JOURNAL OF PALLIATIVE MEDICINE Volume 17, Number XX, 2014 ª Mary Ann Liebert, Inc. DOI: 10.1089/jpm.2013.0658

Palliative Care Referral among Patients Hospitalized with Advanced Heart Failure Daniel T. Greener, MD,1 Timothy Quill, MD,2 Offer Amir, MD,3 Jill Szydlowski, BS,4 and Robert E. Gramling, MD 5

Abstract

Background: Many heart failure (HF) patients experience high symptom burden, but palliative care (PC) services have been used infrequently in this population. Objective: The specific aim of this study was to identify individual-level factors associated with PC referral. Methods: The study sample included adult patients hospitalized at an academic medical center with a primary diagnosis of HF between January 2005 and June 2010. Inpatient records were merged with the PC database to identify HF patients who received PC consultations. The analytical sample included 2647 HF admissions. We used descriptive statistics to characterize HF patients who received and did not receive PC services. Logistic regression analyses were used to identify patient characteristics that predict PC referral. Results: Just over 6% of patients with HF were referred to PC during their hospitalization. We identified the following statistically significant determinants of PC referral: secondary diagnosis of Alzheimer’s disease, receipt of thoracentesis, intensive care unit (ICU) stay, and prior HF-related hospitalizations. Conclusions: Currently, only a fraction of HF patients who are at high risk for morbidity and mortality receive PC services. Additional research is needed to identify factors associated with PC referral that can be prospectively identified, and to develop better prediction models to identify HF patients who may benefit from PC referral.

death occurring at 6 to 9 times the rate seen in the general population.5 To some extent, the decision to refer HF patients for PC consultation may be affected by the difficulty in establishing end-stage prognosis,2,3,6,7 and physicians’ inability to recognize the proximity of death.4,8,9 Today, little is known about why some hospitalized HF patients are referred to PC services whereas others are not. In this study we identified individual-level predictors of PC referral for HF inpatients.

Introduction

S

ince 2005, the guidelines of the American College of Cardiology (ACC) and the American Heart Association (AHA) have included recommendations for ongoing discussions with patients diagnosed with heart failure (HF) about prognosis for functional capacity, goals of care, advance directives, and symptom management.1 The guidelines also state that aggressive procedures such as intubation or implantation of a cardiac defibrillator in the last months of life may not be appropriate, as they do not contribute to patients’ recovery or quality of life. Although HF is the leading cause of death in the United States, fewer than 10% of HF patients receive palliative care (PC) services,2,3 and many receive major treatment interventions in the last 3 days of life.4 Although the ACC and AHA guidelines emphasize PC principles, they do not specifically address when to refer endstage HF patients for PC consultation. The one-year mortality rate for individuals with HF is 20%, with sudden cardiac

Methods Study setting

The study was conducted at an academic medical center providing care to approximately 40,000 inpatients per year. The inpatient PC consulting service was established at the hospital in 2001. Access to PC is by referral from the attending physician. Although such referrals are often based on decisions made by the patients, their families, and their

1 Department of Pathology, 2Department of Medicine, 4Department of Public Health Sciences, 5Department of Family Medicine, University of Rochester School of Medicine, Rochester, New York. 3 Department of Cardiology, Carmel Medical Center, Haifa, Israel. Accepted May 12, 2014.

1

2

physician, the literature suggests that patients and physicians are often reluctant to discuss goals of care and other complex issues that lead to referrals.10 Data sources

We used two sources of data. Hospitalization data came from the administrative database maintained by the Office of Clinical Practice Evaluation (OCPE). Selected variables were obtained for all admissions with the primary diagnosis of HF (per International Classification of Diseases, 9th Revision, Clinical Modification [ICD-9-CM] codes; see Appendix 1) that occurred within the defined time frame. The PC database is maintained separately by the PC program, and was employed to identify HF patients receiving PC consultations during their hospital stay. These two databases were merged using patients’ medical record numbers and their date of hospital admission. For purposes of this analysis, a de-identified data file was used. Institutional Research Subjects Review Board (RSRB) approval for the study was obtained. Sample population

The study sample included adult patients (age > 18 years) who were hospitalized with a primary diagnosis of HF between January 2005 and June 2010. The list of HF diagnostic codes was based on review of the literature and also included input from a consulting cardiologist. The usable analytical sample included 2647 admissions (2001 unique patients). PC records for the same time period totaled 4144. Variables

