Whether anxiety is a risk factor for a range of cardiovascular diseases is unclear. We aimed to determine the association between anxiety and a range of cardiovascular diseases. MEDLINE and EMBASE were searched for cohort studies that included participants with and without anxiety, including subjects with anxiety, worry, posttraumatic stress disorder, phobic anxiety, and panic disorder. We examined the association of anxiety with cardiovascular mortality, major cardiovascular events (defined as the composite of cardiovascular death, stroke, coronary heart disease, and heart failure), stroke, coronary heart disease, heart failure, and atrial fibrillation. We identified 46 cohort studies containing 2,017,276 participants and 222,253 subjects with anxiety. Anxiety was associated with a significantly elevated risk of cardiovascular mortality (relative risk [RR] 1.41, CI 1.13 to 1.76), coronary heart disease (RR 1.41, CI 1.23 to 1.61), stroke (RR 1.71, CI 1.18 to 2.50), and heart failure (RR 1.35, CI 1.11 to 1.64). Anxiety was not significantly associated with major cardiovascular events or atrial fibrillation although CIs were wide. Phobic anxiety was associated with a higher risk of coronary heart disease than other anxiety disorders, and posttraumatic stress disorder was associated with a higher risk of stroke. Results were broadly consistent in sensitivity analyses. Anxiety disorders are associated with an elevated risk of a range of different cardiovascular events, including stroke, coronary heart disease, heart failure, and cardiovascular death. Whether these associations are causal is unclear.
Despite the substantial evidence that depression and general psychological distress are associated with incident cardiovascular disease, the association between anxiety and cardiovascular disease is less clear. Previous results in studies conducted in cardiovascular disease populations, such as heart failure, have shown that anxiety is associated with a higher risk of adverse outcomes. However, these results may be due to reverse causality, that is, more severe cardiovascular disease causing greater anxiety rather than anxiety causing adverse outcomes. In a previous meta-analysis of 20 cohort studies, anxiety disorders were associated with a 26% higher risk of coronary heart disease. However, previous studies have provided conflicting results on whether anxiety disorders are associated with risk of stroke, heart failure, or cardiovascular mortality. Contribution of anxiety disorders to the development of cardiovascular disease would have substantial implications for the estimation of the global and regional burden of anxiety disorders, for prevention and treatment of anxiety disorders, and for future research into the relation between anxiety disorders and cardiovascular disease. We therefore aimed to conduct a comprehensive meta-analysis on the association between anxiety and incident cardiovascular disease and death.
Methods
Cohort studies or case-control studies nested within cohort studies were eligible for inclusion. Studies were required to contain adults with and without anxiety and were required to have a minimum of 3-month follow-up. Anxiety was defined as anxiety symptoms, generalized anxiety disorder, panic, phobia, posttraumatic stress disorder (PTSD), and worry, consistent with previous studies. Only studies that were conducted in a general population were eligible for inclusion; studies conducted on specific populations (e.g., type 2 diabetes and myocardial infarction) were excluded. We restricted our analysis to studies conducted in a general population to minimize the risk of reverse causality, that is, cardiovascular disease causing anxiety symptoms and to ensure that our results were broadly applicable. Eligible studies also reported a measure of relative risk for at least one of the following outcomes: cardiovascular mortality, major cardiovascular events (defined as a composite of cardiovascular death, stroke, coronary heart disease, and incident heart failure), stroke, coronary heart disease (coronary heart disease death, including sudden cardiac death, and nonfatal myocardial infarction), heart failure, atrial fibrillation, and chronic kidney disease. No language restrictions were applied.
MEDLINE and EMBASE were searched from inception (1966) to July 2015 by an experienced research librarian. The following search terms were used for anxiety: “anxiety or posttraumatic stress disorder or tension or anxiety symptoms or anxiety disorder or panic or panic attacks or phobic anxiety or phobia or worry.” Bibliographic review of included studies was also conducted to identify potentially eligible studies. Two researchers screened all abstracts identified in the search in duplicate, excluding those that did not meet eligibility criteria. After this screen, full texts of eligible studies were again assessed in duplicate by 2 separate researchers.
