Plasma Angiopoietin-1 Is Lower After Ischemic Stroke and Associated With Major Disability But Not Stroke Incidence
Background and Purpose—Studies in rodent models suggest that upregulating angiopoietin-1 (Angpt1) improves stroke outcomes. The aims of this study were to assess the association of plasma Angpt1 with stroke occurrence and outcome.
Methods—Plasma Angpt1 was measured in 336 patients who had experienced a recent stroke and 321 healthy controls with no stroke history. Patients with stroke (n=285) were reassessed at 3 months and plasma Angpt1 concentration on admission compared between those with severe and minor disability as assessed by the modified Rankin scale. In a separate cohort of 4032 community-acquired older men prospectively followed for a minimum of 6 years, the association of plasma Angpt1 with stroke incidence was examined.
Results—Median plasma Angpt1 was 3-fold lower in patients who had experienced a recent stroke (6.42, interquartile range, 4.26–9.53 compared with 17.36; interquartile range, 14.01–22.46 ng/mL; P<0.001) and remained associated with stroke after adjustment for other risk factors. Plasma Angpt1 concentrations on admission were lower in patients who had severe disability or died at 3 months (median, 5.52; interquartile range, 3.81–8.75 ng/mL for modified Rankin scale 3–6; n=91) compared with those with minor disability (median, 7.04; interquartile range, 4.75–9.92 ng/mL for modified Rankin scale 0–2; n=194), P=0.012, and remained negatively associated with severe disability or death after adjusting for other risk factors. Plasma Angpt1 was not predictive of stroke incidence in community-dwelling older men.
Conclusions—Plasma Angpt1 concentrations are low after ischemic stroke particularly in patients with poor stroke outcomes at 3 months. Interventions effective at upregulating Angpt1 could potentially improve stroke outcomes.
One of the current limitations in stroke management is the lack of reliable biomarkers for diagnosis and prognosis.1 Angiopoietin-1 (Angpt1) was originally identified for its role in vascular development; however, more recently it has been demonstrated to play key roles in promoting angiogenesis and enhancing healthy endothelial function in adults.2,3 Studies in animal models suggest that Angpt1 protects against cerebral ischemia.3–7 A single nucleotide in the 3′ untranslated region of the gene encoding Angpt1, which is associated with higher circulating concentrations of Angpt1, has also been associated with reduced risk of developing stroke.8 We are aware of no previous studies that have assessed the association of circulating Angpt1 concentrations with stroke.
The aims of this study were (1) to assess the association of plasma Angpt1 concentration with ischemic stroke, (2) to examine the association of plasma Angpt1 concentration with stroke outcome, and (3) to assess the association of plasma Angpt1 concentration with stroke incidence.
To examine aim 1, we undertook a case–control study. For aim 2, we assessed stroke cases from aim 1 at 3 months. For aim 3, we examined subjects from the Health In Men Study (HIMS).
Subjects for Aims 1 and 2
We obtained stroke cases from a previously described study in which consenting patients admitted to any of 4 acute stroke units located in the principal referral hospitals in the Central Coast and Hunter regions of New South Wales, Australia, were recruited.9 For inclusion in the current study, cases had to have a diagnosis of ischemic stroke (defined by World Health Organization clinical criteria) classified by the Trial of Org 10172 in Acute Stroke Treatment system as large artery atherosclerotic, small vessel occlusion, or cardioembolic.10,11 Exclusion criteria included age <18 years, diagnosis of hemorrhagic stroke or transient ischemic attack, and inability to undergo baseline brain imaging. Computed tomographic or MRI brain scans were obtained in all cases to confirm an ischemic stroke; however, stroke volumes were not measured. Stroke mechanism was investigated using clinically appropriate tests including transoesophageal echocardiography and carotid duplex. In some patients, stroke severity was assessed using the National Institutes of Health Stroke Scale and later divided into minor (0–4), moderate (5–15), moderate to severe (16–20), and severe (21–42) stroke as previously described.9 Controls were obtained from the Hunter Community Study (HCS) that has previously been described in detail.12 HCS recruited community-dwelling men and women aged 55 to 85 years from Newcastle, New South Wales, based on listings in the electoral roll between 2004 and 2007. Controls were selected randomly who had no history of any type of stroke. Written informed consent was obtained from each participant as per the University of Newcastle and Hunter New England Area Health ethics committee requirements. Ethics approval was provided by the appropriate committees. At the time of recruitment, subjects were interviewed by research staff using standardized questionnaires.9,12 Prior history of ever smoking, hypertension, diabetes mellitus, myocardial infarction, and atrial fibrillation was noted.
