Natella et al.

BMC Geriatrics (2017) 17:34

DOI 10.1186/s12877-017-0419-2

RESEARCH ARTICLE Open Access

Long-term mortality in older patients

discharged after acute decompensated
heart failure: a prospective cohort study
Pierre-André Natella1,2*, Philippe Le Corvoisier3,4, Elena Paillaud1,5, Bertrand Renaud6, Isabelle Mahé7,8,
Jean-François Bergmann7,8, Hervé Perchet9, Dominique Mottier10,11, Olivier Montagne3,4†
and Sylvie Bastuji-Garin1,2,12†

Abstract
Background: Data are available on short- and intermediate-term mortality rates after discharge for acutely
decompensated heart failure (ADHF). However, few studies specifically addressed ADHF outcomes in patients aged
75 years or over, who contribute more than half of all ADHF admissions. Our objectives here were to estimate the
long-term mortality of patients aged 75 years or over who were discharged after admission for ADHF and to
identify factors, especially geriatric findings, independently associated with 2-year mortality.
Methods: This prospective cohort study in five French hospitals included consecutive patients aged 75 years or
older and discharged after emergency-department admission for ADHF meeting Framingham criteria (N = 478;
median age, 85 years; 68% female). Kaplan-Meier 1-year and 2-year survival curves were plotted. Admission
characteristics independently associated with overall 2-year mortality were identified using multivariable Cox
proportional-hazards regression.
Results: Mortality was 41.7% (95% confidence interval [95% CI], 37.2%–53.5%) after 1 year and 56.0% (95% CI, 51.
5%–60.7%) after 2 years. By multivariable analysis, independent predictors of 2-year mortality were male sex
(hazard ratio [HR], 1.36; 95% CI, 1.00–1.82), age >85 years (HR, 1.57; 95% CI, 1.19–2.07), higher number of impaired
activities of daily living (HR, 1.11 per impaired item; 95% CI, 1.05–1.17), recent weight loss (HR, 1.61; 95% CI, 1.14–
2.28), and lower systolic blood pressure (HR, 0.86 per standard deviation increase; 95% CI, 0.74–0.99). Creatinine
clearance ≤30 mL/min showed a trend toward an association with 2-year mortality (HR, 1.36; 95% CI, 0.97–2.00).
Conclusion: Functional impairment before admission is associated with higher long-term mortality in patients ≥75 years
admitted for ADHF. This study focused on geriatric markers not traditionally collected in heart-failure patients but did not
analyse all cardiologic parameters associated with outcomes in other studies. Nevertheless, our findings may contribute to
identify those patients admitted for ADHF who have the worst prognosis.
Keywords: Acute decompensated heart failure, Long-term mortality, Elderly

* Correspondence: pierre.natella@hmn.aphp.fr
†
Equal contributors
1
Université Paris Est (UPEC), A-TVB DHU, IMRB, EA7376, CEpiA Clinical
Epidemiology and Ageing unit, Créteil, France
2
AP-HP, Hôpital Henri Mondor, Service de Santé Publique, Créteil, France
Full list of author information is available at the end of the article

© The Author(s). 2017 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0
International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and
reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to
the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver
(http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.
Natella et al. BMC Geriatrics (2017) 17:34 Page 2 of 10

Background Strengthening the Reporting of Observational Studies in

Heart failure (HF) affects more than 15 million people Epidemiology (STROBE) statement [22].
in Europe, many of whom are elderly [1]. The heavy We estimated 1-year and 2-year mortality rates and we
burden of morbidity and mortality associated with HF is looked for associations linking admission characteristics,
comparable to that of many advanced malignancies [2]. including geriatric findings and co-morbidities, to 2-year
The mean life expectancy of patients discharged after mortality among hospital survivors.
admission for HF has been estimated at 5.5 years [3].
Several studies investigated the short- and intermediate-
Patients
term risk of death after discharge for acutely decompen-
As previously described [21], consecutive patients aged
sated heart failure (ADHF). In patients aged 65 years or
75 years or older and admitted to an emergency depart-
more, overall mortality ranged from 25% to 40%
ment with a diagnosis of ADHF during a 1-year study
after 1 year [4–15] and from 22% to 52.9% after
period were included in the ELISA cohort if they met
2 years [16–18]. However, few studies specifically ad-
Framingham criteria for HF (at least two major criteria
dressed ADHF outcomes in patients aged 75 years
or one major criterion plus two minor criteria) [23];
or over [9, 10], who contribute more than half of all
dyspnoea at rest or with minimal exertion; and an
patients admitted for ADHF [19]. Moreover, only
expected hospital stay duration ≥24 h. Exclusion cri-
two of these studies reported the associations linking
teria were ventricular arrhythmia at admission and
geriatric syndromes to 1-year mortality in elderly pa-
transfer to another hospital after the initial evaluation.
tients with HF [6, 16], and none investigated clinical
For the present study focusing on survival after hos-
and laboratory variables concomitantly with geriatric
pital discharge, we excluded the patients who died in
findings. These knowledge gaps need to be filled, as
the hospital.
the prevalence of chronic diseases increases with
age, so that most patients older than 75 years have
multiple co-morbidities [20]. We previously reported Data collection
co-morbidities and functional impairments independ- Included patients underwent a standardised clinical
ently associated with in-hospital mortality of older evaluation at the emergency department. However, some
patients admitted for ADHF (ELISA survey) [21]. Better parameters such as the get-up-and-go test were collected
knowledge of risk factors for long-term mortality may as early as possible after the relief of ADHF symptoms.
help to define follow-up and management goals and may Baseline data included socio-demographic characteristics,
improve treatment decision-making. medical history, clinical characteristics, and laboratory test
Here, our primary objective was to assess the long- results. Trained clinical research assistants recorded geri-
term survival of patients aged 75 years or over and atric parameters using validated tools. Nutritional status
discharged after in-patient treatment of ADHF. We was assessed using the Mini Nutritional Assessment-Short
also aimed to identify factors, especially geriatric find- Form [24] to classify patients into three nutritional risk
ings, independently associated with 2-year mortality. categories (≥12, well-nourished; 8-11, at risk; and <8, mal-
To achieve these objectives, we used data from the nourished) [25]. Weight loss over the last three months
previously described ELISA cohort study of severe (none, 1 to 3 kg, or >3 kg) was recorded from the patients
heart failure designed to identify patients with the or relatives then validated by medical record review. For
worst prognosis as early as possible during their hos- cognitive status, assessed using the MMSE, the cut-off of
pital stay [21]. 17 indicating severe cognitive impairment was chosen
[26–28]. Mood was assessed using the 15-item Geriatric
Methods Depression Scale (GDS), with scores ≥5 indicating a risk
Study design of depression [29, 30]. Functional status was assessed
ELISA is a prospective longitudinal cohort study of 680 using the Katz activities of daily living scale (ADL) with
older patients with ADHF seen at the emergency depart- six items (bathing, dressing, toileting, transferring, contin-
ments of five French hospitals [21], between October ence, and feeding) scored from 2 (able to perform the
2004 and December 2007. The ELISA cohort was activity) to 0 (unable to perform the activity) [31]. The
established in compliance with good clinical practice number of impaired ADL items was recorded, as well as
guidelines and was approved by the appropriate ethics the presence of functional impairment defined as a need
committee (institutional review board of the Henri- for assistance for at least one ADL (i.e., ADL score <12).
Mondor Teaching Hospital, Créteil, France). Cohort Impaired mobility was defined as a get-up-and-go test
participants or their relatives gave their written informed time >20 s or an inability to perform the test [32]. Glom-
consent for the collection of personal data to be used in erular filtration rate was calculated using the abbreviated
further analyses. The present report complies with the Modification of Diet in Renal Disease.
Natella et al. BMC Geriatrics (2017) 17:34 Page 3 of 10

