Abstract
Background
HIV infection and depression are each associated with increased ischemic stroke risk. Whether depression is a risk factor for stroke within the HIV population is unknown.
Methods and Results
We analyzed data on 106 333 (33 528 HIV‐positive; 72 805 HIV‐negative) people who were free of baseline cardiovascular disease from an observational cohort of HIV‐positive people and matched uninfected veterans in care from April 1, 2003 through December 31, 2014. International Classification of Diseases, Ninth Revision (ICD‐9) codes from medical records were used to determine baseline depression and incident stroke. Depression occurred in 19.5% of HIV‐positive people. After a median of 9.2 years of follow‐up, stroke rates were highest among people with both HIV and depression and lowest among those with neither condition. In Cox proportional hazard models, depression was associated with an increased risk of stroke for HIV‐positive people after adjusting for sociodemographic characteristics and cerebrovascular risk factors (hazard ratio [HR], 1.18; 95% CI: 1.03–1.34; 0.014). The depression‐stroke relationship was attenuated by alcohol use disorders, cocaine use, and baseline antidepressant use, and unaffected by combined antiretroviral therapy use or individual antiretroviral agents. A numerically higher HR of depression on stroke was found among those younger than 60 years.
Conclusions
Depression is associated with an increased risk of stroke among HIV‐positive people after adjusting for sociodemographic characteristics, traditional cerebrovascular risk factors, and HIV‐specific factors. Alcohol use disorders, cocaine use, and baseline antidepressant use accounted for some of the observed stroke risk. Depression may be a novel, independent risk factor for ischemic stroke in HIV, particularly among younger people.
Keywords: combined antiretroviral therapy, depression, HIV, ischemic stroke, stroke risk
Subject Categories: Ischemic Stroke
Nonstandard Abbreviations and Acronyms
- cART
combined antiretroviral therapy
- VA
Veterans Affairs
- VACS
Veterans Aging Cohort Study
- VIReC
VA Information Resource Center
Clinical Perspective
What Is New?
Depression is associated with an increased risk of incident ischemic stroke among veterans living with HIV.
What Are the Clinical Implications?
Understanding the possible causes by which depression increases the risk of stroke among veterans may help identify opportunities to mitigate stroke risk.
Whether the results are generalizable to other populations remains unknown.
Given impressive improvement in life expectancy in the era of combined antiretroviral therapy (cART), HIV‐positive people are living long enough to develop vascular risk factors associated with aging (eg, hypertension and atrial fibrillation).1, 2, 3, 4, 5 This improved survival also comes with the increased risk of developing measures of subclinical cardiovascular disease (CVD), which tend to be more unfavorable compared with healthy comparators (eg, greater carotid intima‐media thickness, greater coronary artery calcification, faster pulse‐wave velocity, and poorer endothelial function),6 and the potential sequelae of living longer with a chronic viral infection, including an HIV‐associated chronic inflammatory state. Not surprisingly, being HIV‐positive has been independently associated with incident CVD events, including acute myocardial infarction (AMI), congestive heart failure (CHF), peripheral arterial disease, and ischemic stroke.7, 8, 9, 10, 11, 12, 13
Many cerebrovascular risk factors are observed less frequently among HIV‐positive versus HIV‐negative people (eg, hypertension, diabetes mellitus), suggesting that other factors may increase stroke risk in this patient population.11 Other potential mechanisms for elevated cerebrovascular risk among HIV‐positive patients include direct damage to the arterial wall from HIV viral particles, arterial damage due to chronic inflammation, absence of repair due to loss of CD4+ cells, or secondary effects due to cART, including cART‐induced metabolic syndrome14 and dysfunctional endothelium and platelets.13, 15, 16, 17 One potentially modifiable stroke risk factor that has not previously been examined among HIV‐positive people is depression.18
Depression, one of the most common comorbidities of HIV infection, is predictive of incident AMI and CHF among HIV‐positive people.12, 13 Depression is accompanied by many vascular and systemic changes which are hypothesized to underlie the observed relationship between depression and vascular disease. Depression is associated with subclinical measures of vascular disease among HIV‐positive people.19 Moreover, systemic changes occurring in response to chronic stress are altered in depression, including growth factors, inflammatory markers, endocrine markers, and metabolic markers.20 In addition to physiological changes, depression is associated with maladaptive behavioral changes, such as decreased treatment adherence, physical inactivity, poor dietary habits, and cigarette smoking, all of which can contribute to vascular risk.18, 19, 21, 22, 23 Some cART regimens have been linked with depression.24
Data regarding the potential relationship between depression and stroke risk in HIV are nonexistent. Studies conducted in the general population have demonstrated an independent association between depression and future ischemic stroke.18, 21, 22, 23 The presence of stable high depressive symptoms was predictive of incident stroke among younger but not older people, suggesting a moderating effect of age on depression‐stroke risk relationship in the general population.18, 23
Although HIV and depression share physiological pathways that increase stroke risk and despite depression having been shown to increase AMI and CHF risk in HIV, current literature lacks observational studies examining whether depression similarly increases ischemic stroke risk in people living with HIV. To address this knowledge gap, we tested the prospective association between depression and incident ischemic stroke in HIV‐positive people from the VACS (Veterans Aging Cohort Study), while exploring potential influences of age, antidepressant medication, and cART use on the depression‐stroke risk relationship.
Methods
Data Availability Statement
Due to Veterans Affairs (VA) regulations and our ethics agreements, the analytic data sets used for this study are not permitted to leave the VA firewall without a Data Use Agreement. This limitation is consistent with other studies based on VA data. However, VA data are made freely available to researchers with an approved VA study protocol. For more information, please visit https://www.virec.research.va.gov or contact the VA Information Resource Center at VIReC@va.gov.
Study Sample
The VACS virtual cohort is a prospective, multisite cohort of HIV‐positive adults and age, race/ethnicity, and clinical site matched to 2 HIV‐negative adults enrolled in the same calendar year in the US Department of VA system.25 Participants in VACS have been continually selected for inclusion since 1998 by using an existing validated algorithm from the VA national electronic medical record system. Baseline was defined as the date of a participant's first clinic visit on or after April 1, 2003. The participants were followed up from their baseline date until an ischemic stroke event, death, the date of last follow‐up, or censored on December 31, 2014. The University of Pittsburgh, Yale University, and West Haven VA Medical Center institutional review boards approved this study. Subject informed consent was waived.
For the current analysis, participants with prevalent CVD (ie, coronary heart disease, AMI, cardiovascular revascularization, heart failure, or stroke [ischemic/hemorrhagic]) at baseline (n=22 712) and those who seroconverted during the follow‐up period (n=589) were excluded. The final analytic sample consisted of 106 333 veterans (33 528 HIV‐positive; 72 805 HIV‐negative).
