Introduction

Heart failure (HF) significantly contributes to mortality, morbidity, and reduced quality of life, while imposing a considerable financial burden on health care systems worldwide.1-4 International guidelines provide clear recommendations for introducing evidence-based drugs for patients with every phenotype of HF, including HF with reduced ejection fraction (HFrEF), HF with mildly reduced EF (HFmrEF), and HF with preserved EF (HFpEF).5 Elderly patients represent a constantly increasing part of the HF population.6 They are often characterized by multimorbidity, frailty, and cognitive impairment, which can limit the range of therapeutic options available to them.7,8 Observational studies consistently demonstrate that many patients do not receive potentially life-saving therapies, such as percutaneous coronary interventions and coronary artery bypass grafts.7,9,10 Furthermore, guideline-recommended therapy for HF is insufficiently implemented in individuals aged 80 years or older.11,12 Despite high prescription rates, patients commonly fail to reach target doses of recommended HF medications.7,11 Numerous randomized clinical studies have demonstrated that suboptimal dosing may negatively impact cardiovascular outcomes.7,13-15

Recent studies suggest that initiating and continuing angiotensin-converting enzyme inhibitors and β-blockers reduce 1-year mortality but not readmission rates in both North American and Japanese populations.16,17 However, there are limited data on how often cardiovascular medications were prescribed in the elderly population in each HF phenotype, and appropriate analysis of all-cause HF mortality is lacking.

Several studies have explored the differences in HF outcomes across various age groups.6,18,19 Data from the MAGGIC (Meta-Analysis Global Group in Chronic Heart Failure) database20 and the CHARM (Candesartan in Heart Failure: Assessment of Reduction in Mortality and Morbidity) program21 indicate that elderly HF patients exhibit higher mortality rates in comparison with younger patients.

Moreover, advanced age is associated with an increased risk of dementia, and it has been shown that screening diagnosis for dementia following HF decompensation is associated with left ventricular (LV) geometry changes and increased risk of rehospitalization for HF.22,23

To address the need for a longer-term, nationwide perspective, the Lesser Poland Cracovian Heart Failure (LECRA-HF) registry was established. The study involved a large population from a single university hospital, the HF center in Kraków, Poland. The objective of this analysis was to demonstrate the influence of cardiovascular comorbidities, HF laboratory assessments, echocardiographic evaluations, and HF treatments on all-cause mortality across various HF phenotypes within the elderly HF population. We hypothesized that that the population of at least 80 years or older would have worse long-term prognosis, more comorbidities, and that female sex would be associated with lower mortality.

Methods

We conducted a retrospective observational investigation involving patients with HF who were admitted to our tertiary medical facility due to HF decompensation between 2009 and 2019. These patients were enrolled in the LECRA-HF registry. The detailed methodology of the LECRA-HF registry has already been published.24,25 The study protocol complied with the Declaration of Helsinki and was approved by the local Ethics Committee of the Jagiellonian University Medical College (1072.6120.349.2022).

For all patients analyzed, the initial HF symptoms were assessed using the New York Heart Association functional classification system, while the baseline levels of N-terminal pro–B-type natriuretic peptide (NT-proBNP) were measured on admission. Furthermore, the etiology of HF was determined for every patient, and the details regarding the prescribed treatment upon discharge were documented.

Transthoracic echocardiogram examination was conducted on admission for each patient by a certified echocardiographer following current standards.26,27 LV ejection fraction (LVEF) was evaluated using the biplane Simpson method. Invasive coronary angiography was performed during the index HF hospitalization, when clinically indicated. This included cases with ambiguous etiology of HF deterioration, absence of angina or prior coronary artery disease, HF with systolic dysfunction accompanied by angina or regional wall motion abnormalities, and / or evidence of reversible myocardial ischemia (eg, with heart scintigraphy) when revascularization was being considered.

Baseline characteristics were collected during the index HF hospitalization. They included patient demographics, anthropometric measurements, cardiovascular risk factors, previous cardiovascular treatment, other comorbidities, and laboratory test results on admission.

Renal failure was diagnosed when the estimated glomerular filtration rate (eGFR), calculated using the Chronic Kidney Disease Epidemiology Collaboration formula, was lower than 60 ml/min /1.73 m2.