Data were extracted using a standardized form in duplicate. General study characteristics, including the population under study, number of participants with and without anxiety, duration of follow-up, mean age, number of men, and number of participants with cardiovascular disease at baseline, were extracted. Maximally adjusted measures of relative risk and associated 95% CIs for outcomes of interest (cardiovascular mortality, major cardiovascular events, stroke, coronary heart disease, heart failure, atrial fibrillation, and chronic kidney disease) were extracted, as well as the published covariates that investigators included in the regression model. Studies were then categorized as unadjusted, minimally adjusted, or adequately adjusted, as previously performed. Unadjusted studies did not adjust for any potential confounders and were excluded from analysis. Minimally adjusted studies adjusted for age and were excluded in a sensitivity analysis. Adequately adjusted studies adjusted for age, gender, and at least 2 established cardiovascular risk factors (blood pressure, cholesterol, smoking status, body mass index, and diabetes). For studies that reported multiple anxiety categories (e.g., high symptoms of anxiety, moderate symptoms of anxiety, and no anxiety symptoms), the relative risk for the most severe category was used. For studies that reported risk of cardiovascular disease per standard deviation of anxiety symptoms or by quantile (e.g., top fourth relative to the bottom fourth of anxiety symptoms), a normal distribution was assumed and the measure of relative risk standardized to correspond to the top 13.5% of the study population compared with the bottom 86.5% of the study population (13.5% being the mean proportion of the population with anxiety among included studies). Risk of bias in included studies was assessed using the Newcastle–Ottawa Scale.
Inverse variance–weighted random-effects meta-analysis was used to derive overall summary estimates for all analyses. For studies that reported separate measures of relative risk for subgroups (e.g., men and women), inverse variance–weighted fixed-effects meta-analysis was used to derive a study-level relative risk before random effects meta-analysis. Heterogeneity was quantified using the I 2 statistic and the Q-test. Publication bias was assessed using forest plots and Egger’s test for coronary heart disease, cardiovascular mortality, and stroke (the only outcomes reported in more than 5 studies).
Five sensitivity analyses were undertaken. First, studies were stratified by baseline year of assessment. Second, studies were stratified by length of follow-up. Third, studies were stratified by mean participant age. Tests for trend were performed for these continuous outcomes using meta-regression. Fourth, studies were stratified by the type of anxiety under investigation (generalized or unspecified anxiety, PTSD, panic disorder, phobic anxiety, or worry) to examine whether different anxiety types differed in their association with cardiovascular disease. Fifth, studies were stratified by level of adjustment (minimally adjusted vs adequately adjusted) to examine whether further adjustment, as well as adjustment for potential mediators such as hypertension, attenuated the observed association between anxiety and cardiovascular disease. These 5 sensitivity analyses were again restricted to coronary heart disease, cardiovascular mortality, and stroke, as other outcomes were reported in fewer than 5 studies.
Results
A total of 1,804 studies were identified in the search, and 1,564 studies excluded in the abstract screen ( Figure 1 ). After excluding an additional 194 studies in the full-text screen, 46 studies were included in the meta-analysis. These 46 studies included 2,017,126 participants, in total, and 222,253 participants with anxiety. Although we included chronic kidney disease as an outcome of interest in our protocol, no studies were identified that examined the association between anxiety and chronic kidney disease. Characteristics of included studies are provided in Table 1 . Eleven studies were minimally adjusted, whereas 35 studies were adequately adjusted ( Supplementary Table 1 ). Most studies were at low risk of bias when assessed using the Newcastle–Ottawa scale ( Supplementary Table 2 ).
| Author | Name of Cohort | Description of Cohort | Location | Baseline Year | Follow Up (yrs) | Participants (n) | Anxiety | Instrument | Age (yrs) | Men (n) | Cardiovascular Disease (n) |
|---|---|---|---|---|---|---|---|---|---|---|---|
| Generalized anxiety or anxiety unspecified | |||||||||||
