European Journal of Heart Failure (2015) 17, 1032–1041 doi:10.1002/ejhf.290
The 30-day metric in acute heart failure revisited: data from IN-HF Outcome, an Italian nationwide cardiology registry Giuseppe Di Tano1*, Renata De Maria2, Lucio Gonzini3, Nadia Aspromonte4, Andrea Di Lenarda5, Mauro Feola6, Marco Marini7, Massimo Milli8, Gianfranco Misuraca9, Andrea Mortara10, Fabrizio Oliva11, Giovanni Pulignano12, Giulia Russo5, Michele Senni13, and Luigi Tavazzi14, on the behalf of the IN-HF Outcome Investigators† 1 Cardiology
Department, Istituti Ospitalieri, Cremona, Italy; 2 CNR Clinical Physiology Institute, Cardiothoracic and Vascular Department, Azienda Ospedaliera Niguarda Ca’ Granda, Milan, Italy; 3 ANMCO Research Centre, Florence, Italy; 4 Cardiology Department, Ospedale San Filippo Neri, Rome, Italy; 5 Cardiovascular Center, Health Authority n. 1 and University of Trieste, Trieste, Italy; 6 Cardiovascular Rehabilitation, Heart Failure Unit, Ospedale Maggiore SS, Trinità, Fossano, Italy; 7 Cardiology Department, Ospedali Riuniti, Umberto I–Lancisi–Salesi, Ancona, Italy; 8 Cardiology, Ospedale Santa Maria Nuova, Florence, Italy; 9 Cardiology Department, Ospedale dell’Annunziata, Cosenza, Italy; 10 Cardiology Department, Policlinico di Monza, Monza, Italy; 11 Cardiologia 2 Heart Failure and Heart Transplant Programme, ‘A. De Gasperis’ Cardiothoracic and Vascular Department, Azienda Ospedaliera Niguarda Ca’ Granda, Milan, Italy; 12 Heart Failure Clinic, 1st Cardiology/CCU Unit, Cardiovascular Department, San Camillo Hospital, Rome, Italy; 13 Cardiovascular Department, Cardiology 1, Papa Giovanni XXIII Hospital, Bergamo; and 14 Maria Cecilia Hospital–GVM Care&Research–E.S. Health Science Foundation, Cotignola, Italy Received 29 December 2014; revised 17 April 2015; accepted 27 April 2015 ; online publish-ahead-of-print 27 May 2015
Aims
Unplanned readmissions early after a discharge from acute heart failure hospitalization are common and have become a reimbursement benchmark and marker of hospital quality. However, the competing risk of short-term post-discharge mortality is substantial. ..................................................................................................................................................................... Methods Using data from the prospective, nationwide Registry IN-HF Outcome, we analysed the incidence and predictors and results of 30-day mortality or readmissions and associated days-alive-out-of-hospital (DAOH) in 1520 patients discharged alive after admission for acute heart failure. Within 30 days after discharge, 94 patients (6.2%) were readmitted (91% for cardiovascular causes; 60% recurrent heart failure) and 42 (2.8%) died, 10 of which occurred during readmission. Overall, 126 patients (8.3%) met the combined endpoint. By multivariable logistic regression, worsening chronic heart failure as clinical presentation [odds ratio (OR) 1.83, 95% confidence interval (CI) 1.21–2.77, P = 0.005), inotropes during admission (OR 2.19, 95% CI 1.40–3.43, P = 0.0006), length of stay (OR 1.02, 95% CI 1.01–1.04, P = 0.002) and renin–angiotensin system inhibitors at discharge (OR 0.52, 95%CI 0.35–0.77, P = 0.001) independently predicted 30-day all-cause mortality and/or readmission (c-statistic = 0.695). Per cent 30-day DAOH was lower in patients with in-hospital inotrope use, no renin–angiotensin system inhibitors prescription at discharge, New York Heart Association III–IV class at discharge, and correlated inversely with length of stay and age. ..................................................................................................................................................................... Conclusion A clinical and biohumoral profile consistent with chronic advanced heart failure and end-organ damage identifies acute heart failure patients discharged home from cardiology units, who are at highest risk of early death and/or readmission. These findings have practical implications for tailoring specific follow-up.
.......................................................................................................... Keywords
Acute heart failure •
mortality •
30-Day readmission •
Days-alive-out-of-hospital
*Corresponding author: Tel: +39 0372 405323; Fax: +39 0372 433787; E-mail: [email protected], [email protected] † See Appendix S1 for a complete list of participating centres and investigators.
