Marital status impact on the outcomes of patients admitted for acute decompensation of heart failure: A retrospective, single‐center, analysis

Abstract Background Conflicting evidence exists regarding the association between marital status and outcomes in patients with heart failure (HF). Further, it is not clear whether type of unmarried status (never married, divorced, or widowed) disparities exist in this context. Hypothesis We hypothesized that marital status will be associated with better outcomes in patients with HF. Methods This single‐center retrospective study utilized a cohort of 7457 patients admitted with acute decompensated HF (ADHF) between 2007 and 2017. We compared baseline characteristics, clinical indices, and outcomes of these patients grouped by their marital status. Cox regression analysis was used to explore the independency of the association between marital status and long‐term outcomes. Results Married patients accounted for 52% of the population while 37%, 9%, and 2% were widowed, divorced, and never married, respectively. Unmarried patients were older (79.8 ± 11.5 vs. 74.8 ± 11.1 years; p < 0.001), more frequently women (71.4% vs. 33.2%; p < 0.001), and less likely to have traditional cardiovascular comorbidities. Compared with married patients, all‐cause mortality incidence was higher in unmarried patients at 30 days (14.7% vs. 11.1%, p < 0.001), 1 year, and 5 years (72.9% vs. 68.4%, p < 0.001). Nonadjusted Kaplan‐Meier estimates for 5‐year all‐cause mortality by sex, demonstrated the best prognosis for married women, and by marital status in unmarried patients, the best prognosis was demonstrated in divorced patients while the worst was recorded in widowed patients. After adjustment for covariates, marital status was not found to be independently associated with ADHF outcomes. Conclusions Marital status is not independently associated with outcomes of patients admitted for ADHF. Efforts for outcomes improvement should focus on other, more traditional risk factors.


| INTRODUCTION
Heart failure (HF) is a leading cause of morbidity and mortality worldwide. It is estimated that >6 million adult Americans were diagnosed with HF between 2015 and 2018, and similar trends are seen in Europe. 1,2 Socioeconomic status (SES) impacts both the incidence and the outcome of HF. 3,4 Since socioeconomic status is multifactorial, the exact impact of different determinants of SES on HF is less clear. Marriage is one of the closest and most intimate social support environments, nevertheless, conflicting evidence exists regarding the association between marital status and outcomes in patients with HF. [5][6][7][8][9] We thus explored the impact of marital status on both the short and long-term outcomes of patients admitted with acute decompensated HF (ADHF).

| METHODS
This single-center, observational retrospective cohort study utilized data from all adult patients admitted to Shamir Medical Center with ADHF between January 1, 2007, and December 31, 2017. The last date for all-cause mortality was December 31, 2020. Eligible patients were those older than 18 years admitted to an internal medicine department (IMD), who were clinically diagnosed with ADHF upon admission (ICD-9 codes: 428.xx, 429.xx, and 514) and subsequently discharged with a similar diagnosis. We elected to include only HF patients admitted to IMD, and not patients admitted to cardiology departments, as we showed in a previous study that these two populations are distinctively different from each other, with patients admitted to IMD being more reflective of the general population of HF patients. 10 This study was approved by the local institutional review board at Shamir Medical Center and patient consent was waived because of the retrospective nature of the data and analysis.
Demographic, clinical, laboratory, and follow-up data including readmission within 30 days were extracted from the hospital's electronic medical record, and all-cause mortality from Israel's ministry of interior affairs database. As described earlier, 10 a separate data set was created to consolidate medical therapy by drug groups both on admission and at discharge.
For the present analysis, patients were grouped based on selfreported marital status as either "married" (including married or attached) or "unmarried" (including single (i.e., never married), divorced, and widowed Logistic regression analysis was used to explore if marital status was independently associated with 30-day readmission and 30-day all-cause mortality. Cox logistic regression analysis was further utilized to explore indices associated with 5-year all-cause mortality.
Co-variates used in all models were forced based on their relevance to our research question (marital status, sex) or their known association with outcomes of HF patients (age, 11 diabetes, ischemic heart disease, 12 chronic kidney disease, 13 atrial fibrillation, 14 chronic obstructive pulmonary disease, 15 anemia, 16

