Abstract
Objective
To determine if depression, cognitive impairment without dementia (CIND), and/or dementia are each independently associated with risk of ischemic stroke and to identify characteristics that could modify these associations.
Methods
This retrospective-cohort study examined a population-based sample of 7,031 Americans > 50 years old participating in the Health and Retirement Study (HRS) (1998-2008) who consented to have their interviews linked to their Medicare claims. The 8-item Center for Epidemiologic Studies Depression Scale and/or International Classification of Disease, Ninth Revision, Clinical Modification (ICD-9-CM) depression diagnoses were used to identify baseline depression. The Modified Telephone Interview for Cognitive Status and/or ICD-9-CM dementia diagnoses were used to identify baseline CIND or dementia. Hospitalizations for ischemic stroke were identified via ICD-9-CM diagnoses.
Results
After adjusting for demographics, medical comorbidities, and health-risk behaviors, CIND alone (Odds Ratio [OR]: 1.37, 95%CI: 1.11, 1.69) and co-occurring depression and CIND (OR: 1.65, 95%CI: 1.24, 2.18) were independently associated with increased odds of ischemic stroke. Depression alone was not associated with odds of ischemic stroke (OR: 1.11, 95%CI: 0.88, 1.40) in unadjusted analyses. Neither dementia alone (OR: 1.09, 95%CI: 0.82, 1.45) nor co-occurring depression and dementia (OR: 1.25, 95%CI: 0.89, 1.76) were associated with odds of ischemic stroke after adjusting for demographics.
Conclusions
CIND as well as co-occurring depression and CIND are independently associated with increased risk of ischemic stroke. Individuals with co-occurring depression and CIND represent a high-risk group that may benefit from targeted interventions to prevent stroke.
Keywords: depression, cognitive impairment, dementia, ischemic stroke
INTRODUCTION
Each year, nearly 800,000 Americans experience a stroke (1). In 2008, strokes accounted for nearly $19 billion in healthcare costs (1). They are a leading cause of mortality among older adults with one in six strokes being fatal (1). Furthermore, a substantial proportion of stroke survivors face functional disability and cognitive impairment (2, 3), and nearly one-third develop depression (4). In light of the enormous public health burden imposed by strokes, increasing understanding of potentially modifiable risk factors is important.
Depression, cognitive impairment without dementia (CIND) and dementia are prevalent among older adults (5-7). Prior studies have examined depression or dementia as potential risk factors for stroke and have found that both are individually associated with increased risk of stroke (8-18). One of these studies suggested that the association of depression with risk of stroke may be modified by female sex (14). However, while depression is frequently co-occurring with mild cognitive impairment and dementia (19), no study to date has considered the impact of co-occurring depression and mild cognitive impairment or dementia on the risk of stroke. Further, some prior studies have been limited by reliance on administrative data alone (13, 15, 16), or on self-report data alone (9). Finally, if depression, CIND and dementia are indeed risk factors for strokes, then identifying sub-groups of older adults with depression, CIND or dementia at heightened risk is important for intervention targeting.
The present study uses data from an ongoing longitudinal investigation of health outcomes in adults over age 50 to examine if depression, CIND, and dementia are independently associated with risk of ischemic stroke. We hypothesized that depression, CIND, and dementia would each be independently associated with increased risk of a future ischemic stroke, and that co-occurring depression and CIND or dementia would be associated with greater risk of ischemic stroke than that conveyed by any one disorder alone. We also sought to determine if any associations found between baseline depression, CIND or dementia and subsequent ischemic stroke risk were modified by sex, lower socioeconomic status, or relevant medical comorbidities.
METHODS
Population
This study uses nationally representative data from Americans over age 50 participating in the Health and Retirement Study (HRS). Beginning in 1992, the HRS has enrolled over 31,000 individuals. Subjects are interviewed every two years with a follow-up rate exceeding 90-95% (including proxies) (20). In addition to both high initial and re-interview response rates, non-response adjusted participant-level sample weights for each biennial interview are used to maintain representativeness over time as well as minimize the risk of healthy survivor bias (5). In addition, over 80% of eligible respondents have consented to linkage of their Medicare claims records with study data (20).
Our sample consisted of the 7,031 HRS respondents interviewed in 1998 or 2000 who consented to linkage of their Medicare claims records. We followed them through death (n = 3,546) or the 2008 interview.
Standard Protocol Approvals, Registrations, and Patient Consents
The HRS protocol was approved by the University of Michigan Institutional Review Board. Participants provided informed consent upon enrollment and again for linkage to Medicare claims.
Primary Independent Variable
The primary independent variable in our analyses was depression, CIND or dementia status at baseline defined categorically as no disorder, depression alone, CIND alone, dementia alone, co-occurring depression and CIND, or co-occurring depression and dementia.
We defined depression as either a score of ≥ 4 on the 8-item Center for Epidemiologic Studies Depression Scale (CES-D-8) (21) at the baseline HRS interview or a depression diagnosis in the Medicare claims in the previous two years based on International Classification of Disease, Ninth Revision, Clinical Modification (ICD-9-CM) codes 296.2, 296.3, 298.0, 300.4, or 311.0. The CES-D-8 cutoff score of ≥ 4 was shown to be comparable to the cutoff score of ≥ 16 on the full CES-D (21), and has been used in other relevant studies (5, 22).
Cognitive impairment was assessed in the HRS using the modified Telephone Interview for Cognitive Status (TICSm), a composite measure with items testing immediate and delayed recall, working memory with serial sevens subtraction, and backwards counting testing attention and processing speed (23). The version of the TICSm used in the HRS was validated against neuropsychiatric interview in the Aging, Demographics and Memory Study, and found to have a weighted accuracy of 69.2% in correct classification of individuals as either having normal cognition, CIND, or dementia (23). We defined CIND as a TICSm score of 7 to 11 at the baseline HRS interview (23). Dementia was defined as either a TICSm score of ≤ 6 (23) or dementia diagnosis (ICD-9-CM codes 290.0-290.42, 291.2, 294.1, 294.8, 331.0, 331.1, 331.11, 331.19, or 331.82) in the Medicare claims in the two years before baseline.
