Abstract
Aims
Non-cardiac comorbidities are highly prevalent in patients with heart failure (HF). Our objective was to define the association between non-cardiac comorbidity burden and clinical outcomes, costs of care, and length of stay within a large randomized trial of acute HF patients.
Methods and results
Patients with complete medical history for the following comorbidities were included: diabetes mellitus, chronic obstructive pulmonary disease, chronic liver disease, history of cancer within the last 5 years, chronic renal disease (baseline serum creatinine >3.0 mg/mL), current smoking, alcohol abuse, depression, anaemia, peripheral arterial disease, and cerebrovascular disease. Patients were classified by overall burden of non-cardiac comorbidities (0, 1, 2, 3, and 4+). Hierarchical generalized linear models were used to assess associations between comorbidity burden and 30-day all-cause death or HF hospitalization and 180-day all-cause death in addition to costs of care and length of stay. A total of 6945 patients were included in the final analysis. Mean comorbidity number was 2.2 (± 1.34). Patients with 4+ comorbidities had higher rates of 30-day all-cause death/HF hospitalization as compared with patients with no comorbidities [odds ratio (OR) 3.32, 95% confidence interval (CI) 1.61–6.84; P < 0.01]. Similar results were seen with respect to 180-day death (OR 2.13, 95% CI 1.33–3.43; P < 0.01). Higher comorbidity burden was associated with higher 180-day costs of care and length of stay.
Conclusions
Higher comorbidity burden is associated with poor clinical outcomes, higher costs of care, and extended length of stay. Further studies are needed to define the impact of comorbidity management programmes on outcomes for HF patients.
Keywords: Heart failure, Comorbidities, Length of stay
Introduction
Heart failure (HF) is a chronic disorder characterized by frequent acute exacerbations often requiring hospitalization. The condition is highly prevalent and increases in prevalence with age.1,2 As the population with HF has aged,3 comorbidity burden has also increased and patients with HF often have non-cardiac comorbidities that require assessment, management and treatment.
The association between non-cardiac comorbidities and preventable hospitalizations in a chronic HF with reduced ejection fraction (HFrEF) population has been previously established.4,5 Comorbidities such as anaemia, chronic kidney disease, and diabetes have also independently been associated with increased mortality in HFrEF patients.6–9 Despite this, there are relatively limited data to describe the impact of the overall burden of non-cardiac comorbidities on clinical outcomes and hospitalizations across ejection fraction. Furthermore the impact of the burden of non-cardiac comorbidities on costs of care and length of stay (LOS) is poorly understood.
We utilized the Acute Study of Clinical Effectiveness of Nesiritide in Decompensated Heart Failure (ASCEND-HF) randomized clinical trial to examine the impact of non-cardiac comorbidities on these outcomes in patients admitted with acute HF across ejection fractions.
Methods
Study overview
The design and results of the ASCEND-HF trial have been previously reported.10,11 Briefly, ASCEND-HF was a global, randomized, double-blind, placebo-controlled trial evaluating the efficacy of nesiritide vs. placebo, both in addition to standard care, in 7141 patients with acute HF. Detailed inclusion and exclusion criteria have been described previously.11 Patients were included irrespective of ejection fraction. Enrolled patients were required to have dyspnoea at rest or with minimal activity, ≥1 accompanying sign, and ≥1 objective measure of HF. The primary endpoint of the trial was a composite of all-cause mortality or HF readmission at 30 days post-randomization.
Study definitions and endpoints
The present analysis included patients enrolled in the ASCEND-HF clinical trial intention-to-treat population with complete medical history for the following 11 non-cardiac comorbidities: diabetes mellitus, chronic obstructive pulmonary disease (COPD), chronic liver disease, history of cancer within the last 5 years, chronic renal disease (baseline serum creatinine >3.0 mg/mL), current smoker, alcohol abuse, depression, anaemia, peripheral arterial disease, and cerebrovascular disease. Anaemia was defined as haemoglobin <13.5 g/L. All other non-cardiac comorbidities were assessed by the medical record or patient-provided data. Comorbidity selection was based on available data collected during the trial period and consensus among authors. Patients with missing medical history data with respect to the above non-cardiac comorbidities were excluded from our study population. Patients who died in-hospital or were still in hospital at 30 days post-randomization were excluded from all endpoint analyses.
The pre-specified primary clinical endpoints were the composite post-discharge 30-day all-cause death or HF hospitalization and 180-day all-cause death. Secondary clinical endpoints included 30-day all-cause death, 30-day cardiovascular (CV) death and hospitalization at 30 days for the following reasons: all-cause, CV, and HF. An independent blinded clinical events committee adjudicated all-cause of death and hospitalization within 30 days (University of Glasgow, Glasgow, Scotland, UK).
We also examined the following non-adjudicated measures: total cumulative hospital costs from randomization to day 180 and LOS. Cumulative hospitalization costs included inpatient hospital costs, nesiritide treatment costs, emergency room visits within first 30 days, as well as costs associated with all rehospitalizations between day 30 and day 180. LOS was defined as number of days between date of randomization into study and date of discharge from initial hospitalization or death if in-hospital.