| RESULTS
A total of 7457 patients admitted with a diagnosis of ADHF between 2007 and 2017 with reported marital status, were included in the analysis. Of those, 3904 (52%) patients were married. In the unmarried group, most patients (2722;77.1%) were widowed, 636 (18%) were divorced and 173 (4.9%) were never married.
As detailed in Table 1, compared with patients in the married population, unmarried patients were older (79.8 ± 11.5 vs. 74.8 ± 11.1 years; p < 0.001), and had a higher prevalence of females (71.4% vs. 33.2%; p < 0.001). Further, compared with married patients, unmarried patients were less likely to have traditional cardiovascular comorbidities such as diabetes mellitus, smoking, peripheral vascular disease, chronic kidney disease, and ischemic heart disease and were less likely to be chronically prescribed with angiotensin receptor binders, mineralocorticoid receptor antagonist, statins and antithrombotic medications. Patients in the unmarried group had a higher prevalence of normal left-ventricular function compared with married patients.
Procedures performed during admission and medication prescribed at discharge are detailed in Table 2. Save for percutaneous coronary intervention which was performed with higher prevalence in the married patients group (5.1% vs. 3.4%, p < 0.001), no differences were noted in the prevalence of procedures performed during admission. Beyond the differences in medical therapy prescribed on admission, although the rate of diuretic and betareceptor blocker prescription at discharge was high, compared with T A B L E 1 Baseline demographic, pharmacological, clinical, and laboratory indices of patients admitted with acute decompensated heart failure. Mild dysfunction (40%-50%), n (%) 193 (16) 119 (12) Moderate dysfunction (30%-40%), n (%) 297 (25) 198 (21) married patients, unmarried patients were less likely to be prescribed these drugs.
The outcomes of the two study groups are detailed in Table 3. differences in 5-year all-cause mortality between the two groups demonstrating higher mortality rates for unmarried patients. As depicted in Figure 2, stratifying outcomes by gender and marital status, married women had the highest survival rates at 5 years, followed by married men and unmarried patients (log rank < 0.001). Multivariate analyses for indices independently associated with outcomes are presented in Table 4. While anemia and chronic obstructive pulmonary disease (COPD) were both independently associated with both major outcome indices (30-day readmission, and 5-year all-cause mortality), marital status was not associated with any of these outcomes.

| DISCUSSION
The current study, including over 7000 patients admitted for ADHF  F I G U R E 1 Kaplan-Meier survival estimate for patients admitted for acute decompensated heart failure stratified by marital status.
demonstrating in-hospital mortality of 2%-12% and 1-year all-cause mortality of 12%-28%. 18,19 The fact that about 70% of patients died within 5 years intensifies the need for identifying and focusing on patients at increased risk.
Beyond traditional risk factors associated with poor outcomes, psychosocial factors impact patients' outcomes as demonstrated in various cardiovascular diseases, and as acknowledged by the American Heart Association. 20 These factors include, for example, a low socioeconomic status which was previously associated with poor outcomes and adverse events in HF patients. 3 Marital status, as a form of social support, emerged as an independent factor associated with outcomes in patients admitted for an acute coronary syndrome. [21][22][23] The association between marital status and transcatheter aortic valve implantation is more complex as married women were demonstrated to have poor long-term outcomes while married men had a higher probability of survival. 24 As opposed to the F I G U R E 2 Kaplan-Meier survival estimate for patients admitted for acute decompensated heart failure stratified by marital status and gender.
F I G U R E 3 Kaplan-Meier survival estimate for patients admitted for acute decompensated heart failure stratified by four optional marital status.
aforementioned associations, prior studies were not able to associate between marital status and HF outcomes. Verma et al. 7 25 In their study of patients with HF, worse outcomes were recorded in unmarried patients, but in a multivariate analysis, after forcing medication adherence into the model, marital status was not an independent predictor of outcome.
This may indicate that other indices (e.g., medication adherence, perhaps encouraged by partners), are stronger factors associated with outcome, rather than marital status.
In the present analysis, unmarried patients had nonadjusted poor outcomes compared with married patients. The fact that unmarried patients were roughly 5-year-older than married patients may explain why these differences were not sustained after multivariate adjustment. Further, unmarried patients had a higher prevalence of preserved LV function compared with married patients. Their admission with ADHF puts many of these patients in the HF with preserved ejection fraction (HFpEF) group, as opposed to HF with reduced ejection fraction (HFrEF) patients, and current evidence supports worse outcomes for HFpEF patients. 26 It is also plausible that in these patients with multiple comorbidities, the impact of age, anemia, and chronic obstructive pulmonary disease is higher compared with marital status.
The lack of association between marital status and outcomes of ADHF also highlights the need for a better determinant of psychosocial indices which may impact care and hopefully outcomes.
It can be postulated that marital status as a single determinant of psychosocial status is obsolete. Learning from the process of evaluating the eligibility of HF patients for a heart transplant and ventricular assist devices, it seems that combining socioeconomic status, cognitive assessment, medication adherence, and social support may be a more relevant approach to the association between psychosocial indices and outcomes. 27 Several limitations of this study should be acknowledged. First, this was a retrospective analysis and as such, the analyses may have been skewed by unknown confounders. Second, unmarried patients represent a heterogenic group of patients including never married, widowed, and divorced. The unadjusted survival analysis demonstrated a better outcome for divorced patients compared with married and a dire outcome for widowed patients. Unfortunately, the relatively small sample size in these groups limited the ability to draw conclusions regarding this observation. Third, this study utilized data from a single center, which may limit its findings generalizability.
However, the large size of the cohort, and the heterogeneity of the population in the region of this center, may somewhat mitigate this limitation.
In conclusion, marital status was not found to be independently associated with both short and long-term outcomes of patients admitted for ADHF. Efforts for outcomes improvement should focus on more traditional risk factors such as anemia and COPD.

CONFLICT OF INTEREST STATEMENT
The authors declare no conflict of interest.

DATA AVAILABILITY STATEMENT
Data were generated at Shamir Medical center. Derived data supporting the findings of this study are available from the last author (S. M.) on request. Abbreviations: AF, atrial fibrillation; CI, confidence interval; CKD, chronic kidney disease; COPD, chronic obstructive pulmonary disease; DM, diabetes mellitus; IHD, ischemic heart disease; PVD, peripheral vascular disease.