Demographics, Medical Comorbidities and Health-Risk Behaviors
Demographic information (e.g., age, sex, race, education, marital/partnered status, and dual Medicare-Medicaid eligibility [individuals with Medicare whose income is low enough to qualify for additional health insurance coverage through Medicaid]) was obtained from the HRS interviews. Elixhauser (24) and Charlson comorbidity diagnoses (25) from the two years before baseline found to be associated with depression, cognitive impairment and stroke risk (26-28) were obtained from Medicare claims. Information on alcohol use and smoking came from the HRS interviews.
Outcome of Interest
Our outcome of interest was ischemic stroke. We used ICD-9-CM codes 433, 434 and 436 to identify hospitalizations for which the principal discharge diagnosis was for ischemic stroke (29).
Statistical Analysis
We present descriptive data as means and standard deviations (SDs) or proportions. Since we did not have data on time from our exposures of interest to ischemic stroke, we used logistic regression models with robust error variances to estimate odds ratios (ORs) and 95% Confidence Intervals (95%CIs) for the association of baseline depression, CIND or dementia with ischemic stroke. First, we tested this association without adjustment. We then sequentially adjusted for potential confounders chosen a priori based on prior research identifying their associations with depression, CIND or dementia and stroke risk (8-18). Non-normally distributed covariates were categorized. The sequence of adjustments was: 1) demographic characteristics (e.g., age categorized by deciles, sex, race categorized as white versus non-white, education categorized as < high school graduate versus ≥ high school graduate, marital/partnered status categorized as married/partnered versus single/separated/widowed, dual Medicare-Medicaid eligibility); 2) comorbid conditions (myocardial infarction, cerebrovascular disease, congestive heart failure, valvular disease, pulmonary circulation disease, peripheral vascular disease, other neurological disorders, diabetes mellitus, and hypertension); and 3) health-risk behaviors (e.g., alcohol use categorized by the number of drinks per day, smoking status).
To determine if any associations found between baseline depression, CIND or dementia status and ischemic stroke risk were modified by sex, dual Medicare-Medicaid status, or important medical comorbidities (e.g., baseline cerebrovascular disease, diabetes, hypertension), we entered interaction terms (e.g., depression, CIND or dementia status x sex) one at a time into our adjusted regression models.
We conducted four sensitivity analyses. First, we repeated our regression analyses excluding individuals with a baseline history of cerebrovascular disease. Second, we estimated the association of baseline depression, CIND or dementia status with odds of ischemic stroke using only the CES-D-8 and TICSm thresholds to define cases of depression or dementia. Since the CES-D-8 was only administered to self-respondents (21), this sensitivity analysis only included the 6,256 eligible self-respondents. Third, we examined if our results were affected by using a CES-D-8 cut-off score of ≥ 3 in our depression definition since a prior study examining depression as a risk factor for stroke utilizing HRS data used this threshold for the CES-D-8 (9). Finally, we repeated our final regression model with propensity score adjustment to examine whether our results were biased by attrition due to death during follow-up (30). In this analysis, we initially ascertained bivariate associations between all covariates (e.g., demographics, comorbid conditions and health-risk behaviors) and death during follow-up (30). Then, we fit a logistic regression model that predicted whether a participant would die during follow-up as a function of all significant covariates from bivariate analyses (30). Since all of the covariates had significant bivariate associations with death during follow-up, they were all included in this model. The predicted probabilities from this model were then used as a propensity score adjustment to our final logistic regression model for ischemic stroke (30).
We used two-sided significance tests for all analyses with statistical significance set at P < 0.05. Analyses were performed with appropriate components of the STATA 12 (Stata Corporation, College Station, TX) statistical software program.
RESULTS
Table 1 presents the baseline demographic characteristics, medical comorbidities, and health-risk behaviors of the entire sample and grouped by depression, CIND or dementia status. At baseline, 12.0% were depressed (188 [22.3%] by an ICD-9-CM depressive disorder diagnosis alone, 482 [57.2%] by CES-D-8 alone, and 172 [20.4%] by both) without clinically notable cognitive impairment, 13.4% had CIND without co-occurring depression, 7.6% had dementia (188 [35.2%] by an ICD-9-CM dementia diagnosis alone, 325 [60.9%] by TICSm alone, and 21 [3.9%] by both) without co-occurring depression, 5.5% had co-occurring depression (72 [18.5%] by an ICD-9-CM depressive disorder diagnosis alone, 262 [67.2%] by CES-D-8 alone, and 56 [14.3%] by both) and CIND, and another 4.3% had co-occurring depression (43 [14.1%] by an ICD-9-CM depressive disorder diagnosis alone, 154 [50.5%] by CES-D-8 alone, and 108 [35.4%] by both) and dementia (97 [31.8%] by an ICD-9-CM dementia diagnosis alone, 187 [61.3%] by TICSm alone, and 21 [6.9%] by both).
Table 1.
