Meta-analysis of the association between marital status and cardiovascular diseases

Abstract

Significant morbidity and death are connected to cardiovascular disease (CVD), with classic risk factors such as age, sex, smoking, hypertension, hyperlipidemia, and diabetes accounting for 80% of the risk. The other 20% of risk variables are poorly understood. The influence of marital status is one more element whose function is unclear. A comprehensive literature review was conducted on PubMed, Scopus, and Web of Science databases and data from the included articles were extracted and analyzed. A meta-analysis was conducted using the findings of the 13 publications. In this study, we evaluated the incidence of CVD across different marital statuses by analyzing 13 cohort studies from 11 countries with 1,809,825 participants. Our results showed no significant differences in CVD incidence among married (HR 1.00, 95% confidence interval [CI] 0.84–1.20) and unmarried individuals (HR 1.24, 95% CI 0.95–1.62). Divorced individuals had a 28% higher incidence (HR 1.28, 95% CI 1.14–1.44), and widowed individuals showed a 28% increase (HR 1.28, 95% CI 1.11–1.48). The results show no discernible difference in the incidence of CVD between married and single people. However, divorced and widowed people have a noticeably increased risk of CVD incidence. The high heterogeneity of the research highlights the complexity of the link. These results emphasize how crucial it is to take marital status into account when evaluating CVD risk and developing preventative measures.

Introduction

There is an association between cardiovascular disease (CVD) and significant morbidity and mortality.^1,2 About 80% of the risk of developing CVD, including age, sex, smoking, hyperlipidemia, hypertension, and diabetes mellitus, are related to traditional cardiovascular risk factors, and 20% of risk factors are undetermined, ³ So for reducing the burden of CVD it is essential to know the risk factors.

Marital status is a noteworthy factor that is related to socioeconomic determinants that may affect CVD morbidity and mortality. ⁴ The impact of marital or partner status on the long-term risk of readmission for younger persons with AMI is not well understood, even though sociodemographic and psychological factors have been proposed to be significant predictors of the risk of readmission within a year. ⁵

Based on the previous studies, there is an association between lower mortality and marriage and health.^6,7 This might be because of better health behaviors of married people, which may be the consequence of both selection and direct marriage effects; lastly, the stress of divorce or loss of a spouse; and the selection of healthier people, who are more likely to marry or remain married, into the married group (selection theory). In addition, there is an association between marital satisfaction or happiness and remarriage with CVD outcomes. ⁸ Furthermore, it has been shown by the most current research that being married and living with a spouse do not seem to reduce the incidence of CVD. ⁹

The findings from two reviews conducted by Manfredini et al. (2017) ² and Wong et al. (2018) ¹ indicate that marriage is associated with lower risk factors and an improved health status. However, recent studies have produced controversial findings regarding cardiovascular risk among different populations. For instance, Wei (2022) ¹⁷ indicated a significant risk for the unmarried population, while Zhu (2024) ¹⁵ countered this assertion, stating that the risk is not significant. Conversely, Marcus (2019) ²³ reported an association between marriage and a lower risk of CVD; however, Wei (2022) ¹⁷ suggested that the married population may, in fact, be at a higher risk. Due to these controversial findings of the recent studies a new meta-analysis was needed to the impact of marital status on CVD risk by adding recently published studies and our analysis investigate the relation between marital status and incident risk of hypertension.

Methods

This systematic review and meta-analysis investigated the relationship between married status and CVD, considering age and gender as factors. The study protocol was registered on PROSPERO with the registration code CRD42024529235, and our research complies with PRISMA principles.

Literature search

This review article was found by searching PubMed, Scopus, and Web of Science databases up to November 6, 2024, to gather relevant articles. The search strategy included keywords and Medical Subject Headings related to marital status and CVDs. The keywords were combined using the “AND” operator without any date, publication type, or language restrictions. The complete search strategy is detailed in Table 1. The inclusion criteria included (a) patients with CVD, (b) cohort studies, and (c) assessing the relation between marital status and CVD, and the exclusion criteria included (a) cross-sectional studies, (b) review articles, (c) commentaries, (d) editorial studies, (e) animal studies, and (f) irrelevant outcomes. Only 19 publications were left after the literature review was finished, and a meta-analysis was conducted using the findings of these investigations.

