Icks A,
- Claessen H,
- Morbach S,
- Glaeske G,
- Hoffmann F.
- Okura Y,
- Urban LH,
- Mahoney DW,
- Jacobsen SJ,
- Rodeheffer RJ.