Independent and Stratifying Variables
Depression at baseline was defined as a diagnosis of major depressive disorder (at least one inpatient or 2 outpatient International Classification of Diseases, Ninth Revision [ICD‐9] codes 296.2 or 296.3) or dysthymic disorder (ICD‐9 code 300.4).12, 26 HIV was defined as at least one inpatient or 2 outpatient ICD‐9 codes for HIV in the VA Immunology Case Registry. Baseline HIV infection was defined as participants having ≥1 inpatient and/or ≥2 outpatient ICD‐9 codes for the diagnosis. The algorithm has a high sensitivity (90%), specificity (99.9%), and positive predictive value (88%) in identifying HIV‐positive participants.3 The HIV‐specific factors of HIV‐1 RNA level and CD4 cell count were obtained from the VA Corporate Data Warehouse, whereas cART use (yes/no) obtained through pharmacy data as part of clinical care within a window of 180 days before baseline through 7 days post baseline. Participants were categorized into 4 groups: HIV‐positive with depression, HIV‐positive without depression, HIV‐negative with depression, and HIV‐negative without depression (reference group).
Dependent Variable
The primary outcome of this study was incident ischemic stroke, which was defined as at least one inpatient or 2 outpatients ICD‐9 codes from medical records for any of the following conditions: occlusion and stenosis of precerebral arteries (433.x1), occlusion of cerebral arteries excluding thrombosis/embolism without infarction (434, excluding 434.x0), and acute but ill‐defined cerebrovascular disease (436).11
Covariates
Covariates were obtained closest to the baseline date and have been previously described.9, 27 Briefly, sociodemographic variables of age, sex, and race/ethnicity were determined through administrative data. Hypertension was categorized as "no hypertension" (<140/90 mm Hg and no antihypertensive medication), "controlled hypertension" (<140/90 mm Hg with antihypertensive medication), or "uncontrolled hypertension" (≥140/90 mm Hg).28 Blood pressure was defined by averaging 3 routine outpatient systolic blood pressure and diastolic blood pressure measurements. Low‐density lipoprotein cholesterol, high‐density lipoprotein (HDL) cholesterol, and triglycerides were obtained from laboratory data. Statin use was defined as a prescription receipt of a 3‐hydroxy‐3‐methyl‐glutaryl‐coenzyme A reducatase inhibitor from pharmacy records from the participant's baseline enrollment date.11 Diabetes mellitus (yes/no) was identified using a previously validated metric that incorporates glucose measurements, hemoglobin A1c, antidiabetic agent use, and/or at least one inpatient or 2 outpatient ICD‐9 codes for diabetes mellitus.29 Body mass index, calculated as kg/m2, was assessed using the VA Health Factor Dataset. Estimated glomerular filtration rate was calculated using the Chronic Kidney Disease Epidemiology Collaboration equation based on outpatient and clinical laboratory data.30 Hemoglobin levels were obtained from laboratory data. Histories of atrial fibrillation, alcohol use disorders (ie, alcohol abuse or dependence), and cocaine use were defined using ICD‐9 codes.9, 11 Smoking status was categorized as never, current, or past and obtained from the VA Health Factor Dataset.31 Hepatitis C virus seropositivity (yes/no) was defined as a positive Hepatitus C virus antibody test result or at least one inpatient or 2 outpatient ICD‐9 codes. Baseline antidepressant use was defined as documentation of a filled prescription for a selective serotonin uptake inhibitor, a tricyclic antidepressant, or another antidepressant type from VA pharmacy records within 180 days of baseline enrollment date.
Statistical Analysis
Baseline characteristics are presented in groups defined by depression status and HIV status. We report means (SD) for continuous variables and frequencies (percentage) for categorical variables. We constructed Kaplan‐Meier event‐free survival curves, including the number of patients at risk and the number of censored patients, and performed log‐rank tests to compare the various groups. We calculated incident rates per 1000 person‐years stratified by age category. Separately, we constructed Cox proportional hazards regression models to estimate the adjusted hazard ratio (aHR) and 95% CIs for the association between depression and incident ischemic stroke. Cox PH modeling was not used to calculated incident rates. We constructed several models stratified by HIV status: (1) Model One: adjusting sociodemographic factors (ie, age, sex, and race/ethnicity); (2) Model Two: Model One plus CVD risk factors (ie, SBP, DBP, LDL cholesterol, HDL cholesterol, triglycerides, statin use, diabetes mellitus, body mass index, smoking status, estimated glomerular filtration rate, hemoglobin, and Hepatitis C virus); (3) Model Three: Model Two plus atrial fibrillation; (4) Model Four: Model Three plus HIV‐specific factors (ie, HIV‐1 RNA level, CD4 cell count, and cART use); (5) Model Five: Model Three plus alcohol use disorders; (6) Model Six: Model Three plus cocaine use; (7) Model Seven: Model Three plus alcohol use disorders and cocaine use; (8) Model Eight: Model Three plus antidepressant medication use. Model Three served as the primary model for this analysis. Continuous predictors were modeled using restricted cubic splines with 3 knots to allow a nonlinear relationship between the variable and outcome. To investigate age as a potential moderator of the relationship between depression‐incident stroke, we included a depression x age interaction term in Models One, Two, and Three. We report the P value of interactions and present the data graphically.
The proportional hazards assumption was tested by the Schoenfeld residuals and including the interaction term of the covariates by time as evaluated by the “cox.zph” function in R. To handle missing values, multiple imputations using chained equations with 5 separate imputed data sets were generated based on predictive mean matching using the "mice" library of R programming language. Multiple imputations were made for missing values. Participants without an ischemic stroke or death were censored at the end of the study (see Online data supplement). Cox survival models were fitted in each imputed data set and then combined to obtain pooled hazard ratios and standard errors. Variation inflation factors were calculated to assess for multicollinearity between depression and antidepressant use variables in Model Eight. All analyses were performed using R software (version 3.3.3; www.r‐project.org).
Results
The prevalence of depression at baseline in our cohort was 19.0% overall and was similar among HIV‐positive and HIV‐negative people (19.5% versus 18.8%, respectively). Percentages of people in the 4 groups were 6.2% for HIV‐positive with depression, 25.4% for HIV‐positive without depression, 12.9% for HIV‐negative with depression, and 55.6% for HIV‐negative without depression (Table 1).
Table 1.