Table 1.

Search strategy.

Search engine	Number	Search strategy	Results
PubMed	#1	((marital status) OR (Married) OR (married [Title/Abstract]) OR (Widowed[Title/Abstract]) OR (widower) OR (divorced) OR (spouse[Title/Abstract]))	124378
	#2	(“CVD” OR “cardiovascular disease"[Title/Abstract] OR “Coronary heart disease"[Title/Abstract] OR “CHD"[Title/Abstract] OR “myocardial infarction"[Title/Abstract] OR “MI"[Title/Abstract] OR “Chronic heart failure"[Title/Abstract] OR “Sudden cardiac arrest"[Title/Abstract])	517147
	#1 and #2	((marital status) OR (Married) OR (married[Title/Abstract]) OR (Widowed[Title/Abstract]) OR (widower) OR (divorced) OR (spouse[Title/Abstract])) AND (“CVD” OR “cardiovascular disease"[Title/Abstract] OR “Coronary heart disease"[Title/Abstract] OR “CHD"[Title/Abstract] OR “myocardial infarction"[Title/Abstract] OR “MI"[Title/Abstract] OR “Chronic heart failure"[Title/Abstract] OR “Sudden cardiac arrest"[Title/Abstract])	2272
Scopus	#1	(TITLE-ABS-KEY (marital) AND TITLE-ABS-KEY (status) OR TITLE-ABS-KEY (married) OR TITLE-ABS-KEY (unmarried) OR TITLE-ABS-KEY (divorced) OR TITLE-ABS-KEY (widowed))	63329
	#2	(ALL (cardiovascular) AND ALL (disorder) OR ALL (cvd) OR ALL (coronary) AND ALL (heart) AND ALL (disease) OR ALL (chd) OR ALL (myocardial) AND ALL (infarction) OR ALL (mi) OR ALL (heart) AND ALL (failure) OR ALL (chf) OR ALL (chronic) AND ALL (heart) AND ALL (failure) OR ALL (sudden) AND ALL (cardiac) AND ALL (arrest) OR ALL (cardiac AND arrest) AND ALL (arrest))	46052
	#1 and #2	((TITLE-ABS-KEY (marital) AND TITLE-ABS-KEY (status) OR TITLE-ABS-KEY (married) OR TITLE-ABS-KEY (unmarried) OR TITLE-ABS-KEY (divorced) OR TITLE-ABS-KEY (widowed))) AND ((ALL (cardiovascular) AND ALL (disorder) OR ALL (cvd) OR ALL (coronary) AND ALL (heart) AND ALL (disease) OR ALL (chd) OR ALL (myocardial) AND ALL (infarction) OR ALL (mi) OR ALL (heart) AND ALL (failure) OR ALL (chf) OR ALL (chronic) AND ALL (heart) AND ALL (failure) OR ALL (sudden) AND ALL (cardiac) AND ALL (arrest) OR ALL (cardiac AND arrest) AND ALL (arrest)))	34
Web of Science	#1	(((((TI = (marital status)) OR TI = (Married)) OR TI = (Widowed)) OR TI = (widower)) OR TI = (divorced)) OR TI = (spouse)	23881
	#2	(((((((ALL = (CVD)) OR ALL = (cardiovascular disease)) OR ALL = (Coronary heart disease)) OR ALL = (CHD)) OR ALL = (myocardial infarction)) OR ALL = (MI)) OR ALL = (Chronic heart failure)) OR ALL = (Sudden cardiac arrest)	2442061
	#1 and #2	(((((TI = (marital status)) OR TI = (Married)) OR TI = (Widowed)) OR TI = (widower)) OR TI = (divorced)) OR TI = (spouse) And (((((((ALL = (CVD)) OR ALL = (cardiovascular disease)) OR ALL = (Coronary heart disease)) OR ALL = (CHD)) OR ALL = (myocardial infarction)) OR ALL = (MI)) OR ALL = (Chronic heart failure)) OR ALL = (Sudden cardiac arrest)	1007

Bold values indicate the search results (number of records retrieved) from each database during the systematic literature search.