| Berecki-Gisolf 2012 | Australian Longitudinal Study on Women’s Health | Prospective cohort of women from a random sample of the Medicare Australia database | Australia | 1996 | 15 | 11828 | 595 | NA | 50 | 0 | NA |
| Boyle 2006 | Air Force Health Study | Air Force veterans of Vietname, half of whom had been exposed to Agent Orange | United States | 1982 | 15 | 2105 | NA | Minnesota Multiphasic Personality Inventory | 47 | 2105 | NA |
| Butnoriene 2015 | Primary Health Care Centre | Population-based cohort in Lithuania | Lithuania | 2003-2004 | 9.3 | 553 | 182 | Mini Neuropyshiatric Interview / DSM-IV | 62 | 0 | NA |
| Davidson 2010 | Canadian Nova Scotia Health Survey | Sample of Nova Scotia adult population | Nova Scotia, Canada | 1995 | 8.6 | 1739 | NA | State-Anxiety Scale | 46 | 862 | NA |
| Denollet 2009 | Eindhoven Perimenopausal Osteoporosis Study | Women born between 1941 and 1947, living in Eindhoven, Netherlands | Eindhoven, Netherlands | 1994-1995 | 10 | 5073 | 1664 | Edinburgh Depression Scale | 50 | 0 | 0 |
| Eaker 1992 | Framingham Study | Cohort of Massachussets residents | Massachussets, United States | 1965-1967 | 20 | 362 | NA | NA | NA | 0 | NA |
| Eaker 2005 | Framingham Offspring Study | Cohort of Massachussets residents | Massachussets, United States | 1984-1987 | 10 | 3682 | NA | Framingham Scales | 49 | 1769 | NA |
| Garfield 2014 | Veteran Health Administration | Veterans aged 50-90 | United States | 1999-2000 | 5.7 | 236079 | 22457 | ICD-9-CM | 63 | 220529 | 0 |
| Gustad 2014 ( European Heart Journal ) | Nord-Trøndelag Health Study | Citizens of Nord-Trøndelag County | Nord-Trøndelag County, Norway | 1995-1997 | 11 | 57953 | 3014 | Hospital Anxiety and Depression Scale | 48 | 26551 | 0 |
| Gustad 2014 ( European Journal of Heart Failure ) | Nord-Trøndelag Health Study | Citizens of Nord-Trøndelag County | Nord-Trøndelag County, Norway | 1995-1997 | 11.3 | 62567 | NA | Hospital Anxiety and Depression Scale | 49 | 29318 | 1871 |
| Jakobsen 2008 | Danish Psychiatric Central Research Register | National register of pyschiatric disorders in Denmark with matched controls | Denmark | 1977-2000 | 24 | 75861 | 13970 | ICD-8 and ICD-10 | NA | 20618 | 0 |
| Janszky 2010 | NA | Nationwide survey of Swedish males born 1949-1951 | Sweden | 1969-1970 | 37 | 49321 | 162 | ICD-8 | 19 | 49321 | 0 |
| Laan 2011 | Psychiatric Case Register Middle-Netherlands | Psychiatric patients and a randomly sampled comparison group from Utrecht, Netherlands | Utrecht, Netherlands | 1999-2009 | 9.5 | 129781 | 6919 | DSM-IV | 41 | 60213 | NA |
| Lambiase 2013 | First National Health and Nutrition Examination Survey | Representative longitudinal study of US population | United States | 1971-1975 | 16.3 | 6019 | 1953 | Anxious/tense subscale of General Well-being Schedule | 48 | 2746 | 0 |
| Mykletun 2007 | Nord-Trøndelag Health Study | Random sample of Nord-Trondelag County | Nord-Trondelage County, Norway | 1995-1997 | 4.4 | 60109 | 5769 | Hospital Anxiety and Depression Scale | 48 | NA | NA |
| Nicholson 2005 | Whitehall II Study | Non-industrial civil servants, aged 35-55, working in London | London, United Kingdom | 1985-1988 | 6.8 | 5449 | NA | General Health Questionnaire-30 | 49 | 5449 | NA |
| Ostir 2006 | Health of the Public Study | Population based sample of adults older than 75 in Galveston, Texas | Texas, United States | 1995 | 5 | 506 | 77 | Zung self-rating scale | 81 | 249 | NA |
| Phillips 2009 | Vietnam Experience Study | Male military Vietnam veterans in United States | United States | 1985-1986 | 15 | 4256 | 411 | DSM-3 | 38 | 4256 | NA |
| Scherrer 2010 | Veteran Health Administration | Patients with depression and matched cohort | United States | 1999-2000 | 7 | 355999 | NA | ICD-9-CM | 56 | 314092 | 0 |
| Seldenrijk 2015 | Netherlands Study of Depression and Anxiety | Random sample of households in 90 Dutch municipalities | Netherlands | 2004-2007 | 5.5 | 2541 | 882 | Composite International Diagnostic Interview | 41 | 821 | 0 |
| Shirom 2010 | Clalit Health Services | Participants who underwent screening at the Mor Institute in 1988 | Israel | 1988 | 20 | 968 | 412 | Cornell Medical Index | 42 | 614 | 0 |
| Thurston 2006 | First National Health and Nutrition Examination Survey | Representative longitudinal study of US population | United States | 1971-1975 | 15.1 | 6265 | 937 | General Well-Being Schedule | 48 | 2853 | 0 |
| Tolmunen 2014 | Kuopio Ischemic Heart Disease Study | Randome sampe from rural community in eastern finland | Kuopio, Finland | 1984-1989 | 23.4 | 2388 | NA | Minnesota Multiphasic Personality Inventory | NA | 2388 | NA |