Functional Outcomes in Stroke Cases for Aim 2
Telephone follow-up was conducted in stroke cases, that were recruited to address aim 1, after discharge from hospital at 3 months. Patients and relatives, who could be contacted, were interviewed by experienced clinical research nurses who were trained in conducting interviews by phone and with the administration of assessments by phone. Disability was assessed using the modified Rankin Scale (mRS) as previously described.9,13 In brief, the mRS denotes 7 outcomes: 0, no symptoms; 1, no significant disability despite symptoms; 2, slight disability; 3, moderate disability; 4, moderately severe disability; 5, severe disability; and 6, death.
Subjects for Aim 3
HIMS developed from a population-based randomized trial of screening for abdominal aortic aneurysm conducted in Perth, Western Australia.14,15 Of the 19 352 men invited for initial assessment, 12 203 attended baseline screening between 1996 and 1999 (wave 1). Between 2001 and 2004, surviving men were invited to reattend for a second assessment and have blood collected of which 4248 men accepted (wave 2). Each man was invited to complete a questionnaire assessing aspects of history and lifestyle relevant to cardiovascular disease at the time of both assessments, including smoking history; history of diagnosis of high blood pressure, angina, myocardial infarction, stroke, and diabetes mellitus; and history of treatment for high blood pressure, angina, and diabetes mellitus. This information was used to define prior history or treatment for hypertension, diabetes mellitus, coronary heart disease, and stroke. Smoking was defined as ever or never smoker. Waist and hip circumference were measured in accordance with guidelines of the International Society for the Advancement of Kinanthropometry.16 From the time of blood sampling, men were followed using the Western Australian Data Linkage System that provides electronic linkage to data from the death registry and hospital morbidity data system.17 The diagnosis of stroke from hospital morbidity data was defined according to International Classification of Diseases, Ninth Revision codes 362.3, 433.x1, 434. 01, 434.11, 434.91, 430, 431, and 436 or International Classification of Diseases, Tenth Revision codes H34.1, I60, I61, I63, and I64. Incident stroke, including either first-ever stroke for participants with no prior stroke or recurrent stroke for participants with history of previous stroke, was recorded by Western Australian Data Linkage System any time after blood collection (2001–2004) until December 31, 2010. The validity of data within the Western Australian Data Linkage System has been assessed previously and found to be comparable to adjudication of medical records.17 Written informed consent was obtained from each participant, and ethics approval was provided by the appropriate committees.
Blood Samples and Tests
Peripheral blood samples were collected from stroke cases, controls, and HIMS subjects. Plasma was prepared by centrifugation (15 minutes at 2100g) and stored at −80°C for later analysis. Samples were collected from stroke cases within 8 hours of admission to hospital. Plasma Angpt1 was measured with a commercial assay (R&D Systems, Minneapolis, MN) by an experienced scientist blinded to the case–control category and outcome data. The intra- and interassay coefficient of variation was between 4.2% and 11.7%. In HIMS subjects, serum creatinine, low-density lipoprotein, and high-density lipoprotein were measured using automated assays (Hitachi 917, Roche Diagnostics GmBH, Mannheim, Germany) as previously described.18 The interassay coefficient of variation for these assays was between 2.1% and 4.8%.