The outcome of interest was referral to PC, defined as at least one visit by a PC clinician during a hospital stay. The outcome was defined as a dichotomous variable with 1 indicating referral and 0 denoting none. The independent variables focused on patient characteristics and use of hospital services. They were selected based on literature review and their availability in the databases employed. Age ( < 65; 65–74; ‡ 75), marital status (married or not), gender (male or not), race/ethnicity (white or not), and health insurance (Medicare, Medicaid, other) were defined as categorical variables. The service to which each patient was admitted was characterized as cardiology or another service unit. We also constructed patient-specific clinical indicators that may be associated with PC referral. Based on the distribution of comorbid diagnoses and treatment procedures performed, we identified several diagnostic conditions and procedures (dichotomized as present = 1 or absent = 0). We also constructed a measure to count the number of HF hospital admissions in a prior year, and we calculated a 30-day readmission rate (number of HF readmissions occurring within 30 days of the index admission) in each year (both continuous variables). Prior studies have suggested that both the length of hospital stay and the length of stay in the intensive care unit (ICU) are associated with the rate of PC referral. Hospital length of stay was defined as a continuous variable. We defined ICU stay as a dichotomous variable with 1 indicating an ICU stay and 0 indicating none, because length of ICU stay was highly correlated with the overall length of hospital stay. We also included the year of

GREENER ET AL.

inpatient stay to control for any changes in practice that may have occurred during the study time period. Following discharge from the hospital, severity of illness (SOI) and risk of mortality (ROM) were calculated for each patient. Based on the model developed by 3M Health Information Systems, a patient may be classified into one of four subclasses of SOI and ROM, indicating respectively minor, moderate, major, or severe risk. The assignment of a patient into a risk subclass takes into account all secondary diagnoses, as well as the interaction among these diagnoses and age, principal diagnosis, and presence of certain operative and nonoperative procedures.11 Analytical approach

To describe the study population and identify independent variables for inclusion in the multivariate model, we relied on bivariate analyses. Chi-square tests were used to examine the associations between crude (unadjusted) categorical variables and the outcome of interest (referral or lack of referral to PC). For continuous variables, student’s t test was used. All statistical tests were two-sided, as we had no specific hypotheses about the direction of these associations. P values < 0.05 were considered significant. Because independent variables that are highly correlated should not be used in the same multivariable model, we examined possible collinearity between the predictor variables using Pearson’s correlation coefficients. The probability of PC referral was assessed using patientlevel multivariable logistic regression. Because patients had many diagnoses and procedures, the validity of the model fit when all diagnoses are used may be questionable. To create a more parsimonious model, we excluded variables that were statistically insignificant at 0.2 or higher. All excluded covariates were subject to a joint F-test to verify that they were not jointly statistically significant. The reduced model was then reestimated. We also estimated two alternative models. One model included SOI and the year of hospitalization, whereas the other included ROM and the year of hospitalization. C statistics were calculated to assess the predictive power of all models. All analyses were conducted using the Statistical Analysis System (SAS) software version 9.1 (SAS Institute Inc., Cary, NC). Results Characteristics of heart failure patients

Overall, 6.2% of HF patients had a PC referral during their hospital stay (Table 1). Patients who were referred to PC were, on average, significantly ( p = 0.0074) older (mean = 70.0 years) than patients who were not referred (mean = 66.8 years), and were more likely to be married (57.5% versus 49.4%; p = 0.04). HF patients with a referral had significantly longer stays (19.53 days versus 9.67 days; p < 0.0001), higher risk for mortality (score of 3.31 versus 2.56; p < 0.0001), higher severity of illness (score of 3.30 versus 2.85; p < 0.0001), and more days in the ICU (4.96 days versus 2.01 days; p = 0.03). They were also significantly ( p = 0.0004) more likely to have had prior-year HF admissions, and were more likely to have experienced a readmission within 30 days ( p < 0.0001) compared with