Criteria for inclusion and exclusion studies

The title and abstract of the articles were evaluated first, then the complete content. Only research looking at the relationship between cardiovascular illnesses and marital status was considered. Meta-analyses and review papers were not considered. The inclusion criteria included (a) patients with CVD, (b) cohort studies, and (c) assessing the relation between marital status and CVD, and the exclusion criteria included (a) cross-sectional studies, (b) review articles, (c) commentaries, (d) editorial studies, (e) animal studies, and (f) irrelevant outcomes.

Data extraction and study quality assessment

Each study's abstract and title were evaluated separately by two reviewers to see if it satisfied the requirements to be included in the meta-analysis. After excluding studies that did not fit the requirements, the whole texts of the papers that remained were examined. After that, the data extraction procedure was applied to the eligible trials. Information on research characteristics, patient-specific factors, study design, and outcomes were all gathered as part of the extraction process. For cohort, case-control, and analytical cross-sectional research, critical appraisal checklists from the Joanna Briggs Institute were employed; in cases of inconsistency, a third author was consulted.

Statistical analysis

For our data analysis, we employed StataCorp LP's STATA 13.1 software, which was operated in College Station, TX, USA. Our primary focus was on combining hazard ratios (HRs) into a forest plot, enabling us to visualize the results alongside a 95% confidence interval (CI) to gauge the precision of our estimates. To assess the variability among the included studies, we utilized the I² statistic; when we detected significant heterogeneity indicated by an I² value greater than 50% we opted to apply the random effects model to ensure a more accurate synthesis of the data. In order to investigate the potential presence of publication bias, we conducted Egger's regression analysis and performed a visual inspection for symmetry within the funnel plot. The choice of the random model for our statistical analysis was made by the author, with the general population serving as the focal point of the study. This comprehensive approach allowed us to provide a clearer understanding of the outcomes while addressing the complexities inherent in synthesizing data from multiple studies.

Result

Study selection

A systematic search was conducted to determine the incidence of CVD in various marital statuses across three databases: PubMed, Web of Science, and Scopus. This initial search yielded 3313 articles. Following the removal of duplicates, 3211 unique articles remained. The selection process adhered to the PRISMA guidelines and checklist. We first screened the titles and abstracts of these 3211 articles, followed by a full-text review of the remaining relevant studies. After a thorough assessment for potential bias, 13 studies met the inclusion criteria and were included in our study (Figure 1).

Figure 1.

Preferred reporting items for systematic reviews and meta-analyses (PRISMA) chart.

Study characteristics

Table 2 presents the fundamental characteristics of the various studies included in this meta-analysis. The studies originate from 11 different countries and collectively involve a population of 1,809,825 participants. A range of CVDs, such as myocardial infarction (MI), ischemic heart disease, stroke, significant cardiac events, and CVD mortality, has been examined in relation to different marital statuses. It is noteworthy that all studies included in this review are cohort studies.

Table 2.

Basic characteristics of studies included in this review.