Follow-up after hospital discharge Results

Patients were monitored by telephone calls at three- Among the 680 patients initially assessed for eligibility,
month intervals for 2 years after discharge or until death. 478 were analysed (Fig. 1). Table 1 reports their baseline
Vital status was collected from the next of kin, usual phys- characteristics. Median age was 85 years (range, 75-105
ician, or patient’s residential institution. The primary out- years). Most patients (91%) had one or more cardiovas-
come for the present analysis was overall 2-year mortality cular co-morbidities (myocardial infarction, stroke,
after discharge. hypertension, and/or arrhythmia). More than half of pa-
tients had abnormal nutritional status, function, and/or
mobility. Polypharmacy was the rule (median, 7 drugs).
Statistical analysis The 36 (7.5%) patients lost to follow-up were cen-
Sample size was estimated for the main objective of sored at the date of last information. Median follow-up
the ELISA cohort, namely, the identification of factors was 14.1 months (449 days [132–718 days]; range, 1–
associated with in-hospital mortality [21]. 770 days). Mortality was 41.7% (95% CI, 37.2%–53.5%)
All tests were two-sided, and P values ≤0.05 were after 1 year and 56.0% (95% CI, 51.5%–60.7%) after
considered significant. Analyses were performed using 2 years (Fig. 2). Median overall survival was 19 months
STATA software version 12.0 (StataCorp, College (1.6 years). Most deaths occurred early after hospital
Station, TX, USA). No imputation for missing data was discharge, after a median of about 6 months (184 days
performed. [68–367 days]).
Quantitative data are described as mean ± SD or me-
dian [25th-75th centiles], as appropriate, and categorical Factors associated with 2-year survival
variables as n (%). Overall survival was assessed from By univariate analysis (Table 2), factors significantly as-
discharge to death from any cause, 2 years after hospital sociated with mortality were older age (Fig. 3); male sex;
discharge or last follow-up for censored patients. We anaemia; lower systolic blood pressure; renal failure;
used the non-parametric Kaplan-Meier method to esti- lower body mass index, malnutrition, and recent weight
mate 1-year and 2-year survival rates with their 95% loss; and impairments in functional status and mobility.
confidence intervals (95% CIs). Trends (P < 0.15) were observed for history of myocardial
Characteristics of survivors and non-survivors were infarction and severe cognitive impairment. Therefore,
compared using Cox proportional hazards regression these variables were potential candidates for multivariable
models with estimation of hazard ratios (HRs) and analysis. Neither other cardiovascular co-morbidities
their 95% CIs. Log-linearity was tested for quantitative nor polypharmacy were associated with 2-year mortal-
variables. Analyses were routinely adjusted for age dichot- ity. No centre effect or effect modification by centre
omised based on the median value (>85 versus ≤85 years). was demonstrated.
Variables associated with P values <0.15 were selected for Age, sex, creatinine clearance, anaemia, systolic blood
multivariable analyses. Confounders and interactions pressure, myocardial infarction, weight loss, body mass
were tested in bivariate models. We also investigated a index, MNA-SF score, ADL score, Timed Get-Up-and-
potential centre effect. To avoid introducing corre-
lated variables into multivariable models, correlations
were assessed using Cramer’s V for categorical vari-
ables [33] and Spearman’s non-parametric rank correl-
ation (Rho) [34] for quantitative variables; values
above 0.50 were considered to indicate correlations.
The most relevant variables were identified based on
clinical relevance, number of missing values, and the
Akaike information criterion (AIC) [35, 36]. The pro-
portional hazards assumption was assessed both
graphically and statistically using the Schoenfeld resid-
uals test. This assumption was met for all variables in
the final models. Calibration and discrimination of the
final multivariable model were assessed using the
Gronnesby-Borgan goodness-of-fit test and Harrell’s c-
index, respectively [37, 38]. Model robustness was
tested using a non-parametric bootstrap resampling
procedure. The final Cox model was refit for the 500
Fig. 1 Participant flow chart
bootstrap samples thus created [39].
Natella et al. BMC Geriatrics (2017) 17:34 Page 4 of 10