Statistical methods
The prevalence of each non-cardiac comorbidity in our population was reported. The sum of non-cardiac comorbidities was calculated for each patient. As few patients had 5 or more comorbidities, we categorized the sum of non-cardiac comorbidities into five groups: none (zero), one, two, three, and four or more. The prevalence of comorbidities was described in patients with HFrEF, defined as left ventricular ejection fraction (LVEF) < 40% as compared with patients with HF with preserved ejection fraction (HFpEF), defined as LVEF ≥40%. Patient and hospital characteristics were summarized by the sum of non-cardiac comorbidities. Categorical variables were presented as frequencies and percentages, while continuous variables were presented as medians and interquartile range (IQR). Comparisons across categories were conducted by ANOVA or Kruskal-Wallis tests for continuous variables, and chi-square or Fisher’s exact tests or categorical variable, as appropriate.
The association between the sum of non-cardiac comorbidities and the primary endpoints was assessed using hierarchical multivariate regression modelling. A hierarchical logistic model with random intercept for country was used to assess the association between the sum of non-cardiac comorbidities and all 30-day outcomes including rehospitalizations and death. A frailty Cox proportional hazards model with random effect for country was used for the co-primary outcome of 180-day all-cause death. Models were adjusted for covariates previously identified as being associated with clinical outcomes.12 The model for the outcome of 30-day all-cause death or hospitalization for HF was adjusted for age, baseline blood urea nitrogen (log-transformed), creatinine (log-transformed), sodium, systolic blood pressure, dyspnoea, elevated jugular venous pressure, and hospitalization in prior year. The outcomes of 180-day death was adjusted for all-cause rehospitalization at 30 days, age, baseline sodium, temperature, weight, heart rate, blood urea nitrogen (log-transformed), hospitalization in prior year, day 30 follow-up diastolic blood pressure, visual analogue scale score and worsening HF.
Similarly, multivariate logistic regression was used to assess the association between the sum of non-cardiac comorbidities and the secondary outcomes of 30-day all-cause death, 30-day CV death, 30-day all-cause rehospitalization, 30-day CV rehospitalization, and 30-day HF rehospitalization. The hierarchical logistic regression models for 30-day HF rehospitalization, 30-day CV rehospitalization and 30-day CV death were adjusted for the same covariates as those used in the 30-day all-cause death and HF rehospitalization model. The outcome of 30-day all-cause rehospitalization was adjusted for age, baseline blood urea nitrogen (log-transformed), creatinine (log-transformed), sodium, systolic blood pressure, weight, dyspnoea, elevated jugular venous pressure and hospitalization in prior year. Zero non-cardiac comorbidities served as the reference group for all analyses.
We assessed the association between the sum of non-cardiac morbidities and 180-day cumulative hospitalization costs and index hospital LOS. Derivation of costs data for the ASCEND-HF trial have previously been described.13 Cost estimation used detailed billing data collected from 1493 hospitalization and emergency department visits for patients enrolled in the United States. For derivation of inpatient costs for US patients without data on hospital bills or for those enrolled outside of the United States, a cost-prediction model was used as previously described.
We assessed the association between the sum of non-cardiac morbidities and index hospital LOS via a hierarchical generalized linear model. A ‘LOS ratio’ was used, defined as mean LOS in United States divided by mean LOS in another country in order to control for international variations in LOS.
All statistical tests were two-sided. P-values < 0.05 were considered significant. Zero non-cardiac comorbidities also served as the reference group for these analyses. Analyses were conducted using SAS version 9.4 (SAS Institute, Cary, NC, USA). The Duke Clinical Research Institute (DCRI) performed database management and statistical analysis. Scios Inc. (Mountain View, CA, USA) provided financial and material support for the ASCEND-HF trial.
Results
Cohort analysis
A total of 7007 patients had complete data on all 11 assessed non-cardiac comorbidities. After excluding 62 patients who died in-hospital or were still in hospital at 30 days post-randomization, a total of 6945 patients (99.1%) were included in our final analyses. Baseline anaemia was the most prevalent non-cardiac comorbidity within the study population (60.3%), followed by current smoking (49.1%) and diabetes mellitus (42.5%). Chronic liver disease was least prevalent (2.7%) (Table 1). The mean number of non-cardiac comorbidities was 2.2 (± 1.37). Patients with HFpEF had higher average number of non-cardiac comorbidities (mean number of comorbidities 2.6 ± 1.50) compared with patients with HFrEF (mean number of comorbidities 2.1 ± 1.34) (P < 0.01). Patients with zero non-cardiac comorbidities comprised 8.9% (n = 619) of the overall study population, followed by 25.3% (n = 1759) with one comorbidity, 30.0% (n = 2084) with two comorbidities, 20.1% (n = 1399) with three comorbidities, and 15.6% (n = 1084) with four or more comorbidities (Figure 1). A proportion of 6.2% of patients had five or more comorbidities.
Table 1.