Baseline Characteristics of VACS Virtual Cohort, N=106 333
| Factor | HIV Positive (n=33 528) | HIV Negative* (n=72 805) | ||
|---|---|---|---|---|
| With Depression (n=6554) | Without Depression (n=26 974) | With Depression (n=13 713) | Without Depression (n=59 092) | |
| Age, y, mean (SD)† | 48.0 (8.5) | 48.4 (10.2) | 48.9 (7.9) | 49.3 (10.1) |
| Sex, male | 6246 (95.3) | 26 323 (97.6) | 13 079 (95.4) | 57 463 (97.2) |
| Race/ethnicity | ||||
| White | 2809 (42.8) | 10 204 (37.8) | 5612 (40.9) | 22 086 (37.4) |
| Black§ | 3017 (46.1) | 13 539 (50.2) | 6552 (47.8) | 29 394 (49.7) |
| Hispanic | 603 (9.2) | 2094 (7.8) | 1308 (9.5) | 5093 (8.6) |
| Other | 125 (1.9) | 1138 (4.2) | 241 (1.8) | 2529 (4.3) |
| Hypertension | ||||
| None | 2717 (41.5) | 13 097 (48.6) | 4195 (30.6) | 20 426 (34.6) |
| Controlled | 2344 (35.8) | 7185 (26.6) | 5471 (39.9) | 18 566 (31.4) |
| Uncontrolled | 1442 (22.0) | 6254 (23.2) | 3777 (27.5) | 16 958 (28.7) |
| SBP, mm Hg, median (Q1, Q3)† | 128.0 [119.0, 137.7] | 128.3 [119.0, 138.0] | 130.7 [121.7, 140.3] | 132.0 [123.0, 141.3] |
| DBP, mm Hg, median (Q1, Q3) | 78.0 [72.0, 84.3] | 78.0 [71.7, 84.3] | 79.3 [73.3, 85.7] | 79.3 [73.3, 85.7] |
| Hyperlipidemia | ||||
| LDL cholesterol, mg/dL | ||||
| <100 | 2537 (38.7) | 9993 (37.0) | 3655 (26.7) | 14 278 (24.2) |
| 100–129 | 1654 (25.2) | 6317 (23.4) | 3675 (26.8) | 15 034 (25.4) |
| 130–159 | 827 (12.6) | 3511 (13.0) | 2375 (17.3) | 10 145 (17.2) |
| ≥160 | 410 (6.3) | 1636 (6.1) | 1297 (9.5) | 5307 (9.0) |
| HDL cholesterol, mg/dL | ||||
| <40 | 2863 (43.7) | 11 024 (40.9) | 4450 (32.5) | 16 780 (28.4) |
| 40–59 | 2031 (31.0) | 8290 (30.7) | 5182 (37.8) | 21 633 (36.6) |
| ≥60 | 626 (9.6) | 2451 (9.1) | 1515 (11.0) | 6867 (11.6) |
| Triglyceride, mg/dL ≥150 | 2753 (42.0) | 9938 (36.8) | 4617 (33.7) | 16 640 (28.2) |
| Statin use | 994 (15.2) | 3646 (13.5) | 3474 (25.3) | 13 746 (23.3) |
| Diabetes mellitus | 721 (11.0) | 2413 (8.9) | 2253 (16.4) | 8800 (14.9) |
| BMI, kg/m2 † ≥30 | 1112 (17.0) | 3966 (14.7) | 5231 (38.1) | 21 739 (39.8) |
| Atrial fibrillation | 68 (1.0) | 234 (0.9) | 145 (1.1) | 557 (0.9) |
| Smoking‡ | ||||
| Current | 3146 (48.0) | 9717 (36.0) | 6322 (46.1) | 19 071 (32.3) |
| Former | 698 (10.6) | 2797 (10.4) | 1632 (11.9) | 6965 (11.8) |
| Never | 1058 (16.1) | 5373 (19.9) | 2544 (18.6) | 13 479 (22.8) |
| Substance use | ||||
| Alcohol use disorder | 3300 (50.4) | 5299 (19.6) | 7110 (51.8) | 12 372 (20.9) |
| Cocaine use | 2546 (38.8) | 3887 (14.4) | 4501 (32.8) | 6562 (11.1) |
| eGFR mL/min per 1.73 m2 <60 | 333 (5.1) | 1591 (5.9) | 510 (3.7) | 2413 (4.1) |
| Anemia (hemoglobin <12 g/dL)‡ | 719 (11.0) | 3198 (11.9) | 482 (3.5) | 1796 (3.0) |
| Hepatitis C | 2557 (39.0) | 7398 (27.4) | 2840 (20.7) | 6358 (10.8) |
| HIV specific factors | ||||
| HIV 1 RNA*, copies/mL ≥500 | 3133 (47.8) | 12 335 (45.7) | … | … |
| CD4 cell count*, mm3 | ||||
| <200 | 1183 (18.1) | 5525 (20.5) | … | … |
| 211–499 | 2330 (35.6) | 9457 (35.1) | … | … |
| ≥500 | 2074 (50.2) | 7635 (28.3) | … | … |
| cART* | 3288 (50.2) | 11 760 (43.6) | … | … |
| Integrase inhibitor use* | 42 (0.6) | 334 (1.2) | … | … |
| Efavirenz* | 900 (13.7) | 4442 (16.5) | … | … |
| Abacavir* | 1398 (21.3) | 4707 (17.5) | … | … |
| Antidepressant medication use | ||||
| SSRI | 4833 (73.7) | 5299 (19.6) | 10 136 (73.9) | 10 789 (18.3) |
| TCA | 1786 (27.3) | 3243 (12.0) | 3447 (25.1) | 6201 (10.5) |
| Miscellaneous | 4460 (68.1) | 5102 (18.9) | 9533 (69.5) | 11 027 (18.7) |
BMI indicates body mass index; cART, combined antiretroviral therapy; DBP, diastolic blood pressure; eGFR, estimated glomerular filtration rate; HDL, high‐density lipoprotein; LDL, low‐density lipoprotein; SBP, systolic blood pressure; SSRI, selective serotonin reuptake inhibitor; TCA, tricyclic antidepressant; and VACS, Veterans Aging Cohort Study.
Because HIV‐uninfected veterans do not have HIV‐specific biomarkers or antiretroviral therapy regimens, these cells contain a dashed line.
Data represent mean (SD) for continuous variables and n (%) for categorical variables.
The following variables include fewer than 106 333 patients because of missing data (n missing): hypertension (3901, 3.7 %), SBP (3901, 3.7%), DBP (3901, 3.7%), LDL cholesterol (23 682, 22.3%), HDL cholesterol (22 621, 21.3%), triglycerides (22 164, 20.8%), BMI (6568, 6.2%), smoking (33 531, 31.5%), eGFR (10 315, 9.7%), hemoglobin (10 734, 10.1%), HIV‐1 RNA (5299, 15.8% of HIV‐positive people), and CD4 cell count (5324, 15.9% of HIV‐positive people).
Other includes Indian, Black, Asian, mixed race, Hawaiian, and missing.
During a median of 9.2 (25th–75th percentile, 5.2–11.5) follow‐up years, there were 4355 incident stroke events and an overall stroke rate per 1000 person‐years of 5.0 (4.9–5.2). Incident stroke rates were highest among HIV‐positive participants with depression and lowest among HIV‐negative participants without depression (Table 2). Kaplan‐Meier event‐free survival curves in Figure 1 depict time to first incident ischemic stroke, with HIV‐positive people with depression having the poorest stroke‐free survival of all 4 groups (P value of log‐rank test: 0.001). It is worth noting that, while we detected statistically significant group differences, the absolute group differences in stroke‐free survival were modest.
Table 2.
