Professional Documents
Culture Documents
From the *Department of Neurology, Wake Forest Baptist University Health Sciences, Winston-Salem, North Carolina; †Department of Neu-
rology, Duke University Medical Center, Durham, North Carolina; ‡Department of Neurology, University of Kentucky, Lexington, Kentucky;
§Duke Clinical Research Institute, Duke University Medical Center, Durham, North Carolina; ‖Department of Neurology, UT Southwestern,
Dallas, Texas; and ¶Department of Medicine, Durham Veterans Affairs, Durham, North Carolina.
Received November 10, 2016; revision received February 24, 2017; accepted March 21, 2017.
Sources of Funding: The AVAIL project was supported by unrestricted funds from Bristol-Myers Squibb/Sanofi Pharmaceuticals Partner-
ship and conducted through collaboration with the American Heart Association and the “Get With The Guidelines–Stroke” program. AVAIL
analyses were also supported in part by the Agency for Healthcare Research and Quality cooperative agreement U18HS016964.
Address correspondence to Nada El Husseini, MD, MHS, Wake Forest Baptist Medical Center, Department of Neurology, Division of Vascular
Neurology, 6th Floor Janeway Tower, Medical Center Boulevard, Winston-Salem, NC 27157. E-mail: nelhusse@wakehealth.edu.
1052-3057/$ - see front matter
© 2017 National Stroke Association. Published by Elsevier Inc. All rights reserved.
http://dx.doi.org/10.1016/j.jstrokecerebrovasdis.2017.03.026
Journal of Stroke and Cerebrovascular Diseases, Vol. ■■, No. ■■ (■■), 2017: pp ■■–■■ 1
ARTICLE IN PRESS
2 N. EL HUSSEINI ET AL.
Age (y)
Median (Q1-Q3) 64 (55-74) 66 (57-75) 60 (52-72) 60 (53-70) 56 (49-61) <.0001
Sex
Female (%) 637 (44.1) 458 (42.3) 52 (50.4) 59 (49.8) 68 (50.7) .112
Race or ethnicity
White (%) 1191 (82.4) 915 (84.6) 80 (77.6) 98 (77.7) 98 (73.1) .014
Black or African American (%) 159 (11.0) 97 (8.9) 17 (16.5) 23 (18.2) 22 (16.4)
Hispanic (%) 36 (2.4) 22 (2.0) 4 (3.8) 4 (3.1) 6 (4.4)
Marriage status
Married (%) 869 (60.1) 671 (62.0) 55 (53.4) 74 (58.7) 69 (51.4) <.0001
Rehabilitation (physical, occupational,
speech therapy or rehabilitation)
between 3 and 12 months
Yes 829 (57.4) 609 (56.3) 61 (59.2) 76 (60.3) 83 (61.9) .511
Recurrent stroke or TIA between 3
and 12 months
Yes (%) 79 (5.4) 47 (4.3) 9 (8.7) 12 (9.5) 11 (8.2) .013
Antidepressant use at 3 or 12 months
Yes (%) 263 (18.2) 160 (14.8) 27 (26.2) 33 (26.1) 43 (32.0) <.0001
Medical history of diabetes mellitus
Yes (%) 383 (29.3) 275 (28.3) 31 (31.9) 32 (28.3) 45 (35.7) .343
Medical history of CAD or MI
Yes (%) 307 (23.4) 233 (24.0) 21 (21.6) 30 (26.5) 23 (18.2) .418
NIHSS
Median (Q1-Q3) 3 (1-6) 3 (1-6) 3 (1-6) 3 (1-6) 3 (2-9) .050
mRS at 3 months 2 1 2 2 2.5 <.0001
Median (Q1-Q3) (1-2) (0-2) (1-3) (2-3) (2-3)
mRS at 12 months 2 1 2 2 2 <.0001
Median (Q1-Q3) (1-2) (0-2) (2-3) (1-2) (2-3)
Worsening mRS between 3 and 12 months 349 (24.2) 273 (25.3) 33 (31.7) 17 (13.5) 26 (19.4) .0001
Abbreviations: CAD, coronary artery disease; MI, myocardial infarction; mRS, modified Rankin score; NIHSS, National Institutes of Health
Stroke Scale; TIA, transient ischemic attack.