Non-cardiac comorbidities by left ventricular ejection fraction
| Baseline LVEF | |||||
|---|---|---|---|---|---|
| Non-cardiac comorbidities | All patients (n = 6945) | Preserved LVEF (n = 4189) | Reduced LVEF (n = 1044) | Missing LVEF (n = 1712) | P-value |
| History of diabetes mellitus | 2952 (42.5%) | 1731 (41.3%) | 510 (48.9%) | 711 (41.5%) | <0.01 |
| History of COPD | 1140 (16.4%) | 588 (14.0%) | 259 (24.8%) | 293 (17.1 %) | <0.01 |
| Chronic liver disease | 187 (2.7%) | 108 (2.6%) | 36 (3.4%) | 43 (2.5%) | 0.26 |
| History of cancer (within last 5 years) | 262 (3.8%) | 121 (2.9%) | 72 (6.9%) | 69 (4.0%) | <0.01 |
| Baseline serum creatinine >3.0mg/mL | 204 (2.9%) | 118 (2.8%) | 31 (3.0%) | 55 (3.2%) | 0.715 |
| Current smoker | 3410 (49.1%) | 2043 (48.8%) | 520 (49.8%) | 847 (49.5%) | 0.78 |
| History of alcohol abuse | 628 (9.0%) | 444 (10.6%) | 60 (5.7%) | 124 (7.2%) | <0.01 |
| History of depression | 541 (7.8%) | 285 (6.8%) | 135 (12.9%) | 121 (7.1 %) | <0.01 |
| Anaemia (haemoglobin <13.5 g/L) | 4189 (60.3%) | 2452 (58.5%) | 760 (72.8%) | 977 (57.1 %) | <0.01 |
| Peripheral arterial disease | 716 (10.3%) | 369 (8.8%) | 173 (16.6%) | 174 (10.2%) | <0.01 |
| Cerebrovascular disease | 815 (11.7%) | 435 (10.4%) | 160 (15.3%) | 220 (12.9%) | <0.01 |
COPD, chronic obstructive pulmonary disease; LVEF, left ventricular ejection fraction.
Figure 1.
Distribution of non-cardiac comorbidities across left ventricular ejection fraction (EF). This figure shows the distribution of the sum of non-cardiac comorbidities among patients enrolled in the ASCEND-HF randomized clinical trial, categorized as reduced EF (<40%) or preserved EF (≥ 40%).
Baseline clinical characteristics by burden of non-cardiac comorbidities
Patients with higher comorbidity burden were significantly older and more likely to be male as compared with those with lower comorbidity burden. Patients with high comorbidity burden were also more likely to be white or African American and less likely to be of Asian ethnicity, and were much more likely to be enrolled from North America as compared with those with lower comorbidity burden. Patients with higher comorbidity burden also had higher average body mass index and higher baseline N-terminal pro-B-type natriuretic peptide levels. Higher comorbidity patients were much more likely to have had a HF hospitalization within the last year and tended to have higher mean LVEF as compared with those with fewer non-cardiac comorbidities (Table 2).
Table 2.
Baseline patient characteristics by number of non-cardiovascular comorbidities
| Number of non-CV comorbidities | |||||||
|---|---|---|---|---|---|---|---|
| Characteristic | All patients (n = 6945) |
None (n = 619) |
One (n = 1759) |
Two (n = 2084) |
Three (n = 1399) |
Four plus (n = 1084) |
P-value |
| Demographics | |||||||
| Age, years | 65 ± 14.1 | 63 ±15.9 | 64 ±15.7 | 66 ±13.7 | 66 ±13.3 | 68 ± 11.6 | <0.001 |
| Female gender | 2373 (34.2%) | 231 (37.3%) | 694 (39.5%) | 735 (35.3%) | 403 (28.8%) | 310 (28.6%) | <0.001 |
| White | 3882 (55.9%) | 330 (53.3%) | 895 (50.9%) | 1100 (52.8%) | 820 (58.7%) | 737 (68.0%) | |
| Region | <0.001 | ||||||
| Asia-Pacific | 1723 (24.8%) | 201 (32.5%) | 600 (34.1%) | 587 (28.2%) | 258 (18.4%) | 77 (7.1%) | |
| Central Europe | 964 (13.9%) | 154 (24.9%) | 340 (19.3%) | 287 (13.8%) | 135 (9.6%) | 48 (4.4%) | |
| Latin America | 653 (9.4%) | 96 (15.5%) | 197 (11.2%) | 200 (9.6%) | 116 (8.3%) | 44 (4.1%) | |
| North America | 3124 (45.0%) | 130 (21.0%) | 528 (30.0%) | 867 (41.6%) | 774 (55.3%) | 825 (76.1%) | |
| Western Europe | 481 (6.9%) | 38 (6.1%) | 94 (5.3%) | 143 (6.9%) | 116 (8.3%) | 90 (8.3%) | |
| Medical history | |||||||
| HF hospitalization in past year | 2689 (38.8%) | 188 (30.4%) | 568 (32.3%) | 743 (35.7%) | 630 (45.1%) | 560 (51.7%) | <0.001 |
| LVEF <40% in past year | 1044 (20.0%) | 55 (12.2%) | 196 (14.6%) | 290 (18.5%) | 243 (23.5%) | 260 (31.0%) | <0.001 |
| History of atrial fibrillation/flutter | 2604 (37.5%) | 233 (37.6%) | 642 (36.5%) | 738 (35.4%) | 531 (38.0%) | 460 (42.4%) | 0.003 |
| History of ICD/CRT | 615 (8.9%) | 25 (4.0%) | 90 (5.1%) | 160 (7.7%) | 161 (11.5%) | 179 (16.5%) | <0.001 |
| Laboratory values | |||||||
| Baseline creatinine, mg/dL | 1.2 (1.0–1.6) | 1.2 (1.0–1.4) | 1.2 (1.0–1.5) | 1.2 (1.0–1.5) | 1.3 (1.0–1.7) | 1.4 (1.1–1.9) | <0.001 |