Statistical analysis

Numerical variables were first checked for normal distribution and expressed as median (interquartile range [IQR]), whereas categorical variables were expressed as numbers (percentages). Numerical variables were compared by the t test or the Mann–Whitney test if the distribution was normal or different than normal, respectively. Categorical variables were analyzed using the χ² test or the Fisher exact test. The associations between numerical variables were assessed by the Pearson or the Spearman rank correlation coefficient, respectively. A Kaplan–Meier curve of all-cause mortality in the studied groups was plotted and compared with the log-rank test. Finally, a multiple logistic regression model was performed to find the independent determinants of all-cause long-term mortality. The model was adjusted for sex, coronary angiography, NT-proBNP and baseline creatinine levels, LVEF, left atrial diameter, and tricuspid annular plane systolic excursion (TAPSE). These variables were selected based on a P value below 0.01 in the univariate analysis and their clinical relevance. The receiver operating characteristic (ROC) curve was calculated to assess the discrimination value of performed model. A 2-sided P value below 0.05 was considered significant. All statistical analyses were performed using STATISTICA software, version 13.3 (TIBCO Software Inc., Palo Alto, California, United States) or IBM Statistics software, version 26.0 (IBM Corp, Armonk, New York, United States).

Results

Of the 1394 enrolled patients, 306 (22%) were aged 80 years or more, and among them, 7.8% were over 90 years. The remaining 1088 patients (78%) were younger than 80 years.

Baseline characteristics

Median (IQR) age of the older subgroup was 84 (82–86) years, and of the younger subgroup 68 (61–74) years.

The baseline characteristics of both groups are shown in Table 1. The patients aged 80 years and older were more often women (45.1% vs 30.2%; P <⁠0.001) and had lower median (IQR) body mass index (BMI; 26.9 [24–29.9] vs 28.9 [25.5–32.7]; P <⁠0.001) than the younger population. The individuals aged 80 years and older had higher systolic blood pressure and lower diastolic blood pressure on admission, as compared with the younger group (P <⁠0.001; P = 0.004, respectively). There was a significant difference in the distribution of HF types between the groups (Table 1). Although in both groups, the most frequent HF type was HFrEF, it occurred less frequently in the older than the younger patients (54.4% vs 74.4%; P <⁠0.001). In the elderly population, HFpEF was a more significant component of HF types in comparison with the younger group (39.2% vs 17.7%; P <⁠0.001).

Table 1. Baseline characteristics of the patients

Parameter

Patients ≥80 years (n = 306)

Patients <⁠80 years (n = 1088)

P value

Women

138 (45.1)

328 (30.2)

<⁠0.001

Age, y

84 (82–86)

68 (61–74)

<⁠0.001

Body mass index, kg/m2

26.9 (24–29.9)

28.9 (25.5–32.7)

<⁠0.001

Systolic blood pressure on admission, mm Hg

132 (117–150)

125 (110–145)

<⁠0.001

Diastolic blood pressure on admission, mm Hg

75 (63–85)

76 (68–88)

0.004

NYHA class III/IV

226 (73.9)

800 (73.5)

0.69

HF phenotype

HFrEF

136 (54.4)

664 (74.4)

<⁠0.001

HFmrEF

16 (6.4)

71 (8)

HFpEF

98 (39.2)

158 (17.7)

Length of hospital stay, d, median (IQR)

7 (4–10)

7 (4–11)

0.9

Comorbidities and medical history

Prior myocardial infarction

109 (35.6)

423 (39.1)

0.46

Percutaneous coronary intervention

100 (32.8)

331 (30.7)

0.69

Coronary artery bypass surgery

42 (13.8)

165 (15.4)

0.68

Hypertension

262 (85.6)

853 (78.6)

0.007

Hyperlipidemia

187 (61.1)

690 (63.6)

0.43

Diabetes mellitus

143 (46.9)

452 (41.7)

0.48

Renal failure

151 (49.7)

307 (28.5)

<⁠0.001

Peripheral arterial disease

52 (17)

106 (9.8)

0.005

Atrial fibrillation

187 (61.1)

497 (45.7)

<⁠0.001

Stroke

22 (7.2)

97 (9)

0.49

History of cancer

35 (11.4)

92 (8.5)

0.11

COPD

37 (12.1)

125 (11.6)

0.81

Pacemaker

87 (28.6)

137 (12.6)

<⁠0.001

ICD

24 (7.9)