were associated with worsening mRS (Table 2). Com- between incident and persistent depression (OR = 1.74;
pared with no depression at either time point, persistent 95% CI: .86-3.51) (Table 2).
depression (OR = .85, 95% CI: .53-1.34) and incident de-
pression (OR = 1.48, 95% CI: .95-2.30) were not significantly
Discussion
associated with worsening mRS. In contrast, those with
resolving depression were less likely to have a worsen- Post-stroke depression may worsen functional outcome
ing mRS (OR = .49, 95% CI: .29-0.83). A sensitivity analysis by interfering with both cognitive and physical recov-
excluding those with recurrent stroke or TIA yielded similar ery and impeding participation in rehabilitation. In addition,
results (Table 2). depression may alter brain regulatory mechanisms that
When persistent depression was used as the refer- may in turn affect outcomes; for example, depression is
ence group, depression status at 3 months and 12 months associated with decreased cortical regulation of limbic ac-
remained associated with change in functional outcome tivation in response to various stimuli.26 In a systematic
(P = .023). Compared with persistent depression, resolv- review of the association between post-stroke depres-
ing depression tended to be associated with a lower sion and post-stroke functional outcome (14 cohorts with
likelihood of worsening mRS, although the difference was 4498 participants), post-stroke depression was inversely
not statistically significant (OR .45, 95% CI: .20-1.02). There associated with functional outcome.27 Most of the cohorts
was no significant difference in functional deterioration in this review, however, included patients with intracranial
ARTICLE IN PRESS
DEPRESSION STATUS AND FUNCTIONAL DECLINE 5
Table 2. Multivariate logistic regression analysis of depression and functional decline with “never depressed” and with
“persistently depressed” as the reference groups
Incident depression versus never depressed 1.48 (.95-2.30) .010 1.49 (.94-2.36) .013
Resolving depression versus never depressed .49 (.29-0.83) .49 (.28-0.86)
Persistent depression versus never depressed .85 (.53-1.34) .78 (.48-1.27)
Incident depression versus persistently depressed 1.74 (.86-3.51) .023 2.10 (.98-4.46) .023
Resolving depression versus persistently depressed .45 (.20-1.02) .58 (.24-1.38)
Never depressed versus persistently depressed 1.10 (.64-1.88) 1.26 (.70-2.26)
Abbreviations: CAD, coronary artery disease; CI, confidence interval; MI, myocardial infarction; NIHSS, National Institutes of Health
Stroke Scale; OR, odds ratio; TIA, transient ischemic attack.
Sex, race or ethnicity, post-hospital rehabilitation, recurrent stroke or TIA between 3 and 12 months post stroke, baseline NIHSS, medical
history of diabetes, medical history of CAD or MI, insurance status, marital status, and antidepressant use at 3 or 12 months were excluded
from the final model by backward selection. The analysis was also adjusted for age. The table also includes the results of the sensitivity anal-
ysis excluding subjects with recurrent stroke or TIA.
hemorrhage in addition to those with ischemic stroke and the Barthel Index in the first 52 weeks after stroke, but
assessed functional outcome days to years after the as- functional outcome was assessed at only a single time
sessment of depression. 27 Depression did not affect point.33 Similarly, increases in positive emotion score (as-
functional improvement in 2 studies.28,29 Both of these sessed by 4 items from the Center for Epidemiologic Studies
studies were rehabilitation based and did not evaluate Depression Scale) following stroke compared with no change
whether change in depression affected the trajectory of or a decline was associated with better functional status
functional outcome. One study found that function im- assessed at 3 months.34 Another study of an inpatient medical
proved over time regardless of depression in the first year rehabilitation-based cohort evaluated depression trajec-
after stroke, but functional outcome was better among tory (measured by change in Center for Epidemiologic
those who were not initially depressed.28 In another study, Studies Depression Scale score) in relation to change in
depression was associated with increased disability at both functional status (Functional Independence Measure in-
admission and follow-up in the first 6 months after stroke, strument) in the first year following stroke found greater
but depression was not associated with functional improvement in functional status among those who were
improvement.29 In another study, both early-onset (within not depressed at discharge and had fewer depressive symp-
2 weeks) and late-onset post-stroke depression (after 2 toms on follow-up. This study, however, combined ischemic
weeks and within the first year following stroke) were and hemorrhagic stroke and was limited to subjects aged
independently associated with disability and poor func- 65 years and older, limiting generalizability.35 In contrast
