Atrial fibrillation (AF) and heart failure (HF), common in older adults, are associated with poor outcomes. However, little is known about their impact, independent of each other. We studied 5,673 community-dwelling adults aged ≥65 years in the Cardiovascular Health Study. Baseline prevalent AF and HF were centrally adjudicated, and 116 patients had AF only, 219 had HF only, 39 had both, and 5,263 had neither. The Cox proportional hazards model was used to estimate age-gender-race–adjusted hazard ratio (aHR) and 95% confidence intervals (CIs) for all-cause, cardiovascular (CV), and non-CV mortalities. Participants had a mean age of 73 years (±6 years), 58% were women, and 15% African-American. During 13 years of follow-up, all-cause mortality occurred in 43%, 66%, 74%, and 85% of those with neither, AF only, HF only, and both, respectively. Compared with neither, aHR (95% CIs) for all-cause mortality associated with AF only, HF only, and both was 1.36 (1.08 to 1.72), 2.31 (1.97 to 2.71), and 3.04 (2.15 to 4.29), respectively. Similar associations were observed with CV mortality, but HF only also had greater non-CV mortality (aHR 1.72, 95% CI 1.35 to 2.18). Compared with AF alone, aHR (95% CIs) associated with HF alone for all-cause, CV, and non-CV mortalities was 1.69 (1.29 to 2.23), 1.73 (1.20 to 2.51), and 1.64 (1.09 to 2.46), respectively. Compared with HF alone, those with both conditions had greater CV but not all-cause mortality. In conclusion, community-dwelling older adults with AF have greater mortality than those without but lesser than those with HF, and both conditions were associated with greater CV and all-cause mortalities, whereas only those with HF had greater non-CV mortality.
Both atrial fibrillation (AF) and heart failure (HF) are common in older adults, and both conditions are associated with poor outcomes. However, because each is a common morbidity in the other condition, little is known about the effect of one condition independent of the other. There are conflicting data as to whether AF is independently associated with mortality in patients with HF. In the present study, we examined interactive impact of AF only, HF only, neither, and both on mortality among community-dwelling older adults.
Methods
We used a public-use copy of the Cardiovascular Health Study (CHS) data obtained from the National Heart, Lung and Blood Institute, which also sponsored the study. The CHS is an ongoing, prospective, community-based, epidemiologic study of cardiovascular (CV) disease risk factors among participants aged ≥65 years, the rationale and design of which have been previously reported. The 5,888 Medicare-eligible CHS participants were recruited in 2 phases (1989 to 1990 and 1992 to 1993) from 4 US counties (Forsyth, North Carolina; Sacramento, California; Washington, Maryland; and Pittsburgh, Pennsylvania). The public-use copy of the CHS data is based on 5,795 participants (93 did not consent to be included in these data).
AF was defined based on baseline electrocardiograms that were confirmed by CHS Events Committee. We excluded 158 participants with a history of AF but without electrocardiographic evidence of AF. HF was centrally adjudicated by the CHS Events Committee based on review of medical records for symptoms, signs, medications, and other evidence. Data on other characteristics were collected at baseline. Missing values for covariates were imputed based on values predicted by age, gender, and race. Of the 5,673 participants, 116 had AF only, 219 had HF only, 39 had both, and 5,263 had neither. The primary outcome was all-cause mortality during 13 years of follow-up, which was centrally adjudicated by the CHS Events Committee. Secondary outcomes included CV and non-CV mortalities.
We used the Pearson chi-square test and analysis of variance for categorical and continuous variables, respectively, as appropriate, for descriptive analyses. We then conducted survival analysis and plotted unadjusted Kaplan-Meier curves for all-cause mortality. We then used Cox proportional hazard models to estimate unadjusted, age-gender-race–adjusted, and multivariate-adjusted hazard ratios (aHRs) and 95% confidence intervals (CIs) for all-cause mortality. The multivariate model was adjusted for age, gender, race, smoking, acute myocardial infarction, hypertension, diabetes mellitus, stroke, chronic obstructive pulmonary disease, cancer, arthritis, left ventricular ejection fraction, instrumental activity of daily living, time to walk 15 feet, serum creatinine, and serum C-reactive protein. We repeated the aforementioned process for CV and non-CV mortalities. We then repeated the aforementioned analyses after including those with a history of AF without electrocardiographic evidence. Finally, we then repeated the aforementioned process using AF only and HF only as references. All statistical tests were 2-sided, and tests with p value <0.05 were considered significant. SPSS 22 for Windows (released 2013; IBM Corp, Armonk, New York) was used for data analysis.