Number and Incident Rates of Ischemic Stroke
| Age, y | Overall | HIV Positive | HIV Negative | ||
|---|---|---|---|---|---|
| With Depression | Without Depression | With Depression | Without Depression | ||
| Number of ischemic strokes | |||||
| 20–40 | 186/17 521 | 21/1069 | 66/4973 | 33/1738 | 66/9732 |
| 40–50 | 1220/39 578 | 133/2795 | 283/9630 | 199/5721 | 605/21 432 |
| 50–60 | 1924/37 129 | 123/2271 | 470/9026 | 303/5377 | 1028/20 455 |
| 60–96 | 1025/12 114 | 36/419 | 264/3345 | 79/877 | 646/7473 |
| Overall | 4355/106 333 | 313/6554 | 1083/26 974 | 614/13 713 | 2345/59 092 |
| Incident rate* | |||||
| 20–40 | 1.3 (1.2–1.5) | 2.3 (1.5–3.4) | 1.7 (1.3–2.1) | 2.1 (1.5–3.0) | 0.9 (0.7–1.1) |
| 40–50 | 3.6 (3.4–3.8) | 5.5 (4.6–6.5) | 3.6 (3.2–4.1) | 3.8 (3.3–4.3) | 3.2 (3.0–3.5) |
| 50–60 | 6.4 (6.1–6.7) | 7.0 (5.8–8.3) | 7.1 (6.5–7.8) | 6.5 (5.8–7.2) | 6.0 (5.6–6.4) |
| 60–96 | 12.1 (11.4–12.9) | 13.4 (9.5–18.2) | 12.4 (10.9–13.9) | 12.8 (10–15.6) | 11.9 (11–12.8) |
| Overall | 5.0 (4.9–5.2) | 5.8 (5.2–6.5) | 5.3 (5–5.6) | 5.1 (4.7–5.5) | 4.8 (4.6–5.0) |
Incident rate is per 1000 person‐years.
Figure 1. Unadjusted Kaplan‐Meier survival curves of ischemic stroke by depression and HIV status.
Dep indicates depressed; and NotDep, not depressed.
Cox models adjusted for sociodemographic factors demonstrated that HIV‐positive people with depression, compared with HIV‐positive people without depression, had a 22% higher risk of incident stroke (aHR, 1.22; 95% CI, 1.07–1.38; 0.003; Table 3, Model One). A similar statistically significant, though mildly attenuated, association persisted after further adjusting for cerebrovascular disease risk factors (aHR, 1.18; 95% CI, 1.03–1.34; 0.014, Model Three) and HIV‐specific factors (aHR, 1.18; 95% CI, 1.03–1.35; 0.014, Model Four). The association between depression and incident ischemic stroke in HIV‐positive people was modestly attenuated and fell short of statistical significance when alcohol use disorders and cocaine use were added to Model Three separately (Models Five and Six, respectively) and both within the same model (Model Seven). The association between depression and ischemic stroke risk in unadjusted and adjusted models showed similar effect sizes and degree of attenuation between HIV‐positive and HIV‐negative people.
Table 3.
| HIV Positive‡ | HIV Negative§ | |||
|---|---|---|---|---|
| HR [95% CI] | P Value | HR [95% CI] | P Value | |
| Model one: sociodemographic factors‖ | ||||
| With depression | 1.22 [1.07–1.38] | 0.003 | 1.15 [1.05–1.26] | 0.002 |
| Without depression | 1.0 | … | 1.0 | … |
| Model two: model one+CVD risk factors¶ | ||||
| With depression | 1.19 [1.04–1.35] | 0.010 | 1.11 [1.02–1.22] | 0.023 |
| Without depression | 1.0 | … | 1.0 | … |
| Model three: model two+atrial fibrillation# | ||||
| With depression | 1.18 [1.03–1.34] | 0.014 | 1.11 [1.01–1.22] | 0.026 |
| Without depression | 1.0 | … | 1.0 | … |
| Model four: model three+HIV‐specific factors** | ||||
| With depression | 1.18 [1.04–1.35] | 0.011 | … | … |
| Without depression | 1.0 | … | … | … |
| Model five: model three+alcohol use disorders†† | ||||
| With depression | 1.10 [0.96–1.25] | 0.180 | 1.04 [0.95–1.15] | 0.388 |
| Without depression | 1.0 | … | 1.0 | … |
| Model six: model three+cocaine use‡‡ | ||||
| With depression | 1.13 [0.99–1.30] | 0.066 | 1.08 [0.99–1.19] | 0.093 |
| Without depression | 1.0 | … | 1.0 | … |
| Model seven: model three+alcohol use disorders, and cocaine use§§ | ||||
| With depression | 1.10 [0.96–1.26] | 0.162 | 1.04 [0.95–1.15] | 0.367 |
| Without depression | 1.0 | … | 1.0 | … |
| Model eight: model three+antidepressant medication use variables‖‖ | ||||
| Depression | 1.16 [1.00–1.35] | 0.056 | 0.97 [0.87–1.08] | 0.607 |
| Without depression | 1.0 | … | 1.0 | … |
CVD indicates cardiovascular disease; and HR, hazard ratio.
All covariates were measured at baseline. In HIV‐positive people, N=33 528 (ischemic stroke, n=1396). In HIV‐negative people, N=72 805 (ischemic stroke, n=2959). For continuous predictors, restricted cubic splines with 3 knots were applied in order to allow a nonlinear relationship between the covariate and outcome. For variables with missing values multiple imputations with 5 imputed data sets were generated based on predictive mean matching method using "mice" library of R programming language.
P values of interaction between HIV status and depression: 0.503 (Model One), 0.460 (Model Two), 0.445 (Model Three).
P values of interaction between age and depression among HIV‐positive people: 0.190 (Model One), 0.175 (Model Two), and 0.145 (Model Three).
P values of interaction between age and depression among HIV‐negative people: 0.417 (Model One), 0.337 (Model Two), 0.333 (Model Three). As HIV‐uninfected people do not have HIV‐specific biomarkers or antiretroviral therapy regimens, these cells contain a dashed line.
Model one: adjusted for sociodemographic factors (ie, age, sex, and race/ethnicity).
Model two: adjusted for variables in Model One and CVD risk factors (ie, systolic blood pressure [SBP], diastolic blood pressure [DBP], low‐density lipoprotein [LDL] cholesterol, high‐density lipoprotein [HDL] cholesterol, triglycerides, statin use, diabetes mellitus, BMI, smoking status, estimated glomerular filtration rate [eGFR], hemoglobin, and hepatitis C infection).
Model three: adjusted for variables in model two and atrial fibrillation.
Model four: adjusted for variables in model three and HIV‐specific factors (ie, viral load, CD4 count, and antiretroviral therapy).
Model five: adjusted for variables in model three and alcohol use disorders.
Model six: adjusted for variables in model three and cocaine use.
Model seven: adjusted for variables in model three and alcohol use disorders, and cocaine use.
Model Eight: adjusted for variables in model three and the three antidepressant medication use variables (ie, selective serotonin uptake inhibitor [SSRI] use, tricyclic antidepressant [TCA] use, and miscellaneous antidepressant use). Variance inflation factors (VIF) were calculated to determine whether multicollinearity existed between the depression and antidepressant use variables. VIFs from Model 8: For HIV‐positive: depression: 1.5; SSRI: 1.4; TCA: 1.1; miscellaneous antidepressant: 1.4. For HIV‐negative: depression 1.4; SSRI: 1.5; TCA: 1.1; miscellaneous antidepressant: 1.5. As VIFs were <10, there was no evidence of multicollinearity.
In separate supplemental models, we also sought to explore the potential influences of age, antidepressant medication, and specific ARTs on the depression‐incident stroke relationship. We explored whether age moderated the relationship between depression and incident stroke by testing for interaction terms and using a graphical approach. Among HIV‐positive people, age was not a statistically significant moderator of the relationship between depression and stroke risk (Model One [P=0.190], Model Two [P=0.175], Model Three [P=0.145], Table 3, footnote§). However, when displayed graphically, a declining association between depression and incident stroke as age increases was evident. Moreover, the depression‐stroke risk association was significant for people younger than 60 years but not older (Figure 2).