tional outcome at 1 year after stroke.30 Recovery from to these findings, another smaller rehabilitation based-
depression within 1 year decreased but did not elimi- study of 64 patients comparing those with and without
nate the adverse impacts of post-stroke depression on depression assessed at 1 time point within weeks of first
functional outcome and quality of life. In this study, out- stroke found similar improvements in functional status
comes including disability (mRS ≥ 2) and quality of life between the 2 groups, but greater functional impairment
(Short Form-36 Health Survey) were assessed at 1-year among those who were depressed.10
follow-up but not longitudinally.30 Even a decade after Because post-stroke depression is dynamic, with inci-
stroke in individuals aged 18-50 years, depressive symp- dent and resolving depression occurring in the first year
toms and anxiety were more prevalent and associated with after stroke, we determined if changes in depression status
poor functional outcome.31 were associated with functional decline. We found that
Whereas multiple studies showed an association between the improvement in depression following hospitaliza-
post-stroke depression and poor functional outcome, fewer tion for ischemic stroke was associated with a lower
studies evaluated changes in psychological symptoms in likelihood of having a worsening in functional status
relation to changes of functional outcome after stroke.32 between 3 months and 12 months post stroke. The lower
One study found an association between the trajectory likelihood of functional deterioration with resolving de-
of psychological symptoms assessed by the General Health pression in comparison with no depression may be due
Questionnaire and the functional outcome measured by to a worse initial mRS in those who were initially
ARTICLE IN PRESS
6 N. EL HUSSEINI ET AL.
depressed. We previously found that poor functional prospectively and longitudinally evaluated depressive
outcome at 3 months was associated with depression at symptoms, antidepressant use, and functional outcome
12 months (OR 1.49, 95% CI: 1.29-1.73 for each 1-unit in- in the first year post stroke. In AVAIL, subjects were re-
crease in mRS) and with persistent depression (OR 1.93, cruited from geographically diverse hospitals participating
95% CI: 1.62-2.30).6 In the current analysis, worsening in in the American Heart Association GWTG–Stroke program.
functional outcome did not occur in patients who were Hospitals participating in the GWTG–Stroke program do
persistently depressed, likely because functional outcome tend to be larger, urban, and teaching centers. Despite
was “poor” (mRS was high) at both time points in this these differences, a study of the representativeness of the
group.10 GWTG–Stroke registry indicates that the data are gen-
Our findings could indicate that the resolution of de- erally representative of national fee-for-service Medicare
pression prevents worsening in functional status. ischemic stroke populations, providing support for ex-
Alternatively, no worsening or improvement in function- ternal validity.38
al status might lead to improved mood. In addition, In contrast to cross-sectional or only baseline data, our
although heavily motor weighted, the mRS may be af- study provides a longitudinal analysis of the changes in
fected by nonphysical attributes such as post-stroke mood both depression and functional outcome in the first year
disturbances that can contribute to perceived disability.7,36 following ischemic stroke. Our observations suggest that
The positive association between incident depression recovery from post-stroke depression is associated with
at 12 months and worsening mRS scores was not sig- a decrease in post-stroke functional decline. Although func-
nificant. The lack of significance may be due to the tional decline occurred in those without depression and
relatively small sample size, but it is also plausible that in those with persistent depression at a similar rate, de-
the effect of late-onset depression on functional outcome pression was associated with worse functional outcome.
may be qualitatively different from early onset depres- These results point to the importance of the longitudi-
sion, as most stroke recovery takes place over the first nal evaluation of depression in the first year following
several months. Surprisingly, recurrent stroke or TIA was stroke.
not associated with functional decline. This could be at-
tributed to how recurrent stroke or TIA was ascertained
in AVAIL (participant self-report). References
This study has several limitations. It is a secondary anal-
1. Dhamoon MS, Moon YP, Paik MC, et al. Long-term
ysis of a prospective cohort registry. The majority of the
functional recovery after first ischemic stroke: the Northern
cohort had mild disability (mRS < 3), likely explaining the Manhattan Study. Stroke 2009;40:2805-2811.