| Baseline NT-proBNP, pg/mL | 4501 (2100–9200) | 4038 (1965–8355) | 4334 (2018–8174) | 4496 (2006–9640) | 5059 (2451–10154) | 4823 (2247–10151) | 0.001 |
| Baseline BNP, pg/mL | 991 (543–1873) | 1030 (585–1892) | 939 (442–1776) | 1000 (561–1990) | 967 (537–1649) | 1030 (627–1903) | 0.033 |
| Medication at baseline | |||||||
| ACEi or ARB | 4211 (60.6%) | 344 (55.6%) | 1002 (57.0%) | 1254 (60.2%) | 911 (65.2%) | 700 (64.6%) | <0.001 |
| Beta-blockers | 4037 (58.1%) | 301 (48.6%) | 867 (49.3%) | 1188 (57.0%) | 899 (64.3%) | 782 (72.1%) | <0.001 |
| Aldosterone antagonists | 1934 (27.8%) | 195 (31.5%) | 518 (29.4%) | 578 (27.7%) | 389 (27.8%) | 254 (23.4%) | 0.002 |
ACEi, angiotensin-converting enzyme inhibitor; ARB, angiotensin receptor blocker; BNP, B-type natriuretic peptide; CRT, cardiac resynchronization therapy; HF, heart failure; ICD, implantable cardioverter-defibrillator; LVEF, left ventricular ejection fraction; NT-proBNP, N-terminal pro-B-type natriuretic peptide.
Association between the burden of non-cardiac comorbidities and 30-day all-cause death/heart failure hospitalization
Using hierarchical logistic regression, patients with four or more non-cardiac comorbidities were associated with higher rates of 30-day all-cause death/HF hospitalization as compared with patients with no non-cardiac comorbidities in both unadjusted and adjusted analyses [adjusted odds ratio (OR) 3.32, 95% confidence interval (CI) 1.61–6.84; P < 0.01]. Similarly, there was an association with the presence of two or three non-cardiac comorbidities with the primary outcomes of 30-day all-cause death/HF hospitalization when compared with no non-cardiac comorbidities (adjusted OR 2.45, 95% CI 1.21–4.94; P = 0.01 for two vs. no comorbidities; adjusted OR 2.81, 95% CI 1.37–5.78; P < 0.01 for three vs. no comorbidities) (Figure 2). There were no significant differences between one vs. no non-cardiac comorbidities in adjusted analysis.
Figure 2.
Association between non-cardiac comorbidity burden and 30-day all-cause death/heart failure (ACD/HF) hospitalization and death outcomes. This figure shows the adjusted odds-ratio for the association between non-cardiac comorbidity burden and the co-primary outcomes and 30-day death outcomes. CI, confidence interval; CV, cardiovascular; HR, hazard ratio; OR, odds ratio.
Association between the burden of non-cardiac comorbidities and death rates
There were no significant differences with respect to the outcome of 30-day all-cause death when comparing patients with one, two, three, or four or more non-cardiac comorbidities to patients with no non-cardiac comorbidities in adjusted analyses. Similarly, rates of 30-day CV death were similar across all groups (Figure 2). With respect to 180-day death, patients with four of more non-cardiac comorbidities had significantly higher rates of death as compared with patients with no non-cardiac comorbidities (adjusted OR 2.13, 95% CI 1.33–3.43; P <0.01) (Figure 3). Patients with two or three non-cardiac comorbidities had higher event rates of 180-day death, though non-significant after adjustment, as compared with those with no non-cardiac comorbidities.
Figure 3.
Association between non-cardiac comorbidity burden and 180-day all-cause death. This figure shows the association of comorbidity burden (0, 1, 2, 3, 4+) and the co-primary endpoint of 180-day all-cause death (P < 0.01).
Association between individual non-cardiac comorbidities and primary outcomes
Given the awareness that each non-cardiac comorbidity may differ in its impact on clinical outcomes, we engaged in a univariate model to assess the association of each non-cardiac comorbidity individually with the primary outcome measures. With respect to the outcomes of 30-day all-cause death/HF hospitalization, renal disease had the strongest association with increased risk (unadjusted OR 2.60, 95% CI 1.75–3.87), followed by history of depression (unadjusted OR 2.08, 95% CI 1.59–2.74) (Table 3). With respect to 180-day death outcomes, renal disease had the strongest association with increased events (unadjusted OR 2.39, 95% CI 1.61–3.56), followed by history of cerebrovascular disease (unadjusted OR 1.73, 95% CI 1.36–2.21). History of liver disease, recent history of cancer, and history of alcohol abuse were not independently associated with either primary outcome in univariate analysis.
Table 3.