Figure 2. Effect of depression on stroke risk across age strata.
Graphic depiction of the effect of depression on stroke risk across age strata to explore age as a potential moderator of the depression‐incidence stroke relationship.
When baseline antidepressant use was added to Model Three, the previously statistically significant association between depression and incident stroke in HIV‐positive people fell just short of significance (aHR, 1.16; 95% CI, 1.00–1.36; P=0.055; Model Eight, Table 3). Of note, there was no evidence of multicollinearity among the depression and antidepressant use variables in this model (see variation inflation factors in Table 3, footnote‖ ‖).
The addition of HIV specific factors to Model Three showed results similar to our primary model (aHR, 1.18; 95% CI, 1.04–1.35; P=0.011; Model Four, Table 3). To explore influence of specific ARTs having mechanisms that may confer increased cerebrovascular risk,15, 16, 17, 32 we reran Model Three and replaced all cART use with efavirenz and abacavir use. While the depression‐incident stroke relationship remained significant in these models, there was no statistically significant association between specific cART agents and stroke risk (Table S1).
Discussion
In a large sample of veterans with 9 years of follow‐up, the presence of a depressive disorder at baseline was associated with an 18% increased risk of ischemic stroke among HIV‐positive people after adjusting for sociodemographic characteristics, cerebrovascular disease risk factors, and HIV‐specific factors, with people living with both depression and HIV having the highest stroke risk compared with people with one or neither of these conditions. Adjustment for alcohol use disorders and cocaine use, both individually and in combination, attenuated the association between depression and stroke risk, suggesting that these factors account for some of the depression‐incident stroke relationship. In considering supplemental models, while the interaction terms between age and depression were not statistically significant, graphical displays suggest a declining association between depression and incident stroke as age increases. In addition, the depression‐stroke risk relationship was present in people younger than 60 years but not in those older than 60 years. Baseline antidepressant use slightly attenuated the relationship, whereas cART use did not alter the depression‐stroke relationship. The association between depression and ischemic stroke showed similar effect sizes between those with and without HIV were similar. Altogether, our findings suggest that: (1) depression may be a novel independent risk factor for ischemic stroke in people living with HIV; (2) alcohol use disorders, cocaine use, and antidepressant use may serve as areas of future research to determine prospectively how addressing substance use and treating depression and may mitigate stroke risk; (3) the depression‐incident stroke relationship may be more important in younger people; (4) cART and specific cART agents that theoretically may increase stroke risk are not associated with increased stroke risk, and; (5) other HIV specific factors, including HIV status, CD4 count, and viral load may not strongly influence the depression‐incident stroke association.
When considering the ke risk factors as hypertension, diabetese, there is pronounced need to effectively treat known risk factors and to identify novel and underappreciated contributors to stroke risk.5, 21 In the era of cART, HIV‐positive people are living longer2, 3 and are at increased risk of developing clinically important vascular events (eg, AMI, CHF, peripheral arterial disease, ischemic stroke),7, 9, 10, 11, 12, 13, 33 subclinical CVD (eg, endothelial dysfunction),6 and conditions associated with increased cerebrovascular risk (eg, hypertension).1, 5 HIV has previously been independently associated with ischemic stroke in VA‐based3 and community cohorts.10, 34 Ours is the first study to examine the association between depression, incident ischemic stroke, and HIV status in a large, national, contemporary cohort, while adjusting for sociodemographic characteristics, known cerebrovascular risk factors, HIV‐specific factors, alcohol and cocaine use, and antidepressant use.
Mechanisms underlying the association between depression and incident ischemic stroke among HIV‐positive people have not been thoroughly investigated though likely are multifactorial. Chronic HIV infection has been associated with a chronic, systemic inflammatory state, coagulopathies, platelet dysfunction, CD4 cell depletion with resultant lack of vascular repair, cART associated dyslipidemia, and direct vascular damage by HIV.8, 9, 11, 14, 35 While our results support the hypothesis that depression increases stroke risk among HIV‐positive people, this association was attenuated by controlling for alcohol and cocaine use. The American Heart Association/American Stroke Association Primary Prevention of Ischemic Stroke Guidelines consider alcohol and illicit drugs, including cocaine, “less well documented or potentially modifiable risk factors,” in part because of the complexities of their association with stroke risk and comorbidity with depression. Alcohol use disorders are associated with increased risk of developing such stroke risk factors as hypertension, diabetes mellitus, hyperlipidemia, cardiomyopathy, and atrial fibrillation and is known to modulate platelet aggregation, impair synthetic liver function, cause hyperhomocysteinemia, and predispose to thromboembolism.36 A “J shape” association between alcohol consumption and ischemic stroke risk has been described in the HIV‐negative population, with alcohol having a protective effect at low and moderate levels of consumption and a noxious effect at higher levels.37 Cocaine use has also been linked to factors increasing stroke risk, including reversible vasospasm, hypertensive surges, drug‐induced arteritis, cardiac arrhythmias, cardiomyopathy, increased platelet aggregation, and thromboembolism.38 Higher rates of unhealthy alcohol and cocaine use have been described in both depressed39, 40 and HIV‐positive populations.9, 12, 13 Increased stroke risk among people with depression and HIV using alcohol and/or cocaine may also be explained by poorer medication adherence, less healthy behavior modification, toxic effects of alcohol on multiple aspects of physiology and vascular health, and increased rates of other vascular risk conditions (eg, smoking).41 Given that many of these vascular risk factors are not specific to HIV, cART was not associated with increased stroke risk, and the effect sizes from Cox models were similar between those with and without HIV, it appears that most of the association between depression and stroke risk is not driven by HIV or HIV‐specific factors.
Depression has been associated with incident stroke in the general population. A meta‐analysis of 17 prospective studies involving 206 641 participants and 6086 stroke cases demonstrated a positive association between depression and subsequent stroke risk (pooled relative risk, 1.34; 95% CI, 1.17–1.54) after adjustment for potential confounders.21 Another meta‐analysis and systematic review of 28 prospective studies involving 317 540 participants and 8478 reported stroke cases also demonstrated, among 6 studies focusing on ischemic stroke, a positive association between depression and ischemic stroke (pooled aHR ratio: 1.25; 95% CI, 1.11–1.40). Also like HIV‐infection, depression has been associated chronic inflammation, platelet dysfunction, dysregulation of both the autonomic nervous system and hypothalamic‐pituitary‐adrenal axis, and the development of atherosclerosis, hypertension, diabetes mellitus, and arrhythmias.21, 42, 43, 44, 45, 46 Behavioral mechanisms seen with increasing frequency among people with depression may also contribute to increase stroke risk, including sedentary lifestyle, cigarette smoking, and poorer medication adherence.5, 21 In our current analysis, while some risk factors occurred more commonly among HIV‐positive people with depression (eg, hypertension), others occurred similarly across groups (eg, atrial fibrillation). Combining these results with the findings that adjusting for cerebrovascular risk factors did not attenuate the observed depression‐stroke risk association suggests that depression does not solely operate through increasing the prevalence of established vascular risk factors.