Sample demographic, comorbid conditions and health-risk behavioral characteristics
| Entire cohort (n = 7,031)) |
No depression, CIND, or dementia (n = 4,015) |
Depression alone (n = 842) |
CIND alone (n = 945) |
Dementia alone (n = 534) |
Depression and CIND (n = 390) |
Depression and dementia (n = 305) |
|
|---|---|---|---|---|---|---|---|
| Demographics | |||||||
| Age | |||||||
| <65 | 390 (6%) | 187 (5%) | 117 (14%) | 34 (4%) | 9 (1%) | 33 (8%) | 10 (3%) |
| 65-74 | 3,127 (44%) | 2,132 (53%) | 376 (44%) | 321 (34%) | 100 (19%) | 140 (36%) | 58 (19%) |
| 75-84 | 2,527 (36%) | 1,362 (34%) | 267 (32%) | 396 (42%) | 213 (40%) | 148 (38%) | 141 (46%) |
| ≥85 | 987 (14%) | 334 (8%) | 82 (10%) | 194 (20%) | 212 (40%) | 69 (18%) | 96 (32%) |
| Sex | |||||||
| Male | 2,959 (42%) | 1,853 (46%) | 247 (29%) | 422 (45%) | 204 (38%) | 128 (33%) | 105 (34%) |
| Female | 4,072 (58%) | 2,162 (54%) | 595 (71%) | 523 (55%) | 330 (62%) | 262 (67%) | 200 (66%) |
| Race | |||||||
| White | 6,064 (86%) | 3,649 (91%) | 750 (89%) | 778 (82%) | 375 (70%) | 283 (73%) | 229 (75%) |
| Non-white | 966 (14%) | 365 (9%) | 92 (11%) | 167 (18%) | 159 (30%) | 107 (27%) | 76 (25%) |
| Education | |||||||
| ≥ High school graduate | 4,280 (61%) | 2,839 (71%) | 531 (63%) | 459 (49%) | 190 (36%) | 160 (41%) | 101 (33%) |
| < High school graduate | 2,746 (39%) | 1,171 (29%) | 311 (37%) | 486 (51%) | 344 (64%) | 230 (59%) | 204 (67%) |
| Marital status | |||||||
| Married/partnered | 3,996 (57%) | 2,617 (65%) | 406 (48%) | 492 (52%) | 208 (39%) | 149 (38%) | 124 (41%) |
| Single/separated/widowed | 3,030 (43%) | 1,397 (35%) | 435 (52%) | 453 (48%) | 325 (61%) | 240 (62%) | 180 (59%) |
| Dual Medicare-Medicaid | |||||||
| No | 5,635 (80%) | 3,558 (89%) | 638 (76%) | 719 (76%) | 333 (62%) | 220 (56%) | 167 (55%) |
| Yes | 1,396 (20%) | 457 (11%) | 204 (24%) | 226 (24%) | 201 (38%) | 170 (44%) | 138 (45%) |
| Comorbid Conditions | |||||||
| Myocardial infarction | 260 (4%) | 118 (3%) | 43 (5%) | 39 (4%) | 22 (4%) | 22 (6%) | 16 (5%) |
| Cerebrovascular disease | 359 (5%) | 153 (4%) | 39 (5%) | 37 (4%) | 56 (10%) | 30 (8%) | 44 (14%) |
| Congestive heart failure | 490 (7%) | 165 (4%) | 78 (9%) | 66 (7%) | 69 (13%) | 51 (13%) | 61 (20%) |
| Valvular disease | 272 (4%) | 117 (3%) | 44 (5%) | 40 (4%) | 24 (4%) | 21 (5%) | 26 (8%) |
| Pulmonary circulation disease | 87 (1%) | 40 (1%) | 17 (2%) | 9 (1%) | 7 (1%) | 6 (1%) | 8 (3%) |
| Peripheral vascular disease | 251 (4%) | 106 (3%) | 37 (4%) | 28 (3%) | 28 (5%) | 24 (6%) | 28 (9%) |
| Other neurological disorders | 203 (3%) | 69 (2%) | 15 (2%) | 13 (1%) | 51 (10%) | 13 (3%) | 42 (14%) |
| Diabetes without chronic | 465 (7%) | 185 (5%) | 76 (9%) | 58 (6%) | 54 (10%) | 46 (12%) | 46 (15%) |
| complications | |||||||
| Diabetes with chronic | 119 (2%) | 49 (1%) | 16 (2%) | 14 (1%) | 13 (2%) | 14 (6%) | 13 (4%) |
| complications | |||||||
| Coagulopathy | 72 (1%) | 29 (1%) | 1 (1%) | 9 (1%) | 7 (1%) | 3 (1%) | 5 (2%) |
| Hypertension | 1,240 (18%) | 529 (13%) | 188 (22%) | 160 (17%) | 134 (25%) | 103 (26%) | 126 (41%) |
| Health-risk behaviors | |||||||
| Alcohol use | |||||||
| Daily drinker | 1,664 (24%) | 1,182 (29%) | 147 (17%) | 199 (21%) | 58 (11%) | 57 (15%) | 21 (7%) |
| # of drinks per day among | 1.8 (1.4) | 1.7 (1.3) | 1.8 (1.2) | 1.8 (1.7) | 1.9 (1.6) | 1.8 (1.3) | 2.4 (2.9) |
| daily drinkers | |||||||
| Smoking status | |||||||
| Never smoked | 3,015 (43%) | 1,664 (42%) | 339 (40%) | 440 (47%) | 251 (48%) | 172 (44%) | 149 (49%) |
| Former smoker | 3,197 (46%) | 1,887 (47%) | 386 (46%) | 403 (43%) | 242 (46%) | 157 (41%) | 122 (40%) |
| Current smoker | 766 (11%) | 434 (11%) | 113 (14%) | 96 (10%) | 33 (6%) | 59 (15%) | 31 (10%) |
All values are N(%) or mean (SD).
Abbreviation: CIND = cognitive impairment without dementia.
During the follow-up period (mean: 6.8 years, SD: 3.1 years), 875 (12.4%) participants were hospitalized at least once for an ischemic stroke. Of these individuals, 99 (11.3%) were depressed alone, 149 (17.0%) had CIND alone, 75 (8.6%) had dementia alone, 74 (8.5%) had co-occurring depression and CIND, and 47 (5.4%) had co-occurring depression and dementia.
Associations of Baseline Depression, CIND or Dementia with Odds of Stroke
In unadjusted analyses, CIND alone (OR: 1.55, 95%CI: 1.27, 1.90), dementia alone (OR: 1.36, 95%CI: 1.04, 1.77), co-occurring depression and CIND (OR: 1.95, 95%CI: 1.48, 2.56) and co-occurring depression and dementia (OR: 1.51, 95%CI: 1.09, 2.10) were all associated with greater odds of having an ischemic stroke during the follow-up period. Depression alone was not associated with odds of ischemic stroke (OR: 1.11, 95%CI: 0.88, 1.40).
After adjusting for participant demographic characteristics, CIND alone (OR: 1.39, 95%CI: 1.13, 1.71) and co-occurring depression and CIND (OR: 1.76, 95%CI: 1.33, 2.32) remained associated with increased odds of ischemic stroke (Table 2), while dementia alone and co-occurring depression and dementia were no longer significantly associated with odds of ischemic stroke. These associations persisted after adjusting for medical comorbidities (CIND alone: OR: 1.39, 95%CI: 1.13, 1.72; co-occurring depression and CIND: OR: 1.68, 95%CI: 1.27, 2.23) and health-risk behaviors (CIND alone: OR: 1.37, 95%CI: 1.11, 1.69; co-occurring depression and CIND: OR: 1.65, 95%CI: 1.24, 2.18).