Table 1 Baseline characteristics of elderly patients admitted for Table 1 Baseline characteristics of elderly patients admitted for
acute decompensated heart failure acute decompensated heart failure (Continued)
Characteristics Total Non-solid nutrition (blended or minced) 82 (18.1)
n = 478 (n = 215/238)
Socio-demographic characteristics Function and mobility
Age >85 years 222 (46) Number of impaired ADL items (n = 201/226) 2 [0–5]
Male sex 155 (32.4) ADL score <12, (n = 201/226) 268 (62.8)
Living arrangements (n = 220/244) Timed get-up-and-go >20 sc, (n = 178/210) 296 (76.2)
Lives alone 386 (83.2) Cognition
Institutionalised (16.8) MMSE ≤17, severe impairment (n = 221/251) 97 (20.6)
Current or former smoker (n = 211/242) 110 (24.3) Depression
Medical status GDS score ≥5 (n = 143/148) 142 (48.8)
Co-morbidities (past or current): Quantitative variables are expressed as median [25th - 75th centiles] and
categorical variables as N (%)
Arteritis (n = 207/234) 53 (12.0) (n= /) indicates the number of patients in each group in case of missing data
MNA-SF, Mini Nutritional Assessment-Short Form; ADL, activities of daily living
Myocardial infarction (n = 220/251) 97 (20.6)
scale; MMSE, Mini Mental State Examination; GDS, Geriatric Depression Scale
Stroke (n = 218/245) 51 (11.0) a
Hypertension was defined as blood pressure ≥140/90 mmHg or treatment
for hypertension
Anaemia (n = 209/233) 116 (26.2) b
Creatinine clearance was calculated using the abbreviated Modification of
Hypertensiona (n = 223/254) 354 (74.2) Diet in Renal Disease formula, glomerular filtration rate (mL/min/1.73 m2) =
186.3 × [creatinine (μmol/L) /88.4] -1.154 × [Age (years)]-0.203 × 0.742 (if female) ×
Cardiac arrhythmia (n = 215/245) 313 (68.0) 1.21 (if black)
c
Timed Get-Up-and-Go test >20 s or patient unable to perform the test
Diabetes (n = 218/253) 84 (17.8)
Systolic blood pressure, mmHg (n = 210/249) 146 [128–165] Go score, and MMSE score were available for the multi-
Number of drugs 7 [5–9] variable Cox model. Mobility impairment, lower body
≥ 5 drugs per day 382 (80.0) mass index, and MNA-SF category were not introduced
Minnesota Living with Heart Failure into multivariable models, as they correlated with func-
Questionnaire (n = 100/102) tional impairment and recent weight loss, respectively
< 24 (good quality of life) 54 (26.7) (correlation index >0.5, P < 0.05). Because of their as-
[24–45] (intermediate quality of life) 97 (48.0)
sociation with other parameters, anaemia, myocardial
infarction, and cognitive impairment were not inde-
> 45 (poor quality of life) 51 (25.2)
pendently associated with death in the multivariable
Laboratory parameters at admission model (P > 0.15). No significant interactions were
Sodium, mmol/L (n = 282/188) 138 [135–141] found between variables associated with 2-year mor-
Haemoglobin, g/dL (n = 212/247) 12.3 (11.1–13.6) tality. By multivariable analysis (Table 3), five factors
Creatinine clearance, mL/minuteb (n = 218/250) were independently associated with 2-year mortality,
≥ 60 213 (45.5)
namely, male sex, age older than 85 years, higher
number of impaired ADL items, recent weight loss,
]30–60[ 196 (41.9)
and lower systolic blood pressure. A trend was noted
≤ 30 59 (12.6) for renal failure. The final model had good calibration (P
Nutritional parameters value of the goodness-of-fit test >0.20) and acceptable dis-
Body mass index, Kg/m2 (n = 146/166) crimination (Harrell’s c-index, 0.64). The HRs estimated
< 19 10 (3.2) after bootstrap resampling were close to those of the ori-
[19–21 [ 18 (5.8)
ginal model, suggesting excellent internal validity. All five
factors were also independently associated with 1-year
[21–23 [ 39 (12.5)
mortality (Additional file 1: Table S1).
≥ 23 245 (78.5)
MNA-SF score (n = 69/81)
≥ 12 (well-nourished) 58 (38.7) Discussion
[8–12] (at risk) 73 (48.7) High 1-year and 2-year mortality rates of 41.7% and
56.0%, respectively, were documented in unselected eld-
< 8 (malnourished) 19 (12.7)
erly patients discharged alive after treatment for ADHF.
Recent weight loss >3 Kg (<3 months) (n = 201/226) 64 (15.0)
Independent risk factors for death within 2 years were
older age, male sex, prior functional impairment, low
Natella et al. BMC Geriatrics (2017) 17:34 Page 5 of 10

Fig. 2 Kaplan-Meier survival distribution, with 95% confidence intervals, of 478 patients aged 75 years or over and discharged after admission for
acutely decompensated heart failure