lower rate of depression in this cohort as compared with 2. Mandip S, Dhamoon MSE. Disability trajectories before
other studies.4 About 82.6% were white, which may limit and after vascular events: the Cardiovascular Health
the generalizability to other race or ethnic groups. The Study. International Stroke Conference oral
abstracts—session title: clinical rehabilitation and recovery
demographics of the AVAIL population, however, did not
oral abstracts II: abstract 142. Stroke 2015;46:A142.
differ substantially from the overall GWTG–Stroke cohort. 3. Dhamoon MS, Moon YP, Paik MC, et al. Trajectory of
Because subjects rather than proxies were required to com- functional decline before and after ischemic stroke: the
plete the PHQ-8, 526 patients were excluded. For this Northern Manhattan Study. Stroke 2012;43:2180-2184.
reason, the results may be biased toward those with mild 4. Hackett ML, Pickles K. Part I: frequency of depression
after stroke: an updated systematic review and meta-
disability including those without aphasia or significant
analysis of observational studies. Int J Stroke 2014;9:1017-
cognitive impairments. In addition, prestroke depres- 1025.
sion, shown to be strongly associated with poststroke 5. Hackett ML, Yapa C, Parag V, et al. Frequency of
depression, was not assessed in our study.27 Given this depression after stroke: a systematic review of
limitation, we cannot exclude the possibility that a pro- observational studies. Stroke 2005;36:1330-1340.
6. El Husseini N, Goldstein LB, Peterson ED, et al.
portion of subjects in the cohort had a prestroke history
Depression and antidepressant use after stroke and
of depression. A confounding effect of selective sero- transient ischemic attack. Stroke 2012;43:1609-1616.
tonin reuptake inhibitors that are often used for the 7. Dennis M, O’Rourke S, Lewis S, et al. Emotional outcomes
treatment of post-stroke depression on motor recovery after stroke: factors associated with poor outcome. J
is possible.37 Although antidepressant medications in this Neurol Neurosurg Psychiatry 2000;68:47-52.
8. Donnellan C, Hickey A, Hevey D, et al. Effect of mood
study were not analyzed by category, we adjusted for
symptoms on recovery one year after stroke. Int J Geriatr
antidepressant use in the regression analyses, and only Psychiatry 2010;25:1288-1295.
antidepressants used at a commonly prescribed dose for 9. Andersen G, Vestergaard K, Ingemann-Nielsen M, et al.
depression indication based on a previously validated al- Risk factors for post-stroke depression. Acta Psychiatr
gorithm were included to help ensure that they were being Scand 1995;92:193-198.
10. Sinyor D, Amato P, Kaloupek DG, et al. Post-stroke
used for depression and not for other indications.21
depression: relationships to functional impairment, coping
The strengths of this study were the large cohort of strategies, and rehabilitation outcome. Stroke 1986;17:1102-
patients with stroke with complete follow-up who had 1107.
ARTICLE IN PRESS
DEPRESSION STATUS AND FUNCTIONAL DECLINE 7
11. Schwab-Malek S, Vatankhah B, Bogdahn U, et al. De- 26. Anand A, Li Y, Wang Y, et al. Activity and connectivity
pressive symptoms and quality of life after thrombolysis of brain mood regulating circuit in depression: a functional
in stroke: the TEMPiS study. J Neurol 2010;257:1848-1854. magnetic resonance study. Biol Psychiatry 2005;57:1079-
12. Willey JZ, Disla N, Moon YP, et al. Early depressed mood 1088.
after stroke predicts long-term disability: the Northern 27. Kutlubaev MA, Hackett ML. Part II: predictors of
Manhattan Stroke Study (NOMASS). Stroke 2010;41:1896- depression after stroke and impact of depression on stroke
1900. outcome: an updated systematic review of observational
13. Schmid AA, Kroenke K, Hendrie HC, et al. Poststroke studies. Int J Stroke 2014;9:1026-1036.
depression and treatment effects on functional outcomes. 28. Bacher YK-BN, Mayo N, Becker R, et al. A longitudinal
Neurology 2011;76:1000-1005. study of depression among stroke patients participating
14. Bushnell C, Zimmer L, Schwamm L, et al. The Adherence in a rehabilitation program. Can J Rehabi 1990;4:27-37.
eValuation After Ischemic Stroke Longitudinal (AVAIL) 29. van de Weg FB, Kuik DJ, Lankhorst GJ. Post-stroke
registry: design, rationale, and baseline patient depression and functional outcome: a cohort study
characteristics. Am Heart J 2009;157:428-435. e422. investigating the influence of depression on functional