Quantitative data were presented as median and interquartile range and compared by Mann–Whitney U test. Nominal data were presented as number and percentages and compared by χ2 test. For aim 1, the association of plasma Angpt1 with stroke was assessed using multiple regression analysis adjusting for age, male sex, previous myocardial infarction, hypertension, diabetes mellitus, smoking, and atrial fibrillation. For these analyses, subjects were grouped on the basis of plasma Angpt1 quartiles defined as <9.685, 9.685 to 11.866, >11.866 to 14.998, and >14.998 ng/mL. Receiver operating characteristic curves were generated to assess the ability of plasma Angpt1 to predict stroke. For the latter analyses, 100 minus plasma Angpt1 values were used so as to assess the ability of low Angpt1 concentrations in predicting stroke. For aim 2, stroke outcome was compared between patients with severe disability or death (mRS, 3–6) and those with minor or no disability (mRS, 0–2) using χ2 and Mann–Whitney U tests. To assess the association of plasma Angpt1 (per 10 ng/mL) with severe stroke disability, logistic regression analysis was performed adjusting for age (per 9 years), atrial fibrillation, and stroke severity on admission (graded as minor, moderate, moderate to severe, or severe), which were noted to be associated with mRS score 3 to 6 on univariate analyses. For aim 3, the association of plasma Angpt1 (per 30 ng/mL and expressed as angpt1 quartiles, defined as ≤5.38, >5.38–9.38, >9.38–16.71, and >16.71 ng/mL) with stroke incidence was investigated using Cox proportional hazard analysis. Analyses were adjusted for age (per 3 years), diabetes mellitus, hypertension, coronary heart disease, ever smoking, past stroke, creatinine (per 30 μmol/L), low-density lipoprotein (per 0.85 mmol/L), high-density lipoprotein (per 0.35 mmol/L), and waist-to-hip ratio (per 0.35). Units of continuous numbers used in these analyses were based on the approximate SD of the numbers in the HIMS subjects. Stroke incidence was estimated using Kaplan–Meier analysis and compared between men with plasma Angpt1 in different quartiles using the log-rank test.
Plasma Angpt1 Is Lower in Subjects With a Recent Ischemic Stroke
Stroke cases were older and more likely to have a history of hypertension, diabetes mellitus, and atrial fibrillation (Table 1). Median Angpt1 concentrations were ≈3× lower in strokes cases (6.42 interquartile range 4.26–9.53 ng/mL in cases compared with 17.36 interquartile range 14.01–22.46 ng/mL in controls; P<0.001). In a multivariate analysis, Angpt1 level was associated with stroke after adjustment for other risk factors including age, sex, smoking, atrial fibrillation, and medical comorbidities (Table 2). Using subjects with plasma Angpt1 in the highest quartile of values as the reference group, those with Angpt1 in the lower 3 quartiles were much more likely to be stroke cases. Low Angpt1 concentrations were highly predictive of stroke (area under the receiver operating characteristic curve, 0.946; 95% confidence intervals, 0.928–0.963; P<0.001; Figure 1).
Relationship Between Plasma Angpt1, Stroke Classification, Stroke Severity, and Outcome
Plasma Angpt1 concentrations varied by Trial of Org 10172 in Acute Stroke Treatment classification, being lowest in subjects with small vessel occlusion (median, 5.48; interquartile range, 3.77–8.00 ng/mL; n=91), intermediate in those with large artery atherosclerosis (median, 6.44; interquartile range, 4.75–9.79 ng/mL; n=118), and highest in those with cardioembolic stroke (median, 7.13; interquartile range, 4.69–10.15 ng/mL; n=127), P=0.005. Data on stroke severity were recorded at the time of admission in 301 of the 336 (90%) patients. Plasma Angpt1 concentrations were not associated with stroke severity on admission. Median (interquartile range) plasma concentrations of Angpt1 were 6.53 (4.46–9.74; n=158), 6.88 (4.33–9.84; n=110), 6.52 (4.00–9.23; n=18), and 6.06 (3.91–8.20; n=15) ng/mL in patients with minor, moderate, moderate to severe, and severe strokes, respectively, P=0.833. Stroke outcome was available on 285 of the 336 (85%) patients at 3 months and was classified according to the mRS. Ninety-one patients had severe disability (n=71) or were dead (n=20), whereas 194 patients had no symptoms (n=10) or minor disability (n=184). Plasma Angpt1 concentrations on admission were lower in patients who had severe disability or were dead at 3 months (median, 5.52; interquartile range, 3.81–8.75 ng/mL for mRS 3–6; n=91) compared with those with minor disability (median, 7.04; interquartile range, 4.75–9.92 ng/mL for mRS 0–2; n=194), P=0.012. Patients who had severe disability or were dead at 3 months were also older, more likely to have atrial fibrillation and had more severe strokes on admission (Table 3). After adjustment for age, atrial fibrillation, and stroke severity on admission, an increase in plasma Angpt1 of 10 ng/mL (approximate SD) was associated with an odds ratio of 0.42 (95% confidence intervals, 0.18–0.96) of having a mRS of 3 to 6 at 3 months, P=0.040 (Table 4).