Univariate analysis of individual non-cardiac comorbidities and association with primary outcomes
| Non-cardiac comorbidity | OR (95% CI) |
|---|---|
| 30-day all-cause death/HF rehospitalization | |
| History of diabetes | 1.34 (1.11 −1.61) |
| History of COPD | 1.54 (1.23–1.92) |
| Chronic liver disease | 1.01 (0.57–1.80) |
| History cancer (within last 5 years) | 1.07 (0.67–1.71) |
| Renal disease | 2.60 (1.75–3.87) |
| Current smoking | 1.39 (1.16–1.68) |
| History of alcohol abuse | 1.16 (0.85–1.57) |
| Anaemia (haemoglobin <13.5 g/L) | 1.71 (1.39–2.09) |
| Peripheral arterial disease | 1.42 (1.09–1.86) |
| Cerebrovascular disease | 1.80 (1.41 −2.29) |
| History of depression | 2.08 (1.59–2.74) |
| 180-day all-cause death | |
| History of diabetes | 1.37 (1.13–1.65) |
| History of COPD | 1.42 (1.13–1.79) |
| Chronic liver disease | 0.96 (0.53–1.74) |
| History of cancer (within last 5 years) | 1.43 (0.94–2.17) |
| Renal disease | 2.39 (1.61 −3.56) |
| Current smoking | 1.00 (0.83 −1.20) |
| History of alcohol abuse | 0.97 (0.70–1.35) |
| Anaemia (haemoglobin <13.5 g/L) | 1.65 (1.34–2.03) |
| Peripheral arterial disease | 1.71 (1.32–2.21) |
| Cerebrovascular disease | 1.73 (1.36–2.21) |
| History of depression | 1.64 (1.22–2.19) |
CI, confidence interval; COPD, chronic obstructive pulmonary disease; HF, heart failure; OR, odds ratio.
Association between the burden of non-cardiac comorbidities and rehospitalization
Patients with higher burden of non-cardiac comorbidities were associated with higher rates of 30-day all-cause hospitalization as compared with no non-cardiac comorbidities. Higher comorbidity burden was associated with all-cause hospitalization (adjusted OR 3.54, 95% CI 2.08–6.03; P <0.01 for 4+ vs. none; adjusted OR 2.95, 95% CI 1.74–4.99; P < 0.01 for three vs. none). Similarly, higher comorbidity burden was associated with greater number of 30-day CV and HF hospitalization as compared with no non-cardiac comorbidity burden in adjusted analyses. There were no significant differences between HF hospitalization in patients with one vs. no non-cardiac comorbidities, though this approached significance in adjusted analysis (Figure 4).
Figure 4.
Association between non-cardiac comorbidity burden and 30-day rehospitalization outcomes. CI, confidence interval; CV, cardiovascular; HF, heart failure; OR, odds ratio.
Association between the burden of non-cardiac comorbidities and cumulative costs
Mean ± standard deviation cumulative costs were $16 470 ± 190 at 30 days and $25 462 ± 360 at 180 days. Using an ordinary least squares means estimator, patients with four or more comorbidities had significantly higher cumulative costs at 180 days (mean difference $7892, 95% CI $5042–10 742; P < 0.01) as compared with patients with no non-cardiac comorbidities. Patients with two or three non-cardiac comorbidities also had higher 180-day costs as compared with patients with no non-cardiac comorbidities (Figure 5).
Figure 5.
Association between non-cardiac comorbidity burden and total 180-day costs.
Association between the burden of non-cardiac comorbidities and hospital length of stay
Overall, mean ± standard deviation hospital LOS was 8 ± 0.1 days. Using least square means estimates, patients with higher comorbidity burden had significantly longer index hospital LOS than patients with fewer comorbidities. Patients with four or more comorbidities had a mean LOS of 9.6 days (95% CI 8.5–10.8) as compared with patients with no comorbidities who had a mean LOS of 8.1 days (95% CI 7.2–9.2; P < 0.01).
Discussion
Overall, in patients hospitalized with acute HF in a large randomized clinical trial, non-cardiac comorbidities were highly prevalent. Older patients and those with HFpEF generally had higher comorbidity burden as compared with patients with HFrEF. Greater comorbidity burden was associated with higher rates of 30-day all-cause death/HF hospitalization, though rates of 30-day and 180-day death alone were similar across comorbidity burden. Higher comorbidity burden was associated with higher rates of hospitalization, irrespective of cause. Finally, higher comorbidity burden was associated with greater cumulative costs and longer hospital LOS.
We found differences in baseline characteristics between patients with high vs. low comorbidity burden. Expectedly, older male patients with previous hospitalizations for HF had higher rates of comorbidity. Patients with high comorbidity burden were overwhelming white or African American and enrolled from North America. For example, while North American patients accounted for 45% of the overall study population, they accounted for 76% of patients with four of more non-cardiac comorbidities. In contrast, patients enrolled from Asia-Pacific accounted for 25% of the overall population, but only 7% of patients with four or more comorbidities. Reasons for these findings are likely multifactorial, and may, in part, be representative of reporting practices and more aggressive screening for comorbid conditions. They may also reflect differences in diet intake, body mass index, stress levels, and overall lifestyle. Our finding compliment Centers for Disease Control and Prevention compiled data which show that over 50% of Medicare beneficiaries coded with HF have five or more comorbidities.14 In contrast, data from the European HF pilot survey found that only 36% patients had more than one non-cardiac comorbidity.8 Furthermore, the growth in HFpEF, a disease more readily seen in the elderly and multi-morbidity population, continues to rise significantly in the Western world.