| Weitoft 2005 | Swedish Survey of Living Conditions | Random sample of Swedish population | Sweden | 1980-1996 | 5 | 34511 | 4643 | No instrument: self-perceived | 42 | 17084 | NA |
| Whang 2012 | Women’s Health Study | Female health professionals in the United States | United States | 1993 | 10.4 | 30746 | 475 | Subset of MHI-5 questionnaire | 59 | 0 | 0 |
| Yasuda 2002 | Otsuki-town, Kochi prefectur | Individuals in Shikoku Island, aged 65-84 | Shikoku Island, Japan | 1991 | 7.5 | 980 | NA | General Health Questionnaire | 72 | 382 | NA |
| Panic disorder | |||||||||||
| Batelaan 2014 | Netherlands Health Survey and Incidence Study – 2 | Stratified random sample of general population of Netherlands | Netherlands | 2007-2009 | 3 | 5149 | 467 | DSM-IV | 41 | 2305 | 0 |
| Chen 2009 | Taiwan National Health Insurance Research Database | Random sample of Taiwan national health register | Taiwan | 2004 | 1 | 38564 | 9641 | ICD-9-CM | NA | 14820 | 4868 |
| Chen 2010 | Taiwan National Health Insurance Research Database | Random sample of Taiwan national health register | Taiwan | 2002-2003 | 3 | 23346 | 3891 | ICD-9-CM | NA | 8844 | NA |
| Cheng 2013 | National Health Insurance Research Database | Random sample of Taiwan national health register | Taiwan | 2000 | 7 | 42768 | 3888 | ICD-9-CM | 46 | 15407 | 8513 |
| Gomez-Caminero 2005 | Integrated Health Care Information Services Database | Managed care database of 30 US health plans | United States | 1997-2002 | NA | 78580 | 39290 | ICD-9 | 36 | 28716 | 0 |
| Smoller 2007 | Myocardial Ischaemia and Migraine Study | Community-dwelling post-menopausal women | United States | 1997-2000 | 5.3 | 3243 | 330 | DSM-IV | 66 | 0 | 149 |
| Walters 2008 | General Practice Research Database | Sample of general population using electronic health records in United Kingdom | United Kingdom | 1990-2002 | 2 | 404654 | 57615 | GP Diagnosis | 43 | 110894 | 5149 |
| Phobic anxiety | |||||||||||
| Albert 2005 | Nurses’ Health Study | Registered nurses in United States, aged 30-55 | United States | 1976 | 12 | 72359 | 20021 | Crown-Crisp Experiental Index | 54 | 0 | 0 |
| Haines 1987 | Northwick Park Heart Study | Prospective study from three occupational groups in North-West London | Northwick Park, United Kingdom | 1972-1978 | 6.7 | 1457 | 367 | Crown-Crisp Experiental Index | NA | 1457 | NA |
| Kawachi 1994 | Health Professionals Follow Up Study | Longitudinal study of US health professionals | United States | 1988 | 2 | 33999 | 2307 | Crown-Crisp Experiental Index | NA | 33999 | 0 |
| PTSD | |||||||||||
| Boscarino 2006 | Vietnam War Era | United States Army veterans who served in Vietnam | United States | 1985 | 16 | 15288 | 1050 | Research Triangle Institute-PTSD Scale | NA | 15288 | NA |
| Boscarino 2008 | NA | Male US Army Veterans from Vietnam War | United States | 1985 | 15.1 | 4328 | 323 | DSM-3 | 54 | 4328 | NA |
| Chen 2015 | Taiwan National Health Insurance Research Database | Random sample of Taiwan national health register | Taiwan | 2002-2009 | NA | 26085 | 5217 | ICD-9 | 37 | 5450 | NA |
| Crum-Cianflone 2014 | Millennium Cohort | Military personnel serving in October 2000 | United States | 2001-2003 | 5.6 | 60025 | 3331 | PTSD Checklist, Civilian Version | 34 | 43914 | NA |
| Jordan 2013 | World Trade Center Health Registry | Individuals exposed to World Trade Centers attack | New York, United States | 2003-2004 | 6.5 | 46346 | 8102 | PTSD Checklist | 41 | 27667 | 0 |
| Kubzansky 2007 | Normative Aging Study | Veterans from Greater Boston Area | Greater Boston, United States | 1986, 1990 | 11.4 | 1002 | NA | Mississippi Scale for Combat Related PTSD | 63 | 1002 | NA |
| Kubzansky 2009 | Baltimore cohort of Epidemiological Catchment Area | Women in Baltimore participating in a survey of psychiatric conditions in the general population | Baltimore, United States | 1980-1983 | 14 | 1059 | 42 | NIMH Diagnostic Interview Schedule using DSM-3 Criteria | 44 | 0 | NA |
| Roy 2015 | Veterans Affairs Pacific Islands Health Care System | Veterans in the Pacific Islands | Pacific Islands, USA | 2005 | 7.2 | 8248 | 1712 | ICD-9 | 64 | 7878 | NA |
| Vaccarino 2013 | Vietnam Era Twin | Sample of twins from military who served in the Vietname era | United States | 1987-1992 | 13 | 562 | 137 | Diagnostic Interview Schedule / DSM-3 | 43 | 562 | 0 |
| Worry | |||||||||||
| Vogt 1994 | Kaiser Permanente | Random sample of Northwest Region of Kaiser Permanente | Oregon, United States | 1970-1971 | 15 | 2573 | NA | Bradburn Worries Index | NA | 1187 | NA |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