155 (14.3)

0.003

CRT

13 (4.3)

63 (5.8)

0.3

Coronary invasive diagnostics and significant coronary stenosis

Coronary angiography

74 (24.2)

343 (31.5)

0.04

Significant coronary stenosis

27 (9.2)

84 (7.9)

0.9

Treatment at discharge

ACEI

189 (67.5)

716 (69.7)

0.68

β-Blocker

269 (90.6)

963 (90.6)

0.99

Mineralocorticoid receptor antagonist

129 (42.4)

544 (50.3)

0.01

Loop diuretic

240 (81.1)

834 (78.5)

0.33

Statin

215 (70.5)

710 (65.5)

0.1

Data are presented as number (percentage) or median (interquartile range).

Abbreviations: ACEI, angiotensin-converting enzyme inhibitor; COPD, chronic obstructive pulmonary disease; CRT, cardiac resynchronization therapy; HF, heart failure; HFmrEF, heart failure with mildly reduced ejection fraction; HFpEF, heart failure with preserved ejection fraction; HFrEF, heart failure with reduced ejection fraction; ICD, implantable cardioverter-defibrillator; NYHA, New York Heart Association

The patients aged 80 years and older were more frequently diagnosed with comorbidities, including atrial fibrillation (AF; 61.1% vs 45.7%; P <⁠0.001), hypertension (85.6% vs 78.6%; P = 0.007), peripheral arterial disease (PAD; 17% vs 9.8%; P = 0.005), and renal failure (49.7% vs 28.5%; P <⁠0.001). The elderly patients had more pacemakers implanted (28.6% vs 12.6%; P <⁠0.001), but received an implantable cardioverter-defibrillator (ICD) less frequently than the younger population (7.9% vs 14.3%; P = 0.003).

Coronary angiography was performed less frequently in the older group (24.2% vs 31.5%; P = 0.039), with similar rates of significant coronary stenosis observed in both groups. Regarding medical treatment upon discharge, the elderly patients received mineralocorticoid receptor antagonists less frequently than the younger individuals (42.4% vs 50.3%; P = 0.01). There were no significant differences in the prescribed treatments between the groups.

Laboratory parameters

In Table 2, we summarized the laboratory test results on admission. In the participants aged 80 years and older, median (IQR) NT-proBNP level was higher than in the younger group (4497 [1879–8808] vs 3291 [1399–7275] pg/ml; P = 0.002). Additionally, impaired kidney function was observed in the older group, as evidenced by higher median (IQR) creatinine levels on admission (112 [93–148] vs 101 [84–128] mmol/l; P <⁠0.001) and median (IQR) peak creatinine levels (125 [101–173] vs 109 [90–146] mmol/l; P <⁠0.001), along with a lower median (IQR) eGFR than in the younger population (41 [31–54] vs 70 [49–93] ml/min/1.73 m²; P <⁠0.001).

Table 2. Laboratory parameters of the study patients

Parameter

Patients ≥80 years (n = 306)

Patients <⁠80 years (n = 1088)

P value

NT-proBNP, pg/ml

4497 (1879–8808)

3291 (1399–7275)

0.002

Creatinine on admission, mmol/l

112 (93–148)

101 (84–128)

<⁠0.001

eGFR on admission, ml/min/1.73 m2

41 (31–54)

70 (49–93)

<⁠0.001

Peak creatinine, mmol/l

125 (101–173)

109 (90–146)

<⁠0.001

Sodium, mmol/l

140 (138–142)

140 (137–142)

0.37

Potassium, mmol/l

4.4 (4–4.8)

4.4 (4.1–4.8)

0.88

Hemoglobin, g/l

12.1 (11–13.3)

13.4 (12–14.6)

<⁠0.001

White blood cells, × 103/µl

7.4 (5.9–9.1)

7.8 (6.4–9.6)

0.002

Platelet count, × 103/µl

193 (156–229)

207 (171–252)

<⁠0.001

Glucose, mmol/l

6.2 (5.6–7.7)

6.1 (5.3–7.3)

0.03

Total cholesterol, mmol/l

3.6 (3–4.3)

3.7 (3.1–4.5)

0.32

LDL cholesterol, mmol/l

2.1 (1.6–2.6)

2.2 (1.7–2.9)

0.09

HDL cholesterol, mmol/l

1.2 (1–1.5)

1.1 (0.9–1.4)

<⁠0.001

Triglycerides, mmol/l

0.9 (0.7–1.3)

1.1 (0.8–1.5)

<⁠0.001

Data are presented as median (interquartile range).