Author (reference)	Country	Year of publication	Study design	Number of participants	Sex	Marital status	Cardiovascular diseases	Mean age (SD) at baseline	Follow-up duration	Risk of bias assessment (JBL)
Carla Fornari 11	Italy	2010	Cohort	20,647	Men (n = 7520) Women (n = 13,127)	Married or cohabitating (6643 men, 10,767 women) Other living arrangements (814 men, 2276 women) Missing (63 men, 84 women)	Fatal (coronary event, stroke event) and non-fatal major cardiovascular events, coronary end-points (definite and possible myocardial infarctions (MIs), coronary revascularization, silent MIs, sudden cardiac deaths), stroke end-points (definite stroke, carotid arteriotomy)	35–69	Average period of follow-up: 11 years	11/11
Kozo Tanno 12	Japan	2013	Cohort	1064	Men (n = 678) Women (n = 386)	Married 795 (63.5% male) Single 146 (74.7% male) Divorced /widowed 123 (52.0% male)	CVD mortality (cardiac death, death from pulmonary embolism, stroke death, vascular death and sudden death)	men 61.4 (aged 30–95) Women 61.2 (aged 30–95)	Period of follow-up: 5y Mean duration of follow-up: 4y	10/11
Matthew E. Dupre 13	The U.S.	2015	Cohort	15,827	Men (n = 7264) Women (n = 8563)	married 8908 (men 4278, women 4630), divorced 2571 (men 989, women 1582), widowed 724 (men 94, women 630), remarried 3624 (men 1903, women 1721)	AMI	54.3 (men 55.1, women 53.5)	18y (1992–2010)	11/11
Venetia Notara 14	Greece	2015	Cohort	2172	Men (n = 1649) Women (n = 523)	Married (n = 1711) Single (n = 102) Divorced (n = 57) Widowed (n = 220)	acute coronary syndrome (ACS) prognosis	Men 65 ± 13 women 62 ± 11	10y	11/11
Cenjing Zhu 15	United States	2024	prospective cohort	2974	Men (n = 972) Women (n = 2002)	married/partnered (1675) unpartnered (divorced/separated, widowed, or single) (1299)	Readmission to the hospital for any reason related to the heart during the year following the myocardial infarction	48 [interquartile range, 44– 52] years)	1y	10/11
Robin L A Smits 16	Netherlands	2022	retrospective cohort	1,688,285	Men (n = 821,451) Women (n = 866,834)	Not married (women: 227,684 Men: 272,656) Married (women: 438,221 Men: 446,055) Divorced (women: 109,610 Men: 80,021) Widowed (women: 91,293 Men: 22,697)	out-of-hospital cardiac arrest	Women: 49.3 (37.9–62.8) Men: 48.2 (37.7–60.8)	Mean: 2395.717 days (SD: 504.4625) or 6.5y Median: 2556 days (IQR:2556–2556) or 7y	10/11
Zhi-Yao Wei 17	China	2022	cohort	19,912 (4991)	Men (n = 14,921) Women (n = 4991)	Married (18,702) Unmarried (1210)	2-year rate of major adverse cardiac and cerebrovascular events (all-cause death, myocardial infarction, or stroke)	62.4 (53.7–71.7)	2y	11/11
kim 18	korea	2017	cohort	344	Men (n = 208) Women (n = 136)	780 married patients and 200 widowed/divorced/single patients	major cardiovascular events (MACE), coronary artery bypass graft (CABG), cerebral infarction, cerebral bleeding, major bleeding, mortality	Married 67 ± 15. , Widowed/divorced/single. 68 ± 16	1y	10/11
Schultz 19	United state	2017	prospective cohort	6051	Men (n = 3879) Women (n = 2172)	Married = 4088 Never married = 451 Divorced = 842 Widowed = 670	fatal MI, ischemic stroke, or sudden death secondary to presumed cardiovascular cause in this high-risk CAD population	Married = 63(11) Never married = 53(14) Divorced = 60(11) Widowed = 73(11)	Median: 3.7y (IQR: 1.7–6.7Y)	10/11
Karri Silventoinen 20	United kingdom	2024	Retrospective cohort	35,444	Men (n = 16,659) (47%) Women (n = 18,785) (53%)	Married (men = 10,552, women = 11,306) Unmarried (men = 2137, women = 1980) Cohabiting (men = 2356, women = 2457) divorced (men = 1342, women = 2368) widowed (men = 171, women = 775)	Fatal and non-fatal CHD, PGS-CHD (polygenic score of CHD)	30–70	participants conducted between 1992 and 2012. follow-up until 2020,	9/11
Ramezankhani 21	Iran	2019	Retrospective cohort	9737	Men (n = 4382) Women (n = 5355)	Married (men = 4153, women = 4433) Never married (men = 213, women = 204) Widowed/divorced (men = 46 Widowed women = 593 Divorced Woman = 95)	CHD, MI, CHF	47.6 (12.7)	6 y in two-phase	10/11
Vujcic 22	Serbia	2015	Retrospective cohort	135	Men (n = 5562) Women (n = 1671)	Married = 106 Unmarried = 3 Divorced/Widowed = 26	MI, Peripheral vascular disease	57.82 (10.8)	4 y	10/11
Marcus 23	Israel	2019	Retrospective cohort	7233	Men (n = 5562) Women (n = 1671)	Married = 5643 Non married = 1590	CHF, Peripheral vascular disease, MI,	63.96 (12.73)	12 y	11/11