Plasma Angpt1 Does Not Predict Incident Stroke in Older Men
Samples were available from 4032 of the 4248 men (95%) who reattended in wave 2 of HIMS to measure plasma Angpt1. Men with plasma Angpt1 in the first quartile were slightly older, more likely to have a prior history of coronary heart disease, had a larger waist to hip ratio, and a slightly lower serum low-density lipoprotein, high-density lipoprotein, and creatinine (Table I in the online-only Data Supplement). Participants were followed from the time of blood collection (2001–2004) using Western Australian Data Linkage System until death or the end of 2010. During this time, 174 men experienced a stroke. Overall, the incidence of stroke was 0.4, 1.4, 3.0, and 5.3% at 1, 3, 5, and 8 years, respectively. The incidence of stroke at 5 years was 2.9%, 2.5%, 3.9%, and 2.7% for men with plasma Angpt1 in the first, second, third, and fourth quartiles, respectively, P=0.381 (Figure 2). Plasma Angpt1 quartile was not associated with stroke incidence after adjusting for other risk factors using Cox proportional hazard analysis (Table II in the online-only Data Supplement). Findings were similar when plasma Angpt1 was included in a Cox proportional hazard model as a continuous number (hazard ratio, 0.92; 95% confidence interval, 0.74–1.14 per 30 ng/mL; P=0.435).
Previous animal studies have suggested that Angpt1 can protect against ischemic stroke.3–7 Focal cerebral ischemia has been shown to downregulate Angpt1 acutely but upregulate vascular endothelial growth factor in rodent models.19 Upregulation of Angpt1 using an adenovirus vector prior to middle cerebral artery occlusion has been reported to reduce blood–brain barrier leakage and cerebral ischemic volume in mice.4 Findings from several rodent studies suggest that Angpt1 inhibits the ability of vascular endothelial growth factor to stimulate blood–brain barrier leakiness and thereby reduces cerebral ischemic and neuronal damage.4,5 Worse outcomes after middle cerebral artery occlusion in a mouse model of diabetes mellitus have been associated with downregulation of Angpt1,6 whereas improved stroke outcome after exercise has been linked to upregulation of Angpt1 in a rat model of middle cerebral artery occlusion.7 These findings suggested that Angpt1 concentrations might provide important prognostic information on stroke outcome.
In the current study, in keeping with the animal studies, patients with a recent stroke had median plasma concentrations of Angpt1 that were ≈3-fold lower than controls. The association of low plasma Angpt1 concentration with stroke remained statistically significant after adjusting for other risk factors. Patients with the worst stroke outcomes at 3 months had lower plasma Angpt1 concentrations within 8 hours of their stroke admission. Our findings, when interpreted alongside the results of rodent models of stroke, suggest that Angpt1 is downregulated after ischemic stroke and that the degree of downregulation is associated with overall stroke outcome. Another possible explanation for our data is that subjects with lower circulating Angpt1 are at greater risk of incident stroke.