Age is a well‐described, non‐modifiable risk factor for incident ischemic stroke.1, 5 Outside of the HIV literature, people with stable high depressive symptoms aged 50 to 64 years have been reported to have the highest risk of incident stroke (aHR, 1.87; 95% CI, 1.10–3.16), whereas those ≥65 years with a similar degree of depression demonstrated no increased stroke risk (aHR, 1.32; 95% CI, 0.99–1.77).22 Similarly, among a cohort of 3852 stroke‐free people older than 55 years of age, baseline depression was associated with increased stroke risk (aHR, 2.84; 95% CI, 1.11–7.29) among people aged 55 to 64 years but not among people older than 65 years (aHR, 1.20; 95% CI, 0.80–1.79).23 Conversely, a meta‐analysis and systematic review of depression and stroke risk conducted across 28 prospective studies containing 317 540 people found positive associations between depression and stroke risk among those <65 years and those ≥65 years. Of note, the association was more pronounced among younger people.18 We similarly report that depression may have a greater association with incident stroke among younger people living with HIV.
In HIV‐positive people, we found that the association between depression and incident stroke was slightly attenuated after adjustment for baseline antidepressant use. Interestingly, a meta‐analysis conducted on depression and stroke risk commented that, “it is expected that depression treatment would reduce the risk of development of stroke.”21 However, the literature does not completely support this contention. One case‐control study reported an increased risk of incident stroke among patients receiving pharmacological treatments for depression.47 This relationship, however, may have been confounded by depression severity, given that patients with more severe depression are more likely to receive pharmacological treatment. In our analysis, we present information for baseline antidepressant medication use, rather than information related to longitudinal medication adherence or depression severity. Noting the limitations of the current data, while it would be premature for this observational study to claim that pharmacologic treatment of depression reduces stroke risk, it does suggest the need for randomized controlled trial data to address whether successful treatment of depression reduces stroke risk among HIV‐positive people.
The strengths of our study include the larger sample size, the longer follow‐up period, and ability to control for many potential confounders. Limitations of the current work should also be noted. First, given that this is an observational study, we cannot completely exclude unmeasured or residual confounding, and we can only comment on association rather than causation. Second, given the use of ICD‐9 codes to identify depression and stroke, misclassification could have occurred, although stroke ICD‐9 codes have been shown to be sensitive and specific within the Veteran population and have high agreement with formally adjudicated stroke outcomes.11, 48 Depression misclassification may have occurred; however, if we classified some patients as not having a depressive disorder when they did have one, this would have biased our results towards the null and may have attenuated the depression‐incident stroke relationship.13 Third, we did not review imaging data on all patients diagnosed with an ischemic stroke. As such, we do not know the cause of ischemic stroke or stroke subtype, as these are not routinely available in administrative data. Fourth, these data are applicable to incident stroke rather than recurrent ischemic stroke. Finally, as our cohort is comprised of predominantly male veterans, these results may less generalizable to other populations.
In conclusion, we used a large, contemporary database to conduct the first prospective cohort study to determine whether there is an independent association between depression and incident ischemic stroke in people with HIV. While depression has been associated with incident AMI and CHF among people with and without HIV, and though HIV has been associated with incident ischemic stroke, studies examining the association between depression and incident stroke within the HIV‐positive population are lacking. Ischemic stroke is a leading cause of morbidity and mortality worldwide, with an increasing prevalence among HIV‐positive people. Therefore, investigating novel cerebrovascular risk factors that are also potentially modifiable may have important treatment implications for HIV‐positive people. In this study, not only was depression independently associated with incident stroke, we also report that incident stroke risk is partially accounted for, to varying degrees, by alcohol use disorders, cocaine use, and baseline antidepressant use. Furthermore, depression may be a more important contributor to stroke risk for HIV‐positive people younger versus older than 60 years. Apart from understanding the underlying, and likely interconnected, mechanisms by which depression increases stroke risk among people with and without HIV, future work should explore the relationship between the pharmacologic and non‐pharmacologic treatment of depression and cerebrovascular risk reduction. This would include examining specific types of therapies and classes of antidepressant medications, treatment adherence, and depression severity. Providers caring for HIV‐positive people should be aware that depression may be an important comorbidity as it relates to future stroke risk and consider counseling their patients with HIV regarding the increased stroke risk among those with depression.
Sources of Funding
This work was supported by the National Institutes of Health grants R01 HL126557 and U10AA13566 (to Dr Freiberg). This work was also supported by the US Department of Veterans Affairs Health Services Research and Development grant HX001388‐01A1 (to Dr Sico), the Agency for Healthcare Research & Quality grants R18HS03258 and R18HS023464‐01 (to Dr Crystal), and the Emory Center for AIDS Research grant P30AI050409 (for Dr Marconi). The views expressed in this article are those of the authors and do not necessarily represent the view of the Department of Veterans Affairs.
Disclosures
Dr Sico reports grants from Department of Veterans Affairs, during the conduct of the study; other from American Academy of Neurology, outside the submitted work. Dr Gupta reports grants from the National Institutes of Health and Indiana University School of Medicine; consulting/advisory fees from Gilead Sciences and GlaxoSmithKline/ViiV; and travel support from Gilead Sciences during the conduct of this study. Dr Crystal reports grants from the National Institutes of Health, Agency for Healthcare Research & Quality, and Patient Centered Outcomes Research Institute. Dr Marconi reports grants from the National Institutes of Health. Dr Stewart reports grants from the National Institutes of Health and Indiana University. The remaining authors have no disclosures to report.
Supporting information
Table S1
(J Am Heart Assoc. 2021;10:e06300. DOI: 10.1161/JAHA.119.017637.)
Supplementary Material for this article is available at https://www.ahajournals.org/doi/suppl/10.1161/JAHA.119.017637
For Sources of Funding and Disclosures, see pages 10 and 11.