Table 2.
Adjusted associations of depression, CIND and dementiaa at baseline with odds of ischemic stroke
| Adjusted for demographics (n = 7,021) |
Adjusted for demographics and comorbid conditions (n = 7,021) |
Adjusted for demographics, comorbid conditions and health-risk behaviors (n = 6,969) |
|
|---|---|---|---|
| Odds Ratio (95% Confidence Interval) | |||
|
| |||
| Depression alone | 1.13 (0.89, 1.43) | 1.08 (0.85, 1.38) | 1.09 (0.85, 1.38) |
| CIND alone | 1.39 (1.13, 1.71)† | 1.39 (1.13, 1.72)† | 1.37 (1.11, 1.69)† |
| Dementia alone | 1.09 (0.82, 1.45) | 1.08 (0.81, 1.43) | 1.08 (0.81, 1.44) |
|
Co-occurring depression
and CIND |
1.76 (1.33, 2.32)‡ | 1.69 (1.27, 2.23)‡ | 1.65 (1.24, 2.18)‡ |
|
Co-occurring depression
and dementia |
1.25 (0.89, 1.76) | 1.16 (0.82, 1.65) | 1.16 (0.82, 1.65) |
| Age | |||
| 65-74 | 1.45 (0.99, 2.12) | 1.46 (0.99, 2.14) | 1.53 (1.03, 2.26)* |
| 75-84 | 1.91 (1.30, 2.80)† | 1.88 (1.28,2.77)† | 1.99 (1.34, 2.97)† |
| ≥ 85 | 2.15 (1.42, 3.25)‡ | 2.15 (1.41, 3.26)‡ | 2.28 (1.48, 3.51)‡ |
| Female | 0.83 (0.71, 0.97)* | 0.83 (0.71, 0.97)* | 0.84 (0.71, 0.99)* |
| Non-white | 1.13 (0.92, 1.39) | 1.12 (0.91, 1.38) | 1.11 (0.90, 1.37) |
| < High school graduate | 0.99 (0.84, 1.16) | 0.96 (0.82, 1.13) | 0.96 (0.81, 1.13) |
| Single/separated/widowed | 1.17 (1.00, 1.38) | 1.18 (1.00, 1.39) | 1.17 (0.99, 1.38) |
| Dual Medicare-Medicaid | 1.12 (0.92, 1.36) | 1.08 (0.89, 1.32) | 1.08 (0.89, 1.32) |
| Myocardial infarction | 0.84 (0.57, 1.25) | 0.86 (0.58, 1.28) | |
| Cerebrovascular disease | 1.57 (1.17, 2.11)† | 1.57 (1.17, 2.11)† | |
| Congestive heart failure | 0.87 (0.65, 1.18) | 0.89 (0.66, 1.20) | |
| Valvular disease | 1.18 (0.83, 1.69) | 1.18 (0.82, 1.69) | |
| Pulmonary circulation disease | 0.77 (0.39, 1.52) | 0.79 (0.40, 1.56) | |
| Peripheral vascular disease | 1.42 (1.01, 2.00)* | 1.42 (1.01, 2.00)* | |
| Other neurological disorders | 0.55 (0.34, 0.87)* | 0.56 (0.35, 0.89)* | |
| Diabetes without chronic complications |
1.34 (1.00, 1.78) | 1.35 (1.01, 1.80)* | |
| Diabetes with chronic complications |
1.08 (0.63, 1.82) | 1.08 (0.63, 1.83) | |
| Coagulopathy | 1.57 (0.89, 2.76) | 1.59 (0.91, 2.80) | |
| Hypertension | 1.22 (0.99, 1.50) | 1.23 (1.00, 1.52) | |
| Alcohol consumption | |||
| 1 drink/day | 1.04 (0.83, 1.31) | ||
| 2 drinks/day | 1.06 (0.78, 1.44) | ||
| 3 drinks/day | 0.73 (0.39, 1.34) | ||
| ≥ 4 drinks/day | 1.60 (0.98, 2.62) | ||
| Smoking | |||
| Former smoker | 0.93 (0.79, 1.10) | ||
| Current smoker | 1.15 (0.89, 1.48) | ||
Abbreviation: CIND = cognitive impairment without dementia.
P < 0.05
P < 0.01
P < 0.001
The comparison group for analyses of the association of depression, CIND or dementia with odds of ischemic stroke is those subjects with no disorder
The prevalence rates of ischemic stroke for all baseline characteristics found to be significantly associated with odds of stroke in the fully adjusted logistic regression model are available in Table S1 (Supplemental Digital Content 1).
Analyses of Effect Modification
There were significant interactions between baseline depression, CIND or dementia status with female sex and dual Medicare-Medicaid status (Table 3). In our cohort, the odds of ischemic stroke were significantly greater among women with CIND alone at baseline (OR: 1.63, 95%CI: 1.08, 2.48). Dual Medicare-Medicaid status among participants with CIND alone (OR: 1.75, 95%CI: 1.04, 2.94) or co-occurring depression and dementia (OR: 2.19, 95%CI: 1.06, 4.54) was associated with significantly greater odds of ischemic stroke.
Table 3.