We also found that patients with HFpEF had higher comorbidity profiles than those with HFrEF. Similar findings with regards to COPD, anaemia, diabetes, and renal disease have been shown in a number of major US based HF registries.15 In contrast, an analysis from the European HF pilot survey found comorbidity burden to be similar between groups, though some of the comorbidities studied differed from those in our analysis.8 In the absence of guideline-directed medical therapies for patients with HFpEF, comorbidity management and prevention have become a mainstay of therapy in this difficult to treat population. Specialized lifestyle modification clinics/programmes may be beneficial in lower comorbidity burden, hospitalization and death, though these have been incompletely studied.16
Our analysis found a strong relationship between higher comorbidity burden and the risk of all-cause death/HF hospitalization at 30 days. Interestingly, this effect appeared to be driven by an increase in HF hospitalization, as no significant differences were observed between groups with respect to 30-day all-cause death. Similar associations were seen in a large cross-sectional analysis of Medicare beneficiaries.5 At 180 days, however, higher comorbidity burden was associated with greater all-cause death. One potential explanation of these findings are that comorbidities alter the natural history of HF, promoting hospitalizations, each of which increases the risk of overall mortality in HF patients. The cumulative effect of increased hospitalizations on mortality is expressed when a longer outlook is undertaken with respect to the outcome of death. Comorbidity prevention, through lifestyle modifications, weight management, tobacco/alcohol cessation, therefore, may prevent avoidable hospitalization and create a more favourable long-term prognosis for patients with HF.
To the best of our knowledge, this is the first analysis to assess the association between overall comorbidity burden, cumulative costs of care, and initial hospital LOS in a large population of patients hospitalized for acute HF. Our findings suggest that higher comorbidity burden creates important increases in resource utilization and may alter the complexity and cost of the patient’s hospitalization for HF. These issues are expected to worsen with the aging of the population.17 These findings are particularly important not only to hospital administrators and population health experts, but also to programmes aimed at prevention, lifestyle modification, and comorbidity management. Investments in these programmes may have favourable return on investment by reducing costs of hospitalization for a disease state that is estimated to cost $69.7 billion annually by 2030, though these programmes should be empirically evaluated for cost-effectiveness.1
Previous analyses have found that LOS is correlated linearly with all-cause and cause-specific readmission once index LOS exceeds 6 days.18 Similarly, prolonged LOS is associated with increases in mortality. In addition, data from the OPTIMZE-HF registry found COPD was independently associated with longer hospital LOS.19 Our data find that mean LOS in the trial population was >8 days, with consistent association between increased comorbidity burden and longer LOS. These findings suggest that proactive management and prevention of comorbidities in the HF patient may be associated with reduced hospitalizations and shorter LOS when hospitalization does occur.
Our findings promote particular interest in the development of programmes that not only monitor HF metrics in the outpatient setting, but also ones that are dedicated to comorbidity prevention or reduction. There are limited data with respect to development and evaluation of these programmes, but rigorous analysis and potential integration of preventative strategies into existing HF disease management programmes should be further studied. The development of such programmes may alter the natural history of HF hospitalizations and long-term mortality, and potentially could also be fiscally favourable for health systems.
Limitations
There are several limitations that should be acknowledged. First, this analysis was post-hoc and is subject to the biases of exploratory analyses. Second, we defined a set of non-cardiac comorbidities based on data readily obtained from medical history collected in the ASCEND-HF trial and based on authors’ consensus. While the selected comorbidities are not exhaustive, they do reflect common and prevalent conditions, many of which are modifiable in nature. Third, the authors realize that within each comorbidity, there is a spectrum of disease (e.g. mild vs. very severe COPD). Unfortunately, severity of comorbidities was not collected; thus, we are not able to account for the relative severity of the comorbidities on the outcomes we examined. Fourth, the authors acknowledge that there may be differential weighting between comorbidities with respect to their association with outcomes that was apparent in the univariate associations we reported. There were incomplete data in this analysis to use previously validated indices of comorbidity such as the Charlson comorbidity index, which is similarly seen in other analyses from randomized trials assessing comorbidity. Fifth, given the trial protocol, long-term (180-day) cause-specific death and rehospitalization outcomes were not specifically collected or adjudicated, limiting our assessment of long-term outcomes. Sixth, substantial missing data with respect to numerical ejection fraction allowed only for valid comparisons between those with LVEF < 40% and LVEF ≥40%. This may have potentially mischaracterized those patients with mid-range ejection fraction in the preserved ejection fraction category. Seventh, a clinical trial population may vary in the number and extent of comorbidity as compared with a ‘real-world’ population, and therefore our findings on the prevalence of comorbidity may be overestimated or underestimated due to bias in trial enrolment. Finally, the cost analysis included randomized treatment cost, though rates of treatment allocation were similar between groups.
Conclusion
Non-cardiac comorbidities are highly prevalent in patients hospitalized with HF. Higher comorbidity burden was strongly associated with higher rates of long-term all-cause death and short-term all-cause hospitalization. Higher comorbidity burden was also associated with extended hospital LOS and total cumulative costs. These data suggest that comorbidity management/prevention programmes may be important interventions for HF patients. Further prospective studies are needed to assess whether these programmes may be cost-effective and improve outcomes for patients with HF.
Funding
Scios Inc. provided financial and material support for the ASCEND-HF trial. Database management and statistical analysis were performed by the Duke Clinical Research Institute.