Abbreviations: eGFR, glomerular filtration rate; HDL, high-density lipoprotein; LDL, low-density lipoprotein; NT-proBNP, N-terminal pro–B-type natriuretic peptide

Important differences were also observed in blood morphology parameters. The patients aged 80 years and older had lower median (IQR) hemoglobin levels (12.1 [11–13.3] vs 13.4 [12–14.6] g/dl; P <⁠0.001) and white blood cell counts (7.4 [5.9–9.1] vs 7.8 [6.4–9.6] × 10³/µl; P = 0.002), but a higher platelet count than the younger individuals (193 [156–229] vs 207 [171–252] × 10³/µl; P <⁠0.001).

The older patients also had higher median (IQR) glucose levels than the younger group (6.2 [5.6–7.7] vs 6.1 [5.3–7.3] mmol/l; P = 0.03). Interestingly, they exhibited a more favorable lipid profile, with higher median (IQR) high-density lipoprotein cholesterol (HDL-C) levels (1.2 [1–1.5] vs 1.1 [0.9–1.4] mmol/l; P <⁠0.001) and lower triglyceride levels than the younger population (0.9 [0.7–1.3] vs 1.1 [0.8–1.5] mmol/l; P <⁠0.001).

Echocardiographic parameters

In Table 3, we outlined the echocardiographic measurements. In the patients aged 80 years and older, the prevalence of HFrEF was lower than in the younger group, as shown in Table 1 (54.4% vs 74.4%; P <⁠0.001). Additionally, median (IQR) LVEF was markedly higher in the older group (41% [26%–55%] vs 30% [20%–45%]; P <⁠0.001).

Table 3. Echocardiographic parameters of the study patients

Parameter

Patients ≥80 years (n = 306)

Patients <⁠80 years (n = 1088)

P value

LVEF at baseline, %

41 (26–55)

30 (20–45)

<⁠0.001

LV end-diastolic volume, ml

134 (91–182)

190 (140–250)

<⁠0.001

LV end-diastolic diameter, mm

52 (46–59)

61 (53–68)

<⁠0.001

LV end-systolic diameter, mm

36 (29–47)

48 (38–58)

<⁠0.001

Interventricular septum, mm

12 (10–14)

11 (9–13)

<⁠0.001

Left posterior wall, mm

11 (9–12)

10 (9–12)

0.003

Left atrium, mm

47 (44–52)

50 (44–55)

0.001

Right ventricular diameter, mm

33 (29–37)

34 (30–40)

0.008

TAPSE, mm

18 (15–21)

17 (14–20)

0.02

Data are presented as median (interquartile range).

Abbreviations: LV, left ventricular; LVEF, left ventricular ejection fraction; TAPSE, tricuspid annular plane systolic excursion

Considerable differences in LV dimensions were observed in the older cohort, as compared with the younger one. The patients aged 80 years and older had a smaller median (IQR) end-diastolic LV diameter (52 [46–59] vs 61 [53–68] mm; P <⁠0.001), end-systolic LV diameter (36 [29–47] vs 48 [38–58] mm; P <⁠0.001), and LV end-diastolic volume than the younger participants (134 [91–182] vs 190 [140–250] ml; P <⁠0.001).

Moreover, the older patients exhibited smaller  median (IQR) left atrial diameter than the younger group (47 [44–52] vs 50 [44–55] mm; P = 0.001). Median (IQR) interventricular septal thickness and posterior LV wall thickness were greater in the elderly group (12 [10–14] vs 11 [9–13] mm; P <⁠0.001, and 11 [9–12] vs 10 [9–12] mm; P = 0.003, respectively).

Measurements of right ventricular (RV) dimensions also differed between the groups. Median (IQR) RV diameter was slightly larger in the patients aged 80 years and older (33 [29–37] vs 34 [30–40] mm; P = 0.008), whereas their TAPSE was lower than in the younger group (18 [15–21] vs 17 [14–20] mm; P = 0.02).