PALLIATIVE CARE IN PATIENTS HOSPITALIZED WITH HEART FAILURE

3

Table 1. Characteristics of HF Patients by PC Referral Status: CY2005–2010 No palliative care (n = 2480) Patient characteristics Age at admission Gender (male) Minority (nonwhite) Majority (white) Married Insurance Medicare Medicaid Other Percent admitted to cardiology designated floor LOS (days) Intensive care unit LOS (days) Admission/readmissions No. of admissions in last year £ 30 days Risk of mortality score (ROM) Severity of illness score (SOI) Discharge disposition Deceased Hospice Home care with services Home self-care Other Treatment/Procedure Cardiac catheterization Dialysis Bilevel positive airway pressure (BPAP) Thoracentesis Ventilation Heart transplant Packed cell transfusion Amputation Ventricular assist device (VAD) Diagnosis Chronic renal failure COPD Lower respiratory disease Diabetes Acute renal failure Anxiety Depression Atherosclerosis Cerebrovascular disease Alzheimer’s disease

Palliative care (n = 167)

Mean

SD

Mean

SD

66.80 62.66% 23.06% 76.71% 49.44%

15.64

70.03 62.87% 18.56% 81.44% 57.49%

14.90

28.27% 3.71% 68.02% 70.86% 9.67 2.01 0.18 6.57% 2.56 2.85

18.03 10.17 0.38 0.85 0.76

29.34% 2.99% 67.66% 72.46% 19.53 4.96 0.32 14.97% 3.31 3.30

P value 0.0074 0.9561 0.1794 0.0440 0.8670

27.73 17.46 0.47 0.72 0.71

0.4674 < 0.0001 0.0322 0.0004 < 0.0001 < 0.0001 < 0.0001

5.14% 0.24% 41.56% 42.48% 19.58%

43.71% 8.38% 20.36% 5.39% 22.16%

< 0.0001 < 0.0001 < 0.0001 < 0.0001 < 0.0001

16.49% 3.27% 5.89% 1.98% 1.37% 1.57% 1.53% 0.12% 1.94%

19.16% 2.99% 5.99% 7.19% 2.40% 1.20% 2.99% 0.00% 0.60%

0.3701 0.8478 0.9572 < 0.0001 0.2815 0.7041 0.1481 0.6529 0.2148

33.39% 16.85% 8.02% 46.13% 15.12% 9.19% 10.16% 6.77% 1.61% 0.69%

52.10% 18.56% 7.19% 46.71% 35.33% 12.57% 6.59% 9.58% 1.80% 2.40%

< 0.0001 0.5692 0.6986 0.8848 < 0.0001 0.1474 0.1350 0.1675 0.8559 0.0159

COPD, chronic obstructive pulmonary disease; CY, calendar year; HF, heart failure; LOS, length of stay; PC, palliative care; SD, standard deviation.

patients without PC referral. We found no statistically significant differences between these two groups with regard to gender, health insurance, race/ethnicity, or the admitting service. The two groups were significantly different with respect to several clinical characteristics. HF patients with PC referral were statistically significantly more likely to have chronic (52.1% versus 33.4%) and acute renal failure (35.3% versus 15.1%) and Alzheimer’s disease (2.40% versus 0.69%), to be deceased at discharge (43.71% versus 5.14%) or to be

discharged to hospice care (8.38% versus 0.24%), and to undergo thoracentesis (7.19% versus 1.98%). Determinants of palliative care referral

Using a multivariable logistic regression model, we examined the odds of PC referral, while simultaneously controlling for sociodemographics, comorbid conditions, high-risk procedures, and prior hospitalizations (Table 2). Older (age ‡ 75 years) and married patients had significantly higher odds of PC referral

4

GREENER ET AL.

Table 2. The Likelihood of Patients Hospitalized with HF Being Referred for PC

Independent variables

Odds ratio

Year of PC referral (ref = 2005) 2006 0.748 2007 0.772 2008 1.277 2009 0.977 2010a 1.001 Age (ref = £ 65) 65–74 1.122 75 and older 1.916 Gender (ref = female) Male 0.905 Race (ref = nonwhite) White 1.095 Marital status (ref = single) Married 1.421 Primary insurance (ref = Medicare) Medicaid 1.228 Other 1.334 Comorbid conditions Chronic renal failure 1.419 Acute renal failure 2.386 Anxiety 1.387 Depression 0.652 Alzheimer’s 4.531 High-risk procedures Cardiac catheterization/ 1.263 Angioplasty/CABG Bilevel positive airway 1.006 pressure (BPAP) Thoracentesis/Paracentesis 4.125 Ventilation 1.231 Packed cell transfusion 2.106 Ventricular assist device 0.124 (VAD) Hospitalization Number of HF admissions 1.460 in last year ICU stay 2.485 C statistic