Conflict of interest: A.P.A. was supported by National Heart, Lung, and Blood Institute (NHLBI) T32 post-doctoral training grant 5T32HL069749. A.D.D. has received research funding from the American Heart Association, Amgen, and Novartis; and is a compensated consultant for Novartis. J.B. has received research support from U.S. National Institutes of Health (NIH), European Union, and Patient Centered Outcomes Research Institute; and is a compensated consultant for Amgen, AstraZeneca, Bayer, Boehringer Ingelheim, Bristol-Myers Squibb, CardioCell, Janssen, Novartis, Relypsa, ZS Pharma, Medtronic, Merck, and CVRx. A.V. is a compensated consultant for and has received research grants from Alere, Amgen, Bayer, Boehringer Ingelheim, Cardio3BioSciences, Celladon, GlaxoSmithKline, Merck Sharp and Dohme, Novartis, Servier, Singulex, Sphingotec, Stealth Peptides, Trevena, Vifor, and ZS Pharma. R.S. is a compensated consultant for Novartis, BioControl, and Medtronic; and is owner/partner in CardioMEMS; has received research support from U.S. NIH, Medtronic, Biotronik, Novartis, and Thoratec; and has received honoraria from American Board of Internal Medicine. J.A.E. is a compensated consultant for Pfizer, Abbott Laboratories, and Servier; and has received research support from Amgen and Johnson & Johnson. M.M. is a compensated consultant for Bayer, Novartis, and Servier. A.F.H. is a compensated consultant for Sanofi, Johnson & Johnson, AstraZeneca, and Corthera; and has received research support from Amylin and Scios/Johnson & Johnson. C.O’C. is a compensated consultant for Novella and Amgen; is owner/partner in Biscardia, LLC; and has received research support from Otsuka, Roche Diagnostics, BG Medicine, Critical Diagnostics, Astellas, Gilead, GE Healthcare, and ResMed. R.J.M. has received research support from Amgen, AstraZeneca, Bristol-Myers Squibb, GlaxoSmithKline, Gilead, Novartis, Otsuka, and ResMed; and has received honoraria from Thoratec. All other authors have nothing to disclose.
References
- 1.Benjamin EJ, Virani SS, Callaway CW, Chamberlain AM, Chang AR, Cheng S, Chiuve SE, Cushman M, Delling FN, Deo R, de Ferranti SD, Ferguson JF, Fornage M, Gillespie C, Isasi CR, Jiménez MC, Jordan LC, Judd SE, Lackland D, Lichtman JH, Lisabeth L, Liu S, Longenecker CT, Lutsey PL, Mackey JS, Matchar DB, Matsushita K, Mussolino ME, Nasir K, O’Flaherty M, Palaniappan LP, Pandey A, Pandey DK, Reeves MJ, Ritchey MD, Rodriguez CJ, Roth GA, Rosamond WD, Sampson UKA, Satou GM, Shah SH, Spartano NL, Tirschwell DL, Tsao CW, Voeks JH, Willey JZ, Wilkins JT, Wu JH, Alger HM, Wong SS, Muntner P; American Heart Association Council on Epidemiology and Prevention Statistics Committee and Stroke Statistics Subcommittee. Heart disease and stroke statistics – 2018 update: a report from the American Heart Association. Circulation 2018;137:e67–e492. [DOI] [PubMed] [Google Scholar]
- 2.Bui AL, Horwich TB, Fonarow GC. Epidemiology and risk profile of heart failure. Nat Rev Cardiol 2011;8:30–41. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Lang CC, Mancini DM. Non-cardiac comorbidities in chronic heart failure. Heart 2007;93:665–671. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Brown AM, Cleland JG. Influence of concomitant disease on patterns of hospitalization in patients with heart failure discharged from Scottish hospitals in 1995. Eur Heart J 1998;19:1063–1069. [PubMed] [Google Scholar]
- 5.Braunstein JB, Anderson GF, Gerstenblith G, Weller W, Niefeld M, Herbert R, Wu AW. Noncardiac comorbidity increases preventable hospitalizations and mortality among Medicare beneficiaries with chronic heart failure. J Am Coll Cardiol 2003;42:1226–1233. [DOI] [PubMed] [Google Scholar]
- 6.Groenveld HF, Januzzi JL, Damman K, van Wijngaarden J, Hillege HL, van Veldhuisen DJ, van der Meer P. Anemia and mortality in heart failure patients: a systematic review and meta-analysis. J Am Coll Cardiol 2008;52:818–827. [DOI] [PubMed] [Google Scholar]
- 7.Smith DH, Thorp ML, Gurwitz JH, McManus DD, Goldberg RJ, Allen LA, Hsu G, Sung SH, Magid DJ, Go AS. Chronic kidney disease and outcomes in heart failure with preserved versus reduced ejection fraction: the Cardiovascular Research Network PRESERVE study. Circ Cardiovasc Qual Outcomes 2013;6:333–342. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.van Deursen VM, Urso R, Laroche C, Damman K, Dahlström U, Tavazzi L, Maggioni AP, Voors AA. Co-morbidities in patients with heart failure: an analysis of the European Heart Failure Pilot survey. Eur J Heart Fail 2014;16: 103–111. [DOI] [PubMed] [Google Scholar]