References
- 1.Kernan WN, Ovbiagele B, Black HR, Bravata DM, Chimowitz MI, Ezekowitz MD, Fang MC, Fisher M, Furie KL, Heck DV, American Heart Association Stroke Council, Council on Cardiovascular and Stroke Nursing, Council on Clinical Cardiology, and Council on Peripheral Vascular Disease , et al. Guidelines for the prevention of stroke in patients with stroke and transient ischemic attack. A guideline for healthcare professionals from the American Heart Association/American Stroke Association. Stroke. 2014;45:2160–2236. DOI: 10.1161/STR.0000000000000024. [DOI] [PubMed] [Google Scholar]
- 2.Samji H, Cescon A, Hogg RS, Modur SP, Althoff KN, Buchacz K, Burchell AN, Cohen M, Gebo KA, Gill MJ, et al. Closing the gap: increases in life expectancy among treated HIV‐positive individuals in the United States and Canada. PLoS One. 2013;8:e81355. DOI: 10.1371/journal.pone.0081355. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Bedimo RJ, McGinnis KA, Dunlap M, Rodriguez‐Barradas MC, Justice AC. Incidence of non‐AIDS‐defining malignancies in HIV‐infected versus noninfected patients in the HAART era: impact of immunosuppression. J Acquir Immune Defic Syndr. 2009;52:203–208. DOI: 10.1097/QAI.0b013e3181b033ab. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.d'Arminio Monforte A, Sabin CA, Phillips A, Sterne J, May M, Justice A, Dabis F, Grabar S, Ledergerber B, Gill J, et al. The changing incidence of AIDS events in patients receiving highly active antiretroviral therapy. Arch Intern Med. 2005;165:416–423. [DOI] [PubMed] [Google Scholar]
- 5.Meschia JF, Bushnell C, Boden‐Albala B, Braun LT, Bravata DM, Chaturvedi S, Creager MA, Eckel RH, Elkind MSV, Fornage M, American Heart Association Stroke Council; Council on Cardiovascular and Stroke Nursing; Council on Clinical Cardiology; Council on Functional Genomics and Translational Biology; Council on Hypertension , et al. Guidelines for the primary prevention of stroke: a statement for healthcare professionals from the American Heart Association/American Stroke Association. Stroke. 2014;45:3754–3832. DOI: 10.1161/STR.0000000000000046. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Longenecker CT, Hoit BD. Imaging atherosclerosis in HIV: carotid intima‐media thickness and beyond. Transl Res. 2012;159:127–139. [DOI] [PubMed] [Google Scholar]
- 7.Beckman JA, Duncan MS, Alcorn CW, So‐Armah K, Butt AA, Goetz MB, Tindle HA, Sico JJ, Tracy RP, Justice AC, et al. Association of human immunodeficiency virus infection and risk of peripheral artery disease. Circulation. 2018;138:255–265. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Butt AA, Chang C‐C, Kuller L, Goetz MB, Leaf D, Rimland D, Gibert CL, Oursler KK, Rodriguez‐Barradas MC, Lim J, et al. Risk of heart failure with human immunodeficiency virus in the absence of prior diagnosis of coronary heart disease. Arch Intern Med. 2011;171:737–743. DOI: 10.1001/archinternmed.2011.151. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Freiberg MS, Chang CC, Kuller LH, Skanderson M, Lowy E, Kraemer KL, Butt AA, Bidwell Goetz M, Leaf D, Oursler KA, et al. HIV infection and the risk of acute myocardial infarction. JAMA Intern Med. 2013;173:614–622. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Chow FC, Regan S, Feske S, Meigs JB, Grinspoon SK, Triant VA. Comparison of ischemic stroke incidence in HIV‐infected and non‐HIV‐infected patients in a U.S. health care system. J Acquir Immune Defic Syndr. 2012;60:351–358. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Sico JJ, Chang CC, So‐Armah K, Justice AC, Hylek E, Skanderson M, McGinnis K, Kuller LH, Kraemer KL, Rimland D, et al. HIV status and the risk of ischemic stroke among men. Neurology. 2015;84:1933–1940. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.White JR, Chang C‐C, So‐Armah KA, Stewart JC, Gupta SK, Butt AA, Gibert CL, Rimland D, Rodriguez‐Barradas MC, Leaf DA, et al. Depression and human immunodeficiency virus infection are risk factors for incident heart failure among veterans: Veterans Aging Cohort Study. Circulation. 2015;132:1630–1638. DOI: 10.1161/CIRCULATIONAHA.114.014443. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Khambaty T, Stewart JC, Gupta SK, Chang CH, Bedimo RJ, Budoff MJ, Butt AA, Crane H, Gibert CL, Leaf DA, et al. Association between depressive disorders and incident acute myocardial infarction in human immunodeficiency virus–infected adults: Veterans Aging Cohort Study. JAMA Cardiol. 2016;1:929–937. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Grinspoon SK. Metabolic syndrome and cardiovascular disease in patients with human immunodeficiency virus. Am J Med. 2005;118(suppl 2):23s–28s. DOI: 10.1016/j.amjmed.2005.01.047. [DOI] [PubMed] [Google Scholar]
- 15.Faltz M, Bergin H, Pilavachi E, Grimwade G, Mabley JG. Effect of the anti‐retroviral drugs efavirenz, tenofovir and emtricitabine on endothelial cell function: role of PARP. Cardiovasc Toxicol. 2017;17:393–404. DOI: 10.1007/s12012-016-9397-4. [DOI] [PubMed] [Google Scholar]
- 16.Falcinelli E, Francisci D, Belfiori B, Petito E, Guglielmini G, Malincarne L, Mezzasoma A, Sebastiano M, Conti V, Giannini S, et al. In vivo platelet activation and platelet hyperreactivity in abacavir‐treated HIV‐infected patients. Thromb Haemost. 2013;110:349–357. DOI: 10.1160/TH12-07-0504. [DOI] [PubMed] [Google Scholar]
- 17.Desai M, Joyce V, Bendavid E, Olshen RA, Hlatky M, Chow A, Holodniy M, Barnett P, Owens DK. Risk of cardiovascular events associated with current exposure to HIV antiretroviral therapies in a US veteran population. Clin Infect Dis. 2015;61:445–452. DOI: 10.1093/cid/civ316. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Pan A, Sun Q, Okereke OI, Rexrode KM, Hu FB. Depression and the risk of stroke morbidity and mortality: a meta‐analysis and systematic review. JAMA. 2011;306:1241–1249. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Sher Y, Lolak S, Maldonado JR. The impact of depression in heart disease. Curr Psychiatry Rep. 2010;12:255–264. [DOI] [PubMed] [Google Scholar]
- 20.Schmidt HD, Shelton RC, Duman RS. Functional biomarkers of depression: diagnosis, treatment, and pathophysiology. Neuropsychopharmacology. 2011;36:2375–2394. DOI: 10.1038/npp.2011.151. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Dong JY, Zhang YH, Tong J, Qin LQ. Depression and risk of stroke: a meta‐analysis of prospective studies. Stroke. 2012;43:32–37. [DOI] [PubMed] [Google Scholar]
- 22.Gilsanz P, Walter S, Tchetgen Tchetgen EJ, Patton KK, Moon JR, Capistrant BD, Marden JR, Kubzansky LD, Kawachi I, Glymour MM. Changes in depressive symptoms and incidence of first stroke among middle‐aged and older US adults. J Am Heart Assoc. 2015;4:e001923. DOI: 10.1161/JAHA.115.001923. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Seifert CL, Poppert H, Sander D, Feurer R, Etgen T, Ander KH, Purner K, Bronner M, Sepp D, Kehl V, et al. 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]