systolic blood pressure at admission, and recent weight admission, were also associated with in-hospital mortal-
loss. A trend was observed for renal failure. ity in the ELISA cohort [21].
One-year mortality rates close to 40% have been re- Several studies previously documented an independent
ported in subgroups of elderly patients in Canada, the adverse effect of male sex and older age on mortality in
United States, and Israel [5, 7, 10, 11, 14], in keeping elderly patients with HF, in keeping with our results [4,
with the 41.7% rate in our ELISA cohort. Other studies 5, 7, 8, 11, 16]. Similarly, lower systolic blood pressure
obtained variable results. In a Scottish study, 1-year [9–11, 13, 16–18] and renal dysfunction [9–11] were
mortality rates were 49% and 56% in the groups aged also previously associated with mortality.
75–84 years and >84 years, respectively [4]. In contrast, Few studies have assessed pre-admission functional
several studies from Spain, Canada, the United States, impairment as a prognostic marker in elderly patients
and Europe (EHFS II survey) found 1-year mortality with ADHF. The association of this factor with 2-year
rates of about 30% [6, 8, 9, 13]. These discrepancies may mortality in our population is consistent with two stud-
reflect differences in inclusion criteria regarding age, ies assessing associations between findings from a global
ADHF versus newly diagnosed HF, and presentation to geriatric assessment and 1-year mortality [6, 12]. These
the emergency department versus elsewhere. The 2-year two studies and ours produced similar results despite
mortality rate of 56% in our study is consistent with pre- using different functional assessment tools (Katz ADL,
vious reports from Brazil and the United States [17, 18]. Barthel index, or instrumental ADL), a fact that supports
Conversely, an Italian study [16] estimated 2-year mor- the validity of our results. That neither of the two previ-
tality at 22.9%. This difference may be ascribable to the ous studies found associations of mortality with other
younger mean age of 74 years and lesser severity of HF factors, including older age, male sex, systolic blood
(NYHA class II). pressure, and renal impairment, may be ascribable to
We identified several admission characteristics that limited statistical power (88 and 162 patients). Further-
predicted 2-year mortality in elderly patients admitted more, in the octogenarians of the EHFS II survey, dis-
for ADHF. The discrimination level of our final multi- ability (‘self-care problems’) independently predicted
variable model (Harrell’s c-index, 0.64) suggests an influ- long-term mortality in the multivariable analysis [9].
ence on long-term mortality of additional variables not Our results are consistent with those of studies in other
evaluated in our study. Indeed, other factors previously clinical settings, in which functional status predicted
associated with increased mortality in patients with HF, mortality independently from the underlying medical
e.g., type and duration of HF or medical therapy, may conditions. Interestingly, the 2-year risk of death in our
also have affected patient outcomes in our study. Three population increased by 11% for each additional im-
of the six parameters independently associated with 1- paired ADL item.
year and 2-year mortality, namely, lower systolic blood Another important finding from our study is the influ-
pressure, renal failure, and functional impairment before ence of nutritional status. Not only recent weight loss,
Natella et al. BMC Geriatrics (2017) 17:34 Page 6 of 10

Table 2 Comparison of survivors and non-survivors using age-adjusted Cox proportional hazards regression models
Characteristics Survivors Non-survivors Age-adjusted analysisa P value
n = 224 n = 254 HR [95%CI]
Socio-demographic characteristics
Age >85 years 88 (39.0) 134 (53) 1.49 [1.17–1.91] 0.01
Male sex 63 (28.1) 92 (36.2) 1.32 [1.02–1.72] 0.03
Living arrangements (n = 220/244)
Lives alone 188 (85.5) 198 (81.1) 1
Institutionalised 32 (14.5) 46 (18.9) 1.17 [0.84–1.61] 0.35
Current or former smoker (n = 211/242) 53 (25.1) 57 (23.6) 1.00 [0.74–1.34] 0.99
Medical status
Co-morbidities (past or current):
Arteritis (n = 207/234) 19 (9.2) 34 (14.5) 1.28 [0.89–1.85] 0.18
Myocardial infarction (n = 220/251) 36 (16.4) 61 (24.3) 1.27 [0.95–1.69] 0.11
Stroke (n = 218/245) 20 (9.2) 31 (12.7) 1.29 [0.86–1.83] 0.24
Anaemia (n = 209/233) 44 (21.1) 72 (30.9) 1.34 [1.02–1.77] 0.04
b
Hypertension (n = 223/254) 166 (74.4) 188 (74.0) 0.96 [0.72–1.27] 0.76
Cardiac arrhythmia (n = 215/245) 141 (65.6) 172 (70.2) 1.02 [0.77–1.35] 0.87
Diabetes (n = 218/253) 43 (19.7) 41 (16.2) 0.93 [0.66–1.31] 0.70
Systolic blood pressure, mmHg (n = 210/249)c 150 [135–170] 141 [121–162] 0.83 [0.73–0.96] 0.01
Number of drugs 7 [5–9] 7 [5–9]
≥ 5 drugs per day 171 (76.3) 211 (83.1) 1.26 [0.91–1.75] 0.17
Minnesota Living with Heart Failure Questionnaire (n = 100/102)
< 24 (good quality of life) 29 (29.0) 25 (24.5) 1 0.69
[24–45] (intermediate quality of life) 46 (46.0) 51 (50.0) 1.16 [0.72–1.88]
> 45 (poor quality of life) 25 (25.0) 26 (25.5) 1.27 [0.73–2.21]
Laboratory parameters at admission
Sodium, mmol/L (n = 282/188) 138 [135–141] 138 [135–141] 0.99 [0.97–1.01] 0.43
Haemoglobin, g/dL (n = 212/247) 12.5 [11.3–13.7] 12.1 [10.9–13.4] 0.97 [0.92–1.04] 0.40
Creatinine clearance, mL/minuted (n = 218/250)
≥ 60 117 (53.7) 96 (38.4) 1
[30–60] 83 (38.1) 113 (45.2) 1.28 [0.97–1.68] 0.08
≤ 30 18 (8.3) 41 (16.4) 1.81 [1.25–2.61] <0.01
Nutritional parameters
Body mass index, Kg/m2 (n = 146/166)
< 19 2 (1.4 8 (4.8) 2.30 [1.11–4.74] 0.03
[19–21[ 5 (3.4) 13 (7.8) 1.50 [0.84–2.67] 0.19
[21–23[ 14 (9.6) 25 (15.1) 1.50 [0.98–2.31] 0.08
≥ 23 125 (85.6) 120 (72.3) 1
MNA-SF score (n = 69/81)
≥ 12 (well-nourished) 33 (47.8) 25 (30.9) 1
[8–12[(at risk) 31 (44.9) 42 (51.9) 1.64 [0.99–2.70] 0.05
< 8 (malnourished) 5 (7.2) 14 (17.3) 3.01 [1.52–5.97] <0.01
Recent weight loss >3 Kg (<3 months) (n = 201/226) 22 (10.9) 42 (18.6) 1.59 [1.22–2.06] <0.01
Non-solid nutrition (blended or minced) (n = 215/238) 31 (14.4) 51 (21.4) 1.60 [0.94–1.76] 0.11
Natella et al. BMC Geriatrics (2017) 17:34 Page 7 of 10