© 2015 The Authors European Journal of Heart Failure © 2015 European Society of Cardiology
1033
Introduction Hospital admission for heart failure (HF) is the most common cause of hospitalization in patients aged over 65 years, both in the USA and in Europe, and the largest contributor to direct HF-related health costs.1 Each hospitalization may be associated with organ damage, contribute to progression of HF, and represents one of the strongest negative predictors of outcome.2 In the USA almost one-quarter of Medicare beneficiaries hospitalized with HF are readmitted within 30 days; these readmissions are often considered to be a marker of poor health care and have become a benchmark for reimbursement, and an indicator of hospital quality.1,2 The emphasis on 30-day readmission as a HF performance measure has boosted the development of readmission prediction tools to identify high-risk patients and target preventive strategies. However, focus on early readmissions may be misleading. The decision to hospitalize can be driven by factors unrelated to the patient’s clinical status, as threshold for hospitalization and length of stay are highly variable across and within regions of the world and influence both in-hospital event rates and post-discharge outcomes. As models of health-care delivery evolve in response to economic pressures, care may shift to shorter stay in hospital or outpatient HF clinics. Furthermore, the competing risk of short-term post-discharge mortality is substantial and an inverse, although modest correlation, between 30-day mortality and 30-day readmission rates has been demonstrated.3 – 5 Furthermore, while hospitalized HF patient risk stratification tools for 30-day mortality based on administrative claims have fair predictive ability, which is improved by adding clinical data, most risk prediction models for readmissions show modest accuracy that remains unchanged when clinical variables are considered.6 As both mortality and rehospitalization require similar consideration of aggressive therapies and end-of-life care, the use of composite measures such as early mortality and/or readmission events,7 or total days alive out of hospital (DAOH), combining mortality and hospitalizations burden,8 has been advocated. By analysing the incidence and predictive factors of early mortality and/or readmissions in patients discharged home alive from cardiology units after an acute HF hospitalization, we aimed to derive information tailored to the cardiology setting as a guide to implement strategies directed at improving early post-discharge outcomes.
Methods IN-HF Outcome was a prospective, multicentre, observational, nationwide study, which involved 61 Italian cardiology centres including academic and community hospitals, well distributed over the whole country. The aims of the registry and eligibility criteria have been previously reported.9 Briefly, patients were enrolled at the time of hospital admission for acute decompensated HF, as diagnosed by attending physicians from rapid onset/change in typical signs and symptoms requiring urgent treatment.10 Both patients with reduced and preserved left ventricular (LV) ejection fraction (≥50%) were enrolled; the latter had to show evidence of pulmonary congestion on chest X-ray. Those younger than 18 years or unwilling to participate were excluded. Patients were asked to consent to the anonymous management of
........................................................................................................................................................................
The 30-day metric in acute heart failure
© 2015 The Authors European Journal of Heart Failure © 2015 European Society of Cardiology
their individual data. Local institutional review boards were informed of the study according to national rules. Data were collected using a web-based system and stored in a central database. No specific protocols or recommendations for HF management were used. Participating physicians were strongly invited to comply with current HF guidelines10 discussed during investigators’ meetings, before patients’ enrolment. Drug prescriptions and indications to perform diagnostic/therapeutic procedures were left to the participating cardiologists. Consecutiveness of enrolment was strictly recommended but was not checked by an admission log. The clinical status at 30 days post-discharge was ascertained by a telephone interview in patients not attending a clinical visit. Cause of death or hospital admission was adjudicated unblinded by hospital records, death certificates, and autopsy records, or by contacting the patients’ physician or referring cardiologist. The registry enrolled 1855 patients with acute HF. The present report analyses the 30-day mortality and/or rehospitalization of the 1520 patients discharged home. The proportion of DAOH within 30 days from discharge was derived by setting as 100% patients who were alive and had not been readmitted at 30 days and by the equation %DAOH = (DAOH/30) × 100 for those who met the end-point.