Results
Participants (n = 5,673) had a mean (±SD) age of 73 years (±6), 58% were women, and 15% were African-American. Compared with those with neither condition, those with both AF and HF were more likely to be women, with lesser income, self-reported poor health, and greater prevalence of CV risk factors ( Table 1 ).
| Mean (±SD) or n (%) | Neither AF or HF (n = 5263) | AF Only (n = 116) | HF Only (n = 219) | AF and HF (n = 39) | p-Value |
|---|---|---|---|---|---|
| Age (years) | 73 (±6) | 76 (±6) | 75 (±6) | 76 (±7) | <0.001 |
| Female | 3058 (58%) | 49 (42%) | 115 (53%) | 20 (51%) | 0.002 |
| Non-white | 837 (16%) | 11 (10%) | 55 (25%) | 5 (13%) | 0.001 |
| Education college plus | 2255 (43%) | 49 (42%) | 77 (35%) | 17 (44%) | 0.164 |
| Income >25 K | 1944 (37%) | 44 (38%) | 47 (22%) | 10 (26%) | <0.001 |
| Self-reported general health fair to poor | 1028 (23%) | 36 (31%) | 138 (63%) | 25 (64%) | <0.001 |
| Smoking (pack years) | 17 (±27) | 20 (±29) | 19 (±29) | 17 (±30) | 0.486 |
| Alcohol (drinks per week) | 3 (±6) | 3 (±7) | 1 (±5) | 1 (±4) | 0.005 |
| Medical problems | |||||
| Hypertensive | 3065 (58%) | 69 (60%) | 141 (64%) | 30 (77%) | 0.033 |
| Coronary artery disease | 896 (17%) | 19 (16%) | 143 (65%) | 19 (49%) | <0.001 |
| Acute myocardial infarction | 413 (8%) | 9 (8%) | 97 (40%) | 8 (21%) | <0.001 |
| Diabetes mellitus | 803 (15%) | 24 (21%) | 70 (32%) | 12 (31%) | <0.001 |
| Stroke | 197 (4%) | 11 (10%) | 24 (11%) | 8 (21%) | <0.001 |
| Chronic obstructive pulmonary disease | 641 (12%) | 21 (18%) | 43 (20%) | 6 (15%) | 0.003 |
| Cancer | 749 (14%) | 20 (17%) | 26 (12%) | 5 (13%) | 0.589 |
| Arthritis | 2725 (52%) | 130 (49%) | 57 (60%) | 24 (62%) | 0.084 |
| Left ventricular hypertrophy by electrocardiogram | 218 (4%) | 14 (12%) | 29 (13%) | 8 (21%) | <0.001 |
| Left ventricular dysfunction by echocardiogram | 377 (7%) | 22 (19%) | 76 (35%) | 11 (28%) | <0.001 |
| Clinical parameters | |||||
| Systolic blood pressure (mm Hg) | 137 (±22) | 137 (±21) | 136 (±27) | 139 (±26) | 0.899 |
| Diastolic blood pressure (mm Hg) | 71 (±11) | 73 (±13) | 67 (±12) | 68 (±14) | <0.001 |
| Pulse (beats/minute) | 68 (±11) | 70 (±14) | 70 (±13) | 74 (±12) | <0.001 |
| Body mass index (kg/m 2 ) | 27 (±4) | 27 (±4) | 27 (±5) | 26 (±5) | 0.061 |
| Instrumental activity of daily living | 0.3 (±0.7) | 0.4 (±0.9) | 0.9 (±1.2) | 1.2 (±0.2) | <0.001 |
| Time to walk 15 feet (seconds) | 6 (±2) | 6 (±2) | 7 (±3) | 8 (±4) | <0.001 |
| Laboratory parameters | |||||
| Serum creatinine (mg/dl) | 1.0 (±0.4) | 1.0 (±0.3) | 1.2 (±0.7) | 1.0 (±0.4) | <0.001 |
| Hemoglobin (g/dl) | 14 (±1) | 15 (±1) | 14 (±2) | 14 (±2) | <0.001 |
| Serum cholesterol (mg/dl) | 213 (±39) | 190 (±38) | 200 (±38) | 182 (±35) | <0.001 |
| Serum low density lipoprotein (mg/dl) | 131 (±35) | 114 (±35) | 121 (±34) | 108 (±34) | <0.001 |
| Serum high density lipoprotein (mg/dl) | 55 (±16) | 51 (±15) | 50 (±14) | 47 (±14) | <0.001 |
| Serum triglyceride (mg/dl) | 140 (±76) | 123 (±54) | 147 (±77) | 145 (±99) | 0.062 |
| Serum albumin (g/dl) | 4 (±0) | 4 (±0) | 4 (±0) | 4 (±0) | 0.067 |
| Serum uric acid (mg/dl) | 6 (±2) | 6 (±2) | 7 (±2) | 7 (±2) | <0.001 |
| Serum fibrinogen (mg/dl) | 323 (±66) | 317 (±71) | 353 (±78) | 337 (±75) | <0.001 |