Analyses of effect modification of the association of depression, CIND and dementia with odds of ischemic stroke
|
Odds Ratio
(95% Confidence Interval) |
|
|---|---|
| Depression/CIND/Dementia × Sex (n = 6,969) | |
| Depression alone × Female | 1.34 (0.80, 2.24) |
| CIND alone × Female | 1.63 (1.08, 2.48)* |
| Dementia alone × Female | 1.60 (0.91, 2.80) |
| Depression + CIND × Female | 1.27 (0.71, 2.26) |
| Depression + Dementia × Female | 1.83 (0.89, 3.74) |
|
Depression/CIND/Dementia × Dual Medicare-Medicaid
status (n = 6,969) |
|
| Depression alone × Dual Medicare-Medicaid | 1.43 (0.79, 2.58) |
| CIND alone × Dual Meaicare-Medicaid | 1.75 (1.04, 2.94)* |
| Dementia alone × Dual Medicare-Medicaid | 1.43 (0.78, 2.62) |
| Depression + CIND × Dual Medicare-Medicaid | 1.24 (0.67, 2.30) |
| Depression + Dementia x Dual Medicare-Medicaid | 2.19 (1.06, 4.54)* |
|
Depression/CIND/Dementia × Baseline cerebrovascular
disease (n = 6,969) | |
| Depression alone × Prior cerebrovascular disease | 0.97 (0.35, 2.67) |
| CIND alone × Prior cerebrovascular disease | 2.13 (0.91, 5.01) |
| Dementia alone × Prior cerebrovascular disease | 2.43 (1.08, 5.45)* |
| Depression + CIND × Prior cerebrovascular disease | 1.13 (0.42, 3.04) |
| Depression + Dementia × Prior cerebrovascular disease | 1.09 (0.43, 2.79) |
| Depression/CIND/Dementia × Baseline diabetes (n = 6,969) | |
| Depression alone × Baseline diabetes | 0.73 (0.30, 1.75) |
| CIND alone × Baseline diabetes | 1.02 (0.47, 2.22) |
| Dementia alone × Baseline diabetes | 1.50 (0.65, 3.44) |
| Depression + CIND × Baseline diabetes | 1.79 (0.81, 3.96) |
| Depression + Dementia × Baseline diabetes | 0.77 (0.30, 2.01) |
|
Depression/CIND/Dementia × Baseline hypertension
(n = 6,969) |
|
| Depression alone × Baseline hypertension | 0.76 (0.43, 1.37) |
| CIND alone × Baseline hypertension | 1.27 (0.76, 2.13) |
| Dementia alone × Baseline hypertension | 1.56 (0.85, 2.84) |
| Depression + CIND × Baseline hypertension | 0.91 (0.49, 1.68) |
| Depression + Dementia × Baseline hypertension | 1.46 (0.72, 2.96) |
Abbreviation: CIND = cognitive impairment without dementia
P < 0.05
In our cohort, cerebrovascular disease at baseline modified the association of dementia alone and odds of ischemic stroke (OR: 2.43, 95%CI: 1.08, 5.45). There appeared to be an interaction between CIND alone and baseline cerebrovascular disease which neared statistical significance (OR: 2.13, 95%CI: 0.91, 5.01). There was no evidence of effect modification by diabetes or hypertension.
Sensitivity Analyses
When we excluded individuals with cerebrovascular disease at baseline, CIND alone (OR: 1.29, 95%CI: 1.03, 1.60) and co-occurring depression and CIND (OR: 1.59, 95%CI: 1.18, 2.14) remained independently associated with increased odds of having an ischemic stroke.
Limiting the definitions of baseline depression and baseline dementia to a CES-D-8 score ≥ 4 alone and a TICSm score ≤ 6 alone, respectively, resulted in somewhat attenuated associations between CIND alone (OR: 1.24, 95%CI: 1.00, 1.52) and co-occurring depression and CIND (OR: 1.34, 95%CI: 0.98, 1.83) with odds of ischemic stroke.
When we used a CES-D-8 cutoff score of ≥ 3 in our depression definition, co-occurring depression and CIND continued to be independently associated with increased odds of ischemic stroke (OR: 1.51, 95%CI: 1.13, 2.02), while the association of CIND alone with odds of ischemic stroke was attenuated slightly (OR: 1.23, 95%CI: 0.95, 1.61). Depression alone (unadjusted OR: 1.03, 95%CI: 0.85, 1.26), dementia alone (OR adjusted for demographics: 1.11, 95%CI: 0.84, 1.46) and co-occurring depression and dementia (unadjusted OR: 1.21, 95%CI: 0.95, 1.54) continued to not have significant associations with odds of ischemic stroke.
Finally, CIND alone (OR: 1.35, 95%CI: 1.10, 1.67) as well as co-occurring depression and CIND (OR: 1.63, 95%CI: 1.22, 2.17) remained independently associated with increased odds of ischemic stroke after propensity score adjustment to account for death during follow-up.
DISCUSSION
In this nationally representative cohort of Americans over age 50, we identified a complex relationship between depression, CIND and dementia with subsequent risk of stroke. We found that CIND alone was independently associated with 37% greater odds of ischemic stroke, and co-occurring depression and CIND was independently associated with 65% greater odds of ischemic stroke. However, contrary to our original hypothesis, as well as the findings of prior studies (8-18), depression and dementia were not individually associated with increased odds of ischemic stroke in our entire cohort. The discordance between our findings and those of previous studies could be due to methodological differences such as our use of both self-report and administrative data versus either one alone, different measures used to ascertain depression and cognitive impairment across studies, or that many studies (8, 10, 13-16, 18) did not account for pre-existing cognitive impairment in examining depression as a risk factor and vice versa. Nevertheless, our results regarding the association of co-occurring depression and CIND with ischemic stroke risk add to existing research identifying individuals with co-occurring depression and cognitive impairment as a uniquely at-risk population for a wide-range of adverse outcomes including cognitive decline, institutionalization, and mortality (31, 32).
Although dementia alone and co-occurring depression and dementia were not independently associated with risk of ischemic stroke in our entire cohort, we did identify significant associations among relevant sub-groups. We found that dual Medicare-Medicaid status modified the association between CIND alone as well as co-occurring depression and dementia with risk of ischemic stroke. To the extent that dual Medicare-Medicaid status is an indication of lower socioeconomic status, particularly among the 45% of dual Medicare-Medicaid eligible individuals with co-occurring depression and dementia in our cohort, a possible explanation could be reduced adherence to preventive therapies due to financial concerns (33). Furthermore, prior work has identified that lower socioeconomic status may modify the impact of psychological distress on risk of stroke-related mortality (34). In addition, female sex modified the association of CIND alone with ischemic stroke risk. Building upon prior work that identified women at increased stroke risk compared to men (35), our data suggest that cognitively impaired women may be a sub-group at elevated risk. Overall, our results reinforce the findings of prior studies highlighting demographic factors that may identify important sub-groups of patients at risk for stroke that warrant prevention efforts including risk factor modification (14, 36).