Long-term mortality

Median (IQR) follow-up was 46 (15–81) months. Long-term all-cause mortality was notably higher in the patients aged 80 years and older than the younger group (81% vs 67.4%; P <⁠0.001; Figure 1). Furthermore, in the older cohort, women had a higher long-term survival rate than men (P = 0.01; Figure 1).

Figure 1. Kaplan–Meier survival curves of the patients stratified by age (A) and sex (B)

For the patients aged 80 years and older, we designed a predictive model for all-cause mortality. The model included all variables from the univariate analysis, that is, female sex, coronary angiography results, NT-proBNP and baseline creatinine levels, LVEF at baseline, LA diameter (parasternal long axis), and TAPSE value (Table 4). The ROC curve analysis and area under the curve (AUC) confirmed the meaningful performance of the described model for the prediction of all-cause mortality (AUC = 0.71; P <⁠0.0001; Figure 2). Female sex and TAPSE were identified as predictors of lower all-cause mortality in the patients aged 80 years and older (odds ratio [OR], 0.47; 95% CI, 0.26–0.83; P = 0.01 and OR, 0.91; 95% CI, 0.84–0.97; P = 0.01, respectively). Baseline creatinine level was a predictor of higher all-cause mortality (OR, 1.01; 95% CI, 1–1.02; P = 0.011). The multivariable analysis identified female sex as an independent predictor of higher overall survival in the patients aged 80 years and older (OR, 0.33; 95% CI, 0.12–0.82; P = 0.02; Table 4).

Table 4. Independent predictors of all-cause mortality in the patients aged 80 years or older

Independent variable

Univariate model

Multivariable model

OR (95% CI)

P value

OR (95% CI)

P value

Female sex

0.47 (0.26–0.83)

0.01

0.32 (0.12–0.82)

0.02

Coronary angiography

0.58 (0.32–1.11)

0.09

–

–

NT-proBNP, pg/ml

0.84 (0.67–1.05)

0.13

–

–

Baseline creatinine, mmol/l

1.01 (1–1.02)

0.01

–

–

LVEF at baseline, %

0.98 (0.96–1)

0.05

–

–

Left atrium diameter (PLAX), mm

1.04 (0.99–1.09)

0.07

–

–

TAPSE, mm

0.91 (0.84–0.97)

0.01

–

–

Abbreviations: PLAX, parasternal long axis; OR, odds ratio; others, see Tables 2 and 3

Figure 2. Receiver operating characteristic curve for the predictive model of all-cause mortality

Abbreviations: AUC, area under the curve

Discussion

Our study indicates that, while HFrEF was the most common HF phenotype in both groups, HFpEF was more frequently observed in the elderly than the younger patients. The patients aged 80 years and older were characterized by a higher burden of comorbidities, such as hypertension, AF, PAD, and renal failure. They also had a higher overall mortality, as compared with the individuals younger than 80 years. Moreover, we found that female sex was an independent predictor of lower long-term mortality.

Among the patients enrolled in the LECRA-HF registry, 22% were 80 years old and older, and among them, 7.8% were over 90 years, representing 1.7% of the overall HF population. In the Euro Heart Failure Survey II, patients aged 80 years and older constituted 21% of the population hospitalized for HF.28 An individual patient meta-analysis of 8 clinical trials showed that octogenarians accounted for 13.1% of the patients hospitalized for HF.29 Our results align with those findings, indicating that our cohort may be representative of the HF population. Currently, there are limited data on HF among patients over 90 years in the literature; therefore, its prevalence in this population is not well established.

In both studied groups, the most prevalent HF phenotype was HFrEF. Although HFrEF is the most frequent HF phenotype in the entire HF population,30,31 a recent meta-analysis showed that HFpEF is the most common phenotype in patients aged 80 years and older. In our study, HFpEF was more commonly diagnosed in the elderly patients, a finding consistent with the existing literature indicating that the aging population is more prone to developing HFpEF.32 The increasing stiffness of the heart muscle with age impairs diastolic function. Chronic changes decrease cardiac output, leading to HFpEF.33

In the current study, the patients aged 80 years and older had lower BMI that the younger group. Studies have shown that low BMI in the elderly is associated with lower functional capacity and an increased risk of mortality reference. Malnutrition is also a common issue in the elderly, and it is widely recognized as an independent predictor of both morbidity and mortality.34