Table 2 Comparison of survivors and non-survivors using age-adjusted Cox proportional hazards regression models (Continued)
Function and mobility
Number of impaired ADL items (n = 201/226)e 1 [0–4] 2 [0–5] 1.10 [1.05–1.17] <0.01
ADL score <12, (n = 201/226) 106 (52.7) 162 (71.7) 1.61 [1.20–2.16] <0.01
Timed get-up-and-go >20 sf, (n = 178/210) 125 (70.2) 171 (81.4) 1.49 [1.05–2.11] 0.03
Cognition
MMSE ≤17, severe impairment (n = 221/251) 38 (17.2) 59 (23.5) 1.26 [0.94–1.69] 0.12
Depression
GDS score ≤5 (n = 143/148) 71 (49.7) 71 (48.0) 0.96 [0.70–1.33] 0.81
HR, hazards ratio; CI, confidence interval; MNA-SF, Mini Nutritional Assessment-Short Form; ADL, activities of daily living scale; MMSE, Mini Mental State
Examination; GDS, Geriatric Depression Scale
a
Hazards ratios and confidence intervals were estimated using Cox proportional models adjusted for age (≤85 years versus >85 years)
b
Hypertension was defined as blood pressure ≥140/90 mmHg or treatment for hypertension
c
Hazards ratios and confidence intervals per increase by 1 standard deviation
d
Creatinine clearance was calculated using the abbreviated Modification of Diet in Renal Disease formula, glomerular filtration rate (mL/min/1.73 m2) =
186.3 × [creatinine (μmol/L) /88.4] -1.154 × [Age (years)]-0.203 × 0.742 (if female) × 1.21 (if black)
e
Hazards ratios and confidence intervals per each additional impaired ADL item
f
Timed Get-Up-and-Go test >20 s or patient unable to perform the test

but also malnutrition or risk of malnutrition, predicted recent weight loss is part of the Fried phenotype, and
mortality independently from age. The only previous functional impairment is related to the slowness and low
study assessing the prognostic value of MNA results in physical activity criteria [43]. Furthermore, impairment
elderly patients with HF found no association with mor- of one ADL item corresponds to the 6th and 7th cat-
tality [6]. However, in older adults, malnutrition is often egories of the Canadian Study of Health and Aging
related to adverse health outcomes and strongly predicts frailty scale [44]. Frailty assessed using both a frailty
mortality [40, 41]. Weight loss is common in end-stage index and the Fried phenotype predicted mortality
HF and may reflect numerous mechanisms including among community-dwelling patients with HF (mean
neurohormonal dysregulation and an imbalance between age, 71 years) [45]. These results suggest that elderly
anabolism and catabolism [42]. Our results show that re- patients with HF may require a specific functional and
cent weight loss is a major prognostic marker in elderly nutritional evaluation to allow the development of a cus-
patients successfully treated for an episode of ADHF. tomised treatment plan aimed at improving outcomes. A
Two of the six factors independently associated with number of measures might prove useful. For example,
mortality are recognized indicators of fragility. Thus, studies in chronic HF showed that exercise training pro-
duced statistically significant improvements in self-
reported health status among both younger adults [46]
and elderly individuals [47], although mortality was not
significantly affected. Furthermore, the programmes
for elderly patients focused mainly on aerobic and re-
sistance training and did not include a routine assess-
ment of nutritional and functional status [47]. The
cost-effectiveness of a multidisciplinary disease man-
agement programme has been evaluated in elderly
outpatients with HF and mild-to-moderate frailty [48].
Interventional studies might be useful to assess whether
similar programmes decrease mortality and/or improve
function and quality of life of elderly patients discharged
after an ADHF episode.
In a previous study, cognitive impairment was associated
with higher long-term mortality in elderly patients with
chronic HF, but no multivariable analysis was reported [49].
In our cohort, a crude association was observed, but only a
trend towards an association persisted in the age-adjusted
analysis and there was no significant association by multi-
variable analysis. Other prognostic factors in elderly pa-
Fig. 3 Kaplan-Meier survival distributions according to age
tients with ADHF have been reported [4, 5, 7, 9–11, 17].
Natella et al. BMC Geriatrics (2017) 17:34 Page 8 of 10

Table 3 Factors independently associated with 2-year mortality by multivariable analysis (n = 399)
Characteristics Model developmenta Parameter estimates after bootstrapping methods
HR 95% CI P value Mean HR 95% CI P value
Male sex 1.36 [1.00–1.82] 0.05 1.36 [1.00–1.83] 0.05
Age >85 years 1.57 [1.19–2.07] <0.01 1.58 [1.19–2.08] <0.01
Number of impaired ADLb items 1.11 [1.05–1.17] <0.01 1.11 [1.04–1.18] <0.01
c
Recent weight loss 1.61 [1.14–2.28] <0.01 1.61 [1.12–2.32] 0.01
Systolic blood pressure (mmHg)d 0.86 [0.74–0.99] 0.04 0.85 [0.73–1.00] 0.05
Creatinine clearance ≤30 mL/minute e
1.36 [0.97–2.00] 0.09 1.38 [0.89–2.08] 0.12
HR, hazards ratio; CI, confidence interval; ADL, activities of daily living scale
a
Hazards ratios and confidence intervals were estimated using Cox proportional models simultaneously adjusted for all variables listed in the table
b
per additional impaired ADL item
c
>3 Kg within the 3 months preceding admission
d
per increase by 1 standard deviation
e
Creatinine clearance was calculated using the abbreviated Modification of Diet in Renal Disease formula, glomerular filtration rate (mL/min/1.73 m2) =
186.3 × [creatinine (μmol/L) /88.4] -1.154 × [Age (years)]-0.203 × 0.742 (if female) × 1.21 (if black)