Co-occurring depression and cognitive impairment may be associated with increased risk of ischemic stroke through several mechanisms. A biologically plausible mediator could be endothelial dysfunction and impaired regulation of coagulation pathways. Mild cognitive impairment and depression have been associated with fibrinogen dysregulation and exaggerated platelet reactivity (37, 38), both of which could predispose to vascular pathology such as ischemic stroke (39). A further mediator of the associations presented here could be functional disability due to major depression and/or cognitive dysfunction (40, 41). Our study suggests that additional research is warranted to elucidate the pathophysiological mechanisms that may link co-occurring depression and cognitive impairment with risk of stroke.
An important implication of identifying that co-occurring depression in older adults with CIND is independently associated with increased risk of ischemic stroke is that depression in cognitively impaired patients is treatable. Evidence-based treatments exist for older adults with co-occurring depression and cognitive impairment (42), and collaborative care interventions have been shown to increase access to these treatments, reduce depressive symptoms, and improve chronic medical illness management (43, 44). Further research is needed to examine if these interventions could aid in the prevention of incident strokes.
Our study has several limitations. Our baseline assessments cannot exclude the possibility that CIND in some participants may have been indicative of a transient process such as delirium. Also, depression status may have changed over the course of follow-up, raising the possibility that other more proximate factors may have played a greater role in conveying risk for ischemic stroke than co-occurring depression and CIND. However, it is important to note that prior work has established that depression in older adults with medical illnesses is frequently chronic (45), and rates of depression among HRS participants have been found to be fairly stable over time (5). Furthermore, the probable effect of dementia as a risk factor for ischemic stroke is likely greater than presented here since 10-20% of older adults with CIND progress to dementia annually (6). An additional limitation is that we lacked information on other possible risk factors for ischemic stroke such as atrial fibrillation, as well as data on treatment for depression in our cohort to be able to infer whether appropriate therapies could modify the association of co-occurring depression and CIND with risk of ischemic stroke presented here. A further limitation of our study is that we did not have data on time from our exposures of interest to ischemic stroke occurrence and therefore could not use Cox proportional hazards regression models to estimate hazard ratios for ischemic stroke.
Furthermore, we acknowledge that our assessments of baseline depression, CIND or dementia could be subject to misclassification bias since diagnosing depression in older adults with cognitive impairment is difficult, and cognitive impairment can be a sign of depression itself. Our baseline assessments also preclude consideration of depressive symptom severity as a potential mediator. Although the CES-D-8 has not been specifically validated in cognitively impaired patients nor has the TICSm been specifically validated in depressed individuals, both have been used previously in other relevant studies with high rates of these disorders (22, 46). While the version of the TICSm used in the HRS has modest discriminatory capabilities to identify dementia, our dementia definition was based on the TICSm threshold and also supplemented with ICD-9-CM dementia diagnoses. Finally, residual confounding remains a possibility, as in any observational study.
In conclusion, using a nationally representative sample of Americans over age 50, we have found that CIND as well as co-occurring depression and CIND are independently associated with increased odds of ischemic stroke. We also identified that demographic and socioeconomic characteristics such as sex and dual Medicare-Medicaid status may modify the association of depression, CIND and dementia with risk of ischemic stroke. Further research that elucidates the pathophysiology underlying the bi-directional relationship between depression, CIND and dementia with stroke, as well as develops cost-effective interventions preventing stroke that take into account depression and cognitive impairment, is needed in light of the adverse consequences of strokes for patients, their loved ones and society.
Supplementary Material
ACKNOWLEDGEMENTS
The HRS is performed at the Institute for Social Research, University of Michigan. We appreciate the expert programming of Laetitia Shapiro at the University of Michigan.
Sources of Funding
This work was supported by grants KL2 TR000421, R01 AG030155, and U01 AG09740 from the National Institutes of Health.
Abbreviations
- CES-D-8
8-item Center for Epidemiologic Studies Depression Scale
- CIND
cognitive impairment without dementia
- HRS
Health and Retirement Study
- ICD-9-CM
International Classification of Disease, Ninth Revision, Clinical Modification
- OR
odds ratio
- SD
standard deviations
- TICSm
modified Telephone Interview for Cognitive Status
- 95%CI
95% Confidence Interval
Footnotes
Publisher's Disclaimer: Disclaimer: The views expressed in this article are those of the authors and do not necessarily reflect the position or policy of the Department of Veterans Affairs, the National Institutes of Health, or the US government.
Conflicts of Interest: Dr. Katon has received honorariums for CME lectures from Eli Lilly, Forest, and Pfizer pharmaceutical companies. Drs. Davydow, Levine, Zivin and Langa have no relevant potential conflicts of interest to disclose.