- 9.Streng KW, Nauta JF, Hillege HL, Anker SD, Cleland JG, Dickstein K, Filippatos G, Lang CC, Metra M, Ng LL, Ponikowski P, Samani NJ, van Veldhuisen DJ, Zwinderman AH, Zannad F, Damman K, van der Meer P, Voors AA. Non-cardiac comorbidities in heart failure with reduced, mid-range and preserved ejection fraction. Int J Cardiol 2018;271:132–139. [DOI] [PubMed] [Google Scholar]
- 10.O’Connor CM, Starling RC, Hernandez AF, Armstrong PW, Dickstein K, Hasselblad V, Heizer GM, Komajda M, Massie BM, McMurray JJ, Nieminen MS, Reist CJ, Rouleau JL, Swedberg K, Adams KF Jr, Anker SD, Atar D, Battler A, Botero R, Bohidar NR, Butler J, Clausell N, Corbalán R, Costanzo MR, Dahlstrom U, Deckelbaum LI, Diaz R, Dunlap ME, Ezekowitz JA, Feldman D, Felker GM, Fonarow GC, Gennevois D, Gottlieb SS, Hill JA, Hollander JE, Howlett JG, Hudson MP, Kociol RD, Krum H, Laucevicius A, Levy WC, Méndez GF, Metra M, Mittal S, Oh BH, Pereira NL, Ponikowski P, Tang WH, Tanomsup S, Teerlink JR, Triposkiadis F, Troughton RW, Voors AA, Whellan DJ, Zannad F, Califf RM. Effect of nesiritide in patients with acute decompensated heart failure. N Engl J Med 2011;365:32–43. [DOI] [PubMed] [Google Scholar]
- 11.Hernandez AF, O’Connor CM, Starling RC, Reist CJ, Armstrong PW, Dickstein K, Lorenz TJ, Gibler WB, Hasselblad V, Komajda M, Massie B, McMurray JJV, Nieminen M, Rouleau JL, Swedberg K, Califf RM. Rationale and design of the Acute Study of Clinical Effectiveness of Nesiritide in Decompensated Heart Failure trial (ASCEND-HF). Am Heart J 2009;157:271–277. [DOI] [PubMed] [Google Scholar]
- 12.Khazanie P, Heizer GM, Hasselblad V, Armstrong PW, Califf RM, Ezekowitz J, Dickstein K, Levy WC, McMurray JJV, Metra M, Wilson Tang WH, Teerlink JR, Voors AA, O’Connor CM, Hernandez AF, Starling R. Predictors of clinical outcomes in acute decompensated heart failure: Acute Study of Clinical Effectiveness of Nesiritide in Decompensated Heart Failure outcome models. Am Heart J 2015;170:290–297. [DOI] [PubMed] [Google Scholar]
- 13.Reed SD, Kaul P, Li Y, Eapen ZJ, Davidson-Ray L, Schulman KA, Massie BM, Armstrong PW, Starling RC, O’Connor CM, Hernandez AF, Califf RM. Medical resource use, costs, and quality of life in patients with acute decompensated heart failure: findings from ASCEND-HF. J Card Fail 2013;19:611–620. [DOI] [PubMed] [Google Scholar]
- 14.Centers for Disease Control and Prevention, National Center for Health Statistics. Underlying Cause of Death, 1999–2017. 2019. https://wonder.cdc.gov/ ucd-icd10.html (7 January 2019). [PubMed] [Google Scholar]
- 15.Mentz RJ, Kelly JP, von Lueder TG, Voors AA, Lam CS, Cowie MR, Kjeldsen K, Jankowska EA, Atar D, Butler J, Fiuzat M, Zannad F, Pitt B, O’Connor CM. Noncardiac comorbidities in heart failure with reduced versus preserved ejection fraction. J Am Coll Cardiol 2014;64:2281–2293. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Aggarwal M, Bozkurt B, Panjrath G, Aggarwal B, Ostfeld RJ, Barnard ND, Gaggin H, Freeman AM, Allen K, Madan S, Massera D, Litwin SE; American College of Cardiology’s Nutrition and Lifestyle Committee of the Prevention of Cardiovascular Disease Council. Lifestyle modifications for preventing and treating heart failure. J Am Coll Cardiol 2018;72:2391–2405. [DOI] [PubMed] [Google Scholar]
- 17.Sharma A, Zhao X, Hammill BG, Hernandez AF, Fonarow GC, Felker GM, Yancy CW, Heidenreich PA, Ezekowitz JA, DeVore AD. Trends in noncardiovascular comorbidities among patients hospitalized for heart failure: insights from the Get With the Guidelines-Heart Failure Registry. Circ Heart Fail 2018;11: e004646. [DOI] [PubMed] [Google Scholar]
- 18.Sud M, Yu B, Wijeysundera HC, Austin PC, Ko DT, Braga J, Cram P, Spertus JA, Domanski M, Lee DS. Associations between short or long length of stay and 30-day readmission and mortality in hospitalized patients with heart failure. JACC Heart Fail 2017;5:578–588. [DOI] [PubMed] [Google Scholar]
- 19.Mentz RJ, Fiuzat M, Wojdyla DM, Chiswell K, Gheorghiade M, Fonarow GC, O’Connor CM. Clinical characteristics and outcomes of hospitalized heart failure patients with systolic dysfunction and chronic obstructive pulmonary disease: findings from OPTIMIZE-HF. Eur J Heart Fail 2012;14:395–403. [DOI] [PubMed] [Google Scholar]