| Serum interleukin-6 (pg/ml) | 2 (±2) | 3 (±2) | 3 (±2) | 3 (±1) | <0.001 |
| Serum coagulation factor-VII (%) | 124 (±29) | 104 (±24) | 119 (±34) | 105 (±26) | <0.001 |
| Serum C-reactive protein (mg/dl) | 5 (±8) | 7 (±16) | 8 (±12) | 7 (±8) | <0.001 |
| Serum insulin (μIU/ml) | 17 (±24) | 18 (±14) | 27 (±56) | 26 (±62) | <0.001 |
During 13 years of follow-up, all-cause mortality occurred in 43%, 66%, 74%, and 85% of those with neither, AF only, HF only, and both, respectively ( Table 2 ). Compared with the neither condition, age-gender-race–aHRs (95% CIs) for all-cause mortality associated with AF only and HF only were 1.36 (1.08 to 1.72) and 2.31 (1.97 to 2.71), respectively (p for interaction = 0.878; Table 2 and Figure 1 ). Respective HRs for AF only and HF only in the cohort including those with a history of AF without electrocardiographic evidence were 1.36 (1.08 to 1.71) and 2.29 (1.96 to 2.68). Compared with the neither condition, age-gender-race–aHR (95% CI) for all-cause mortality associated with both was 3.04 (2.15 to 4.29), which was 3.02 (2.14 to 4.27) after inclusion of those with a history of AF without electrocardiographic evidence. The association with AF only lost significance after multivariate adjustment, whereas that for HF only and both remained significant ( Table 2 ). Both AF only and HF only had significant multivariate-adjusted associations with CV mortality; however, patients with HF only had significantly greater non-CV mortality ( Table 2 ).
| Category | % (Events/Total) | Unadjusted Hazard Ratio (95% Confidence Interval) | Age-Sex-Race Adjusted Hazard Ratio (95% Confidence Interval) | Adjusted Hazard Ratio ∗ (95% Confidence Interval) |
|---|---|---|---|---|
| All-cause mortality | ||||
| Neither | 43 (2283/5263) | 1.00 (Reference) | 1.00 (Reference) | 1.00 (Reference) |
| Atrial fibrillation | 66 (76/116) | 1.85 (1.47–2.33); p <0.001 | 1.36 (1.08–1.72); p = 0.008 | 1.15 (0.91–1.45); p = 0.236 |
| Heart failure | 74 (163/219) | 2.83 (2.42–3.32); p <0.001 | 2.31 (1.97–2.71); p <0.001 | 1.45 (1.22–1.72); p <0.001 |
| Both | 85 (33/39) | 3.95 (2.80–5.57); p <0.001 | 3.04 (2.15–4.29); p <0.001 | 1.70 (1.20–2.41); p = 0.003 |
| Cardiovascular mortality | ||||
| Neither | 17 (905/5263) | 1.00 (Reference) | 1.00 (Reference) | 1.00 (Reference) |
| Atrial fibrillation | 35 (41/116) | 2.49 (1.82–3.41); p <0.001 | 1.85 (1.35–2.54); p <0.001 | 1.40 (1.01-1.93); p = 0.042 |
| Heart failure | 42 (91/219) | 3.92 (3.16–4.87); p <0.001 | 3.20 (2.58–3.98); p <0.001 | 1.60 (1.27–2.03); p <0.001 |
| Both | 59 (23/39) | 6.76 (4.47–10.23); p <0.001 | 5.21 (3.43–7.90); p <0.001 | 2.36 (1.54–3.63); p <0.001 |
| Non-cardiovascular mortality | ||||
| Neither | 26 (1371/5263) | 1.00 (Reference) | 1.00 (Reference) | 1.00 (Reference) |
| Atrial fibrillation | 30 (35/116) | 1.43 (1.02–2.00); p = 0.037 | 1.05 (0.75–1.47); p = 0.792 | 0.96 (0.68–1.35); p = 0.811 |
| Heart failure | 33 (72/219) | 2.10 (1.66–2.67); p <0.001 | 1.72 (1.35–2.18); p <0.001 | 1.30 (1.01–1.67); p = 0.039 |
| Both | 26 (10/39) | 2.03 (1.09–3.78); p <0.026 | 1.56 (0.84–2.92); p = 0.160 | 0.99 (0.53–1.86); p = 0.978 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