15. Vandenbroucke JP, von Elm E, Altman DG, et al. recovery from stroke. Clin Rehabil 1999;13:268-272.
Strengthening the Reporting of Observational Studies in 30. Shi YZ, Xiang YT, Yang Y, et al. Depression after minor
Epidemiology (STROBE): explanation and elaboration. stroke: the association with disability and quality of life—a
Ann Intern Med 2007;147:W163-W194. 1-year follow-up study. Int J Geriatr Psychiatry
16. American Psychiatric Association. Diagnostic and statistical 2016;31:421-427.
manual of mental disorders. 4th ed. Washington, DC: 31. Maaijwee NA, Tendolkar I, Rutten-Jacobs LC, et al.
American Psychiatric Association, 1994. Long-term depressive symptoms and anxiety after
17. Kroenke K, Strine TW, Spitzer RL, et al. The PHQ-8 as transient ischaemic attack or ischaemic stroke in young
a measure of current depression in the general population. adults. Eur J Neurol 2016;23:1262-1268.
J Affect Disord 2009;114:163-173. 32. Ghosal MK, Burman P, Singh V, et al. Correlates of
18. Wells TS, Horton JL, LeardMann CA, et al. A comparison functional outcome among stroke survivors in a
of the PRIME-MD PHQ-9 and PHQ-8 in a large military developing country—a prospective community-based
prospective study, the Millennium Cohort Study. J Affect study from India. J Stroke Cerebrovasc Dis 2014;23:2614-
Disord 2013;148:77-83. 2621.
19. Williams LS, Brizendine EJ, Plue L, et al. Performance 33. West R, Hill K, Hewison J, et al. Psychological disorders
of the PHQ-9 as a screening tool for depression after after stroke are an important influence on functional
stroke. Stroke 2005;36:635-638. outcomes: a prospective cohort study. Stroke 2010;41:
20. Lowe B, Kroenke K, Herzog W, et al. Measuring 1723-1727.
depression outcome with a brief self-report instrument: 34. Seale GS, Berges IM, Ottenbacher KJ, et al. Change in
sensitivity to change of the Patient Health Questionnaire positive emotion and recovery of functional status
(PHQ-9). J Affect Disord 2004;81:61-66. following stroke. Rehabil Psychol 2010;55:33-39.
21. Damush TM, Jia H, Ried LD, et al. Case-finding algorithm 35. Ostir GV, Berges IM, Ottenbacher A, et al. Patterns of
for post-stroke depression in the veterans health change in depression after stroke. J Am Geriatr Soc
administration. Int J Geriatr Psychiatry 2008;23:517-522. 2011;59:314-320.
22. Berges IM, Kuo YF, Ottenbacher KJ, et al. Recovery of 36. Banks JL, Marotta CA. Outcomes validity and reliability
functional status after stroke in a tri-ethnic population. of the modified Rankin scale: implications for stroke
PM R 2012;4:290-295. clinical trials: a literature review and synthesis. Stroke
23. Hankey GJ, Spiesser J, Hakimi Z, et al. Rate, degree, and 2007;38:1091-1096.
predictors of recovery from disability following ischemic 37. Chollet F, Tardy J, Albucher JF, et al. Fluoxetine for motor
stroke. Neurology 2007;68:1583-1587. recovery after acute ischaemic stroke (FLAME): a
24. Jorgensen HS, Reith J, Nakayama H, et al. What randomised placebo-controlled trial. Lancet Neurol
determines good recovery in patients with the most severe 2011;10:123-130.
strokes? The Copenhagen Stroke Study. Stroke 1999;30: 38. Reeves MJ, Fonarow GC, Smith EE, et al.
2008-2012. Representativeness of the Get With The Guidelines–Stroke
25. Mikami K, Jorge RE, Adams HP Jr, et al. Effect of Registry: comparison of patient and hospital characteristics
antidepressants on the course of disability following among Medicare beneficiaries hospitalized with ischemic
stroke. Am J Geriatr Psychiatry 2011;19:1007-1015. stroke. Stroke 2012;43:44-49.