The prevalence of comorbidities, such as AF, hypertension, PAD, and renal failure, was higher in the older patients with HF, which is consistent with other studies on comorbid conditions in the elderly.35,36 Research indicates that approximately 80% of the elderly population has 3 or more chronic conditions.35 Since each chronic disease is typically managed by a different specialist, this division of care can result in ineffective collaboration among health care providers. Consequently, the presence of multiple comorbidities in elderly patients with HF often contributes to a reduced quality of life and increased mortality risk.35

We did not find significant differences between the groups in terms of medical treatment, except for a lower frequency of mineralocorticoid receptor antagonists prescribed to the elderly group, which has been previously observed. However, the underprescription is a well-documented phenomenon in the literature, with studies indicating that HF management in the elderly is deficient.28,29

Our laboratory findings demonstrated deteriorated kidney function in the group of patients aged 80 years and older, as compared with the younger population. The poorer kidney function in this group is naturally linked to the progressive aging of the kidneys.37 Higher NT-proBNP values in the older group indicate that, in this cohort, factors other than HF contribute to elevated NT-proBNP levels. It has been reported that in patients over 85 years old, comorbidities, such as abnormal renal function, obesity, and AF, have a significant effect on NT-proBNP levels.38

In the group aged 80 years and older, we also observed lower overall hemoglobin levels. Apart from age, factors contributing to lower hemoglobin levels include chronic kidney disease, which is more prevalent in the elderly, as shown in our study.39 Patients with chronic kidney disease produce less erythropoietin that stimulates red blood cell production.39

An interesting finding was better lipid profile in the older group, as compared with the younger population. The elderly exhibited higher HDL-C and lower triglyceride levels. It was proposed that a selection bias against HDL-C– lowering genetic variants might be responsible for its higher levels among the elderly patients.40 The poorer lipid profile in the younger group may also have been associated with a higher prevalence of overweight and obesity, as evidenced by BMI measurements.

In our study, long-term mortality was markedly higher in the patients aged 80 years and older. This is in line with previous studies pointing to age as an independent predictor of mortality in the HF population.41-45 The individuals aged 80 years and older had a 40% higher risk of cardiovascular mortality and twice as high risk of noncardiovascular mortality than the the younger population.29 In HF patients, mortality remains high and stable, despite advances in HF therapy.46 Nevertheless, Garred et al47 found a decrease in mortality in patients with HF over a 25-year observation period across all studied age groups in the Danish society; however, this trend was less visible in the patients aged 80 years and older.

We showed that in the older cohort, women not only had a significantly higher long-term survival rate than men, but female sex also was an independent predictor of better overall survival in the elderly. Data from the Global Congestive Heart Failure registry showed that women had a lower adjusted risk of mortality across various countries after stratifying by EF.48 The cause of this phenomenon is unclear, since the use of HF medications and cardiac tests, as well as hospitalization rates were comparable between the sexes.48 The results from the MAGGIC individual patient level meta-analysis also showed that survival was better for women with HF than men, irrespective of EF.49 While the results from the Euro Heart Failure Survey I did not show any difference between sex-related death rates in octogenarians,41 individual participant meta-analysis demonstrated that women had better survival and event-free survival rates than men in both over-80-year-old and below-80-year-old populations.29 In the CHART-2 (Chronic Heart Failure Analysis and Registry in the Tohoku District-2) study,43 the proportion of noncardiovascular deaths increased with age with significantly higher incidence of cancer and pneumonia deaths among men, as compared with women. Higher frequency of those deaths among men may partly explain increased all-cause mortality in this population.

Our study has several limitations. First, the registry contains data from a single tertiary center; therefore, our results should not be directly extrapolated to other populations. Nevertheless, our outcomes are comparable with other large-scale studies. Second, we did not analyze the causes of death as cardiovascular and noncardiovascular-related. Third, the infrequent use of sodium-glucose cotransporter 2 inhibitors during the study prevented us from analyzing the impact of this drug class on outcomes in the elderly patients.

Conclusions

The patients aged 80 years and older were characterized by a higher burden of comorbidities, including AF, hypertension, PAD, and renal failure. They were more often qualified for cardiac pacemakers, but less often for ICDs and invasive coronary procedures. The most common HF phenotype in this group was HFrEF, with a significant increase in HFpEF. They exhibited markedly higher overall mortality, as compared with the individuals below 80 years old. Finally, female sex was an independent predictor of lower long-term mortality.