We found no significant associations with long-term mor- Conclusion

tality for other factors, such as diabetes, myocardial infarc- Independent predictors of 2-year mortality in unselected
tion, stroke, and hyponatremia. These results may be elderly patients discharged after inhospital treatment for
related to the inclusion of functional impairment in our ADHF included not only older age, male sex, lower systolic
multivariable model and to differences in inclusion criteria blood pressure, and lower creatinine clearance; but also
across studies. geriatric markers of frailty, such as functional impairment
and recent weight loss. An assessment of frailty should
therefore be considered an integral part of the evaluation
Strength and limitations of elderly patients admitted for ADHF. For this frail popu-
The inclusion in a vast multicentre cohort of unse- lation, an interdisciplinary approach targeting the multidi-
lected elderly patients with successful in-hospital mensional aspects of health may improve outcomes.
treatment for ADHF and the low lost-to-follow-up
rate over 2 years (7.5%) support the validity of our re-
Additional file
sults. Furthermore, the inclusion of consecutive pa-
tients limited the risk of selection bias, and the main Additional file 1: Table S1. Factors independently associated with
endpoint (overall mortality) is a robust criterion that 1-year mortality by multivariable analysis (n = 399). (DOC 40 kb)
leaves little room for classification bias. Finally, we
routinely adjusted all analyses for age as a potential
Abbreviations
confounder, and we took interactions and confounding ADHF: Acutely Decompensated Heart Failure; ADL: Activities of Daily Living
into account in the analyses. The similar HR values scale; AIC: Akaike Information Criterion; BNP: Brain Natriuretic Peptide;
obtained after bootstrapping resampling procedures CI: Confidence Interval; GDS: Geriatric Depression Scale; HF: Heart Failure;
HR: Hazard Ratio; MMSE: Mini Mental State Examination; MNA-SF: Mini
further support the robustness of our findings. More- Nutritional Assessment-Short Form; NT-proBNP: N-Terminal pro-Brain Natri-
over, few data exist on long-term outcomes of patients uretic Peptide; NYHA: New York Heart Association classification; SD: Standard
older than 75 years after successful inhospital ADHF Deviation; STROBE: Strengthening the Reporting of Observational Studies in
Epidemiology
management. A limitation of our study is the absence
in the analysis of several factors having previously Acknowledgements
reported associations with increased mortality, such We thank Katia Le Dudal for technical support and A Wolfe, MD, for editing
as, ventricular function, BNP or NT-proBNP concen- the manuscript.
tration, and treatments. Several studies suggest that a
score including NT-proBNP may predict post-discharge Funding
This work was supported by a grant from the Hospital Clinical Research
1-year mortality in ADHF populations with a mean age of Program of the French Ministry of Health (PHRC AOM 02-109). The funders
about 60-75 years; therefore, in our population of very had no role in study design, data collection and analysis, decision to publish,
elderly patients, using this score to adjust the fragility or preparation of the manuscript.
indicators associated with post-discharge mortality
would have been of interest [15]. However, our findings Availability of data and materials
Assistance Publique-Hôpitaux de Paris (AP-HP) owns the data, of which any
demonstrate the prognostic value of geriatric characteris- use or transmission to a third party cannot be made without its prior
tics in elderly patients with ADHF. agreement.
Natella et al. BMC Geriatrics (2017) 17:34 Page 9 of 10

Authors’ contributions 9. Komajda M, Hanon O, Hochadel M, Lopez-Sendon JL, Follath F, Ponikowski