To investigate such a possibility we used subjects from the HIMS. We found no association between plasma Angpt1 and subsequent stroke incidence during an extended follow-up of >4000 older men. Taken together, our findings suggest the possibility that interventions able to upregulate Angpt1 could improve stroke outcomes should this relationship be causal. Measuring plasma Angpt1 soon after ischemic stroke may also provide new prognostic information, although this requires further examination in larger stroke cohorts using prognostic modeling.
Our study has several strengths and weaknesses. Limitations include the possibility of reverse causality, residual confounding, and reduced power because of limited sample sizes. We examined relatively well-characterized groups of subjects including >300 stroke cases and controls and >4000 older men; however, stroke incidence rate was low, and type II error may have affected our results. We reported total stroke incidence rather than focusing on ischemic stroke incidence alone because a large number of strokes were coded as of indeterminate type. Our stroke outcome analysis was restricted to only 285 subjects (85% of the original cohort), and validation in a larger cohort is needed to confirm the contribution of plasma Angpt1 to current models of stroke outcome. Furthermore, stroke severity was unavailable in 10% of patients, and we had no information on cerebral infarct volume. We sought to adjust associations demonstrated for potential confounding factors; however, it is possible that variables we did not measure may have affected the associations we demonstrated. An interventional study directly upregulating Angpt1 would be needed to confirm a protective role of this peptide after ischemic stroke. Given the interaction between Angpt1 and vascular endothelial growth factor, it would have been desirable to measure plasma concentrations of both proteins; however, that was not possible in the current study. Finally, we did not consistently assess ethnicity, and stroke incidence was only examined in men; therefore, our findings may have less relevance to some ethnic groups and women.
In conclusion, the current study suggests that plasma Angpt1 concentrations are lower in patients who have experienced a recent ischemic stroke and that lower concentrations of Angpt1 are associated with worse stroke outcomes. Plasma Angpt1 is not predictive of stroke incidence in healthy older men. The findings when interpreted along with previous rodent studies provide a rationale for further investigations to examine whether upregulating Angpt1 improves stroke outcome.
Sources of Funding
Funding from the Queensland Government, the Townsville Hospital Private Practice Fund, and National Health and Medical Research Council supported this work. J. Golledge and Dr Levi hold Practitioner Fellowships from the National Health and Medical Research Council, Australia. J. Golledge holds a Senior Clinical Research Fellowship from the Office of Health and Medical Research, Queensland.
The authors have been awarded several grants to fund research relevant to this study as detailed under the sources of funding. The authors have no other relevant conflicts of interest.
Guest Editor for this article was Costantino Iadecola, MD.
The online-only Data Supplement is available with this article at http://stroke.ahajournals.org/lookup/suppl/doi:10.1161/STROKEAHA.113.004339/-/DC1.
- Received November 30, 2013.
- Revision received January 7, 2014.
- Accepted January 16, 2014.
- © 2014 American Heart Association, Inc.
- Chen J,
- Yang T,
- Yu H,
- Sun K,
- Shi Y,
- Song W,
- et al
- Maguire J,
- Thakkinstian A,
- Levi C,
- Lincz L,
- Bisset L,
- Sturm J,
- et al
- 10.↵World Health Organization Task Force. Recommendations on stroke prevention, diagnosis and therapy: report of the WHO Task Force on Stroke and Other Cerebrovascular Disorders. Stroke. 1989;20:1407–1431.
- Adams HP Jr.,
- Bendixen BH,
- Kappelle LJ,
- Biller J,
- Love BB,
- Gordon DL,
- et al
- McEvoy M,
- Smith W,
- D’Este C,
- Duke J,
- Peel R,
- Schofield P,
- et al
- Quinn TJ,
- Dawson J,
- Walters MR,
- Lees KR
- Norman PE,
- Jamrozik K,
- Lawrence-Brown MM,
- Le MT,
- Spencer CA,
- Tuohy RJ,
- et al
- Norman PE,
- Flicker L,
- Almeida OP,
- Hankey GJ,
- Hyde Z,
- Jamrozik K
- Golledge J,
- Clancy P,
- Jamrozik K,
- Norman PE