Outcomes

Our analysis revealed that the incidence of CVDs among the married population did not present any statistically significant differences. The overall findings, with a 95% CI, were determined to be 1.00 (95% CI 0.84–1.20). The Egger's test used for publication biased which yielded a bias coefficient of −1.83 (SE = 0.75, t = −2.42, p = 0.060, 95% CI: −3.77 to 0.11). The forest plot illustrated in Figure 2 demonstrates an overall I² value of 86.97%. The funnel plot for the studies was asymmetrical.

Figure 2.

Forrest plot of CVD's hazard ratio (HR) in the married population vs. general population.

According to the findings of this meta-analysis, the incidence of CVDs among the divorced population is elevated by 28% compared to the general population, with a 95% CI ranging from 14% to 44%. The overall I² statistics, depicted in the forest plot presented in Figure 3, is recorded at 71.68%. The Egger's test revealed slope coefficient = 2.59, SE = 0.014, t = 206.19, p < 0.001; bias coefficient = −4.73, SE = 0.47, t = −10.13, p < 0.001). Furthermore, the funnel plot of the included studies demonstrates an asymmetrical distribution.

Figure 3.

Forrest plot for hazard ratio (HR) of CVDs in the divorced population vs. general population.

Based on this meta-analysis, the incidence of CVDs among the widowed population is elevated by 28% compared to the general population (with a 95% CI of 11% to 48%). The forest plot in Figure 4 indicates an overall I²-squared value of 70.50%. The Egger's test showed (slope coefficient = 3.18, SE = 0.015, t = 211.88, p < 0.001; bias coefficient = −5.23, SE = 0.54, t = −9.69, p < 0.001). Furthermore, the funnel plot for these studies is asymmetrical.

Figure 4.

Forrest plot for hazard ratio (HR) of CVDs in the widowed population vs. general population.

According to our analysis, the incidence of CVDs within the unmarried population did not exhibit any significant differences. The overall result, with a 95% CI, was established at 1.24 (95% CI 0.95–1.62). The Egger's test results showed (slope coefficient = 2.45, SE = 0.03, t = 80.70, p < 0.001; bias coefficient = −2.59, SE = 0.90, t = −2.86, p = 0.017). The 95% CIs for the slope and bias were [2.38, 2.52] and [−4.62, −0.57], respectively. The forest plot presented in Figure 5 indicates an overall I² value of 96.04%. Additionally, the funnel plot for these studies is asymmetrical.

Figure 5.

The forest plot shows the hazard ratio (HR) of CVDs among the unmarried population vs. general population.