Statistical analysis Categorical variables are presented as number and percentages, while continuous variables are presented as their means and standard deviation (SD). Categorical variables were compared by the chi-square test and continuous variables by the t-test, if normally distributed, or by the Mann–Whitney U-test, if not. Backward logistic regression models were used to identify the independent predictors of all-cause mortality and/or hospital readmission and of hospital readmission within 30 days after discharge. For continuous variables, missing values ranging from 10–25% were replaced by the median value calculated on patients with data available. The models were adjusted by variables with
The 30-day metric in acute heart failure revisited: data from IN-HF Outcome, an Italian nationwide cardiology registry Giuseppe Di Tano1*, Renata De Maria2, Lucio Gonzini3, Nadia Aspromonte4, Andrea Di Lenarda5, Mauro Feola6, Marco Marini7, Massimo Milli8, Gianfranco Misuraca9, Andrea Mortara10, Fabrizio Oliva11, Giovanni Pulignano12, Giulia Russo5, Michele Senni13, and Luigi Tavazzi14, on the behalf of the IN-HF Outcome Investigators† 1 Cardiology
Department, Istituti Ospitalieri, Cremona, Italy; 2 CNR Clinical Physiology Institute, Cardiothoracic and Vascular Department, Azienda Ospedaliera Niguarda Ca’ Granda, Milan, Italy; 3 ANMCO Research Centre, Florence, Italy; 4 Cardiology Department, Ospedale San Filippo Neri, Rome, Italy; 5 Cardiovascular Center, Health Authority n. 1 and University of Trieste, Trieste, Italy; 6 Cardiovascular Rehabilitation, Heart Failure Unit, Ospedale Maggiore SS, Trinità, Fossano, Italy; 7 Cardiology Department, Ospedali Riuniti, Umberto I–Lancisi–Salesi, Ancona, Italy; 8 Cardiology, Ospedale Santa Maria Nuova, Florence, Italy; 9 Cardiology Department, Ospedale dell’Annunziata, Cosenza, Italy; 10 Cardiology Department, Policlinico di Monza, Monza, Italy; 11 Cardiologia 2 Heart Failure and Heart Transplant Programme, ‘A. De Gasperis’ Cardiothoracic and Vascular Department, Azienda Ospedaliera Niguarda Ca’ Granda, Milan, Italy; 12 Heart Failure Clinic, 1st Cardiology/CCU Unit, Cardiovascular Department, San Camillo Hospital, Rome, Italy; 13 Cardiovascular Department, Cardiology 1, Papa Giovanni XXIII Hospital, Bergamo; and 14 Maria Cecilia Hospital–GVM Care&Research–E.S. Health Science Foundation, Cotignola, Italy Received 29 December 2014; revised 17 April 2015; accepted 27 April 2015 ; online publish-ahead-of-print 27 May 2015
Aims
Unplanned readmissions early after a discharge from acute heart failure hospitalization are common and have become a reimbursement benchmark and marker of hospital quality. However, the competing risk of short-term post-discharge mortality is substantial. ..................................................................................................................................................................... Methods Using data from the prospective, nationwide Registry IN-HF Outcome, we analysed the incidence and predictors and results of 30-day mortality or readmissions and associated days-alive-out-of-hospital (DAOH) in 1520 patients discharged alive after admission for acute heart failure. Within 30 days after discharge, 94 patients (6.2%) were readmitted (91% for cardiovascular causes; 60% recurrent heart failure) and 42 (2.8%) died, 10 of which occurred during readmission. Overall, 126 patients (8.3%) met the combined endpoint. By multivariable logistic regression, worsening chronic heart failure as clinical presentation [odds ratio (OR) 1.83, 95% confidence interval (CI) 1.21–2.77, P = 0.005), inotropes during admission (OR 2.19, 95% CI 1.40–3.43, P = 0.0006), length of stay (OR 1.02, 95% CI 1.01–1.04, P = 0.002) and renin–angiotensin system inhibitors at discharge (OR 0.52, 95%CI 0.35–0.77, P = 0.001) independently predicted 30-day all-cause mortality and/or readmission (c-statistic = 0.695). Per cent 30-day DAOH was lower in patients with in-hospital inotrope use, no renin–angiotensin system inhibitors prescription at discharge, New York Heart Association III–IV class at discharge, and correlated inversely with length of stay and age. ..................................................................................................................................................................... Conclusion A clinical and biohumoral profile consistent with chronic advanced heart failure and end-organ damage identifies acute heart failure patients discharged home from cardiology units, who are at highest risk of early death and/or readmission. These findings have practical implications for tailoring specific follow-up.
.......................................................................................................... Keywords
Acute heart failure •
mortality •
30-Day readmission •
Days-alive-out-of-hospital
*Corresponding author: Tel: +39 0372 405323; Fax: +39 0372 433787; E-mail: [email protected], [email protected] † See Appendix S1 for a complete list of participating centres and investigators.