REFERENCES
- 1.National Heart Lung and Blood Institute . Morbidity and Mortality: 2012 Chart book on cardiovascular, lung and blood diseases. National Institutes of Health; Bethesda, MD: 2012. [Google Scholar]
- 2.U.S. Burden of Disease Collaborators The state of U.S. health, 1990-2010: burden of diseases, injuries, and risk factors. JAMA. 2013;310:591–608. doi: 10.1001/jama.2013.13805. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Desmond DW, Moroney JT, Sano M, Stern Y. Incidence of dementia after ischemic stroke: results of a longitudinal study. Stroke. 2002;33:2254–2260. doi: 10.1161/01.str.0000028235.91778.95. [DOI] [PubMed] [Google Scholar]
- 4.Ayerbe L, Ayis S, Wolfe CD, Rudd AG. Natural history, predictors and outcomes of depression after stroke: systematic review and meta-analysis. Br J Psychiatry. 2013;202:14–21. doi: 10.1192/bjp.bp.111.107664. [DOI] [PubMed] [Google Scholar]
- 5.Zivin K, Pirraglia PA, McCammon RJ, Langa KM, Vijan S. Trends in depressive symptom burden among older adults in the United States from 1998 to 2008. J Gen Intern Med. 2013;28:1611–1619. doi: 10.1007/s11606-013-2533-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Plassman BL, Langa KM, Fisher GG, Heeringa SG, Weir DR, Ofstedal MB, Burke JR, Hurd MD, Potter GG, Rodgers WL, Steffens DC, McArdle JJ, Willis RJ, Wallace RB. Prevalence of cognitive impairment without dementia in the United States. Ann Intern Med. 2008;148:427–434. doi: 10.7326/0003-4819-148-6-200803180-00005. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Hurd MD, Martorell P, Delavande A, Mullen KJ, Langa KM. Monetary costs of dementia in the United States. N Engl J Med. 2013;368:1326–1334. doi: 10.1056/NEJMsa1204629. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Arbelaez JJ, Ariyo AA, Crum RM, Fried LP, Ford DE. Depressive symptoms, inflammation, and ischemic stroke in older adults: a prospective analysis in the Cardiovascular Health Study. J Am Geriatr Soc. 2007;55:1825–1830. doi: 10.1111/j.1532-5415.2007.01393.x. [DOI] [PubMed] [Google Scholar]
- 9.Glymour MM, Maselko J, Gilman SE, Patton KK, Avendaño M. Depressive symptoms predict incident stroke independently of memory impairments. Neurology. 2010;75:2063–2070. doi: 10.1212/WNL.0b013e318200d70e. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Pan A, Okereke OI, Sun Q, Logroscino G, Manson JE, Willett WC, Ascherio A, Hu FB, Rexrode KM. Depression and incident stroke in women. Stroke. 2011;42:2770–2775. doi: 10.1161/STROKEAHA.111.617043. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Pan A, Sun Q, Okereke OI, Rexrode KM, Hu FB. Depression and risk of stroke morbidity and mortality: a meta-analysis and systematic review. JAMA. 2011;306:1241–1249. doi: 10.1001/jama.2011.1282. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Dong JY, Yong-Hong Z, Tong J, Qin LQ. Depression and risk of stroke: a meta-analysis of prospective studies. Stroke. 2012;43:32–37. doi: 10.1161/STROKEAHA.111.630871. [DOI] [PubMed] [Google Scholar]
- 13.Li CT, Bei YM, Tu PC, Lee YC, Huang YL, Chen TJ, Chang WH, Su TP. Major depressive disorder and stroke risks: a 9-year follow-up population-based, matched cohort study. PLoS One. 2012;7:e46818. doi: 10.1371/journal.pone.0046818. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Seifert CL, Poppert H, Sander D, Feurer R, Etgen T, Ander KH, Pürner K, Brönner M, Sepp D, Kehl V, Förstl H, Bickel H. Depressive symptoms and the risk of ischemic stroke in the elderly—influence of age and sex. PLoS One. 2012;7:e50803. doi: 10.1371/journal.pone.0050803. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Tolppanen AM, Lavikainen P, Solomon A, Kivipelto M, Soininen H, Hartikainen S. Incidence of stroke in people with Alzheimer disease: a national-register based approach. Neurology. 2013;80:353–358. doi: 10.1212/WNL.0b013e31827f08c5. [DOI] [PubMed] [Google Scholar]
- 16.Chi NF, Chien LN, Ku HL, Hu CJ, Chiou HY. Alzheimer disease and risk of stroke: a population-based cohort study. Neurology. 2013;80:705–711. doi: 10.1212/WNL.0b013e31828250af. [DOI] [PubMed] [Google Scholar]
- 17.Imfeld P, Bodmer M, Schuerch M, Jick SS, Meier CR. Risk of incident stroke in patients with Alzheimer disease or vascular dementia. Neurology. 2013;81:910–919. doi: 10.1212/WNL.0b013e3182a35151. [DOI] [PubMed] [Google Scholar]
- 18.Jackson CA, Mishra GD. Depression and risk of stroke in midaged women: a prospective longitudinal study. Stroke. 2013;44:1555–1560. doi: 10.1161/STROKEAHA.113.001147. [DOI] [PubMed] [Google Scholar]
- 19.Lyketsos CG, Lopez O, Jones B, Fitzpatrick AL, Breitner J, DeKosky S. Prevalence of neuropsychiatric symptoms in dementia and mild cognitive impairment: results from the Cardiovascular Health Study. JAMA. 2002;288:1475–1483. doi: 10.1001/jama.288.12.1475. [DOI] [PubMed] [Google Scholar]
- 20.Health and Retirement Study: Sample sizes and response rates. http://hrsonline.isr.umich.edu/sitedocs/sampleresponse.pdf. Accessed July 13, 2014.