Individual contribution of each author: Study conception and design: OM, P, et al. Contemporary management of octogenarians hospitalized for heart
BR, and PLC. Acquisition of data: BR, IM, JFB, HP, DM, and OM. Analysis and failure in Europe: Euro Heart Failure Survey II. Eur Heart J. 2009;30(4):478–86.
interpretation of data: PAN, SBG, PLC, OM and EP. Statistical expertise: SBG. 10. Barsheshet A, Shotan A, Cohen E, Garty M, Goldenberg I, Sandach A, et al.
Drafting the manuscript: PAN, SBG, PLC, EP and OM. Critical revision of the Predictors of long-term (4-year) mortality in elderly and young patients with
manuscript: all authors. All authors read and approved the final manuscript. acute heart failure. Eur J Heart Fail. 2010;12(8):833–40.
11. Kociol RD, Horton JR, Fonarow GC, Reyes EM, Shaw LK, O’Connor CM, et al.
Competing interests Admission, Discharge, or Change in B-Type Natriuretic Peptide and Long-
The authors declare that they have no competing interests. Term Outcomes Data From Organized Program to Initiate Lifesaving
Treatment in Hospitalized Patients With Heart Failure (OPTIMIZE-HF) Linked
Consent for publication to Medicare Claims. Circ Heart Fail. 2011;4(5):628–36.
Not applicable as our manuscript does not contain data from any individual 12. Delgado Parada E, Suárez García FM, López Gaona V, Gutiérrez Vara S,
person. Solano Jaurrieta JJ. Mortality and functional evolution at one year after
hospital admission due to heart failure (HF) in elderly patients. Arch
Ethics approval and consent to participate Gerontol Geriatr. 2012;54(1):261–5.
The ELISA cohort was established in compliance with good clinical practice 13. Huynh BC, Rovner A, Rich MW. Long-term Survival in Elderly Patients
guidelines and was approved by the appropriate ethics committee Hospitalized for Heart Failure: 14-Year Follow-up From a Prospective
(institutional review board of the Henri-Mondor Teaching Hospital, Créteil, Randomized Trial. Arch Intern Med. 2006;166(17):1892.
France). Cohort participants or their relatives gave their written informed 14. Saczynski JS, Darling CE, Spencer FA, Lessard D, Gore JM, Goldberg RJ. Clinical
consent for the collection of personal data to be used in further analyses. Features, Treatment Practices, and Hospital and Long-Term Outcomes of Older
Patients Hospitalized with Decompensated Heart Failure: The Worcester Heart
Author details Failure Study. J Am Geriatr Soc. 2009;57(9):1587–94.
1
Université Paris Est (UPEC), A-TVB DHU, IMRB, EA7376, CEpiA Clinical 15. Scrutinio D, Ammirati E, Guida P, Passantino A, Raimondo R, Guida V, et al.
Epidemiology and Ageing unit, Créteil, France. 2AP-HP, Hôpital Henri The ADHF/NT-proBNP risk score to predict 1-year mortality in hospitalized
Mondor, Service de Santé Publique, Créteil, France. 3Inserm, Centre patients with advanced decompensated heart failure. J Heart Lung
d’Investigation Clinique, 1430 Créteil, France. 4AP-HP, Hôpital Henri Mondor, Transplant. 2014;33(4):404–11.
Pôle Vigilance Recherche Méthodologie & Information Médicale, 94010 16. Cacciatore F, Abete P, Mazzella F, Viati L, Della Morte D, D’Ambrosio D, et al.
Créteil, France. 5AP-HP, Hôpital Henri Mondor, Département de Médecine Frailty predicts long-term mortality in elderly subjects with chronic heart
Interne et Gériatrie, Créteil, France. 6AP-HP, Hôpital Henri Mondor, Structure failure. Eur J Clin Invest. 2005;35(12):723–30.
des Urgences, Créteil, France. 7AP-HP, Hôpital Lariboisière-Fernand Widal, 17. VillaCorta H, Mesquita ET, Cardoso R, Bonates T, Maia ER, Silva AC, et al.
Département de Médecine Interne, Paris, France. 8Université Paris 7, EA Emergency department predictors of survival in decompensated heart
REMES Université Paris Diderot, Sorbonne Paris Cité, Paris, France. 9Centre failure patients. Rev Port Cardiol. 2003;22(4):495–507.
hospitalier de Meaux, Service de Cardiologie, Meaux, France. 10CHU Brest, 18. Goldberg RJ. Long-term Survival After Heart Failure: A Contemporary
Hôpital Cavale Blanche, Département de Médecine Interne et de Population-Based Perspective. Arch Intern Med. 2007;167(5):490.
Pneumologie, Brest, France. 11Université de Bretagne Occidentale, EA 3878 19. Hugli O, Braun JE, Kim S, Pelletier AJ, Camargo CA. United States Emergency
(GETBO), Brest, France. 12AP-HP, Hôpital Henri Mondor, Unité de Recherche Department Visits for Acute Decompensated Heart Failure, 1992 to 2001.
Clinique, Créteil, France. Am J Cardiol. 2005;96(11):1537–42.
20. Stenholm S, Westerlund H, Head J, Hyde M, Kawachi I, Pentti J, et al.
Received: 18 June 2016 Accepted: 13 January 2017 Comorbidity and Functional Trajectories From Midlife to Old Age: The
Health and Retirement Study. J Gerontol A Biol Sci Med Sci. 2014;24:glu113.
21. Le Corvoisier P, Bastuji-Garin S, Renaud B, Mahé I, Bergmann J-F, Perchet H,
References et al. Functional status and co-morbidities are associated with in-hospital
1. McMurray JJV, Adamopoulos S, Anker SD, Auricchio A, Böhm M, Dickstein K, mortality among older patients with acute decompensated heart failure: a
et al. ESC guidelines for the diagnosis and treatment of acute and chronic multicentre prospective cohort study. Age Ageing. 2014;12.
heart failure 2012: The Task Force for the Diagnosis and Treatment of Acute 22. von Elm E, Altman DG, Egger M, Pocock SJ, Gøtzsche PC, Vandenbroucke
and Chronic Heart Failure 2012 of the European Society of Cardiology. JP. The Strengthening the Reporting of Observational Studies in
Developed in collaboration with the Heart Failure Association (HFA) of the Epidemiology (STROBE) Statement: Guidelines for Reporting Observational
ESC. Eur J Heart Fail. 2012;14(8):803–69. Studies. Ann Intern Med. 2007;147(8):573–7.
2. Askoxylakis V, Thieke C, Pleger ST, Most P, Tanner J, Lindel K, et al. Long- 23. McKee PA, Castelli WP, McNamara PM, Kannel WB. The natural history of
term survival of cancer patients compared to heart failure and stroke: a congestive heart failure: the Framingham study. N Engl J Med. 1971;285(26):
systematic review. BMC Cancer. 2010;10:105. 1441–6.
3. Alter DA, Ko DT, Tu JV, Stukel TA, Lee DS, Laupacis A, et al. The Average 24. Rubenstein LZ, Harker JO, Salvà A, Guigoz Y, Vellas B. Screening for
Lifespan of Patients Discharged from Hospital with Heart Failure. J Gen Undernutrition in Geriatric Practice Developing the Short-Form Mini-
Intern Med. 2012;27(9):1171–9. Nutritional Assessment (MNA-SF). J Gerontol A Biol Sci Med Sci. 2001;56(6):
4. MacIntyre K, Capewell S, Stewart S, Chalmers JWT, Boyd J, Finlayson A, et al. M366–72.
Evidence of Improving Prognosis in Heart Failure Trends in Case Fatality in 25. Kaiser MJ, Bauer JM, Ramsch C, Uter W, Guigoz Y, Cederholm T, et al.
66 547 Patients Hospitalized Between 1986 and 1995. Circulation. 2000; Validation of the Mini Nutritional Assessment short-form (MNA-SF): a
102(10):1126–31. practical tool for identification of nutritional status. J Nutr Health Aging.
5. Jong P, Vowinckel E. Prognosis and determinants of survival in patients 2009;13(9):782–8.
newly hospitalized for heart failure: A population-based study. Arch Intern 26. Folstein MF, Folstein SE, McHugh PR. “Mini-mental state”. A practical
Med. 2002;162(15):1689–94. method for grading the cognitive state of patients for the clinician.
6. Formiga F, Chivite D, Solé A, Manito N, Ramon JM, Pujol R. Functional J Psychiatr Res. 1975;12(3):189–98.
outcomes of elderly patients after the first hospital admission for 27. Derouesne C, Poitreneau J, Hugonot L, Kalafat M, Dubois B, Laurent B. Mini-
decompensated heart failure (HF). A prospective study. Arch Gerontol Mental State Examination:a useful method for the evaluation of the
Geriatr. 2006;43(2):175–85. cognitive status of patients by the clinician. Consensual French version.
7. Curtis LH, Greiner MA, Hammill BG, Kramer JM, Whellan DJ, Schulman KA, et Presse Med. 1999;28(21):1141–8.
al. Early and Long-term Outcomes of Heart Failure in Elderly Persons, 2001- 28. Bastuji-Garin S, Joly P, Lemordant P, Sparsa A, Bedane C, Delaporte E, et al.
2005. Arch Intern Med. 2008;168(22):2481. Risk Factors for Bullous Pemphigoid in the Elderly: A Prospective Case–
8. Ezekowitz JA, Bakal JA, Kaul P, Westerhout CM, Armstrong PW. Acute heart Control Study. J Invest Dermatol. 2011;131(3):637–43.
failure in the emergency department: Short and long‐term outcomes of 29. D’Ath P, Katona P, Mullan E, Evans S, Katona C. Screening, detection and
elderly patients with heart failure. Eur J Heart Fail. 2008;10(3):308–14. management of depression in elderly primary care attenders. I: The
Natella et al. BMC Geriatrics (2017) 17:34 Page 10 of 10