- 24.Munoz‐Moreno JA, Fumaz CR, Ferrer MJ, Gonzalez‐Garcia M, Molto J, Negredo E, Clotet B. Neuropsychiatric symptoms associated with efavirenz: prevalence, correlates, and management. A neurobehavioral review. AIDS Rev. 2009;11:103–109. [PubMed] [Google Scholar]
- 25.Fultz SL, Skanderson M, Mole LA, Gandhi N, Bryant K, Crystal S, Justice AC. Development and verification of a "virtual" cohort using the National VA Health Information System. Med Care. 2006;44:S25–S30. DOI: 10.1097/01.mlr.0000223670.00890.74. [DOI] [PubMed] [Google Scholar]
- 26.Khambaty T, Stewart JC, Gupta SK, Chang CH, Bedimo RJ, Budoff MJ, Butt AA, Crane H, Gibert CL, Leaf DA, et al. Association between depressive disorders and incident acute myocardial infarction in human immunodeficiency virus‐infected adults: Veterans Aging Cohort Study. JAMA Cardiol. 2016;1:929–937. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Armah KA, McGinnis K, Baker J, Gibert C, Butt AA, Bryant KJ, Goetz M, Tracy R, Oursler KK, Rimland D, et al. HIV status, burden of comorbid disease, and biomarkers of inflammation, altered coagulation, and monocyte activation. Clin Infect Dis. 2012;55:126–136. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Chobanian AV, Bakris GL, Black HR, Cushman WC, Green LA, Izzo JL, Jones DW, Materson BJ, Oparil S, Wright JT, et al. Seventh report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure. Hypertension. 2003;42:1206–1252. [DOI] [PubMed] [Google Scholar]
- 29.Butt AA, McGinnis K, Rodriguez‐Barradas MC, Crystal S, Simberkoff M, Goetz MB, Leaf D, Justice AC; Veterans Aging Cohort S . HIV infection and the risk of diabetes mellitus. AIDS. 2009;23:1227–1234. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Levey AS, Stevens LA, Schmid CH, Zhang YL, Castro AF III, Feldman HI, Kusek JW, Eggers P, Van Lente F, Greene T, et al. A new equation to estimate glomerular filtration rate. Ann Intern Med. 2009;150:604–612. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.McGinnis KA, Brandt CA, Skanderson M, Justice AC, Shahrir S, Butt AA, Brown ST, Freiberg MS, Gibert CL, Goetz MB, et al. Validating smoking data from the Veteran’s Affairs Health Factors dataset, an electronic data source. Nicotine Tob Res. 2011;13:1233–1239. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Nan C, Shaefer M, Urbaityte R, Oyee J, Hopking J, Ragone L, Perger T, Win B, Vangerow H, McCoig C, et al. Abacavir use and risk for myocardial infarction and cardiovascular events: pooled analysis of data from clinical trials. Open Forum Infect Dis. 2018;5:ofy086. DOI: 10.1093/ofid/ofy086. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Mateen FJ, Post WS, Sacktor N, Abraham AG, Becker JT, Smith BR, Detels R, Martin E, Phair JP, Shinohara RT; For the Multicenter ACSI . Long‐term predictive value of the Framingham Risk Score for Stroke in HIV‐positive vs HIV‐negative men. Neurology. 2013;81:2094–2102. DOI: 10.1212/01.wnl.0000437296.97946.73. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Marcus JL, Leyden WA, Chao CR, Chow FC, Horberg MA, Hurley LB, Klein DB, Quesenberry CP Jr, Towner WJ, Silverberg MJ. HIV infection and incidence of ischemic stroke. AIDS. 2014;28:1911–1919. [DOI] [PubMed] [Google Scholar]
- 35.Riddler SA, Smit E, Cole SR, Li R, Chmiel JS, Dobs A, Palella F, Visscher B, Evans R, Kingsley LA. Impact of HIV infection and HAART on serum lipids in men. JAMA. 2003;289:2978–2982. [DOI] [PubMed] [Google Scholar]
- 36.Fernandez‐Sola J. Cardiovascular risks and benefits of moderate and heavy alcohol consumption. Nat Rev Cardiol. 2015;12:576–587. [DOI] [PubMed] [Google Scholar]
- 37.Patra J, Taylor B, Irving H, Roerecke M, Baliunas D, Mohapatra S, Rehm J. Alcohol consumption and the risk of morbidity and mortality for different stroke types—a systematic review and meta‐analysis. BMC Public Health. 2010;10:258. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Toossi S, Hess CP, Hills NK, Josephson SA. Neurovascular complications of cocaine use at a tertiary stroke center. J Stroke Cerebrovasc Dis. 2010;19:273–278. [DOI] [PubMed] [Google Scholar]
- 39.Ramsey SE, Engler PA, Stein MD. Alcohol use among depressed patients: the need for assessment and intervention. Prof Psychol Res Pr. 2005;36:203–207. DOI: 10.1037/0735-7028.36.2.203. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40.Grant BF. Comorbidity between DSM‐IV drug use disorders and major depression: results of a national survey of adults. J Subst Abuse. 1995;7:481–497. DOI: 10.1016/0899-3289(95)90017-9. [DOI] [PubMed] [Google Scholar]
- 41.Samet JH, Horton NJ, Meli S, Freedberg KA, Palepu A. Alcohol consumption and antiretroviral adherence among HIV‐infected persons with alcohol problems. Alcohol Clin Exp Res. 2004;28:572–577. DOI: 10.1097/01.ALC.0000122103.74491.78. [DOI] [PubMed] [Google Scholar]
- 42.Howren MB, Lamkin DM, Suls J. Associations of depression with C‐reactive protein, IL‐1, and IL‐6: a meta‐analysis. Psychosom Med. 2009;71:171–186. DOI: 10.1097/PSY.0b013e3181907c1b. [DOI] [PubMed] [Google Scholar]
- 43.Carney RM, Freedland KE, Veith RC. Depression, the autonomic nervous system, and coronary heart disease. Psychosom Med. 2005;67(suppl 1):S29–S33. DOI: 10.1097/01.psy.0000162254.61556.d5. [DOI] [PubMed] [Google Scholar]
- 44.Vreeburg SA, Hoogendijk WJ, van Pelt J, Derijk RH, Verhagen JC, van Dyck R, Smit JH, Zitman FG, Penninx BW. Major depressive disorder and hypothalamic‐pituitary‐adrenal axis activity: results from a large cohort study. Arch Gen Psychiatry. 2009;66:617–626. [DOI] [PubMed] [Google Scholar]
- 45.Taylor WD, Aizenstein HJ, Alexopoulos GS. The vascular depression hypothesis: mechanisms linking vascular disease with depression. Mol Psychiatry. 2013;18:963–974. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 46.Davidson K, Jonas BS, Dixon KE, Markovitz JH. Do depression symptoms predict early hypertension incidence in young adults in the CARDIA study? Coronary Artery Risk Development in Young Adults. Arch Intern Med. 2000;160:1495–1500. [DOI] [PubMed] [Google Scholar]
- 47.Chen Y, Guo JJ, Li H, Wulsin L, Patel NC. Risk of cerebrovascular events associated with antidepressant use in patients with depression: a population‐based, nested case‐control study. Ann Pharmacother. 2008;42:177–184. [DOI] [PubMed] [Google Scholar]
- 48.Ives DG, Fitzpatrick AL, Bild DE, Psaty BM, Kuller LH, Crowley PM, Cruise RG, Theroux S. Surveillance and ascertainment of cardiovascular events. The Cardiovascular Health Study. Ann Epidemiol. 1995;5:278–285. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Table S1
Data Availability Statement
Due to Veterans Affairs (VA) regulations and our ethics agreements, the analytic data sets used for this study are not permitted to leave the VA firewall without a Data Use Agreement. This limitation is consistent with other studies based on VA data. However, VA data are made freely available to researchers with an approved VA study protocol. For more information, please visit https://www.virec.research.va.gov or contact the VA Information Resource Center at VIReC@va.gov.