95% confidence interval 0.398 0.411 0.700 0.530 0.497

1.408 1.448 2.329 1.799 2.016

0.706 1.256

1.781 2.924

0.634

1.291

0.697

1.720

1.001

2.018

0.448 0.885

3.367 2.013

0.916 1.650 0.831 0.340 1.402

2.198 3.451 2.316 1.248 14.649

0.812

1.964

0.498

2.033

2.023 0.396 0.772 0.016

8.411 3.824 5.749 0.987

1.232

1.731

1.654 0.753

3.733

a partial year. CABG, coronary artery bypass graft; HF, heart failure; ICU, intensive care unit; PC, palliative care.

(odds ratio [OR] = 1.916 and OR = 1.421, respectively). Patients with acute renal failure and those with an Alzheimer’s disease diagnosis also had statistically significantly higher odds of PC referral (OR = 2.386 and OR = 4.531, respectively) than patients without these comorbid conditions. Other significant risk factors for PC referral included thoracentesis (OR = 4.125; 95% confidence interval [CI] = 2.023-8.411), prior HF admissions in the last year (OR = 1.460; CI = 1.232-1.731), and ICU stay (OR = 2.485; CI = 1.654-3.733). The statistical fit of this multivariable model was good (C statistic = 0.753). We also estimated two alternative models, each including only the year of hospitalization and risk level for SOI and ROM, respectively. Both SOI and ROM are calculated after a

hospital event is concluded, and they are not available prospectively. Therefore, they cannot currently be used as predictors of PC referral. However, we wanted to examine how well these measures alone discriminate between HF patients who are and are not referred to PC. To assure that these findings were not confounded by possible time-related trends in PC referral or changes in SOI and ROM over time, we included the year of hospitalization as a covariate. The measures of ROM and SOI were not available prior to calendar year 2007, so the analyses were conducted on data available since 2007. The statistical fit of the SOI model was modest with a C statistic of 0.669, and it was a good for the ROM with a C statistic of 0.741 (Table 3). The model for SOI discriminated well between patients with and without PC referral only when the SOI was at the highest possible risk level, compared with patients whose SOI was minor (OR = 8.532; CI = 1.153-63.115). In the ROM model, patients at major (OR = 19.376) and severe (OR = 47.627) risk for mortality were significantly more likely to have been referred to PC, compared with patients whose risk for mortality was determined to have been minor. Discussion

HF patients experience high levels of life-limiting illness and report a significant symptom burden.12–14 Although prior studies have called for, and practice guidelines have recommended PC for HF patients,15 the literature documenting the provision of PC to this population has been sparse and empirical evidence has been lacking. We found a relatively low PC referral rate (6%) among HF inpatients, and no substantial changes in the referral pattern over the study time period. This rate of PC referral is consistent with prior studies of the HF population,15,16 as well as of other noncancer populations.17 We found that several patient factors may be predictive of PC referral. When all other characteristics are controlled for, older patients > 90% more likely to be referred compared with patients younger than 65. These older patients probably had more prior acute disease exacerbations and their illness severity or functional status, for which we were unable to adequately adjust, may have been key factors in PC referral. Marital status was another important PC predictor, potentially suggesting that spouses may be important advocates in this process or that they themselves may have been in need of the support that PC provides. Adjusting for other covariates, patients with secondary diagnoses of renal failure (particularly the acute condition) and patients with Alzheimer’s disease had a significantly higher likelihood of being referred to PC. Similarly, patients receiving thoracentesis (probably a marker for refractory fluid accumulation) were much more likely to receive referral to PC. It may also be possible that these patients were actively dying, and that thoracentesis was used as a last resort to relieve symptoms. All else being equal, patients with HF admissions in a prior year, and those with an ICU stay, had a significantly higher likelihood of PC referral. Each additional prior year’s HF hospitalization resulted in an approximately 46% increase in the odds of PC referral. The alternative models, which included only the SOI and ROM measures, suggested that comorbid conditions and their severity are important in identifying candidates for PC referral. Unfortunately, these measures are currently only