acceptability and performance of the 15 item Geriatric Depression Scale

(GDS15) and the development of short versions. Fam Pract. 1994;11(3):260–6.
30. de Craen AJM, Heeren TJ, Gussekloo J. Accuracy of the 15-item geriatric
depression scale (GDS-15) in a community sample of the oldest old.
Int J Geriatr Psychiatry. 2003;18(1):63–6.
31. Katz S, Ford AB, Moskowitz RW, Jackson BA, Jaffe MW. Studies of illness in
the aged. The index of ADL: a standardized measure of biological and
psychological function. JAMA J Am Med Assoc. 1963;185:914–9.
32. Podsiadlo D, Richardson S. The timed “Up & Go”: a test of basic functional
mobility for frail elderly persons. J Am Geriatr Soc. 1991;39(2):142–8.
33. Cramér H. Mathematical methods of statistics. Princeton: Princeton
University Press; 1946.
34. Spearman C. The Proof and Measurement of Association between Two
Things. Am J Psychol. 1904;15(1):72.
35. Akaike H. Information theory and an extension of the maximum likelihood
principle. Second Int Symp Inf Theory. 1973;267–81
36. Akaike H. A new look at the statistical model identification. IEEE Trans
Autom Control. 1974;19(6):716–23.
37. May S, Hosmer DW. Hosmer and Lemeshow type goodness-of-fit statistics
for the Cox proportional hazards model. In: Handbook of statistics 23:
advances in survival analysis. Amsterdam: Elsevier B. V.; 2004. pp. 383-94.
38. Royston P, Altman DG. External validation of a Cox prognostic model:
principles and methods. BMC Med Res Methodol. 2013;13:33.
39. Chen CH, George SL. The bootstrap and identification of prognostic factors
via Cox’s proportional hazards regression model. Stat Med. 1985;4(1):39–46.
40. Kiesswetter E, Pohlhausen S, Uhlig K, Diekmann R, Lesser S, Uter W, et al.
Prognostic differences of the Mini Nutritional Assessment short form and
long form in relation to 1-year functional decline and mortality in
community-dwelling older adults receiving home care. J Am Geriatr Soc.
2014;62(3):512–7.
41. Kagansky N, Berner Y, Koren-Morag N, Perelman L, Knobler H, Levy S. Poor
nutritional habits are predictors of poor outcome in very old hospitalized
patients. Am J Clin Nutr. 2005;82(4):784–91.
42. Pureza V, Florea VG. Mechanisms for cachexia in heart failure. Curr Heart Fail
Rep. 2013;10(4):307–14.
43. Fried LP, Tangen CM, Walston J, Newman AB, Hirsch C, Gottdiener J, et al.
Frailty in Older Adults Evidence for a Phenotype. J Gerontol A Biol Sci Med
Sci. 2001;56(3):M146–57.
44. Rockwood K, Song X, MacKnight C, Bergman H, Hogan DB, McDowell I, et
al. A global clinical measure of fitness and frailty in elderly people. CMAJ
Can Med Assoc J. 2005;173(5):489–95.
45. McNallan, SM, Chamberlain AM, Gerber Y, Singh M, Kane RL, Weston SA, et
al. Measuring frailty in heart failure: a community perspective. Am Heart J.
2013;166(4):768–74.
46. Flynn KE, Piña IL, Whellan DJ, Lin L, Blumenthal JA, Ellis SJ, et al. Effects of
Exercise Training on Health Status in Patients With Chronic Heart Failure:
Findings From the HF-ACTION Randomized Controlled Trial. JAMA J Am
Med Assoc. 2009;301(14):1451–9.
47. Chen YM, Li Y. Safety and efficacy of exercise training in elderly heart failure
patients: a systematic review and meta-analysis. Int J Clin Pract. 2013;67(11):
1192–8.
48. Pulignano G, Del Sindaco D, Di Lenarda A, Tarantini L, Cioffi G, Gregori
D, et al. Usefulness of frailty profile for targeting older heart failure
patients in disease management programs: a cost-effectiveness, pilot
study. J Cardiovasc Med Hagerstown Md. 2010;11(10):739–47.
49. Zuccalà G, Pedone C, Cesari M, Onder G, Pahor M, Marzetti E, et al. The
effects of cognitive impairment on mortality among hospitalized patients
with heart failure. Am J Med. 2003;115(2):97–103. Submit your next manuscript to BioMed Central
and we will help you at every step:
• We accept pre-submission inquiries
• Our selector tool helps you to find the most relevant journal
• We provide round the clock customer support
• Convenient online submission
• Thorough peer review
• Inclusion in PubMed and all major indexing services
• Maximum visibility for your research

Submit your manuscript at

www.biomedcentral.com/submit