© 2015 The Authors European Journal of Heart Failure © 2015 European Society of Cardiology
1033
Introduction Hospital admission for heart failure (HF) is the most common cause of hospitalization in patients aged over 65 years, both in the USA and in Europe, and the largest contributor to direct HF-related health costs.1 Each hospitalization may be associated with organ damage, contribute to progression of HF, and represents one of the strongest negative predictors of outcome.2 In the USA almost one-quarter of Medicare beneficiaries hospitalized with HF are readmitted within 30 days; these readmissions are often considered to be a marker of poor health care and have become a benchmark for reimbursement, and an indicator of hospital quality.1,2 The emphasis on 30-day readmission as a HF performance measure has boosted the development of readmission prediction tools to identify high-risk patients and target preventive strategies. However, focus on early readmissions may be misleading. The decision to hospitalize can be driven by factors unrelated to the patient’s clinical status, as threshold for hospitalization and length of stay are highly variable across and within regions of the world and influence both in-hospital event rates and post-discharge outcomes. As models of health-care delivery evolve in response to economic pressures, care may shift to shorter stay in hospital or outpatient HF clinics. Furthermore, the competing risk of short-term post-discharge mortality is substantial and an inverse, although modest correlation, between 30-day mortality and 30-day readmission rates has been demonstrated.3 – 5 Furthermore, while hospitalized HF patient risk stratification tools for 30-day mortality based on administrative claims have fair predictive ability, which is improved by adding clinical data, most risk prediction models for readmissions show modest accuracy that remains unchanged when clinical variables are considered.6 As both mortality and rehospitalization require similar consideration of aggressive therapies and end-of-life care, the use of composite measures such as early mortality and/or readmission events,7 or total days alive out of hospital (DAOH), combining mortality and hospitalizations burden,8 has been advocated. By analysing the incidence and predictive factors of early mortality and/or readmissions in patients discharged home alive from cardiology units after an acute HF hospitalization, we aimed to derive information tailored to the cardiology setting as a guide to implement strategies directed at improving early post-discharge outcomes.
Methods IN-HF Outcome was a prospective, multicentre, observational, nationwide study, which involved 61 Italian cardiology centres including academic and community hospitals, well distributed over the whole country. The aims of the registry and eligibility criteria have been previously reported.9 Briefly, patients were enrolled at the time of hospital admission for acute decompensated HF, as diagnosed by attending physicians from rapid onset/change in typical signs and symptoms requiring urgent treatment.10 Both patients with reduced and preserved left ventricular (LV) ejection fraction (≥50%) were enrolled; the latter had to show evidence of pulmonary congestion on chest X-ray. Those younger than 18 years or unwilling to participate were excluded. Patients were asked to consent to the anonymous management of
........................................................................................................................................................................
The 30-day metric in acute heart failure
© 2015 The Authors European Journal of Heart Failure © 2015 European Society of Cardiology
their individual data. Local institutional review boards were informed of the study according to national rules. Data were collected using a web-based system and stored in a central database. No specific protocols or recommendations for HF management were used. Participating physicians were strongly invited to comply with current HF guidelines10 discussed during investigators’ meetings, before patients’ enrolment. Drug prescriptions and indications to perform diagnostic/therapeutic procedures were left to the participating cardiologists. Consecutiveness of enrolment was strictly recommended but was not checked by an admission log. The clinical status at 30 days post-discharge was ascertained by a telephone interview in patients not attending a clinical visit. Cause of death or hospital admission was adjudicated unblinded by hospital records, death certificates, and autopsy records, or by contacting the patients’ physician or referring cardiologist. The registry enrolled 1855 patients with acute HF. The present report analyses the 30-day mortality and/or rehospitalization of the 1520 patients discharged home. The proportion of DAOH within 30 days from discharge was derived by setting as 100% patients who were alive and had not been readmitted at 30 days and by the equation %DAOH = (DAOH/30) × 100 for those who met the end-point.
Statistical analysis Categorical variables are presented as number and percentages, while continuous variables are presented as their means and standard deviation (SD). Categorical variables were compared by the chi-square test and continuous variables by the t-test, if normally distributed, or by the Mann–Whitney U-test, if not. Backward logistic regression models were used to identify the independent predictors of all-cause mortality and/or hospital readmission and of hospital readmission within 30 days after discharge. For continuous variables, missing values ranging from 10–25% were replaced by the median value calculated on patients with data available. The models were adjusted by variables with
