Acute heart failure (AHF) is a life-threatening condition that requires urgent evaluation and management.1 The physician in charge of first care, whether in the pre-hospital or the in-hospital setting,2,3 must identify the high-risk patient, start immediate treatment as required, and evaluate whether admission to intensive care is warranted.1,4 The decision to admit the patient to the Intensive Care Unit (ICU) is usually established on an individual basis, and should be fundamented upon risk models corresponding to the specific disease process involved.5,6 In clinical practice, such decision making can be very complex, particularly in elderly patients without severe dependency for activities of daily living and no advance directives referred to the limitation of therapeutic effort.7,8

Acute heart failure is one of the most common situations seen in the hospital Emergency Department (ED), and is the leading cause of hospital admission in elderly people.9 The definition of an “elderly person” is arbitrary, and is fundamented upon sociodemographic factors. Biological age is a term that encompasses the consequences of changes related to aging and associated disease, and is therefore able to characterize the elderly population better than chronological age.10 Frailty in turn is a marker of the biological age of an individual, and is defined as a state of maximum vulnerability in the face of situations of stress. Frailty is associated to poor outcomes.11 Different authors have considered that frailty should be included in the risk models12,13 and ought to be taken into account in the making of decisions referred to patient care.1,14–17

The Identification of Senior at Risk (ISAR) scale is a multidimensional frailty screening tool that has been found to be useful in stratifying the short term risk of adverse outcomes in elderly patients discharged from hospital.18–21 To date, there is no evidence of the usefulness of the ISAR scale in predicting short-term mortality risk in elderly individuals with AHF at the time of first care in the ED. The present study was therefore carried out to assess the usefulness of the ISAR scale in establishing the 30-day prognosis among elderly patients with AHF seen in the ED.

The present study included 1059 patients of the 1100 cases of the OAK-3 register. Forty-one patients lacking data referred to the ISAR scale were excluded (Fig. 1).

Figure 1.

Flowchart of the patients included in the study.

(0.1MB).

The mean age of the patients included in the study was 85 (SD 5.9) years; 604 (57.2%) were women; and 701 (66.4%) had suffered a previous AHF episode. The mean ISAR score was 2.9 (SD 1.3) points. Eleven subjects (1.0%) yielded an ISAR score of 0 points, 149 (14.1%) a score of 1 point, 278 (26.3%) a score of 2 points, 260 (24.6%) a score of 3 points, 209 (19.7%) a score of 4 points, 122 (11.5%) a score of 5 points, and 30 (2.8%) a score of 6 points. Ninety-five patients (9.0%) died in the first 30 days following care in the ED. Tables S2a and S2b of the supplementary material (Appendix B annex 2) show the characteristics of the patients included in the study.

Tables S3a and S3b of the supplementary material (Appendix B annex 2) report the distribution of elderly patients with AHF according to the categories established by the ISAR scale. As the ISAR score increased, patient age was seen to rise, with a greater presence of women, arterial hypertension, chronic renal failure, cerebrovascular disease, previous heart failure episodes, dementia, severe comorbidity, physical frailty and dependency for activities of daily living, basal NYHA functional class III–IV, systolic blood pressure <110mmHg, tachypnea, basal oxygen saturation <90%, renal failure, anemia, hospital admission and 30-day mortality.

Fig. 2 shows the increase in frequency of 30-day mortality in relation to the ISAR category. The AUC of ISAR was 0.703 (95%CI: 0.655–0.751; p<0.001) for 30-day mortality. Table 1 reports the measures of effect, both crude and adjusted for the risk groups of the EFFECT scale, corresponding to each of the ISAR categories for the adverse event. Fig. 3 documents the predicted probabilities of the ISAR categories stratified by EFFECT scale risk groups. Table 2 shows the characteristics of the ISAR scale according to the different established cut-off points.

Figure 2.

Thirty-day mortality due to all causes in patients≥65 years of age seen for acute heart failure in the Emergency Departments according to the ISAR score.