PALLIATIVE CARE IN PATIENTS HOSPITALIZED WITH HEART FAILURE

5

Table 3. The Likelihood of Patients Hospitalized with HF Being Referred for PC: Alternative Models Severity of illness model (n = 1718) Independent variables

Odds ratio

95% confidence interval

Year of PC referral (ref = 2007) 2008 1.634 0.962 2009 1.294 0.760 1.332 0.715 2010a Severity of illness risk level (ref = minor risk) Moderate risk 1.808 0.234 Major risk 3.314 0.449 Severe risk 8.532 1.153 Risk of Mortality level (ref = minor risk) Moderate risk Major risk Severe risk C statistic 0.669

2.777 2.205 2.483

Risk of mortality model (n = 1718) Odds ratio

95% confidence interval

1.617 1.228 1.145

0.951 0.721 0.615

2.750 2.090 2.134

6.183 19.376 47.627 0.741

0.373 1.194 2.938

102.388 314.492 771.963

13.989 24.468 63.115

a partial year. HF, heart failure; PC, palliative care.

generated postdischarge and thus cannot be employed to prospectively identify patients who may benefit from PC referral. These findings suggest that should such measures become available prospectively, a substantially higher proportion of HF patients might receive PC consultations. For example, more than 20% of HF patients were retrospectively characterized as having severe SOI and 50% to 60% as having major-to-severe ROM, both statistically significant risk factors for PC referral. Yet, only 6.2% of HF patients actually received PC referral.

hospitalized with heart failure. Our findings suggest that only a fraction of HF patients who are at high risk for morbidity and at a significant risk for mortality receive PC. Integrating the management of HF with PC, as early as at diagnosis, may be important in assuring that PC is viewed as a complementary service designed to minimize distressing symptoms, and to assist patients and their families in making the difficult but necessary decision about future treatments. Additional research is also needed to develop and validate a clinically useful prediction tool to prospectively identify HF patients who may most benefit from PC referral.

Limitations

Several limitations should be noted. First, the analysis was based on retrospective rather than prospective data, thus it is based on patients for whom the decision to refer or not to refer to PC has already been made. Second, the data for this study came from only one medical center and this may cause two biases: 1) Practice patterns in medical centers vary depending on local culture and the relative institutional strength of palliative care, and 2) PC presence and utilization in this care setting are also likely to be different than in smaller institutions or community hospitals. The setting used for this report has an established PC consultation service, which may represent the best in the current PC practice, and our findings may not be generalizable to other settings. Third, only a subset of patient-specific variables was available for this analysis. This limited our ability to risk-adjust for patient case-mix or for illness severity (e.g., ejection fraction). Prior studies have shown that physician characteristics are also important predictors of PC referral.18 However, the available administrative data did not have any information on physician characteristics. The reliance on an administrative database also precluded us from identifying whether certain procedures (e.g., thoracentesis) preceded or followed referral to PC. Conclusions

This study adds important empirical information to the current body of knowledge about PC use among patients

Author Disclosure Statement

No conflicting financial interests exist. References

1. Hunt S, Abraham W, Chin M, Feldman A, Francis G, et al.: ACC/AHA 2005 Guideline Update for the Diagnosis and Management of Chronic Heart Failure in the Adult: A report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines (Writing Committee to Update the 2001 Guidelines for the Evaluation and Management of Heart Failure): Developed in collaboration with the American College of Chest Physicians and the International Society for Heart and Lung Transplantation: Endorsed by the Heart Rhythm Society. Circulation 2005;112 e154–e235. 2. Rogers AE, Addington-Hall JM, Abery AJ, et al.: Knowledge and communication difficulties for patients with chronic heart failure: Qualitative study. BMJ 2000;321:605–607. 3. Lee DS, Austin PC, Rouleau JL, Liu PP, Naimark D, Tu JV: Predicting mortality among patients hospitalized for heart failure. JAMA 2003;290:2581–2587. 4. Fox E, Landrum-McNiff K, Zhong Z, Dawson NV, Wu AW, Lynn J: Evaluation of prognostic criteria for determining hospice eligibility in patients with advanced lung, heart, or liver disease. SUPPORT Investigators. Study to Understand Prognoses and Preferences for Outcomes and Risks of Treatments. JAMA 1999;282:1638–-1645.

6

GREENER ET AL.