- 21.Steffick DE. Documentation of affective functioning measures in the Health and Retirement Study. Survey Research Center; Ann Arbor, Michigan: 2000. [Google Scholar]
- 22.Davydow DS, Hough CL, Levine DA, Langa KM, Iwashyna TJ. Functional disability, cognitive impairment, and depression after hospitalization for pneumonia. Am J Med. 2013;126:615–624. doi: 10.1016/j.amjmed.2012.12.006. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Crimmins EM, Kim JK, Langa KM, Weir DR. Assessment of cognition using surveys and neuropsychological assessment: The Health and Retirement Study and the Aging, Demographics, and Memory Study. J Gerontol B Psychol Sci Soc Sci. 2011;66(Suppl 1):i162–i171. doi: 10.1093/geronb/gbr048. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Elixhauser A, Steiner C, Harris DR, Coffey RM. Comorbidity measures for use with administrative data. Med Care. 1998;36:8–27. doi: 10.1097/00005650-199801000-00004. [DOI] [PubMed] [Google Scholar]
- 25.Charlson ME, Pompei P, Ales KL. A new method for classifying prognostic comorbidity in longitudinal studies: development and validation. J Chron Dis. 1987;40:373–383. doi: 10.1016/0021-9681(87)90171-8. [DOI] [PubMed] [Google Scholar]
- 26.Katon WJ. Clinical and health services relationships between major depression, depressive symptoms, and general medical illness. Biol Psychiatry. 2003;54:216–226. doi: 10.1016/s0006-3223(03)00273-7. [DOI] [PubMed] [Google Scholar]
- 27.Gorelick PB. Risk factors for vascular dementia and Alzheimer’s disease. Stroke. 2004;25:2620–2622. doi: 10.1161/01.STR.0000143318.70292.47. [DOI] [PubMed] [Google Scholar]
- 28.Greenlund KJ, Croft JB, Mensah GA. Prevalence of heart disease and stroke risk factors in persons with prehypertension in the United States, 1999-2000. Arch Intern Med. 2004;164:2113–2118. doi: 10.1001/archinte.164.19.2113. [DOI] [PubMed] [Google Scholar]
- 29.Roumie CL, Mitchel E, Gideon PS, Varas-Lorenzo C, Castellsague J, Griffin MR. Validation of ICD-9 codes with a high positive predictive value for incident strokes resulting in hospitalization using Medicaid health data. Pharmacoepidemiol Drug Safe. 2008;17:20–26. doi: 10.1002/pds.1518. [DOI] [PubMed] [Google Scholar]
- 30.Rao RS, Sigurdson AJ, Doody MM, Graubard BI. An application of a weighting method to adjust for nonresponse in standardized incidence ratio analysis of cohort studies. Ann Epidemiol. 2005;15:129–136. doi: 10.1016/j.annepidem.2004.05.007. [DOI] [PubMed] [Google Scholar]
- 31.Rosenberg PB, Mielke MM, Appleby BS, Oh ES, Geda YE, Lyketsos CG. The association of neuropsychiatric symptoms in MCI with incident dementia and Alzheimer disease. Am J Geriatr Psychiatry. 2013;21:685–695. doi: 10.1016/j.jagp.2013.01.006. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Okura T, Plassman BL, Steffens DC, Llewellyn DJ, Potter GG, Langa KM. Neuropsychiatric symptoms and the risk of institutionalization and death: the Aging, Demographics, and Memory Study. J Am Geriatr Soc. 2011;59:473–481. doi: 10.1111/j.1532-5415.2011.03314.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Levine DA, Morgenstern LB, Langa KM, Piette JD, Rogers MA, Karve SJ. Recent trends in cost-related medication nonadherence among stroke survivors in the United States. Ann Neurol. 2013;73:180–188. doi: 10.1002/ana.23823. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Lazzarino AI, Hamer M, Stamatakis E, Steptoe A. Low socioeconomic status and psychological distress as synergistic predictors of mortality from stroke and coronary heart disease. Psychosom Med. 2013;75:311–316. doi: 10.1097/PSY.0b013e3182898e6d. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Tsadok MA, Jackevicius CA, Rahme E, Humphries KH, Behlouli H, Pilote L. Sex differences in stroke risk among older patients with recently diagnosed atrial fibrillation. JAMA. 2012;307:1952–1958. doi: 10.1001/jama.2012.3490. [DOI] [PubMed] [Google Scholar]
- 36.Levine DA, Langa KM, Rogers MA. Acute infection contributes to racial disparities in stroke mortality. Neurology. 2014;82:914–921. doi: 10.1212/WNL.0000000000000214. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Song F, Poljak A, Kochan NA, Raftery M, Brodaty H, Smythe GA, Sachdev PS. Plasma protein profiling of mild cognitive impairment and Alzheimer’s disease using iTRAQ quantitative proteomics. Proteome Sci. 2014;12:5. doi: 10.1186/1477-5956-12-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Musselman DL, Tomer A, Manatunga AK, Knight BT, Porter MR, Kasey S, Marzec U, Harker LA, Nemeroff CB. Exaggerated platelet reactivity in major depression. Am J Psychiatry. 1996;153:1313–1317. doi: 10.1176/ajp.153.10.1313. [DOI] [PubMed] [Google Scholar]
- 39.Fitzgerald DJ. Vascular biology of thrombosis: the role of platelet-vessel wall adhesion. Neurology. 2001;57(Suppl 2):S1–S4. doi: 10.1212/wnl.57.suppl_2.s1. [DOI] [PubMed] [Google Scholar]
- 40.Bruce ML. Depression and disability in late life: directions for future research. Am J Geriatr Psychiatry. 2001;9:102–112. [PubMed] [Google Scholar]
- 41.Dodge HH, Kadowaki T, Hayakawa T, Yamakawa M, Sekikawa A, Ueshima H. Cognitive impairment as a strong predictor of incident disability in specific ADL-IADL tasks among community-dwelling elders: the Azuchi Study. Gerontologist. 2005;45:222–230. doi: 10.1093/geront/45.2.222. [DOI] [PubMed] [Google Scholar]
- 42.Orgeta V, Qazi A, Spector AE, Orrell M. Psychological treatment for depression and anxiety in dementia and mild cognitive impairment. Cochrane Database Sys Rev. 2014;1 doi: 10.1002/14651858.CD009125.pub2. CD009125. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43.Unützer J, Katon W, Callahan CM, Williams JW, Jr., Hunkeler E, Harpole L, Hoffing M, Della Penna RD, Noël PH, Lin EH, Areán PA, Hegel MT, Tang L, Belin TR, Oishi S, Langston C, IMPACT Investigators Improving Mood-Promoting Access to Collaborative Treatment. Collaborative care management of late-life depression in the primary care setting: a randomized controlled trial. JAMA. 2002;288:2836–2845. doi: 10.1001/jama.288.22.2836. [DOI] [PubMed] [Google Scholar]
- 44.Katon WJ, Lin EH, Von Korff M, Ciechanowski P, Ludman EJ, Young B, Peterson D, Rutter CM, McGregor M, McCulloch D. Collaborative care for patients with depression and chronic illnesses. N Engl J Med. 2010;363:2611–2620. doi: 10.1056/NEJMoa1003955. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 45.Katon WJ, Von Korff M, Lin EH, Simon G, Ludman E, Russo J, Ciechanowski P, Walker E, Bush T. The Pathways Study: a randomized trial of collaborative care in patients with diabetes and depression. Arch Gen Psychiatry. 2004;61:1042–1049. doi: 10.1001/archpsyc.61.10.1042. [DOI] [PubMed] [Google Scholar]
- 46.Iwashyna TJ, Ely EW, Smith DM, Langa KM. Long-term cognitive impairment and functional disability among survivors of severe sepsis. JAMA. 2010;304:1787–1794. doi: 10.1001/jama.2010.1553. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.