(0.08MB).

Figure 3.

Predicted 30-day mortality due to all causes in patients≥65 years of age seen for acute heart failure in the Emergency Departments according to the ISAR scale stratified by EFFECT scale risk categories.

(0.11MB).

The present study offers important results: (1) the ISAR scale has good 30-day mortality predictive capacity in patients≥65 years of age with AHF seen in the ED; (2) the ISAR scale should not be applied with a pre-established cut-off point but rather considering that as the total score increases, there is a greater probability of a poorer short-term outcome; (3) the category of the ISAR scale is correlated to the degree of comorbidity, physical and cognitive frailty, and the severity of the episode; and (4) the ISAR scale affords information on variables related to the functional, cognitive and sensory situation of the patient, polypharmacy and previous contact with the healthcare system, generally not contemplated in AHF risk models, and which could prove useful in stratifying short-term risk in this group of patients.

The ISAR scale classically has been used as a frailty screening tool in elderly patients, with a view to predicting short-term adverse effects in individuals discharged directly from the ED, or even to select those patients that might benefit from integral geriatric assessment.19–21,25,26 Such integral assessment is the reference frailty diagnostic test, and has been shown to increase the probability of returning home and of being alive after 12 months among the hospitalized elderly.27 The consumption of human resources and the time needed to complete such assessment limit its application in the ED, however. Consequently, in the case of acute situations, it is preferable to use scales based on self-reporting questions or on the subjective clinical judgment of the healthcare professionals, and to postpone integral geriatric assessment until clinical stabilization has been achieved.14,15 Different frailty screening scales have been described, such as the ISAR scale,18 the Frail scale,28 the Clinical Frailty scale29 or the Tilburg Frailty Indicator.30 Despite the lack of evidence of these scales in application to AHF, the ISAR scale is recommended as the tool of choice in screening for frailty among elderly patients with AHF seen in the ED.14

The present study for the first time evaluates the validity of the ISAR scale in predicting the short-term outcome among elderly patients with AHF seen in the ED. The results show that the scale has good 30-day mortality predictive capacity. This fact is of clinical relevance, since this outcome variable has classically been used to derive the risk models that are of help in decision making referred to issues such as the need for hospital admission or the best place for patient admission.31 Furthermore, the scale has been found to contribute information on frailty to the EFFECT risk model, which focuses on cardiological variables.24 In this respect, it is known that risk models that do not consider variables related to biological age have only limited predictive capacity in the elderly population with AHF.31

The ISAR scale meets the requirements of simplicity and briefness for application in the emergency care setting. The scale comprises 6 dichotomic questions that are answered by the patient and/or caregiver at the time of first care, and which allow us to quantify the degree of multidimensional frailty at the patient bedside, without the need for a physical performance test (e.g., gait speed, up and go test, or the Short Physical Performance Battery), or for lengthy functional scales (e.g., the Barthel index) or short geriatric assessments (e.g., the Multidimensional Prognostic Index), which are sometimes difficult to apply in decision making in a life-threatening scenario.14 For example, it has been estimated that administration of the Barthel index takes about 5min. Therefore, the incorporation of this functional variable to AHF risk models such as the BI-EFFECT scale,32 the FBI-EFFECT scale13 or the MEESSI-AHF scale,33 largely complicates the applicability of these tools at the time of first care in patients with AHF in the emergency care setting.

To date, the ISAR scale has been used in clinical practice as a screening tool with a cut-off score of 2 points, which affords high sensitivity and limited specificity for the adverse outcome variables.19–21,34 Two recent meta-analyses have shown that this cut-off score of 2 points on the ISAR scale has a sensitivity of 97–99% and a specificity of 24–27%20,21 for 30-day mortality. The results of our study are consistent with this. Furthermore, we found a cut-off score of 3 to yield the greatest predictive capacity for this outcome variable. Modification of the ISAR cut-off score to 3 has already been commented by other authors.19–21 However, in our opinion, the scale should be regarded as a frailty index, i.e., as the sum of positive questions increases, so does the 30-day mortality risk.