Discussion

In this systematic review and meta-analysis, we examined the association between marital status and various CVDs, including MI, cardiac arrest, acute coronary syndrome (ACS), heart failure (HF), peripheral vascular disease, and coronary artery disease (CAD). Our analysis incorporated data from 1,809,825 individuals across 13 studies. Our findings indicate that different marital statuses are linked to an increased prevalence of CVDs, with widowed and divorced individuals facing significantly higher risks. While previous research has explored the correlation between marital status and CVD risk, the underlying mechanisms remain insufficiently understood. Several studies suggest that psychological factors and health-related behaviors may play a crucial role. For instance, married men generally exhibit better sleep quality, lower stress levels, improved mood, and healthier dietary habits compared to never-married men. Moreover, supportive romantic relationships have been associated with increased longevity.^24,25 The overall analysis showed no significant difference in CVD incidence among married individuals, with a HR of 1.00 (95% CI: 0.84–1.20). However, studies conducted in different populations have yielded varying results. For instance, Ramezankhani et al. conducted a 12-year longitudinal study in Iran with 9737 participants (55% female) and found a higher risk of CVDs in married individuals compared to those who were never married. ²¹ In contrast, an Italian study that followed 2832 participants (56% female) over 20 years reported no significant initial difference between married and unmarried individuals but later found an association between marriage and CHD in the presence of cardiovascular risk factors. ²⁶ A study by Zhang et al. in the United States, which analyzed 4502 women and 3745 men, found that 8.7% of married women and 13.5% of married men had CVDs. This study also categorized remarried individuals separately, reporting similar CVD prevalence. ²⁷ Moreover, in various studies comparing married individuals to widowed and never-married individuals, CVD-related mortality was significantly lower among the married population.^10,28 Divorced individuals had a 28% higher incidence of CVDs (HR 1.28, 95% CI 1.14–1.44). Our study highlights a particularly significant finding: the incidence of CVDs among widowed individuals is elevated by 28% compared to the general population (HR 1.28, 95% CI 1.11–1.48). Furthermore, divorced individuals demonstrated an even higher risk (HR 1.28, 95% CI 1.14–1.44). However, in a study involving 13,889 participants from Scotland (mean age: 52.3 years), widowed and divorced women exhibited higher CVD mortality rates compared to their married counterparts. Prior research suggests that widowhood is associated with greater psychological stress and reduced social support, which may contribute to adverse cardiovascular outcomes. ²⁹ Similarly, Zhang et al. found that 10.8% of widowed females and 16.5% of widowed males had CVDs, aligning with our analysis.

Notably, our results did not reveal a statistically significant difference in CVD incidence among unmarried individuals (HR 1.24, 95% CI 0.95–1.62). This aligns with Zhang's study in the U.S., where 8.4% of unmarried women and 13% of unmarried men had CVDs. ²⁷ However, a study by Yang in China suggested an association between being unmarried and increased CVD mortality (HR 1.65, 95% CI 1.07–2.54). ³⁰ Another UK-based study involving 13,889 participants (mean age: 52.3 years, 56.1% female) found that being unmarried or divorced significantly increased cardiovascular mortality among women. In single men, cardiovascular mortality risk was 2.97 (95% CI 1.85–4.78), while in never-married women, it was 2.23 (95% CI 1.17–4.24). ²⁹

Strengths and limitation

This meta-analysis also has a number of advantages. It is more broadly applicable because it is based on a sizable sample size of more than 1.8 million participants from various geographic locations. The temporal association between marital status and CVD is strongly supported by the cohort designs of all the included studies. Its methodological strength was further enhanced by the use of specialized statistical software (such as I² and Egger's test) and random-effects models. There should be some disclaimers, though. Study population variability, definitions of marital status, or CVD outcomes can all be blamed for the very heterogeneity that has been noted across studies. The limitations of published aggregate data prevented individual-level control for confounding variables, such as comorbidities, lifestyle factors, or socioeconomic status.

Furthermore, differences between unmarried subgroups (such as never married versus cohabiting) were not always reported. Finally, subgroup analyses based on age and sex were not conducted, even when these variables were reported.

Conclusion

According to the result of study, marital status affects the risk of CVD, especially for those who are divorced or widowed, who have a much higher prevalence of the condition. Other sociodemographic and psychological factors may moderate the effects of marital status, as seen by the lack of substantial variations in CVD incidence between married and unmarried individuals. These findings highlight the necessity of focused interventions and support networks for high-risk groups, including widowed and divorced people, to lessen their susceptibility to cardiovascular events.