5. Lloyd-Jones D, Adams R, et al.: Heart Disease and Stroke Statistics—2009 Update: A report from the American Heart Association Statistics Committee and Stroke Statistics Subcommittee. Circulation 2009;119:e21–e181. 6. Haga K, Murray S, Reid J, et al.: Identifying community based chronic heart failure patients in the last year of life: A comparison of the Gold Standards Framework Prognostic Indicator Guide and the Seattle Heart Failure Model. Heart 2012;98:579–583. 7. Allen LA, Stevenson LW, Grady KL, et al.: Decision making in advanced heart failure: A scientific statement from the American Heart Association. Circulation 2012;125:1928–1952. 8. Kheirbek RE, Alemi F, Citron BA, Afaq MA, Wu H, Fletcher RD: Trajectory of illness for patients with congestive heart failure. J Palliat Med 2013;16:478–484. 9. Bakitas M, Macmartin M, Trzepkowski K, et al.: Palliative care consultations for heart failure patients: How many, when, and why? J Cardiac Failure 2013;19:193–201. 10. Quill TE: Initiating end-of-life discussions with seriously ill patients. JAMA 2000;284:2502–2507. 11. Hughes J: 3M Health Information Systems (HIS) APRDRG Classification Software-Overview. In Mortality Measurement February 2009. Agency for Healthcare Research and Quality, Rockville, MD. http:www.ahrq.gov/ professionals/quality-patient-safety/quality-resources/tools/ mortality/Hughes.html. The History of Medical Coding. Last accessed July 18, 2014. 12. Solano JP, Gomes B, Higginson IJ: A comparison of symptom prevalence in far advanced cancer, AIDS, heart disease, chronic obstructive pulmonary disease and renal disease. J Pain Symptom Manage 2006;31:58–69.

13. Janssen DJA, Spruit MA, Wouters EFM, Schols J: Daily symptom burden in end-stage chronic organ failure: A systematic review. Palliat Med 2008;22:938–948. 14. Rutledge T, Reis VA, Linke SE, Greenberg BH, Mills PJ: Depression in heart failure: A meta-analytic review of prevalence, intervention effects, and associations with clinical outcomes. J Amer Coll Cardiol 2006;48:1527–1537. 15. Jaarsma T, Beattie JM, Ryder M, et al.: Palliative care in heart failure: A position statement from the palliative care workshop of the Heart Failure Association of the European Society of Cardiology. Eur J Heart Failure 2009;11:433–443. 16. Johnson MJ, Houghton T: Palliative care for patients with heart failure: Description of a service. Palliat Med 2006;20: 211–214. 17. Holloway RG, Ladwig S, Robb J, Kelly A, Nielsen E, Quill TE: Palliative care consultations in hospitalized stroke patients. J Palliat Med 2010;13:407–412. 18. Low J, Pattenden J, Candy B, Beattie JM, Jones L: Palliative care in advanced heart failure: An international review of the perspectives of recipients and health professionals on care provision. J Cardiac Failure 2011;17:231–252.

Address correspondence to: Daniel T. Greener, MD Department of Pathology University of Rochester School of Medicine 601 Elmwood Avenue Rochester, NY 14642 E-mail: [email protected]

Appendix 1. Congestive Heart Failure: ICD-9-CM Diagnostic Codes Code

Descriptor

398.91 Rheumatic heart disease (congestive) 402 Hypertensive heart disease 402.01 Malignant with heart failure 402.11 Benign with heart failure 402.91 Unspecified with heart failure 404 Hypertensive heart and chronic kidney disease 404.01 Malignant, with heart failure and with chronic kidney disease stage I through IV or unspecified 404.03 Malignant, with heart failure and chronic kidney disease stage V or end-stage renal disease 404.11 Benign, with heart failure and with chronic kidney disease stage I through IV or unspecified 404.13 Benign, with heart failure and chronic kidney disease stage V or end-stage renal disease 404.91 Unspecified, with heart failure and with chronic kidney disease stage I through IV or unspecified 404.93 Unspecified with heart failure and chronic kidney disease stage V or end-stage renal disease 428 Heart failure 428.0 Congestive heart failure, unspecified 428.1 Left heart failure 428.20 Systolic heart failure, unspecified 428.21 Systolic heart failure, acute 428.22 Systolic heart failure, chronic 428.23 Systolic heart failure, acute on chronic 428.30 Diastolic heart failure, unspecified