With regard to the frequency of the different items of the ISAR scale, the item referred to serious memory problems was the least frequent (17.4%), while the use of three or more drugs was identified in practically all of the sample (96.3%). This latter observation caused the question referred to chronic medication to be non-discriminative in our sample. Previous studies have modified the polypharmacy cut-off point to 5, 6 or 8 drugs, without demonstrating clear improvement in the discriminative capacity of this scale.20,34,35

Our results are consistent with the data found in the literature regarding the high frequency of comorbidity, frailty and disability in elderly individuals with AHF.36,37 As in previous studies, the ISAR scale showed increasing scores to be associated to a greater frequency of elderly patients with important comorbidity and physical and cognitive frailty.34 Comorbidity, the cognitive situation and the degree of frailty and disability are poor outcome factors in AHF.16,32 This could explain the direct relationship observed between the ISAR score and the short-term mortality risk after adjusting for the risk categories of the EFFECT model, which includes variables related to age, comorbidity and data referred to the acute episode.

A notorious observation was the fact that two out of every three of the elderly patients had no serious dependency referred to activities of daily living. This shows that there is a potentially large group of elderly individuals who a priori could be admitted to intensive care, and that new parameters related to biological age therefore should be incorporated to the patient assessment process. An evident example of this circumstance is the application of the EAHFE-3D scale as a tool of help in identifying patients at high life-threatening risk,38 and therefore amenable to admission to the ICU. An age of 75 years or older is one of the 7 items on that scale, and is one of the three – along with respiratory failure and inotropic or vasopressor drug use – that contributes most points to the total score. In this regard, the universal application of this prognostic model, without considering frailty, could result in an inappropriate number of admissions to intensive care. In turn, the determination of this frailty index, in combination with other variables or risk models commonly used in clinical practice, could be of great help in decision making and for establishing individualized management plans for elderly patients with AHF.14,15

The present study has limitations. It is an exploratory study based on a multipurpose register – a fact that has limited the number of events required for the number of categories investigated. Selection bias cannot be ruled out, since systematic sampling was performed in which geriatric assessment adapted to the ED setting was made from 8a.m. to 10p.m. on working days. The ISAR data were extracted from the short geriatric assessment of the patient during acute phase care by different investigators in each center; some degree of interindividual variability therefore may exist. The information referred to prognostic biomarkers such as troponin or lactate was not available, since these parameters are not routinely compiled in all patients seen for AHF in Spanish EDs.39 The treatments prescribed during the acute phase and after discharge were left to the criterion of the supervising physicians, and this consequently may have influenced the results.40 Nevertheless, this circumstance allows the findings to be more real and thus more generalizable to clinical practice.

In conclusion, the ISAR scale is a simple tool that could be considered in the initial assessment of elderly patients with AHF in the ED, with a view to stratifying 30-day mortality risk.

FJMS designed the study. FJMS and OM obtained funding for the investigation. FJMS supervised conduction of the study and data collection. All the investigators of the OAK register recruited patients. FJMS managed the data, including the quality controls. FJMS, XR and CFP performed the statistical analysis of the data. FJMS wrote the manuscript, and XR, OM, MGC, ADR contributed substantially to its review. FJMS is responsible for the manuscript as a whole.

The present study has been possible in part thanks to aids PI15/00773, PI15/01019, PI11/01021 and PI10/01918 of the Instituto de Salud Carlos III with funds from the Spanish Ministerio of Sanidad, Servicios Sociales e Igualdad (MSSSI) and the European Regional Development Fund (ERDF); aids of the Generalitat de Catalunya for Consolidated Research Groups (GRC 2009/1385, 2014/0313 and 2017/1424); and the La Marató aid of TV3 (20152510). Xavier Rossello has been financed by the SEC-CNIC CARDIOJOVEN fellowship.

The authors declare that they have no conflicts of interest.